Problem set for Control Engineering

Indian Institute of Technology Madras

 Revision
1. Differentiate
(i) x1/3

(ii) 1 + x2
(iii) sin(1/x)
(iv) 4x
(v) 2x /(2x 3x )
(vi) xx

2. Given x1 ,...,xn , find R+ to maximize

n
Y
exi .
i=1

Sketch this curve.

3. Given x1 ,...,xn , find R and R+ to maximize

n
Y 1 2 2
e(xi ) /2 .
i=1 2 2

Sketch this surface.

4. Define Taylor series for a function that depends on a single variable. Using the definition, calculate
the Taylor series expansion at x = 0 for the following function

f (x) = ex .

5. Define Taylor series in the case of two variables, and calculate the Taylor series expansion around
the point (a, b) = (0, 0) of the function

f (x, y) = ex log(1 + y).

6. Find the eigenvalues and eigenvectors of A, A2 , A1 and A + 4I.

 
2 1
A=
1 2

7. Find the eigenvalues and eigenvectors of the following matrices:

(a)  
cos sin
A=
sin cos

(b)
2 0 2
B= 0 0 2
2 2 1

(c)
1 1 0 0
0 2 1 0
C=
0

0 3 1
0 0 0 4

2
8. Find the eigenvalues for the matrix  
2 1
A= .
0 2
Show that for matrix A, the algebraic multiplicity and geometric multiplicity of its eigenvalue is
not the same.
9. Find the eigenvalues of matrices A, B and A + B.
   
3 0 1 1
A= B=
1 1 0 3

Are the eigenvalues of A + B equal to sum of eigenvalues of A and eigenvalues of B?

10. Find the eigenvalues of matrices A, B, AB and BA.

   
1 0 1 2
A= B=
1 1 0 1

(a) Are the eigenvalues of AB equal to eigenvalues of A times eigenvalues of B?

(b) Are the eigenvalues of AB equal to the eigenvalues of BA?

11. The eigenvalues of A equal the eigenvalues of AT . This is because det(A I) equals det(AT I).
That is true because . Show by an example that the eigenvectors of A and AT are not the
same.

12. Let X1 and X2 be the eigenvectors corresponding to distinct eigenvalues of the symmetric matrix
A (A = AT ). Show that X1 is orthogonal to X2 (X1T X2 = X2T X1 = 0).

13. Show that inverse A1 exists if and only if none of the eigenvalues 1 , 2 . . . n is zero, and then
A1 has the eigenvalues 1/1 , 1/2 . . . 1/n .

14. Let 1 , 2 . . . n be the eigenvalues of matrix A. Show that the eigenvalues of eA are given by e1 ,
e 2 . . . e n .

15. Find the eigenvalues and eigenvectors of matrix

0 1 0
A= 0 0 1 .
6 11 6

Find the eigenvectors X1 , X2 and X3 corresponding to the distinct eigenvalues 1 , 2 and 3 of

matrix A. Create the 3 3 matrix P = [X1 X2 X3 ], and show that

1 0 0
AP = P 0 2 0 .
0 0 3

16. Diagonalize the following matrices and find the eigenvector of A = P 1 AP .

7.3 0.2 3.7 5 10 10
(1) 11.5 1.0 5.5 (3) 10 5 20
17.7 1.8 9.3 5 5 10

7 0 3  
2 1 8 1
(2) 1 (4)
5 2
2 0 2

3
17. Given the following state space model

x1 0 1 0 x1 0
x2 = 0 0 1 x2 + 0 u
x3 6 11 6 x3 6

and
  x1
y= 1 0 0 x2 .
x3
Transform the above state model using x = P z.

18. Find the matrix P such that following state space equation is transformed into diagonal canonical
form:
x1 a1 1 0 x1 0
x2 = a2 0 1 x2 + 0 u.
x3 a3 0 0 x2 1

19. Obtain a state transition matrix (t) of the following system,

      
x1 0 1 x1 0
= + u.
x2 2 3 x2 1

If u(t) is a step input function, find the state response of the system.

20. Find the transfer function of the following state space model,
        
x1 1 1 x1 0   x1
= + u and y= 1 1 .
x2 0 1 x2 1 x2

21. For the following set of differential equations, write down the canonical state space model:

(i)
d3 y(t) d2 y(t) dy
2 + 3 +5 + 2y(t) = u(t),
dt3 dt2 dt
(ii)
t
d3 y(t) d2 y(t) dy
Z
3
+5 +3 + y( )d = u(t).
dt dt2 dt 0

22. Find f(A)=A10 , where  

0 1
A= .
2 3

Chapter 1
23. Astrom and Murray, Exercise 1.2

24. Astrom and Murray, Exercise 1.3

25. Astrom and Murray, Exercise 1.5

4
 Chapter 2
26. Astrom and Murray, Exercise 2.1

27. Astrom and Murray, Exercise 2.3(a)(b)

Chapter 3
11. Read the following models outlined in Chapter 3 of Astrom and Murray.

(a) Cruise Control

(b) Bicycle Dynamics
(c) Operational Amplifier Circuits
(d) Computing Systems and Networks
(e) Atomic Force Microscopy
(f) Drug Administration and
(g) Population Dynamics.

Analyze, and comment, on the above models through the lens of the following guidelines.
As part of your discussion outline the mathematical model and clearly identify all the parameters.
What is modeling?

Going back and forth between

Nature/RealWorld Model Analysis Design

What is a model?

Approximate/Exact mathematical representation of a system

Different purposes of a model?

Identify parameters that really matter

Identify regimes that really matter
Broad outlines to make predictions
Concentrate on the most distinctive features
Design, to help build systems
Explain causes of the behavior that we observe
Persuade others of your point of view

Model Development

Be clear on the purpose of the model

Two options for model development: (i) First Principles (ii) Experiments

Model Derivation

Offer a clear derivation of the model

What is the key idea behind the derivation
If possible offer an alternate derivation for the model
Make use of diagrams: block diagrams / schematic diagrams

5
Model Assumptions

Identify and state the assumptions, and suggest modifications to the assumptions
Always understand (and test) the simplifying assumptions
Always understand the limitations of the model
Have simple assumptions, then refine model by removing the simple assumptions
Make simplifications to gain insight; then develop realistic models

Model Analysis

First derive a tractable model, even on simplifying/ideal assumptions

Results with simple assumptions may not be completely accurate, but they serve as a first
approximation to which refinements can be made.
Have a simple model, then enhance the mathematical understanding of the model
Use numerical examples to illustrate ideas/results
Always try and benchmark with data/simulations/experiments

Model Parameters

Minimize the number of parameters in the model

Do not leave any arbitrary parameters in the model: can all the parameters either have an
analytical basis, or be estimated empirically?
Offer design guidelines for all the parameters

What are good models?

Able to explicitly model the relationships between all the key variables.
Can be adjusted to deal with some of the short comings/simple assumptions.

Good modeling practice

Always aim to try and identify / uncover relationships, and then plot them
Draw plots to show relationships between quantities of interest
Some early stage modeling questions

What is the problem we would like to solve?

What is the simplest problem we can solve?
Is the problem formulation elegant? simple, yet general enough?

Some late stage modeling questions: So what?

What relationships are uncovered by the analysis?

Did we offer design guidelines for the system?
Can any principles be generalized?
How is the subject area better informed by your work?

Finally

Use multiple models, perhaps under different assumptions, to

Understand the problem from a multiplicity of view points.
Use a combination of:
Analysis, Computations, Simulations (simple, detailed), test bed experiments and mea-
surements to build a comprehensive view.

6
28. Consider a damped spring-mass system with dynamics

mq + cq + kq = F.
p
Let 0 = k/m be the natural frequency and = c/(2 km) be the damping ratio.
(a) Show that by rescaling the equations, we can write the dynamics in the form

q + 20 q + 02 q = 02 u,

where u = F/k. This form of the dynamics is that of a linear oscillator with natural frequency
0 and damping ratio .
(b) Show that the system can be further normalized and written in the form
dz1 dz2
= z2 , = z1 2z2 + v.
d d
The essential dynamics of the system are governed by a single damping parameter . The
Q-value defined as Q = 1/2 is sometimes used instead of .
(c) Show that the solution for the unforced system (v = 0) with no damping ( = 0) is given by

z1 ( ) = z1 (0) cos + z2 (0) sin , z2 ( ) = z1 (0) sin + z2 (0) cos .

Invert the scaling relations to find the form of the solution q(t) in terms of q(0), q(0), and 0 .
(d) Consider the case where = 0 and u(t) = sin t, > 0 . Solve for z1 ( ), the normalized
output of the oscillator, with initial conditions z1 (0) = z2 (0) = 0 and use this result to find
the solution for q(t).
29. The motion of an ideal pendulum is described by

+ g sin = 0,

where is the angle between the pendulums position and vertical, and g is the gravitational
acceleration.
(a) Using the small angle approximation sin , solve for an expression for (t), written in
terms of the initial conditions (0) = 0 , (0) = 0 and the parameter g.
(b) Plot the pendulums motion in three different environments: Earth (g = 9.8 m/s2 ), the moon
(g = 1.6 m/s2 ), and on Temple I the comet that the Deep Impact mission collided with
on in July 2005 (g = 0.00004 m/s2 ). Assume that the pendulum is given a small initial
starting angle (0) = 0.05 radians (about 3 degrees) and then released with no initial velocity
((0) = 0). Note that this is an idealized equation of motion and damping is not included, so
there are no frictional forces to slow the pendulum down over time.
(c) If the pendulum is pushed with a force u(t), the equation of motion becomes

+ g sin = u(t).

Apply the small angle approximation and assume (0) = 0 and (0) = 0. Solve for (t) when
u(t) = sin t.

Chapter 4
30. Consider the following dynamical system
dw(t) 1 w(t)x(t)p(t)
= ,
dt 2
where x(t) = w(t)/ , p(t) = f (x(t)), and > 0 . Assume that the above dynamical system has a
unique equilibrium point.

7
 (i) Linearize the system around the equilibrium point.
(ii) Derive the characteristic equation.
(iii) Discuss the stability properties of the dynamical system.
(iv) Repeat (i), (ii) and (iii) for the following cases.
 B
(a) f (x(t)) = x(t)
C
x(t)C
(b) f (x(t)) = x(t) .

(v) Repeat (i), (ii), (iii) and (iv) for the following dynamical systems:

(a)
dw(t)   w(t)
= i(w(t)) d(w(t))p(t)
dt

(b)
dw(t) 1 w(t)x(t )p(t )
=
dt 2

(c)
dw(t)   w(t )
= i(w(t)) d(w(t))p(t )
dt
where x(t) = w(t)/ , p(t) = f (x(t)), > 0 and where i(w) and d(w) are arbitrary functions of w.

31. For each of the following systems, locate the equilibrium points for the system and indicate whether
each is asymptotically stable, stable but not asymptotically stable) or unstable. To determine
stability, you can either use a phase portrait (if appropriate), analyze the linearization or simulate
the system using multiple nearby initial conditions to determine how the state evolves.

(a) Nonlinear spring mass. Consider a nonlinear spring mass system with dynamics

mq = k(q aq 3 ) cq,

where m = 1000 kg is the mass, k = 250 kg/s2 is the nominal spring constant, a = 0.01
represents the nonlinear softening coefficient of the spring and c = 100 kg/s is the damping
coefficient. Note that this is very much similar to the spring mass system we have studied in
section 2.2, except for the nonlinearity.
(b) Genetic toggle switch. Consider the dynamics of two repressors connected together in a cycle.
It can be shown (Exercise 2.9) that the normalized dynamics of the system can be written as

dz1 dz2
= z1 v 1 , = z2 v 2 .
d 1 + z2n d 1 + z1n

where z1 and z2 represent scaled versions of the protein concentrations, v1 and v2 represent
external inputs and the time scale has been changed. Let = 2.16, n = 2 and v1 = v2 = 0.

32. Write a mathematical essay on Lyapunov functions and do Astrom and Murray, Excercise 4.4.

33. Write a mathematical essay on the KrasovskiLasalle Invariance Principle and do Astrom and Mur-
ray, Excercise 4.5.

34. Astrom and Murray, Exercise 4.10. Consider the linear system

dx 0 1 1

= x + u, y= 1 0 x,


dt 0 3 4

8
 with the feedback u = ky. Plot the location of the eigenvalues as a function the parameter k.
Identify the approximate gains at which the system becomes unstable and label these on your plot.
(To generate your plot, you should compute the eigenvalues at multiple values of k and plot these
on the complex plane. Label the locations of the eigenvalues for k = 0 with and the locations
for k
with an if they converge to a finite value. Choose the units on your graph so that
key features are visible and use arrows on your plot to indicate which direction corresponds to
increasing gain, similar to Figure 4.18b in the text.)

Chapter 5
35. Astrom and Murray, Exercise 5.1

36. For each of the following linear systems, determine whether the origin is asymptotically stable and,
if so, plot the step response and frequency response for the system. If there are multiple inputs or
outputs, plot the response for each pair of inputs and outputs.

(a) Coupled mass spring system. Consider the coupled mass spring system from Example 5.5 with
m = 250, k = 50 and c = 10. The input u(t) is the force applied to the right-most spring.
(b) Bridged Tee Circuit. Consider the following electrical circuit, with input vi and output y = v0 .

C2

1 2 3

R1 R2
Vi Vo
C1

The dynamics are given by

   
1 1 1
C11R2 1 1 1

d v c1 =
C R1 + R2
v c1

C1 R1 + R2

+ vi ,

1





dt vc2 1 C21R2 vc2 1



C2 R 2 C2 R 2

vc1

y= 0 1 + vi ,


vc2

where vc1 and vc2 are the voltages across the two capacitors. Assume that R1 = 100 ,
R2 = 100 and C1 = C2 = 1 106 F.

37. Astrom and Murray, Exercise 5.5

38. Astrom and Murray, Exercise 5.6

39. Astrom and Murray, Exercise 5.7

40. Astrom and Murray, Exercise 5.8(a)b)

9
41. Astrom and Murray, Exercise 5.8, parts (a) through (c). For part (a), note that the equation can
also be written as.
k1
X
x[k] = CAk x0 + CAi Bu[k 1 i] + Du[k]
i=0

For part (b), you can assume that the matrix A has a full basis of eigenvectors. For part (c), use
inputs u[k] = sin k.

42. Astrom and Murray, Exercise 5.10

Chapter 6
43. Write down the definition of reachability.

44. Draw the block diagram of a system in reachable canonical form.

45. (Exercise 6.7)(Reachability matrix for reachable canonical form) Consider a system in reach- able
canonical form. Show that the inverse of the reachability matrix is given by

1 a1 a2 . . . an
0 1 a1 . . . an1
..

..
Wr1 = 0 0 1 . . .
. .

.. . .. ..

0 0 0 ... 1

46. (Exercise 6.8)(Non-maintainable equilibria)Consider the normalized model of a pendulum on a cart

d2 x d2
= u, = + u.
dt2 dt2
where x is cart position and is pendulum angle. Can the angle = 0 for 0 6= 0 be maintained?

47. Astrom and Murray, Exercise 6.10

Chapter 7
48. Write down the definition of observability.

49. Draw the block diagram of a system in observable canonical form.

50. Show that the inverse of the observability matrix has a form given by

1 0 0 ... 0
a1 1 0 ... 0

1 a2 a 1 ... 0
Wo = 1 .
.. .. .. .. ..
. . . . .
an1 an2 an3 . . . 1

10
51. (Example 7.1)Consider the two-compartment model in which drug with concentration c0 is injected
in compartment 1 at a volume flow rate of u and that the concentration in compartment 2 is the
output. Let c1 and c2 be the concentrations of the drug in the compartments, V1 and V2 be the
volumes of the compartments and q be the outflow rate. Introducing the variables k0 = q0 /V1 , k1 =
q/V1 , k2 = q/V2 and b0 = c0 /V1 , the system can described by the linear system and its model is,
     
dC k0 k1 k1 b0
= C+ u,
dt k2 k2 0

and,  
y= 1 0 C.

52. (Example 7.2)Consider the compartment model characterized by the matrices

   
k0 k1 k1 b0  
A= , B= , C= 1 0 .
k2 k2 0
(a) Find the condition for observability of the system.
(b) Let the desired characteristic polynomial for the observer be s2 + p1 s + p2 = 0. Obtain the
observer gain (L) so that the desired characteristic polynomial is achieved.

53. (Exercise 7.1)Consider a system under a coordinate transformation z = T x, where T Rnn is

an invertible matrix. Show that the observability matrix for the transformed system is given by
W0 = W0 T 1 and hence observability is independent of the choice of coordinates.
54. (Exercise 7.3) Show that if a system is observable, then there exists a change of coordinates z = T x
that puts the transformed system into observable canonical form.
55. (Exercise 7.4)(Bicycle dynamics) The linearized model for a bicycle is given in equation (3.5), which
has the form

d2 Dv0 d mv02 h
J 2
= mgh +
dt b dt b
where is the tilt of the bicycle and is the steering angle. Give conditions under which the
system is observable and explain any special situations where it loses observability.

Chapter 8
56. Consider the following linear dynamical system:
dx
= Ax + Bu, y = Cx + Du.
dt

Let z = T x be the new set of coordinates, where T is invertible.

(a) Represent the dynamical system in terms of the new coordinates z.
(b) Show that the transfer function remains invariant to changes of the coordinates in the state
space.
57. Explain, with the help of an example the definition and idea of Bode plot.
58. Draw Bode plots for the following dynamical systems:

a 02
(a) G(s) = (b) G(s) =
s+a s2 + 20 s + 02
20 s s2
(c) G(s) = (d) G(s) = 2 .
s2 + 20 s + 02 s + 20 s + 02

11
 d n

r e u v y
F (s) C(s) P (s) r

59. For the closed-loop feedback system in the figure, show that following are the transfer functions:
F 1 P
(a) Ger = (b) Gen = (c) Ged =
1 + PC 1 + PC 1 + PC
P CF P 1
(d) Gyr = (e) Gyd = (f) Gyn =
1 + PC 1 + PC 1 + PC

60. Show that the Bode plot for the transfer function G(s) = (s + a)/a can be approximated by
(
0 <a
log |G(i|
log log a > a,

0
<a
G(i) 45 + 45(log log a) a/10 < < 10a

90 > 10a.

61. Let G(s) be the transfer function for a linear system. Show that if we apply an input
u(t) =
A sin(t), then the steady-state output is given by y(t) = |G(i)|A sin t + arg G(i) .
62. Consider the system
dx
= ax + u.
dt
Compute the exponential response of the system and use this to derive the transfer function from
u to x. Show that when s = a, a pole of the transfer function, the response to the exponential
input u(t) = est is x(t) = eat x(0) + teat .
63. Neglecting the damping and linearizing the inverted pendulum around the upright position gives
a system characterized by the matrices
   
0 1 0  
A= , B= , C= 1 0 , D = 0.
mgl/Jt 0 1/Jt

Determine the transfer function of the system.

64. Consider the following differential equation
dn y dn1 y dn1 u dn2 u
+ a 1 + . . . + a n y = b 1 + b 2 + . . . + bn u.
dtn dtn1 dtn1 dtn2
(a) Let be a root of the characteristic polynomial

sn + a1 sn1 + . . . + an = 0.

Show that if u(t) = 0, the differential equation has the solution y(t) = et .
(b) Let be a zero of the polynomial

b(s) = b1 sn1 + b2 sn2 + . . . + bn .

Show that if the input is u(t) = et , then there is a solution to the differential equation that
is identically zero.

12
 Chapter 9
65. Write down the definition of loop transfer function.
66. Draw the Nyquist plot for the following transfer function
1
L(s) = .
(s + a)3

67. Consider the following transfer function

k
L(s) = .
s(s + 1)2

(a) Draw the Nyquist plot of the given transfer function.

(b) Comment on the stability of the system as the value of parameter k is varied.
68. Write the statement of the general Nyquist stability criterion.
69. The loop transfer function of an inverted pendulum with a proportional-derivative controller having
transfer function C(s) = k(s + 2) is given by

k(s + 1)
L(s) = .
s2 1
Using the Nyquist stability criterion, find the condition for stability of the closed-loop system.

Chapter 10
70. Show that a controller with integral action gives zero error if the closed-loop system reaches steady
state.
71. Consider the following closed-loop dynamical systems represented by the block diagrams. Assume
that the process has the transfer function P (s) = b/(s + a) and show that the transfer functions
from r to y are

r
ki /s ki /s

r e u y u y
r

kp P (s) r r

kp P (s) r

kd s r

kd s

1 1

(a) PID using error feedback (b) PID using two degrees of freedom

bkd s2 + bkp s + bki

(a) Gyr (s) = ,
(1 + bkd )s2 + (a + bkp )s + bki
bki
(b) Gyr (s) = .
(1 + bkd )s2 + (a + bkp )s + bki

72. Consider a second-order process with transfer function

b
P (s) = .
s 2 + a1 s + a2

13
 The closed-loop system with a PI controller is a third-order system. Show that it is possible to
position the closed-loop poles as long as the sum of the poles is a1 . Give equations for the
parameters that give the closed-loop characteristic polynomial

(s + 0 )(s2 + s0 0 s + 02 ).

73. Consider a system with the transfer function P (s) = (s + 1)2 . Find an integral controller that
gives a closed-loop pole at s = a and determine the value of a that maximizes the integral gain.

74. Consider a system with the transfer function P (s) = es /s. Determine the parameters of P, PI
and PID controllers using Ziegler-Nichols step and frequency response methods. Compare the
parameter values obtained by the different rules and discuss the results.

Quiz I
All questions from Quiz I.

Quiz II
All questions from Quiz II.

14