Chapter 4

Interconnection of LTI Systems

4.1 INTRODUCTION

Block diagrams and signal flow graphs are commonly used to describe a large feedback
control systems. Each block in the system is represented by a transfer function, which
indicates a proportional relationship between two Laplace-transformed signals known as
transmittance. The transmittance relates the incoming and outgoing signals as indicated
within each block of the system.

MATLAB® has numerous functions that can be used to interconnect two or more blocks to
form a larger feedback control system. With these function, you will be able to design a
large feedback control system without solving for the closed-loop transfer function.

4.2 THE TRANSFER MATRIX

The transfer matrix is a matrix whose entries are the transfer functions relating an input to
an output, and has the form

 G11 ( s ) G12 ( s ) " G1n ( s ) 

 
G ( s ) G22 ( s ) " G2 n ( s ) 
G ( s ) =  21 . (4.1)
 # # % # 
 
Gm1 ( s ) Gm 2 ( s ) " Gmn ( s ) 

The matrix equation realization of a MIMO system with m outputs and n inputs is

Y (s) = G (s) X (s)

 Y1 ( s )   G11 ( s ) G12 ( s ) " G1n ( s )   X 1 ( s ) 

     (4.2)
 Y2 ( s )  =  G21 ( s ) G22 ( s ) " G2 n ( s )   X 2 ( s ) 
 #   # # % #  # 
    
Ym ( s )  Gm1 ( s ) Gm 2 ( s ) " Gmn ( s )   X n ( s ) 

40
where, Y ( s ) = Y1 ( s ) Y2 ( s ) " Ym ( s )  and X ( s ) =  X 1 ( s ) X 2 ( s ) " X n ( s ) 
T T

are the output and input vectors, respectively. If the transfer matrix is square and diagonal,
then the subsystems in the matrix are said to be independent from each other, i.e., having
n simultaneous transfer functions independent from each other.

A diagonal transfer matrix can be generated in MATLAB® using the function

syst=append(sys1,sys2, " ,sysn), where sys1, sys2, " ,sysn are the
subsystems to be appended. Listing 4.1 shows an example of a script that generates a
3 × 3 transfer matrix of the system model

 s 
G11 ( s ) = s + 1 0 0 
 Y1 ( s )     X1 ( s ) 
   1   X ( s ) .
Y2 ( s )  =  0 G22 ( s ) = 2
s + 3s − 2
0
 2 
Y3 ( s )  
s + 2   3 ( )
X s 
 0 0 G33 ( s ) = 2
 s − 5s + 5 

Listing 4.1
>> sys1=tf([1 0],[1 1]);
>> sys2=tf(1,[1 3 -2]);
>> sys3=tf([1 2],[1 -5 5]);
>> syst=append(sys1,sys2,sys3)

Transfer function from input 1 to output...

s
#1: -----
s + 1

#2: 0

#3: 0
Transfer function from input 2 to output...
#1: 0

1
#2: -------------
s^2 + 3 s - 2

#3: 0

Transfer function from input 3 to output...

#1: 0

#2: 0

s + 2
#3: -------------
s^2 - 5 s + 5

41
For a single-input-multiple-output system, the function tf can also be used. The first step
is to define all numerator and denominator vectors for each subsystem. Then use tf to
generate the SIMO system defined by your subsystems. If the transfer matrix contains 3
subsystems sys1, sys2, and sys3, then the tf function can be used as follows

syst = tf({num1;num2;num3},{den1,den2,den3}).

Alternatively, the SIMO system can be generated by defining a column vector of

individual transfer functions.

syst = [sys1;sys2;sys3]

Listing 4.2 shows an example of a script that generates a single-input-three-output system

of the system model

 X1 ( s ) 
 s 1 s+2  
Y ( s ) = G1 ( s ) = G2 ( s ) = 2 G3 ( s ) = 2   X 2 ( s ) .
 s +1 s + 3s − 2 s − 5s + 5 
 X 3 ( s ) 

Listing 4.2
>> % Using the subsystems previously developed.
>> syst=[sys1;sys2;sys3]

Transfer function from input to output...

s
#1: -----
s + 1

1
#2: -------------
s^2 + 3 s - 2
s + 2
#3: -------------
s^2 - 5 s + 5

The function tf can also be used to generate a MIMO system. The first step is to define
all the subsystems and plug it into the tf function as to make a transfer matrix. Listing
4.3 shows an example of a script that generates a MIMO system of the model

 1 1 s   X1 ( s ) 
Y1 ( s )  G11 ( s ) = s + 1 G12 ( s ) = s 2 − 4 G13 ( s ) =
s+5  
 =   X 2 ( s ) .
Y2 ( s )   G ( s ) = 1 G22 ( s ) =
s +1
G23 ( s ) = 2
2s + 3   
 s + 3s + 4   X 3 ( s ) 
21
s s+2

42
Listing 4.3
>> sys11=(tf(1,[1 1]))

Transfer function:
1
-----
s + 1

>> sys12=(tf(1,[1 0 -4]))

Transfer function:
1
-------
s^2 - 4

>> sys13=(tf([1 0],[1 5]))

Transfer function:
s
-----
s + 5

>> sys21=(tf(1,[1 0]))

Transfer function:
1
-
s

>> sys22=(tf([1 1],[1 2]))

Transfer function:
s + 1
-----
s + 2

>> sys23=(tf([2 3],[1 3 4]))

Transfer function:
2 s + 3
-------------
s^2 + 3 s + 4

>> syst=[sys11 sys12 sys13; sys21 sys22 sys23]

Transfer function from input 1 to output...

1
#1: -----
s + 1

1
#2: -
s

43
 Transfer function from input 2 to output...
1
#1: -------
s^2 - 4

s + 1
#2: -----
s + 2

Transfer function from input 3 to output...

s
#1: -----
s + 5

2 s + 3
#2: -------------
s^2 + 3 s + 4

4.3 SERIES INTERCONNECTION OF SYSTEMS

sys1 sys2 " sysn

Fig. 4.1 Systems connected in series.

Fig. 4.1 shows an example of systems connected in series. If all the systems are single-
input-single-output, then the total system transfer function can be expressed as

T ( s ) = G1 ( s ) G2 ( s )" Gn ( s ) (4.3)

where, T(s) is the total transfer function, and G1 ( s ) , G2 ( s ) ," , Gn ( s ) are the individual
SISO transfer functions. However, if two MIMO systems are interconnected as shown in
Fig 4.2, then the multiplication of transfer functions is not anymore applicable.

U1 ( s )
Y1 ( s )
W1 ( s ) sys2
X1 ( s ) Y2 ( s )
sys1 U2 ( s)
X2 (s)
W2 ( s )
Fig. 4.2 Two systems connected in cascade.

44
The MATLAB® function syst=series(sys1,sys2,outputs1,inputs2) is
used to interconnect two systems in series of any configuration. sys1 and sys2 are the
two systems to be connected in series, outputs1 is a vector consisting of outputs of
sys1 that will be connected to the inputs of sys2 defined by the vector inputs2.
Listing 4.4 shows an example of a script that interconnects two systems with the
following matrix equations:

 1 1 
Y1,1 ( s )   G11,1 ( s ) = s 2 G12,1 ( s ) =
s + 4   X 1,1 ( s ) 
System 1:  =  
Y2,1 ( s )  G21,1 ( s ) = s   X 2,1 ( s ) 
0
 s2 + 3 
 1 
G11,2 ( s ) = s 2 − 3s + 2   X 1,2 ( s ) 
System 2: Y1,2 ( s ) =   
G ( s ) = s   X 2,2 ( s ) 
 21,2 s2 + 6 

The two systems are interconnected as shown in Fig. 4.3.

X 1,2 ( s )

Y1,1 ( s ) sys2 Y1,2 ( s )

X 1,1 ( s )
sys1 X 2,2 ( s )
X 2,1 ( s )
Y2,1 ( s )

Fig 4.3 An example of two systems connected in cascade.

Listing 4.4
>> sys111 = tf(1,[1 0 0])
Transfer function:
1
---
s^2

>> sys121 = tf(1,[1 4])

Transfer function:
1
-----
s + 4

>> sys211 = tf([1 0],[1 0 3])

45
Transfer function:
s
-------
s^2 + 3

>> syst1 = [sys111 sys121 ; sys211 0]

Transfer function from input 1 to output...

1
#1: ---
s^2

s
#2: -------
s^2 + 3
Transfer function from input 2 to output...
1
#1: -----
s + 4

#2: 0

>> sys112 = tf(1,[1 -3 2])

Transfer function:
1
-------------
s^2 - 3 s + 2

>> sys212 = tf([1 0],[1 0 6])

Transfer function:
s
-------
s^2 + 6

>> syst2 = [sys112 sys212]

Transfer function from input 1 to output:

1
-------------
s^2 - 3 s + 2

Transfer function from input 2 to output:

s
-------
s^2 + 6

>> systotal = series(syst1,syst2,1,2)

46
 Transfer function from input 1 to output:
s
-----------
s^4 + 6 s^2

Transfer function from input 2 to output:

s
----------------------
s^3 + 4 s^2 + 6 s + 24

Interconnecting the two systems with the configuration shown in Fig. 4.3 will result to
two transfer functions:

Y1,2 ( s )  1  1  1
T1 ( s ) = =  2  2 = 4 , and
X 1,1 ( s )  s   s + 6  s + 6s 2
Y (s)  1  1  s
T2 ( s ) = 1,2 =  2 = 3 2 .
X 2,1 ( s )  s + 4   s + 6  s + s − 10s + 8

Thus, the overall system model becomes

 1 
 s + 6s
4 2   X1 ( s ) 
Y (s) =   .
 s   X 2 ( s ) 
 s 3 + s 2 − 10s + 8 

4.4 PARALLEL INTERCONNECTION OF SYSTEMS

sys1

sys2

# #

sys3

Fig. 4.4 Systems connected in parallel.

47
Fig. 4.4 shows an example of systems connected in parallel. If all the systems are single-
input-single-output, then the total system transfer function can be expressed as

T ( s ) = G1 ( s ) + G2 ( s ) + " + Gn ( s ) (4.4)

where, T(s) is the total transfer function, and G1 ( s ) , G2 ( s ) ," , Gn ( s ) are the individual
SISO transfer functions. However, if two MIMO systems are interconnected as shown in
Fig 4.5, then the addition of transfer functions is not anymore applicable.

X1 ( s ) Y1 ( s )
sys1

U (s) W (s)

sys2
X2 (s) Y2 ( s )

Fig. 4.5 Two systems connected in parallel.

The MATLAB® function syst=parallel(sys1,sys2,in1,in2,out1,out2)

is used to interconnect two systems in parallel of any configuration. sys1 and sys2 are
the two systems to be connected in parallel, in1 and in2 are the inputs to be tied
together, and out1 and out2 are the outputs to be summed together. Listing 4.5 shows
an example of a script that interconnects two systems with the following matrix equations
given in the previous section. The two systems are interconnected as shown in Fig. 4.6.

X 1,1 ( s ) Y1,1 ( s )
U1 ( s ) W1 ( s )
X 2,1 ( s ) sys1 Y2,1 ( s )

U2 (s) W2 ( s )

X 1,2 ( s )
sys2
U3 ( s ) Y1,2 ( s )
X 2,2 ( s )

Fig. 4.6 An example of two systems connected in parallel.

48
Listing 4.5
>> %Use the previous subsystems generated in the previous
section.

>> sys111 = tf(1,[1 0 0])

Transfer function:
1
---
s^2

>> sys121 = tf(1,[1 4])

Transfer function:
1
-----
s + 4
>> sys211 = tf([1 0],[1 0 3])

Transfer function:
s
-------
s^2 + 3

>> syst1 = [sys111 sys121 ; sys211 0]

Transfer function from input 1 to output...

1
#1: ---
s^2

s
#2: -------
s^2 + 3

Transfer function from input 2 to output...

1
#1: -----
s + 4

#2: 0

>> sys112 = tf(1,[1 -3 2])

Transfer function:
1
-------------
s^2 - 3 s + 2

49
 >> sys212 = tf([1 0],[1 0 6])

Transfer function:
s
-------
s^2 + 6

>> syst2 = [sys112 sys212]

Transfer function from input 1 to output:

1
-------------
s^2 - 3 s + 2

Transfer function from input 2 to output:

s
-------
s^2 + 6

>> systotal = parallel(syst1,syst2,2,1,2,1)

Transfer function from input 1 to output...

1
#1: ---
s^2

s
#2: -------
s^2 + 3

Transfer function from input 2 to output...

1
#1: -----
s + 4
1
#2: -------------
s^2 - 3 s + 2

Transfer function from input 3 to output...

#1: 0

s
#2: -------
s^2 + 6

Interconnecting the two systems with the configuration shown in Fig. 4.6 will result to six
transfer functions since the overall system has three inputs and two outputs. The transfer
functions are:

W1 ( s ) 1
T1 ( s ) = =
U1 ( s ) s 2

50
 W1 ( s ) 1
T2 ( s ) = =
U2 ( s) s + 4
W1 ( s )
T3 ( s ) = =0
U3 ( s )
W2 ( s ) s
T4 ( s ) = = 2
U1 ( s ) s + 3
W2 ( s ) 1 1
T5 ( s ) = = 0+ 2 = 2
U2 (s) s − 3s + 2 s − 3s + 2
W2 ( s ) s
T6 ( s ) = = 2 .
U3 ( s ) s + 6

Thus, the overall system model becomes

0   X1 ( s ) 
 1 1 
 Y1 ( s )   s 2 s+4  
 =   X 2 ( s ) .
Y2 ( s )   s 1 s  
 s 2 + 3 s − 3s + 2
2
s + 6   X 3 ( s ) 
2

4.5 FEEDBACK INTERCONNECTION OF SYSTEMS

sys1

sys2

Fig. 4.7 Systems connected as a feedback pair.

Fig. 4.7 shows an example of two systems connected as a feedback pair. If all the systems
are single-input-single-output, then the total system transfer function can be expressed as

G1 ( s )
T (s) = (4.4)
1 + G1 ( s ) G2 ( s )

assuming negative feedback, where, T(s) is the total transfer function, and G1 ( s ) and
G2 ( s ) are the individual SISO transfer functions. However, if two MIMO systems are
interconnected as shown in Fig 4.8, then the feedback formula is not anymore applicable.

51
 X 1,1 ( s ) Y1,1 ( s )
U1 ( s ) W1 ( s )
X 2,1 ( s ) sys1 Y2,1 ( s )
U2 (s) W2 ( s )

sys2
Y1,2 ( s ) X 1,2 ( s )

Fig. 4.8 An example of two systems connected as a feedback pair.

The MATLAB® function

syst=feedback(sys1,sys2,feedin,feedout,sign)

is used to interconnect two systems in feedback of any configuration. sys1 and sys2
are the two systems to be connected in feedback, feedin is a vector containing indices
into the input vector of sys1 and specifies which inputs are involved in the feedback
loop, and feedout specifies which outputs of sys1 are used for feedback. If sign=1,
then the function uses negative feedback, otherewise (sign=0), positive feedback is
used. Listing 4.6 shows an example of a script that interconnects two systems with the
following matrix equations given in the previous section. The two systems are
interconnected as shown in Fig. 4.8 with the following system models:

 1 1 
Y1,1 ( s )   G ( s ) = G ( s ) =
s + 4   X 1,1 ( s ) 
11,1 2 12,1
s
System 1:  =  
Y2,1 ( s )  G ( s ) = s 0   X 2,1 ( s ) 
 21,1 s2 + 3 
 s +1 
System 2: Y (s) =   X (s)
 s+5

Listing 4.6
>> % Using sys1 in the previous example.

>> syst1

Transfer function from input 1 to output...

1
#1: ---
s^2

52
 s
#2: -------
s^2 + 3

Transfer function from input 2 to output...

1
#1: -----
s + 4

#2: 0

>> syst2=tf([1 1],[1 5])

Transfer function:
s + 1
-----
s + 5
>> feedback(syst1,syst2,2,2)

Transfer function from input 1 to output...

8 s^3 + 23 s^2 + 27 s + 60
#1: --------------------------------------
s^6 + 9 s^5 + 23 s^4 + 27 s^3 + 60 s^2

s
#2: -------
s^2 + 3

Transfer function from input 2 to output...

1
#1: -----
s + 4

#2: 0

The overall system resulted to a two-input-two output model

 8s 3 + 23s 2 + 27 s + 60 1 
 Y1 ( s )   s 6 + 9s 5 + 23s 4 + 27 s 3 + 60s 2 s + 4   X 1 ( s )  .
 =  
Y2 ( s )   s
0    X 2 ( s )
 s +3
2


53
4.5 EXERCISES

Given the system models

System 1
1 s
G11 ( s ) = G12 ( s ) =
s +1 s+2
s 1
G21 ( s ) = 2 G22 ( s ) =
s +3 s
 G ( s ) G12 ( s ) 
G ( s ) =  11 
G21 ( s ) G22 ( s ) 

System 2
1 1
H11 ( s ) = H12 ( s ) =
s+2 2
s +2
s 1
H 21 ( s ) = H 22 ( s ) = 2
s+2 s
3
H 31 ( s ) = 0 H 32 ( s ) = 2
s
 H11 ( s ) H12 ( s ) 
 
H ( s ) =  H 21 ( s ) H 22 ( s ) 
 H 31 ( s ) H 32 ( s ) 

System 3
1
W (s) =
s+5

+
sys3 Y1 ( s )
+

X1 ( s )
Y2 ( s )
X2 (s) sys1 + sys2
Y3 ( s )
−

sys3

Fig. 4.9 A complex feedback control system for No. 2.

54
1. Generate the three systems.

2. Connect the three systems as shown in Fig. 4.9 and determine all the transfer
functions in the overall system.

3. Determine the transfer matrix.

4. What is the difference between a transfer function and a transfer matrix?

5. If a MIMO system has 3 inputs and 4 outputs, how many combinations of transfer
functions can be derived?

55