Inner/outer control design for vertical takeoff and landing aircraft¶
This script demonstrates the use of the python-control package for analysis and design of a controller for a vectored thrust aircraft model that is used as a running example through the text Feedback Systems by Astrom and Murray. This example makes use of MATLAB compatible commands.
Code¶
1# pvtol-nested.py - inner/outer design for vectored thrust aircraft
2# RMM, 5 Sep 09
3#
4# This file works through a fairly complicated control design and
5# analysis, corresponding to the planar vertical takeoff and landing
6# (PVTOL) aircraft in Astrom and Murray, Chapter 11. It is intended
7# to demonstrate the basic functionality of the python-control
8# package.
9#
10
11import os
12import matplotlib.pyplot as plt # MATLAB-like plotting functions
13import control as ct
14import numpy as np
15
16# System parameters
17m = 4 # mass of aircraft
18J = 0.0475 # inertia around pitch axis
19r = 0.25 # distance to center of force
20g = 9.8 # gravitational constant
21c = 0.05 # damping factor (estimated)
22
23# Transfer functions for dynamics
24Pi = ct.tf([r], [J, 0, 0]) # inner loop (roll)
25Po = ct.tf([1], [m, c, 0]) # outer loop (position)
26
27#
28# Inner loop control design
29#
30# This is the controller for the pitch dynamics. Goal is to have
31# fast response for the pitch dynamics so that we can use this as a
32# control for the lateral dynamics
33#
34
35# Design a simple lead controller for the system
36k, a, b = 200, 2, 50
37Ci = k * ct.tf([1, a], [1, b]) # lead compensator
38Li = Pi * Ci
39
40# Bode plot for the open loop process
41plt.figure(1)
42ct.bode_plot(Pi)
43
44# Bode plot for the loop transfer function, with margins
45plt.figure(2)
46ct.bode_plot(Li)
47
48# Compute out the gain and phase margins
49gm, pm, wcg, wcp = ct.margin(Li)
50
51# Compute the sensitivity and complementary sensitivity functions
52Si = ct.feedback(1, Li)
53Ti = Li * Si
54
55# Check to make sure that the specification is met
56plt.figure(3)
57ct.gangof4(Pi, Ci)
58
59# Compute out the actual transfer function from u1 to v1 (see L8.2 notes)
60# Hi = Ci*(1-m*g*Pi)/(1+Ci*Pi)
61Hi = ct.parallel(ct.feedback(Ci, Pi), -m * g *ct.feedback(Ci * Pi, 1))
62
63plt.figure(4)
64plt.clf()
65plt.subplot(221)
66ct.bode_plot(Hi)
67
68# Now design the lateral control system
69a, b, K = 0.02, 5, 2
70Co = -K * ct.tf([1, 0.3], [1, 10]) # another lead compensator
71Lo = -m*g*Po*Co
72
73plt.figure(5)
74ct.bode_plot(Lo) # margin(Lo)
75
76# Finally compute the real outer-loop loop gain + responses
77L = Co * Hi * Po
78S = ct.feedback(1, L)
79T = ct.feedback(L, 1)
80
81# Compute stability margins
82gm, pm, wgc, wpc = ct.margin(L)
83print("Gain margin: %g at %g" % (gm, wgc))
84print("Phase margin: %g at %g" % (pm, wpc))
85
86plt.figure(6)
87plt.clf()
88ct.bode_plot(L, np.logspace(-4, 3))
89
90# Add crossover line to the magnitude plot
91#
92# Note: in matplotlib before v2.1, the following code worked:
93#
94# plt.subplot(211); hold(True);
95# loglog([1e-4, 1e3], [1, 1], 'k-')
96#
97# In later versions of matplotlib the call to plt.subplot will clear the
98# axes and so we have to extract the axes that we want to use by hand.
99# In addition, hold() is deprecated so we no longer require it.
100#
101for ax in plt.gcf().axes:
102 if ax.get_label() == 'control-bode-magnitude':
103 break
104ax.semilogx([1e-4, 1e3], 20*np.log10([1, 1]), 'k-')
105
106#
107# Replot phase starting at -90 degrees
108#
109# Get the phase plot axes
110for ax in plt.gcf().axes:
111 if ax.get_label() == 'control-bode-phase':
112 break
113
114# Recreate the frequency response and shift the phase
115mag, phase, w = ct.freqresp(L, np.logspace(-4, 3))
116phase = phase - 360
117
118# Replot the phase by hand
119ax.semilogx([1e-4, 1e3], [-180, -180], 'k-')
120ax.semilogx(w, np.squeeze(phase), 'b-')
121ax.axis([1e-4, 1e3, -360, 0])
122plt.xlabel('Frequency [deg]')
123plt.ylabel('Phase [deg]')
124# plt.set(gca, 'YTick', [-360, -270, -180, -90, 0])
125# plt.set(gca, 'XTick', [10^-4, 10^-2, 1, 100])
126
127#
128# Nyquist plot for complete design
129#
130plt.figure(7)
131plt.clf()
132ct.nyquist_plot(L, (0.0001, 1000))
133
134# Add a box in the region we are going to expand
135plt.plot([-2, -2, 1, 1, -2], [-4, 4, 4, -4, -4], 'r-')
136
137# Expanded region
138plt.figure(8)
139plt.clf()
140ct.nyquist_plot(L)
141plt.axis([-2, 1, -4, 4])
142
143# set up the color
144color = 'b'
145
146# Add arrows to the plot
147# H1 = L.evalfr(0.4); H2 = L.evalfr(0.41);
148# arrow([real(H1), imag(H1)], [real(H2), imag(H2)], AM_normal_arrowsize, \
149# 'EdgeColor', color, 'FaceColor', color);
150
151# H1 = freqresp(L, 0.35); H2 = freqresp(L, 0.36);
152# arrow([real(H2), -imag(H2)], [real(H1), -imag(H1)], AM_normal_arrowsize, \
153# 'EdgeColor', color, 'FaceColor', color);
154
155plt.figure(9)
156Tvec, Yvec = ct.step_response(T, np.linspace(0, 20))
157plt.plot(Tvec.T, Yvec.T)
158
159Tvec, Yvec = ct.step_response(Co*S, np.linspace(0, 20))
160plt.plot(Tvec.T, Yvec.T)
161
162plt.figure(10)
163plt.clf()
164P, Z = ct.pzmap(T, plot=True, grid=True)
165print("Closed loop poles and zeros: ", P, Z)
166
167# Gang of Four
168plt.figure(11)
169plt.clf()
170ct.gangof4_plot(Hi * Po, Co)
171
172if 'PYCONTROL_TEST_EXAMPLES' not in os.environ:
173 plt.show()
Notes¶
1. The environment variable PYCONTROL_TEST_EXAMPLES is used for testing to turn off plotting of the outputs.