Linear System

Theory and Design

主讲：朱芳来
Part 1
 Chapter 1: Introduction
1.1 Introduction
™ Two methods of study and design of
physical systems
Empirical method:
Various signals are applied to a physical
system and its responses are measured. If
the performance is not satisfactory, some of
the parameters are adjusted to improve its
performance.
 Analytical method:
The analytical study of physical systems
consists of four parts: modeling, development
of mathematical descriptions, analysis, and
design.
™ Modeling
A physical system is a device or a collection
of devices existing in the real word;
A physical system may have different
models depending on the questions asked
 A physical system may modeled differently
in different operational ranges
In order to develop a suitable model for a
physical system, a thorough understanding of
the physical system and its operational range
is essential.
A system is a model of a physical system
™ Mathematical description
Once a system (or model) is selected for a
physical system, the next step is to apply
various physical laws to develop mathematical
equations to describe the system.
The equations that describe systems may be:
Linear equations;
Nonlinear equations;
Integral equation;
Difference equations;
Differential equations
™ Analysis
Quantitative analysis:
In quantitative analysis, we are
interested in the responses of systems
excited certain inputs;
Qualitative analysis:
we are interested in the general
properties of systems, such as stability,
controllability, and observability
 Chapter 2: Mathematical
Descriptions of Systems
2.1 Introduction
The class of systems studied in this text is
assumed to have some input terminals and
output terminals as shown in Fig. 2.1

Figure 2.1 System

™ We assume that if an excitation or input is
applied to the input terminals, a unique
response or output signal can be measured
at the output terminals.
™ A system with only one input terminal and
only one output terminal is called a single-
input single-output (SISO) system.
™ A system with two or more input terminals
and/or two or more output terminals is called
a multivariable (SIMO,MISO,MIMO ) system.
™A system is called a continuous-time system
if it accepts continuous-time signals as its
input and generates continuous-time signals
as its output.
™ A system is called a discrete-time system if
it accepts discrete-time signals as its input
and generates discrete-time signals as its
output.
2.1.1 Causality and Lumpedness
™ A system is called a memoryless systems if
its output y(t0) depends only on the input
applied at t0; it is independent of the input
applied before or after t0.
™ A system is called a cause or nonanticipatory
system if its current output depends on past
and current inputs but not on future input.
™ If a system is not causal, its current output will
depend on future input. No physical system
has such capability. This text studies only
causal systems.
 Definition 2.1 The state x(t0) of a system at
time t0 is the information at t0 that, together
with the input u(t), for t≥t0, determines
uniquely the output y(t) for all t≥t0.

™ The state summarizes the effect of past input

on future output. So, if we know the state at t0,
there is no more need to know the input u(t)
applied before t0 in determining the output y(t)
after t0.
™ An example for the network shown in Fig.2.2

Figure 2.2 Network with 3 variables

™ For the network shown in Fig.2.2, if we know the
voltages x1(t0) and x2(t0) across the two capacitors
and the current x3(t0) passing through the inductor,
then for any input applied on and after t0, we can
determine uniquely the output for t≥t0. Thus the
state of the network at time t0 is
⎡ x1 (t0 ) ⎤
x(t0 ) = ⎢⎢ x2 (t0 )⎥⎥
⎢⎣ x3 (t0 ) ⎥⎦

The entries of x are called state variables

™ Using the state at t0, we can express the input and
output of a system as

x(t 0 ) ⎫
⎬ → y (t ), t ≥ t 0 (2.1)
u (t ), t ≥ t 0 ⎭

™ It means that the output is partly excited by the

initial state at t0 and partly by the input applied at
and after t0.
™ A system is said to be lumped if its number of state
is a finite vector. A system is called a distributed
system if its state has infinitely variable
™ Example 2.1 Consider the unit-time delay
system defined by
y(t)=u(t-1)
The output is simply the input delayed by one
second. In order to determine {y(t),t≥t0} from
{u(t),t≥t0}, we need the information {u(t),t0-
1≤t<t0}, Therefore the initial state of the
system is {u(t),t0-1≤t<t0}. Thus the unit-time
delay system is a distributed system.
2.2 Linear System
™ A system is called a Linear system if for
every t0 and any two state-input-output pairs
x i (t0 ) ⎫
⎬ → y i (t ), t ≥ t0
u i (t ), t ≥ t0 ⎭

for i=1,2, we have

x1 (t0 ) + x 2 (t0 ) ⎫
⎬ → y1 (t ) + y 2 (t ), t ≥ t0 (additivity)
u1 (t ) + u 2 (t ), t ≥ t0 ⎭
and
 αx1 (t0 ) ⎫
⎬ → αy1 (t ), t ≥ t0 (homogeneity)
αu1 (t ), t ≥ t0 ⎭

for any real constant α . The first property is

called the additivity property, the second, the
homogeneity property. The two properties
can be combined as
α1x1 (t ) + α 2 x 2 (t ) ⎫
⎬ → α1y1 (t ) + α 2 y 2 (t ), t ≥ t0
α1u1 (t ) + α 2u 2 (t ), t ≥ t0 ⎭

for any real constants α 1 andα 2 , and it is called

 the superposition property.
™ A system is called nonlinear system if the
superposition property does not hold.
™ If the input u(t) is identically zero for t≥t0,
the output will be excited exclusively by the
initial state x(t0). This output is called the
zero-input response and will be denoted as
yzi or
x(t0 ) ⎫
⎬ → y zi (t ), t ≥ t0
u(t ) ≡ 0, t ≥ t0 ⎭
™ Ifthe initial state x(t0) is zero, the output will
be excited exclusively by the input. This
output is called the zero-state response and
will be denoted as yzs or
x(t0 ) = 0 ⎫
⎬ → y zs (t ), t ≥ t0
u(t ), t ≥ t0 ⎭

The additivity property implies

 ⎧x(t0 ) ⎧x(t0 )
Output due to ⎨ =output due to ⎨
⎩u(t ), t ≥ t0 ⎩u(t ) ≡ 0, t ≥ t0

⎧x(t0 ) = 0
+ output due to ⎨
⎩u(t ), t ≥ t0

or
Response=zero-input response
+zero-state response
™ Input-output description We develop a
mathematical equation to describe the zero-
state response of linear systems. The SISO
linear system is considered first.
™ Let δ Δ (t − t1 ) be the pulse shown in Fig 2.3

1
Δ
Δ

t1 t1+ Δ

Figure 2.3 Pulse at t1

The description of δ Δ (t − t1 ) is
⎧1
⎪ t1 ≤ t ≤ t1 + Δ
δ Δ (t − t1 ) = ⎨ Δ
⎪⎩0 others

™ For
any input u(t), it can be expressed
approximately
u (t ) ≈ ∑ u (t i )δ Δ (t − t i )Δ
i
as shown in Figure 2.4.
Figure 2.4 Approximation of input signal
™ Let g Δ (t , t i ) be the output at time t excited by the
pulse u (t ) = δ Δ (t − ti ) applied at time ti. Then we
have
δ Δ (t − t i ) → g Δ (t , t i )

δ Δ (t − t i )u (t i )Δ → g Δ (t , t i )u (t i )Δ (homogeneity)

∑δ
i
Δ (t − t i )u (t i )Δ → ∑ g Δ (t , t i )u (t i )Δ
i
(additivity)
™ Thus the output y(t) excited by the input u(t)
can be approximated by
y (t ) ≈ ∑ g Δ (t , t i )u (t i )Δ (2.2)
i
™ Now if Δ approaches zero, the pulse δ Δ (t − t1 )
becomes an impulse at ti, denoted by δ (t − t i )
that is
⎧∞ t = t0
δ (t − t 0 ) = lim δ Δ (t − t 0 ) = ⎨
Δ →∞
⎩0 t ≠ t0
or
⎧∞ t=0
δ (t ) = lim δ Δ (t ) = ⎨
Δ →∞
⎩0 t≠0

™ The δ (t )
is called impulse function. It has the
property as follows:
∞
∫ −∞
f (t )δ (t − t 0 )dt = f (t 0 )
™ As Δ approaches zero, the approximation
in (2.2) becomes an equality and the
summation becomes an integration:
∞
y (t ) = lim ∑ g Δ (t , ti )u (ti )Δ = ∫ g (t ,τ )u (τ )dτ
Δ →0 −∞
i

That is
∞
y (t ) = ∫ g (t ,τ )u (τ )dτ (2.3)
−∞

™ The function g (t ,τ ) is the response at time t

excited by an impulse δ (t ) at time τ , and it
is called impulse response.
™A system is said to be relaxed at t0 if its initial
state is 0. Every linear system that is
causal and relaxed at t0 can be described
by
τ
y (t ) = ∫ g (t ,τ )u (τ )dτ (2.4)
0 t

™ we call the above equation the input-output

description.
™ If a linear system has p input terminals and q
output terminals, the input-output description
of it can be then extended to
 τ
y (t ) = ∫ G (t ,τ )u(τ )dτ
t0
where

⎛ g11 (t ,τ ) g12 (t ,τ ) L g1 p (t ,τ ) ⎞
⎜ ⎟
⎜ g 21 (t ,τ ) g 22 (t ,τ ) L g 2 p (t ,τ ) ⎟
G (t ,τ ) = ⎜ ⎟
M M M
⎜ ⎟
⎜ g (t ,τ ) g (t ,τ ) L g (t ,τ ) ⎟
⎝ q1 q2 qp ⎠
and g ij (t ,τ ) is the response at time t at the ith output
terminals due to an impulse applied at time τ at
the jth impulse terminal, the inputs at other
terminals being identically zero. The G is called
the impulse response matrix of the system
™ State-space-description
™ Everylinear lumped system can be described
by a set of equations of the form
x& (t ) = A (t ) x(t ) + B (t )u (t ) (2.6)
y (t ) = C(t ) x(t ) + D(t )u (t ) (2.7)

dx(t )
where x& (t ) := , x∈R ,
n
u ∈ R p
, y ∈ R q

dt
2.3 Linear Time-Invariant (LTI) Systems
™ A system is said to be time invariant if for
every state-input output pair

x(t 0 + T ) ⎫
⎬ → y (t − T ), t ≥ t 0 + T
u(t − T ), t ≥ t 0 + T ⎭

™ Input-output description If the system is

time invariant, we have
g (t ,τ ) = g (t + T ,τ + T ) = g (t − τ ,0) = g (t − τ )

for any T.
™ For linear time-invariant system, the (2.4)
reduces to
y (t ) = ∫ g (t − τ )u (τ )dτ = ∫ g (τ )u (t − τ )dτ (2.8)
t t

0 0
The above integration is called a convolution
integral.
™ Example 2.2 The unit-time delay system
studied in Example 2.1 is a device whose
output equals input delayed by 1 second. If
the impulse δ (t ) is applied at the input terminal,
the output is δ (t − 1) . Thus the impulse
response δ (t − 1) .
™ Example 2.3 Consider the unit-feedback
system shown in Fig. 2.5(a).

r(t) + u(t) Unit-time y(t)

a delay element
+

Figure 2.5 Positive and negative feedback system

™ If r (t ) = δ (t ) , then the output is

 g f (t ) = aδ (t − 1) + a 2δ (t − 2) + a 3δ (t − 3) + L
∞
= ∑ a iδ (t − i ) (2.9)
i =1

™ Letr(t) be any input with r (t ) ≡ 0 for t<0; then

the output is given by
∞
y (t ) = ∫ g f (t − τ ) r (τ ) dτ = ∑ a
t t
∫ δ (t − τ − i )r (τ )dτ
i
0 0
i =0
∞ ∞
= ∑ a i r (τ ) = ∑ a i r (t − i )
τ =t −i
i =1 i =1
™ Transfer-function matrix The Laplace
transform is an important tool in the study of
linear time-invariant (LTI) systems. Let yˆ ( s ) be
the Laplace transform of y(t), that is
∞
yˆ ( s) = L( y (t )) = ∫ y (t )e − st dt
0

™ For causal systems, we have g (t ) = 0 for t < 0 or

g (t − τ ) = 0 for τ > t . So the Laplace transform of
y(t) in (2.8) is
 ∞ ∞
yˆ ( s ) = ∫ (∫ g (t − τ )u (τ ) dτ )e − s ( t −τ ) e − sτ dt
t =0 τ =0

∞ ∞
= ∫ （∫ g (t − τ )e − s ( t −τ ) dt )u (τ )e − sτ dτ
τ =0 t =0

∞ ∞
=∫ (∫ g (v )e − sv dv )u (τ )e − sτ dτ
τ =0 v = −τ

∞ ∞
=∫ (∫ g (v )e − sv dv )u (τ )e − sτ dτ
τ =0 v =0

∞ ∞
=∫ g (v )e − sv
dv ∫ u (τ )e − sτ dτ
v =0 τ =0
™ That is
yˆ ( s ) = gˆ ( s )uˆ ( s )

where ∞
gˆ ( s ) = L( g (t )) = ∫ g (t ) s − st dt
0

is called the transfer function of the system.

™ The transfer function is the Laplace transform
of the impulse response and, conversely, the
impulse response is the inverse Laplace
transform of the transfer function.
™ Fora p-input q-output system, (2.10)can be
extended as
⎛ yˆ1 ( s) ⎞ ⎛ gˆ 11 ( s) gˆ 12 ( s) L gˆ 1 p ( s ) ⎞⎛ uˆ1 ( s ) ⎞
⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ yˆ 2 ( s) ⎟ ⎜ gˆ 21 ( s ) gˆ 22 ( s ) L gˆ 2 p ( s ) ⎟⎜ uˆ 2 ( s ) ⎟
⎜ M ⎟=⎜ M M M ⎟⎜ M ⎟
⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ yˆ ( s) ⎟ ⎜ gˆ ( s ) gˆ ( s) L gˆ ( s ) ⎟⎜ uˆ ( s) ⎟
or ⎝ q ⎠ ⎝ q1 q2 qp ⎠⎝ p ⎠

yˆ ( s ) = G
ˆ ( s )uˆ ( s )
where gˆ ij ( s) is the transfer function from jth
input to ith output. The q × p matrix Gˆ ( s) is
called the transfer-function matrix of the
system.
™ Example 2.4 Consider the unit-time delay
system
y (t ) = u (t − 1)

Its impulse response is δ (t − 1) . Therefore its

transfer function is
∞
gˆ ( s ) = L(δ (t − 1)) = ∫ δ (t − 1)e − st dt = e − st = e −s
0 t =1

This transfer function is an irrational function of s

™ Example 2.5 Consider the feedback system
shown in Fig.2.5(a). The transfer function from
r to y can be computed directly from the block
diagram as
 ae − s
gˆ f ( s ) =
1 − ae − s

™ Thiscan also be obtained by taking the

Laplace transform of the impulse response,
which was computed in (2.9) as
∞
g f (t ) = ∑ a i δ (t − i )
i =1

™ Because L(δ (t − i )) = e −is , the Laplace transform

of gf(t) is
 ∞ ∞
gˆ f ( s ) = L( g f (t )) = ∑ a e
i −is
= ae −s
∑ )
( ae −s i

i =1 i =0

ae − s
=
1 − ae − s

™ The transfer function in (2.12) is an irrational

function of s. This is because the feedback
system is a distributed system.
™ If a linear time-invariant system is lumped,
its transfer function will be a rational function
of s.
™ Every rational transfer function can be
expressed as
N (s)
gˆ ( s ) =
D( s)

where N(s) and D(s) are polynomials of s .

The gˆ ( s) can be classified as follows:
™ gˆ ( s ) proper ⇔ degD(s)≥degN(s) ⇔

ĝ (∞) = zero or nonzero constant

™ gˆ ( s ) strictly proper ⇔ degD(s)>degN(s)

⇔ gˆ ( s ) = 0
 ™ gˆ ( s ) biproper ⇔ degD(s)=degN(s)
⇔ ĝ (∞) = nonzero constant
™ gˆ ( s ) improper ⇔ degD(s)<degN(s)
⇔ ĝ (∞ ) = ∞

State-space equation Every linear time-

invariant lumped system can be described
by a set of equation of the form
x& (t ) = Ax(t ) + Bu (t ) (2.13)
y (t ) = Cx(t ) + Du (t )
 where A ∈ R n×n , B ∈ R n× p , C ∈ R q×n and D ∈ R q× p
™ Applying the Laplace transform to (2.13)
yields
sxˆ ( s ) − x(0) = Axˆ ( s ) + Buˆ ( s )
yˆ ( s ) = Cxˆ ( s ) + Duˆ ( s )

If the initial state x(0) is zero, then we have

yˆ ( s ) = [C( sI − A ) −1 B + D]uˆ ( s )
So, the transfer function of this system is
ˆ ( s ) = C( sI − A ) −1 B + D
G (2.16)
 2.5 Examples In this section, we use
example to illustrate how to develop transfer
function and state-space equation for physical
system.

™ Example 2.6 Consider the mechanical system

in Fig. 2.8.
Figure 2.8 Mechanical system
™ Based on Newton’s law, we have
m&y& = u − k1 y& − k 2 y
Applying the Laplace transform and assuming
zero initial conditions, we obtain
ms 2 yˆ ( s) = uˆ ( s ) − k1 syˆ ( s ) − k 2 yˆ ( s )
which implies
1
yˆ ( s ) = uˆ ( s )
ms + k1 s + k 2
2

that is
1
gˆ ( s ) =
ms 2 + k1 s + k 2
™ Ifm=1, k1=3, k2=2, the impulse response of
the system is
⎡ −1 1 ⎤ −1 ⎡ 1 1 ⎤ −t − 2t
g (t ) = L ⎢ 2 = L − = e − e
⎣ s + 3s + 2 ⎥⎦ ⎢⎣ s + 1 s + 2 ⎥⎦

and the convolution description of the system

is
∞ ∞
y (t ) = ∫ g (t − τ )u (τ )dτ = ∫ (e −( t −τ ) − e − 2 ( t −τ ) )u (τ )dτ
0 0
™A state-space description for the system:
Selecting the displacement and the velocity
of the block as state variables, that
is, x1 = y, x2 = y&. Then we have
x&1 = x 2
mx& 2 = −k 2 x1 − k1 x 2 + u

In matrix form, they become

⎛ x&1 (t ) ⎞ ⎛ 0 1 ⎞⎛ x1 (t ) ⎞ ⎛ 0 ⎞
⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟⎜⎜ ⎟⎟ + ⎜⎜ ⎟⎟u (t )
⎝ x& 2 (t ) ⎠ ⎝ − k 2 / m − k1 / m ⎠⎝ x2 (t ) ⎠ ⎝1 / m ⎠
⎛ x1 (t ) ⎞
y (t ) = (1 0 ) ⎜⎜ ⎟⎟
⎝ x2 (t ) ⎠
2.7 Concluding Remarks
System type Internal description External description

Distributed, t
y (t ) = ∫ G (t ,τ )u (τ )dτ
linear t0

Lumped, x& = A(t ) x + B (t )u t

y (t ) = ∫ G (t ,τ )u (τ )dτ
linear y = C (t ) x + D (t )u t0
t
Distributed, y (t ) = ∫ G (t − τ )u (τ )dτ
0
linear, time-
invariant yˆ ( s ) = Gˆ ( s )uˆ ( s ), Gˆ ( s ) irrational
t
Lumped, x& = Ax + Bu y (t ) = ∫ G (t − τ )u (τ )dτ
0
linear, time-
y = Cx + Du yˆ ( s ) = Gˆ ( s )uˆ ( s ), Gˆ ( s ) rational
invariant