Control theory

Kim Mathiassen

15.02.2011

Control theory
Mass spring damper system
Modeling
Open loop vs. closed loop
Second order system
Stability
PID control
P - Proportional
I - Integral
D - Derivative
Optimal control
LQR
15.02.2011

Mass spring damper system

From Wikimedia Commons

x = displacement [m]
f = force applied [kg m/s 2 ]
15.02.2011

m = mass of the block [kg ]

B = damping constant [kg /s]
k = spring constant [kg /s 2 ]

Mass spring damper system

Using Newtons second law

Spring force: f1 = kx

Damping force: f2 = f

External force: f3 = u

fi = ma. We have three forces

x
t

= f x

This gives the equation

m
x = kx f x + u

Differential equation for mass spring damper system

x +
15.02.2011

f
m x

k
mx

1
mu
4

Modeling domains
Frequency domain (Transfer functions)
x(s)=h(s)u(s)

h(s)=

1
m
f
k
s2+ m
s+ m

State space domain

x=Ax
+ Bu

15.02.2011

x 1 =x2
k
x 2 = m
x1

f
m x2

1
mu

Block diagrams
u
-

1
m

x 2

x2 = x 1

x1

15.02.2011

SISO and MIMO

Single-Input Single-Output (SISO)
The system has one input u and one output x
Multiple-Input Multiple-Output (MIMO)
The system has multiple input u and multiple
output x
Single-Input Multiple-Output (SIMO)
Can be regarded as several SISO systems
Multiple-Input Single-Output (MISO)
Can be regarded as several SISO systems

15.02.2011

Process

Process

Open loop vs. closed loop

Open-loop
r

Controller

Process

Closed-loop
r

Controller

Process

15.02.2011

Mesurements

Second order systems

H(s) =

s2 +

1
m
f
ms

k
m

1
m

(s 1 )(s 2 )

Solution
The generic solution gives three cases depending on pole
placemend. The three cases are called under-damped, over-damped
and critially damped
!
r
f
km
{1,2} =
1 14 2
(1)
2m
f
15.02.2011

Second order systems

Damping ratio
=

(1 +2 )

2 1 2

Over-damped, > 1 (1 and 2 real and distinct)

Slow system responce

Critically damped, = 1 (1 = 2 )
Fastes system responce without oscillations

Under-damped, < 1 (1 and 2 complex conjugates)

Fast system responce, but with oscillations
15.02.2011

10

Second order system responce

15.02.2011

11
From Wikimedia Commons

Stability
Consider the system y (s) = h(s)y0 (s) where y0 (s) has finite length
and amplitude

Asymptotically stable
The system is asymptotically stable if y 0 when t

Marginally stable
The system is marginally stable if |y | < for all t 0

Unstable
If the system is not stable, it is unstable
15.02.2011

12

PID control
We want to make the system stable and controllable with a
controller. The PID controller is a simple controller that may
acheive this goal. The PID controller is often analyzed in the
frequency domain.

PID controller
Z
u = Kp e + Ki

15.02.2011

e( )d + Kd e

13

Proportional
I

A pure proportional controller will have a steady-state error

Adding a integration term will remove the bias

High gain (Kp ) will produce a fast system

High gain may cause oscillations and may make the system
unstable

High gain reduces the steady-state error

15.02.2011

14

Proportional

15.02.2011

15
From Wikimedia Commons

Integral
I

Removes steady-state error

Increasing Ki accelerates the controller

High Ki may give oscillations

Increasing Ki will increase the settling time

15.02.2011

16

Integral

15.02.2011

17
From Wikimedia Commons

Derivative
I

Larger Kd decreases oscillations

Improves stability for low values of Kd

May be highly sensitive to noise if one takes the derivative of a

noisy error

High noise leads to instability

15.02.2011

18

Derivative

15.02.2011

19
From Wikimedia Commons

PIDstop

From http://www.pidstop.com/demo

PID games
http://www.pidstop.com/demo

15.02.2011

(K1 = -110 K2 = 0.728)

20

Optimal control
I

Optimal controll is another control approach than PID

The idea is to specify a cost function and then find the

optimal input

The Dynamics of the system is used to design the controller

For non-linear system it is not always possible to find the

optimal solution

A special case is for linear systems with a quadradic cost

function

The optimal controller must have all states as input

Most often used with an observer to estimate the states that

are not measured

15.02.2011

21

Optimal control

e Controller

Process

15.02.2011

Observer

Mesurements

22

Linear-quadratic regulator (LQR)

I

The feedback is given as u = G 1 x + G 2 r

r is the reference function

The matrix G 1 and G 2 is found based on the system dynamics

and the cost function using Pontryagins Maximum principle

When following a trajectory the function r (t) must be known

for all future timesteps in order to find the optimal solution

Cost function
J=
15.02.2011

1
2

e T Qe + u T Pudt

t
23

References
J. B. Balchen, T. Andresen, and B. A. Foss.
Reguleringsteknikk.
Institutt for teknisk kybernetikk, 2004.
PID controller.
http://en.wikipedia.org/wiki/pid_controller, February 2011.
Damping.
http://en.wikipedia.org/wiki/damping, February 2011.
O.A. Solheim and Norges tekniske hgskole Institutt for teknisk
kybernetikk.
Optimalregulering.
Tapir, 1976.
15.02.2011

24