ECE-171B: Linear Control System Theory

Homework - 2
Fall 2021

⼝
* Deadline is Wednesday Oct 27th, 11:59PM
** Please mark how much time you spent on this homework
at the first page of your submission.
1. Plot phase portraits of the systems covered in the Lecture-8 Exercise A and B.
See whether it matches the stability conclusion in lecture.
(a) System Model of Exercise A:
ẋ1 = x1 x2 2 , ẋ2 = 2x1 x2 2 x1 x2
(b) System Model of Exercise B:
ẋ1 = x2 , ẋ2 = 1 x11 2 x2
2. Consider the following dynamic system:

x2
ẋ1 = x1 [(K x1 ) + ]
x2 + 1
x2 x1 x2
ẋ2 = +
2 x2 + 1
(a) Determine the equilibrium point(s) of the above system in terms of K.
(b) Find the value of K = K ⇤ for which the system has only one equilibrium
point besides (0,0).
(c) Linearize the non-linear system around the equilibrium points obtained by the
K = K ⇤.
(d) Determine whether the equilibrium points are stable for the linearized system
obtained in part-(c). (Note: if any of the eigenvalues of the matrix A is zero,
you have the right to say “don’t know whether the system is stable or not”)
3. (PI cruise-control) Consider the following cruise-control example with dynamics
mv̇ = av + u + !
where u is the control input (force applied by engine) and ! the disturbance input
(force applied by hill, etc.), which will be ignored below, i.e., != 0. An open-loop
control strategy to achieve a given reference speed vref would be to choose
u = âvref
where â is your estimate of a, which may not be accurate. Assume m, a and â are
all positive.
I
plot

b
Plot
 悲道機桁

t
以 义 冰 k x.it 淼 0 ⼼⼼0 dofor⺮怑 ⼼
加 ⼆式 0 i
缒㖳 EQELE.tn
Uk卌 芯 0
⼗
K htk defInAk
Put Mlk 刈⼗点
f
h
i ki
⾮ ⼊
E
到
K X it
if 0
ZXIKXFif2iiii
zxitzxi.li0
⼆⼋
⼗⼆ ⼋ ⼀
避政 0

2XR lzktzsx.tn 2K1

⼊1 o ⼊2
ti
⼀
⼆
HE unstable
总重签器 forKh.tt
i there are only one equilibriumpointbesides too EQEE.tn
壜圃湖
G forKKY HE 㒊三点⼝ 回 到
1 器 堬 ⼝ ⼼
1焱 iii
1
⼆ ⽐⼆
蕊 藏 N 避政 0
def E NEQLE.tt ⼊no xzi
笸炎惠 囊 蕊蕊 unstable

匲圃湖
iii
i

for1 姡 1 E 德i

蕊 两

ii
嚻匳 颧剗
 (a) Compute the equilibrium points (which is also called as steady-state) for both
the open-loop strategy above, and for the feedback law
u= kp (v vref )
and compare the steady-states as a function of = a/â when kp = 10â. (You
should solve the problem analytically, and then plot the response vss /vref as a
function of for both the open-loop and proportional-gain feedback law. Here
vss stands for the steady states.)
(b) Now consider a proportional-integral (PI) control law
Z t
u = kp (v vref ) ki (v vref ) dt
0
Rt
Let us introduce a new variable q = 0 (v vref ) dt, then q̇ = v vref and
u = kp (v vref ) ki q. We can write the closed-loop dynamics in the form of
a kp ki
v̇ = v (v vref ) q,
m m m
q̇ = v vref
Compute the steady state solution for the proportional-integral (PI) control law.
Then compare the response with the proportional feedback law in (a).
4. (Predator-Prey Model) Let’s consider a predator-prey ecological system. In an iso-
lated island, there are only two kinds of animals: tigers (predator) and hares (prey),
and tigers have no choice but to hunt hares for food. Let H(t) represent the num-
ber of hares and G(t) represent the number of tigers. For simplification, H(t) and
G(t) are treated as continuous variables instead of integer variables. The population
dynamics of this predator-prey system are modeled as
Ḣ(t) = rH(t) aG(t), H(t) 0 , (1a)
bH(t)G(t)
Ġ(t) = eG(t), G(t) 0 (1b)
c + H(t)
Where, r is the growth coefficient of the hares, a represents the diminishing coefficient
by the hunt of tigers, b represents the growth coefficient of the tigers, c is a constant
parameter that controls the prey consumption rate for low hare population, and e
represents the mortality rate of the tigers. In our following problem, r, a, c, b, e are
all positive constant parameters with b > e.
 ⇤
H
(a) Find the equilibrium points for the above predator-prey dynamic system
G⇤
above1 . The number of equilibrium points may be more than one. Please write
down the answer using the parameters.
1
You might notice that this model is di↵erent from the model in AM Chapter 3. Again, any model
is only an approximation. In this homework assignment, we stick with abstract model in equation (1)

Page 2
3 ldmir
avtut TEkpcv
we V
ivti ltiN.it
V ref
equilibrium
盐商
如
以堦州 0 ⼆

V55
丽
he kpiloa hd
ks
lre Nssi
piaG j.li
器 前
b V 出V iN wefl
V
vre ifit.si
ti q
eady_state
iv
崔品 kwef
ifv qeE.liref
by comparing w la steady state
goesnothing to do with
Ipl
4 as equilibrium point 邓蕊咒
rHHI
aGHF btp.ec 0
⼆
点
㶙品 儸 ⼆
器
i.IE
br o.ie ⼼⼼
喈晶
邀 问 吃些 ⼀
品
 for 笓 8
linearized model
HH o 1114 0.5G4
Get 0.1G4 义
洲
XH
燚愁灯
A
沿 ⼷⻔
for
HH
瀏 吢了
0.5GH
o 1H4
G4 o ol HH
灬 出
只 器了 义 洲
A
品 怒了
C for A 治 志⻔
PEAK0 ⼊z
n 0.1
0.1
i unstable

for A 0
品⼚ 啊
HIM 0 ⼊ o 05 ⼟005

i Unstable
4
d UH WHY 7
foil 义 我
X AHUH1
Āiiǐ 啊
⼊ A 0 ⼋otwxtoooF
V co.lt⼼⼊ to 005
stable
let thesystembe
⼼ 170 Was I
(b) 
Let r = 0.1, e = 0.1, c = 100, b = 0.2, a = 0.5. Calculate the equilibrium points
H⇤
with these given constant values. Linearize the system model around the
G⇤
equilibrium points
 and write down the linearized model in the form of ẋ = Ax
H(t) H ⇤
where x(t) = . (You may obtain multiple di↵erent equilibrium
G(t) G⇤
points, please derive the linear system models for each equilibrium point)
(c) Study the stability of the equilibrium points calculated in question (b) by check-
ing the eigenvalues of the matrix A you obtain in question (b). (Please write
down the steps how you get your conclusion so we can give you credits even if
the final answer is not right.)
(d) Suppose that you are a biologist entering this island and want to improve the
ecological stability of the predator-prey system. Your e↵orts can be regarded
as a control
 input u(t) to the system. Recall that the state variable x(t) is
H(t) H ⇤
x(t) = in the linear system ẋ = Ax you obtained in question (b).
G(t) G⇤
Now consider the following system with a feedback controller,

w(H(t) H ⇤)
ẋ = Ax + u(t) , u(t) =
0
where ẋ = Ax is the linearized state-space model that you derive in question
(b) based on the nonzero equilibrium point, w is the coefficient representing the
additional growth rate of hares by your e↵orts. Write down the new system
model in an autonomous form of ẋ = Āx by plugging u(t) into ẋ = Ax + u(t)

Page 3