ELC 4351: Digital Signal Processing

Liang Dong

Electrical and Computer Engineering

Baylor University
liang [email protected]

September 6, 2016

Liang Dong (Baylor University) Difference Equations September 6, 2016 1 / 15

Discrete-time Systems Described by Difference Equations

A LTI system is characterized by its unit sample response h(n).

The output y (n) of the system for any given input x(n) is determined by
∞
X
y (n) = h(k)x(n − k)
k=−∞

y (n) = h(n) ⊗ x(n)

FIR systems vs. IIR systems

Liang Dong (Baylor University) Difference Equations September 6, 2016 2 / 15

Recursive and Nonrecursive Discrete-time Systems
e.g. Cumulative average of signal x(n)

n
1 X
y (n) = x(k)
n+1
k=0

n−1
X
(n + 1)y (n) = x(k) + x(n)
k=0
= ny (n − 1) + x(n)

n 1
y (n) = y (n − 1) + x(n)
n+1 n+1
where y (n0 − 1) is the initial condition for the system at time n = n0 .
Liang Dong (Baylor University) Difference Equations September 6, 2016 3 / 15
LTI Systems Characterized by Constant-Coefficient
Difference Equations

A recursive system:
y (n) = αy (n − 1) + x(n)
where α is a constant.

y (0) = αy (−1) + x(0)

y (1) = αy (0) + x(1) = α2 y (−1) + αx(0) + x(1)
.. ..
. .
Xn
y (n) = αn+1 y (−1) + αk x(n − k), n ≥ 0
k=0

Liang Dong (Baylor University) Difference Equations September 6, 2016 4 / 15

LTI Systems Characterized by Constant-Coefficient
Difference Equations

n
X
n+1
y (n) = α y (−1) + αk x(n − k), n ≥ 0
k=0
1 The system is initially relaxed at time n = 0, i.e., y (−1) = 0.
Zero-state response or forced response
n
X
yzs (n) = αk x(n − k), n ≥ 0
k=0

Liang Dong (Baylor University) Difference Equations September 6, 2016 5 / 15

LTI Systems Characterized by Constant-Coefficient
Difference Equations

n
X
n+1
y (n) = α y (−1) + αk x(n − k), n ≥ 0
k=0
1 The system is initially relaxed at time n = 0, i.e., y (−1) = 0.
Zero-state response or forced response
n
X
yzs (n) = αk x(n − k), n ≥ 0
k=0
2 The input x(n) = 0, ∀n.
Zero-input response or natural response
yzi (n) = αn+1 y (−1), n ≥ 0

Liang Dong (Baylor University) Difference Equations September 6, 2016 5 / 15

LTI Systems Characterized by Constant-Coefficient
Difference Equations

n
X
n+1
y (n) = α y (−1) + αk x(n − k), n ≥ 0
k=0
1 The system is initially relaxed at time n = 0, i.e., y (−1) = 0.
Zero-state response or forced response
n
X
yzs (n) = αk x(n − k), n ≥ 0
k=0
2 The input x(n) = 0, ∀n.
Zero-input response or natural response
yzi (n) = αn+1 y (−1), n ≥ 0
3 The total response of the system
y (n) = yzs (n) + yzi (n)
Liang Dong (Baylor University) Difference Equations September 6, 2016 5 / 15
LTI Systems Characterized by Constant-Coefficient
Difference Equations

For LTI systems, a general form of the input-output relationship.

N
X M
X
y (n) = − ak y (n − k) + bk x(n − k)
k=1 k=0

N
X M
X
ak y (n − k) = bk x(n − k), a0 ≡ 1
k=0 k=0

The integer N is the order of the difference equation or the order of the
system.

Liang Dong (Baylor University) Difference Equations September 6, 2016 6 / 15

Linear Systems Requirements

A Linear System:

The total response is equal to the sum of the zero-state and

zero-input responses

y (n) = yzs (n) + yzi (n)

Liang Dong (Baylor University) Difference Equations September 6, 2016 7 / 15

Linear Systems Requirements

A Linear System:

The total response is equal to the sum of the zero-state and

zero-input responses

y (n) = yzs (n) + yzi (n)

The principle of superposition applies to the zero-state response.

Liang Dong (Baylor University) Difference Equations September 6, 2016 7 / 15

Linear Systems Requirements

A Linear System:

The total response is equal to the sum of the zero-state and

zero-input responses

y (n) = yzs (n) + yzi (n)

The principle of superposition applies to the zero-state response.

The principle of superposition applies to the zero-input response.

Liang Dong (Baylor University) Difference Equations September 6, 2016 7 / 15

Solution of Linear Constant-Coefficient Difference
Equations

The direct solution

y (n) = yh (n) + yp (n)
| {z } | {z }
homogeneous solution particular solution

Liang Dong (Baylor University) Difference Equations September 6, 2016 8 / 15

The Homogeneous Solution of A Difference Equation
The homogeneous difference equation:
N
X
αk y (n − k) = 0
k=0

We assume that the solution is in the form of an exponential, i.e.,

yh (n) = λn .

Liang Dong (Baylor University) Difference Equations September 6, 2016 9 / 15

The Homogeneous Solution of A Difference Equation
The homogeneous difference equation:
N
X
αk y (n − k) = 0
k=0

We assume that the solution is in the form of an exponential, i.e.,

yh (n) = λn .
Substituting this in the equation, we obtain the polynomial equation
N
X
αk λn−k = 0
k=0

λn−N (λN + α1 λN−1 + · · · + αN−1 λ + αN ) = 0

| {z }
characteristic polynomial

Liang Dong (Baylor University) Difference Equations September 6, 2016 9 / 15

The Homogeneous Solution of A Difference Equation
The homogeneous difference equation:
N
X
αk y (n − k) = 0
k=0

We assume that the solution is in the form of an exponential, i.e.,

yh (n) = λn .
Substituting this in the equation, we obtain the polynomial equation
N
X
αk λn−k = 0
k=0

λn−N (λN + α1 λN−1 + · · · + αN−1 λ + αN ) = 0

| {z }
characteristic polynomial

The characteristic polynomial of the system has N roots:

λ1 , λ2 , . . . , λN .
Liang Dong (Baylor University) Difference Equations September 6, 2016 9 / 15
The Homogeneous Solution of A Difference Equation

If the N roots are distinct, the general solution to the homogeneous

difference equation is

yh (n) = C1 λn1 + C2 λn2 + · · · + CN λnN

where C1 , C2 , . . . , CN are weighting coefficients.

These coefficients are determined from the initial conditions of the system.

yh (n) is the zero-input response of the system.

Liang Dong (Baylor University) Difference Equations September 6, 2016 10 / 15

The Homogeneous Solution of A Difference Equation

If the characteristic polynomial contains multiple roots, e.g. λ1 is a root of

multiplicity m, then

hh (n) = C1 λn1 + C2 nλn1 + · · · + Cm nm−1 λn1 + Cm+1 λnm+1 + · · · + CN λnM

Liang Dong (Baylor University) Difference Equations September 6, 2016 11 / 15

The Particular Solution of A Difference Equation

The particular difference equation for a specific input signal x(n):

N
X M
X
ak yp (n − k) = bk x(n − k), a0 ≡ 1
k=0 k=0

Input Signal x(n) Particular Solution yp (n)

A K
AM n KM n
AnM K0 nM + K1 nM−1 + · · · + KM
An nM A (K0 nM + K1 nM−1 + · · · + KM )
n

A cos ω0 n K1 cos ω0 n + K2 sin ω0 n

A sin ω0 n K1 cos ω0 n + K2 sin ω0 n

Liang Dong (Baylor University) Difference Equations September 6, 2016 12 / 15

The Total Solution of A Difference Equation

y (n) = yh (n) + yp (n)

Liang Dong (Baylor University) Difference Equations September 6, 2016 13 / 15

The Impulse Response of a LTI Recursive System

The impulse response h(n) is equal to the zero-state response of the

system when the input x(n) = δ(n) and the system is initially relaxed.
n
X
yzs (n) = h(k)x(n − k), n ≥ 0
k=0

When x(n) = δ(n), yzs (n) = h(n).

Liang Dong (Baylor University) Difference Equations September 6, 2016 14 / 15

The Impulse Response of a LTI Recursive System

The impulse response h(n) is equal to the zero-state response of the

system when the input x(n) = δ(n) and the system is initially relaxed.
n
X
yzs (n) = h(k)x(n − k), n ≥ 0
k=0

When x(n) = δ(n), yzs (n) = h(n).

If the excitation is an impulse, the particular solution is zero, since

x(n) = 0, ∀n > 0. That is yp (n) = 0.

The response of the system to an impulse consists only of the solution

to the homogeneous equations.

Liang Dong (Baylor University) Difference Equations September 6, 2016 14 / 15

The Impulse Response of a LTI Recursive System

Nth-order linear difference equation.

The solution of the homogeneous equation is

N
X
yh (n) = Ck λnk .
k=1

Hence, the impulse response of the system is

N
X
h(n) = Ck λnk .
k=1

Liang Dong (Baylor University) Difference Equations September 6, 2016 15 / 15