EE580, Fall 2023

Homework one
Due on September 15, 2023, 4:00PM. Upload your solution to Canvas.
Please write down name, student ID and course code on the first page of the solution.

Problem 1: (10 points) Consider a system whose input and output are related by
(
u(t)2
u(t−1) , u(t − 1) 6= 0
y(t) =
0 u(t − 1) = 0
for all t. Show that the system satisfies the homogeneity property but not the additivity property.
(αu(t))2 u(t)2
Solution: αu(t−1) = α u(t−1) = αy(t) implies the homogeneity property. In addition,
u1 (t)2 u2 (t)2 (u1 (t) + u2 (t))2
+ 6= .
u1 (t − 1) u2 (t − 1) u1 (t − 1) + u2 (t − 1)
Thus the additivity property does not hold.

Problem 2: (15 points) Show that if the additivity property holds, then the homogeneity property holds for all
rational number α. More specifically, let H(u, x0 ) be the system output when initial state is x0 and input is u. By
using the additivity property, show H(αu, αx0 ) = αH(u, x0 ) for any rational number α.
Hint: Given any rational number α, there are two integers m and n such that α = m n . We also know H(0, 0) = 0.
Solution: Since H(u, x0 ) = H(u + 0, x0 + 0) = H(u, x0 ) + H(0, 0), then H(0, 0) = 0.
Let n be a positive integer. By the additivity property, we have H(nu, nx0 ) = H(u, x0 ) + · · · + H(u, x0 ) =
nH(u, x0 ).
Since 0 = H(nu+(−nu), nx0 +(−nx0 )) = H(nu, nx0 )+H(−nu, −nx0 ), then H(−nu, −nx0 ) = −H(nu, nx0 ) =
−nH(u, x0 ).
Thus, the additivity property implies H(nu, nx0 ) = nH(u, x0 ) for any integer n.
1 1 1 1 1 1 1
Let m be an integer. Since H(u, x0 ) = H(m( m u), m( m x0 )) = mH( m u, m x0 ), then H( m u, m x0 ) = m H(u, x0 ).
n n n
Pick any real number α and let α = m for some integers m and n. Then we have H(αu, αx0 ) = H( m u, m x0 ) =
1 1 n
nH( m u, m x0 ) = m H(u, x0 ) = αH(u, x0 ).

Problem 3: (20 points) Find a state-space equation to describe the following circuit. Find also its transfer function
by using impedances.

Solution: Applying KCL at node A yields C2 ẋ2 = x3 and at node B yields u−x R = C1 ẋ1 + C2 ẋ2 = C1 ẋ1 + x3 .
1

Applying KVL to the right-hand-side loop yields Lẋ3 = x1 − x2 . The output is given by y = Lẋ3 = x1 − x2 . So
a state-space equation is given by:
1
0 − C11
   1 
− RC 1 RC1
1
ẋ =  0 0 C2
x +  0 u
1
L − L1 0 0
y = [1 − 1 0]x + 0 · u.
 1
We use impedances to compute the transfer function. The impedance of the parallel connection of C1 s and
Ls + C12 s is
1
(Ls + C12 s ) LC2 s2 + 1
Z = C11 s 1 = .
C1 s + Ls + C2 s
(C1 + C2 )s + LC1 C2 s3
Z Ls LC2 s2 Z
Thus the voltage at B is uB (s) = R+Z û(s) and ŷ(s) = Ls+ C1 s
uB (s) = LC2 s2 +1 R+Z û(s). Thus the transfer
2
function ĝ(s) is
ŷ(s) LC2 s2
ĝ(s) = = .
û(s) RLC1 C2 s3 + LC2 s2 + R(C1 + C2 )s + 1

Problem 4: (15 points) A synchronous generator connected to an infinite bus can be represented by
M δ̈ = P − Dδ̇ − η1 Eq sin δ
τ Ėq = −η2 Eq + η3 cos δ + EF D ,
where δ is an angle in radius, Eq is voltage, P is mechanical input power, EF D is field voltage (input), D is
damping coefficient, M is inertial coefficient, τ is time constant, and η1 , η2 and η3 are constant parameters.
(P1) Using δ , δ̇ and Eq as state variables, find the state equation.
(P2) Let P = 0.815, EF D = 1.22, η1 = 2.0, η2 = 2.7, η3 = 1.7, τ = 6.6, M = 0.0147 and D/M = 4. Find all
equilibrium points.
(P3) Suppose that τ is relatively large so that Ėq ≈ 0. Show that assuming Eq to be constant reduces the model to
a pendulum equation.
Solution: (P1) Let x1 = δ , x2 = δ̇ and x3 = Eq and u = EF D , and we have
ẋ1 = x2
P D η1
ẋ2 = − x2 − x3 sin x1
M M M
η2 η3 1
ẋ3 = − x3 + cos x1 + u.
τ τ τ
(P2) The equilibrium points are the roots of the equations:
0 = x2
0 = 0.815 − Dx2 − 2x3 sin x1
0 = −2.7x3 + 1.7 cos x1 + 1.22.
0.4075
Notice that x2 = 0 and x3 = . Substituting x3 in the third equation yields (1.22 + 1.7 cos x1 ) sin x1 −
sin x1
1.10025 = 0. The foregoing equation has two roots x1 = 0.4067 and x1 = 1.6398 in the interval −π ≤ x1 ≤ π .
Due to the periodicity, 0.4067 + 2nπ are also roots for n = ±1, ±2, · · · . Each root x1 = x gives an equilibrium
point (x, 0, 0.4075/ sin x).
(P3) With constant Eq , the model reduces to
ẋ1 = x2
P D η1
ẋ2 = − x2 − Eq sin x1
M M M
which is a pendulum equation with an input torque.

Problem 5: (15 points) The pendulum system in Figure 1 is useful to model one link robotic manipulator. Apply
Newton’s law in the tangent direction and we have u cos θ − mg sin θ = mlθ̈.
(P1) Find a state-space equation of the system.
(P2) Find linearization at θ = 0.
Fig. 1. Pendulum system of Problem 5

Solution: Define x1 = θ and x2 = θ̇. Then we have

g u
ẋ1 = x2 = f1 (x1 , x2 ), ẋ2 = − sin x1 + cos x1 = f2 (x1 , x2 , u).
l ml
This is a nonlinear system.
Note that f1 is linear and
∂f2 g u g
|x1 =0 = (− cos x1 − sin x1 )|x1 =0 = − ,
∂x1 l ml l
∂f2 1 1
|x =0 = cos x1 |x1 =0 = .
∂u 1 ml ml
So the linearization is
   
0 1 0
ẋ = x+ u.
− gl 0 1
ml
Problem 6: (25 points) In general, linearizations around trajectories lead to linear time-varying systems because
the partial derivatives need to be computed along the trajectories. However, for some nonlinear systems, there are
trajectories for which linearizations lead to LTI systems. For models of vehicles; e.g., cars, airplanes, helicopters
and submarines, trajectories lead to LTI linearizations are called trimming trajectories. They often correspond to
motion along lines, circumferences and helices.
Consider a single-wheel cart (unicycle) Figure 2 below moving on the plane with linear velocity v and angular
velocity w can be modeled by the nonlinear system:
ṗx = v cos θ, ṗy = v sin θ, θ̇ = w,
where (px , py ) denote the Cartesian coordinates of the cart and θ denote its orientation. The system inputs are v
and w.
(P1) Construct a nonlinear state-space model for this system with state
   
x1 px cos θ + (py − 1) sin θ
x =  x2  =  −px sin θ + (py − 1) cos θ 
x3 θ
input u = [v w]T and output y = [x1 x2 ]T .
(P2) Compute a linearization for the system in (P1) around the equilibrium x∗ = 0 and u∗ = 0.
(P3) Show that w(t) = v(t) = 1, px (t) = sin t, py (t) = 1 − cos t, θ(t) = t is a solution to the system in (P1).
(P4) Show that the linearization of the system in (P1) around the trajectory in (P3) leads to an LTI system.
Solution: (P1) ẋ1 = ṗx cos θ−px sin θθ̇+ṗy sin θ+py cos θθ̇−cos θθ̇ = v+w(−px sin θ+py cos θ−cos θ) = v+wx2 .
ẋ2 = −ṗx sin θ − px cos θθ̇ + ṗy cos θ − py sin θθ̇ + sin θθ̇ = −w(px cos θ + py sin θ − sin θ) = −wx1 . The state-space
equation is given by  
v + wx2
ẋ =  −wx1 
w
Fig. 2. Unicycle of Problem 6

(P2) The linearization around the equilibrium is given by

   
0 0 0 1 0  
˙ 1 0 0
δx = 0
 0 0  δx +  0 0  δu, δy = δx.
0 1 0
0 0 0 0 1
(P3) In the x-coordinate, the candidate solution
     
px cos θ + (py − 1) sin θ sin t cos t − cos t sin t 0
x(t) =  −px sin θ + (py − 1) cos θ  =  − sin t sin t − cos t cos t  =  −1  .
θ t t
Therefore,    
0 v + wx2
ẋ(t) =  0  =  −wx1 
1 w
(P4) The linearization along the trajectory is given by
   
0 1 0 1 −1  
˙ =  −1 0 0  δx +  0 0  δu, 1 0 0
δx δy = δx.
0 1 0
0 0 0 0 1