Dynamical Systems 2010
Class 1
Department of Electrical EngineeringEindhoven University of Technology
Siep Weiland
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 1 / 42
Part I
Organization of the course
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 2 / 42
Outline of Part I
1 Organization of the courseMaterialScheduleExams and gradingsPurpose
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 3 / 42
Organization Material
website and course material
Website:
http://w3.ele.tue.nl/en/cs/education/courses/dynamical systems/
Material:
BookNonlinear Dynamics and Chaos:with applications to Physics, Biology, Chemistry andEngineering
Steven H. Strogatz,about EURO 50.
Slides and handoutsClass 1 (TUE) Dynamical Systems 2010 Siep Weiland 4 / 42
Organization Schedule
schedule course 2010
1 August 30: Organization; phase flows; stability
2 September 3: Existence and uniqueness; potential functions; Matlab
3 September 6: Bifurcations
4 September 13: Applications of bifurcations
5 September 17: Two dimensional flows
6 September 20: Stability, dissipativity and Lyapunov
7 September 27: Hamiltonian systems
8 October 1: Limit cycles
9 October 4: Poincare-Bendixson and oscillators
10 October 11: Discrete time evolutions
11 October 15: Chaos: strange attractors, sensitivity
12 October 18: Chaos: demonstrations and applications
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 5 / 42
Organization Exams and gradings
exams and gradings
ExercisesWe will make exercises in and outside class hoursMay need to bring your notebook (will be announced)Solutions posted on website.
ExamsYour choice of two projects (take home style)Details follow
GradingDetermined by grade of take home exam, possibly averaged withsome exercises that need to be handed in.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 6 / 42
Organization Purpose
purpose of this course
After this course you can
distinguish among particular classes of NL systems analyze stability, periodicity, chaotic behavior, bifurcations, hysteresis,
attractors, repellers, limit cycles
efficiently simulate NL evolution laws appreciate this research area
Topics:
existence and uniqueness of solutions of (nonlinear) differentialequations.
periodic, quasi-periodic and chaotic systems. fixed points, limit sets, periodic orbits, Lyapunov functions and
stability, Bendixson theorem, Poincare maps, bifurcations.
Chaotic attractors, Lyapunov exponents, iteration of functions andperiodic points.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 7 / 42
Organization Purpose
purpose of this course
After this course you can
distinguish among particular classes of NL systems analyze stability, periodicity, chaotic behavior, bifurcations, hysteresis,
attractors, repellers, limit cycles
efficiently simulate NL evolution laws appreciate this research area
Topics:
existence and uniqueness of solutions of (nonlinear) differentialequations.
periodic, quasi-periodic and chaotic systems. fixed points, limit sets, periodic orbits, Lyapunov functions and
stability, Bendixson theorem, Poincare maps, bifurcations.
Chaotic attractors, Lyapunov exponents, iteration of functions andperiodic points.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 7 / 42
Organization Purpose
purpose of this course
After this course you can
distinguish among particular classes of NL systems analyze stability, periodicity, chaotic behavior, bifurcations, hysteresis,
attractors, repellers, limit cycles
efficiently simulate NL evolution laws appreciate this research area
Topics:
existence and uniqueness of solutions of (nonlinear) differentialequations.
periodic, quasi-periodic and chaotic systems. fixed points, limit sets, periodic orbits, Lyapunov functions and
stability, Bendixson theorem, Poincare maps, bifurcations.
Chaotic attractors, Lyapunov exponents, iteration of functions andperiodic points.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 7 / 42
Organization Purpose
purpose of this course
After this course you can
distinguish among particular classes of NL systems analyze stability, periodicity, chaotic behavior, bifurcations, hysteresis,
attractors, repellers, limit cycles
efficiently simulate NL evolution laws appreciate this research area
Topics:
existence and uniqueness of solutions of (nonlinear) differentialequations.
periodic, quasi-periodic and chaotic systems. fixed points, limit sets, periodic orbits, Lyapunov functions and
stability, Bendixson theorem, Poincare maps, bifurcations.
Chaotic attractors, Lyapunov exponents, iteration of functions andperiodic points.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 7 / 42
Organization Purpose
purpose of this course
After this course you can
distinguish among particular classes of NL systems analyze stability, periodicity, chaotic behavior, bifurcations, hysteresis,
attractors, repellers, limit cycles
efficiently simulate NL evolution laws appreciate this research area
Topics:
existence and uniqueness of solutions of (nonlinear) differentialequations.
periodic, quasi-periodic and chaotic systems. fixed points, limit sets, periodic orbits, Lyapunov functions and
stability, Bendixson theorem, Poincare maps, bifurcations.
Chaotic attractors, Lyapunov exponents, iteration of functions andperiodic points.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 7 / 42
Part II
Todays lecture
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 8 / 42
Outline of Part II
2 Motivating examples
3 Linear systemsLinearization
4 Nonlinear systemsGeneral structureFixed points
5 Stability of fixed pointsStable fixed pointsUnstable fixed pointsVerifying stability of fixed points
6 Summary
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 9 / 42
Motivating examples
Outline
2 Motivating examples
3 Linear systemsLinearization
4 Nonlinear systemsGeneral structureFixed points
5 Stability of fixed pointsStable fixed pointsUnstable fixed pointsVerifying stability of fixed points
6 Summary
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 10 / 42
Motivating examples
Example 1: computational fluid dynamics
Control variables:
temperature velocity
Constraints
maximum temp. fuel constraints emission constraints temp. gradients
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 11 / 42
Motivating examples
Example 2: phase-lock loops (PLLs)
Classical PLL control loop configuration
VCOLFPD
1/N
- - -vovi
6
-
PD: phase detector LF: loop filter VCO: voltage controlled
oscillator
1/N: Divider
Used in many, many applications for
carrier synchronization carrier recovery frequency division and multiplication demodulation schemes.
Aim: lock frequency of output voltage vo to frequency of input voltage viClass 1 (TUE) Dynamical Systems 2010 Siep Weiland 12 / 42
Motivating examples
Example 3: Laser beams
Monochromatic, coherent anddirectional light produced viastimulated photon emission(1958)
Application of lasers in
CD/DVD players eye surgery optical communication welding, cutting, blasting concerts dental drills . . .
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 13 / 42
Motivating examples
Example 4: Meteorology
Atmospheric turbulence Tropical cyclone path prediction
Tropical storm Gustav path forecast, Thursday August 28, 2008
Turbulence and cyclone path predictions are difficult for good reasons
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 14 / 42
Motivating examples
Example 5: Unpredictable circuit behavior
A very simple electronic circuit
1 nonlinear diode characteristic
Its voltage behavior
2.5 2 1.5 1 0.5 0 0.5 1 1.5 2 2.50.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 15 / 42
Motivating examples
Example 6: Electrical power networks
Main trends
Liberalization of power market From monopolistic to competitive
market Increase of complexity
Increase of distributed and renewablepower generation
wind turbines, photovoltaic cells,. . . contribute to power generation but
not to stabilization changes of transmission structures
Aim:Stable operation of power net
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 16 / 42
Linear systems
Outline
2 Motivating examples
3 Linear systemsLinearization
4 Nonlinear systemsGeneral structureFixed points
5 Stability of fixed pointsStable fixed pointsUnstable fixed pointsVerifying stability of fixed points
6 Summary
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 17 / 42
Linear systems Linearization
linear systems
So far in your E curriculum:
All systems, physical components, models were assumed to haveidealized linear dynamics
You have seen different formats: State space models
x = Ax + Bu, y = Cx + Du
Transfer function models
Y (s) = H(s)U(s)
Models of differential equations
my + by + ky = u
What means linearity precisely and how realistic is this property
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 18 / 42
Linear systems Linearization
linear vs. nonlinear systems
Example: Pendulum dAAAAAh?
Fgravity
Model of pendulum of length L
+g
Lsin() = u
Linear ?? Nonlinear ??
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 19 / 42
Linear systems Linearization
linear vs. nonlinear systems
Definition
By a dynamical system we mean any collection B of functions w : TWdefined on a time set T R and producing values in a signal space W.A dynamical system is linear (over R) if this collection is a linear space:if w1,w2 B then also 1w1 + 2w2 B for any 1, 2 R.
Example: Pendulum modelLet (1, u1) and (2, u2) be two solutions and set = 1 + 2. Then
1 +gL sin(1) = u1
2 +gL sin(2) = u2
}6 + g
Lsin() = u1 + u2
So the pendulum model is not linear for this reason.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 20 / 42
Linear systems Linearization
linear vs. nonlinear systems
Definition
By a dynamical system we mean any collection B of functions w : TWdefined on a time set T R and producing values in a signal space W.A dynamical system is linear (over R) if this collection is a linear space:if w1,w2 B then also 1w1 + 2w2 B for any 1, 2 R.
Example: Pendulum modelLet (1, u1) and (2, u2) be two solutions and set = 1 + 2. Then
1 +gL sin(1) = u1
2 +gL sin(2) = u2
}6 + g
Lsin() = u1 + u2
So the pendulum model is not linear for this reason.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 20 / 42
Linear systems Linearization
linearization and state representations
With x1 = and x2 = this gives:
nonlinear model in state space form:(x1x2
)=
(x2
gL sin(x1) + u)
=
(f1(x1, x2, u)f2(x1, x2, u)
) linearized model in differential form:
Around = 0 gives sin() so that
+g
L = u
linearized model in state space form:(x1x2
)=
(0 10 gL
)
A
(x1x2
)+
(01
)B
u
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 21 / 42
Linear systems Linearization
linearization -definitions from earlier courses
Linear systems often obtained from linearization of nonlinear system
x = f (x , u), y = g(x , u)
Set
x(t) = x0 + (t), u(t) = u0 + (t), y(t) = y0 + (t)
with (x0, u0, y0) a linearization point and (, , ) a perturbation ofstate, input and output.
Taylor expansion of f and g around (x0, u0, y0) yields:
f (x , u) = f (x0, u0) +f
x(x0, u0)[x x0] + f
u(x0, u0)[u u0] + . . .
g(x , u) = g(x0, u0) +g
x(x0, u0)[x x0] + g
u(x0, u0)[u u0] + . . .
where . . . stands for higher order terms [x x0]2, [u u0]2 etc.Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 22 / 42
Linear systems Linearization
linearization -definitions from earlier courses
Linear systems often obtained from linearization of nonlinear system
x = f (x , u), y = g(x , u)
Set
x(t) = x0 + (t), u(t) = u0 + (t), y(t) = y0 + (t)
with (x0, u0, y0) a linearization point and (, , ) a perturbation ofstate, input and output.
Taylor expansion of f and g around (x0, u0, y0) yields:
f (x , u) = f (x0, u0) +f
x(x0, u0)[x x0] + f
u(x0, u0)[u u0] + . . .
g(x , u) = g(x0, u0) +g
x(x0, u0)[x x0] + g
u(x0, u0)[u u0] + . . .
where . . . stands for higher order terms [x x0]2, [u u0]2 etc.Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 22 / 42
Linear systems Linearization
linearization -definitions from earlier courses
Linear systems often obtained from linearization of nonlinear system
x = f (x , u), y = g(x , u)
Set
x(t) = x0 + (t), u(t) = u0 + (t), y(t) = y0 + (t)
with (x0, u0, y0) a linearization point and (, , ) a perturbation ofstate, input and output.
Taylor expansion of f and g around (x0, u0, y0) yields:
f (x , u) = f (x0, u0) +f
x(x0, u0)[x x0] + f
u(x0, u0)[u u0] + . . .
g(x , u) = g(x0, u0) +g
x(x0, u0)[x x0] + g
u(x0, u0)[u u0] + . . .
where . . . stands for higher order terms [x x0]2, [u u0]2 etc.Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 22 / 42
Linear systems Linearization
linearization -definitions
Assume that (x0, u0, y0) is a fixed point, that is
f (x0, u0) = 0 and y0 = g(x0, u0)
Since x = , = x x0, = u u0 and = y y0, we have
=f
x(x0, u0) +
f
u(x0, u0)+ . . .
=g
x(x0, u0) +
g
u(x0, u0)+ . . .
Ignoring the higher order terms yields a model of the form
= A + B, = C + D
where
A =f
x(x0, u0),B =
f
u(x0, u0),C =
g
x(x0, u0),D =
g
u(x0, u0)
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 23 / 42
Linear systems Linearization
linearization -definitions
Definition
The model = A + B, = C + D defined on the previous frame isthe linearization of the nonlinear model x = f (x , u), y = g(x , u) aroundthe fixed point (x0, u0, y0).
It represents an approximation of the dynamic behavior of thenonlinear model for small perturbations around the linearization point(x0, u0, y0). So, it has local validity.
Equivalently represented by its transfer function
H(s) = C (Is A)1B + D
its frequency response H(i), or its impulse responseh(t) = C exp(At)B + D(t), etc.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 24 / 42
Linear systems Linearization
linearization -definitions
Definition
The model = A + B, = C + D defined on the previous frame isthe linearization of the nonlinear model x = f (x , u), y = g(x , u) aroundthe fixed point (x0, u0, y0).
It represents an approximation of the dynamic behavior of thenonlinear model for small perturbations around the linearization point(x0, u0, y0). So, it has local validity.
Equivalently represented by its transfer function
H(s) = C (Is A)1B + D
its frequency response H(i), or its impulse responseh(t) = C exp(At)B + D(t), etc.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 24 / 42
Linear systems Linearization
linearization -example
Example
Linearize the nonlinear model
x = f (x , u) = x2 + u sin(x) + u2x2 1y = g(x , u) = x2 + sin(u) exp(x)
around point (x0, u0, y0) = (1, 0,1).
Solution:
Compute partial derivatives of f and g :
f
x= 2x + u cos(x) + 2xu2;
f
u= sin(x) + 2ux2
g
x= 2x + sin(u) exp(x); g
u= cos(u) exp(x)
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 25 / 42
Linear systems Linearization
linearization -example
Evaluate these at fixed point (x0, u0, y0) = (1, 0,1):f
x(1, 0) = 2;
f
u(1, 0) = sin(1);
g
x(1, 0) = 2; g
u(1, 0) = exp(1).
Hence,A = 2, B = sin(1), C = 2, D = exp(1)
yields the linearized model
= 2 + sin(1), = 2 + exp(1) Equivalently, the transfer function
H(s) = 2(s 2)1 sin(1) + exp(1)with pole in 2 and zero in 2 + 2 sin(1)/ exp(1).
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 26 / 42
Linear systems Linearization
linearization -pros and cons
Why are linearizations so important?
Linear models have local validity(only valid for small perturbations)
We can easily analyse linear models(freq. responses, stability, interconnections, robustness)
Allow many equivalent representations(state space, transfer functions, differential equations, convolutions)
Extremely suitable for control system designBut:
We ignore global dynamics We ignore phenomena beyond small perturbations
(periodicity, attractors, chaos, bifurcations)
Qualitatitive properties of nonlinear dynamics not (always) capturedin linearized models
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 27 / 42
Nonlinear systems
Outline
2 Motivating examples
3 Linear systemsLinearization
4 Nonlinear systemsGeneral structureFixed points
5 Stability of fixed pointsStable fixed pointsUnstable fixed pointsVerifying stability of fixed points
6 Summary
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 28 / 42
Nonlinear systems General structure
general structure of nonlinear systems
We consider the general form
x = f (x , u), y = g(x , u)
wherex(t) Rn, u(t) Rm, y(t) Rp
Important special cases:
homogeneous or autonomous system: no input u. one-dimensional flows: case n = 1.
Hence, state is one dimensional vector.
linear system: both f and g linear in x and u.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 29 / 42
Nonlinear systems General structure
some questions
x = f (x , u), y = g(x , u)
What do we mean by a solution ??
Do solutions x(t) exist for any input, init. condition and time ??
Can we compute them ??
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 30 / 42
Nonlinear systems General structure
some questions
x = f (x , u), y = g(x , u)
What do we mean by a solution ??
Do solutions x(t) exist for any input, init. condition and time ??
Can we compute them ??
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 30 / 42
Nonlinear systems General structure
some questions
x = f (x , u), y = g(x , u)
What do we mean by a solution ??
Do solutions x(t) exist for any input, init. condition and time ??
Can we compute them ??
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 30 / 42
Nonlinear systems General structure
example
Determine solutions of the autonomous system
x = sin(x)
Solution by separation of variables:
dt =dx
sin x
Integrate:
t + C = log |1 + cos xsin x
|
Hence, if x(0) = x0 then C = log |1+cos x0sin x0 | so that
t = log | [1 + cos(x0)] sin(x)sin(x0)[1 + cos(x)]
|
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 31 / 42
Nonlinear systems General structure
example
Determine solutions of the autonomous system
x = sin(x)
Solution by separation of variables:
dt =dx
sin x
Integrate:
t + C = log |1 + cos xsin x
|
Hence, if x(0) = x0 then C = log |1+cos x0sin x0 | so that
t = log | [1 + cos(x0)] sin(x)sin(x0)[1 + cos(x)]
|
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 31 / 42
Nonlinear systems General structure
example (ctd)
With x0 = pi/3 we can determine x(t) for t = 12 by solving
12 = log | [1 + cos(pi/3)] sin(x)sin(pi/3)[1 + cos(x)]
|
for x .This is no easy task!!
Questions:
is it solvable at all? if it is, is solution unique? what happens with x(t) as t ?
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 32 / 42
Nonlinear systems Fixed points
fixed points -definition
Definition
A point x is a fixed point of the homogeneous flow x = f (x) if f (x) = 0.
x is fixed point means that constant x(t) = x, t R is solution ofx = f (x) with initial condition x(0) = x.
Fixed points are also called equilibrium points, constant solutions,working points, steady solutions, stagnation points.
Example
x = sin(x) has x = kpi with k Z as its fixed points.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 33 / 42
Nonlinear systems Fixed points
fixed points -definition
Definition
A point x is a fixed point of the homogeneous flow x = f (x) if f (x) = 0.
x is fixed point means that constant x(t) = x, t R is solution ofx = f (x) with initial condition x(0) = x.
Fixed points are also called equilibrium points, constant solutions,working points, steady solutions, stagnation points.
Example
x = sin(x) has x = kpi with k Z as its fixed points.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 33 / 42
Stability of fixed points
Outline
2 Motivating examples
3 Linear systemsLinearization
4 Nonlinear systemsGeneral structureFixed points
5 Stability of fixed pointsStable fixed pointsUnstable fixed pointsVerifying stability of fixed points
6 Summary
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 34 / 42
Stability of fixed points Stable fixed points
stability of fixed points
Homogeneous flow x = sin(x) flows
to the right if sin(x) > 0 (velocity is positive) to the left if sin(x) < 0 (velocity is negative)
6 4 2 0 2 4 61.5
1
0.5
0
0.5
1
1.5
Fixed points at x = kpi, k odd: x is stable fixed point. k even: x is unstable fixed point.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 35 / 42
Stability of fixed points Stable fixed points
stability of fixed points
Definition
A fixed point x is stable if any solution x(t) stays near x for all t 0 ifthe initial condition x0 starts near enough to x
. Precisely, if for all > 0there exist > 0 such that for all initial condition x0 with |x0 x| < wehave that |x(t) x| for all t 0.
Important to note that
stability is a local property of a fixed point. may depend on , that is some initial conditions should be chosen
closer to x than others. The definition does not say that x(t) x as t . x is also called Lyapunov stable
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 36 / 42
Stability of fixed points Stable fixed points
stability of fixed points
Definition
A fixed point x is stable if any solution x(t) stays near x for all t 0 ifthe initial condition x0 starts near enough to x
. Precisely, if for all > 0there exist > 0 such that for all initial condition x0 with |x0 x| < wehave that |x(t) x| for all t 0.
Important to note that
stability is a local property of a fixed point. may depend on , that is some initial conditions should be chosen
closer to x than others. The definition does not say that x(t) x as t . x is also called Lyapunov stable
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 36 / 42
Stability of fixed points Unstable fixed points
instability of fixed points
Definition
A fixed point x is unstable if it is not stable.
Example
Consider x = x2 1. Fixed points are x = 1 and x = 1. Phasediagram tells us that x = 1 is unstable, x = 1 is stable.
Stable or unstable??
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 37 / 42
Stability of fixed points Unstable fixed points
solutions of x = sin(x)
Flow patterns x(t) for 0 t 10 of x = sin(x) for various initialconditions:
0 1 2 3 4 5 6 7 8 9 1010
8
6
4
2
0
2
4
6
8
10
t
solu
tion
x
solutions of xdot=sin(x)
Note stable and unstable fixed points on vertical axis!
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 38 / 42
Stability of fixed points Unstable fixed points
some refinements on stability definitions
Definition
A fixed point x is said to be attractive if there exist > 0 such that limt |x(t) x| = 0
whenever the initial condition x0 satisfies |x0 x| . asymptotically stable if it is both stable and attractive.
Remark: there exist examples of stable fixed points that are not attractive,and examples of attractive fixed points that are not stable.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 39 / 42
Stability of fixed points Verifying stability
how to verify stability?
Theorem
If f is differentiable at a fixed point x of the flow x = f (x), then x is stable if f (x) < 0. unstable if f (x) > 0.
So stability can be inferred from sign of f (x). No statement on casewhere f (x) = 0.For example, verify stability of fixed points of x = x3 or x = x3.
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 40 / 42
Summary
Outline
2 Motivating examples
3 Linear systemsLinearization
4 Nonlinear systemsGeneral structureFixed points
5 Stability of fixed pointsStable fixed pointsUnstable fixed pointsVerifying stability of fixed points
6 Summary
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 41 / 42
Summary
summary
Linear systems admit representations in state space, transfer function,differential equation and convolution format.
Nonlinear systems only allow representations in differential and statespace format. No transfer functions!!
Focused on autonomous (no inputs) and one-dimensional (state hasdimension 1) nonlinear systems.
We defined fixed points of nonlinear dynamical systems Phase diagrams are helpful to decide about stability of fixed points Introduced precise definitions of stability Can verify stability of fixed points through sign of f (x).
to next class
Class 1 (TUE) Dynamical Systems 2010 Siep Weiland 42 / 42
Organization of the courseOrganization of the courseMaterialScheduleExams and gradingsPurpose
Today's lectureMotivating examplesLinear systemsLinearization
Nonlinear systemsGeneral structureFixed points
Stability of fixed pointsStable fixed pointsUnstable fixed pointsVerifying stability of fixed points
Summary