 INPUT/OUTPUT STABILITY

• Boundedness

• Spaces

• Stability

• Small Gain Theorem- Basic Form

• Small Gain Theorem – Incremental Form

• Robustness

1
• Input/Output Representation

2
• Input/Output Representation

• Could be Linear/Nonlinear

3
• Input/Output Representation

• Could be Linear/Nonlinear

• Time Varying/Time Invariant

4
• Input/Output Representation

• Could be Linear/Nonlinear

• Time Varying/Time Invariant

• Static/Dynamic

5
• Input/Output Representation

• Could be Linear/Nonlinear

• Time Varying/Time Invariant

• Static/Dynamic

• Continuous Time/ Discrete-Time

6
• Input/Output Representation

• Could be Linear/Nonlinear

• Time Varying/Time Invariant

• Static/Dynamic

• Continuous Time/ Discrete-Time

7
• If input is “bounded”, does the output exist?

8
• If input is “bounded”, does the output exist?

• If input is “bounded”, does the output remain “bounded”?

9
• If input is “bounded”, does the output exist?

• If input is “bounded”, does the output remain “bounded”?

• Need “function spaces”

10
• If input is “bounded”, does the output exist?

• If input is “bounded”, does the output remain “bounded”?

• Need “function spaces”

• Need a measure of “boundedness” ( NORM)

11
• If input is “bounded”, does the output exist?

• If input is “bounded”, does the output remain “bounded”?

• Need “function spaces”

• Need a measure of “boundedness” ( NORM)

12
 • If input is “bounded”, does the output exist?

• If input is “bounded”, does the output remain “bounded”?

• Need “function spaces”

• Need a measure of “boundedness” ( NORM)

• Important Ones :

13
 • If input is “bounded”, does the output exist?

• If input is “bounded”, does the output remain “bounded”?

• Need “function spaces”

• Need a measure of “boundedness” ( NORM)

• Important Ones :

14
 • If input is “bounded”, does the output exist?

• If input is “bounded”, does the output remain “bounded”?

• Need “function spaces”

• Need a measure of “boundedness” ( NORM)

• Important Ones :

15
• Examples :

16
• Examples :

17
• Examples :

18
• Examples :

• FACT : spaces are linear vector spaces

19
• Examples :

• FACT : spaces are linear vector spaces

• NORM :

20
• FACT : with p-norm is a Banach Space,

21
• FACT : with p-norm is a Banach Space,

22
• FACT : with p-norm is a Banach Space,

• spaces are NOT SUFFICIENT, because unstable signals

are not included.

23
 • EXTENDED SPACES

• TRUNCATION FUNCTION :

24
 • EXTENDED SPACES

• TRUNCATION FUNCTION :

25
 • EXTENDED SPACES

• TRUNCATION FUNCTION :

26
• is a Linear Vector Space

27
• is a Linear Vector Space

• is a non decreasing function of T

28
• is a Linear Vector Space

• is a non decreasing function of T

29
• is a Linear Vector Space

• is a non decreasing function of T

• A mapping is called if :

30
• is a Linear Vector Space

• is a non decreasing function of T

• A mapping is called if :

• for some nonnegative k and b

31
• is a Linear Vector Space

• is a non decreasing function of T

• A mapping is called if :

• for some nonnegative k and b

• is called THE GAIN OF H

32
• is a Linear Vector Space

• is a non decreasing function of T

• A mapping is called if :

• for some nonnegative k and b

• is called THE GAIN OF H

• Bounded Input Bounded Output (BIBO) stability

33
• Example : LTI systems

STABLE

34
• Example : LTI systems

STABLE

35
• Example : LTI systems

STABLE

36
• Example : LTI systems

STABLE

PARSEVAL EQUALITY

37
• Example : LTI systems

STABLE

PARSEVAL EQUALITY

38
• Example : LTI systems

STABLE

PARSEVAL EQUALITY

39
40
• FACT :
• An LTI system is stable for one

• An LTI system is stable for all

• H(s) is a stable transfer function…

42
 • FACT :
• An LTI system is stable for one

• An LTI system is stable for all

• H(s) is a stable transfer function…

• Example : An LTI system which is not stable

43
•

•
?
•

• NOT STABLE

44
 •

•
?
•

• NOT STABLE

• STABLE…

45
• FEEDBACK SYSTEMS

46
 • FEEDBACK SYSTEMS

• , are possibly nonlinear operators (ANY)

47
 • FEEDBACK SYSTEMS

• , are possibly nonlinear operators (ANY)

48
 • FEEDBACK SYSTEMS

• , are possibly nonlinear operators (ANY)

49
 • FEEDBACK SYSTEMS

• , are possibly nonlinear operators (ANY)

50
 • FEEDBACK SYSTEMS

• , are possibly nonlinear operators (ANY)

51
• BASIC QUESTIONS :
• BASIC QUESTIONS :
• BASIC QUESTIONS :
• BASIC QUESTIONS :
• BASIC QUESTIONS :
 • BASIC QUESTIONS :

• SMALL GAIN THEOREM (BASIC FORM)

• ASSUMPTIONS :

57
 • BASIC QUESTIONS :

• SMALL GAIN THEOREM (BASIC FORM)

• ASSUMPTIONS :

58
 • BASIC QUESTIONS :

• SMALL GAIN THEOREM (BASIC FORM)

• ASSUMPTIONS :

• RESULT : IF

59
 • BASIC QUESTIONS :

• SMALL GAIN THEOREM (BASIC FORM)

• ASSUMPTIONS :

• RESULT : IF

60
61
62
63
64
65
66
67
68
69
70
71
• REMARKS :

• Theorem is quite general; applies to linear/nonlinear; time-varying/time

invariant; continuous, discrete time cases,…

72
• REMARKS :

• Theorem is quite general; applies to linear/nonlinear; time-varying/time

invariant; continuous, discrete time cases,…

• Theorem assumes that for , the closed-loop system HAS

SOLUTIONS and . EXISTENCE IS ASSUMED !

73
• REMARKS :

• Theorem is quite general; applies to linear/nonlinear; time-varying/time

invariant; continuous, discrete time cases,…

• Theorem assumes that for , the closed-loop system HAS

SOLUTIONS and . EXISTENCE IS ASSUMED !

• SMALL GAIN : is the GAIN OF , is THE GAIN OF .

is THE LOOP GAIN. : SMALL GAIN

74
• REMARKS :

• Theorem is quite general; applies to linear/nonlinear; time-varying/time

invariant; continuous, discrete time cases,…

• Theorem assumes that for , the closed-loop system HAS

SOLUTIONS and . EXISTENCE IS ASSUMED !

• SMALL GAIN : is the GAIN OF , is THE GAIN OF .

is THE LOOP GAIN. : SMALL GAIN

• If LTI, , are STABLE TRANSFER FUNCTIONS….

Related to Nyquist Criterion….

75
• SMALL GAIN THEOREM (INCREMENTAL FORM)

76
 • SMALL GAIN THEOREM (INCREMENTAL FORM)

77
 • SMALL GAIN THEOREM (INCREMENTAL FORM)

• are Lipschitz

78
 • SMALL GAIN THEOREM (INCREMENTAL FORM)

• are Lipschitz

79
80
81
• FIXED POINT PROBLEM !

• UNIQUE SOLUTION IF CONTRACTION

82
83
84
85
• AN APPLICATION : LTI SYSTEMS

86
• AN APPLICATION : LTI SYSTEMS
• AN APPLICATION : LTI SYSTEMS

88
• AN APPLICATION : LTI SYSTEMS

89
• AN APPLICATION : LTI SYSTEMS

90
• ADD A NONLINEAR PERTURBATION

91
• ADD A NONLINEAR PERTURBATION

• Assume stability when

92
• ADD A NONLINEAR PERTURBATION

• Assume stability when

93
• ADD A NONLINEAR PERTURBATION

• Assume stability when

94
• ADD A NONLINEAR PERTURBATION

• Assume stability when

95
96
97
98
99
100
101
• INPUT/OUTPUT STABILITY vs . LYAPUNOV STABILITY

• A : Stable, f : continuous

102
• INPUT/OUTPUT STABILITY vs . LYAPUNOV STABILITY

• A : Stable, f : continuous

103
• INPUT/OUTPUT STABILITY vs . LYAPUNOV STABILITY

• A : Stable, f : continuous

104
• INPUT/OUTPUT STABILITY vs . LYAPUNOV STABILITY

• A : Stable, f : continuous

105
• INPUT/OUTPUT STABILITY vs . LYAPUNOV STABILITY

• A : Stable, f : continuous

106
• If A is a stable matrix and f does not contain any linear term,
then by linearization x=0 is locally asymptotically stable.

• Could we deduce GLOBAL ASYMPTOTIC STABILITY ?

 • If A is a stable matrix and f does not contain any linear term,
then by linearization x=0 is locally asymptotically stable.

• Could we deduce GLOBAL ASYMPTOTIC STABILITY ?

• FACT : If is stable, then x=0 globally asymptotically stable for

 • If A is a stable matrix and f does not contain any linear term,
then by linearization x=0 is locally asymptotically stable.

• Could we deduce GLOBAL ASYMPTOTIC STABILITY ?

• FACT : If is stable, then x=0 globally asymptotically stable for

109
 • If A is a stable matrix and f does not contain any linear term,
then by linearization x=0 is locally asymptotically stable.

• Could we deduce GLOBAL ASYMPTOTIC STABILITY ?

• FACT : If is stable, then x=0 globally asymptotically stable for

110
 • If A is a stable matrix and f does not contain any linear term,
then by linearization x=0 is locally asymptotically stable.

• Could we deduce GLOBAL ASYMPTOTIC STABILITY ?

• FACT : If is stable, then x=0 globally asymptotically stable for

111
112
113
114
115
116
117
118
119
120
121
122
123
124