Applied Nonlinear Control Nguyen Tan Tien - 2002.

4
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

4. Advanced Stability Theory

The objective of this chapter is to present stability analysis for

non-autonomous systems. Definition 4.2 The equilibrium point 0 is asymptotically stable
at t 0 if
4.1 Concepts of Stability for Non-Autonomous Systems
• it is stable
Equilibrium points and invariant sets • ∃r (t 0 ) > 0 such that x(t 0 ) < r (t 0 ) ⇒ x(t ) → 0 as
For non-autonomous systems, of the form t →∞

x& = f (x, t ) (4.1) The asymptotic stability requires that there exists an attractive
region for every initial time t 0 .
equilibrium points x* are defined by
Definition 4.3 The equilibrium point 0 is exponentially stable
f (x* , t ) ≡ 0 ∀t ≥ t 0 (4.2) if there exist two positive numbers, α and λ , such that for
sufficiently small x(t 0 ) ,
Note that this equation must be satisfied ∀t ≥ t 0 , implying that x(t ) ≤ α x 0 e −λ (t −t0 ) ∀t ≥ t 0
the system should be able to stay at the point x* all the time.
For instance, we can easily see that the linear time-varying
Definition 4.4 The equilibrium point 0 is global stable ∀x(t 0 ) ,
system
x(t ) → 0 as t → ∞
x& = A(t ) x (4.3)
Example 4.2 A first-order linear time-varying system_______
has a unique equilibrium point at the origin 0 unless A(t) is
always singular. Consider the first-order system x& (t ) = −a(t ) x(t ) . Its solution is
t

Example 4.1________________________________________ x(t ) = x(t 0 ) e

− ∫t0 a (r )dr . Thus system is stable if
The system ∞
a (t ) ≥ 0, ∀t ≥ t 0 . It is asymptotically stable if a (r )dr = +∞ .
∫ 0
a (t ) x It is exponentially stable if there exists a strictly positive
x& = − (4.4)
1+ x2 t +T
number T such that ∀t ≥ 0 ,
∫ a(r )dr ≥ γ , with γ
t
being a
has an equilibrium point at x = 0 . However, the system positive constant.
For instance
a (t ) x
x& = − + b(t ) (4.5) • The system x& = − x /(1 + t ) 2 is stable (but not
1+ x2 asymptotically stable)
• The system x& = − x /(1 + t ) is asymptotically stable
with b(t ) ≠ 0 , does not have an equilibrium point. It can be
• The system x& = −t x is exponentially stable
regarded as a system under external input or disturbance b(t ) .
x
Since Lyapunov theory is mainly developed for the stability of Another interesting example is the system x& (t ) = −
nonlinear systems with respect to initial conditions, such 1 + sin x 2
problem of forced motion analysis are more appropriately t x
treated by other methods, such as those in section 4.9.
− ∫t0 1+sin x2 (r ) dr
The solution can be expressed as x(t ) = x(t 0 ) e
__________________________________________________________________________________________
t x t − t0
Extensions of the previous stability concepts Since
∫ 1 + sin x (r ) dr ≥
t0 2 2
, the system is exponentially

Definition 4.1 The equilibrium point 0 is stable at t 0 if for convergent with rate 1 / 2 .
__________________________________________________________________________________________
any R > 0 , there exists a positive scalar r ( R, t 0 ) such that
Uniformity in stability concepts
x (t 0 ) < r ⇒ x(t ) < R ∀t ≥ t 0 (4.6) The previous concepts of Lyapunov stability and asymptotic
stability for non-autonomous systems both indicate the
importance effect of initial time. In practice, it is usually
Otherwise, the equilibrium point 0 is unstable. desirable for the systems to have a certain uniformity in its
behavior regardless of when the operation starts. This
The definition means that we can keep the state in ball of motivates us to consider the definitions of uniform stability
arbitrarily small radius R by starting the state trajectory in a and uniform asymptotic stability. Non-autonomous systems
ball of sufficiently small radius r . with uniform properties have some desirable ability to
___________________________________________________________________________________________________________
Chapter 4 Advanced Stability Theory 19
 Applied Nonlinear Control Nguyen Tan Tien - 2002.4
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

withstand disturbances. The behavior of autonomous systems Definition 4.7 A scalar time-varying function V (x, t ) is locally
is dependent of the initial time, all the stability properties of an positive definite if V (0, t ) = 0 and there exits a time-variant
autonomous system are uniform.
positive definite function V0 (x) such that
Definition 4.5 The equilibrium point 0 is locally uniformly
stable if the scalar r in Definition 4.1 can be chosen ∀t ≥ t 0 , V (x, t ) ≥ V0 (x) (4.7)
independent of t 0 , i.e., if r = r (R) .
Thus, a time-variant function is locally positive definite if it
Definition 4.6 The equilibrium point at the origin is locally dominates a time-variant locally positive definite function.
uniformly asymptotically stable if Globally positive definite functions can be defined similarly.
• it is uniformly stable
• there exits a ball of attraction B R0 , whose radius is Definition 4.8 A scalar time-varying function V (x, t ) is said to
be decrescent if V (0, t ) = 0 , and if there exits a time-variant
independent of t 0 , such that any trajectory with initial
positive definite function V1 (x) such that
states in B R0 converges to 0 uniformly in t 0 .
∀t ≥ 0, V (x, t ) ≥ V1 (x) (4.7)
By uniform convergence in terms of t 0 , we mean that for
all R1 and R2 satisfying 0 < R2 < R1 ≤ R0 , ∃T ( R1 , R2 ) > 0 such In other word, a scalar function V (x, t ) is decrescent if it is
dominated by a time-invariant p.d. function.
that, ∀t ≥ t 0
Example 4.4________________________________________
x(t 0 ) < R1 ⇒ x(t ) < R2 ∀t ≥ t 0 + T ( R1 , R2 )
Consider time-varying positive definite functions as follows
i.e., the trajectory, starting from within a ball B R1 , will
V (x, t ) = (1 + sin 2 t )( x12 + x22 )
converges into a smaller ball B R2 after a time period T which
V0 (x) = x12 + x22
is independent of t 0 .
V1 (x) = 2( x12 + x22 )
By definition, uniform asymptotic stability always implies
asymptotic stability. The converse (ñaû o ñeà ) is generally not ⇒ V (x, t ) dominates V0 (x) and is dominated by V1 (x)
true, as illustrated by the following example. because V0 (x) ≤ V (x, t ) ≤ V1 (x) .
__________________________________________________________________________________________
Example 4.3________________________________________
x Given a time-varying scalar function V (x, t ), its derivative
Consider the first-order system x& = − . This system has
1+ t along a system trajectory is
1 + t0
general solution x(t ) = x(t 0 ). The solution asymptotically d V ∂V ∂V ∂V ∂V
1+ t = + x& = + f ( x, t ) (4.8)
converges to zero. But the convergence is not uniform. dt ∂t ∂x ∂t ∂x
Intuitively, this is because a larger t 0 requires a longer time to
Lyapunov theorem for non-autonomous system stability
get close to the origin.
__________________________________________________________________________________________ The main Lyapunov stability results for non-autonomous
systems can be summarized by the following theorem.
The concept of globally uniformly asymptotic stability can be
defined be replacing the ball of attraction B R0 by the whole Theorem 4.1 (Lyapunov theorem for non-autonomous
systems)
state space.
Stability: If, in a ball B R0 around the equilibrium point 0,
4.2 Lyapunov Analysis of Non-Autonomous Systems there exits a scalar function V (x, t ) with continuous partial
derivatives such that
In this section, we extend the Lyapunov analysis results of
chapter 3 to the stability of non-autonomous systems. 1. V is positive definite
2. V& is negative semi-definite
4.2.1 Lyapunov’sdirect method for non-autonomous systems
then the equilibrium point 0 is stable in the sense of
The basic idea of the direct method, i.e., concluding the
Lyapunov.
stability of nonlinear systems using scalar Lyapunov functions,
can be similarly applied to non-autonomous systems. Besides
Uniform stability and uniform asymptotic stability: If,
more mathematical complexity, a major difference in non-
furthermore
autonomous systems is that the powerful La Salle’s theorems
do not apply. This drawback will partially be compensates by 3. V is decrescent
a simple result in section 4.5 called Barbalat’s lemma.
then the origin is uniformly stable. If the condition 2 is
Time-varying positive definite functions and decrescent strengthened by requiring that V& be negative definite, then
functions the equilibrium point is asymptotically stable.
___________________________________________________________________________________________________________
Chapter 4 Advanced Stability Theory 20
 Applied Nonlinear Control Nguyen Tan Tien - 2002.4
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Global uniform asymptotic stability: If, the ball B R0 is A simple result, however, is that the time-varying system
(4.17) is asymptotically stable if the eigenvalues of the
replaced by the whole state space, and condition 1, the
strengthened condition 2, condition 3, and the condition symmetric matrix A(t ) + A T (t ) (all of which are real) remain
strictly in the left-half complex plane
4. V (x, t ) is radially unbounded
∃λ > 0, ∀i, ∀t ≥ 0, λi ( A(t ) + A T (t )) ≤ −λ (4.19)
are all satisfied, then the equilibrium point at 0 is globally
uniformly asymptotically stable. This can be readily shown using the Lyapunov function
Similarly to the case of autonomous systems, if in a certain V = x T x , since
neighborhood of the equilibrium point, V is positive definite
and V& , its derivative along the system trajectories, is negative V& = x T x& + x& T x = x T ( A(t ) + A T (t )) ≤ −λ xT x& = −λ V
semi-definite, then V is called Lyapunov function for the non-
autonomous system. so that ∀t ≥ 0, 0 ≤ xT x = V (t ) ≤ V (0) e − λt and therefore x
tends to zero exponentially.
Example 4.5 Global asymptotic stability_________________ It is important to notice that the result provides a sufficient
Consider the system defined by
condition for any asymptotic stability.

x&1 (t ) = − x1 (t ) − e −2t x2 (t ) Perturbed linear systems

x& 2 (t ) = x1 (t ) − x2 (t ) Consider a linear time-varying system of the form

x& = [ A1 + A 2 (t )] x (4.20)
Chose the Lyapunov function candidate

where A1 is constant and Hurwitz and the time-varying matrix

V (x, t ) = x12 + (1 + e −2t ) x 22
A 2 (t ) is such that A 2 (t ) → 0 as t → ∞ and
This function is p.d., because it dominates the time-invariant
∞
p.d. function x12 + x 22 . It is also decrescent, because it is
∫ 0
A2 (t ) dt < ∞ (i.e., the integral exists and is finite)
dominated by the time-invariant p.d. function x12 + 2x 22 .
Furthermore, Then the system (4.20) is globally stable exponentially stable.

V& (x, t ) = −2 [ x12 − x1 x2 + x 22 (1 + e −2t )] Example 4.8________________________________________

Consider the system defined by
This shows that
x&1 = −(5 + x 25 + x38 ) x1
V& (x, t ) ≤ −2 ( x12 − x1 x2 + x 22 ) = −( x1 − x2 ) 2 − x12 − x 22
x& 2 = − x 2 + 4x32
x&3 = −(2 + sin t ) x3
Thus, V& (x, t ) is negative definite, and therefore, the point 0 is
globally asymptotically stable.
Since x3 tends to zero exponentially, so does x32 , and
4.2.2 Lyapunov analysis of linear time-varying systems therefore, so does x 2 . Applying the above result to the first
Consider linear time-varying systems of the form
equation, we conclude that the system is globally
exponentially stable.
x& = A(t ) x (4.17) __________________________________________________________________________________________

Since LTI systems are asymptotically stable if their Sufficient smoothness conditions on the A(t ) matrix
eigenvalues all have negative real parts ⇒ Will the system
(4.17) be stable if any time t ≥ 0 , the eigenvalues of A(t ) all Consider the linear system (4.17), and assume that at any time
have negative parts ? t ≥ 0 , the eigenvalues of A(t ) all have negative real parts

Consider the system ∃α > 0, ∀i, ∀t ≥ 0, λi [ A(t )] ≤ −α (4.21)

 x&1  − 1 e 2t   x1 
 x& 2  =  0 − 1   x 2  (4.18) If, in addition, the matrix A(t ) remains bounded, and
 
∞
Both eigenvalues of A(t ) equal to -1 at all times. The solution
−t t
∫ 0
A T (t ) A(t ) dt < ∞ (i.e., the integral exists and is finite)
of (4.18) can be rewritten as x2 = x 2 (0) e , x&1 + x1 = x 2 (0) e .
Hence, the system is unstable. then the system is globally exponentially stable.
___________________________________________________________________________________________________________
Chapter 4 Advanced Stability Theory 21
 Applied Nonlinear Control
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
4.2.3 The linearization method for non-autonomous systems
Lyapunov’s linearization method can also be developed for
non-autonomous systems. Let a non-autonomous system be
described by (4.1) and 0 be an equilibrium point. Assume that
f is continuously differentiable with respect to x. Let us denote
 ∂f 
A(t ) =  
 ∂ x  x =0
The for any fixed time t (i.e., regarding t as a parameter), a
Taylor expansion of f leads to
x& = A(t ) x + f h.o.t . (x, t )
If f can be well approximated by A(t ) x for any time t , i.e.,
f h.o.t . (x, t )
lim sup =0
x →0 x
then the system
x& = A(t ) x
is said to be the linearization (or linear approximation) of the
nonlinear non-autonomous system (4.1) around equilibrium
point 0.
⊗ Note that:
- The Jacobian matrix A thus obtained from a non-
autonomous nonlinear system is generally time-varying,
contrary to what happens for autonomous nonlinear
systems. But in some cases A is constant. For example,
the nonlinear system x& = − x + x 2 / t leads to the linearized
system x& = − x .
- Our late results require that the uniform convergence
condition (4.23) be satisfied. Some non-autonomous
systems may not satisfy this condition, and Lyapunov’s
linearization method cannot be used for such systems.
For example, (4.23) is not satisfied for the system
x& = − x + t x 2 .
Given a non-autonomous system satisfying condition (4.23),
we can assert its (local) stability if its linear approximation is
uniformly asymptotically stable, as stated in the following
theorem:
Therem 4.2 If the linearized system (with condition (4.23)
satisfied) is uniformly asymptotically stable, then the
equilibrium point 0 of the original non-autonomous system is
also uniformly asymptotically stable.
⊗ Note that:
- The linearized time-varying system must be uniformly
asymptotically stable in order to use this theorem. If the
linearized system is only asymptotically stable, no
conclusion can be draw about the stability of the original
nonlinear system.
- Unlike Lyapunov’s linearization method for autonomous
system, the above theorem does not relate the instability
of the linearized time-varying system to that of the
nonlinear system.
___________________________________________________________________________________________________________
Chapter 4 Advanced Stability Theory
Nguyen Tan Tien - 2002.4
Therem 4.3 If the Jacobian matrix A(t ) is constant,
A(t ) = A 0 , and if (4.23) is satisfied, then the instability of the
linearized system implies that of the original non-autonomous
nonlinear system, i.e., (4.1) is unstable if one or more of the
eigenvalues of A 0 has a positive real part .
(4.22) 4.3 Instability Theorems
4.4 Existence of Lyapunov Functions
4.5 Lyapunov-Like Analysis Using Barbalat’s Lemma
Asymptotic stability analysis of non-autonomous systems is
generally much harder than that of autonomous systems, since
it is usually very difficult to find Lyapunov functions with a
negative definite derivative. An important and simple result
which partially remedies (khaé c phuï c) this situation is
Barbalat’s lemma. When properly used for dynamic systems,
∀t ≥ 0 (4.23) particularly for non-autonomous systems, it may lead to the
satisfactory solution of many asymptotic stability problem.
4.5.1 Asymptotic properties of functions and their derivatives
(4.24) Before discussing Barbalat’s lemma itself, let us clarify a few
points concerning the asymptotic properties of functions and
their derivatives. Given a differentiable function f of time t ,
the following three facts are important to keep in mind
• f& → 0 ≠> f converges
The fact that f& → 0 does not imply that f (t ) has a limit
as t → ∞ .
• f converges ≠> f& → 0
The fact that f (t ) has a limit as t → ∞ does not imply
that f& → 0 .
• If f is lower bounded and decreasing ( f& ≤ 0) , then it
converges to a limit.
4.5.2 Barbalat’s lemma
Lemma 4.2 (Barbalat) If the differentiable function f (t ) has
a finite limit as t → ∞ , and if f& is uniformly continuous, then
f& (t ) → 0 as t → ∞ .
⊗ Note that:
- A function g (t ) is continuous on [0, ∞) if
∀t1 ≥ 0, ∀R > 0, ∃η ( R, t1 ) > 0, ∀t ≥ 0, t − t1 < η
⇒ g (t ) − g (t1 ) < R
- A function g (t ) is said to be uniformly continuous on
[0, ∞) if
∀R > 0, ∃η ( R ) > 0, ∀t1 ≥ 0, ∀t ≥ 0, t − t1 < η
⇒ g (t ) − g (t1 ) < R
or in other words, g (t ) is uniformly continuous if we can
always find an η which does not depend on the specific
point t1 - and in particular, such that η does not shrink
as t → ∞. t and t1 play a symmetric role in the definition
of uniform continuity.
22
 Applied Nonlinear Control Nguyen Tan Tien - 2002.4
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

- A simple sufficient condition for a differentiable function This implies that V (t ) ≤ V (0) , and therefore, that e and θ are
to be uniformly continuous is that its derivative be bound. bounded. But the invariant set cannot be used to conclude the
This can be seen from the finite different theorem: ∀t , t1 , convergence of e , because the dynamics is non-autonomous.
∃t 2 (t ≤ t 2 ≤ t1 ) such that g (t ) − g (t1 ) = g& (t 2 )(t − t1 ) . To use Barbalat’s lemma, let us check the uniform continuity
And therefore, if R1 > 0 is an upper bound on the of V& . The derivative of V& is V&& = −4e (−e + θ w) . This shows
function g& , we can always use η = R / R1 independently that V&& is bounded, since w is bounded by hypothesis, and e
of t1 to verify the definition of uniform continuity. and θ were shown above to be bounded. Hence, V& is
uniformly continuous. Application of Babarlat’s lemma then
Example 4.12_______________________________________ indicates that e → 0 as t → ∞ .
Note that, although e converges to zero, the system is not
Consider a strictly stable linear system whose input is bounded. asymptotically stable, because θ is only guaranteed to be
Then the system output is uniformly continuous. bounded.
__________________________________________________________________________________________

Indeed, write the system in the standard form

x& = A x + B u ⊗ Note that: Such above analysis based on Barbalat’s lemma
shall be called a Lyapunov-like analysis. There are two
y = Cx important differences with Lyapunov analysis:
Since u is bounded and the linear system is strictly stable, thus
the state x is bounded. This in turn implies from the first - The function V can simply be a lower bounded function
equation that x& is bounded, and therefore from the second of x and t instead of a positive definite function.
equation that y& = C x& is bounded. Thus the system output y is - The derivative V& must be shown to be uniformly
uniformly continuous. continuous, in addition to being negative or zero. This is
__________________________________________________________________________________________
typically done by proving that V&& is bounded.
Using Barbalat’s lemma for stability analysis
To apply Barbalat’s lemma to the analysis of dynamic systems, 4.6 Positive Linear Systems
one typically uses the following immediate corollary, which
looks very much like an invariant set theorem in Lyapunov In the analysis and design of nonlinear systems, it is often
analysis: possible and useful to decompose the system into a linear
subsystem and a nonlinear subsystem. If the transfer function
Lemma 4.3 (Lyapunov-Like Lemma) If a scalar function of the linear subsystem is so-called positive real, then it has
V (x, t ) satisfies the following conditions important properties which may lead to the generation of a
• V (x, t ) is lower bounded Lyapunov function for the whole system. In this section, we
study linear systems with positive real transfer function and
• V& (x, t ) is negative semi-definite their properties.
• V& (x, t ) is uniformly continuous in time
4.6.1 PR and SPR transfer function
then V& (x, t ) → 0 as t → ∞ . Consider rational transfer function of nth-order SISO linear
systems, represented in the form
Indeed, V the approaches a finite limiting value V∞ , such that
V∞ ≤ V (x(0),0) (this does not require uniform continuity). The bm p m + bm−1 p m−1 + K + b0
h( p ) =
above lemma then follows from Barbalat’s lemma. p n + a n−1 p n−1 + K + a0

Example 4.13_______________________________________ The coefficients of the numerator and denominator

polynomials are assumed to be real numbers and n ≥ m . The
Consider the closed-loop error dynamics of an adaptive
difference n − m between the order of the denominator and
control system for a first-order plant with unknown parameter
that of the numerator is called the relative degree of the system.
e& = −e + θ w(t )
Definition 4.10 A transfer function h(p) is positive real if
θ& = −e w(t )
Re[h( p )] ≥ 0 for all Re[ p ] ≥ 0 (4.33)
where e and θ are the two states of the closed-loop dynamics,
representing tracking error and parameter error, and w(t ) is a It is strictly positive real if h( p − ε ) is positive real for
bounded continuous function. some ε > 0
Consider the lower bounded function
Condition (4.33) is called the positive real condition, means
V = e2 +θ 2 that h( p ) always has a positive (or zero) real part when p has
positive (or zero) real part. Geometrically, it means that the
Its derivative is
rational function h( p ) maps every point in the closed RHP (i.e.,
V& = 2e [−e + θ w(t )] + 2θ [−e w(t )] = −2e 2 ≤ 0 including the imaginary axis) into the closed RHP of h( p) .

___________________________________________________________________________________________________________
Chapter 4 Advanced Stability Theory 23
 Applied Nonlinear Control Nguyen Tan Tien - 2002.4
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

Example 4.14 A strictly positive real function_____________ which shows that h4 is SPR (since it is also strictly stable). Of
course, condition (4.34) can also be checked directly on a
1 computer.
Consider the rational function h( p ) = , which is the
p+λ __________________________________________________________________________________________

transfer function of a first-order system, with λ > 0 . ⊗ The basic difference between PR and SPR transfer
Corresponding to the complex variable p = σ + jω , functions is that PR transfer functions may tolerate poles on
the jω axis, while SPR functions cannot.
1 σ + λ − jω
h( p ) = =
(σ + jω ) + λ (σ + λ ) 2 + ω 2 Example 4.16_______________________________________
1
Consider the transfer function of an integrator h( p ) = . Its
Obviously, Re[ h( p )] ≥ 0 if σ ≥ 0 . Thus, h( p ) is a positive real p
function. In fact, one can easily see that h( p ) is strictly σ − jω
value corresponding to p = σ + jω is h( p ) = . We
positive real, for example by choosing ε = λ / 2 in Definition σ 2 +ω 2
4.9. can easily see from Definition 4.9 that h( p ) is PR but not SPR.
__________________________________________________________________________________________
__________________________________________________________________________________________

Theorem 4.10 A transfer function h( p) is strictly positive real

Theorem 4.11 A transfer function h( p ) is positive real if, and
(SPR) if and only if
only if,
i. h( p) is a strictly stable transfer function
• h( p ) is a stable transfer function
ii. the real part of h( p ) is strictly positive along the jω axis,
• The poles of h( p ) on the jω axis are simple (i.e., distinct)
i.e.,
and the associated residues are real and non-negative
∀ω ≥ 0 Re[h( jω )] > 0 (4.34)
• Re[ h( jω )] ≥ 0 for any ω ≥ 0 such that jω is not a pole of
h( p )
The above theorem implies necessary conditions for asserting
whether a given transfer function h( p) is SPR:
The Kalman-Yakubovich lemma
• h( p ) is strictly stable If a transfer function of a system is SPR, there is an important
• The Nyquist plot of h( jω ) lies entirely in the RHP. mathematical property associated with its state-space
Equivalently, the phase shift of the system in response to representation, which is summarized in the celebrated
sinusoidal inputs is always less than 900 Kalman-Yakubovich (KY) lemma.
• h( p) has relative degree of 0 or 1
• h( p ) is strictly minimum-phase (i.e., all its zeros are in Lemma 4.4 (Kalman-Yakubovich) Consider a controllable
linear time-invariant system
the LHP)
x& = A x + b u
Example 4.15 SPR and non-SPR functions______________
y = cT x
Consider the following systems
The transfer function
p −1 p −1
h1 ( p ) = h2 ( p ) =
p 2 + ap + b p2 − p +1 h( p ) = cT [ pI − A]−1 b (4.35)
1 p +1
h3 ( p ) = h4 ( p ) = is strictly positive real if, and only if, there exist positive
p 2 + ap + b p2 + p +1
matrices P and Q such that

The transfer function h1 , h2 and h3 are not SPR, because h1 is

A T P + PA = -Q (4.36a)
non-minimum phase, h2 is unstable, and h3 has relative degree
Pb = c (4.36b)
larger than 1.
Is the (strictly stable, minimum-phase, and of relative degree
In the KY lemma, the involved system is required to be
1) function h4 actually SPR ? We have
asymptotically controllable. A modified version of the KY
lemma, relaxing the controllability condition, can be stated as
jω + 1 ( jω + 1)(−ω 2 − jω + 1) follows
h4 ( jω ) = 2
=
− ω + jω + 1 (1 − ω 2 ) 2 + ω 2
Lemma 4.5 (Meyer-Kalman-Yakubovich) Given a scalar
γ ≥ 0 , vector b and c , any asymptotically stable matrix A ,
(where the second equality is obtained by multiplying
numerator and denominator by the complex conjugate of the and a symmetric positive definite matrix L , if the transfer
denominator) and thus function
−ω 2 +1+ ω 2 1 γ
Re[h4 ( jω )] = = H (p ) = + c T [ pI − A]−1 b
(1 − ω 2 ) 2 + ω 2 (1 − ω 2 ) 2 + ω 2 2
___________________________________________________________________________________________________________
Chapter 4 Advanced Stability Theory 24
 Applied Nonlinear Control Nguyen Tan Tien - 2002.4
_________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________

is SPR, then there exist a scalar ε > 0 , a vector q , and a

symmetric positive definite matrix P such that

A T P + P A = -q q T − ε L
Pb = c + γ q

This lemma is different from Lemma 4.4 in two aspects.

• the involved system now has the output equation
γ
y = cT x +
u
2
• the system is only required to be stabilizable (but not
necessary controllable)

4.6.3 Positive real transfer matrices

The concept of positive real transfer function can be
generalized to rational positive real matrices. Such generation
is useful for the analysis and design of MIMO systems.

Definition 4.11 An m × m matrix H ( p ) is call PR if

• H ( p ) has elements which are analytic for Re( p ) > 0

• H ( p ) + H T ( p*) is positive semi-definite for Re( p ) > 0

where the asterisk * denote the complex conjugate transpose.

H ( p ) is SPR if H ( p − ε ) is PR for some ε > 0 .

4.7 The Passivity Formalism

4.8 Absolute Stability

4.9 Establishing Boundedness of Signal

4.10 Existence and Unicity of Solutions

___________________________________________________________________________________________________________
Chapter 4 Advanced Stability Theory 25