Contact me in order to access the whole complete document. Email: smtb98@gmail.

com
WhatsApp: https://wa.me/message/2H3BV2L5TTSUF1 Telegram: https://t.me/solutionmanual

INTRODUCTION 2
ssm
Chapter 1: Introduction
1.3.5 Mobile Robot Engineering
mtt
1.3.5.1 Mobility
Comment in one or two sentences for each subsystem on how the goal of mobility
requires that a mobile robot have such a subsystem.
• position estimation
bb99
• perception
• control
• planning
• locomotion
• power/computing
1.3.5.1 Mobility: Solution
88@
a) Closed loop directed motion requires position feedback. Open loop control not usually
feasible.
b) Must perceive environment in order to compute terrainability or avoid obstacles, or to
satisfy mission objectives (not mobility).
@
c) Must be able to execute motion commands on real hardware.
d) Must look ahead to predict consequences of decisions now. In this way, avoid getting
trapped.
e) Must move.
f) Mobile robots must carry their own power and smarts wherever they go.
ggm
maa
iill..cc
oom
m

complete document is available on https://unihelp.xyz/ *** contact me if site not loaded

MATH FUNDAMENTALS 3

Chapter 2: Math Fundamentals

2.2.9 Matrices
2.2.9.1 Matrix Exponential
Show that:

  6
exp  6 0  = e 0
 0 –2  0 e
–2

2.2.9.1 Solution
Using the formula for matrix exponential:

  2
1
exp  6 0  = I + 6 0 + ----- 6 0 + 
 0 –2  0 – 2 2! 0 – 2

1 2
 6 0  1 + 6 + ---  6  +  0 6
exp  = 2 = e 0

 0 –2  1 2
0 e
–2
0 1 – 2 + ---  – 2  + 
2

2.2.9.2 Jacobian Determinant

You probably learned in multivariable calculus that the ratio of volumes in a 2D linear
mapping is given by the Jacobian determinant. Consider the matrix:

A = a b
c d
T T
and the vectors dx = dx 0 and dy = 0 dy . Using the result for the area of a
parallelogram defined by two vectors in Figure 2.8, show that the area formed by the
vectors du = Adx and dv = Ady is det  A dxdy .
2.2.9.2 Solution
The output vectors are:

du = Adx = a b dx = adx dv = Ady = a b 0 = bdy

c d 0 cdx c d dy ddy
The area of the output region is:
Area = u x v y – u y v x = adxddy – cdxbdy =  ad – cb dxdy

2.2.9.3 Fundamental Theorem and Projections

2
A matrix P is called a projection matrix if it is symmetric and P = P which is called
the property of idempotence.
MATH FUNDAMENTALS 4
(i) What happens if you compute p = Px and then p = Pp ? An important
1 2 1
projection matrix can be derived from a general n  m matrix A (where m  n ) as
T –1 T
follows: PA = A  A A  A .
(ii) Show that P A satisfies both requirements of a projection matrix.
(iii)Note that p = P A x must reside in the column space of A . The orthogonal
1
complement Q A of P A is defined as Q A = I – P A . Note that
Q A P A = P A Q A = 0 . In what subspace does q = Q A x reside?
1
2.2.9.3 Solution
a) You get p = p . The second projection does nothing.
2 1
b) The requirements are:
T T –1 T
PA = A  A A  A = PA
T –1 T T –1 T T –1 T
PA PA = A  A A  A AA A A
= PA = AA A A
c) Q A extracts the component of x which has no projection onto the rows of A . Hence, it
is in the nullspace of A .

2.2.9.4 Derivative of the Inverse

–1
Suppose a square– 1matrix A  t  depends on a scalar, say t . Differentiate A A and find an
expression for A· .
2.2.9.4 Solution
The time derivative of the matrix inverse is relatively easy to compute. Because:

–1
d A A d I 
= = 0
dt dt
We have:
–1 –1
A· A + A A· = 0
Which gives the time derivative of the matrix inverse as:
–1 –1 –1
A· = – A A· A

2.3.5 Fundamentals of Rigid Transforms

2.3.5.1 Specific Homogeneous Transforms
What do the following transforms do? If it is not obvious, transform the corners of a
square to find out.
1 0 0 a a 0 0 0 1 0 a 0 1 0 0 0 1 0 0 0
0 1 0 b 0 b 0 0 0 1 b 0 0 1 0 0 0 1 0 0
0 0 1 c 0 0 c 0 0 0 1 0 0 0 1 0 0 0 0 0
0 0 0 1 0 0 0 1 0 0 0 1 0 0 1d 1 0 0 0 1

2.3.5.1 Solution
From left to right: translation, scale, xy shear (preserves z), perspective projection to plane
at z = d, orthographic projection to plane at z = 0.
MATH FUNDAMENTALS 5
2.3.5.2 Operators and Frames
2D Homogeneous transforms work just like 3D ones except that a rigid body in 2D has
only three degrees of freedom – translation along x or y or rotation in the plane. Consider
the transform:
0 –1 7
T = 1 0 3
0 0 1
(i) Recall that the unit vectors and origin of a frame can be represented in its own
coordinates as an identity matrix. Let such a matrix represent frame a . Consider
the T matrix to be an operator and operate on the unit vectors and origin of frame
a expressed in its own coordinates to produce another frame, called b . Write
explicit vectors down for the unit vectors and origin of the new frame b . Use a
notation that records the coordinate system in which they are expressed.
(ii) Visualization of the New Frame. When a transform is interpreted as an operator,
the output vector is expressed in the coordinates of the original frame. Get out
some graph paper or draw a grid in your editor with at least 10  10 cells. Draw a
set of axes to the bottom left of the paper called frame a . Draw the transformed
frame, called b in the right place with respect to frame a based on the above
result. Label the axes of both frames with x or y .
(iii)Homogeneous Transforms as Frames. Consider the coordinates of the unit vectors
and origin of the transformed frame when expressed with respect to the original
frame. Compare these coordinates to the columns of the homogeneous transform.
How are they related? Explain why this means homogeneous transforms are also
machines to convert coordinates of general points under the same relationship
between the two frames involved. HINT: how is a general point related to unit
vectors and origin of any frame.
2.3.5.2 Solution
(i) The unit vectors and origin expressed in frame ‘b’ are:

ˆa 0 –1 7 ˆa 0 –1 7 1 0
xb = 1 0 3 xa = 1 0 3 0 = 1
0 0 1 0 0 1 0 0

ˆa 0 –1 7 ˆ a 0 –1 7 0 –1
yb = 1 0 3 ya = 1 0 3 1 = 0
0 0 1 0 0 1 0 0

ˆa 0 –1 7 ˆ a 0 –1 7 0 7
ob = 1 0 3 oa = 1 0 3 0 = 3
0 0 1 0 0 1 1 1
MATH FUNDAMENTALS 6
(ii) Frames ‘a’ and ‘b’ can be visualized as follows:

ˆa
xb

ˆa
yb

ˆa
ya

ˆa
xa
(iii) The first column is the transformed x axis, second is the y, third is the origin. Any
general point is a linear combination of the basis vectors. Since these basis vectors are
transformed correctly, any point will be as well.

2.3.5.3 Pose of a Transform and Operating on a Point

(i) Solving for the Relative Pose. The parameters of the compound homogeneous
transform that relates the frames in question 2.3.5.2 can be found using the
techniques of inverse kinematics. Write an expression (in the form of a
homogeneous transformT with three degrees of freedom (or “parameters” in
operator form) a b  for the general relationship between two rigid bodies in
2D, equate it to the above transform.
(ii) Solve the above expression by inspection for the “parameters.”
2.3.5.3 Solution
a = 7
c – s a 0 –1 7
a) b)
s c b = 1 0 3 b = 3
0 0 1 0 0 1  = 2

2.3.5.4 Rigid Transforms

Operating on a general point is no different
T than operating on the origin.
(i) Operate on the point p = 3 2 1 with the transform T from Section 2.3.5.2 and
write the coordinates p ' of the new point. Copy your last figure and draw p and
p ' on it. Label each.
(ii) How do the coordinates of p ' in the new frame (called b ) compare to the
coordinates of p in the old frame (called a )?
(iii)What property of any two points is preserved when they are operated upon by an
orthogonal transform and what does this imply about any set of points?
MATH FUNDAMENTALS 7
2.3.5.4 Solution
0 –1 7 0 –1 7 3 5
a) p' = 1 0 3 p = 1 0 3 2 = 6
0 0 1 0 0 1 1 1

p p'

ˆa
xb

ˆa
yb

ˆa
ya

ˆa
xa

b) The coordinates of the transformed point in the transformed frame are identical to the
coordinates of the original point in the original frame. In other words:
a b
p = p'
c) Lengths are preserved. This shows that the transform moves all of space rigidly.

2.3.5.5 Homogeneous Transforms as Transforms

(i) Transforming a General Point. This exercise is worth extra attention. It illustrates
the basic duality of operators and transforms upon which much depends. Copy the
last figure including
T points p and p' on a fresh sheet. Draw the point
b
q = 4 1 1 . The notation superscript b means the point has been specified
with respect to frame b , so make sure to draw it in its correct position with respect
to frame b .
a
(ii) Write out the multiplication of this point by “the transform” and call the result q .
b a
Using a different symbol than the one drawn for q , draw q in its correct
position with respect to frame a .
(iii)Earlier, when p was moved to p' , p was expressed in frame a and so was p' .
b a
Here you expressed q in frame b to produce a result q expressed in frame a .
Now the following is the key point. Discuss how interpreting the input differently
(i.e in different coordinates) leads to a different interpretation of the function of
the matrix.
(iv)How can the function performed on the point p be reinterpreted as a different
function applied, instead, to p' ?
 MATH FUNDAMENTALS 8
ssm
2.3.5.5 Solution

0 –1 7 0 –1 7 4 6
b) a b
q = 1 0 3 q = 1 0 3 1 = 7
mtt
0 0 1 0 0 1 1 1

Both q’s are in same place q a) b)

p
p'
bb99
ˆa
xb

ˆa
yb
88@
ˆa
@
ya

ˆa
xa

iii) The key difference was that the input was specified wrt frame b, so the matrix is being
ggm
used as a transform. In all earlier cases, the input was expressed with respect to frame a
and the matrix was used as an operator.
iv) The previous operation on point p can be interpreted as a conversion of coordinates
for p' from frame b to frame a.
maa

2.4.5 Kinematics of Mechanisms

2.4.5.1 Three Link Planar Manipulator
Every roboticist should code manipulator kinematics at least once. Using your favorite
programming environment, code the forward and inverse kinematics for the three link
manipulator. Pick some random angles and draw a figure. Then compute the end effector
pose from the angles and verify that the inverse solution regenerates the angle from the
iill..cc

pose. What does the second solution look like? If you are ambitious, try a case near
singularity, and experiment with the mechanism Jacobian.

2.4.5.2 Pantograph Leg Mechanism

The pantograph is a four bar linkage that can be used to multiply motion. This one was
used on the Dante II robot that ascended and entered an active volcano on Mount Spur in
Alaska in 1994. Triangles ABD, ACF, and DEF are all similar. When the actuator pushes
oom
m

complete document is available on https://unihelp.xyz/ *** contact me if site not loaded