7.

1 Solving Linear Differential Equations

Dynamics of time invariant, linear, continuous-time systems is described by th

order linear differential equations with constant coefficients


 


 
 
 
 

 

  
 
 

where and represent, respectively, the system input and output signals,
and  are constant coefficients.

This differential equation corresponds to a real physical system for . The

initial conditions that carry information about and its derivatives at






the initial time , that is, are

assumed to be known. The reason for specifying the system initial conditions at


instead of at is twofold. Firstly, we have to be able to find the

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–1
system impulse response due to the impulse delta signal located at the time origin,

, which requires integration from  in order to be able to completely include

the impulse delta signal within the integration limits. Secondly, the system zero-

input response is contributed by the system initial conditions that are present in the

system before the input is applied to the system. Assuming that the initial time at

which the system input is applied is , the system initial conditions must be

defined before , that is at  .

It is known from elementary theory of differential equations that the solution

of a differential equation has two components:  –particular solution and

 –homogeneous solution, that is

 

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–2
 The homogeneous solution satisfies the corresponding homogeneous differential

equation, obtained by setting , that is

      
     

  

The particular solution is any solution that satisfies the given differential equation

regardless of the system initial conditions.

Homogeneous Solution

The characteristic equation of the given differential equation is defined by

        

The solutions of the characteristic equation (roots of the characteristic polynomial

) are called the system characteristic values. The characteristic values are

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–3
also known under the name system eigenvalues.

Depending on the nature of the solutions of the characteristic equation, we can

get in a systematic way the expression for the homogeneous solution.

Let us first assume that the solutions of the characteristic equation are distinct,

that is

! "

with ! " , then the homogeneous differential equation has a

solution given by

# $&%(' ! $*)+' " $-,.'

The constant coefficients / have to be determined from the

system initial conditions in the manner explained later on in this section. It can be

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–4
shown that the form of the given homogeneous solution guarantees the existence

of the unique constants 0 .

Example 7.1: Consider a continuous-time linear system whose mathematical

model is described by the following homogeneous linear differential equation

1 2 2
2
1

The characteristic equation of this differential equation is

1

which implies the following characteristic values 3 1 . The

corresponding homogeneous solution is given by

2 1 475
3 465

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–5
 Note that a pair of complex characteristic values, say 89;: , represents

two distinct characteristic values. Hence, we can use the previously given formula

to find the corresponding homogeneous solution, that is

=?>@ACBEDGF =?>IHJA.BEDGF >EF ACBKF HJA.BKF
< 8 : 8 :

However, this solution can be further simplified by using Euler’s formula and

determining the constants 8 and : from the initial conditions. Since <
is a real function of time, its final expression can not contain complex numbers. In

general, a pair of complex conjugate characteristic values implies oscillatory system

response so that we must obtain sine and cosine functions in the final expression

for < . That is why, a more convenient form for the homogeneous solution in

the case of complex conjugate characteristic values is

>EF >EF
< 8 :

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–6
 Example 7.2: The homogeneous linear differential equation given below has

the complex conjugate characteristic values

L M M
M
L

that is
L

which implies N L . The corresponding homogeneous

solution is given by
M L O6P
N O6P

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–7
 In the case when we have a multiple root, say the characteristic value Q has
multiplicity , and the remaining characteristic values are distinct, the homogeneous

solution has the form

R Q S T S U WU V Q XZY\[

U^] Q X`_(a Y [ b X`cC[

In general, a characteristic value d of multiplicity d generates a term of the form

d d] Q ]d UfegV S U e V S ]d UhegV Q U e V Q X e[

in the homogeneous solution of the corresponding differential equation. It can be

shown that the chosen form for R facilitates the existence of unique coefficients
i .

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–8
 In the case of multiple complex characteristic values, for example the charac-

teristic value j j j has multiplicity , the corresponding part of the

homogeneous solution has the form

jk lnm\o jp jq
j j
lrmso tj uj
j j
q
l mo jv wj
j j
p
xWy lrmzo jq p jq lrmzo
xWy j x j

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–9
 Example 7.3: The homogeneous linear differential equation
{ | } | |
|
{ }

has the characteristic equation

{ } }

and the characteristic values ~ } { . The corresponding

homogeneous solution is given by

|
~ 6€ }  }€ { }
 €

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–10
 Example 7.4: Consider the homogeneous linear differential equation with a

triple characteristic value

 ‚ ƒ ‚ ‚
‚
 ƒ

The characteristic equation is given by

 ƒ 

The triple characteristic values are „ ƒ  so that the corresponding

homogeneous solution is
‚
„ …6† ƒ …6†  ƒ …6†

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–11
 Particular Solutions

In Chapters 4 and 6, we have already dealt with a special form of a particular

solution given by the convolution integral

ˆ
‡ ‹Œ
‰^Š

where , representing the system impulse response, is the solution of the

considered differential equation for . The lower integration limit

in the convolution integral is taken at Ž in order to accommodate the impulse

delta function at . Note that when , the system impulse response

contains the impulse delta function at the origin, as demonstrated in the next

example. The delta impulses can also be generated due to the system differentiation

of input signals as demonstrated in Example 4.21. It should be emphasized that

the particular solution defined by the convolution integral represents the system

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–12
zero-state response,  ‘ .

In general, the particular solution,  , is any solution that satisfies the given

differential equation regardless of the system initial conditions. For any  at

the initial time the following relations must be satisfied

 ’
“G”–• “—”–• “”•
 ’

“™˜š”›• “œ˜š”›• “?˜š”•

 ’

This formula requires a bit more explanation, since the system initial conditions
š
are specified at . Let us solve a simple first-order example in order to get

a better understanding of a potential problem caused by the system differention of

the input signal.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–13
 Example 7.5: Consider the following linear differential equation

žŸ  ž

We first solve this differential equation by using the Laplace transform method

from Chapter 4, which leads to

ž
ž

ž ¡›  ž ¡  ž Ÿ  ž ¡›  ž Ÿ 

Note that at ¢ we have ¢ . Hence, an instant

jump occurs at the initial time from

ž to ¢ . This jump in

the initial condition is contributed by the system differentiation of the input signal,

which in general generates the impulse delta function as an additional input to

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–14
the system. The impulse delta function is able to change system initial conditions

instantaneously. Note that £ ¤¥ £ ¤¥ £ ¤¥ .

Let us now find the homogeneous and particular solutions of the given differential

equation. The homogeneous differential equation is given by

¦
¦

Its solution is easily obtained as ¦ § £ ¨¥ . In order to find a

particular solution, we first guess its form as © £ ¤¥ . Substituting this

particular solution in the original differential equation, we find that ,

hence © £ ¤¥ . It follows that

¦ © § £ ¨¥ £ ¤¥

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–15
If we use information about ª we will get either « (if we assume
that the particular solution is not present at ª ) or « , which

implies « . Obviously, both values for « are different from and

hence they produce wrong expressions for . The way to get the correct answer

for « is to use information about ¬ , which is (in this particular example)

equal to ¬ (see the solution obtained using the Laplace transform). That

information will produce « « , leading to the correct

answer for . However, information about the value of ¬ is not given,

in general.

It seems that we are in a trouble while using the classical method for solving a

linear differential equation that corresponds to a real physical linear system since

we need to know the system initial condition at ¬ . In general, it is difficult to

determine the system initial conditions at ¬ based on information about
The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–16
the system differential equation and the system initial conditions at ­ .
However, that problem can be completely avoided if we use the convolution formula

to obtain the particular solution. In such a case, the particular solution represents

the zero-state solution (it does not depend on the system initial conditions) and the

homogeneous solution represents the zero-input solution (it depends only on the

system initial conditions).

In this example, the system impulse response is easily found as

­® ­® ­ ¯°

so that the particular solution obtained from the convolution integral is given by
° °
± ­ ¯¶µ ­ ·^¸°f­µ`¹
²^³ ²´³

­ ¯° ­ ·° º»

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–17
 We can find the constant ¼ in the expression for the homogeneous solution as
½ ¾ ¼ ¾ ¼ ½ ¾ ¿›À Á´Â

which leads to the result obtained using the Laplace transform

½ à Á´Â ÁÄ

¾ ¿›À ¾ ¿À ¾śÀ ¾ ¿À ¾ śÀ

We can conclude from this example that in the case when the system differentiates

the input signal, the classical method for solving linear differential equations with

constant coefficients can be used for finding the response of time invariant linear

systems only if either the particular solution is obtained by using the convolution

formula or if the system initial conditions are known at Æ (which is hardly

known and difficult to determine from information about system’s initial conditions

at ¾ ).

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–18
 Hence, in the process of finding the system response, we have to use the

particular solution obtained through the convolution formula. In such a case,

Ç È´É so that all constants in Ç are obtained from
Ç Ì Ì
ÍÎÏ Í—Î–Ï
ÈË Ç Ì Ì
Ê
Í?Ð Ì ÎÏ ÍœÐ Ì Î›Ï
Ç Ì Ì

In this course, we find the particular solution of linear differential equations

representing continuous-time linear systems through the convolution procedure. In

that direction, we will present in Section 7.4 a time domain method for finding

the continuous-time system impulse response. Note that the frequency domain

technique for finding the continuous-time system impulse response by using the

inverse Laplace transform has been already presented in Chapter 4.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–19
 Note that a procedure for obtaining a particular solution of a linear differential

equation with constant coefficients known as the method of undetermined coeffi-

cients is studied in detail in elementary differential equations courses. That method

is based on guessing a particular solution, usually as a linear combination of poly-

nomial, exponential, and sinusoidal functions (a linear combination of the forcing

function and its derivatives). However, that procedure is not systematic and some-

times has difficulties finding a particular solution. Another systematic method for

finding particular solutions of linear differential equations with constant coefficients

is the method of variation of parameters. That method is computationally much

more involved than the method of undetermined coefficients.

In the next two examples we indicate difficulties that one is faced with while

finding a particular solution using the method of undetermined coefficients.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–20
 Example 7.6: In this example, we demonstrate difficulties one is faced with in

the process of guessing a particular solution of a second-order differential equation.

Consider first a differential equation given by

Ñ
Ñ
Ñ

A natural choice will be to take for a candidate for a particular solution the following
Ñ
expression Ò (a linear combination of the forcing function

and its derivatives) and to try to determine unknown constant coefficients .

By taking the first and second derivatives of Ò and plugging them into the

differential equation, it can be checked easily that no exist such that the
Ñ
guessed solution is a particular solution of this differential equation.

We have to try a more complex form that also includes an integral of the forcing
Ñ Ó
function, that is Ò .

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–21
 This will produce the required solution as Ô ÖÕ ÖÕ × ØÕ Ù ,

with being an arbitrary constant. For simplicity, we may set so that

Ô × Ù

The particular solution obtained through the convolution formula is given by

Ú × Ù Û ×Ü
Ô

In the last formula we have used the upper script to indicate that the particular

solution is obtained through the convolution method. Note that the system impulse

response is Û ×Ü . It should be emphasized that the

particular solution obtained through the convolution formula by integrating a product

of the system input and the system impulse response carries information about both

the actual system input and the internal system dynamics represented in the system

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–22
impulse response by a weighted sum of Ý`Þàß terms (note the term â á ãä ß in the
particular solution obtained using the convolution formula, and its absence from the

particular solution obtained using the method of undetermined coefficients).

The previous example was not so difficult. However, if we change the original

differential equation slightly into

ä
ä ã ß
ä

then the problem of obtaining a particular solution by guessing its form becomes

much more difficult. The students may try several choices. However, we can obtain

the particular solution of this differential equation using the convolution formula in

a pretty straight forward manner (see Problem 7.2).

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–23
 Example 7.7: Consider a continuous-time linear system represented by the

differential equation defined in Example 7.6, that is

å
å
å

and the initial conditions given by æ çèé æ . The characteristic

values of this differential equation are å so that its homogeneous

è
solution is ê å æ å ë . Using the particular solution obtained in
è
Example 7.6 via the method of undetermined coefficients, the solution of this

differential equation has the form

ê ì å æ å›ë å í
è

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–24
 From the given initial conditions, we obtain a set of linear algebraic equations

for determining the constants î and ï

ð
î ï
ñ îò ð
ï

which leads to ï and î ï . The required solution,

obtained via the method of undetermined coefficients, is given by

ð ï›ó ï ô

If we use the particular solution obtained through the convolution, then the

constants î and ï have to be determined from the following set of linear

algebraic equations

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–25
 õ ö õ ÷ ø
ù ÷ú õ ù ÷ú õ ø
ö

ø ÷
which produces and . The solution obtained this way is identical

to the solution obtained using the method of undetermined coefficients since

öû û õ øfý ø þ õ øý
ü
ø þ õ øý

In conclusion, in the case when the linear time invariant system does not

differentiate the input signal, we can find the system response using any method

for determining a particular solution of the corresponding linear constant coefficient

differential equation. However, in the case when the system differentiates the input

signal the particular solution need to be found through the convolution procedure.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–26
7.4 Continuous-Time System Impulse Response

The continuous-time system impulse response is obtained using the impulse delta

function as the system input signal and assuming that the system initial conditions

are zero. It is important to observe that the impulse delta function is present only

at the initial time and that during that time it sets the system zero initial

conditions to nonzero values.

The linear continuous-time system impulse response is defined by

ÿ ÿ 
ÿ ÿ
 ÿ
  
  
       

with
 
ÿ 

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–27
 For  , we have the homogeneous differential equation (an input free
system)
  
 
 
  

whose initial conditions   have to be determined.

Finding the initial conditions   can be done

similarly like in the discrete-time domain by using either the direct method or

the method based on system linearity and time invariance. In the continuous-time

domain, in general, the direct method is pretty cumbersome so that we will limit

our presentation to the method based on the linearity and time invariance principles.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–28
 We define first the elementary continuous-time system impulse response by
  
  
   
 
    
        
  

Since the delta impulse function sets at some of the system initial con-

ditions to non zero values and then it disappears, the above system becomes the

homogeneous one for , that is

  
  
   
 

The initial conditions

!  "
for this system can be

obtained by using the following reasoning. The right-hand side of the equation for

at is equal to the delta impulse function so that the left-hand side

must be also equal to the impulse delta function.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–29
 Recalling the notion of generalized derivatives from Chapter 2, the above
$%
& $')(*+&
observation means that none of the terms # # # can have

a jump discontinuity at since the derivative of a function with a jump

discontinuity produces the impulse delta function. Hence, we must have for all of
$,& ( $,& -
them # # . However, in order to get
$'(% &
the delta impulse function on the left-hand side, the term # must contain

a jump discontinuity at . In addition that jump must be equal to one since

$'.&
after being differentiated in the # term, it produces the impulse delta function

multiplied by one on the left-hand side so that the equality between the left- and

right-hand sides holds. We conclude that the initial conditions for the elementary
-
system impulse response at are given by
- $%
& - $'(*
& - $'(% & -
# # # #

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–30
Using linearity and time invariance, the original impulse response becomes
/ 0 2
0 0
0132

Note that we have assumed that . The case will be discussed soon.

The obtained result is graphically presented in Figure 7.3.

b0 δ ( t ) Linear b0 h 0 ( t )
System

b1 δ ( t )
(1) (1)
Linear b1 h0 ( t )
System
...
.

bmδ ( t )
(m) (m)
Linear bm h0 ( t )
System

Σ bδ Σ bh
m m
(i) (i)
i=0
i ( t) Linear i=0
i 0 ( t )= h ( t )
System

Figure 7.3: Determination of the continuous-time linear system impulse response

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–31
 We need to justify a general claim that the linear time invariant system at rest

under the input signal equal to the th derivative of the system input produces on

its output the th derivative of the original system output. This can be justified as

follows. Let 4 be the solution of the elementary impulse response differential

equation. If we take the th derivative of the corresponding differential equation

we will obtain
5 6 5
78 6 6
5 6 4 5)78 578 6 4 8 6 4
6 6
4 6 4 6

Assuming that all initial conditions are zero, which is the case for the impulse
6 6
response, it follows that 4 is the solution (the system output) of the
6 6
above differential equation due to the forcing function given by .

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–32
 The above derivations are also valid when is any differentiable signal (in the

generalized sense) so that we have the following general result valid for continuous

time invariant linear system

9;:
9;:
< <
9;:
< <

= =
9;:
= =

Note that in the case when , the obtained formula for the impulse

response has to be slightly modified. Namely, integrating -times, we conclude

that in such a case at we have > . The modified formula for

becomes

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–33
 ?
)@ CED B
@ AB EC F D
?
G )@ CED B
@ AB EC F D

Example 7.19: Consider the problem of finding the system impulse response

solved in Example 4.20 by using the Laplace transform. The system is defined by
H
H

so that the system impulse response satisfies

H
H
I
I

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–34
 The impulse response is given by

J J

where the elementary system impulse response J is obtained from

K J J
K J
J L
J L

The solution for J is given by

J M NPO K N KO

Using the initial conditions at L set by the impulse delta function, we obtain
M K . Hence, the elementary system impulse response is given by

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–35
 Q RPS RT S

so that the system impulse response sought is

Q Q RUS RTVS RUS RTVS RUS RTVS

which is identical to the result obtained in Example 4.20.

Example 7.20: Consider now the case when . The impulse

response of the following system

T T
T T

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–36
satisfies the differential equation
W W
W W
X
X

It follows that
W Y Y
Y
W

Y
where the elementary system impulse response is obtained from
W Y Y
Y
W
Y Z
Y Z

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–37
 The solution for [ is given by

[ \ ]_^ ` ] `^

Using the initial conditions at a , we obtain \ ` . Hence, the

elementary system impulse response is given by

[ ]U^

so that the system impulse response is

`
[ [ ` [

]U^ ]U^ ]U^ ]U^ ]U^ ]U^

]U^ ]U^

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–38
7.5 Complete Continuous-Time System Response

In Chapter 1, we have initiated discussion about the relationship between the zero-

state and zero-input components of the system response and the particular and

homogeneous solutions of the corresponding differential equation. Now, we know

that the particular solution, obtained through the convolution procedure, is equal to

zero-state system response. In such a case, the homogeneous solution that satisfies

the system initial conditions is the zero-input system response.

Based on the discussion presented at the end of Section 7.1, we conclude that the

classical method for solving linear differential equations with constant coefficients

can be used for finding the complete response of linear time invariant systems even

in the case when the system differentiate the input signal. Even more, the same

method can be used for finding the zero-state and zero-input components of the

complete system response. To do so, we must perform the following steps.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–39
 Step 1: Find the impulse response in terms of the elementary system

impulse response.

Step 2: Find the particular solution using the convolution formula, that is
b . That particular solution is the system zero-state response,
b c;d .

Step 3: Find the general form of a homogeneous solution and determine its

constants using the following initial conditions

e f f
gh
i gjhki
e f f
b c;d
gl f hVi gml f h i
e f f

The homogeneous solution obtained represents the system zero-input response, that

is, e con .

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–40
 Example 7.21: The above steps are demonstrated on the second-order system

considered in Example 7.20, that is

p p
p p
qUr
s q tu
v q

The impulse response of this system (Step 1) has been already found in Example

7.20 as
qUs qUs , so that the particular solution (Step 2) obtained

using the convolution formula is given by

s
w
xzy
s
qU{ q|{ q|r}t~sq|{€v
xy

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–41
 ƒ
|‚
ƒ ‚+† ‡
† ‡ˆ†
„z…
|‚
ƒ ‡Vƒ ‡Vƒ ‡Vƒ

|‚
ƒ Uƒ Uƒ ‰‹Š

In Step 3, we find the homogeneous solution that satisfies the system initial

conditions. The homogeneous solution is given by

Œ  Uƒ ‡ Uƒ

Using the system initial conditions, we have

Œ   

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–42
and
Žjk Ž

‘ ’ ’  “ “

Hence
‘ ”o• ’_– ’U–

The complete system response is now given by

”;— ”o• ’U– ’U– ’|˜

–

Very often in linear system theory and practice, we use the notions of the system

transient response and the system steady state response. Sometimes, they are

identified, respectively, as the system zero-input and zero-state responses, which

is only partially true as explained below.

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–43
 The system transient response is determined by –the complete system

response during the short time interval immediately following the initial time, that

is ™›š œ for some small œ , see Figure 1.4. How small œ

should be depends on the system. The transient response is usually characterized

with fast signal changes mostly contributed by ož . The system steady state

response is defined theoretically as follows

ŸŸ
™¡ £¢

Practically, this means that the system response after some large period of time

is considered as the steady state system response, ŸVŸ œ .

Usually, the zero-input response decays to zero rapidly, hence the steady state

response is predominantly determined by the system zero-state response (the forced

response).

The slides contain the copyrighted material from Linear Dynamic Systems and Signals, Prentice Hall 2003. Prepared by Professor Zoran Gajic 7–44