DISCRETE-TIME SYSTEMS DESCRIBED BY DIFFERENCE EQUATIONS

 A linear and time-invariant systems are characterized by the unit sample response h(n).
 To determine the output y(n) of the system for any given input sequence x(n) by means
of the convolution summation, i.e.

y (n)  
k  
h( k ) x ( n  k ) (1)

 In general, linear and time-invariant systems is characterized by the input-output

relationship in Eq.(1).
 In the FIR systems, the realization involves additions, multiplications and a finite
number of memory locations.
 An FIR system is implemented directly, as implied by the convolution summation.
 In the IIR systems, its practical implementation as implied by convolution is clearly
impossible.
 Since it requires an infinity number of memory locations, multiplication and additions.
 In IIR systems, this family of discrete-time systems is more conveniently described by
difference equations.

October 28, 2022 1

 Recursive and Nonrecursive Discrete-Time Systems

 To compute the cumulative average of a signal x(n) in the interval 0 ≤ k ≤ n, defined as

1 n
y (n)   x(k )
n  1 k 0
n  0, 1, ....... (2)

 The computation of y(n) requires the storage of all the input samples x(k) for 0 ≤ k ≤ n.
 Since n is increasing, the memory requirements grow linearly with time.
 The output y(n) can be computed more efficiently by utilizing the previous output value
y(n - 1).
 By a simple algebraic rearrangement of Eq.(2), we obtain
n 1
( n  1) y ( n)   x ( k )  x ( n)
k 0

 ( n  1) y ( n)  ny ( n  1)  x (n )
n 1
 y ( n)  y ( n  1)  x( n) (3)
n 1 n 1

October 28, 2022 2

Recursive and Nonrecursive Discrete-Time Systems
 The computation of y(n) by means of Eq.(3) requires two multiplications, one addition
and one memory location, as illustrated in Fig. 1.

Figure 1: A recursive cumulative averaging system.

 This is an example of a recursive system.
 In general, a system whose output y(n) at time n depends on any number of past
output values y(n - 1), y(n - 2 ), . . . is called a recursive system.
 The computation of the recursive system, with n = 0 and proceed forward in time.
 Thus, according to Eq.(3), we obtain
y (0)  x(0)  n 1 
 y ( n)  y (n  1)  x ( n) 
y (1)  1 y (0)  1 x(1) n 1 n 1 
2 2
y (2)  2 y (1)  1 x (2)
3 3
and so on.
October 28, 2022 3
Recursive and Nonrecursive Discrete-Time Systems
 At n = no, the output begins
n0 1
y (n0 )  y (n0  1)  x(n0 ) (4)
n0  1 n0  1
and proceeds forward intime until sometime, say n = n1.
 The term y(no - 1) is called the initial condition for the system in Eq. (4).
 The output of a causal and practically realizable recursive system can be expressed in
general as
y ( n)  F [( y ( n  1), y ( n  2), ......., y (n  N ), x (n ), x ( n  1), ....., x ( n  M )] (5)
where F[.] denotes some function of its arguments.
 This is a recursive equation specifying a procedure for computing the system output
in terms of previous values of the output and present and past inputs.
 If y(n) depends only on the present and past inputs, then
y ( n)  F [ x ( n), x ( n  1), ....., x (n  M )] (6)

 Such a system is called nonrecursive system.

October 28, 2022 4

 Recursive and Nonrecursive Discrete-Time Systems
 The convolution summation for a causal FIR system is
M
y ( n )   h( k ) x (n  k )
k 0

 h(0) x ( n)  h(1) x ( n  1)  ......  h( M ) x ( n  M ) (7)

 F [ x (n)  x( n  1)  ......  x( n  M )]
where the function F[.] is simply a linear weighted sum of present and past inputs
and the impulse response values h(n), 0 ≤ n ≤ M, constitute the weighting
coefficients.
 The basic differences between nonrecursive and recursive systems are illustrated in
Figure 2.

(a) nonrecursive
(b) recursive system
Figure 2: Basic form for a causal and realizable.
October 28, 2022 5
 Linear Time-Invariant Systems Characterized by Constant-Coefficient
Difference Equations
 A family of linear time-invariant systems described by an input-output relation called
a difference equation with constant coefficients.
 The constant-coefficient linear difference equation systems are a subclass of the
recursive and non-recursive systems.
 Let us consider that we have a recursive system with an input-output equation
y(n)  a y(n  1)  x(n) (8)
where a is a constant.
 Figure 3 shows a block diagram realization of the system.

Figure 3: Block diagram of a recursive system.

6
October 28, 2022
Constant-Coefficient Difference Equations
 We compute successive values of y(n) for n ≥ 0, beginning with y(0).
 Thus
y(0)  ay(1)  x(0)
y(1)  ay(0)  x(1)  a 2 y(1)  ax(0)  x(1)
y(2)  ay (1) x(0)  x(2)  a3 y(1)  a 2 x(0)  ax(1)  x(2)
 
y(n)  ay(n  1)(0)  x(n)  a n 1 y(1)  a n x(0)  a n 1 x(1)    ax(n  1)  x(n)
or, more compactly,
n
y (n)  a n 1
y(1)   a k x(n  k ) n0 (9)
k 0

 The response y(n) of the system as given by the right-hand side of (9) consists of two
parts.
 The first part, which contains the term y(- 1), is a result of the initial condition y(-1) of
the system.
 The second part is the response of the system to the input signal x(n).

October 28, 2022 7

Constant-Coefficient Difference Equations

 If the system is initially relaxed at time n = 0, then its memory should be zero.
 Hence,
y(-1) = 0.
 Thus, a recursive system is relaxed if it starts with zero initial conditions.
 This is called the zero state of the system.
 Its corresponding output is called the zero-state response or forced response, and it is
denoted by yzs(n).
 The zero-state response or forced response of the system is given by
n
yzs (n)   a k x(n  k ) n0 (10)
k 0

 Equation (10) is a convolution summation involving the input signal convolved with the
impulse response
h( n)  a nu ( n) (11)
 The system described by the first-order difference equation in Eq. (8) is causal.
 As a result, the lower limit on the convolution summation in Eq. (10) is k = 0.
October 28, 2022 8
 Constant-Coefficient Difference Equations

 Furthermore, the condition y(-1) = 0 implies that the input signal can be assumed
causal and hence the upper limit on the convolution summation in Eq. (10) is n, since
x(n - 1) = 0 for k > n.
 The relaxed recursive system described by the first-order difference equation in (8) is a
linear time-invariant IIR system with impulse response given by Eq. (11).
 Suppose that, the system described by (8) is initially non-relaxed [i.e., y(- 1) ≠ 0] and
the input x(n) = 0 for all n.
 Then, the output of the system with zero input is called the zero-input response or
natural response and is denoted by yzi(n).
 From Eq. (8), with x(n) = 0 for - ∞ < n < ∞, we obtain

yzi (n)  a n1 y(1) n0 (12)

 The recursive system with nonzero initial condition is nonrelated in the sense that it
can produce an output without being excited.
 The zero-input response is due to the memory of the system.

October 28, 2022 9

Constant-Coefficient Difference Equations
 In general, the total response of the system can be expressed as
y(n)  yzi (n)  yzs (n) (13)
 The general form for recursive systems such an equation is
N M
y (n)   ak y ( n  k )   bk x (n  k )
k 1 k 0 (14)
or, equivalently,
N M

a k y (n  k )   bk x (n  k ), a0  1
k 0 k 0 (15)
 The integer N is called the order of the difference equation or the order of the system.
 The negative sign on the right-hand side of (14) is introduced as a matter of
convenience to allow us to express the difference equation in (15) without any
negative signs.
A system is linear if it satisfies the following three requirements:
1. The total response is equal to the sum of the zero-input and zero-state responses
[i.e. y(n) = yzi(n) + yzs(n)].
2. The principle of superposition applies to the zero-state response (zero-state linear).
3. The principle of superposition applies to the zero-input response (zero-input linear).
October 28, 2022 10
 Solution of Linear Constant-Coefficient Difference Equation

 A linear constant-coefficient difference equation as the input-output relationship

describing a linear time-invariant system.
 The output y(n), n ≥ 0, of the system given a specific input x(n), n ≥ 0, and a set of
initial conditions.
 The direct solution method assumes that the total solution is the sum of two parts:
y ( n )  yh ( n )  y p ( n ) (16)
 The part yh(n) is known as the homogeneous or complementary solution, whereas
yp(n) is called the particular solution.

The homogeneous solution of a difference equation

 The linear constant-coefficient difference equation given by assuming that the input
x(n) = 0.
 Thus, the solution to the homogeneous difference equation
N

a
k 0
k y (n  k )  0
(17)
October 28, 2022 11
The homogeneous solution of a difference equation
 We assume that the solution is in the form of an exponential, that is
yh ( n )   n (18)
where the subscript h on y(n) is used to solve the homogeneous difference equation.
 If we substitute this assumed solution in (17), we obtain the polynomial equation
N

a 
k 0
k
nk
0

or  n  N ( N  a1 N 1  a2  N  2    aN 1  aN )  0 (19)

 The polynomial is called the characteristic polynomial of the system.
 Let us assume that the roots are distinct, i.e. there are no multiple-order roots.
 Then the general solution to the homogeneous difference equation in Eq.(17) is
yh ( n)  C11n  C2 2n    C N Nn (20)
where C1 , C2 ,  CN are weighting coefficients.
 If the characteristic equation contains multiple roots, the form of the solution given in
Eq. (20) must be modified.
 For example, if λ1 is a root of multiplicity m, then Eq. (20) becomes
yh (n)  C11n  C2 n1n  C3 n 2 1n    C N n m 11n  Cm1mn 1  C N n (21)
October 28, 2022 12
The particular solution of the difference equation:
 The particular solution yp(n) is required to satisfy the difference equation Eq. (15) for
the specific input signal x(n), n ≥ 0.
 In other words, yp(n) is any solution satisfying
N M

a
k 0
k y p (n  k )   bk x(n  k )
k 0
a0  1 (22)

 To solve (22), we assume for yp(n), a form that depends on the form of the input x(n).

The total solution of the difference equation:

 The linearity property of the linear constant-coefficient difference equation allows us

to add the homogeneous solution and the particular solution in order to obtain the
total solution.
 Thus
y ( n )  yh ( n )  y p ( n ) (23)

 The resultant sum y(n) contains the constant parameters {Ci} embodied in the
homogeneous solution component yh(n).
 These constants can be determined to satisfy the initial conditions.
October 28, 2022 13
 The Impulse Response of a Linear Time-Invariant Recursive System

 The impulse response of a linear time-invariant system is defined as the response of

the system to a unit sample excitation [i.e., x(n) = δ(n)].
 A recursive system, x(n) is simply equal to the zero-state response of the system when
the input x(n) = δ(n) and the system is initially relaxed.
 Let us consider a simple first-order recursive system, the zero-state response given by
n
y zs (n)   a k x(n  k ) (24)
k 0

 With x(n) = δ(n) is substituted into (24), we obtain

n
y zs (n)   a k  (n  k )  a n n0
k 0

 Hence, the impulse response of the first-order recursive system is

h( n)  a n u ( n) (25)
 In the general case, linear time-invariant recursive system, the zero-state response
expressed in terms of the convolution summation is
n
y zs (n)   h(k ) x( n  k ) n0 (26)
k 0
October 28, 2022 14
The Impulse Response of a Linear Time-Invariant Recursive System

 When the input is an impulse [i.e., x(n) = δ(n)], Eq. (26) reduces to
y zs (n)  h(n) (27)
 Let us consider the problem of determining the impulse response h(n) given a linear
constant-coefficient difference equation description of the system.
 We have established the fact that the total response of the system to any excitation
function consists of the sum of two solutions of the difference equation: the solution to
the homogeneous equation plus the particular solution to the excitation function.
 In the case where the excitation is an impulse, the particular solution is zero, since
x(n) = 0 for n > 0, that is
y p ( n)  0
 Consequently, the response of the system to an impulse consists only of the solution
to the homogeneous equation, with the [Ck] parameters evaluated to satisfy the initial
conditions dictated by the impulse.

October 28, 2022 15

The Impulse Response of a Linear Time-Invariant Recursive System

 When the system is described by an Nth-order linear difference equation, the solution
of the homogeneous equation is
N
yh (n)   Ck kn (28)
k 1

when the roots {λk} of the characteristic polynomial are distinct.

 The impulse response of the system is
N
h(n)   Ck kn (29)
k 1
where the parameters {Ck} are determined by setting the initial conditions
y (-1) = . . . = y( - N ) = 0.
 Since BIBO stability requires that the impulse response be absolutely summable,
then, for a causal system, we have
  N N 

 h( n)    Ck   Ck  k
n
n
k  (30)
n 0 n  0 k 1 k 1 n 0

October 28, 2022 16

The Impulse Response of a Linear Time-Invariant Recursive System

 Now if |λk| < 1 for all k, then




n
k 
n 0

and hence


 h ( n)  
n 0

 On the other hand, if one or more of the |λk| ≥ 1, h(n) is no longer absolutely
summable, and consequently, the system is unstable.

October 28, 2022 17

October 28, 2022 18