EE553 Lect 2
EE553 Lect 2
Lecture 2
K. Hariche
IGEE/UMBB
November 5, 2020
1 Introduction
In this lecture we will consider typical mathematical modelling of nonlinear
systems and some basic de…nitions.
2 System models
Typically a nonlinear system is modelled mathematically either as an internal
( state space) model or as an external ( input/output or I/O ) model.
It may also be described by a characteristic curve representing output versus
input.
dX
The state space model for a nonlinear control system is of the form dt = f [X(t); u(t); t]; X(0) = X0
y(t) = g[X(t); u(t); t]
where X(t) is the state vector of dimension n; u(t) the control input of dimension
m and y(t) the output vector of dimension p:
The system may as well be modelled as an I/O model in the form y(t) =
G[u(t); t]
For a large class of nonlinear control systems , the dynamics are described
by an nth order nonlinear di¤erential equation relating the output y(t) to the
input u(t) as
dn y(t) n 1
dtn = F [y; dy
dt ; ; ddtn y
1 ; u(t); t]
This can be expressed in state space form by de…ning the states as:
x1 (t) = y(t)
x2 (t) = dy(t)
dt
..
. n 1
xn (t) = ddtn 1y
1
0 dx1 1 0 1
dt x2
B dx2 C B x3 C
dX(t) B dt C B C
leading to the state space equation dt =B .. C=B .. C
@ . A @ . A
dxn
F [x1 ; x2 ; ; xn ; u(t); t]
dt
dX
in the form dt = f [X(t); u(t); t]
In this course, we will be mainly concerned with state space models as they
describe the input/output relationship as well as the internal dynamics of the
system. In the analysis part, we are mainly concerned with the dynamic be-
haviour of the system under no control or under some speci…ed control. Hence,
in this part we might as well consider u(t) as a parameter.
If we de…ne fu (X; t) = f (X; u; t) and gu (X; t) = g(X; u; t) we obtain the
model used during the analysis part as:
dX dX
dt = fu (X; t) or simply as dt = f (X; t)
y(t) = gu (X; t) y(t) = g(X; t)
3 Basic de…nitions
dX
= f (X; t)
dt
De…nition 1 The system is said to be autonomous if the
y(t) = g(X; t)
functions f; g do not depend explicitely on time t; otherwise it is non-autonomous
dX
= f (X) dt
Clearly an autonomous system is described simply by :. The
y(t) = g(X)
dependence on t is implicit through the state X(t): Autonomy is the counterpart
of time-invariance for linear systems. It essentially means that that the choice
of time origin is irrelevant in the analysis od the system.
dx1
dt = x21 + tx2
Example 2 The system dx2 is clearly a nonautonomous non-
dt = x21 + x22
linear system of order 2
dx1 2
dt = 1 x1 + 2 x1 x2
Example 3 The system dx2 is an autonomous nonlin-
dt = 1 x1 + 2 x1 x2
ear system of order 2
2
Clearly if Xe is an equilibrium point of the system at time t0 ; then it is
an equilibrium point for all t t0 . Now if the system is autonomous Xe is
an equilibrium point for all t: Hence for an autonomous system : Xe is an
equilibrium point , f (Xe ) = 0 8t:
The expression f (Xe ) = 0 8t means that dXdt = 0 once the state reaches the
equilibrium point Xe . It follows that once an equilibrium reached, the system
will remain there forever under no external force.