Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

Signals and Systems: Final Exam

School of Electrical Eng., Korea Univ.

June 11, 2021

1. (a) Find a discrete-time signal, of which both the Fourier and Z transforms exist.
(b) Find a discrete-time signal, of which only the Z transform exists.
(c) Find a discrete-time signal, of which neither the Fourier transform nor the Z transform
exists.

2. (a) Prove that



X ∞
2π X
e−jkωT = δ(ω − 2πk/T )
T
k=−∞ k=−∞
using

X ∞
1 X j2kπt/T
δ(t − kT ) = e .
T
k=−∞ k=−∞

(b) The signal xs (t) obtained by sampling x(t) at every T sec. can be modeled as

X
xs (t) = x(t)p(t) where p(t) = δ(t − kT ).
k=−∞

Write the Fourier transform, Xs (jω), of xs (t) in terms of X(jω). Sketch Xs (jω) and
X(jω), assuming that X(jω) is band-limited.

3. Consider a causal signal x(t) (i.e. x(t) = 0 for t < 0), of which the Laplace transform is
s3 + 2s2 + 3s + 2
X(s) =
s4 + 2s3 + 2s2 + 2s + 2
(a) Give the region of convergence for X(s). Note that the poles of X(s) are −1.1898 ±
0.6028i and 0.1898 ± 1.0432i, and the zeros of X(s) do not coincide with these poles.
(b) Determine the Laplace transforms of the following signals:
· y(t) = 3x(t/3) · y(t) = tx(t) · y(t) = tx(t − 1)
Rt
· y(t) = dx(t)
dt · y(t) = 0 x(τ )dτ · y(t) = (t − 1)x(t − 1) + dx(t)
dt

You do not have to prove the properties of the Laplace transform that you use to
evaluate these transforms; however do not forget to mention the regions of conver-
gence.

4. Suppose that the input x(t) and the output y(t) of a causal linear time-invariant system
satisfy the following differential equation
N
X N
X −1
dk y(t) dk x(t)
ak = bk , t ≥ 0.
dtk dtk
k=0 k=0

1
(a) Assuming that y(t), dy(t)/dt, · · ·, dN −1 y(t)/dtN −1 are all zero at t = 0−, find the
Laplace transform, H(s), of the impulse response h(t).
(b) Suppose now that some of the initial values are fixed to be nonzero. Explain that
Y (s) is no longer equal to H(s)X(s) where X(s) and Y (s) are the Laplace transforms
of x(t) and y(t).
(c) Note that H(s) can be computed by Y (s)/X(s) under the zero initial conditions (even
when the coefficients of the differential equation are unknown). Do you think you
can still compute H(s) using X(s) and Y (s) under the nonzero initial conditions in
part (b)?

5. Consider a linear time invariant system described by

d2 y(t)
y(t) − = x(t),
dt2
where x(t) and y(t) are the input and output, respectively.

(a) Suppose that this system is stable. What is the impulse response of the system?
Determine also whether this system is causal or not.
(b) Suppose that this system is causal and unstable. What is the impulse response of the
system?
(c) Suppose that this system is noncausal and unstable. What is the impulse response
of the system?

6. Consider the following control system:

Controller Plant
r(t) + e(t) x(t) y(t)
C(s) G(s)
-

where the design objective is to achieve (i) stability and (ii) no steady state error (i.e.
limt→∞ e(t) = 0) when r(t) = u(t).

(a) Consider the following controller


k2
C(s) = k1 + .
s
• Show that the objective can be achieved if G(s) = 1/s.
• Show that the objective cannot be achieved if G(s) = 1/s2 .
(b) Modify the controller in part (a) such that the control objective can be achieved for
G(s) = 1/s2

Note that the roots of s3 + a2 s2 + a1 s + a0 all have negative real parts if a2 a1 > a0 > 0.

You might also like