Lecture 2: Discrete-Time Systems and
z-Transform
• Discrete-time signals v.s. continuous-time signals
• Discrete-time systems v.s. continuous-time systems
• z-Transform
• Inverse z-Transform
• z-Transform for solving linear difference equations
Continuous-Time Signals
• A signal that changes continuously in time:
e(t), −∞ < t < ∞, or t ≥ 0
• Defined at all times t (no gap)
• Signal can take arbitrary values
• Example: temperature, position, velocity,…
1
Discrete-Time Signals
• A signal (sequence, or series) whose
values are defined only at discrete times:
e(k), k = · · · , −1, 0, 1, · · · or k = 0, 1, 2, · · ·
• Signal defined only at integer times k
• Signal can still take arbitrary values
Where do discrete-time signals come from?
• Some come naturally
– Population of a species in different generations
– Annual growth percentage of GDP
– Results of a numerical algorithm in different rounds of iteration
• Some arise by sampling continuous-time signals at
regular time intervals, say, every T seconds
e(t), −∞ < t < ∞
⇒ e(kT ), k = · · · , −1, 0, 1, · · ·
⇒ e(k), k = · · · , −1, 0, 1, · · · , (if T is known from the context)
2
Digital Signals
Actuators System Sensors
y(t)
continuous-time
signal
ŷ(kT ) y(kT )
c(kT )
D/A Computer Quantizer
discrete-time
signal
digital signal
• The process of representing an arbitrary continuous value y(kT) with
binary bits of finite word length is called quantization (A/D)
• Quantized signal is discrete in both time and value, called a digital signal,
• Errors will be incurred as a result of quantization
• The larger the word length, the smaller the average quantization error
• In most of the course, we assume zero quantization error
Continuous-Time LTI Systems
• A continuous-time system has input and output that are
both continuous-time signals
• Continuous-time linear time-invariant (LTI) system given
by differential equation:
n
y(t) = βn d dte(t)
n + · · · + β1 de(t)
dt + β0 e(t)
dn y(t)
−αn dtn − · · · − α1 dy(t)
dt
• Model completely determined by the order n of the
system, the coefficiencts α and β
3
Transfer Functions of Continuous-Time
LTI Systems
• Taking Laplace transform (assuming zero initial conditions):
Y (s) βn sn +···+β1 s+β0
H(s) = E(s)
= αn sn +···+α1 s+1
• H(s) is called the transfer function
• For LTI system, H(s) is a rational function (fraction of two
polynomial functions of s)
• Impulse response is the inverse Laplace transform of H(s)
Discrete-Time LTI Systems
• A discrete-time system has input and output that are
both discrete-time signals
• Discrete-time linear time-invariant (LTI) system given by
linear difference equation
y(k) = bn e(k) + bn−1 e(k − 1) + · · · + b0 e(k − n)
−an−1 y(k − 1) − · · · − a0 y(k − n)
• Current output depends on a finite history of input/output
• Model completely determined by the order n of the
system, and the coefficients a’s and b’s
4
Example of Discrete-Time LTI Systems
Example 1: Population of A Species (Fibonacci)
y(k): population of the k-th generation, with y(0)=1
Assumptions:
(i) each individual starts duplicating itself after two generations
(ii) no death from aging, and no migration
Discrete-Time LTI Systems As
Approximation of Continuous-Time Systems
A continuous-time LTI system:
Sample every T seconds
Approximating discrete-time LTI system: What difference
equation can approximate the behavior of the above
continuous-time system?
5
Euler’s Methods
First order derivative
Euler Forward Approximation:
dy(kT ) y((k+1)T )−y(kT )
dt ≃ T
Euler Backward Approximation:
dy(kT ) y(kT )−y((k−1)T )
dt ≃ T
Higher order derivatives
d2 y(kT )
dt2 ≃
Approximating Discrete-Time System
Example 2: a continuous-time system
Approximating discrete-time system:
6
z-Transform
• Given a discrete-time signal
f (k), k = · · · , −1, 0, 1, · · · or k = 0, 1, 2, · · ·
Its (double-sided) z-transform is defined by
∞
F (z) = Z[f (k)] = k=−∞ f (k)z −k , k = · · · , −1, 0, 1, · · ·
Its (single-sided) z-transform is defined by
∞
F (z) = Z[f (k)] = k=0 f(k)z −k , k = 0, 1, 2, · · ·
• z-transform F(z) is a function of the complex variable z,
and is only well defined on a region of the z-plane (DOC)
• We will always assume single-sided (causal) signals and
single-sided z-transform, unless otherwise stated
Discrete-Time Unit Impulse Function
δ(k) = 1, if k = 0, δ(k) = 0, otherwise
δ(k − n)
(n is an integer)
7
More General Discrete-Time Signals
Discrete-time unit step function
u(k) = 1, k = 0, 1, 2, · · ·
z
U(z) = z−1 DOC: |z −1 | < 1
More generally
f(k) = rk , k = 0, 1, 2, · · ·
F (z) = DOC:
Properties of z-Transform
Linearity: (α,β∈ℜ)
Real Translation (Translation in time): n is a positive integer
(shift right)
(shift left)
Example: f (k) = rk , n = 1
8
Properties of z-Transform
Initial Value Theorem:
f (0) = limz→∞ F (z)
Final Value Theorem:
f (∞) = limk→∞ f(k) = limz→1 (z − 1)F (z) (if exists)
Example:
f (k) = u(k) f (k) = r k
Properties of z-Transform
k
Convolution: let f1 (k) ∗ f2 (k) = l=0 f1 (l)f2 (k − l) be the convolution of
the two discrete-time signals. Then
Z[f1 (k) ∗ f2 (k)] = F1 (z)F2 (z)
Example:
dF (z)
Another property: Z[k f (k)] = −z dz
rz
Example: Z[kr k ] = (z−r)2
Complex Translation: Z[r k f (k)] = F (zr −1 ) for arbitrary complex number r
A list of the properties of z-transform can be found in Table 2-2, pp. 37.
9
Inverse z-Transform
• Given the z-transform F(z), how do we find the sequence
f (k), k = 0, 1, 2, · · ·
so that F(z) is the z-transform of f(k):
∞ −k
F (z) = k=0 f (k)z .
• The above f(k) is called the inverse z-transform of F(z):
f(k) = Z −1 [F (z)]
• Can find inverse z-transform in different ways:
• Power series method (long division)
• Inversion-formula method
• Partial fraction expansion
Power Series (Long Division) Method
A rational F(z) is the ratio of two polynomials in z:
b(z) bm z m +···+b1 z+b0
F (z) = a(z) = an z n +···+a1 z+a0
If m ≤ n, we can use long division to expand F(z) in power series:
∞
F (z) = k=0 f (k)z−k
to obtain the inverse z-transform f(k), k=0,1,…
Example: 1
F (z) = z 2 −3z+2
Drawbacks: good by hand or computer for a few f(k), no idea about general f(k)
10
Inversion Formula Method
The inverse z-transform of F(z) is given by
f(k) = [residue of F (z)z k−1 ]z=p
where the summation is over all the poles p of F(z)zk-1
residue at a simple pole z=a:
(residue)z=a = (z − a)F (z)z k−1 z=a
residue at a pole z=a of multiplicity m:
1 dm−1
(residue)z=a = (m−1)! dz m−1 (z − a)m F (z)z k−1 z=a
z
Example: F (z) = z 2 −3z+2
Example: 1
F (z) = z(z−1)
11
Partial Fraction Expansion Method
z
• If F (z) = z−p , then from previous result
f (k) = Z −1 [F (z)] = pk , k = 0, 1, 2, . . ..
pz
• If F (z) = (z−p)2 , then
f (k) = Z −1 [F (z)] = kpk , k = 0, 1, 2, . . ..
• ….
• Idea: for general rational
b(z) bm z m +···+b1 z+b0
F (z) = a(z) = an z n +···+a1 z+a0 (m ≤ n)
try to find the partial fraction expansion of F(z)/z, and use
the linearity of the z-transform to find the inverse transform
Simple Case
Assume that the denominator of F(z) can be factorized as
bm z m +···+b1 z+b0
F (z) = (z−p1 )···(z−pn )
and the poles p1,…,pn are all distinctive and nonzero. Then
F (z) c0 c1 c2 cn
z = z + z−p1 + z−p2 + ··· + z−pn
where the constants are residues given by
c0 = F (z)z=0 ci = (z − pi ) F z(z) z=pi
Matlab command: residue
As a result,
Z −1 [F (z)] = c0 δ(k) + c1 pk1 + c2 pk2 + · · · + cn pkn
12
Example
1
F (z) = z2 −3z+2
Repeated Poles Case
If F(z) has repeated poles (or pole at 0), for example,
bm z m +···+b1 z+b0
F (z) = (z−p1 )3 (z−p4 )··· (pi
= 0)
Then
F (z) c0 c1 c2 c3 c4
z = z + z−p1 + (z−p1 )2 + (z−p1 )3 + z−p4 + ···
where the constants are residues given by
c0 = F (z)z=0 c4 = (z − p4 ) F z(z) z=p
4
1 d2
3 F (z)
c1 = 2! dz2 (z − p1 ) z z=p1
d
c2 = dz (z − p1 )3 F z(z) z=p
1
c3 = (z − p1 )3 F (z)
z
z=p 1
13
Exercise
1
F (z) = z(z−1) Z −1 [F (z)] =?
Alternative method: find the inverse of z2F(z) first, then shift to the right by two steps
Complex Conjugate Poles Case
Assume that the denominator of F(z) can be factorized as
bm z m +···+b1 z+b0
F (z) = (z−p1 )···(z−pn )
where some poles are complex conjugate pairs.
Example: F (z) = z2z+1 = (z+j)(z−j)
z
F (z) j/2 −j/2 √
z = z+j + z−j (j = −1)
⇒ Z −1 [F (z)] = 2j (−j)k − 2j j k = cos[ π2 (1 − k)]
For general complex poles case, read the textbook pp. 44-45.
14
Solution of Linear Difference Equations
• Given a linear difference equation
y(k) = bn e(k) + bn−1 e(k − 1) + · · · + b0 e(k − n)
−an−1 y(k − 1) − · · · − a0 y(k − n)
• Input e(k) is given for all k=0,1,2,…
• How to compute y(k), k=0,1,2…?
• Essential task for studying discrete-time LTI systems
Solution by Iterations
• The most obvious approach is by iterations
– Easily implemented by computers
f or k=0:N
y(k) = · · ·
..
.
end
– Need to know the initial conditions to jump start the
iterations process:
y(−1), y(−2), . . . , y(−n) and e(−1), e(−2), . . . , e(−n)
– Example: y(k) = e(k) − e(k − 1) − y(k − 1)
15
Zero-State Response by z-Transform
• Assuming zero initial conditions:
y(−1) = y(−2) = . . . = y(−n) = 0
and e(−1) = e(−2) = . . . = e(−n) = 0
• Taking the z-transform of the equation
y(k) = bn e(k) + bn−1 e(k − 1) + · · · + b0 e(k − n)
−an−1 y(k − 1) − · · · − a0 y(k − n)
we have
Y (z) = bn E(z) + bn−1 z −1 E(z) + · · · + b0 z −n E(z)
−an−1 z −1 Y (z) − · · · − a0 z −n Y (z)
bn +bn−1 z −1 +···+b0 z −n
⇒ Y (z) = 1+an−1 z−1 +···+a0 z−n E(z)
• The final output is the inverse z-transform of the above expression
(called zero-state response)
Example (Example 2.9, pp. 39)
Find the output of the system
y(k) = e(k) − e(k − 1) − y(k − 1) k = 0, 1, . . .,
1, if k is even
under the input e(k) = k = 0, 1, . . .,
0, if k is odd
and zero-initial condition: y(−1) = 0, e(−1) = 0
16
Zero-Input Response by z-Transform
Find the output (called zero-input response) of the system
y(k) = e(k) − e(k − 1) − y(k − 1)
under zero input e(k) = 0, k = 0, 1, . . ., and non-zero initial condition
y(−1) = 1, e(−1) = −1
Idea: instead of dealing with double-sided signal y(k) and e(k), we define
ŷ(k) = y(k − 1), ê(k) = e(k − 1)
17
Full Response
• Superposition Law: For an LTI system, the response
under non-zero input and non-zero initial condition is the
sum of the zero-state and zero-input responses
• Example: Find the output of the system
y(k) = e(k) − e(k − 1) − y(k − 1)
under the input
1, if k is even
e(k) = k = 0, 1, . . .,
0, if k is odd
and the initial condition:
y(−1) = 1, e(−1) = −1
18