SOLUTIONS MANUAL NONLINEAR SYSTEMS THIRD EDITION HASSAN K. KHALILAcquisitions Editor: Dorothy Marrero Associate Editor: Alice Dworkin Supplement Editor: Jessica Romeo Executive Managing Editor: Vince O'B: Managing Editor: David A. George Production Editor: Barbara A. Till Supplement Cover Management/Design: Paul Gourhan Manufacturing Buyer: Tene Kahn © 2002, 1996, 1992 _ by Pearson Education, Ine. Pearson Education, Inc. Upper Saddle River, NJ 07458 All rights reserved. No part of this book may be reproduced in any form or by any means, without permission in writing from the publisher. The author and publisher of this book have used their best efforts in preparing this book. These efforts include the development, researc {esting of the theories and programs to determine their effectiveness. The author and publisher make no warranty of any kind, expressed or implied, with regard (0 these programs or the documentation contained in this, book, The author and publisher shall not be liable in any event for incidental or consequential damages in connection with, or arising out of, the furnishing, performance, or use of these programs. Printed in the United States of America 0987654321 ISBN O-13-0b7530-x Pearson Education Ltd., London Pearson Education Australia Py. Lid., Sydney Pearson Education Singapore, Pte. Ltd Pearson Education North Asia Lid., Hong Kong Pearson Education Canada, Ine., Toronto Pearson Educacfon de Mexico, $.A. de C.V. Pearson Education—Japan, Tokyo Pearson Education Malaysia, Pte. Lid, Pearson Education, Upper Saddle River, New JerseyContents 10 cu 12 13 14 Introduction Second-Order Systems Fundamental Properties Lyapunov Stability Input-Output Stability Passivity Frequency Domain Analysis of Feedback Systems Advanced Stability Analysis, Stability of Perturbed Systems Perturbation Theory and Averaging Singular Perturbations Feedback Control Feedback Linearization Nonlinear Design Tools i 45, 59 o1 101 107 127 139 151 161 179 189 207Chapter 1 od Let ay = 9,22 = 9, oy aq = yO) =m Sn = ty fn = alt tiy.-- Ent) 5 yen O12 Let =y, a2 =y), A =m fn = tat faa = YOY Say galt tr 22500 tea) fn = -aalttiyyznadi— (22 rm =!) — galt, 15>, Enaa) (& 00 = Galt iy...) Tn—1,7n + g2-Ju,u) aon , 8 aon ~ (Bee Sete tt 2 ton tv) w noe 918 Let =y, 22 = 9, an Sng mbm tnt ste = YO), ng md, os Bats =m G(B1y-+-5 Tay Zmtay-+-yZa¢mst) = tage atm2 CHAPTER 1 s14 Len =an=h [fer A= om i = G = M7(a)[u-Cler,22)20 ~ Der ~ of) 91.5 Let ay = gn, 2 = Gis ts = 24 = a =m _ _ Mak k i = J sin, — F(21 ~ 23) fs = k 1 B= Gla—as) +50 © 1.6 Let 21 = a1, 22 = thy 22 = 92, 24 = de, where 2; € R™, a —M~'(2,)[h(zi22) + Kes ~ 23)] iy =u f= IK 25) 4 Au ©L.7 Let Art Buy be a state model of the linear system. u=r- (ty) =r—¥(t,Cz) Hence iz — BY(t,0z)+Br, y=Cz © 1.8 (a) Let 21 = 4, 22 = 6, 15 = Ey, and u= Epp. f= b= (b) The equilibrium points are the roots of the equations O=m 0 = 0815 —Drz ~2.0z5sinz, 0 = 2725 +170sz +122 2 =0> 23 oe sing Substituting 25 in the third equation yields (1.22 + 1.7 cosa) sin 21 ~ 1.10025 = 0 ‘The foregoing equation has two roots 2; = 0.4067 and z; = 1.6398 in the interval —x < 21 < x. Due to periodicity, 0.4067 + 2nz and 1.6398 + 2nr are also roots for n = £1,+2,.... Each root 21 = 2 gives anequilibrium point (z,0,0.4075/ sinz). (c) With B, = constant, the model reduces to PD, om Eysin2, he Mt which isa pendulum equation with an input torque. # 1.9 (a) Let 21 = $1, 22 =00. in (b) Let 25 = in, 22 = ve a ‘The model of (a) is more familiar since it is the pendulum equation. 1.20 (a) Let 21 = $2, 22 = 00. an #3 finte notte (b) 22 =0 + Ley +2} =0 + 21 =0. Thereis a unique equilibrium point at the origin. ©1111 (a) #=Ar+Buy y=Cx, u=sine €=6,-6, = -b. z=AztBsine, €=-Cz () 0=Az+Bsine > = Cz + -CA"Bsine G(0) 40 = sine (c) For G(s) = 1/(r8 +1), take A= -1/r, B= 1/7 and tne, n 1 et isine, é=—2 Let 21 =e, 22 = 11s smn psinz ham,4 (CHAPTER 1. 1.12 The equation of motion is Mi = Mg -ky- cay — could] Let 2) =y and 29 =f © 1.13 (a) mij = ~ (ki + ka)y ~ oh + h(v0 ~ 9) where ¢ > 0 is the viscous friction coefficient. (b) h(v) & A(wvo) ~ (wo). mij = (Ri + ka)y — [e+ h'(vo)]i + h(e0) (c) To obtain negative friction, we want c+ h'(up) <0. This can be achieved with the friction characteristic of Figure 1.5(d) if vo is in the range where the slope is negative and the magnitude of the negative slope is greater than c. 1.14 The equation of motion is Mo = F ~ Mgsin@ ~ kysgn(v) ~ kav — kav? ‘where ky, ka, and ky are positive constants. Let 2 =v, u=F, and w = sind. 1 2 = 7 [hisen(@) - haz - bye? +4] - gw © 1.15 (a) ¢ ‘ H = mylu+ Lsin8) = m5 (i + Li cos®) = (i+ Licos0 — 16? sind) mB sin 9 — mL6 cos8 + mg e nee V = mip(Leos0) +mg = m5 (—Lésind) +mg = Substituting V and # in the é-equation yields Té = VLsind - HLcosé = —mL*6(sin8)? — mL?6? sind cos + mgLsin® =m cos — mL*0(cos0)® + mL76? sin 8 cos = —m1?6{(sin 8)? + (cos6)?] + mgLsin# — mLijcos@ = —mL36 + mgLsind ~ mLjcos® Substituting H in the j-equation yields Mi = F — mj + Lbcos0 — 6? sin0) — hy ) ; , é mgL sin Pe 5 ]=[ es mubsne—ry | where I+mL? ml:cos _[f+mL? mbcos P= | beme mre det(D(6)) = (I + mL2)(m-+ M) ~ PL? cos? 8 = A(0)Hence, D“@) = m+M | By -mucwe T'smet 6)_ 1] m+M_ -mbcose mgLsin8 G | BO | -mbeose r+mz? || F+mLésind— Ky © : : s-2 xe [m+ Atymgtsin# ~ mz coso(F + mLé sind — by)] = App lm Mimgbsin 2, ~ mbcosy(u + mLz}sinzy ~ kr] : : r Bigs [em "Hain e080 + (1+ m22)(F + mL sind — bi) = Ape Lm ePasin gs cosas +(F + ml2)(u+ mz} sine, ~ hea)] 2116 Re miles +Lsind) = molec + Léc0s6) = m(, + Léicos0 — L6? sind) R= ne 7a (Lcos@) = m$(-Lisin 8) = ~mLasind ~ mLé* cos poe er er wb ut FyLsind ~ F,Lcos0 = u-mL?6(sin6)' — mL76? sin8 cos8 —mLi. cos — mL6(cos8)? + mL6? sin8 cos® = u—mL6{(sind? + (cos0)?] - mL. cos = u-m1%6 — mL, cos Substituting F, in the Z.-equation yields Mi, = —mé, — mlb cos6 + mL6" sin8 — kx. ‘Thus, . 6 u P| 2 ]=[ nzétsne te. | where I+mL? mL cosé +mL? mbcos PO =| micos@ m+M | (b) det (D(0)) = (I + mL*)(m-+ M) ~ mi? cos? 6 = (0) Hence, D@)= m+M Feel aa mL cos@ I+ mL?6 CHAPTER 1. [2] =aty[ tie TES] [ ncitate ae | (e) i” : i gy [om + Mw ~ miscos mt sind ~ kx.) = ae {(mm + M)u~ mbeozy mba} sine, — kes)] dy = 1 a6 [mu cosy + (I +mL?)(mba3sin zy ~ kzs)] te = % mLucosd + (I + mL?)(mLé? sind ~ kze)] 1 A(z) E (d) Take w= constant. Setting the derivatives 2; = 0, we obtain 22 = 24 =0 and 0 0 (m+ Myu+ mkLzs.cosz, —mLucosz, ~ K(I + mL?)23 Eliminating z3 between the two equations yields ul(m + M)(I + mL) — m?L?cos? 23] =u A(z) = 0 Since A(z1) > 0, equilibrium can be maintained only a: u = 0. Then, x = 0. Thus, the system has an ‘equilibrium set {22 = 25 = 74 = 0} © LIT (a) Let 21 = iy, 22 = ig, and ty =u. constant and vy constant and v, =u. A constant field voltage implies that (at steady state) iy = V;/Ry “ 1, = constant. Hence, the model reduces to the second-oder linear model Re, aly, | a= -By- Mayet é 3, oly ty = -8m+ he, Tet FT (d) Let v +R a= ~ Bt - fe, u a= - Ban Bows t yy fy = — Sat Snare#148 (a) 21 =, 22-9, 29 = 4, UH. “ k. 1 a= -Spso+ tru kL 1 Lost min Balh+ a1 /ae Loar} Imla +x)? di tS] k = 9- im = a ~ at De). oe _ 6 db, To)” PO) w Ri) +f Diy artarar 1 To" oaza ts Tay et eae) (&) The equilibrium equations are O= = Lent £,,- lok - ine +i 0 = -Re, + oats wrap te Set # =r, £3 = Iss, and % = Vyy. Then 2mg(a+r)? we | Yaa = Rl ta) In = 1.19 (a) $( “aaa = w-k/pgh ah AQh = u-kVagh a= ae hype, y= (b) 2 = p— pa = pgh. = apy i V2 v= 2/0) (6) At equilibrium, O= Ue kV Pie, Yer = te =P Hence, tis, = kVPIF8 (CHAPTER 1. + 1.20 (a) From the equations = us ~ we and p= pe + (ou/A, we have a; 09, = M6-1(4p)— Bhs = PE, — a.) = Mlo-*(A) ~ kV BA Using 2 = Ap as the state vatiable, we obtain 28 15-1¢e3 b= Flee) kva] (b) At equilibrium we have OG) = ive Writing 2 = ¢(w;), we can rewrite the previous equation as Ne ) Hence, 0, E) =) ‘The solutions of this equation are given by the intersecticn of the curve (t:/k)? with the curve $(i), which is shown in Figure 1.29 of the text. From the figure, it is clear that there is only one intersection point. © 1.21 (a) We have tr =up—w, aw — we = 9, =, Aa Bin b= Ze wi =hivpi— Pa, w2=haVP— Pa, Pi ~ Pa = Ow) Let 21 = pi ~ pa and 22 = Pa ~ Po. a= Blea) - hv 7] i = 2 (hve - hve] (b) The equilibrium equations are uVa-B baVia ‘From the second equation, we have ‘Substituting this expression in the first equilibrium equation yields Oi) = keqVFi, where keg = ‘Writing #1 = (tp), we can rewrite the previous equation as y= kaa Olas) 5)? ae)” = () -#n ‘The solutions of this equation are given by the intersection of the curve (tp/keq) with the curve $(ip), which is shown in Figure 1.29 of the text. From the figure, it is clear that there is only one intersection point. Hence,+1.22 (a) fy = dmy-dn tn iy = dy -dm—m ‘The assumptions ry = yxy, ra =1/Y = wany/Y, and 247 = 0 yield th = udm 2. = d(tay — 22) ya /Y (b) When 1 = pn z2/(m +22), the equilibrium equations are tints _ 4) 5 o = (fet aa in 0 = dear 23) ~ Fie aa) from the fist equation, . hed yao or it 0 oe re Tim 4 Substituting #1 = O in the second equilibrium equation yields #2 = 22y. Substituting 22 = kmd/ (im —d) in the second equilibrium equation yields . ‘ nav (ay-ie Hence, there are two equilibrium points at (c) When p1 = j4m22/(Km + 22 + ky23), the equilibrium equations are - nia °= (geetna-4)* in 0 = ales -2) fe "Viet +h from the first equation, £, = 0 or 22 is the root of d = (2). Since d < maxz,>0{u(z2)}, the equation d= p(@2) has two roots. Denote these roots by #2, and 25. Substituting 2 = 0 in the second equilibrium equation yields #2 = ray. Substituting 22 = 2. in the second equilibrium equation yields (224 — E20) — H(F2a)21/¥ =0 => 41 =¥ (aay — Fra) since (2a) = d. Similarly, substituting #2 = zap in the second equilibrium equation yields 21 = ¥ (zay — a9). Thus, there are three equilibrium points at (Y (ay — 10)» Fa), (Y (zag — Fz), 220), and (0,224)Chapter 2 © 2.1 (1) O=-n 42x42, O=-z-2 2) 3 O=2n(ef-1) + m1 =0,1or-1 ‘There are three equilibrium points at (0,0), (1,—1), and (—1,1). Determine the type of each point using Tincarzaton, af _[-1+6? 1 =| a ] > hg=—145 = (0,0)is a stable focus Bln | £V8 + (1-1) isa saddle Similarly, (~1,1) isa saddle @) O=a(l+z), 0=-z2+2}3+2122—2) -1 =2(1+22) > 21 =0orzy 2120 505-2423 > m=00rm m=-1 0=2-n-2f = nal There are three equilibrium points at (0,0), (0,1), and (1,1). Determine the type of each point using linearization Of _[ ltm a Oz =| 2—Se} -142e2+2 10 ffs eal BS ee [ZL] F as=21 > (on i unstable node |) -1 > (0,0) isa saddle =-14jv3 = (1,-1) isa stable focus a Alen (3) Dayz: Iva" 0=(1-2)m - nl2 CHAPTER 2. From the second equation, x2 = 0 or 22 = 2(1 +21) m2 =0 > 2 = Dora, 21421) $0=(e +3) > 2 ‘There are four equilibrium points at (0,0), (1,0); (0,2), and (-3, ~4). Notice that we have assumed 1421 # 0; otherwise the equation would not be well defined. fa abe ~2nthy jigenvalues: 1,2 => (0,0) is unstable node <1 -1 ae J) Bicenvaties = 1,25 (1,0) is a saddle 2 = (0,2) is a stable node (4) O=m, 0=-2 +2,(1-2} +0124) ‘There is a unique equilibrium point at (0,0). Determine its type using linearization. oF 0 1 Oz ~ | -1-2n 22 +04z}r. 1-2} +0.12f of | 01 2x00) ale il > ha =(1/2)45V3/2 > (0,0) is unstable focus (5) O= (1 —22)(1— 2} — 23), O= (a1 +22)(1— 2} - 23) {2} +2} = 1) is an equilibrium set and (0,0) is an isolated equilibrium point. a [1-8ef- af 2ne 222-142} +323 1-1 =|1-se}-2f-20122 -2ny22+1-2}-328 Jig [1 1 Eigenvalues are 1 + j; hence, (0,0) is unstable focus. (6) -a}+z, O=2-23 a =a] > 2(1-2f)=0 > 2 =0orat ‘The equation 2f = 1 has two real roots at 2, = +1. Thus, there are three equilibrium points at (0,0), (1,1), CD. , oy [oe oe 1 3231B 2s) = elma [1-3 Similarly, (1,1) is a stable node. 2.2 (1) O= 2, 0=—2 + (1/16)2$— 22 )=0 + O=a(2t-16) > x =0, 2, or -2 ‘There are three equilibrium points at (0,0), (2,0), and (—2,0). Determine the type of each point using linearization. eh } : ar | -14 Greet -1 | > ha =-(1/2)45V3/2 > (0,0) is a stable focus (1/2)£VT7/2 > (2,0) isa saddle Similarly, (~2,0) is a saddle. @) 0=22- 2122, 0= 222-2 2i(2—2,)=0 > x =00rz =2 m=0 3 m=0, 2 =2 3 af=1 5 m=1or-1 There are three equilibrium points at (0,0), (1,2), and (—1,2). Determine the type of each point using linearization. Q-m -n 4-1 ag 20 Fen lo | + Aa=2-1 + (0,0) isa saddle a elke | > d2=-(1/2)#5VI5/2 > (1,2) is a stable focus laa) 2 8 A] 4 ma =-(/2)£5Vi5/2 > (-1,2)is a stable focus Or | 1.2) ian (3) O=22, 0=-22~¥(z1—22) n=0 + vin)=0 + 1=0 ‘There is a unique equilibrium point at (0,0). Determine its type by linearization. of 3r| - 0 1 0 1 loo) | -8(a1— 22)? -0.5 -143(e1- 22)? +05 Joy | -05 -05 ‘The eigenvalues are —(1/4) + jV7/4. Hence, (0,0) is stable focus.u CHAPTER 2. «2.3 (1) The system has three equilibrium points: (0,0) is a stable focus, (1, -1) and (~1,1) are saddle points. The phase portrait is shown in Figure 2.1. The stable trajectories of the saddle form a lobe around the stable focus. All trajectories inside the lobe converge to the stable focus. All trajectories outside it diverge to infinity. (2) The system has three equilibrium points: (0,0) is a seddle, (0,1) is unstable node, and (1, ~1) isa stable focus. The phase portrait is shown in Figure 2.2. The 22-axisis a trajectory itself since 2 (t) = 0 = u(t) = 0. The z-axis is a separatrix. All trajectories in the right Falf converge to the stable focus. All trajectories in the left have diverge to infinity. On the z-axis itself, trajectories starting at z2 <1 converge to the origin, ‘while trajectories starting at 2» > 1 diverge to infinity x0) zo a “aoe 85 ° 3 Figure 2.1: Exercise 2.3(1). Figure 2.2: Exercise 2.3(2) (3) The system has four equilibrium points: (0,0) is unstable node, (1,0) is a saddle, (0,2) isa stable node, and (—3,—4) is a saddle. To avoid the condition z; +1 = 0, we limit our analysis to the right half of the plane, that is, (21 > 0}. This makes sense in view of the fact that the z2-axis is a trajectory since Z(t) = 0 = #(t) = 0. Hence, trajectories starting in {2; > 0} stay there for all time. ‘The phase por- trait is shown in Figure 2.3. Notice that the 21-axis is a trajectory since z(t) = 0 => a(t) = 0. It isa separatrix that divides the half plane {z1 > 0} into two quarters. All trajectories starting in the quarter {21 > 0, a2 > 0} converge to the stable node (0, 2). All trajectories starting in the quarter {2 > 0, +2 <0} diverge to infinity. ‘Trajectories starting on the zz-axis approach the stable node (0,2) if z2(0) > 0 and diverge to infinity if 22(0) <0. ‘Trajectories starting on the z-axis with 21(0) > 0 approach the saddle (1,0). (4) There is a unique equilibrium point at the origin, which is unstable focus. ‘The phase portrait is shown in Figure 2.4. There are two limit cycles. The inner cycle is stable while the outer one is unstable. Al trajectories starting inside the stable limit cycle, except the origin, approach it as ¢ tends to infinity. ‘Trajectories starting in the region between the two limit cycles approach the stable limit cycle. Trajectories starting outside the unstable limit cycle diverge to infinity (5) The system has an equilibrium set at the unit circle and unstable focus at the origin. ‘The phase portrait is shown in Figure 2.5. All trajectories, except the origin, approach the unit circle at ¢ tends to infinity. (6) The system has three equilibrium points: a saddle at (0,0) and stable nodes at (1,1) and (~1,~1). ‘The phase portrait is shown in Figure 2.6. The stable trajectories of the saddle lie on the line 21 + +2 = 0. All trajectories to the right of this line converge to the stable node (1,1) and all trajectories to its left converge to the stable node (—1, ~1). Trajectories on the line x: +2 = 0 itself converge to the origin.15 6 9 a 4 7 2 zal | | - o 123 -s! s ° 5 Figure 2.3: Exercise 2.33 i Figure 2.4: Exercise 2.3(4). 2 l= ol LA 0 “ Ke] “ a rr 4 4 Figure 2.5: Exercise 23(8) Figure 2.6: Exercise 2.36). ‘© 2.4 (1) The system has three equilibrium points at (0,0),(a,0), and (—a,0), where a is the root of a=tan(a/2) + 02.3311 ‘The Jacobian matrix is [12m +204] oz 1=2/[L + (zy +42)"] -2/[1 + (2 + 22)?] o1 ] = = a= Although we have multiple eigenvalues, we can conclude that the origin is a stable node because f(z) is an analytic function of z in the neighborhood of the crigin. 8) Peles.) =-f.o 2 - =| com well = Ang = 0.6892, -1 = (2.3311,0) is a saddle Similarly, (~2.9311,0) is a saddle. ‘The phase portrait is shown in Figure 2.7 with the arrowheads. The stable trajectories of the two saddle points forms two separatrices, which divide the plane into three regions. All trajectories in the middle region converge to the origin as ¢ tends to infinity. All trajectories in the outer16 CHAPTER 2. regions diverge to infinity. (2) 0= (2-22), From the first equation, 21 = 0 or 22 =2. a} — a2 m=0 > m=0 mods del _=41 There are three equilibrium points at (0,0), (1,2), and (~1,2) oF 200 Beles) le Al + Ma=2,-1 > (0,0) isasaddle Soa [2 aa] = se=-armsivis2 = 0,2) sa sable focus 7 \¢n,2) ~ & [ ” =| > Ma =—(1/2)45V15/2 + (-1,2) is a stable focus Oth 12) “ ‘The phase portrait is shown in Figure 2.7 with the arrowheads. The stable trajectories of the saddle lie on the 22-axis. They form a separatrix that divides the plane in two halves. Trajectories in the right half ‘converge to the stable focus (1,2) and those in the half converge to the stable focus (—1, 2). (3) There is a unique equilibrium point at the origin. ofl OF \(00) [2.2] > req crmsivay = (1,2) unstable fous ‘The phase portrait is shown in Figure 2.7 with the arrowheads. There is a stable limit cycle around the origin. All trajectories, except the origin, approach the Limit cycle as t tends to infinity. (4) The equilibrium points are given by the real roots of the equation omy 2 ey where z; = y? and 22 =1—y. It can be seen that the equation has four roots at y = 0,1, (~1:V5)/2. Hence, there are four equilibrium points at (0,1), (1,0), ((3 ~ 75)/2,(3~ v5)/2), and ((3 + v5)/2,(3 + v5)/2). ‘The following table shows the Jacobian matrix and the type of each point, Point Jacobian matrix Eigenvalues Type 2-1 (0,1) =n 2.4142, 0.4142 saddle 0-1 (1,0) es 0.4142, -2.4142 saddle eee | 0.2861, -2.2361 stable node (e-vane-vo [ “TP -19301 3.2361 en 4.2361, 2.2361 unstable node (a+ vay2o+v9/2) [92 ee7 ‘The phase portrait is shown in Figure 2.7 with the arrowheads. The stable trajectories of the saddle points, divide the plane into two regions. The region that contains the stable focus has the feature that all trajectories inside it converge to the stable focus. All trajectories in the other region diverge to infinity. : =] ‘| = : KF f SS x q 7 AN a 4s ° 3 0 @ 2 ql aE 2s| s 2| 15 q %, , 1 os o| 2 = 3 ® Figure 2.7: Phase portraits of Exercise 2.4 +25 (a) Of _[ -ltzizale) — -A(e) + 4a(z) a ~| A(2)-afale) 1 —-zxz20(2) where 1 (2) = ~—_) A(z) = _1_, (23 + 23)(In ay +23)?" infeiees Noting that lim,.40 2:2ja(z) = 0 for i,j = 1,2 and lims +0 A(#) = 0, it can be seen that af] -_f-1 0 ee ee (b) Transform the state equation into the polar ccordinates +a, o=tan (2 Yatra ome (5) ar = r(t) =r to obtain18 CHAPTER 2. and, for 0< 79 <1, 1 1 o> ine “Tay = 0 > Ot) = Op — In In 70] +t) + in(|inrol) Hence, for 0 < ro <1, r(¢) and 6(t) are strictly decreasing and lim... r(t) = 0, limy-4.06(t the trajectory spirals clockwise toward the origin, (©) f(2) is continuously differentiable, but not analytic, in the neighborhood of z = 0. See the discussion on page 54 of the text. # 2.6 (a) The equilibrium points are the real roots of 0 =~m, + az, ~ bayz0 +2, (a+ b)x + bo} — 212 From the second equation we have ai[-(0+8) +2 — 22] =0 + m1 =O0rz =—(0+b) +b, Substitution of 21 = 0 in the first equation yields (22 +a) =0 > 22=0orty ‘Thus, there are equilibrium points at (0,0) and (0,—a). Substitution of 22 = —(a+6) + br; in the first ‘equation yields oases) eRe a0 ae MOD gpa OD Hence, there is an equilibrium point at (422, =), (b) of -1- bee a bey +209 oz (a+b) + 2bx ~ 29 on 1. 2=(0,0) ‘The eigenvalues of A are ‘The equilibrium point (0,0) is a stable focus if 4a(a + 8) > 1, a stable node if 0 < da(a +b) <1, and asaddle if a(a+8) <0. 2. 2=(0,-a) ‘The eigenvalues of A are A= ab and 4 = —1. The equilibrium point (0,~a) is a saddle if b> 0 and a stable node if 6 < 0. oe lex -¥-a-% T+B | (a+)? Ha +)19 ‘The eigenvalues of A are x1+ Vi- Wr) 2 ‘The equilibrium point (4852, ={0322) isa stable focus i 46(a+b) > 1, astable node if 0 < 44(a++8) < 1, and a saddle if bla +8) <0. ‘The various cases are summarized in the following table. (0) a) | Gee. SS) TSO, delat) > 1, Was )> 1 | wable focus | saddle —[ ~ stable focus b> 0, 4a(a+6) > 1, 46a +6) <1 | stable focus | saddle stable node b> 0, da(a +5) <1, 45(a +5) > 1 | stable node saddle ‘stable focus b> 0, 4a(a +b) <1, 4b(a +b) <1 | stable node ‘saddle, stable node <0, a+b>0, dala +8)>1 | stable focus | stable node | — saddle b<0,a+6> 0, 4a(a+b) <1 stable node | stable node saddle, b<0,a+6<0, dats) >1 saddle ‘stable node ‘stable focus b<0,a+6<0, 4a(a+b) <1 saddle, stable node_ stable node If any one of the above conditions holds with equality rather than inequality, we end up with multiple eigenvalues or eigenvalues with zero real parts, in which case linearization fails to determine the type of the equilibrium point of the nonlinear system. (c) The phase portraits of the three cases are shown in Figures 2.8 through 2.10. {a= b=, The equilibrium pints are (60) sae ows (oy) sade (hot) Rabe oc Th ieavnsion at hemdaeis4=[ 9 1]. the stable cuovectris [1] andthe nae sgarcorit[~!] ty awd gh able se ute jr a ‘The stable trajectories form a separatrix that, divides.the plane into two halves, with all trajectories in the right half approaching (1,—1) and all trajectories in the left half approaching (0,0). iia=1, The equilibrium points are (0,0) stable focus (.-1) stable node (BF) saddle ‘The Hinearication a the sade A, where (1+ ##)4 = [ ~({/2) 1/8) |. the stable eigenvector a/s) (4/4) 0.9975 0.0709 is { na ] and the unstable eigenvector is [ rte ] Tet arate eens taececnieteea unstable trajectories of the saddle, ‘The stable trajectories form a separatrix in the form of a lobe. All trajectories outside the lobe approach (0, —1); all trajectories inside the lobe approach (0,0). iii a= 1, = —2. The equilibrium points are (0,0) saddle @-1) stable node @b stable focus ‘The linearization at the saddle is A= | ~} | ‘The stable eigenvector is | _) 8507 | and the20 CHAPTER 2. ;, [ 0.5257, . enti avers [ 255]. te ae eo gmt the sable and ual ont the saddle. The stable trajectories form a separatrix in the form of a lobe. All trajectories outside the lobe approach (0, —1); all trajectories inside the lobe approach (2, 4 4 5 A zo wo 2 “ao 2 5 ° 3 % % Figure 2.8: Exercise 2.6(i. Figure 2.9: Exercise 2.6(i) 5 zo -s 5 ° 3 Figure 2.10: Exercise 2.6( © 2.7 The system a =a, <2; — 2) (-1+432} - x} + 28) ou -11 eigenvalues are 0.5 + 0.8663. Hence the origin is unstable focus. The phase portrait is shown in Figure 2.11. ‘There are three limit cycles. The inner limit cycle is stable, the middle one is unstable, and the outer one is, stable. All trajectories starting inside the middle limit cycle, other than the origin, approach the inner limit cycle as ¢ tends to infinity. All trajectories starting outside the middle limit cycle approach the outer limit cycle as t tends to infinity. Trajectories starting at the urstable focus or on the unstable limit cycle remain there hhas a unique equilibrium point at the origin. Linearization at the origin yields A whose + 2.8 (a) The equilibrium points are the real roots of O=m, O=-n 4 irh— aea Figure 2.11: Exercise 27. Hence 21(16~2$) =0 = 21 =0, 42 ‘There are three equilibrium points at (0,0), (2,0), and (~2,0). of 0 1 Or | a4 het -1 o1 --09 = 4-{ ]--24 5 a1 (0,0) is a stable focus. 0 2 = (2,0) or (-2,0) A [ 4 (2,0) and (—2,0) are saddle points. (b) The phase portrait can be sketched by constructing @ vector field diagram and using the informa- tion about the equilibrium points, especially the directions of the stable and unstable trajectories at the saddle points. The stable and unstable eigenvectors of the linearization at the saddle points are wow[ata [see [ata [a Find the directions of the vector fields on the two axes. On 2 = 0, f = [ 2 field makes an angle —45 deg with the 2, axis and its magnitude increases with |ra|. On z2 = Hence the vector2 CHAPTER 2. 0 4mK,, the equilibrium is a stable node, and if (Ky +2Kava + K,)? < 4mK,, the equilibrium is a stable focus. (c) For the given numerical values, the eigenvalues of the linearization are —0.0289 and ~0.3461. Hence, the equilibrium point is a stable node. ‘The phase portrait is shown in Figure 2.12. All trajectories approach the stable node along the slow eigenvector of the node, which has a small slope. Starting from different initial speeds, the trajectory reaches the desired speed with no (or very little) overshoot. (a) The eigenvalues of the linearization are —0.1875 + 0.2546,; hence the equilibrium point is a stable focus. ‘The phase portrait is shown in Figure 2.13. All trajectories approach the stable focus. Notice the increased overshoot compared with the previous case. For example, starting at the initial state (1, = 15, x2 = 10), the speed reaches about 36 m/sec before approaching the steady-state of 30 m/sec. (e) The phase portrait is shown in Figure 2.14. The local behavior near the equilibrium point is not affected since saturation will not be effective. However, far from the equilibrium point we can see that the state of the integrator, 11, takes large values during saturation, resulting in an increased overshoot. + 2.10 (a) Using the same scaling as in Example 2.1, the state equation is given by 2-21 ~ 0.222 +0.2) 1 =08[-Ha1) +22), te where A(z) is given in Example 2.1. ‘The equilibrium points are the intersection points of the curves 2, = h(a) and 29 = 1~ 521. Figure 2.15 shows that there is a unique equilibrium point. Using the “roots” ‘command of MATLAB, the equilibrium point was determined to be 2 = (0.057,0.7151). Of _[ -O5h(x1) 05 oe -02 -0.04 ay 05 “02 -0.04 0.9343, 0.1518 = stable node (b) The phase portrait is shown in Figure 2.16. All trajectories approach the stable node. This circuit is known as ‘“monostable” because it has one steady-state operating point. |] = Biceaatues23 a a} wo oo 100 ao Figure 2.12: Exercise 2.9(c) Figure 2.13: Exercise 2.9(4). | c 5 Figure 2.14: Exercise 2.9(¢) + 2.11 (a) Using the same scaling as in Example 21, the state equation is given by iy =0.5[-Ma) +22), 2 =0.2(-z1 - 0.21 +04) where h(z1) is given in Example 2.1. The equilibrium points are the intersection points of the curves 2 = h(x) and x = 2—52,. Figure 2.17 shows that there is a unique equilibrium point. Using the “roots” command of MATLAB, the equilibrium point was determined to be # = (0.2582, 0.7091). Of) _ [18173 05 Oz\pap | 0.2 -0.04 (b) The phase portrait is shown in Figure 2.18. The circuit has a stable limit cycle. All trajectories, except the constant solution at the equilibrium point, approach the limit cycle. This circuit is known as “astable.” | > Become = 1.7018,00155 -+ unsiable node 4 , on taal | Es os o5| oa A 02| il , or 8s Figure 2.15: Exercise 2.10: equilibrium point. Figure 2.16: Exercise 2.10: phase portrait.24 CHAPTER 2. 2 15) 4 “2| 15 1 1 ost» og | o| x ° “$s 00s 1s 0 os yt Figure 2.17; Exercise 2.1: equilibrium point. Figure 2.18: Exercise 2.11: phase portrait. + 2.12 (a) Note that Tia = Ta =1 > phy = php = Land Ta = Tan =0 > phy = phy = 0. Hence the state equation is given by y= M(es)fan — 2n(za)] file), da = (ea) [2a — 2n(x2)] fale) where h(z) = \cos*(n2/2) and n(2) = g~1(2) = (2/rA) tan(x2/2). Equilibrium points are the intersection points of the curves = 2n(z1) and 2; = 2y(z). Note that 1/(0) = 1/dand (2) = (1/%) sec*(w2/2) > 1/ ‘Therefore, for A < 2, the two curves intersect only at the origin (0,0). For \ > 2, there are three intersection points at (0,0), (@,a) and (—a,~a) where 0 < a <1 depends on A. This fact can be seen be sketching the curves and using symmetry; see Figure 2.19. The partial derivatives of fi and fz are given by oh oh Gh = weenyier—2nlen]-aneviten), FL = ne) Bie non), BE = wey ~ anton) 2h(ca) es) At equilibrium points, [72 ~ 2n(21)] = 0 and [x1 — 2n(x2)] = 0. Therefore, the Jacobian matrix reduces to oF _ —27/(6) 1 | =00 [1 rey | where b = 0, a, or ~a, depending on the equilibrium point. of 3e| han [2 23] © Benes 28 For A < 2, the unique equilibrium point at (0,0) is a stable node. For A > 2, the equilibrium point (0,0) is adde, For A> 3 there are two other equim points at (a,a) and (aya). af ma) 1 ] , ar Ne = hey{=20/(a) #1 of (0) [Lay | Bienes = Mo)-27) +1 It is not hard to see from the sketch of the curves zz = 2n(z,) and x, = 2n(z2) that at the intersection point (a,a), the slope 2n/(a) > 1. Hence, (a,a) is a stable node. Similarly, it can be shown that (—a,—a) is Eoeable node (b) The phase portrait is shown in Figure 2.20. The stable trajectories of the saddle point at the ori- gin form a separatrix that divides the plane into two regions. ‘Trajectories in each region approach the stable node in that region.1 4 | 05| os} d x -05} 03 “05008 Figure 2.19: Exercise 2.12: eq ium point. + 2.9 (a) Using Kircofts Voltage law, we obtain n= cte—-n- f= gg ate) ‘Using Kirchhoff's Current law, we obtain ee ee ee ana BRM ™ ate} ‘Thus, the state equation is 1 ay [—21 + 22 — 9(z2)] GR : 1 1 1 1 fy = gen glen cht cates GR" GR?” GR" * GR (b) For the given data, the state equation is given by y= tan—g(ta), t= 21 ~ 2x + 9(e2) (22) = 3.2342, ~ 2.19523 + 0.66625 ‘The system has a unique equilibrium point at the origin. ‘The Jacobian at the origin is given by a Ae ]- Eigenvalues = 0.117 + 0.993) Hence the origin is an unstable focus. The phase portrait is shown in Figures 2.21 and 2.22 using two different scales. The system has two limit cycles. The inner limit cycle is stable, while the outer one is unstable. All trajectories starting inside the outer limit cycle, except the origin, approach the inner one. All trajectories starting outside the outer limit. cycle diverge to infinity. © 2.14 The system is given by fy = —kxy — ox — (21,22) where aka, for x2 = 0&|21| < wyng/k =waig signa), for x2 = O&|zy| > pamg/k pamg sign(z2), for za >0 (1,22)26 CHAPTER 2. ° 1 4 NY 2 os Ce = 9 d -2| > ) y -05| -4| oS ° 3 “0500S Figure 2.21: Exercise 2.13. Figure 2.22: Exercise 2.13. For 2 > 0, the state equation is given by a =m fy = kay ~ cra — pag while, for 22 <0, it is given by aA in m key ~ cra + pamg In each half, we can determine the trajectories by studying the respective linear equation. Let us start with 2 > 0. The linear state equation has an equilibrium point at (—sarng/k,0). Shifting the equilibrium to the crn we tan net otin wh mais |, +, | wh hrc agin A 404k = 0, where k and c are positive constants. The equilibrium point is a stable focus when 4k > c? ‘and a stable node when 4k < c®. We shall continue our discussion assuming 4k > c?. Trajectories would tend to spiral toward the equilibrium point (~sxmg/k,0). It will not actually spiral toward the point because the equation is valid only for z2 > 0. Thus, for any point in the upper half, we can solve the linear equation to find the trajectory that should spiral toward the equilibrium point, but follow the trajectory only until it hits the z-axis. For 2» <0, we have a similar situation except that trajectories tend to spiral toward the point (jxmg/k,0). On the 2,-axis itself, we should distinguish between two regions. Ifa trajectory hits the y-axis within the interval [—y.mg/k, y4mg/k}, it will rest at equilibrium. If it hits outside this interval, it will have ¢ # 0 and will continue motion. Notice that trajectories reaching the z1-axis in the interval 2, > juamg/k will be coming from the upper half of the plane and will continue their motion into the lower half. By symmetry, trajectories reaching the z-axis in the interval 2; < —y4smg/k will be coming from the lower half of the plane and will continue their motion into the upper half. Thus, a trajectory starting far from the origin, will spiral toward the origin, until it hits the z-axis within the interval [—yamg/k, usmg/B ‘The phase portrait is sketched in Figure 2.23. ‘© 2.15 The solution of the state equation fy = 2, 1(0)=20 fz =k 22l0) where k = £1, is given by a(t) = kt+20 nl) = fee +20t +2107 Figure 2.23: Exercise 2.14 Eliminating t between the two equations, we obtain 1 n= gite where ¢ = 210 ~ 24/(2k). This is the equation of the trajectories in the 24~r2 plane. Different trajectories correspond to diferent. values of c. Figures 2.24 and 2.25 show the phase portraits for u = 1 and w= ~1, respectively. The two portraits are superimposed in Figure 2.26. From Figure 2.26 we see that trajectories can reach the origin through only two curves, which are highlighted. The curve in the lower half corresponds to u = 1 and the curve in the upper half corresponds to u = —1. We will refer to these curves as the switching curves. To move any point in the plane to the origin, we can switch between :+1. For example, to move the point A to the origin, we apply u = —1 until the trajectory hits the switching curve, then we switch to u = 1. Similarly, to move the point B to the origin, we apply u = 1 until the trajectory hits the switching curve, then we switch to u = —1. When the trajectory reaches the origin we can keep it there by switching to u=0 which makes the origin an equilibrium point. 4————— 4 aN a 7] “eo “a x * Figure 2.24: Exercise 2.15. Figure 2.25: Exercise 2.15 ‘+ 2.16 (a) The equilibrium points are the roots of O=2(1-2—or2), 0=bz2(01- 22) From the first equation, we have 2; = 0 or x = 1~az2. Substitution of z = 0 in the second equation results in z2 = 0. Substitution of x; = 1 azz in the second equation results in z2(1— azz — x2) = 0 which yields28 CHAPTER 2. Figure 2.26: Exercise 2.15. 29 = 0 or x = 1/(1 +a). Thus, there are three equilibrium points at (0,0), (1,0), and (1/(1+a), 1/(1+a)). The Jacobian matrix is given by as lanes Sees mary ] 24 3] af ba bay ~ 2ba2 [0 0} shy Lo F) Slew FL? 3] an low Flaw Tl? 3 At the equilibrium point (0,0) the matrix has a zero eigenvalue; hence linearization fails to determine the type of the equilibrium point. At (1,0), the equilibrium point is a saddle. At (745, 2 sia rfz)> the eigenvalues (1+ 8) + VI- 254 tab 30a) dug Hence, (sets) is a stable node if 1 — 2b + b? ~ 4ab > 0 and a stable focus if 1 — 2b+ b? — dab <0. ‘The phase portrait is shown in Figure 2.27. The equilibrium point ($, ) is a stable focus that attracts all trajectories in the first quadrant except trajectories on the z-axis or the 2 axis. Trajectories on the z-axis ‘move on it approaching the saddle point at (1,0). Motion on the 2,-axis corresponds to the case when there are no predators, in which case the prey population settles at 21 = 1. Motion on the 2 axis corresponds to the case when there are no preys, in which the case the predator population settles at 22 = 0; ie., the predators vanish. In the presence of both preys and predators, their populations reach a balance at the equilibrium point (3,4) © 2.17 (1) Assume > O and let 21 =y, 29 = jy and V(2) = 23 +23 02, fy =—2 +e0x(1—2]- 23) $2) V (2) = e241 ~ 29 29) = 2e23(1-V)Figure 2.27: Exercise 2.16. Hence, f(z)- VV (2) < 0 for V(z) > 1. In particular, all trajectories starting in M = {V(z) < 1} stay in M for all future time. M contains only one equilibrium point at the origin. Linearization at the origin yields ‘the matrix [ ©, 1). Hence, the origin is unstable node or unstable focus. By the Poincaré Bendixson criterion, there isa periodic orbit in M (2) Let V(2) = 23 +23. Se) WV (a) = 2a}(2 — 325 ~ 209) = 429(0 — 2} — 28) — 222} < 409 - 2} 29) Hence, f(z) VV(z) < 0 for 2} +2} > 1. In particular, all trajectories starting in M = {V(2) < 1} stay in M for all future time. M contains only one equilibrium point at the origin. Linearization at the origin yields the matrix | °°} } whose eigenvalues are 1 and 1. Since f(z) is an analytic function of 2, we conclude that the origin is unstable node, By the Poincaré-Bendixson criterion, there isa periodic orbit in M. (8) Let V(z) = 323 + 22422 + 22} f(z) VV(z) (G2; + 2x22 + (2x1 + 422)[—21 + 22 — 2(21 + 2x2)23) — 2a} + dzyz5 + 62} — A(x + 222)?2 =2(x} + 28) + ray + 20a) ~ A(x + 2x2)*2} = -2(a} +23) +1 [1 - 2xo(e1 + 202)) < 22 +23) 410, forz?+a3>3 Choose a constant ¢ > 0 such that the surface V(x) = c contains the circle {z} + 23 = }} in its interior. ‘Then, all trajectories starting in M = {V(z) < c} stay in M for all future time. M contains only one Of Je wioce genta 1a are (1:4 jV3)/2. Henee, the origin is unstable focus. By the Poincaré-Bendixson criterion, there is a periodic orbit in Mf, equim point a the origin. Linearization at he origin yes the matic [ (4) The equilibrium points are the roots of O= 21 422-2, max{lzi[[2zl), 0= 221 +22 —22max{lail|22l} From the first equation, we have 2 = 23(max{[r,[r2|} ~1). Substitution in the second equation results in 0 = ~221 ~ 24 (max{|rs|,|22]} — 1)? [2+ (max{|ni[,l2al}- 1] + 21 =0 22 =030 CHAPTER 2. Hence there is a unique equilibrium point at the origin. Linearization at the origin yields the matrix jt | whose eigenvalues are 1 + j1.4142. Hence, the origin is an unstable focus. Now consider ital wef 2x, (zy + 22 — 2; max |x|, |22|}) + 2z2(—2x + 22 — x2 max{|z4|,|z2]}) = Qe} — 2x9 + 2x} — 2x} + 23) max{ ln], |22l} 2fo"Pe—lol max(tsh fs], where P=[ 5, “2° ] The matrix P is positive definite with maximum eigenvalue 1.5. Therefore, WV F< 2L-bllall3— IIell3 max{le|,|2al}] <0, for max{|z|,|22]} > 1.5 ‘Thus, by choosing c large enough, VV - f will be negative on the surface {V(z) = c}. Hence, all trajectories starting in the set M = {V(2) 0, the equation V(z) = c defines a closed curve that encloses the origin. Since V = 0, a trajectory starting on the curve must remain on the curve for all . Moreover, from #1 = 72, we see that the trajectory can only move in the clockwise direction. Hence, a trajectory starting at. any point on the closed curve V(z) = ¢ must move around the curve until it comes back to the starting point. Thus, the trajectory is a periodic orbit, (c) Extension’ of (b) because V(x) = c is a closed curve forall c > 0. constant. At z= (4,0), V =G(A). Thus £V2G(A) - Gi())) (d) V(x) = 423 + G(x) HAW +61) = G(A) > 22(0 () Starting from i, = 22, we have a= a for x2 > 0. Calculating the line integral of the right-hand side in the upper half of the plane from (~A,0) to (4,0), we obtain Tr 271 yaeccaan 7 yeaa where we have used the fact that G(c1) is an even function. (f) We can generate the trajectories using the equation in part (d). For each value of A, we solve the equation to find x2 as a function of 2. The function G(r) has a minimum, a maximum, or a point of inflection at each equilibrium point of the system. In particular, It has a minimum at x; = 0 corresponding to the equilibrium point at the origin. Starting from small values of A, the equation will have a solution defining a closed orbit. As we increase the value of A, the equation will continue to define a closed orbit until A reaches the level of a maximum point of G(r). For values of A higher than the maximum, the curves ‘will not be closed. Depending on the shape of G(si), the equation may have multiple solutions defining trajectories in different parts of the plane. The conditions of part(c) ensure that G(z,) will have a global minimum at 2 = 0.31 © 2.19 The phase portraits can be generated by solving the equation of the previous exercise either graph- ically or using a computer. We will only give the function G(x) and calculate the period of the trajectory through (1,0). qa) cosy Gu) = fine a tt feosy — 8 1 T=2V2 lo ® , ow = [er dea hie Gly) + 00 as ly] + 00 and zg(z) > 0 for all z. Hence, every solution is periodic. T=28 lo ® , ow = [ de G(y) > co as |y| + 00 and z9(z) > 0 for all z. Heace, every solution is periodic. rao f b Ji +220 @) Oh Oi Ox, © Ba. By Bendixson’s criterion, there are no periodic orkits. (2) The equilibrium points are the roots of ay(-1+23+23), 0=2,(-1+2} +23) The system has an isolated equilibrium point at the origin and a continuum of equilibrium points on the unit circle 2} +23 = 1. It can be checked that the origin is a stable node. Transform the system into the polar coordinates 21 =r cos6, x2 = rsind. It can be verified that fe-r-7) For r < 1, every trajectory starting inside the unit circle approaches the origin as t + oo. For r > 1, every trajectory starting outside the unit circle escapes to co as t -+ oo. Thus, there are no limit cycles. (3) The equilibrium points of the system are the roots of ° -nzi, 0=n ‘These equations have no real roots. Thus, there are no equilibrium points. Since, by Corollary 2.1, a closed orbit must enclose an equilibrium point, we conclude that there are no closed orbits.32 CHAPTER 2. (4) The 21-axis is an equilibrium set. ‘Therefore, a periodic orbit cannot cross the z-axis; it must lie entirely in the upper or lower halves of the plane. However, there are no equilibrium points other than the z-axis. Since, by Corollary 2.1, a periodic orbit must enclose an equilibrium point, we conclude that there are no periodic orbits. (5) The equilibrium points are the roots of O=s008z, O=sinzy The equilibrium point are (nr,0) for n = 0,1,2,--». Linearization at the equilibrium points yields the (p | where @ = 41. Hence, all equilibrium points are saddles. Since, by Corollary 2.1, a periodic orbit must enclose equilibrium points such that N — $ = 1, we conclude that there are no periodic orbits. matrix °2.21 (a) Let mtb n+ ‘The function V(z) is negative in D and the curve V(z) = 0 is the boundary of the set D. V(z) = 22- (2) VV (2) = ~ers (21 +0) + ny + 29(z1 +2) ~ 8] Evaluating f(z) - VV(z) on the curve V(z) =0 yields $2) Wy (eyn Hence, trajectories on the boundary of D must move into D, which shows that trajectories starting in D cannot leave it. (b) en (1 +a) <0, Yee aD Of , Of, _ Bor * Gea = By Bendixson's criterion, there canbe no closed orbits entirely in D. Since trajectories starting in D cannot Teave it, a closed orbit through any point in D must lie entirely in D. Thus, we conclude that there are no closed orbits through any point in D. L42,<0, VrEeD + 2.22 (a) The value of Zy on the z-axis is #y = baf >0. Thus, trajectories starting in D cannot leave it. (b) ah . of gh + 52 =a-m-0< -(c~a) z Ber * Beg $-(e-a) <0, VreD By Bendixson’s criterion, there can be no closed orbits entirely in D. Since trajectories starting in D cannot leave it, a closed orbit through any point in D must lie entirely in D. Thus, we conclude that there are no closed orbits through any point in D. + 2.23 (a) -2ab for [21] >1 =a(2b—k) for fas] <1 Oh , Of kdb Fit 5 <0 VeBy Bendixson’s criterion, there are no closed orbits. (b) Consider the case k > 2b. We shall show the existence of a periodic orbit by imitating the steps of Example 2.9. We track the trajectory starting at point A = (0,p); see Figure 2.28. At the starting point, % FP Figure 2.28: Exercise 2.23. fi = 22 >0 and f= (k - 26)aza > 0. So, the trajectory starts with a positive slope, Within the segment 0.< 21 <1, the trajectory will continue to have a positive slope, provided p is large enough, until it arrives at B = (1,8(p)). As the trajectory leaves B, we save fy = —2abr2 — a,z; < 0. Thus, the trajectory will turn around forming the curve BCD. Let D = (1,~7(p)) and consider the motion on the curve BCD. Let V(z) = a°z? +23. Then V = 20%e122 + 2r9(—0?21 — 2abr2) = —dabz3 <0 vio)-viB)= fv a Noting that V(D) ~ V(B) = a? + 72(p) ~ a? — f(p), we obtain F) - 0) As p increases, the arc BCD moves to the right and the domain of integration increases. It follows that, 7°(p) — 8?(p) decreases as p increases and lim, (0°@) ~ B°()} + ~00 a p00 ‘Thus, for sufficiently large p, by the time the trajectory reaches the point E' = (0,~8(p)) on the zp-axis, we have 4(p) < p. Similar to Example 2.9, we form a closed curve of the are ABCDE, its reflection through the origin, and segments on the z-axis connecting these arcs. Let M be the region enclosed by this closed curve. Every trajectory starting in M stays in M for all future time. Linearization at the origin yields the matrix | ©, 15, |» which has both eigenvalues in the right-half plane. Thus, the origin is unstable node or unstable focus. By the Poincaré-Bendixson criterion, there is a periodic orbit in M. ata [23% dry = dab 3 ars lap da lop Wa + 2abez34 (CHAPTER 2. # 2.24 Suppose M does not contain an equilibrium point, Then, by the Poincaré-Bendixson criterion, there is a periodic orbit in M. But, by Corollary 2.1, the periodic orbit must contain an equilibrium point: A contradiction. Thus, M contains an equilibrium point. ° 2.25 Verifying Lemma 2.3 by examining the vector fields is simple, but requires drawing several sketches. Hence, it is skipped. (2) Unasaton ath ists [8%]. eh ii ot pee, The nerf origin is zero. This can be easily seen by noting that fy = 2? is always nonnegative. Clearly, the vector field cannot make a full rotation as we encircle the origin because this will require f; to be negative. (2) tomentose eign ts [© 9] ace te og ont Tar ne is two. This can be seen by sketching the vector field along a closed curve around the origin. + 2.27 (1) The equilibrium points are the real roots of O=22, O= plz; +22)— 22-2} - 32722 of 8 |,0.0) 2 =e Ot V0) —H Similarly, (~ 7,0) is a stable node. For 1 <0, there is a unique equilibrium point at (0,0). as o4 | > Agen, # = (0,0) isa stable node Thus, there is supercritical pitchfork bifurcation at y= 0. (2) The equilibrium points are the real roots of, O=mzp tae, O= —(1+y?)ey + 2uze — way + 222 — prs)? ar {-1+ (ef wll + 22h(ef — oI} For all values of , there is an equilibrium point at (0,0). At = 0, there are two other equilibrium points at (a,a*) and (—a,—a®), where a® = 0.5. It can be checked that these two equilibrium points are saddles. By continuous dependence of the roots of a polynomial equation on its parameters, we see that there is a range of values of 1 around zero for which these two saddle points will persist. We will limit our attention to such values of ys and study local bifurcation at y = 0. pi ar neat = wei oo [ats a] > Msn-01 o6| os | oa 02 02 wo wo 0 -02 -o4 04 06 -06 08 ay a rT a =-005 2008 og 06] o4| oa o2| oI wo 2 of -02| 02] 06 -05| oe -oal 5 OCS a 5 % Figure 2.29: Exercise 2.27(2). Hence, the origin is a stable focus for 1 < 0 and unstable focus for 4 > 0. The phase portrait for different values of 1 is shown in Figure 2.29. For js < 0, there is a stable focus at the origin and unstable limit cycle around the origin. The size of the limit cycle shrinks as y tends to zero. For > 0, the origin is an unstable focus and the limit cycle disappears. Hence, there is a subcritical Hopf bifurcation at = 0. (3) The equilibrium points are the real roots of O=2, O=p—22-2}-2x22 = de m= When 1 <0, there are no equilibrium points. When 1 > 0, there are two equilibrium points at (/7,0) and (-v0) as o 1 Bel yao) lve aa > Aa=-1, -2Vii > (Vii0) is a stable node of -[.° 1 _ Bele leva -e-21m] + Mg=-l, 2H > (Vi0) isa saddle ‘There is a saddle-node bifurcation at = 0. (4) The equilibrium points are the real roots of Osa, O= (+i )z + Que + wat — ahee36 CHAPTER 2. 2=0 > 0=-(14 ux + u2} For all values of ps, there is an equilibrium point at (0,0). For > 0 there are two other equilibrium points at (2,0) and (~a,0), where a = (1+ #)/u of [ 0 | ee, (oo) oe —(+H) Hence, the origin is a stable focus for 4 < 0 and unstable focus for > 0 Bela on L2ats) case | > wi it) Hence, (a,0) and (~a,0) are saddle points. The phase portrait for different values of ys is shown in Fig- ure 2.30. For jz <0, there is a stable focus at the origin. For 1 > 0, the origin is an unstable focus and there is a stable limit cycle around the origin. The size of the limit cycle shrinks as 4 tends to zero. Hence, there is a supercritical Hopf bifur 0. Also, as becomes positive, the saddle points appear on the sasis a 5, = 2/(LE APY] The aude points sart at indnity and they move toward the origin 86 1 increases, until they reach +2 at = 1. Then they move again toward infinity. For 4 = 0.2, the phase portrait is shown in a larger area that includes the saddle points. peat =o. wo wo 4 4 ° +e rr % 4 n=02 n=02 2 5 20 wo -s eee $5 3 3 Figure 2.30: Exercise 2.27(4). (5) The equilibrium points are the real roots of Om, 0 = ple +22)- 22-2} + 32}z237 m=0 3 0=n(u-2) For all values of jz, there is an equilibrium point at (0,0). For 1 > 0 there are two other equilibrium points at (,0) and (—J/H,0) ooL wopnl she Joo Hence, the origin is a stable node for < 0 and a saddle for > 0. $4 (= tu) & JOP =H oz (vB) [ teen | he ‘The following table gives the type of the equilibrium points (v/7,0) and (— 7,0) for various positive values of Range ‘Type O< pu < 0.067 ‘Stable node 067 < p< 035 ‘Stable focus Unstable focus Unstable node 0.25

0. Belin |“ -ory] 2 oz Hence, (11,0) is a saddle for 1 < 0 and a stable node for « > 0. There is a transcritical bifurcation at j= 0. bt © 2.28 (a) The equilibrium points are the real roots of, = te +tanh(Az) —tanh(Azz), 2 + tanh(Azy) + tanh(Az2) By adding and subtracting the two equations, we see that the equilibrium points are the intersections of the ‘two curves . zy= m1 +2rtanh(Azy), 1 = x2 — 2rtanh(dz2) Clearly there is an equilibrium point at the origin (0,0). By plotting the two curves for different values of Ar (see Figure 2.32), it can be seen that the origin is the only intersection point. In fact, the two curves touch each other asymptotically as Ar —+ 00. Thus, we conclude that the origin is the only equilibrium point. Next ‘we use linearization to determine the type of the equilibrium point. -sstes ~danle remo — t+ am ony +A =A Qr-1) Hence, the origin is a stable focus for Ar < 1 and unstable focus for 4r > 1. To apply the Poincare-Bendixson criterion when Ar > I, we need to find a set M that satisfies the conditions of the criterion. We do it by transforming the equation into the polar coordinates 22) a 24+ cos(6)[tanh(Ar cos(0)) ~ tan(Ar sin(0))] + sin(@)ftanh(r cos()) + tanh sin(6)] rayatte}, 0 = tan Using | tanh(-)| < 1, |cos(-)] < 1, and | sin(-)| < 1, we see that ten teea Choosing r = c > 4r, we conclude that on the circle r =o, # <0. Hence, vector fields on r = point to the inside of the circle. Thus, the set M = {r 1, there is an unstable focus at the origin ‘and a stable limit cycle around the origin. Hence, there is a supercritical Hopf bifurcation at Ar = 139 Aeos.cet ae Figure 2.33: Exercise 2.28. * 2.29 (a) The equilibrium points are the real roots of EA O=a-n~ 2. ote (- ry) From the second equation we have 2 = 0 or +2 = 1+}. The frst equation cannot be satisfied with 21 = 0 Substitution of 22 = 1+ 22 in the first equation results in x; = a/5. Thus, there is a unique equilibrium point at ((a/5),1+ (0/5)®). The Jacobian matrix is given by yt tee of O(1- af) + Gee - sf ar - [sree ~4(0/5) 1, Bz | cajsyaetasy) Pe C@/S | 2(a/5)?— -H(a/5) | ~ T+ (a/5*40 CHAPTER 2. ‘The characteristic equation of B is 2 +Bs+7=0 where B=5-3(a/5)? +b(a/5), 7 = 5[1 + (a/5)*}6(a/5) We have 7 > 0. If 8 < 0, the eigenvalues will be real and positive or complex with positive real parts; hence the equilibrium point will be unstable node or unstable focus. 6 is negative if } < 3(a/5) ~ 25/a. To apply the Poincare-Bendixson criterion, we need to choose the set M. Figure 2.34 sketches the two curves whose intersection determines the equilibrium point. On the sketch we identify a rectangle with vertices at 4, B, C and D. On the line AB, #2 > 0; hence the vector fields point upward. On the line BC, #1 < 0; hence the vector fields point to the left. On the line CD, 2 <0; hence the vector fields point downward. On the line ‘DA, #1 > 0; hence the vector fields point to the right. ‘Thus, taking the set M to be the rectangle ABCD, we see that every trajectory starting in M stays in M for all future time. Moreover, M is closed, bounded, and contains only one equilibrium point which is unstable node or unstable focus. By the Poincare-Bendixson criterion, we conclude that there is a periodic orbit in M. 5 a Figure 2.34: Exercise 2.29. (b) For a = 10 and b = 2, we have 6 < 3/5 ~ 25/a. The equilibrium point is (2,5) and it is unstable focus. The phase portrait is shown in Figure 2.35. The system has a stable limit cycle. All trajectories, except the equilibrium solution x = (2,5), approach the limit cycle asymptotically (c) For a = 10 and b = 4, we have 6 > 3a/5~ 25/a. The equilibrium point is (2,5) and it is a stable focus. The phase portrait is shown in Figure 2.35. All trajectories approach the equilibrium point asymp- totically, (A) For & < 30/5 - 25/a, 8 is negative. Moreover, when 6 is close to 34/5 ~ 25/a, 8 will be close to zero. Hence, 4+ > 6? and the equilibrium point is unstable focus. As we saw from the phase portrait, there41 a=10,b=2 15 15 10 1] S 10 Figure 2.95: Exercise 2.20. is a stable limit cycle around the equilibrium poin:. For b > 3a/5 — 25/a, 8 is positive. Once again, when & is close to 3a/5 — 25/a, 8 will be close to zero. Hence, 4y > ? and the equilibrium point is a stable focus. ‘There is a supercritical Hopf bifurcation at b = 3a/5 ~ 25/a. © 2.30 (a) The equilibrium points are the roots of tina = deny — 9) - HOT ° (ea )m O= dear —#2)~ 7g, +29) ‘The first equation has two solutions: 2; = 0 or the solution of d = jin 2/(km +22), which we denote by a. When d < jim, there is a unique solution @. Substitution of x = 0 in the second equation yields x2 = 22. Substitution of z2 = a in the second equation yields x; = Y(z27 — a), which will be a feasible solution if @& < af; that i8, d < pi 2g (Km + 224). Thus, when d < jmt24/(km + 227), there are two equilibrium points at (0,22) and (Y(z2y —a),a). When d > im22y /(Km + 224), there is a unique equilibrium point at (0,224). oy [oRR-4 dete ‘ea 0 a(S | Vintesy —4 Hence, (0,727) is a saddle if d < maz /(km + 224) and a stable node if d > pmz2s/(km + 224): (204-0) ivienr-are) | gd — Bega ‘The eigenvalues are —d and ~kmptm(22y ~ @)/ (Kn. +.0)®. For d < pmtay/(km + 224); (¥ (224 ~@),0) is stable node. For the given numerical data, m2//(Km + 22) = 0.4878. When d > jm, there is a unique equilibrium point at (0,227), which is a stable node. a be (b) The bifurcation diagram is shown in Figure 2.36. As d increases toward 0.4878, the saddle at (0,22) and the stable node at (¥ (22 — a), a) collide and bifurcate into a stable node at (0,22).42 CHAPTER 2. Saddle Stable node 4 4 0.4878 % Figure 2.36: Exercise 2.30: Bifurcation diagram. Figure 2.37: Exercise 2.30 (©) For d = 0.4, @ = kmd/(jm ~d) = 04. Since d = 0.4 < 0.4878, there is a saddle at (0,4) and a stable node at (1.44,0.4). The phase portrait is shown in Figure 2.37. +281 (a) Let © m=z Kin + 22 + iad was The equilibrium points are the roots of z1n(ea) O=z[u(z2)—d], 0 = d(zay — 22) — From the first equation, 2 = 0 or u(z2) = d. ny win) =d m1 =¥ (ay —m) By sketching the function (2), it can be seen that if d < maxz,>0{u(z2)}, the equation d = (72) will have ‘two solutions. Let us denote them by a1 and az. In this case, there are three equilibrium points at (0,27), (¥ (x2 — a1), 1), and (¥ (22g ~ a2), a2), provided zy ~ax and 227 —a3 are nonnegative numbers. If one of these numbers is negative, the corresponding equilibrium point is not feasible. If d > maxz,>o{j4(22)}, the equation d= (22) has no solutions. In this case, there is only one equilibrium point at (0,227). The plot of 22s a function of zz is shown in Figure 2.38. By differentiation, it can be seen that has a maximum value (um y/Em7Bx)/ (2km + \/Fim/Ex) = 0.3455 at 22 = y/km/Fi = 0.4472. When d > 0.3455, there is a unique ‘equilibrium point at (0,4), and when d < 0.3456 there are three equilibrium points at (0,4), (0.4(4—a1),a1), and (0.4(4—az),g), where a < 0.4472 and a2 > 0.4472 are the solutions of d= (22). In the case of az, the equilibrium point (0.4(4— az), <2) is not feasible if az > 4. It. can be checked that 1 = 0.1653 at 22 = 4 Hence, for d < 0.1653, there are only two equilibrium points at (0,4) and (0.4(4 ~ a).04). The Jacobian ‘matrix is given by : pnzilha =e ose o.s(01te af _| ~o* Rea Win toa m3 4+ open Wises t0e! oe aa pazilte-te) | nosy osm 01-ts) vaste 4 ¥(ierssteat? wap «© ~4 aaoiea ros At 2 = (0,4), the Jacobian matrix is -d+0.1653 0 * -d43 When d > 0.1653, the equilibrium point is a stable node. When d < 0.1653, it is a saddle. At 2 = (0.4(4 — a1), a), the Jacobian matrix is 0 ay -d/Y -d-@8 where B = (227 — 01)(m — hxa2)/s4ma? > 0. The eigenvalues of this matrix are ~d and ~d8. Hence, the equilibrium point is a stable node, At x = (0.4(4~ a3), a2) with az <4, the Jacobian matrix is 0 ey -a/¥ -d-@y where 7 = (224 ~ 2)(km — !103)/1m03 <0. The eigenvalues of this matrix are ~d and —d?7. Hence, the equilibrium point is a saddle. In summary, we have the following three cases: ‘¢ When d > 0.3455, there is one equilibrium point at (0,4) which is a stable node. ¢ When 0.1653 < d < 0.3455, there are three equilibrium points: a stable node at (0,4), a stable node at (0.4(4 — a1),an), and a saddle at (0.4(4 ~ a2), 02) ‘¢ When d < 0.1653, there are two equilibrium points: a saddle at (0,4) and a stable node at (0.4(4 — %),01). as oas Ea 0.05| % 0.1653 Figure 2.38: Exercise 2.31 Figure 2.39: Exercise 2.31 (b) The bifurcation diagram is shown in Figure 239. There is a saddle-node bifurcation at d= 0.3455. At d = 0.1653 there is a type of bifurcation that is not shown in Figure 2.28. A saddle point bifurcates into a stable node and a new saddle is created. (c) When d = 0.1, there are two equilibrium points: (0,4) is a saddle and (1.59,0.0251) is a stable node. The phase portrait is shown in Figure 2.40. The stable trajectories of the saddle are on the z-axis. All trajectories in the first. quadrant approach the stable node. (a) When d = 0.25, there are three equilibrium points: (0,4) is a stable node, (1.5578, 0.1056) is a stable node, and (0.8426, 1.8936) is a saddle. The phase portrait is shown in Figure 2.41. The stable trajectories Of the saddle form a separatrix which divides the frst quadrant into two halves. All trajectories in the right half approach the stable node (1.5578, 0.1056), while all trajectories in the left half approach the stable node (0,4).“4 a-01 4=025 5 5 3 3 * 1 1 q ° 2 4 oT s 4s Figure 2.40: Exercise 2.31(c) Figure 2.41: Exercise 231(d).. 4-05 V7 al\ “| 1 o| oF sas Figure 2.42: Exercise 2.31(¢). (e) When d = 0.5, there is one equilibrium point at (0,4) which is a stable node. The is shown in Figure 2.42. All trajectories in the first quadrant approach the stable node. CHAPTER 2, phase portrait,Chapter 3 3.1 (1) The term |z| is not continuously differentiable at x = 0, but it is globally Lipschitz. ‘The term 2? iy continuvusly differeutiable, but its partial derivative is uot globally bounded. Thus f= 2? + |2| is not continuously differentiable at x = 0. It is continuously differentiable on a domain that does not include 2 =0. It is locally Lipschitz, hence continuous, but not globally Lipschitz. (2) The term sgn(z) is discontinuous at z= 0. Thus, f(z) = 2+ sgn(z) does not have any of the four properties in a domain that contains x = (3) f(z) = sin(z) sgn(z) is globally Lipschitz, This can be seen as follows. If both x and y are nonnegative, we have If (2) — f(y)| = |sin(z) — sin(y)| < |x - yl x >0 and y <0, we have IFC) ~ FW) = |sin(z) + sin(y)| = [2sin(3(x + y)) cos(}(@ — y))] < |e - ul Other cases can be dealt with similarly to conclude that |f(z) ~ f(y)| < [2 — yl for all 2,y. It follows that f is both locally Lipschitz and continuous. It is not continuously differentiable at z = 0 because Timg-sos. {"(2) = +1 while lim, 40- f"(z) = -1. (4) f(2) = ~2 + asinz is continuously differentiable. Hence, itis locally Lipschitz and continuous. $f = 1 + acosz is globally bounded. Hence, itis globally Lipschitz, (5) f(@) = 2 + 2lc| is not continuously differentiable. It is globally Lipschitz because both x and |z| are 80. Hence, it is locally Lipschitz. (6) f(2) = tan(z) is continuously differentiable in the open interval —x/2 < 2 < 1/2. Hence, it is locally Lipschitz and continuous in the same interval. Its derivative sec®(z) is not globally bounded; hence, itis not globally Lipschitz. (7) The function tanh(y) is continuously differentiable and its derivative 1/cosh*(y) is globally bounded; hence it is globally Lipschitz. Clearly, the linear finction y is both continuously differentiable and globally Lipschitz. Hence, f has all four properties. (8) f is not continuously differentiable due to the term [za] in f;. Check the Lipschitz property component, by component. fi is globally Lipschitz as can be easily checked. f is continuously differentiable, but its partial derivatives are not globally bounded. Henc» fy is locally Lipschitz but not globally so. Since both fi and fz are locally Lipschitz, so is J. Since f is locally Lipschitz, it is continuous. f is not globally Lipschitz since fo is not so. + 3.2 (1) 2. Of 0 1 10)*[ psn, best et | *3e7 [geen —t | {8f /22] is globally bounded. Hence f is globally Lipschitz, which implies that it is locally Lipschitz on D, for any r > (2) vo abet |Z EP A] 4546 CHAPTER 3. [of /2z} is continuous everywhere; hence it is bounded on the bounded set D,. Thus f is locally Lipschitz on D, for any finite r > 0. It is not globally Lipschitz since [9/82 is not globally bounded. (3) _ 2 = [_ a2, ght anenen | n(e1,22) is discontinuous at 1 = 0. Hence it is not locally Lipschitz at the origin. This means itis not locally Lipschitz on D, for any r > 0 @ 8 0 1 f@)= [ -n-e(1-a))my ] > or [ -1-2eme —e(1— 2) | [@f/@z] is continuous on Dy; hence f is locally Lipschitz on D, for any r > 0. (9f/8:] is not glotally bounded; hence f is not globally Lipschitz. (5) Let x = [eo 41,2)”. etree) of [ m+ hyes prt) men) (t) aoe (8) 0 0 F(t,2) = ~rair(t) ae | —721(21 + um(t)) —72r1 +um(t)) 0 0 {0f/0z] is continuous and bounded on D, for bounded r(t) and ym(t). Hence f is locally Lipschitz. It is not globally Lipschitz since [8/82] is not globally bounded. ©) 5 f(z) = Ax — BY(Cz) ‘where y\(-) is a dead-zone nonlinearity. The dead-zone nonlinearity is globally Lipschitz. Hence f is globally Lipschitz, which implies that it is locally Lipschitz on D, for any r > 0. © 3.3 For each 20 € R, there exist positive constants r, L1, La, ki, and ke such that \al@) - A@)I < Lala vl, fale) ~ faly)| < Zale ~ yl, [fu@)| Shr, Wfel2)| < be for all x,y € {2 € R| [2 —20| 0 Aw) = { and ¥(2) = (2)|LKall bh ife y and v2 > p, we have Vi(ae2) — A(Wa)| = If de > wand ty to. © 8.8 It can be easily seen that f() is continuously differentiable and MF ()I < ki + kallzll, V2 € B® for some positive constants ky and kz. Apply Exercise 3.6. # 3.9 Due to uniqueness of solution, trajectories in the plane cannot intersect. Therefore, all trajectories starting in the region enclosed by the limit cycle must remain in that region. The closure of this region is a ‘compact set. Therefore, the solution must stay in a compact set. Apply Theorem 3.3. 23.10 aadenetn a =d(-m-Re ty T where R= 1.5, u= 1.2, and the nominal values of C and L are 2 and 5, respectively. Let A= [C, LJ". The Jacobian matrices [8f /Az] and [8 /AA), are given by St [> aha) 2, | 8 _[ - del-h(a) +22] ° ar a BR} @ 0 ~ dr(-21 — Rea +u) Evaluate these Jacobian matrices at the nominal values @ = 2 and L = 5. Let oz Ts Ts $= BX onion [2 =] Then oy of $= HL BL 9 ‘The augmented equation (3.7) is given by fy = 05[-h(z1) +29] fy = 0.2(-1 - 1522 +12) fy = O.8[-H'(xy)r5 +24) - 0.25{-h(z1) + 22] fy = 0.2(-25~ 1.524) ds = O.5[-K'(a)2s +26] ie 0.2(—as — 1.526) — 004(—xy — 1.522 + 1.2) with the initial conditions 21(0) = 210, 22(0) = 220, #5(0) = 24(0) = 25(0) = 26(0) =49 3.41 hem, f= -n +e(1-2f)m Denote the nominal values of ¢ by ¢o. The Jacobian matrices {9 /0z] and [0 /de), are given by Be | aatenes at an |+ 57 [G20 | Lua [E] of of) = Felons S* Blenaa 8° ‘The augmented equation (3.7) is given by Let B sa% ‘Then a =m fy = -nteo(l-ai)n By = — [14 2eoriza)as + eo(l — af) + (1 22)z2 with the initial conditions 21(0) = 210, 22(0) = x20, 3(0) = x4(0) =0 = (a mt #) Denote the nominal values of € by ¢o. The Jacobian matrices (9 /82] and [8 /de], are given by 23.12 n= pt t= -e e(1-23) ~ (m1 - 22 + $23) let ee BE laominas | TA Thea St ot » $0) =0 JE \nominal ‘The augmented equation (3.7) is given by ade eo (21 - 22+ 18 fn = -o(n-at be dy = 4x - Se, » = dn Bn 1 fe = ~en teo(1-2)nu— (1-22 + Let) with the initial conditions 21(0) = 210, 22(0) = 220, 3(0) = x4(0) = 050 CHAPTER 3. 23.18 Stan" '(azy)~ 222, 2 = ba} - cr Let A = [a,b,¢J_ The nominal values are ao = 1, by =0, and cp = 1. The Jacobian matrices [8 /0z] and [0f AX], are given by a [se 7] oi [= 0 "] Re : tb =e 0 af -m Let a [saa lnominal (74 6 a a 8 Fea 3 anna! 8978 ‘The augmented equation (3.7) is given by tan '(23) — 2122 a rs el (aeen ares ig = ata? : conditions 24(0) = 210, 22(0) = 220, 29(0) = 4(0) = 25(0) + 8.14 (a) Let p= [5 ] be the vector of parameters a2 | -(/r) + cosh®(an)) —_ -A/cosh(Azz) or Af cosh* (Az) (1/7) +A/ cosh?(Az2) pat ee Resta oe | -(/r?)t2 — 21/cosh(Ax1) + 22/ cosh? (Aza) ‘The sensitivity equation is given by $= AoS + Bo, $(0)=0 where Ao and Bo are evaluated at the nominal parameters. This equation should be solved simultaneously with the nominal state equation.51 (b) ri = aii + ante —(1/r)r? + zi{tanh(Azy) ~ tanh(Azs)] + za[tanh(Azy) + tanh(Az3)] ~(1/r)r? + r-cos(@) tanh(Az1) — tanh(Az2)] +rsin(@){tanh(Azi) + tanh(Az1)) (1/2)? + 2r(|e0s(6)] + |sin(@))) (frye? + Vr Win (c) By the comparison lemma, r(t) < u(t) where w satisfies the scalar differential equation (I/rju+2v2, u(0) =r(0) = lle(O)lle ‘The solution of this differential equation is ult) = exo(-t/r (Ol +f exnl-(t~0)/raV3 ae = exp(-t/r)|lx(Olls + 2V2r[1 — exp(-t/7)] 3.15 Let V = lle} = 23 +23. Then Azza, devz2 Tra * Isa Vo = 2a +2ené = —22} — 20} + teal Tea 24 Lea, ale < -W +4 < av sedi! (steely <3) < -W+2vIVV (since llalh < vallelle) ‘Taking W = VV = |lz\l2, we see that for V #0, At V =0, we have W(e+h)— WO] _ Wee A) _ 1 ee RT = Fillet + all Similar to Example 3.9 of the textbook, it can be shown that oy ptt im 5 [ Ms(e())a dr = 0 ‘Thus D+ W(t) < -W(t) + V2 for all t > 0. Let u(t) be the solution of the differential equation = ut vi, u(0) =|[2(O)l2 By the comparison lemma, llz(@)ll2 < u(t) = e~*[lx(O)Ila + V2 (1 - e*)52 CHAPTER 3. © 3.16 Let v= 2asint = Ore = 227 + bears Tee? <-W+1 let u(t) be the solution of the differential equation a= -2u41, u(0)=2 Then ' ae oft) ult) = Det [e209 gy = LSE (2) < u(t) =2 +f ara th ‘Thus a kol= Va x (4 #3.7 (a) fo ons da lem Gare = tae = 227 f(z) |g2Pe| < otetns( ns < 201218 (b) Let V(t) = 27(t)z(t) and Vo = 29'zp, then from part (a) we have 2LV(t) < V(t) < 2LV() ‘ide through by V(t), multiply by dt, and integrate to obtain y a [asl Fs [a ito. vy ¥ to =20(t= to) ming ry. Since f(z) is uniformly bounded on S, we have IIf@)- FON SC, VeyeS,C>0 ‘Therefore, whenever ||x — y|| > min;r;, we have Mf(z) - fos _ Take emne {tata ate} «20 Wehne I Se) - fll < Lile~yll, Vay ew Given e > 0, let 6 = ¢/L. For all [lx ~ yl] < 6, we have |If(2) ~ f(y) < L6 = , which implies uniform continuity. # 3.21 The vector y is defined only for x # 0. For x =0, we can take y as any vector with [lylle = 1. Now, for x #0 we have Allie lll oP sign (af) 1 Sp Tel Drew Wiel For p = co, take 1, = argmax|z/ Py { rgmax [24] Then, y7z = |lz\l20 and |lyll 23.22 If [Zeal SL, V(t2) € [a,8] x RP ‘then, from Lemma 3.1, US(t,y) — (t,2)11 < Lille — vl, ¥ (tz) € [0,0] x R™54 CHAPTER 3. Alternatively, suppose f(t,z) is globally Lipschitz. By the mean value theorem Aba) ~ fdts2) = 22) y-2) where = az +(1—a)y and 0 1, we have z = [a+ (1-a)é]z “ yz, and | Fer) (-n4] Gomey — S VEE REE [al Thus on [ere] VreRVte [a0] By letting @ approach 1, we conclude that f (t,2) 2| [eel SL, VIER VLE [a8] which shows that ae) 0 such that Ils(t,2()I| to. Then ‘ ‘ aul = [ slo.a(e) ds+ [Fe ¥(e) - Sle.2(00) ds to te We have l/(s,2(0)I| 0 and pis odd. In the neighborhood of the origin, sign(f(z)) = sign(z). Hence, a trajectory starting near x = 0 will be always moving away from z = 0. This shows that the origin is unstable. When p is even, a similar behavior will take place on one side of the origin; namely, ‘on the side 2 > 0 when a > 0 and x <0 when a <0. Therefore, the origin is unstable. 4.3 (1) Let V(c) = (1/2)(22 + 23). Ven(-n+n2)-2 In the set {\l2[lz <7}, we have |zy| 1/2, V is positive. Hence, trajectories starting in the region V(z) > 1/2 cannot approach the origin. In fact, they grow unbounded. ‘Thus, the origin is not globally asymptotically stable. (8) Let V(x) = 2? Pz = puzt + 2pi2zi22 + pooh, where P is a positive definite symmetric matrix. V = ~2pyazf + 2(pur — Pra ~ Poa) 122 — 2(P22 — Pi2)3 + Higher order terms 5960 CHAPTER 4. Near the origin, the quadratic term dominates the higher-order terms. Thus, V will be negative definite in the neighborhood of the origin if the quadratic term is negative definite. Choosing pyz = 1, por = 2, and pir = 3 makes V(z) positive definite and V(x) negative definite. Hence, the origin is asymptotically stable. Itis not globally asymptotically stable since the origin is not the unique equilibrium point. The set {x3 = 1} is an oquilbrinin st. (4) Let V(e) = 29 + (/2)23. V =~ 22} ~ Qeyre + Qayr — 27 watt Hence, the origin is globally asymptotically stable. #44 (a) Take V(o) = (1/2)(4u8 + Jou} + Jou8) as a Lyapunov function candidate Vi = Iyuyin + Junin + Jawan (Ja = Ja)orwatrs + (Ja — Ji)aywsaiea + (Jt ~ Ja)oywoawrs =0 ‘The origin is stable. It is not asymptotically stable since Vis identically zero. (b) The closed-loop state equation is Suis = (Ja Js)uriag — hres Jain = (Jy — Sioa — kaw Jaitg = (Si = Ja)anien — hows ‘Using the same function V() as in part (a), we obtain V = kya? — baud - ye ‘Thus, the origin is globally asymptotically stable. © 4.5 Let o(z) = VV; gi(t) = OV/Oxy. Similarly 005 _ 8 ov Ox; Bx; Oz; Hence $4 2; Alternatively, suppose Define Vee) FF sbn.0, sO) din +f gt este 0.020) dn set LM onlertas tnaith) din6 av 2 A Fae = red s0h fF Ceram, 0) dn * 00 tek [tenn tnt) de = (21,0, 0} [Ce 90.0.0) dn sek [eran tert) ie = (21,0,--.,0) + 91(t19250,--- Ole” be E124. ta Tnd|G” = n(2) Similarly, it can be shown that cg Fa He Vi +46 Try . any + Ba a= ae | ‘To meet the symmetry requirement, take = 8. V (a) = (az + Bra)ze — (624 + )[(01 + 22) + h(a + 22)] V(z) = Br} + (a — 26)xy22 — Bla +22)h(z + 22) 26 and 6 > 0 yields V (a) = ~ Bat - B(ar + 22)h(e1 + 22) which is negative definite for all zr € R?. Now acy=a[} p]et re = ves [ote a= jetPs where P is positive definite. Thus, V(z) is a radially unbounded Lyapunov function and the origin is globally asymptotically stable. 97 (2)Q¢(2). Choose 9(z) = Pz so that V(2) = (1/2)27Px. We ~2™ PQ¢(2) is negative definite. Choosing P = Q~! yields Yo aesle) V is negative definite in the neighborhood of the origin because yp(y) > 0 for y # 0. Hence, the origin is asymptotically stable. . (b) The function V(2) is radially unbounded. The origin will be globally asymptotically stable if V is negative definite for all x. This will be the case if yps(y) > 0 for all y # 0. (c) The function d» satisfies the condition y6,(y) > 0 for all y # 0. The function 1 satisfies the condition only near y = 0 because 4i(y) vanishes at y = 1. Thus, we can only show asymptotic stability of the origin using the Lyapunov function V(2) = 2" Q-z = 23 + 2ryz2 + 223. 4.7 (a) Let VV(e) = (2). Then, V = need to choose P = PT > 0 such that ~a7 (2) =62 CHAPTER 4. 248 (a) V(0) =0 and V(z) > 0 for x #0. ta [= + 2» sole 22) - - Ba 4 <0, VreR’, 240 (b) The slope of the tangents to the hyperbola is given by de —— dy @—V2P pat Vie +1 On the other hand, fa = («1 +2) Fi lnypervat 1 + 2a + a ~ Rie 2Vie +1 For 2; > v2 2} + 2V2x, +1 > Be? - V2r, +1 Hence Moreover - as +622; + 208 + 40 Sibypertots = + 5B = aay yy Hence, the vector field on the branch of the hyperbola in the first quadrant always points to the right of the hyperbola. (c) Since trajectories starting to the right of the hyperbola do not reach the origin, the origin is not globally asymptotically stable. © 4.9 (a) + 00 a8 |22| + co 00 a8 |x] + 00 (b) On the line 22 = 21, we have at Viz) + Las [21] + 00 Tat + 4.10 (a) aT Pia) +J"@)P2 = pf’ $eox)2 do [2 [Rea] do Pa63 (b) V(z) = f*(z)P/(2) is positive semidefinite. To show that it is positive definite, we need to show that V=0 @& 2=0. Since P is positive definite, we need to show that f(z) = 0 if and only if x = 0; that is, the origin is the unique equilibrium point. Suppose, to the contrary, that there is p #0 such that f(p) = 0. Then, P's -[p"PS(p) + f7(p)Pp] =0 > which is a contradiction. Hence, the origin is the unique equilibrium point. To show that V is radially unbounded, we note that for any € R” 2™Pf(2) 1 Welg 2Ue fe" Pf(2) + f"(@)Pa]<-5, Vee R” Suppose now that [|/(2)llz < as Izll2 ++ 00. Then Wet P FCI WTA < PLLC 5 as iy + oo Tel Tei S Tap Ol > But this is a contredition since z™Pf(2) 1 Sa <- 5) Veter” Welz ~ 2 Thus, as [[2z + 00, the magnitude of at least one component of f(z) must approach co. This shows that V(@) 7 00 as [lll2 > o. (c) We have shown that V(x) is positive definite and radially unbounded. V@)= 7) { P [Le] + [Ze] } f@) $ “2 Since f(z) = 0 # = 0, V(z) <0, for all x € R", ££ 0. Thus, the origin is globally asymptotically stable. «4.11 Since V;(z) is not negative semidefinite, there exists a point zr arbitrarily close to the origin such that Vi(zo) > 0. Let U = {x € B, | Vi(z) > 0}, where B, C D. Since Vi(2) is positive definite, we have Vi(a) > 0 for all x € U. Hence, the origin is unstable. # 4.12 Since Vj (2) is not negative semidefinite, there exists a point zo arbitrarily close to the origin such that Vj (29) > 0. Let U = {x € B, | Va(z) > 0}, where B, CD. Since W(z) > 0 for all z € D, we have Vile) = W(e) + Miz) >0, VeeU Hence, the origin is unstable. © 4.13 (1) Apply Chetaev’s theorem with V(x) = (1/2)(2} - 23). The function V is positive at points arbitrarily close to the origin on the 2;-axis. V(e) = x(a} +23z2) — 22(-22 +23 +2122 ~ 23) = (af tare)? + 23(1 — 22-21-23) For any 0 < c < 1, there is a domain around the origin where I-m-n-2}>e>0 Hence, in this domain, we have V(a) > (@} + xiz0)* + x}64 CHAPTER 4. tp [%e 08 os} 7 5 04 W/ Mage, 02 ° es 0 Os 1 Figure 4.1: Exercise 4.13 (2) ‘The righthand side of the preceding inequality is positive definite; hence all the conditions of Chetaev’s theorem are satisfied and the equilibrium point at the origin is unstable. (2) The system ttm=fla), #2 =2f-23 = f(z) has two equilibrium points at (0,0) and (1,1). The set I'= {0 < 21 <1} {22 > a3} M {22 < 23} is shown, in Figure 4.1. On the boundary z2 = 2, fz = 0 and fy > 0; hence, all trajectories on this boundary move into T, On the boundary 22 = 23, fy = 0 and fz > 0; hence, all trajectories on this boundary move into T. Thus, [is positively invariant. Inside I, both f, and fz are positive. Thus all trajectories move toward the equilibrium point (1,1). Since this happens for trajectories starting arbitrarily close to the origin, we conclude that the origin is unstable, cy wana 2 4, EF vate) av [vd = 58 ‘Therefore, rfid V(a) > fa? taiza +29 w(t 2] ‘The matrix of the quadratic form is postive definite. Hence, V (c) is postive definite for all x, and radially unbounded Vi = 219(t1)a1 + 212 + yd) + Zarate = o(es)(2r22 ~ 2} ~ 2129 ~ 2ey2n ~ 223) +2 ~a(2s)(o} + 22129 +208) +28 ~o(zr)2™Qz +23 where Q [ na ] is positive definite. Since g(z1) > 1 and 2™Qz > 0, we have V <-Gf + Qeize + 228) +23 (af + 2evae + 29) = —(21 +22)? This shows that Vis negative semidefinite. We need to apply the invariance principle. V=0 + 0<-(r +22)? > 02 (e422)? + nt+22=065, 0+ nit x(t) +22(t) =0 > ii(t) +in(t) =0 = x(t Since V (2) is radially unbounded and all the assumptions hold globally, we conclude that the origin is globally asymptotically stable. '* 4.15 (a) The equilibrium points are the roots of the equations O= a9, O=—hi(z1)—z2-ha(zs), O=22—z5 2=0 3 2=0 + h(n) =0 + m=0 Hence, there is a unique equilibrium point at the origin. (b) V(a2) is the sum of nonnegative terms; hence V(2) > 0. To show that it is positive definite for all x, ‘we need to show that V(z) =0 + 2 Since yhi(y) > 0 for all y # 0, the integrals ff" hi(y) dy and fo? ha(y) dy vanish only at x; = 0 and 23 = 0, respectively. Hence, V(z) is positive definite (©) V = hy(a)a2 + x2[—hy (21) — 22 — ha(zs)] + ha(23)(n2 ~ 23) = —2} — asha(as) V (a) is negative semidefinite for all 2, but not negative definite because V(z) = 0 when 2» = zy = 0 for any 21. We apply the invariance principle. x2(t) =0 and z(t) =0 > hi(zi(t)) =0 > a(t) =0 Hence, the origin is asymptotically stable. (A) To show global asymptotic stability we need V(2) to be radially unbounded. This will be the case if the integrals f° hi(y) dy, 1 = 1,2, tend to infinity as |z| -+ oo. © 4.16 Let V(x) = (1/4)2f + (1/2)z3. V = afr - stm ~2$ = V(z) is negative semidefinite for all z. =0 > qw= aa(2) > a(t) 50 By the invariance principle, the origin is globally asymptotically stable. 2417 (a) —9(21) ~ h(zi)e2 22, At equilibrium, 22 =0 & (i) +A(2;)z2 = 0 > 22 =0& g(r) =0 ‘Assume that 9(z1) = 0 has an isolated root at the origin. Then, the origin is an isolated equilibrium point. () With V(z) = JG" ou) dy + 323, we have V(e) = (ar) — zag(ea) ~ h(as)23 = -(ay)3 <0 ‘Assume that h(x;) > 0 ¥ 21 € D (a domain that contains the origin). Then, V <0 and > g(zi(t)) =0 = n()=066 (CHAPTER 4. Hence, by LaSalle’s theorem (Corollary 4.1), the origi is asymptotically stable. (c) With V(z) =} [ra + JG Aly) dy]? + JZ" o(y) dy, we have V = [ae [a0 de e+ 2 )8) + ote) [oo [mon ay] on) ~ateren + eon + nae) = ater) [ma ay Assume that o(21) fG" A(y) dy > O and {* A(y) dy #0. Then , 7 <0 and V=0> g(x(t)) =0 + x(t) =0 + 22(t)=0 Hence, by LaSalle’s theorem (Corollary 4.1), the origin is asymptotically stable. ‘+ 4.18 The system has an equilibrium point at y = Mg/t and jj = 0. Let 21 = y— Mg/k and 22 = 9. Aay- ay — 2 y|n0] fa- Shey Sealey Take V(2) = a2} +623, with a,b > 0. V(z) is positive definite and radially unbounded. Qe Derg _ 2bee ) ayn -- FA} 4 V(e) = -c12} — coz} ao] <0, Var fam, iy apt — BE sttea| Ve) =2(a- ‘Taking a = k/2 and b= M/2, we obtain Moreover, ve = n() 20> n= Using LaSalle's theorem (Corollary 4.2), we conclude that the origin is globally asymptotically stable. ‘* 4.19 The equation of motion is M(a)G + C(a,4)4 + Da + 9(@) where M(q) = M7(g) is positive definite, Mf — 2C is skew symmetric, D = D? is positive semidefinite, 9(q) = 0 has an isolated root at ¢ = 0, 9(g) = [AP(q)/8q]7, and P(q) is positive definite. The 2m- dimensional state vector can be taken as 2 = | $ (a) Let u=0 and V = }g"M(a)d + P(Q). V is a positive definite function of x. Lewy 1 +-¢°D4 V mgt Mat 5a? Md + 0M = Fa (at — 204 GP DG P99 Hence, the origin is stable. (b) With u = —Kad, we have V = —47 (Ka + D)g <0. Moreover, og > g(a) =0 >67 Hence, by LaSalle’s theorem (Corollary 4.1), the crigin is asymptotically stable. (©) With w= 9() — Kp(q— 9") — Kad, the equation of motion is given by M(q)G + C(q,4)¢ + Dg + Ky(q— 9") + Kad = 0 ‘There is an equilibrium point at q = 9" and @=0. Take an [ a | ast 4 Je Met JeT Kyeis postive definite. 1 Mae Ma a gt = ME 20) ga — A at De Kt He DESO ve0seé 0> Ke=0 > Hence, by LaSalle’s theorem (Corollary 4.1), the equilibrium point (q°,0) is asymptotically stable. '* 4.20 According to LaSalle’s theorem, z(t) approaches M as ¢ -+ oo. Equivalently, given ¢ > 0 there is T > 0 such that Bh le@®)-vlT Choose ¢ so small that the neighborhood V(p,2«) of p € M contains no other points in M. Claim: ||x(¢) ~ pl] <¢, for all ¢ > T, for some p € M. ‘The claim can be proved by contradiction. At t =t; > T, let p, € M be a point for which [lz(t1) — pill ty such that |[e(¢s) ~ pil] = ¢. Let p # ps be any other point of M. Then l2(a) ~ rll = lla(ta) - pr +m — Pll 2 lip: ~pll-llx(te) - pil] > 2e- > inf let - olde ‘The last statement contradicts the fact that infycas z(t) ~ yl] < ¢ for all t > T', which proves the claim. Since this claim is true for any, sufficiently small, « > 0, it is equivalent to 2(t) > p as t -+ 0, for some peM. 2421 @ waa % V(z) =04 VV(z)=084=0 —(VV)T(VV) <0 Hence, V(x) =0 if and only if z is an equilibrium point. (b) Every solution starts in a set. with ¢ > V(zo). Since V <0 in 2, the solution remains in Me for all £0. Since M- is compact, we conclude by Theorem 3.3 that the solution is defined for allt > 0. (c) By LaSalle’s theorem, x(t) -+ M = {p1,...,p,} a8 t + 00. Since the points pi,...,p+ are isolated, we ‘conclude from Exercise 4.20 that 2(t) + pi as t+ oo for some py € M. 24.22 Sufficiency: Suppose there is P = PT > 0 such that PA+ATP =-CTC68 CHAPTER 4 Let V(z) = 27 Pe. V(a) =-27C% cz <0 V(z) = 0 > Cx(t) = 0 > Cexp(At)zo since the pair (A,C) is observable. By LaSalle’s Theorem (Corollary 4.2), the origin is globally asymptotically stable. Necessity: Suppose A is Hurwitz. Let 2 =0 P= [* en(AT 0" expat) dt P is symmetric and positive semidefinite by construction. To show that itis positive definite, suppose itis not so. Then, there is x #0 such that 2" Pz =0. Consequently o>2=0 f aT exp(AT1)CTC exp(At)ex dt =0 > Cexp(At) Hence, P is positive definite. Similar to the proof of Theorem 4.6, it can be shown that P satisfies the equation PA+A™P=-CTC ‘and that it is the unique solution. 24.23 (1) Let V(z) = 27 Pr. V(z) =27[P(A- BR" BTP) + (A- BR“ B™P)" Piz Using the Riccati equation, we obtain V(@)= 27(Q+ PBRB?P)z Q>0>Q+PBR*BTP >0 Hence, ¥(z) is negative definite and the origin is globally asymptotically stable. (2) When Q = CTC, we can only conclude that V(2) is negative semidefinite, But V(@)=0 + 27(0(Q+PBR-B™P)x(t) =0 > Cx(t) =0 and RBT Pa(t) Due to the second identity, the state equation simplifies to Ar — BRB" Px = Az and it solution is given by 2(t) = Cexp(At)z9. Thus Cexp(At)xo = 0 29 =0 since (4,0) is observable. By LaSalle’s theorem (Corollary 4.2), we conclude that the origin is globally asymptotically stable. 24.24 Ve) = Fert) - HGR renee) (L)© Substitute for [9V/Az]f using the Hamilton-Jacobi-Bellman equation. ev, = av\? Ve) = -ata)- (4) Fowrrecra) (F) If q(e) is positive definite and & > 1/4, we conclude that V(2) is negative definite; hence, the origin is asymptotically stable. If q(x) is only negative semidefinite and k > 1/4, we can only conclude that V(z) is negative semidefinite. But ov r V(ay=0 = ate) ana oe) (2) =0 + #=s() Since the only solution of = f(x) can stay identically in the set {q(z) = 0} is the zero solution, we see that aa) =0 + 2()=0 By LaSalle’s theorem (Corollary 4.1), we conclude that the origin is asymptotically stable. The origin will be globally asymptotically stable if all the assumptions hold globally and V(z) is radially unbounded '* 4.25 Since (A,B) is controllable, the controllability Gramian W = fy e~4*BBT [e~4t]" dt is positive definite. Hence, W~' is positive definite. Note that AW + WAT f {ae@ BB? fe? ten BBT [eA] AT} at lo 0 0 = LF {eB [e-44]"} at = BBT eA" BBT fe]? Hence, (A- BK)W + W(A— BK)? = AW +WAT — 2BBT = —e~4" BBT [e~47]" — BBT With V(2) = 27W~12, we have V(q) = 27W-(A- BK)W + W(A- BK) Wt Two {eATBBT [eA] + BBT} W-'r <0 Hence, the origin of ¢ = (A ~ BK) is stable; all eigenvalues of (4 — BKK) satisfy Re[A] <0. Now we want to show that Re(A} < 0. Let 1 have Re{A] = 0 and let v be the left eigenvector of (A — BK) corresponding tod. Then v*(A— BK) = Av" and (A- BK)Tv = tv. We have : (A- BKW + W(A- BK)? = -e-*"BBT [e~*"]" — BBT v*(A- BK)We + v"W(A— BK)" v = -"e-* BBY [e~4"]" v — BBTy Wt Mut We = oe BBT [eA] v — BBM = 2(Re[A])v* We = -v*e-AT BB? [e“AT]" vy — v* BBT ‘Thus Ref] =0 + v'BBTy=0 3 vB =0 which contradicts the controllability of (A, B). ‘Thus, all the eigenvalues of (A — BK) have negative real parts.70 CHAPTER 4. + 4.26 (a). Suppose z = 0 is an isolated equilibrium point. Clearly z = 0 is an equilibrium point. We can show that itis igolated by contradiction. Suppose itis not isolated. ‘Then there is # # 0, arbitrarily close to 0, such that f(z) = 0. Define = T-1(2). Then, f(z) = [@T/Oz]~ f(z) = 0; that is, 2 is an equilibrium point. By continuity of T-1(.), we can make 2 arbitrarily close the origin, which contradict the fact that the origin is an isolated equilibrium point. Clearly the argument works the other way around. Hence, 2 = 0 is an isolated equilibrium point if and only if z = 0 i an isolated equilibrium point. (b) Suppose x =0 is a stable equilibrium point. Then, given ex > 0 there is 6; > 0 such that We(0)Il < 61 > Hla(@i 0 there is r > 0 such that Well 0 such that Hx(O)|] <6 => lle(ll 0 such that, Well 0 there is T, > 0 such that |[z(¢)|| < ey for all t > Tj. By continuity of T(-): Given ¢2 > 0 there is r > 0 such that Ilell 0 such that lel Te = [ell T Hence 2(t) 40 ast +00 and z = 0 is asymptotically stable. The opposite direction of the proof is done similarly. Now = is stable ¢ z = 0 is stable is equivalent to is unstable ¢ 2 = 0 is unstable © 4.27 (a) The equilibrium points are the roots of the equations O=-mey+1, O=z25-22, 0=23(1-25) From the third equation, 23 = 0 or zs = 1. The first equation cannot be satisfied with 23 = 0. t3=1 3 mal 4+ mn =2=1nm Hence, there is a unique equilibrium point at (1, 1,1). (b) 0 -% -2 0-1-1 sfu a on s{i-11 y [0 0 2-33 Jay LO 0 = The eigenvalues are ~1 and (~14 j¥3)/2. Herce, the origin is asymptotically stable. The third state equation has equilibrium at zs = 0. Starting with the initial condition zs(0) = 0, we have 23(¢) = 0. Then, 4, =1 and z(t) grows unbounded. Thus, the equilibrium point is not globally asymptotically stable. © 4.28 (a) Oma, = (tz ~ 129 + (22-14 2})e2 Substitution of 21 = 0 in the sccond equation yields ~42(1 +93) Hence, the origin is the unique equilibrium point. (b) zl tenes (iste, 26} one -142 ||, [ Hence, the origin is asymptotically stable. (©) Let V(2) = 2122 V(z) (2122 ~ 1)z2} + (2129-14 23)2429 — 222 ae + onze V(a)| = 423 >0 which implies that Tis a positively invariant set. (A) The origin is not globally asymptotically stable since trajectories starting in I’ do not converge to the origin. ‘© 4.29 (a) The equilibrium points are the roots of the equations O=a-2}+22, 0=3r-2 =n > 2(4-23)=0 ‘The equilibrium points are (0,0), (2,6), and (—2, ~6). (b) Of _[t-3e} 1 az 3-1 os 1a 4s of aL} b] = t-teos teas ‘The equilibrium point (0,0) is unstable (saddle). 11.29, -0.71 ua , _ { 55 Al => ¥412A+8=0 5 2 ‘The equilibrium point (2,6) is asymptotically stable (stable node). Flaw l 3 4]2 CHAPTER 4. ‘The equilibrium point (—2, ~6) is asymptotically stable (stable node). (©) ter A= [22 3, ] and P be the solution of PA + ATP = ~1, Using Mata, P is found tobe p =| 00088, 177 crm oer the region of attraction of (2,6), shift the equilibrium peint to the origin via the change of variables ‘The eigenvalues of P are Am ) = 0.7266 and Amin(P) = 0.0442. To estimate =n-2, %=2-6 ‘The state equation in the new coordinates is given by 11d, + in - 627 31 — aa We use V = 27P# as a Lyapunov function candidate. Tae derivative V is given by Vo = -a2— 2p + pnte)(6+21)89 S_ ~l813 - 12(p1121 + prata)24 - 2prathte SWB + 12y bh + efi + pall < -(1-24r 0.17717) |I3, for lle Sr ‘Taking r = 0.4, we see that V(2) is negative in {|lzll FPR Choosing csmall enough so that the closed surface {V(2) ~ ¢} is contained inthe ball {x2 < 1/(4AiIPIl)}, wwe see that [9V/Oz] f(z) > U for all z € {V(z) = c}. + 4.34 Proof of Lemma 4.2, + Since a, (.) is continuous and strictly increasing on (0, a), for each given y € [0, a1 (a)) there is a unique such that ay(z) = y. Define x = a~(y). The function a~"(-) is continuous, vanishes at the origin, and is strictly increasing. Hence, a-*(.) is a class K function. ‘ From the previous item, aj"(.) is a class K function. Moreover, since a3() is a class Kay function, a51(.) is defined on (0,00) and a5"(r) ~¥ 00 as r + 00. Hence a5"(-) is a class Kg function Let a(r) = an(a2(r)). Then (0) =0; a(r) >0, forr >0 72> 71 = @a(r2) > a2(r1) = ar(aa(r2)) > ax(a2(rs)) ‘Therefore a(:).is strictly increasing. It is also continuous. Hence, it is a class K function. « From the previous item, a(r) = as(aa(r)) is a class K function. Moreover, itis defined for all r > 0 and 1400 = au(r) + 00 + afr) ++ 00 Hence, a3(a4(:)) is a class Ke function. + For each fixed s, 6(a2(r),§) is a class K function of r. Thus @y(8(a2(r),3)) is a elass K function of 1. Now, for each fixed r, 8(az(r),) decreases as s increases. Hence, ay (3(aa(r),8)) decreases as increases. Moreover, ay (3(aa(r),§)) + 0.as 8+ 00. Hence, ay (3(aa(r),s)) is a class KL function, © 4.35 Ifry > ra, we have ry +12 <2ry. Hence, a(ry +12) < a(2r) < a(2r) + a(2r2) If ra > ry, we have 1 +72 < 2ra. Hence, a(n: +12) < a(2r2) < a(2r) + a(2re) ‘Thus, the inequality ary +2) < a(2r1) + a(2r2) is always satisfied. '¢ 4.36 This exercise is already dealt with in Example 4.18.76 CHAPTER 4. © 4.87 (1) Let V(z) = (2? +23). = a} + alt)nrze + a(t)zy22 ~ 223 < —2} ~ 223 + 2s] [2a] ~[ ELTA 2 ][Ef]s-92et+e ‘where 0.382 is the minimum eigenvalue of the 2x 2 matrix. By Theorem 4.10, the origin is exponentially stable. (2) Let V(z) = 3(2} + 29) v o} — 23 2} + a(t)z22 — a(t)ziz22 — 20} = By Theorem 4.10, the origin is exponentially stable. (8) Take V(z) = put} + 2pi27122 + pra}, where pi1p22 — pip > 0, and let k be an upper bound on a(t)]. V = —2praz} + 2lpis — poo — a(t) pr2}eiee + 2lpr2 — at)paa)2} ‘Take pi = P22 = p > 1 and pr2 = 1. 7 V < 22 + 221} [2a] — 2(2p— 1)23 = ell ies roe el Choose p > (1 + k#/4)/2 so that 4(2p ~ 1) ~ ? > 0. ‘Then, the 2 x 2 matrix is positive definite and, by ‘Theorem 4.10, the origin is exponentially stable. (4) Take V (2) = prix] + 2pr2m22 + pore}, where pi1p22— Piz > 0, and let k be an upper bound on |a(t)|- V = ~2lpua ~ a(t)pi2}e} + 2[-pr2 + a(t)pan — 2pralerz2 — 4paaz} ‘Take py2 = 0 and pap = 1 r V <—2puzt + 2k|2y| [zal - 423 = — (a | [ ou = ] [ (a Choose pir > #?/8 so that Spr: ~ A? > 0. Then, the 2 x 2 matrix is positive definite and, by Theorem 4.10, the origin is exponentially stable. + 4.38 (a) Using the bounds on L(t), C(t) and R(t), we can show that 2hy 241) 2h hs + Ee S8O+ Row St Re 202 02 Be = RO $B Using the upper bounds, it is clear that V(t, z) is decrescent. If we try to use the lower bounds to show that V(t,2) is postive definite, we will have to restrict the corstants to satisfy V(t.2) =27 Pe =o! 2 mo RO + iy 1 Jee [? j >7 It can be shown that the minimum eigenvalue of P is given by Amin(P) = $(a- Va? = 4) where 2 a=R+5 It can be easily seen that there are positive constants ¢, and cy such that a? ~ 4 > ¢1 and Amin(P(t)) > e2, for all t > 0, which shows that V(t,) is positive definite. (b) von=-Zfra(h-9)+8-Ha-3(i+B) a p(L_¢ te ok vk(h-9) +b LR 14 Bee for all ¢ > 0, for some positive constants c, and c2. Then : dep 2a Vea) <- Bat- Bat Hence, V(t,2) is negative definite. This shows that the origin is uniformly asymptotically stable. By linearity, it is exponentially-stable. Alternatively, we can conclude that the origin is exponentially stable by noting that V (t,2) satisfies the conditions of Theorem 4.10. 24.39 (a) Since g(t) > a>, 1+ag(t)—a? >1+0a-a? >1 Hence V(t,2) > (asin; +29)? +1 — cosa, which shows that V is positive definite in the region [z1| < 2x. Since g(t) < 8, V(t2) < $(asinay + 22)? + (1+ a9 +0%)[1 ~ cosas] which shows that V(t,2) is decrescent. (») Vo = -[p(t) ~ acosz}z3 — asin? 2; ~ a®(1 ~ cosz;)z2 sin, + a9(t)(1 — 0082) ~(a~ a)23 +.ay(1 ~ cosas) — asin? 2; + O(\z|*) ~(a~ a)z3 ~ a(2 ~ 9)(1 — cos) + af2(1 ~ cos) ~ sin? 21] + O(| =I) ‘The term [2(1 ~ cosa) ~ sin? 23] is O(|z1|*). Hence (a — a)x3 — a(2 — 7)(1 — cosa) + O(llal*) (c) The preceding inequality shows that V is negative definite, since near the origin the negative definite quadratic term dominates the cubic order term. Thus, the origin is uniformly asymptotically stable.78 CHAPTER 4. 24.40 (a) P(t+T) = exp(Bt+ BT)O(0,t+T) = exp(Bt)exp(BT)#(0,7)8(7,t+7) = exp(BY8(T,t +7) since (0,7) = [®(T, 0)]-? = lexp(BT))"2 AE+T) = AW > &(T,t+ 7) = 80,1) > P(t +7) = ex(Bt)4(0,t) = P(t) (b) PY) expl(t—7)B|P(r) = 4(6,0)exp(-Be) exp(Bt ~ Br) exp(B7)4(0,7) (6,0)8(0,7) = &(t7) (©) P(t) and P-¥(¢) are continuous functions of t. Hence, they are bounded on [0,7]. Since they are periodic, they are bounded for all t > 0. Let ||P(#) < 41 and |[P-"(t)l] < ka. Then H&C,7)Il S Ruball exp[(t - 7) Bi) and Hlexp[(t — 7) Bi] < iy hall®(t,7)IL Hence, ||®(t,7)|| is bounded by ke~7(*-") if and only is B is Hurwitz. 4.41 @) 2 2eyme + 3t +2 — 32y ~ 2t + I)az.= 24 3t +23 AWt+1) = ay ae, at)= [1] eatin (b) Recall from the discussion at the beginning of Section 4.5 that to show asymptotic stability of a solution wwe shift it to the origin and then show asymptotic stability of the origin. Let 21 = 21 —t and 22 = 2 ~ 1. ‘Then Asm, = 2-2-2 We need to show that the origin = = 0 is uniformly asymptotically stable. af 0 1 1422, -2+2z | a(1+2%2)2?z <—a2"s Ta +27 S(a)z+ (272) = where we used the property 27 5(2) © 4.43 The closed-loop system is # = f(z) - oG(2)G?(z)Pz. Use V(z) candidate 0. By Theorem 4.10, the origin is globally exponentially stable. 27 Px as a Lyapunov function V = 22" Pi = 22" Pfc) - 202" PG(2)G"(e)Px < ~727 Px — W(2) < -Wie) By Theorem 4.10, the origin is globally exponentially stable.79 9 4.44 Let V(x) = }(2} + 23). Vo = Wah + ayza+eu(ef + of) sint ~ area — 23 + role} + 29) cost ~(2} +23) + (c? +23)(x1 sint + x2 cost) lle lf + lal Vint)? (cost)? = lla + lel G=rIB, Velasr, for any r <1 < < Hence, by Theorem 4.10, the origin is exponentially stable. Since V(2) can be estimated by the set {Ill < r} for any r <1 4ilzllZ, the region of attraction © 4.45 (a) Use V(z) = 3(c} + 23) as a Lyapunov function candidate. V = ai[h(t)zo — o(t)xi] + 22[—A(t)xi — g(t)23] = —9(t)(et + 23) < —Kat +29) Hence, the origin is uniformly asymptotically stable. Since all assumptions hold globally and V() is radially unbounded, the origin is globally uniformly asymptotically stable, which answers part. (c). (b) The conditions of Theorem 4.10 are not satisfed. Let us linearize the system a9 = £00 [ 0 ae =A) 0 ‘Use V(x) = }(2} + 23) as a Lyapunov function candidate for the linear system. V = zyh(0)r2 ~ eah(t)z1 = 0 ‘This shows that solutions starting on the surface V(2) = ¢ remain on that surface forall t. Hence, the origin of the linear system is not exponentially stable. This implies that the origin of the nonlinear system is not exponentially stable. (4) No. + 4.46 Linearization at the origin yields the matrix Of] _[ 143242} —14+22122 -fa4 Bel,o | 14222 -1+2}+3e} y= | 1 1 whose eigenvalues are ~1 + j. Hence, the matrix is Hurwitz and the origin is exponentially stable. Conse- quently, itis asymptotically stable. © 4.47 Let V(2) = }(b23 + 023). v 0(t)(a1 — azz)? — acy(t)2$ < —bdo(n — a2)? — acyor It can be verified that Wa(z) is positive definite for all z. Hence, by Theorem 4.9, the origin is globally ‘uniformly asymptotically stable. Linearization at the origin yields the linear system ~o(tn + ad(t)ys, ta = b9(t)an — add (tye [2]-6 [4] 0, 42 = d(t)ar — O(t)(1 + ab)z2 which has the solution 2;(t) = constant. This shows that the linear system is not exponentially stable. ‘Therefore, the origin of the nonlinear system is not exponentially stable. a ‘The invertible change of variables results in the system80 CHAPTER 4. 4.48 Let Ay = $£(0) be the linearization of # = f(z). To find the linearization of # = A(2)f(2), set 9(2) = A(z) (2). Then ay p ste) Hence 89%) — pag) Of, , Ob _ a Of FEO = norseo) + Forno) = no-SE0 Hence x= 2260) = WA, Since h(0) > 0, Ay is Hurwitz if and only if Az is Hurwitz. By Theorem 4.15, £= f(z) is exp. stable > Ay is Hurwitz ¢ Ay is Hurwitz <@ 4 = h(z){(2) is exp. stable 4.49 Equilibrium points: at, +b > 2, =d/a 0 = ~cty +24(0~ i122) Substituting 2 = b/a in the second equation, we obtain __abja 22 FS Bibfae ‘Thus the system has a unique equilibrium point at (21,22). Let h=n-h, mon-% It can be verified that th=—an, te =—le+ Bly +21)"]ye + na — Beoy? — 2821720 3 1 7, 1 Vaiite2e+ iu, 1>0 Vo = ~avf—fe+ Stun + £1)? lvs + 1(a~ 282152) ive — Brzavive — ay S_ ~ayf — rev} + -realua! val +-reav? Ival — ay, ¢1 > 0, 2 >0 = 99-22 aa ial |” pat ]- [ary ai au 2a | hal | %L tool JL tool | Lal where a=| a2 3] qo=[ 32/4, ~re2 =yei/2 ye/4 —yea/2 ye/4 Choosing + < min{ac/2c}, 3ac/4c3} ensures that Q; ané Q2 are positive definite. Then — Syz_ Mya _ Bye Vis -gu-yu- Gu 1 2414 < ~min(o.ch (Fat + 38+ jut) —ap¥, a9 = min{a,c} which shows that the origin is globally exponentially statle.81 {ey Srce the origin of the linearization is exponentially stable, by Theorem 4.13, the origin is exponentially stable. Therefore, there exist positive constants fy, 7, and ¢, Such that Lal se“ Ula(to), Hel to)ll Ser ‘Suppose that cr < ||2(to)|| 0 such that A(c,T — to) = cx. Thus { Alllz(to))ht—to), for to which implies that Nzx(@)Il < B(x (¢o)|], Oe" e- ME), W E> to, Ver < fhx(to)Il 1. Define a(r) = het") A(r,0) Clearly, a(:) is a class K function and lz(®Il < a(llz(to)[)e""*“, Vt > to, V fx(Co)II toy W [la(to}ll oo. + 4.51 Using the inequalities Auli" $V 2 >0 wwe see that as long as V > kop", we have [lz] > wand vc BY = Vist) ceO/-V (20)82 CHAPTER 4. Hence, w waco < (LEROY ¢ (Le-tome-rv ia) (detonator) @r enUis/ehae zB ‘The foregoing inequality will hold over the interval [to, ts +] during which V > kou®. For t > to +T,, we have veto s (MEO) < (st) hy )jx(to)|| = ke~%*-)|hx(to) |] iw © 4.52 Let © = maxw,(z} 0 such that for to < t < to-+T, 2(6) will be in A, while for t > to+T, it will be in {V(z) to +T, 2(t) € {V(z) < c}. Since positive definite functions are bounded from below and above by class K functions (Lemma 4.3), any x in {V(z) <¢} can be bounded by a class K function of .. In general, the function will not be aj! (aa(1)) as in (4.43). # 4.58 The proof is exactly the same as that of Theorem 4.18, Notice that the argument used in the proof of Theorem 4.18 holds outside a ball that contains the origin. + 4.54. (1) The system is not input-to-state stable since with u(t) = ¢ > 1 and 2(0) > 0, 2(t) + 00 as tro. (2) Let V2) = 32? v a purt— 28 <—at, V|2|> va By Theorem 4.19, the system is input-to-state stable. (3) The system is not input-to-state stable since with u(t) = and 2(0) > 0, z(t) ++ 00 as t + 00. (4) With w= 0, the origin of = 2 is unstable. Hence, the system is not input-to-state stable. © 4.55 (1) Take V = }(2? +23). Vo = ~(}+23)+22u < —Hlelld + lllelul (1 ~ Olaf ~ ala + Illa < —(1 — Oil, Hells > ful/@ where 0 < @ < 1, Hence, the system is input-to-state stable. (2) Take V = (1/4)zf + (1/2)23 Va -2f — 2} + 29u = ~2f - (1 - 0)2} — 603 +290 < -2f — (1-O)z3, for |2al > [ul/0 where 0<@ <1. When |z2] < 1/8, we have Vs -(1- Ot — bc - 23-+u2/0 < (1 - O24 — a8, for esl > Viu/o83 ‘Thus, V < -(1 = 6)[z4 + 23] for all [jz\loo > p(|ul), where p(r) = max {r/6, VP} Hence, the system is input-to-state stable. (8) Take V = 23 + 22,29 +229 + 2f V = ~22f — 203 + 22yu + dou < —22f — 209 + 2Jeul ful + lao [ul We have zal ful = 2f2n] [2u) < |e2)? + 2ul? Using Young's inequality, we have le ful = fxs] [2ul < Jaal* + (2hul) > ‘Thus V sof a3 + (alu? + sul? Define the class K function ¢(r) = (2r)*/3 + 4r?. Then, Vs -af - 23 + (lu) Since 2{ + 2} is positive definite and radially unbounded, by Lemma 4.3, there exists a class Ky function ‘ag such that ( e) V < ~az(\lzI) + ¥(lul) = (1 — 8)as\(llz||) — Bars (llx|l) + v(Jul) < —(1 - @)as(llzl)), ¥ lel 2 ag” where 0 <@ <1. It follows from Theorem 4.19 that the system is input-to-state stable. (4) With u =0, the system = (z1-22)(@}-1), a2 = (a1 +22)(z} -1) hhas an equilibrium set {27 = 1). Hence, the origin is not globally asymptotically stable. It follows that the system is not input-to-state stable. (8). The unforced system (with u = 0) has equilibrium points at (0,0), (1,1) and (—1,—1). Hence, the origin isnot globally asymptotically stable. Consequently, the system is not input-to-state stable (6) Let V(z) = Hat +29) wa} na ty +2 2h trae wa} a afta tan (1 ~ O)xt — (1 ~ O)x§ — 6x — Ox + [2] |lullao + [22] lIullocy 0 (Hie —(1~ 0)2f — (1-6), for fra] > For [2] < (Ilo/8)"/?, we have v ils)” Mw af 29+ [ea lhulleo + alley”? ~(1=0)2} ~ 24 ~ 023 +o las + Hes u Let p,(r) be the largest positive real root of the pclynomial equation 3 tryt+ ba a0 12084 CHAPTER 4. Tt can be seen that py(r) is a class Kp function of r ard py(r) > 7/9. Hence, for |z2| < (|lullec/8)"/8, we have V < -(-8)2f — 23, for [21] > pr({lullo) For [z1| < ulleo/®, we have A 4 Mello Vg matoag LO tal tls rs (Iu = -a] — (1 0)x} — 025 + [22] [ulloo + Let p2(r) be the largest positive real root of the polynomial equation 2 eee oy! + ryt 0, r20 It can be seen that p2(r) is a class Keo function of r and pa(r) > (r/8)'/%. Hence, for [21] < llulleo/8, we have . V<-2}—(1-@)2, for [22] > pa((ulleo) Define p(r) = max{or(r), pa(r)}. Then, p(r) is a class Keo function of r and Vs -(1-0)z}—(1-8)2§, ¥ [lrll-c > piilul From Theorem 4.19, we conclude that the system is input-to-state stable. (7) Let V(2) = $23 + fo" oly) dy + 323 Vv fe + o(21)|(~21 + 22) + 22[—21 — o(t1) ~ 22 +4] maf — n0(21) — 23 + zou < —[Irlf + [lellalul <-G—O)llzlla, leila > lul/ where 0 <@ <1, From Theorem 4.19, we conclude that the system is input-to-state stable. © 4.56 The system 2, = —2} +22, with 2» as input, is input-to-state stable. The system #y = ~23 has an ‘asymptotically stable equilibrium point at the origin. It follows from Lemma 4.7 that the origin of the full system is globally asymptotically stable. © 4.57 ¥(u) < |Y(u)], where |¥(u)| is continuous and positive definite. It follows from Lemma 4.3 that there is a class K function such that [2}(u)| < o((Iull). Hence Vis ~~ #)as(ll2ll) — Gas(He]l) + o(\lull), 0<@<1 S (= Aas(llell),¥ lel] > ag? (a) Now we can apply Theorem 4.19 to conclude that the system is input-to-state stable. 4 4.58 Given € > 0, find c1 > 0 such that 7(¢1) < ¢/2. Since lims-s.0 u(t) = 0, given €1 there is T, > 0 such that llu(¢)|| < ¢1 for all t > Ty. Take to > Tj. For t > to, we have Hz(E)ll < B(llex(to) Il, t = to) + ex) < Ble, for some c > 0, where we used the fact that z(t) is bounded. Since 8(c,t to) -+ 0 as t + 00, there is T» > 0 such that 8(c,t ~ to) < ¢/2 for all t > Ts. Thus, to) + 6/2 Ill Se, ¥t> T= max{T,T2} which shows that lim 20 2(t) = 0.# 4.59 It follows from Exercise 4.58, but an indepenent solution is given next. Let V(z) = 24 Vez? +r Starting from any time to, we have et < e~® for all t > to. Hence , ey Vs -a8 + [ote <-(1—-6)2", V|e|> () where 0 <@ < 1. It follows from Theorem 4.18 that there is a finite time t; > fo such that lel (sy. on Given 0 < ¢ <1, take tp = In(1/#e*). Then there exists a finite time T such that ||2(t)|| < ¢ for all t > T. ‘This shows that (0) converges to zero as ¢ tends ‘o infinity. * 4.60 Applying the non-global version of Theorem 4.18 shows that for any 2(to) and any input u(t) such that M2(to)ll 0 such that [on (232)] coca, veer Consequently ayiy $-(1-B)z}, Vi>T = ri(t) 40 ast +00 12 + is input-to-state stable. Hence, for any bounded u(t), z2(t) is bounded. 100) fa) [(=232)'-)] a lei(01 < len) +f lex(o)| dr (b) The linear system = a(t) = By Gronwall-Bellman inequality Iei(t)| < Ie ()let Let T be as defined in part (a). nih S-(1- Alef, VERT > In (OLS z(D}, VDT > [er(O1 086 CHAPTER 4. which shows that 2; (t) is bounded. (c) With u(t) = 1 and 22(0) = 1, we have 2(t) = 1. Then oo [(@=3)' =] =0 5 m()=2(0)=0 (d) Suppose the system is input-to-state stable, then there exist a class KC function f and a class K function ‘7 such that Uo < AUleO}8) + (apllue) , ¥Ez0 Applying this inequality tothe solution of part (¢), we obtain Nx(€)l| < 5(\l2(0)|1,#) + (2), VEO Since 6(||2(0)s#) tends to zero as t tends to infinity, there exists a finite time T; such that lle(#)l| < 27), Ve 2 Tr But |[z(0)|lz = Va? FI can be made arbitrarily large by choosing a large enough. This is a contradiction. hence, the system is not input-to-state stable. ‘+ 4.62 The equilibrium point 2 = 0 of 2(k +1) = F(2(K}) is ‘+ stable, if for each ¢ > 0, there is 6 > 0 such that llz(0)|| < 5 = [lel O ‘* unstable, if not stable + asymptotically stable, if stable and 6 can be chosen such that N2(0)l| < 6 = Jim 2(k) = 0 + 4.63 The proof proceeds exactly as in the proofs of Theorems 4.1 and 4.2. + 4.64 Since AV < ~calal? < ~(ex/ex)V, Va * V(2(A) satisfies the inequality Vass ~ Vi $ ~(a/ea)¥i By applying this inequality recursively, t can be shown that Vp < (1~ #8)" ¥o. Tt can be seen thatthe constant (cs/cz) < 1, for if this was not the case V¢ would go negative which is impossible. Therefore, (1-8) <1. Now 2) corr] = (2) 0-8)" eo vote)". fa eons (LEE) < [2 (0 ue ue Set a = (=) and = (1- 2)" to obtain llx(&)I| < allz(0)y*, where a > 1 and 0< 7 <1. The origin will be globally exponentially stable if the inequalities satisfied by V and AV hold globally. we © 4.65 The proof proceeds as in the proof of LaSalle’s theorem (Theorem 4.4)87 + 4.66 (3) = (1) Consider the Lyapunov function candidate V(2) = 27 Pz. Then AV(z) =27(ATPA- P)r = -27Qr <0, V2 40 Hence, the origin is asymptotically stable (1) © @) The solution ofthe state equation is 2(8) = Ate) = MJ*M™2(0) where J = block diag|J1, Jay... Jn] is the Jordan form of A. If J, = 4,, then . Jha 40, 69 k+00 + P< Tf J, = A +N, where N is a Nilpotent matrix, then : I= 0d em= > ( ® ) aire! 0, ask 04> [M<2 mB Thus 2(k)= for all eigenvalues of A. (2) + (8) Let '2(0) = MJ!M~12(0) ++ 0, as k + 00 <=> [A <1 From (2), [|All < C7", for 0 << 1 and C > 0. Therefore Wis eMart < Spgycry* < HIS =F Hence, the infinite summation exists. On the other hand V() =2?Pe=2"Qr4e [= (ary7@ ws] 2>27Q2 Hence, P is positive definite. Now ATPA-P Scatry aca) -F (ay @ (at) x & Yayreay-T ay" (a4) = @ 5 & '* 4.67 Suppose the eigenvalues of A have magnituces less than one, and let P be the solution of ATPA-P = —Q, where Q = Q? > 0. It follows from the previous exercise that P = PT > 0. Take V(x) = 27 Pz asa Lyapunov function candidate. VS(@)) ~V(2) = fTa)Pf(z) — 27 Px = [Art g(x) PlAz t g(z)] -27 Px = -27Qz + 29"(2)PAz + 97(2)P9(2) Avie)88 CHAPTER 4. Given 7 > 0, there is r > 0 such that |Ig(x)ll2 < 7/|2ll2 for all |[xlz min(Q) Then 1 AV (2) S~ FAnin(Q)IleI3 which shows that the origin is asymptotically stable. © 4.68 Let 9(k, x) be the solution of x(k +1) = f(x(k)) that starts at x at time k = 0. Let wa VG) = Doha) 6(k,2) Then ea V(z) = 272+ SY) OM (k,x)o(k,2) > 272 On the other hand Vee) < Soi 0. Thus, the second inequality is satisfied with cy = 1-0?" > 0. Finally, from the continuous differentiability of f(z) we know that f(x) is Lipschitz cover the bounded domain D. Let L be a Lipschitz constant such that 4(2) ~ F(w)ll < Liz — alle Then HO(k + 1,2) — o(k + Ly)lla = IF (OC, 2) — F(R, ¥)) Ila < Lileks z) — o(,y) lle and by induction we obtain No(k,2) ~ 6(k,u)lla < Lj — lleConsider IV(z) - Vw) A 89 oa [67 (&,2)6(&, 2) ~ 37 (k,w)o(k.u)] wa {97,2 [0(h, 2) — o(bw)] + 6 (by) [60 2) — asm wot [Mo(e, z)llllo(, 2) — 9(&, w)lle + llO(K, »)MlllO(&, 2) — o(&, yDlla J iM 2 [los 2)llo + No(s y) be] Lhe — lla M 3 (Cr llelle + Cr*lvlla) Lz — ylle < (E orn) (llella + HvllIkz — wlle = & ‘Thus the last inequality is satisfied with oy = (Cyt cyt).Chapter 5 + 5.1 Let (ui,m1) and (ua,y2) be the input-output pairs of the two systems. We have u =u, y = v2, ts =v, and Ilyerlle $a (Ihtirlle) + Bi, §= 1,2 Then Mivarlle Soa (Iitnelle) + Ba Sa (a1 (llurrll) + 61) + Ba Sa (en (Ileal) + (281) + Bo -2(281) + Ba, to obtain A where we have used Exercise 4.35. Set a= a2 02a; and 8 llvrlle $ o(llurlle) +8 ‘To show finite-gain £ stability, start from 12 Iyirlle Sve lluielle + Bis In this case Iyorlle Uvellen <4 (lll) Hence, the system is Lao stable with zero bias, (>) The two curves |y} = fu and [yl = alul intersect. at the point |ul = (1/a)°/?. ‘Therefore, for Jul < (1/a)*/? we have |y] < (1/a)!/?, while for {ul > (1/2)*/? we have |y| < alu|. Thus, lvl Salul + (1/0), ¥ |u| >0 Setting y = a and 8 = (1/a)}/2, we obtain Iyelle, SrMlelle, +B (c) To show finite-gain stability, we must use nonzero bias. This example shows that a nonzero bias term may be used to achieve finite-gain stability in situations where it is not possible to have finite-gain stability with zero bias. 254 (1) h(0) =0 > |A(u)| < Llu), Vu. For p = co, we have suply(®l < Leuplu(| which shows that the system is finite gain C. stable with zero bias. For p € [1,00), we have [ WP ae < vf lw(OP @ = Ivelle, $Lilulle, Hence, for each p € [1,00], the system is finite-gain Cp stable with zero bias. (2) Let |h(0)| =k > 0. Then, |h(u)| < Llul +k. For p= 00, we have suply(@)| < Lsuplu(t)| + & wo Be which shows that the system is finite-gain Loo stable. For p € (1, 00), the integral J (L|u(t)| +k)? dt diverges as T + 00. The system is not C, stable for p € [1, 00), as it can be seen by taking u(t) = 0. «5.5 The relay characteristics of parts (a), (b), and (d) satisfy |y(¢)| < klult)| for some k > 0. Therefore, sup,so [y(t)| < ksupyzo lu(t)| and Jo" y*(t) dt < A? Jy” u?(t) dt. Thus, in these three cases the system is both finite-gain Cao and finite-gain £2 stable. In case (c), the system is clearly Cq, stable since the output is always bounded. However, itis not £2 stable. For example, the Cz input u(t) = e~* produces the output. y(t) = 1 which is not Lo. #56 Let V(t,2(¢)) =0. D.W = tim sup 51+ iys(e+A)) ~ Wie) = lim sup 1 /WGthzG+m) 2 We have cs V(t+hia(t+h)) < Filet + AIP93 a(t +h) = Alf (t,0) + 9(t,0)] + 00) => [x(t + h)|? ? g(t, O)||? + ho(h) FeV C+ mate+ Hy) < Sheo)F + a 1 ex a [ee Jim sup TER AEM < [Say < Ege since Y/@q/2c > 1. Thus ce SzTPO+ a DW's so WS Th) which agrees with the right hand side of (5.12) at W =0. + 5.7 Following the proof of Theorem 5.1, it can be shown that Hz(2)ll < v2 supllu(dll + Bry where ay Bo Consequently, Hv@ll < (ama +m) sup llu(i +f +s which shows that (5.11) is satisfied with p= oo ard 5.8 Following the proof of Theorem 5.1, it can be shown that lle(@)ll $m supliu(t)Il+ 81, where m = 2, , Bo e1ce Using (5.20), we obtain a ION < a (nsypitae+ 6) +00 (expt) +n A on (2rayptoce) +on(28) + a2 (sup) which shows that (5.22) is satisfied with rw(r) = ax(2nr) +aa(r), Bo = an(261) +0 * 5.9 Consider the system & = A@z+ Bou y = Or+D(u Assume that all matrices are uniformly bounded; that is, IOI Ser, [BOM Ser, NOMI Ses, IDOI Sey ¥E> to ‘and the origin of ¢ = A(t) is exponentially stable so that the transition matrix (t,o) satisfies NC to)l] < he“), WED to Tt can be easly shown that lv(h < esker*#—* oa +f cacaher*—P fury dr + calla From this point on, proceed as in Corollary 5.2.4 CHAPTER 5. # 5.10 (1) Let V(z) = Vs-(tujet<-(-r)et, Viulsralul lesin(u)] < fal, we we have V=2VQV-1)>0 when V>1/2 Thus, solutions starting in lIzla > 1 grow unbounded. This shows that the system is not La. stable. By linearization, it can be seen that the origin of the uaforced system is exponentially stable and all the assumptions of Theorem 5.1 are satisfied locally. Hence, the system is small-signal finite-gain La, stable for ‘sufficiently small ||2(0)Il (4) We saw in Exercise 4.55(6) that the system is input-to-state stable, By Theorem 5.3, it is La. stable By linearization, it can be seen that the origin of the unforced system is exponentially stable and all the assumptions of ‘Theorem 5.1 are satisfied locally. Hence, the system is small-signal finite-gain C.. stable for sufficiently small ||2(0)]| (5) With u = 0, the system has three equilibrium poinis at (0,0), (1,1) and (—1,~1). By linearization, it can be seen that the equilibrium points at (1,1) and (-1,—1) are saddles. By sketching the phase portrait, we can see that there are trajectories that diverge to infinity. Hence, the system is not La, stable. By linearization, it can be seen that the origin of the unforced system is exponentially stable and all the assumptions of Theorem 5.1 are satisfied locally. Hence, the system is small-signal finite-gain £. stable for sufficiently small ||2(0)| (6) We saw in Exercise 4.55(8) that the system is input-to-state stable. By Theorem 5.3; it is Ca. stable. By linearization, it can be seen that the origin of the unforced system is not exponentially stable. Therefore, we cannot apply Theorem 5.1. (7) We view the system as a cascade connection of the linear system &S-m-2, emt man and the delay element y(t) = yi(t - 7). The linear system has a Hurwite matrix. Hence, by Corollary 5.2, it is finite gain £. stable. The time-delay element is finite-gain Ca. stable. Hence, the cascade connection is finite-gain Lop stable.95 1+ 5.12 We start by investigating stability of the unforced system dy se2, f= -(2 +22) —he +m) ‘We use the variable gradient method to find a Lyapunov function. V = 9" (2) f(z) = ou(a)fu(z)-+92(2) fo(2) Since fo(2) = ~(21 +22) ~ (2a +22), let us take gy = 21 +22. From the symmetry condition dg1/8z2 = 92/8, = 1, we take gy = bay +23. Then V(a) = 362i + a7 + 323. The quadratic function V(2) is postive definite for b > 1. V = (bay + 22)t2 — (a1 + 22)? — (a1 + 22)A(e1 + 22) = —2f + (b- 2)zate — (2 +-22)h(z + 22) ‘Taking 6 = 2 yields V = a} - (a1 + an)h(a1 +22) < -2} — ae +29)? = -2T Qe where Q = [ cane globally exponentially stable. Now consider the forced system. The Lyapunov function V satisfies inequalities (5.6)-(5.8) globally. It is easy to see that (5.9) and (5.10) are satisfied globally. It follows from Theorem 5.1 that the system is Cy stable for each p € [1,00]. |: ‘The matrix @Q is positive definite; hence, the equilibrium point at the origin is *5.13 Since W(2) is positive definite and radially unbounded, it follows from Lemma 4.3 that there is a class Kop function a such that W(z) > a(([z|)) for all x € R". Since |y(u)]| is positive definite, it follows from Lemma 4.3 that there is a class K function po such that |¥(u)| < po(|lull) for all ué R™. Hence, Y yeu) S ~a(llzll) + po( ull) < ~ Fat), V [lz|| 2 a (2po({lul)) ‘We conclude from Theorem 4.19 that the system is input-to-state stable. Furthermore, ||h(z, u)|| is a positive definite fonction of [ |. 1 follows fom Lemna 4.3 that there i lass Kae funtion py sch that wceeot 0 for all x € A? oa 2 = en qe fle kal = —Ki (a) [asinz, 2] { wasinz ~ hea ]96 CHAPTER 5. Gee=[osan 2) []= ‘Thus, W(z) satisfies (5.32)-(5.33) globally. It follows from Example 5.9 that the system is finite-gain C2 stable and its Co gain is less than or equal to 1/k. (2) See a = h(a) f(z) Let W(z) = Bax. ~2 M ya=[a m 2] _— (3 — 22) sat(2} - 23) = -h sat(h) 23 sat(z} - 23) 0 Beal nm =} [2 ]-a-a-xe Let D = {x € R? | |h(z)] < 1}. W(2) satisfies (5.32)-(6.33) in D with k = 1. Taking V(2) = W(z) and 7 = 1, it can be verified that (5.28) is satisfied in D. Consider now the unforced system, Ae) = > ta(t)=0 > x9(t) = constant > é2(t) = > u(t) 20 > a(t) 20 > a(t) =0 > x(t) =0 ‘Using of Lemma 5.2, we conclude that, for sufficiently small ||zo|}, the system is small-signal finite-gain £2 stable and its £2 gain is less than or equal to 1. (8) Let D = {x € R? | [22, + 29] < 1}, For z € D, the system is given by A+ Bu, y=Cr 4=[4 4] e-[t] este 9) Since A is Hurwitz, there exist positive constants ky and fy such that for all ||z(0)|| < ky and supy>o |u(t)| < kz, 2(¢) € D for all t > 0. The system is linear time-irvariant and its C2 gain can be determined using ‘Theorem 5.4. The transfer function is where 1 wh HO= Sy eg1 = Faure Th C= 05 sup |H(ju)| = —h— = 2 ook ~ t= 8 VB ‘Thus, for sufficiently small |[zo||, the system is small-signal finite-gain £2 stable and its £2 gain is 2/3. (4) n 1)=[-aeBaeg) = [2 ]- meen ce! [ase7 Monlat =] [8 Jens =a ‘Thus, W(z) satisfies (5.32)-(5.33) globally with k = 1. It follows from Example 5.9 that the system is finite-gain Co stable and its £2 gain is less than or equal to 1. # 5.16 (a) Let V(z) = fo" o(u) dy + Hah +29) Vo = ~af-so(e)—2} +290 < -af—a} tzu S ~~ Olle olla + lelllel < -C ~ Oli, ¥ Ua > lul/o where 0 < 6 <1. It follows from Theorem 4.19 that the system is input-to-state stable. Since |y| = [za] < {f2a, we conclude from Theorem 5.3 that the the system is La. stable. (b) Let V(2) = alf" oly) dy + H(z? +23). _ 1 Ve (YT Ln, H= Gf amor (B) +5hTH = asf nolo) ~ a+ Sat 38 < (-04 G4 h) a Choosing a = 7 = 1 yields W <0. Hence, the system is finite gain C2 stable with Cy gain less than or equal toone. +537 Me, My, Moy 2 1 1 . M+ Mou = ae! Kou- Hos Wo (L+ Wu) + 2o+ Wahl + Wu) = foe warc sways 24 Bout Le Wut btw = -hu+wahc+wy+ {ere} ditL+ peta} ~ pins Deu hu ee ul (21 = J? Ju =- Hb + Ww)" + Way +2 - ght Ruth 3 ut u — Buu eee ae = = pes wus won tne dy WY <0 implies Me Meu liautu- 1yr Belt Be Ouse guy From this point on, proceed as in the proof of Theorem 5.5 (starting from (5.29)). # 5.18 The closed-loop system is given by sar (7 ‘The closed-loop map from w to [ ez ] is given by B= felt) +G.(ew Ye fe(z)98 (CHAPTER 6. where tons-oot (Y's conn w= [2]. M=[ ort] For the closed-loop system, the left-hand side of (5.28) is given by av 1 (av 2 (OV)" Apr ne = Maer a (2) ect (54) tate = © [yar (®%)] 44 () eur (%)” torn gt (OH avy" -¥{ o* (SE) ] +a (SE) x (SE) +45 (5) oo" (52) av, 1a Ov)? Aan = we! +3(3) [ ES) +Qhh <0 From Theorem 5.5, we conchae thatthe clsedoop map from w to [ Y | ie Snitegain Ca stable with Ca gain less than or equal to 7 + 5.19 (a) The existence of 6 follows from the fact that : <2-Vo Bias ‘Thus, by choosing 5 small enough, we can make (1 — ¢/2- v8)/vI=8 > 1- (b) By definition of the Ca gain, we can find u € Lp such that [lulleay = 1 and lfyllcon > Yan(1 ~ €/2). If there was no such u, we would have llvllean < Y2n(1— €/2) for all [lullezy = 1, which contradicts the claim that zr is the Can gain. By choosing 6 of part (a) less than one, there exits time f; such that fhoubOult) at = 6, (We have . , [ ul ua(t) dt = [ uT (u(t) dt =6 = ljuallean = V8 © Flom d= [woul d= 1-5 > Hula = IHS Consequently, Ueellean $ tee V5 ) € Uvallan 2 llllcan ~ Ivellean 2 ver (1~$) ~ VBran ‘Therefore, fallen » PV, > enn (d) Let u(t) = ui(t-+ tr) and y(t) = m(t +h) f WP (butt) dt lb Which implies that |jullcy = [Iusllcan- Similarly, lIylle: = llyillcan- The fact that y(t) is the output corresponding to u(t) follows from linearity. Finally, |lyllc, > (1 — €)72Rllullcs follows from part (). [te +h )u(t+ 4) d= [~ uy(r)Pun(r) dr99 '*.5.20 The closed-loop transfer functions are given by i 2 wetter J | U2 a HO ‘The closed-loop transfer function from (tu; ,uz) to (p1,y2) ( or (€1,€2)) has four components. Due to pole-zero cancellation of the unstable pole s = 1, three of these components do not contain the unstable pole; thus, each component by itself is input-output stable. Ifwe restrict our attention to any one of these components, ‘we miss the unstable hidden mode. By studying all four components we will be sure that unstable hidden modes must appear in at least one component 25.21 Yar = (Hiei)y» Yor = (Hae2), IN Wire < ralselle +4 <1 [Ra tnede + lla +o tna] +64 am n(Br + r261) ot thuaelle + BA oD alarlle + AEE + 8 ne 1 Ba + Br 2 —Husrlle + nm oy ‘The expression for llyarlle can be derived in the same way. «5.22 (a) Let the underlying space be £,. We have dz, dz € Coo. From the analysis of Example 5.14, we have €1,€2,2 € Loo, provided ¢ < 1/7174. From the equation c= A+ A Ber wwe see that 1 € Ceo. ‘Thus, for sufficiently small ¢, the state of the closed-loop system is uniformly bounded. (b) From (5.44), we see that one of the terms on the right-hand side is €7,||dal|c.,. For da(t) = asinwt, this term is given by enlidalic., = wary When the product ew is small, the term is negligible. However, as we increase w, the product ew will no longer be small. At frequencies of the order O(1/e), the product. ew will be of order O(1). While the system remains La, stable, the bound on 2, given by (5.44), will not be close to the bound |ld]_, +8 + 782, as concluded in the example. The conclusion of the example is valid only for frequencies of the order O(1) (or frequencies for which ew is small). © 5.23 (a) Using 4 = wom sub} @ = wt =m we obtain a = ote fy = -s}om-aleu is = m-a100 CHAPTER 5. (b) We search for functions Vj and V2 that satisfy the Hamilton-Jacobi inequality for the systems H, and Ha, respectively. For Hy, let i = ax(324 + 323), a1 > 0 H cu(at +28) + Shad Bot s(-u+gh+})a Choosing a; to minimize 7, we end up with ay = 71 = 1. For Hy, let V2 = (a2/4)z4, a2 > 0. & 3) af Choosing a2 to minimize 72, we end up with ay Since 7172 = } < 1, we conclude by the tmnall-gain theorem that the closed-loop system is Snive- gain‘ stable” ‘To find an upper bound on the £4 «gain, we search for a function V that satisfies the Hamilton-Jacobi identity. We consider V(a) = Vi + Va =n (def + $23) + Sef and we allow ourselves the freedom to change the choice of the positive constants a; and a. ov 4 a am Gel = neatt - a2} - Bead + anzaz} — an2$ ‘Choose aig = a3 /2 to cancel the cross product term. ov. Re = arp Ms 1 1 ogy She bet < (- a+ gh ghaas})a Choosing a; to minimize +, we end up with a1 == 1. Thus, the C2 gain is less than or equal to one. Note that a more conservative upper bound can be obtained by applying (5.40). According to (5.40), an upper bound on the £2 gain from u =u; to y = yr is given by. H= —ayx} — 023 — n am a 3Chapter 6 © 6.1 Define @, d, and g as shown in the next figure. Then, (tu) — Kru, Ku, From [h(¢,u) — Kaul [A(¢,u) — Kau) <0 we have a @-Ku)<0 > g-a <0 > F(-a <0 @ Gar gtoe TL on 40-4 , ] K Figure 6.1: Exercise 6.1 +62 V = ah(z)z = K(2) [= +hne+ 4] = Lh@)tn(e) — bal + hau < yu *6.3 Take V(z) = S[ka*z} + 2kaziz2 + 23] +5 J" h(y) dy, where 6 > 0 and 0 V~ phar} + 2aypiez? + (ka ~ 2p2)2} = V+¥(2) Since py2 < min{2ax, ak/2}, ¥(2) is positive definite. Hence, the system is strictly passive. © 6.5 We have X(s) = (Ms-+K)"1U(s) or (Ms+K)X(s) = U(s). Thus, the state model for the dynamical system is Mé=-Ke+u Let V(z) = JZ AT(@)M do > 0. 7 (a) M: Hence, the system is output strictly passive. Ve = AT (a)(-Kaz +u) < -AT (z)h(z) + AT (2)u = —yTy + yu £66 We have uy = uz =u and y= + yp. Let ayn > Ff leasu) + wea) + of als) + 0s) apn > 2 plese) +a onl) + vale) + ¥ale2) where V, = Vi (x1)‘and V2 = Va(z2). Then, at; =n tm) > Eien) tuo) + vPovan) +08 alm) + 062) where V (2) = Vi(z1) + Valea), (u) = va(u) + a(u), W(x) = a(er) + ¥2(e2), and fle 4) = | Fo) yy ‘The proof follows from this inequality for the cases of passivity, strict passivity, and input strict passivity. For output strict passivity, we require y?'pi(yi) > iu? vi for all ys, for some 6, > 0. Then filziyu) | wf u(r) + uF pa(ve) > Fmin{br,52} vy where we used the fact that (+e) (vr +2) $207 + ¥F ¥2) © 6.7 G(s) is Hurwitz if and only if @ and az are positive. by + jew | buae + (boas = bio cei li Ceara erg =F + jane! Re(G( ju) > 0 for all w € Rif and only if by > 0 and boa, > br Him? RefG iu) = boa — by ‘Thus, G(s) is strictly positive real if and only if all coefficients are positive and by < boa.103 ‘+ 6.8 For any strictly positive real transfer function, D + D™ > 0. Since, D + D™ is nonsingular, we have D+ D? > 0. Hence, W is a square nonsingular matrix. Thus, ITL = LW(D+ DT)" WTL Substituting for WL from (6.15) we obtain I7L = (CT - PB)(D+D")"(C - BP) Substituting this expression in (6.14) yields Pl(c/2)I + A] + [(e/2)1 + ATP + (C7 — PB)(D + DT)-(C - BTP) =0 Factoring out the quadratic term, we obtain the given Riccati equation. + 6.9 Since the system is input strictly passive with o(u) = eu, we have uty> & jew) teuTu, €>0 Since the system is inite-gain Cp stable, we have [Vow an [Pou asa, >0, 620 To arrive at the desired inequality, we need to assume that ; = 0. From the first inequality, we have V(a(m)) - Vien) < f [u" (u(t) -eu7 (Qu(e)] at [Ct evoa-§ [wt oun a-§ [ uP aule) at [Cen a-§ [vr que a-£ [vow a An Since this inequality is valid for all 7 > 71 > 0, we have OY peeyu) suly— Sutu— ey? Ty > & seus Eatus Hy Fel) Sty Sulu say = uhy> Fase) + sutut vy + 6.10 (a) The equations of motion are M(q)G+ C(a,4)4 + Da + 9(a) veg? Lear Pe Vox FM is jatar ; ren ; a lu C4 - D4— gla) ~ HAMA TG = Vu—V™Dy < vu where we have used the property that AM —2C is ¢ skew-symmetric matrix. The inequality V < yu shows that the system is passive. (b) In this case a Vsu"(-Kay +0)104 CHAPTER 6. which shows that the map from v to y is output strictly passive. (©) With v = 0, we have . Vs -u"Kay = -a Kai <0 V=0 = a(t) =0 + a(t) =0 + g(a(t)) =0 = alt) Hence, the origin is asymptotically stable. It will be globally asymptotically stable if q = 0 is the unique root of g(q) = 0 and P(q) is radially unbounded © G11 (a) Let V = $Jiu} + $aud + Jy} Vi = Suyiin + Jawratin + Jgwig (Ja = Js)onsorids + wren + (J5 — Sr)eorwrisy + watta + (dt — Ja)uiunuy +ugug = oT u which shows that the system is lossless. (b) With u=—Kw +, we have eK toy = oW>V 4+ Amin(K)|ld Hence, the map from v to w is finite gain C, stable with C2 gain less than or equal to 1/Amin(K). (©) With u = ~Ku, we have V =~" Kw. V is positive definite and radially unbounded and Vis negative definite for all. Hence, the origin is globally asymptotically stable. 6.12 €1 = t~ ye = ty — hale) ~ Ja(aa)en ep = uty = wth(a) + A(er)er ‘Substitute e, from the second equation into the first one. 1 =m ~ ha(2) — Ja(a)fus + ha (21) + Ja(ea)en} (1 + Jo(22) di aa)]en = un ~ hava} — Ja(va)ua ~ Ja(za)ha (21) 1 = [+ Jawa) Ji (2) "lus — ha (22) ~ Jo(e2)ua — Ja(a)ha(2a)] 2 = ta + ha (21) + Ji(2a)LL + Jala) Ja (01)] "us ~ ha(ee) ~ Ja(ea)ua — Ja(ea)ha(23)] Similarly, Substituting ¢; from the first equation into the second one, we obtain en = [T+ Ji(21)da(a2)) ue + ha (2s) + Ji(ea)un ~ Ji(aa)ha(za)] Substitute the expressions for e, and e2 into the equations #1 = file) +Gr(er)e, 22 = fala) + Galzn)en + 6.13 Let us start with (6.26)-(6.27) 1 =u —ha(enea), —€2 = ua + Ma(zs) substitute ey from the second equation into the first one. 1 =u ~ hala, te + hi(21)) ‘The pair (e1,¢2) is uniquely determined. Consider now (6.30)-(6.31). €1 = ty — halter), ey = ua + har) substitute ey from the second equation into the first one. 1 = ty — haltyua + hi(ai)) ‘The pair (¢1,€2) is uniquely determined.105 © 6.14 (a) Take Vi = (1/2)(2} + 23) and Vo = Jo* ha(s) ds. Y= mm) +mea > na =Vtwk(n) Since hy € (0, 00}, Hh is output strictly passive. Va =—zaha(2s) + eahales) > wer = Va +zsho(2s) Since hz € (0, 0}, Ha is strictly passive. ‘Thus, the feedback connection is passive. (b) With e; = 0, we have 0 a) =0 > nl mit) Hence, H; is zero-state observable and the origin of the feedback connection is asymptotically stable. It is globally asysnptotically stable if Vj aud Vp are radially unbounded. Vj is radially unbounded because itis a ‘quadratic form and V3 is radially unbounded becaise [mo as c = ‘Thus, the origin is globally asymptotically stable. > alt 1 2) 5in(1 +23) + 0 as |23| +00 6.15 (a) Let Vi = fof + 33. W=-sf-abtme Hence, H; is strictly passive. Let Va = 424 Vo =-2$ +e Hence, Hs is strictly passive. It follows from Theorem 6.1 that the feedback connection is passive. (b) Since both systems are strictly passive with redially unbounded storage functions, it follows from The- ‘orem 6.3 that the origin is globally asymptotically stable, 6.16 To study asymptotic stability, take uy =O and uz = 0. Then, e1 = —ys, es = 11. Suppose Hz is input strictly passive. Vise =—viu, Va SFya—eFyalea) = vlna —eFvrles), Fpa(er) >0 Ver #0 Take V = 4%. V<~eFpa(es) <0 and V=0 > = = m=0 By zero-state observability of Hi(—H2), n() 20 > 2) =0 Hence, by LaSalle’s invariance principle, the origin is asymptotically stable. Similarly, suppose H; is output, strictly passive Viselu vial) =—vFn —vfoilu), Veselu=vfm, veor(u)>0 Vm £0 ‘Take V = 4 + Va. V<-vTolu) <0 and V=0 > m1 =0 By zero-state observability of Hy(~H),106 CHAPTER 6. 6.17 To study asymptotic stability, take up = 0 and u; = 0. Then, ex = yn, ¢x = ve. Suppose Hh is input strictly passive Vasu auto, Vi seu —eTyor(er) = Fu —efer(e), Terler) >0 Yer #0 Take V = Ve +i. . V<-cyi(er) <0 and V=0 + 4 =0 > w=0 By zero-state observability of Hp, walt) =0 + 2) =0 Hence, by LaSalle’s invariance principle, the origin is asymptotically stable. Similarly, suppose H? is output strictly passive, Vo S$ €fun ~ vi palve) = wT v2 — vE pa(ve), Vi Sem =v vs» ve pave) >0 Vy #0 Take V=V+Vi- V <-v¥po(y2) <0 and V=0 + w=0 By zero-state observability of Hai, w(t) =0 > z(t) =0 + 6.18 To study asymptotic stability, take uz = 0 and a; = 0. Then, ¢2 = ti, €1 Va Sut Quan +27 Sier + ef Ries, Vo

0. Yas V yr =0 and w=0 Now u(t) =0 > er(t)=0 and y(t) By zero-state observability, salt) 0 > a4(t) =0 and w(t) > z(t) no Hence, by the invariance principle, the origin is asymptotically stable. It will be globally asymptotically stable if Vj and Vp are radially unbounded.Chapter 7 «7.1 (1) The Nyquist plot of G(s) for w > 0 is shown in Figure 7.1. It is a circle centered at (~},0) with radius equal to #. The plot for w <0 is identical; that is, the Nyquist plot traverses the circle twice as w changes from—oo to oo. The transfer function has two poles in the right-half plane. Thus, for absolute stability, the disk D(a, 8) must be inside the circle, so that the Nyquist plot encircles the disk twice in the counterclockwise direction, Clearly, the largest disk is the circle itself. Thus taking “1, -1 gh Gg ~The where €; > 0 and €2 > 0 are arbitrarily small, we conclude that the system is absolutely stable for the sector [1 +e1,1/ea}. (2) The Nyquist plot of G(s) for w > 0 is shown in Figure 7.2. Since G(s) is Hurwitz we can apply case 2 ‘or case 3 of the circle criterion. For case 2, we see that the Nyquist plot lies to the right of the vertical line ‘through (—0.021,0). Thus the system is absolutely stable for the sector [0, 47.62]. For case 3, the Nyquist plot should be contained inside the disk D(a,8). We chose a circle of radius 0.114 centered at 0.0732 ‘Thus, a = -1/0.1871 = 5.34, 8 = 1/0.0408 = 24.51, and the system is absolutely stable for the sector [-5.34,24.51] Nyquist Diagrams 015; ; an \ 2 ons > i of -= 4 E008] 04 , 15 1-05 0 05 01 0 Ore Real Axis eal Acs Figure 7.1: Exercise 7.1(1). Figure 7.2: Exercise 7.1(2) (3) The Nyquist plot is shown in Figure 7.3. Since G(s) is Hurwita we can apply case 2 or case 3 of the circle criterion. For case 2, we see that the Nyquist plot lies to the right of the vertical line through (—0.35,0). ‘Thus the system is absolutely stable for the sector [0, 2.857]. For case 3, the Nyquist plot should be contained inside the disk D(a, 9). We chose a circle of radius 1.07 centered at 0.35. Thus, a = ~1/142 = -0.704, 8 = 1/0.72 = 1.389, and the system is absolutely stable for the sector [-0.704, 1.389] 107108 CHAPTER 7. (A) The transfer function can be simplified to @-1 C0) = BT It has poles on the imaginary axis. We close the loop around G(s) with a. the transformed transfer function Gs) 1-s rls) = TyaG@) ~ F¥la— Dart is Hurwitz for 0 < a < 1. Take a = 0.1. The Nyquist plot of Gr(s) is shown in Figure 7.4. It lies to the right of the vertical line through (—2.76,0). Thus the system is absolutely stable for the sector [0.1, 0.4632] yout Diagrams a5 y im ye Imag Axis Imaginary Ase =| -0.5| 10} 05 0 05 1 os Real Axis Fal Ads Figure 7.3: Exercise 7.1(3). Figure 7.4: Exercise 7.1(4). (5) The Nyquist plot is shown in Figure 7.5. The transfer function is Hurwitz. Therefore, we can apply cases 2 and 3 of the circle criterion. We start with case 2. From the Nyquist plot we find that min{Re{G(jw)]} > =0.6. So we choose 8 = 1/0.6 = 1.67 and conclude that the system is absolutely stable for the sector (0, 1.67]. Now we apply case 3 of the circle criterion. The Nyquist plot should be inside the disk D(a, 8). We locate the center of the disk at (0.2,0) and take the radius as 0.9. The circle is shown in Figure 7.5 (dotted line) Thus, 1/ 0.91, 1/8 =0.7 > 8 = 1.43, and the system is absolutely stable for the sector [-0.91, 1.43} (6) The Nyquist plot is shown in Figure 7.6. Since G(s) has one pole in the right-half plane, we apply case 1 of the circle criterion. The Nyquist plot should encircle the disk D(a,/8) once CCW. We chose a circle of radius 0.09 centered at (—0.16,0). Thus, a = 1/0.25 = 4, B = 1/.07 = 14.29, and the system is absolutely stable for the sector [4,14.29]. (7) The Nyquist plot is shown in Figure 7.7. The transfer function is Hurwitz. Therefore, we can apply cases, 2 and 3 of the circle criterion. We start with case 2. From the Nyquist plot we find that min{Re{G(jw)]} > 0.341. So we choose f = 1/0.341 = 2.93 and conclude that the system is absolutely stable for the sector [0,2.93]. Now we apply case 3 of the circle criterion. The Nyquist plot should be inside the disk D(a, ). We locate the center of the disk at (0.3,0) and take the radius as 0.8. The circle is shown in Figure 7.7 (dotted line). Thus, 1/a =—11 + a= O91, 1/8=0.5 = 8 =2, and the system is absolutely stable for the sector (0.91, 2} (8) The Nyquist plot is shown in Figure 7.8. The transfer function is Hurwitz. Therefore, we can apply cases 2 and 3 of the circle criterion. We start with case 2. From the Nyquist plot we find that min{Re(G(jw)]} > =0.08. So we choose § = 1/0.08 = 12.5 and conclude that the system is absolutely stable for the sector [0.12.5]. Now we apply case 3 of the circle criterion. The Nyquist plot should be inside the disk D(a, 8).109 1 1 05 ; . o5| 3 f 2 2 \ i g 3 of + $0 £ £ -08| -05| 4 4 4-050 08 4 2 45-4 05 0 Real Axis Real Axis Figure 7.5: Exercise 7.1(5) Figure 7.6: Exercise 7.1(6) ‘We locate the center of the disk at (0.1,0) and take the radius as 0.21. The circle is shown in Figure 7.8 (dotted line). Thus, 1/a = -0.31 = a = -3.2%, 1/8 =0.11 + 8 = 9.1, and the system is absolutely stable for the sector [-0.31, 9.1. Nyquist Diagrams Nyquist Diagrams 02 i 2 $ i gu i --24a Bo -- 3c 4 E Boas -o5| 02] a 05 -02 ° 02 Real Ais eal As Figure 7.7: Exercise 7.1(7), Figure 7.8: Exercise 7.1(8) 272 (a) ‘The Nyquist plot is shown in Figure 7.9. The plot lies entirely in the right-half plane. By case 2 of the circle criterion, we conclude that the system is absolutely stable for the sector (0, 5], where 8 can be made arbitrarily large. Consequently, the system is absolutely stable for the sector (0,1). (b) The nonlinearity sat(y) belongs to the sector [0,1]. Hence, the system has a globally exponentially stable ‘equilibrium point at the origin. This implies that there can be no periodic solutions, since every solution ‘must converge to the origin # 7.3 (a) The equilibrium points are the roots of the equations O= Hz — hey +22), O= 2, 22 2h(ay +22) ‘Multiply the first equation by ~2 and add to the second equation to obtain the equation (x1 +22) — 4h(a; + 22)0 CHAPTER 7. os| Imag Axis 2 05 0 05 1 «15 2 Real Axis Figure 7.9: Exercise 7.2. which has the unique solution 2; +z = 0 due to the properties of h. 2 422 =0> A(z; +22) =0 > 2, = 0 m2 =0 Hence, the origin is the unique equilibrium point. (b) Let y= 21 +22 and u= h(x; +22). Rewrite the system equation as a = -n+u fig = ty — an +20 y= mts u = -h(t; +22) “The system is now modeled as a feedback connection with the transfer function ast4 G(s) = (= aay and the nonlinear element ¥(y) = h(y). The Nyquist plot of G(s) lies completely in the right-half plane. Hence, by the circle criterion, the system is absolutely stable for the sector (0, ], where f can be arbitrarily large. This sector includes the nonlinearity A. Thus, the origin is globally asymptotically stable. 67.4 Let y= 21 and u=—gcosut 21. Rewrite the system equation as fp = —(W +0?)q, — Quer tu yon uu = ~geosut y ‘The system is now modeled as a feedback connection with 1 Ge) = 7+ Qys +p? +a? cost y. The function ¥ belongs to the sector (~9,q]. We will apply the result of Example 7.1 and (ts) We also have ‘We have [ys(t,y)| < v2lul with 12 1 Col = Garauu It can be verified that 1 cps = sup |G(iw)| =f A BDO me ngleoo={ EP eee ‘The system is absolutely stable if qq <1. Since the system is linear, this condition ensures that the origin is exponentially stable. © 7.5 Represent the system as @=Az+Bu, y=Cz, E(t)Cz = E(t)y & -v(t,y) ‘This problem is a special case of Example 7.1, with I(t, w)lle < E()|allvlle < Ilvll2 > 2 =1 ‘21 = Sup FmaslG()] <1 wer y (172 < 1) is satisfied. Hence, the origin is uniformly asymptotically +76 (a) The closed-loop state equation is = Av + Blsat(—Fs/L) =(A- BF)x + B[Fz — Lsat(Fz/L)] Setting y = Fz and Uy) = Lsat(y/L) — y, we represent the system as (A-BF)s-Bvly), y= Fe which isin the form of Figure 7.1 of the text, with G(s) = F(s! - A+ BF)". (b) We have Iwill < yh Val < £046) Hence Ital = 4] Sova < Hl ‘This is a special case of Example 7.1, with 6 w= pp7 am n= sup IGG) T+6 Hence, the system is absolutely stable with a finite domain if 5 Tas Sup lOGel <2 © 2s+1 G9) = F508 It can be checked that sup |G(jw)| < 2.55. Hence, the system is absolutely stable with a finite domain if boa Tae < agp 7 F< 006852u2 CHAPTER 7. ‘To estimate the region of attraction, we apply the loop transformation of Figure 7.3 of the text. —Kysitny o=Kyti=ayti The state model of the transformed system is, 8 (A-BF + BF)z + Bu InFrti ) where @ € [0, co]. We need to find a matrix P that satisfies (7.6)-(7.8) for the transformed system. Equations (7.6)-(7.8) take the form = P(A- BF +%3BF)+(A-BF+%BF)'P = -L"L-eP PB =2y,FT - J2L7 It can be verified that 0.4638 0.3757 = 0.002, P= [ O68 03757 |, and L= [0.2859 -0.4764 ] 27 Pz, asin Example 7.4 satily (7.6)-(7.8). The region of attraction can be estimated using V(2 «7-7 (1) The transfer function has poles in the right-half plane. Therefore, we start by applying a loop transformation. We close the loop around G(s) with a. The transformed transfer function is 0) TraG@) #4 (@—Detl Gr(s) ‘Take a = 1.1. The Popov plot of Gr(s) is shown in Figure 7.10. ‘The largest sector is obtained with 7 = 0 so that the line of slope 1/7 is vertical and intersects the horizontal axis at a point arbitrarily close to zero. ‘The transformed system is absolutely stable for the sector (0, ], where k can be arbitrarily large. Hence, the original system is absolutely stable for the sector [1.1, J], where 8 can be arbitrarily large. (2) G(s) is Hurwitz. Its Popov plot is shown in Figure 7.11. The Popov plot approaches the origin tangent to a line whose slope is 4.926. We take +7 = 1/4.926 = 0.208 and k > 0 can be arbitrarily large. Let us verify that = 0.203 is acceptable. Since G(s) has relative degree two, the product CB is zero in any state-space realization of G(s). Thus, the condition 2/k +yCB +7B7C7 > 0 is always satisfied. The poles of G(s) are ~2 and ~3. Thus the condition 1 +7A #0 is satisfied for + = 0.203. We conclude that the given system is absolutely stable for the sector [0, k] where k is arbitrarily large? (3) G(s) is Hurwitz. Its Popov plot is shown in Figure 7.12. The Popov plot approaches the origin tangent to a line whose slope is 1. We take 7 = 1 and k > O can be arbitrarily large. Let us verify that 7 = 1 is acceptable. Since G(s) has relative degree two, the product C'B is zero in any state-space realization of G(s). ‘Thus, the condition 2/k + 7CB + yBTOT > 0 is always satisfied. The poles of G(s) are ~0.5 + j0.5V3. Thus the condition 1+7A # 0 is satisfied for 7 = 1. We conclude that the given system is absolutely stable for the sector (0,] where & is arbitrarily large. We can obtain a sector with a < 0 by applying a loop transformation. Using the Routh-Hurwitz criterion, it can be shown that Gr(s) = G(s)/[1 + aG(s)] is Hurwitz if a > ~1. Take a = ~0.9 to obtain Gr(s) = 1/(s? +s +0.1). The Popov plot of Gr(s) is shown in Figure 7.13. The Popov plot approaches the origin tangent: to a line whose slope is 1 We take y = 1 and & > 0 ‘can be arbitrarily large. Let us verify that 7 = 1 is acceptable. Since Gr(s) has relative degree two, the product CB is zero in any state-space realization of Gr(s). Thus, the condition 2/k+yCB +4BTCT > Ois always satisfied. The poles of Gr(s) are ~0.1127 and ~0.8873. Thus the condition 1+ 7A # 0 is satisfied for 1. We conclude that the given system is absolutely stable for the sector [-0.9, k] where k is arbitrarily large. Compare these sectors with the ones obtained using the circle criterion in Exercise 7.1.113 ° 4 2 005 g g Zo =o 3 3 2 “015 “4 0 2 4 6 6 w 005 0 008 01 015 Oz Reis] Reis) Figure 7:10; Exercise 7.7(1) Figure 7.11. Exercise 7.7(2). . 4 2 i 1 Zo EY esd “ “02 0 02 04 08 08 02 4 6 @ eG) eG] Figure 7-12: Exercise 7.7(3) Figure 7.18: Exercise 7.7(3). (A) The transfer function can be simplified to e-1 C= S57 It has poles on the imaginary axis. We close the loop around G(s) with a. the transformed transfer function Gs) es Gris) 1+aGis) s?+(a-1)s+1 is Hurwite for 0 < a < 1. Take a = 0.1. The Popor plot is shown in Figure 7.14. The Popov plot lies to the right of the vertical line through (~2.76,0). Thus, we take y = 0 and k = 1/2.76 = 0.3623. Therefore, the system is absolutely stable for the sector (0.1, 0.4623]. This is the same sector obtained by using the circle criterion because 7 = 0. (5) The Popov plot is shown in Figure 7.15. It lis to the right of (and tangent to) a line through (—0.5,0) whose slope is 3. Thus, we take 1/3 and 1/0.505 = 1.98. Let us verify that 1/3 is acceptable. Expanding G(s) as a power series in 1/s, it can be seen that the coefficient of 1/2 is ~1. Therefore, in any state-space realization of G(s), CB = ~1. Thus, 2)k-+ yCB-+7B"C? = 0.3433 > 0. The poles of G(s) are land —1. Thus, 1 + yA = 2/3 # 0. Hence, the system is absolutely stable for the sector [0, 1.98]. TWe can also show absolute stability with a sector (a, where a is negative by applying a loop transformation. See the solution of part (3).14 (CHAPTER 7. 2 os os| | 02] img) bows oe wins) -02| -0.4] 3 10 350. 05 1 Fela) Peja) Figure 7.14: Exercise 7.7(4). Figure 7.15: Exercise 7.7(5). (6) G(s) is not Hurwitz, We need to apply a loop transformation with a > 0 such that Gr(s) = G(s)/{1 + G(s)] is Hurwitz. Using the Routh-Hurwitz criterion, it can be shown that G(s) is Hurwitz if'a > 4. Take = 4.1 to obtain G(s) = (6-+1)/(s? + 3s? + 4.16 +0.1). The Popov plot of Gz(s) is shown in Figure 7.16. ‘The Popov plot approaches the origin tangent to a line whose slope is 2. We take 7 = 0.5 and k > 0 can be arbitrarily large. Let us verify that 7 = 0.5 is acceptable. Since G(s) has relative degree two, the product CB is zero in any state-space realization of G(s). Thus. the condition 2/k + CB +yB7CT > 0 is always satisfied. The poles of Gr(s) are ~0.0248 and -1.4876+ 1.3466. Thus the condition 1+ 7A # 0 is satisfied for 7 = 0.5.We conclude that the given system is absolutely stable for the sector (4.1, k] where k is arbitrarily large. Compare this sector with the one obtained using the circle criterion in Exercise 7.1 (7) The Popov plot is shown in Figure 7.17. It lies to the right of (and tangent to) a line through (-0.26, 0) whose slope is 1. Thus, we take 7 = 1 and k = 1/0.2505 = 3.99. Let us verify that +7 = 1 is acceptable. Since G(s) has relative degree four, the product CB is zero in any state-space realization of G(s). Thus, the condition 2/k-+yCB + BCT > 0 is always satisfied. The poles of G(s) are at ~1. Hence, = 1 does not satisfy the condition 1+ 7A # 0. We take 7 = 1.01. It can be verified that the Popov plot lies to the right of the line through (—0.2505,0) whose slope is 1/1.01 = 0.99 (see Figure 7.17). Thus, the system is absolutely stable for the sector [0, 3.99} os 03) 02 fo 0. L g zo zo = 8 3-04} / -o2|/ -03| Ss 0 05 5 ° os 1 Fel] eG) Figure 7.16: Exercise 7.7(6). Figure 7.17: Exercise 7.7(7).. (8) The Popov plot is shown in Figure 7.18. It lies to the right of (and tangent to) a line through (~0.0556, 0) ‘whose slope is 1.6. Thus, we take 7 = 1/1.6 = 0.625 and k = 1/0.056 = 17.86. Let us verify that + = 0.62515 is acceptable. Since G(s) has relative degree four, the product CB is zero in any state-space realization of G(s). Thus, the condition 2/k + 7CB + 7B7CT > 0 is always satisfied. The poles of G(s) are at —1 and 2. Hence, the condition 1+ 7A # 0 is satisfied for 7 = 0.625 . Thus, the system is absolutely stable for the sector [0, 17.86] 0.05 0: 6 Sr 0 0102 03 Fee] Figure 7.18: Exercise 7.7(8) © 7.8 (a) Write V as Vast P eae? +b [ y(o) do 5 Since P = PT > 0 and ad? > 0, the first two terms form a positive definite quadratic function of (z, v). Since b > 0 and y is a first quadrant-third quadrant nonlinearity, the integral term is nonnegative. Hence, V is positive definite. It is also radially unbounded since V(z,v) > 27 Pz + adv? () V = 27(PA+ ATP): ~ 227 PBY — 2ad*wp + bi(CAz — CBY — dy) <_2(PA+ ATP): — 22" PBY — Badly ~ C2)b-+ '/CAz — HOBO? — bay? — A y(y — hy) oe aeatrys—acten pater ~atctye- (seats 2!) = 2™(PA+ATP)z—227(PB— wp? where w= PaTo? tad, y=0CB + bd + OE Choose b such that 7 > 0. Suppose there are P = PT > 0, L and ¢ >0 which satisfy PA+ATP = -LTL-eP PB = w-l'VF ‘Then mez Pz = 27 LT Lz + 227 LT Vb — Ww? we2™ Pa — (2717 - Jw\(Lz ~ V7) < -e2™Pe + wi6 CHAPTER 7. Hence, Vis negative semidefinite. V =05 2(t) =05 v(t) = ‘Thus, by LaSalle’s theorem, the origin is globally asymptotically stable. Now, by Lemma 6.3, the existence of P, L, and e satisfying the foregoing equations is equivalent to the strict positive realness of ls) = wl(sI- AB + by = ad [(C +nCA)(ot — AB + = v(t) 30 v(t) =0 ao E where 1 = 8/2ad, By Lemma 6.1, we want Rel2(jw)] > 0 forall ay that i, +nd+ cn} f+ Rel(C + 1CA) Gur - A)1B + n(d+CB)] >0 Noting that (L49s)0(sf- AB = C(sl— AB +980 [ +5At | B = Ole A)B+n0B +A [ir 4at | B = (C+nCA)(sI= AB + nCB and Re [o+i0n3 wwe obtain Rel(C-+nCA) iol ~ Ay" B + (4-408) = Re{(1+ ion [oGar ate + 4] and the condition for absolute stability reduces to TA Rel + jeMGGu)]>0, Vue R + 7.9 (a) From the block diagram, we have Els) = -¥ (8) ~ KU (6) = “{H(s) +4104 (s) = ~[H) + #120 (6) Hence, the system can be represented as the feedback connection of Figure 7.1 of the text with Langs) 44 = SESE +20-+8) Gls) = SIH(s) + A= ey ye +3) (b) From the Popov criterion (Exercise 7.8), the system is absolutely stable if 1 B ‘To simplify the calculations, choose n = }. Then +Ref(1 + jum)G(ju)| >0, VweR ku? 49k -3 Red + JemGGu)l = Tos whieh is positive for k > 4. Thus, we take ke = 3.uy * 7.10 qa) sin® @ = [}(1 — c0526)]° = 3(1 — 3.00526 + 3.cos? 26 — cos? 26] = } (1 - 300526 + § + $.c0s.40 — cos* 26) ‘The terms c0828, cos? 26, and cos 40 integrate to zero over 0 to . Thus ae w= 2 (2) B 2 [ots 0 a0 5 “a 4 8 J s2 6 a= in cos ~ 5 sind cosd — = cos Upon substituting the limit from 0 to 7, the first two terms vanish and the third term gives (16/15). Thus (3) We can write the given function as (y) = k +A sgn(y). The describing function of the signum nonlinearity is given in Example 7.6. Thus wayar+ 4 (4) Fora (a) =0. For a> A, v(asind) is given by =[% — bor0S0A’ (0) = { ‘The sum ¥; (a) + W2(a) gives the desired describing function. 0, fora< B —k-+kN ($) fora>B ©7211 (1) (iw) jo + jo) ~ jo +a?)us CHAPTER 7. Re[G(jw)] = Im[eGu)) i Im[G(ju)] =0 w= 1, ReG(j)] =—1 For U(y) = y°, we have (a) = 5a*/8. The equation 1 ~ #(a) = 0 has a unique solution a = (2)* = 1.126. ‘There is a possibility of a periodic solution of amplitude close to 1.125 and frequency close to 1 rad/sec. (2) The transfer function is the same as in part (1). The describing function is given by 0, fora 1. By using Matlab, it can be checked that the equation 1 - ¥(a) = 0 has no solution. So, we expect that the system will not have sustained oscillation. (3) 1 00) = Trio = 1-4 6(— jun) + 15(— ju)? + 20( je)? + 15(— ju) + (fa)? + (—i)® ata Le a + 15 = 8 + jf + 209 — 6] = aay Im[G(ju)] = 0 > -6 + 200 — 6. = 05 u? = 3 oru? Hi 1 a7 RelOU)]= A Rolo H)) = - From Example 7.6, we know that &(a) = 4/na. For w? = 3, the equation 1+ (a)Re{G] = 0 has no solution. For u? = }, the equation has a unique root a = 27/1éx. Thus, we expect that the system will have a periodic solution with amplitude close to 27/16 and frequency close to 1//3 rad/sec. (@) Gta) = —248, (6+ ju)(2 — ju)(3— he (24 + 0) — 5(36 = juGo+ Ga +3) w(t + w+ wD wd aO +a) Im[G(ju)] = 0+ 36-4? =0 50 =6 Re{G(j6)] = 1/30 and the equation 1+ ¥(a)ReG = 0 has the unique solution a = 2/15. Thus, we expect that the system will have a periodic solution with amplitude close to 2/15m and frequency close to 6 rad/sec. (5) jw jw w? + ju(1—w) jor wtjw (ut) po? In|G(ju)] =0 > w= Re{G(j)] = -1 and the equation 1+ B(a)ReG = 0 has the unique solution a = (8/5)!/*. Thus, we expect ‘that the system will have a periodic solution with amplitude close to (8/5)"/4 and frequency close to 1 rad/sec. (6) Gu) 5(1+ jw) —_ 5(1 + 4jw)(2— jo)? _ 5(4 + 150) + j200(3 TuO+ iu = Pa a sw=Vv3 Gu) = Im[G(ju)] = 0 (3-219 Re{G(jV3)] = —5/12. The equation 1 + W(a)ReG = 0 yields (a) = 12/5. Wa) =24 422% oe” 75 7 ‘Thus, we expect that the system will have a periodic solution with amplitude close to 10/ and frequency close to 1 rad/sec. (7) From part (6), the equation 1 + 4(a)ReG = 0 has the unique solution a = (8/5)'/*. 0, foract Fo)= fora>1 (0) < 4/ra. Hence, ¥(a) = 12/5 has no solution. Thus, we expect that the system will not have sustained oscillation. (8) From part (6), the equation 1+ (a)ReG = 0 has the unique solution a = (8/3)"/*, ‘The describing function is given in the solution of part (2). By using Matlab, it can be verified that (a) = 12/5 has no solution. Thus, we expect that the system will not have sustained oscillation. @) Gia) = ph, = Le ae) (1+ ju)® (l+u?)? InfGGu)]=0 > v@-u4)=0 = wavs Re{G(jV3)] = -1/8. The equation 1 + U(a)Re{G] = 0 yields W(a) = 8. From Example 7.6, the describing funetion of the signum function is ¥(a) = 4/(7a). The equation ¥(a) = § has the solution a = 1/(2n) ‘Thus, we expect that the system will have a periodic solution with amplitude close to 1/(2n) and frequency close to 8 rad/sec. (10) From part (9), ¥(a) = 8. From Example 7.7, the describing function of the saturation function satisfies ¥(a) < 1. Thus, ¥(a) = 8 has no solution and we expect that the system will not have sustained oscillation, «7.12 With reference to Section 1.2.4, represent the negative resistance oscillator in the z-coordinates Asm 2=-n—Me) where h(v) = —v +0? — tv. including the —v term with the linear system, we can represent the system in the form of Figure 7.1 with ] c=[0 1], vw =v- Fy s 0? + ju(1— a?) For 6)" —“Grapar Im G(jw)=0 > w=1 Re G(j)=-1 + 9a) On the other hand, 2 a! 78 ¥(a) = 2f [lesinay? = sing] sing ao = ‘The equation 22 ~ $f = 1 has two solutions: a = 2 and a = V3, Thus the harmonie balanoe equation hhas two solutions: (i, a,) = (1,2) and (15, a.) = (1, 2). We expect that the system will have two periodic solutions, the first one has amplitude close to 2 and frequency close to 1 rad/sec, and the second one has120 CHAPTER 7. amplitude close to /@ and frequency close to 1 rad/sec ‘The phase portrait (Figure 7.19) shows that the system has two limit cycles. The amplitude of 22 in the inner limit cycle is about 1.4 while that of the ‘outer limit cycle is about 2, ‘These numbers correlate reasonably well with the estimates of the amplitude of the first harmonic obtained using the describing function method. From Figure 7.20, we can estimate the frequency of oscillation by 0.9657 rad/sec, which is close to the frequency of oscillation predicted by the describing function method. 4 2 al 1 wo] } wo 2 a 4 se a a ) % ' Figure 7.19: Exercise 7.12. Figure 7.20: Exercise 7.12, eras 26%? + j2bu(1 Cera In[G(ju)] =0 3 w=1 Re{G(j)] = -2. The equation 1+ ¥(a)ReG = 0 yields (a) = }. The describing function of the saturation nonlinearity is given in Example 7.7. By using Matlab, it can be verified that ¥(a) = } has a solution a= 247. Thus, we expect that the system will have a periodic solution with amplitude close to 2.47 and frequency close to 1 rad/sec. To confirm this conjecture, we apply Theorem 7.4. The saturation nonlinearity satisfies the slope restriction with a = 0 and = 1. lee ae Glia) ~~ 2s 2? Gu) = a) ‘The inverse Nyquist plot is a vertical line through the point (—-0.5,0). The critical circle has its center at the point (~0.5,0) and its radius is 0.5. The inverse Nyquist plot intersects the critical circle at the points (-0.5,-0.5) and (—0.5,0.5). The frequencies at these two points are given, approximately, by w = 1 — 6/2 and w = 1+6/2. So, we limit our analysis to the frequency band w € (1 —b/2,1 + 6/2] lew? abs 2 wal jes? | For w > 1, the right-hand side is a monotonically increasing function of w. Hence, p(w) is calculated at k tobe ou? plu) = For 1 Sw <1+8/2, 9040/2 =1 4 (1498/8) 65(1 + 872) a( sat) 4 3 SS11 ‘Thus, for small b, p(w) is of order O(1/6). Consequently, Oe plw)—0.5 ~ 16 ‘The uncertainty band will be very narrow for sufficiently small 6, It can be verified that the uncertainty Sand defies a complete intereecion. Theorem 74 cones that thee isa ballwave symmetie prod solution. The amplitude of oscillation will be O(6) close to 2.47 and the frequency of oscillation will be O(b) close to 1 rad/sec. o(w) 00) © 7.14 (a) #€ [0,1], Since G(s) is Hurwitz, we apply case 2 of the circle criterion. From the Nyquist plot, ‘we see that Re{G(jw)] > -0.085. Thus, the largest 6 is b = 1/0.085 = 11.76. (b) The Popov piot lies to the right of a line having slope = 1.2 and intersecting the real axis at ~0.06 ‘The slope corresponds to 7 = 1/1.2 = 0.833. Since the eigenvalues of A are at —1 and —2, the condition (1+ Ay) 4 0 is satisfied. Also, since G(s) has relative degree higher than one, CB = 0. Hence, the condition 2+ 2ykCB > 0 is satisfied for any 7. Thus, the choice 7” = 0.833 is acceptable and the largest bis, b= 1/0.06 = 16.67. (©) W is a special case of the piecewise linear function of Example 7.7 with 1 =, s2 = 0, and 6 = 1/b. The describing function is wm] aft), 2 iy? wO=T [o ) + ay () (Lu? ~ 25u)(4-u? ~ 4jw) 2(6 ~ au) 9) 1+ ¥(a)Re[G(jv2)] = 0 > U(a) = 18 Because (a) starts from 6 at a = 0 and decreases after a = 1/b, the equation (a) = 18 has a solution if b> 18. The frequency of oscillation will be close to w = v2. (d) For b = 10, the slope restrictions are a = 0 and 8 = 10. The inverse Nyquist plot and the critical circle are shown in Figure 7.21. The inverse Nyquist plot is plotted only for w € [0,2]. It keeps moving farther away for higher frequencies. Because the inverse Nyquist plot is always outside the critical circle, there is no oscillation. (c) For b = 30, the slope restrictions are a = 0 and = 30. The inverse Nyquist plot and the critical circle are shown in Figure 7.23. The error circle are plotied for the following points: w. T/GG0) | aw) TS] (15:1130,2.418) | 0.7536 1.389 | (-17.3589,0.589) | 0.5911 1.436 | (-18.555,-0.535) | 0.5231 145 | (-18.912,-0.8917) | 0.5048 1.5_| (-20.1875-2.25) | 0.4456 ‘The circles tangent to the real axis correspond to w = 1.389 and w = 1.436, Thus, w = 1.389 rad/sec and w, = 1436 rad/sec. ‘The boundaries of the band of uncertainty intersects the real axis at 17 and =18.75. By plotting (a) versus a, the values of a; and az are determined as a, = .0649 and a = 0.0721. It can be verified that the regularity conditions are satisfied. Therefore, there is oscillation of frequency w € [1.389, 1.436] and the amplitude of the first harmonic of the output is a € (0.0649, 0.0721). At w = 1.389 and a = 0.0721, we have GGGu) Im[GGw)] Osw=v2 =8x10-*12 CHAPTER 7. ‘The energy content in the first harmonic is 2 f Oe 8 situ) 52x 107? (0.0721)? = the ratio is 15.98 x 10~*. Taking the square root, the percentage of the higher harmonics to the first one is 3.9%. The simulation is shown in Figure 7.22. The simulation shows a period of oscillation of about 4.4 sec., that is, w ~ 1.428 rad/sec. The amplitude of oscillation is about 0.068. ‘These results are consistent ‘with the ones we obtained using the describing function method. e a1 , oss i : io i 4 2-10 aoe 18} / G4 @) =e falta Figure 7.21: Exercise 7.14 Figure 7.22: Exercise 7.14 + 7.15 (a) The nonlinearity belongs to the sector (0,0), By using Matlab to find the Nyquist plot of G(jw), we found that min Re(G(jw)] = ~1.1638. Hence, the origin is globally asymptotically stable for 1b <1/1.1638 = 0.8593. (b) By using Matlab, we found that the Popov plot intersects the real axis at —0.5564. With 7 = 1/11 we ‘can draw a line of slope 1/77 which passes through the point (~-0.5564, 0) and is to the let of the Popov plot. ‘The choice ~y = 1.1 satisfies the condition 1+ 7A # 0 since A= —1 or ~2. The condition 1 + ykCB > Ois satisfied for any + since CB = 0 (the transfer function has relative degree two). Thus, the origin is globally asymptotically stable for 6 < 1/0.5564 = 1.7972, (c) W is a special case of the piecewise linear function of Example 7.7 with sy = b, 62 = 0, and 6=1/b. The ‘citing non v2 [a (3)+2)1-(2) ‘The harmonic balance equation is 1 1+GGu)H) =0 & aS +¥(a) =0 hi 24") + jul -47)] + H(a) =0 ‘The imaginary part yields u? = 5; then the real part yields W(a) = 1.8. Because Y(a) starts from b at a= 0 and decreases after a = 1/b, the equation (a) = 18 has a solution if b > 1.8. Thus, the describing function method predicts that there will oscillation for > 1.8. The frequency of oscillation will be close to w = V5 = 2.2361.123 Rss OC ° 5 eal Ads Figure 7.23: Exercise 7.14. © 7.16 (a) . (1-344) + ju? - 3) G+u) Im Gjw) =0 + w= V3 Re Gv3)=- 2 G(jw) = => Ha) 3 With s; = 10, 69 = 0 and 6 = 0.1, the describing function is given by w= [ao (i) + say = CB) Plotting U(a) using Matlab, we determined that {(a) = 8 corresponds to a = 0.146. Thus we predict that there is an oscillation of frequency close to V3 rad/s and amplitude of the first harmonic close to 0.146. (b) —— GGa) ‘The inverse Nyquist plot is shown in Figure 7.24, together with the critical circle and the band of uncertainty. It defines a complete intersection with wy = 1.699, uw = 1.76, a, = 0.139, and az = 0.1535. Thus, by ‘Theorem 7.4 we conclude that there is an oscillation of frequency w € (1.699, 1.76) and amplitude of the first harmonic a € (0.139, 0.1535). (c) Figure 7.25 shows the inverse Nyquist plot with three critical circle corresponding to k = Bu? + ju(3 0) , 7.5, 10. With124 CHAPTER 7. k =4, the inverse Nyquist plot lies outside the critical circle. It follows from the first part of Theorem 7.4 that there is no oscillation. This however is a conservative estimate because we can use the second part of the theorem to conclude that there is no oscillation if the inverse Nyquist plot enters the critical circle but for each point inside the critical circle the error circle does not intersect the real axis. The critical circle with k-=7.5 captures this situation. Therefore, we conclude that the largest slope for which the system does not oscillate is about 7.5. © 7.17 (1) Peay Ga) "Ge +1) ~ “oS + 0?) Im[G(ju)] = 00 =1 Re{G(j)] = -2 and the equation 1+(a)Re(G] = O yields (a) = }. The describing function of the saturation nonlinearity is given in Example 7.7. The equation #(a) = } has a unique solution a = 2.47. Thus, we expect the system to have a periodic solution with amplitude close to 2.47 and frequency close to 1 rad/sec. ‘To confirm this conjecture, we apply Theorem 7.4. The uncertainty band defines a complete intersection. ‘The sketch in Figure 7.26 shows the details of the compte intersection. ‘The two circles (almost) tangent to the real axis have wy = 0.948, ¥(a;) = 0.4255 and uy = 1.032, (az) = 0.5494, From the describing function of the saturation nonlinearity, we find that (a) = 0.4255 gives a = 2.933 and ¥(a) = 0.5494 gives a= 2.238. Thus, Theorem 7.4 confirms that there is a half-wave symmetric periodic solution with frequency in the range (0.948, 1.032] and the amplitude of the first harmonic is in the range [2.238, 2.933]. (2) From Example 7.14, we know that the harmonic balance equation has a solution with w = 2yV and ‘V(a) = 0.8. The inverse Nyquist plot is shown in Figure 721 of the text. For the nonlinearity Y(y) = }siny, the slope restriction is satisfied with a = - and 8 = 3. Therefore, the critical circle is centered at the origin and has a radius of 0.5. Neither the inverse Nyquist plot nor the uncertainty band enter the critical circle. Hence, by the first part of Theorem 7.4, we conclude that there is no oscillation (3) From Example 7.14, we know that the harmonic balance equation has a solution with w = 2vV2 and Y(a) = 0.8. The inverse Nyquist plot is shown in Figure 7-21 of the text. The nonlinearity satisfies the slope restriction with a = —1 and 6 = 1. Therefore, the critical circle is centered at the origin and has a radius Gu) =125 Figure 7.25: Exercise 7.16. ¥ 005 5 E -0.05] - iGo) 06 05 04 03 Real Axis Figure 7.26: Exercise 7.17(1) of 1. The describing function is given by 1 oss? (a) = ¥ (8) sin (2) + VT OTOP 1 ae $ ir (2) -sin(2)] +2 VI= OPP B VIM BOF +1 ads Itis shown in Figure 7.27. The equation #(a) = 0.8 has two roots at a= 2.48 and a = 12.47. Thus, we expect that there are two periodic solutions, one with amplitude close to 2.48 and frequency close to 2/2 rad/sec. and the other with amplitude close to 12.47 and frequency close to 2V2 rad/sec. Next, we apply ‘Theorem 7.4. It is clear from Figure 7.27 that (a) is different than zero at the roots a,. Thus, all the conditions of Theorem 7.4 (case 3) are satisfied and we confirm the existence of the periodic solutions. By calculating and plotting the uncertainty band, it can be seen that the boundaries of the band intersect the real axis at ~0.6481 and ~0,9519, The frequencies of the error circles closest to the real axis are w = 1.75126 CHAPTER 7. and w = 2.905. Using the graph of (a), the set T corresponding to each solution is given by Ty = {(w,a) | 2.75 0. ®) 2-31 of a=] 217 To HL [F Fe] - nt ‘Take y = 25 and | ga(a-w, a= -ta-Setei tla, eat Ns (h(a) = By (u) (bay) ~ a}? + 2ha (4) + Shay) — 9? = [ia(v) — vP? Nal) = Raha) ~ uP — h(a) ~ hal) ha(0) = 0, h4(0) ‘Try oy) =0. Mi(0) = Olly?), N2(0) =0 ‘The origin of the reduced system y = y” + O(|y|*) is unstable. Hence, the origin of the full system is unstable.