Control Techniques for Mechatronical Systems

MATLAB Experiment Guide

1 Simulink
2 Transfer function
2.1 The rational true fraction of the transfer function
𝑪𝑪(𝒔𝒔) 𝒃𝒃𝒎𝒎 𝒔𝒔𝒎𝒎 + 𝒃𝒃𝒎𝒎−𝟏𝟏 𝒔𝒔𝒎𝒎−𝟏𝟏 + ⋯ + 𝒃𝒃𝟏𝟏 𝒔𝒔 + 𝒃𝒃𝟎𝟎
𝑮𝑮(𝒔𝒔) = =
𝑹𝑹(𝒔𝒔) 𝒂𝒂𝒏𝒏 𝒔𝒔𝒏𝒏 + 𝒂𝒂𝒏𝒏−𝟏𝟏 𝒔𝒔𝒏𝒏−𝟏𝟏 + ⋯ + 𝒂𝒂𝟏𝟏 𝒔𝒔 + 𝒂𝒂𝟎𝟎
num1 = [2, 1];
den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1)
sys1 =

2 s + 1
---------------------
s^3 + 2 s^2 + 2 s + 1

Continuous-time transfer function.

2.2 Zero-pole form of the transfer function

(𝒔𝒔 − 𝒛𝒛𝟏𝟏 )(𝒔𝒔 − 𝒛𝒛𝟐𝟐 ) ⋯ (𝒔𝒔 − 𝒛𝒛𝒎𝒎 )
𝑮𝑮(𝒔𝒔) = 𝑲𝑲∗
(𝒔𝒔 − 𝒑𝒑𝟏𝟏 )(𝒔𝒔 − 𝒑𝒑𝟐𝟐 ) ⋯ (𝒔𝒔 − 𝒑𝒑𝒏𝒏 )
zeros = 0;
poles = [1-1i, 1+1i, 2];
k = 2;
sys2 = zpk(zeros, poles, k)
sys2 =

2 s
--------------------
(s-2) (s^2 - 2s + 2)

Continuous-time zero/pole/gain model.

2.3 Conversion between models

num1 = [2, 1];
den1 = [1, 2, 2, 1];
[Z1, P1, K1] = tf2zp(num1, den1)
Z1 =

-0.5000

P1 =

-1.0000 + 0.0000i
-0.5000 + 0.8660i
-0.5000 - 0.8660i

K1 =

zeros = 0;
poles = [1-1i, 1+1i, 2];
gain = 2;
[num2, den2] = zp2tf(zeros, poles, gain)
num2 =

0 0 2 0

den2 =

1 -4 6 -4
3 Time domain analysis
3.1 Typical input signal
num1 = [2, 1];
den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1);
%impulse(sys1);
impulse(num1, den1);

num1 = [2, 1];

den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1);
%step(sys1);
step(num1, den1);
3.2 Stability analysis
num1 = [2, 1];
den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1);
pzmap(sys1);
grid on;

num1 = [2, 1];

den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1);
pole(sys1)
ans =

-1.0000 + 0.0000i
-0.5000 + 0.8660i
-0.5000 - 0.8660i

zero(sys1)
ans =

-0.5000
roots(den1)
ans =

-1.0000 + 0.0000i
-0.5000 + 0.8660i
-0.5000 - 0.8660i

3.3 Transient performance analysis

num1 = [2, 1];
den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1);
%step(sys1);
step(num1, den1);
num1 = [2, 1];
den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1);
[y, t] = step(sys1);
plot(t,y);
[Y, k] = max(y);
tp = t(k)
tp =

2.8552

C = dcgain(sys1);
[Y, k] = max(y);
percent_overshoot = 100 * (Y - C) / C
percent_overshoot =

43.3919

% tr: 0 - 100%
n = 1;
while y(n) < C
n = n + 1;
end
tr = t(n)
tr =

1.5658

% tr: 10% - 90%

n = 1;
while y(n) < 0.1 * C
n = n + 1;
end
m = 1;
while y(m) < 0.9 * C
m = m + 1;
end
tr = t(m) - t(n)
tr =

1.1052

i = length(t);
while (y(i) > 0.95 * C) & (y(i) < 1.05 * C)
i = i - 1;
end
ts = t(i)
ts =

7.2762

3.4 Steady-state error analysis

num1 = [2, 1];
den1 = [1, 2, 2, 1];
sys1 = tf(num1, den1);
C = dcgain(sys1);
Ess = 1 - C
Ess =

0
4 Root Locus Techniques
4.1 Common functions for drawing root locus
rlocus(): Evans root locus.

rlocfind(): Find root locus gains for a given set of roots.

Example 4-1:

zeros = [];
poles = [0, -2];
gain = 1;
GH = zpk(zeros, poles, gain);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);

Example 4-2:
𝑲𝑲∗
𝑮𝑮(𝒔𝒔)𝑯𝑯(𝒔𝒔) =
𝒔𝒔(𝒔𝒔 + 𝟏𝟏)(𝒔𝒔 + 𝟐𝟐)
zeros = [];
poles = [0, -1, -2];
gain = 1;
GH = zpk(zeros, poles, gain);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);

Example 4-3:
𝑲𝑲∗ (𝒔𝒔 + 𝟒𝟒)
𝑮𝑮(𝒔𝒔)𝑯𝑯(𝒔𝒔) =
𝒔𝒔(𝒔𝒔 + 𝟐𝟐)
zeros = [-4];
poles = [0, -2];
gain = 1;
GH = zpk(zeros, poles, gain);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);
Example 4-4:
𝑲𝑲∗ (𝒔𝒔 + 𝟏𝟏)
𝑮𝑮(𝒔𝒔)𝑯𝑯(𝒔𝒔) =
(𝒔𝒔 + 𝟏𝟏. 𝟓𝟓 + 𝒋𝒋)(𝒔𝒔 + 𝟏𝟏. 𝟓𝟓 − 𝒋𝒋)
zeros = [-1];
poles = [-1.5-1i, -1.5+1i];
gain = 1;
GH = zpk(zeros, poles, gain);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);
Example 4-5:
𝑲𝑲∗ (𝒔𝒔 + 𝟏𝟏. 𝟓𝟓)(𝒔𝒔 + 𝟐𝟐 − 𝒋𝒋)(𝒔𝒔 + 𝟐𝟐 + 𝒋𝒋)
𝑮𝑮(𝒔𝒔)𝑯𝑯(𝒔𝒔) =
𝒔𝒔(𝒔𝒔 + 𝟐𝟐. 𝟓𝟓)(𝒔𝒔 + 𝟎𝟎. 𝟓𝟓 − 𝟏𝟏. 𝟓𝟓𝟓𝟓)(𝒔𝒔 + 𝟎𝟎. 𝟓𝟓 + 𝟏𝟏. 𝟓𝟓𝟓𝟓)
zeros = [-1.5, -2+2i, -2-2i];
poles = [0, -2.5, -0.5+1.5i, -0.5-1.5i];
gain = 1;
GH = zpk(zeros, poles, gain);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis([-4, 0.5, -3.5, 3.5]);
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);

Example 4-6:
𝑲𝑲∗
𝑮𝑮𝟎𝟎 (𝒔𝒔) =
𝒔𝒔(𝒔𝒔 + 𝟏𝟏)(𝒔𝒔 + 𝟑𝟑. 𝟓𝟓)(𝒔𝒔 + 𝟑𝟑 − 𝒋𝒋𝒋𝒋)(𝒔𝒔 + 𝟑𝟑 + 𝒋𝒋𝒋𝒋)
zeros = [];
poles = [0, -1, -3.5, -3+2i, -3-2i];
gain = 1;
GH = zpk(zeros, poles, gain);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis([-10, 5, -5, 5]);
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);

Example 4-7:
𝑲𝑲∗ (𝒔𝒔 + 𝟏𝟏)
𝑮𝑮(𝒔𝒔) =
𝒔𝒔(𝒔𝒔 − 𝟏𝟏)(𝒔𝒔𝟐𝟐 + 𝟒𝟒𝟒𝟒 + 𝟏𝟏𝟏𝟏)
num = [1, 1];
den = conv([1, 0], conv([1 -1], [1, 4, 16]));
GH = tf(num, den);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis([-4.5, 3.5, -4.5, 4.5]);
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);
Example 4-8:
𝑲𝑲∗ 𝑲𝑲∗
𝑮𝑮(𝒔𝒔)𝑯𝑯(𝒔𝒔) = =
𝒔𝒔(𝒔𝒔𝟐𝟐 + 𝟔𝟔𝟔𝟔 + 𝟏𝟏𝟏𝟏) 𝒔𝒔(𝒔𝒔 + 𝟑𝟑 − 𝒋𝒋)(𝒔𝒔 + 𝟑𝟑 + 𝒋𝒋)
zeros = [];
poles = [0, -3+1i, -3-1i];
gain = 1;
GH = zpk(zeros, poles, gain);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis([-6, 1, -4, 4]);
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);
Example 4-9:
𝑲𝑲∗
𝑮𝑮(𝒔𝒔)𝑯𝑯(𝒔𝒔) =
𝒔𝒔(𝒔𝒔𝟐𝟐 + 𝟐𝟐𝟐𝟐 + 𝟐𝟐)(𝒔𝒔𝟐𝟐 + 𝟔𝟔𝟔𝟔 + 𝟏𝟏𝟏𝟏)
num = [1];
d1 = [1, 2, 2];
d2 = [1, 6, 13];
den1 = conv(d1, d2);
den = [den1, 0]; % den = [1, 8, 27, 38, 26, 0]
GH = tf(num,den);

rlocus(GH);
sgrid();

h = findobj(gcf,'type','line');
axis([-6, 2, -5, 5]);
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);

[k, p] = rlocfind(GH);

k =

29.5940

p =

-2.8163 + 2.1726i
-2.8163 - 2.1726i
-2.3581 + 0.0000i
-0.0046 + 0.9959i
 -0.0046 - 0.9959i

4.2 Root locus analysis of step response

Example 4-10:
𝑲𝑲
𝑮𝑮(𝒔𝒔) =
𝒔𝒔(𝒔𝒔 + 𝟏𝟏)(𝒔𝒔 + 𝟐𝟐)

Find the system open-loop gain corresponding to the damping ratio 0.707, draw the step
response of a unit feedback closed loop system.
zeros = [];
poles = [0, -1, -2];
gain = 1;
G = zpk(zeros, poles, gain);

rlocus(G);
zeta = [0:0.707:1];
wn = [1:10];
sgrid(zeta, wn);

h = findobj(gcf,'type','line');
axis([-6, 2, -5, 5]);
axis equal;
set(h, 'LineWidth', 2);
set(h, 'markersize', 15);

[k,p] = rlocfind(G);

k =

0.6697
p =

-2.2409 + 0.0000i
-0.3796 + 0.3934i
-0.3796 - 0.3934i

k = 0.6697;
G = zpk(zeros, poles, k);
Gf = feedback(G, 1);
figure();
step(Gf);

h = findobj(gcf,'type','line');
set(h, 'LineWidth', 2);

5 Frequency Response Techniques

5.1 Nyquist Diagram
Example 5-1:
𝟐𝟐𝟐𝟐 + 𝟔𝟔
𝑮𝑮(𝒔𝒔) =
𝒔𝒔𝟑𝟑 + 𝟐𝟐𝒔𝒔𝟐𝟐 + 𝟓𝟓𝟓𝟓 + 𝟐𝟐
Draw the Nyquist diagram, determine system stability.
close all;
clear all;
num = [2, 6];
den = [1, 2, 5, 2];
nyquist(num,den);

% num = [2, 6];

% den = [1, 2, 5, 2];
% [re, im] = nyquist(num,den);
% plot(re, im);
% grid on;
%
% [re, im] = nyquist(num,den,0.5);

h = findobj(gcf,'type','line');
set(h, 'LineWidth', 2);

Example 5-2:

𝝎𝝎𝟐𝟐𝒏𝒏
𝑮𝑮(𝒔𝒔) =
𝒔𝒔𝟐𝟐 + 𝟐𝟐𝟐𝟐𝝎𝝎𝒏𝒏 𝒔𝒔 + 𝝎𝝎𝟐𝟐𝒏𝒏
Sketching the Nyquist diagram.
close all;
clear all;
zetas = [0.3, 0.5, 0.7, 0.9];
wn = 1;
for i = 1:4
num = wn * wn;
den = [1, 2*zetas(i), wn * wn];
nyquist(num,den);
axis('square');
hold on;
end
h = findobj(gcf,'type','line');
set(h, 'LineWidth', 2);

Example 5-3:
𝟐𝟐𝟐𝟐 + 𝟏𝟏
𝑮𝑮(𝒔𝒔) =
𝒔𝒔𝟐𝟐 (𝟎𝟎. 𝟓𝟓𝟓𝟓 + 𝟏𝟏)(𝒔𝒔 + 𝟏𝟏)
Sketching the Nyquist diagram.
num = [2, 1];
den = conv([1, 0, 0], conv([0.5, 1], [1, 1]));
nyquist(num, den, {0.2,10000});

h = findobj(gcf,'type','line');
set(h, 'LineWidth', 2);
5.2 Bode Plots
Example 5-4:
𝟑𝟑𝟑𝟑(𝟎𝟎. 𝟐𝟐𝟐𝟐 + 𝟏𝟏)
𝑮𝑮(𝒔𝒔) =
𝒔𝒔(𝒔𝒔𝟐𝟐+ 𝟏𝟏𝟏𝟏𝟏𝟏 + 𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏)
Draw the Bode plot.
num = [6, 30];
den = conv([1, 0], [1, 16, 100]);
% bode(num, den);

bode(num, den, {0.1,100});

% [mag, phase, w] = bode(num, den, {0.1,100});

% semilogx(w, 20*log10(mag));

grid on;
h = findobj(gcf,'type','line');
set(h, 'LineWidth', 2);
 Bode Diagram
20

-20
Magnitude (dB)

-40

-60

-80
-60

-90
Phase (deg)

-120

-150

-180
-1 1 2
10 10 0 10 10
Frequency (rad/s)

5.3 Gain Margin and Phase Margin

Example 5-5:
𝟏𝟏𝟏𝟏
𝑮𝑮(𝒔𝒔) =
𝒔𝒔𝟑𝟑 + 𝟑𝟑𝒔𝒔𝟐𝟐 + 𝟗𝟗𝟗𝟗
Find the magnitude/gain margin and phase margin.
close all;
clear all;
num = [10];
den = [1, 3, 9, 0];
margin(num,den);
grid on;

[Gm, Pm, Wcg, Wcp] = margin(num,den);