Java程序辅导

A Multidisciplinary Digital
Control Systems Laboratory
Gregory L. Plett, David K. Schmidt
University of Colorado at Colorado Springs
ABSTRACT
We describe a multidisciplinary digital control systems lab-
oratory under development. Its use is shared by Electri-
cal and Computer Engineering (ECE) and Mechanical and
Aerospace Engineering (MAE) students.
The lab comprises interesting and stimulating physical de-
vices: the Educational Control Products Magnetic Levita-
tion and Control-Moment Gyroscope systems.
Control is accomplished using a digital computer running
the Real Time Linux operating system, via MathWorks’ Mat-
lab/ Simulink/ the Real Time Workshop (RTW) and Quality
Real-Time Systems’ Real Time Linux Target (RTLT).
GOALS FOR THE NEW LAB
Hands-on: Promote control-systems education with exper-
imental control of a physical device. Lab projects are de-
signed to complement and synchronize with lecture topics
in order to best reinforce concepts.
Economy: A multidisciplinary ECE/MAE facility allows ef-
ficient use of space and equipment, better use of available
funds, and elimination of overlap among individual depart-
mental labs. Focusing experiments on one or two devices
results in economies of space, money and student time.
CHOICE OF LAB DEVICES
We decided to base our new lab primarily around the Mag-
netic Levitation (MagLev) and Control-Moment Gyroscope
units by Educational Control Products (ECP); see Fig. 1.
Both provide dramatic and interesting demonstrations and
may be configured in a variety of ways; see Table 1.
Figure 1
Main lab devices: ECP MagLev on the left and Gyro on the right.
Table 1
Attractive attributes of the selected dynamical devices.
Desirable dynamic attribute MagLev Gyro
1. Linear single-variable, stable Y Y
2. Linear single-variable, unstable Y Y
3. Nonlinear single-variable, stable Y Y
4. Nonlinear single-variable, unstable Y Y
5. Linear multi-variable, little I/O interaction Y N
6. Nonlinear multi-variable, large I/O interaction N Y
7. Dynamically rich system N Y
8. Electromechanical system Y Y
A generic PC (running RTLinux, Matlab, Simulink, RTLT)
is used as the controller via D2A and A2D circuits connected
to a “breakout box” that the student can access. Power am-
plifier/sensor conditioner boxes drive the devices. See Fig. 2.
This configuration allows investigation of digital effects: dif-
ferent sampling rates, fixed-point math, discrete-time com-
putational structures.
Gyroscope Power Amplifier
Breakout Box
MagLev Power Amplifier
ON
DAC2
DAC/
DAC/
ADC4
ADC/ADC3
ADC/ADC2
ADC/ADC1
DAC/
DAC/
Gyroscope Plant
MagLev Plant
ON
BRAKE
ON
BRAKEec
p 
M
od
el
 7
50
Co
nt
ro
l−
M
om
en
t
Gy
ro
sc
op
e
ec
p 
M
od
el
 7
30
Le
vi
ta
tio
n
De
vi
ce
M
ag
ne
tic
De
vi
ce
AXIS 3 PWR
OFF DAC1
AXIS 4
PWR DAC1
DAC2ADC/
ecp
ecp
Educational Control Products
A
D
C/
A
D
C/
A
D
C/
A
D
C/
A
D
C1
A
D
C2
A
D
C3
A
D
C4
D
A
C1
D
A
C2
D
A
C/
D
A
C/
D
A
C/
D
A
C/
D
A
C1
D
A
C2
Educational Control Products
Model 730
Model 750
Se
rv
oT
oG
o 
M
od
el
 2
 I/
O
 B
oa
rd
Standard PC Hardware
RTLinux Operating System
Matlab Environment
Simulink
RTW
RTLT
St
ud
en
t
Figure 2
System setup including hardware and software integration.
A sample Simulink diagram for discrete time multi-input
multi-output control of the MagLev is shown in Fig. 3. This
control system is compiled and executed without writing a
line of code!
Bottom Coil
Top Coil
    Scale reference input for     
good steady−state error.
Estimate the state
value using            
measurements of  
     the output y[k].          
    Compute state feedback    
    by multiplying state           
  estimate by K.                 
r2−source
r1−source
z
1
Scope
K
Matrix 
Gain (Cd)
K
Matrix 
Gain (Bd)
K
Matrix 
Gain (Ad)
K
Matrix 
Gain 
inv(Kdc)
K
Matrix
Gain (L)
K
Matrix
Gain (K)
HW Adapter
yrawycal
Fix Top
Sensor
yrawycal
Fix Bottom
Sensor
D/A
Digital To Analog
Converter1
D/A
Digital To Analog
Converter
Demux
Demux
y2o
y1ou1o
u2o
A/D
Analog To Digital
Converter1
A/D
Analog To Digital
Converter0
u[k]
kxhat xhat[k] y
Figure 3
Simulink diagram for discrete-time MIMO control.
COURSE ORGANIZATION
We have developed two lab courses: undergraduate Digi-
tal Control Laboratory, and graduate Digital Flight Control.
Both cover similar topics, although the graduate class leaves
more of the theory development to the student. A Gantt chart
showing the relative phasing of undergraduate lecture and
lab curriculum is shown in Fig. 4.
INITIAL EVALUATION
We wanted to test two hypotheses with the new lab:
Hypothesis 1: Hands-on learning improves basic under-
standing. Therefore, students taking the lecture course only
(the lab is not required) will do more poorly in the lecture
course than those students taking the lab as well.
Results: Over four semesters, students taking both the lec-
ture and lab courses did 11% better than students taking only
the lecture course.
Hypothesis 2: An experimental lab provides better learning
than a lab based on simulation.
Results: Students in the experimental lab did 11% better
than students in a simulation-based lab (although not enough
data is available for this result to be statistically significant).
1. Introduction to digital control.
Sp
rin
g 
Br
ea
k V
ac
at
io
n
Unit 4: State−Space Controller Design
In
−
cl
as
s e
xa
m
 I 
(to
pic
s 1
−6
)
In
−
cl
as
s e
xa
m
 II
 (t
op
ics
 7−
9)Review of continuous−time control.
Emulation of analog controllers (I).
The z−transform.
Emulation of analog controllers (II).
Sampling and reconstruction.
Discrete−time systems.
Stability analysis techniques.
Digital controller design (transfer fns).
State space−−continuous time.
State space−−discrete time.
Discrete estimator design.
Review.
Discrete−time simulation with Simulink.
Time−domain controller emulation.
Frequency−domain controller emulation.
Sampling, aliassing, zero−order hold.
Discrete−time plant modeling.
Filter stucture & finite−precision effects.
Frequency−response control design.
Numeric optimal PID control design.
Ragazzini’s direct control design method.
State−space model verification (disc.−time).
State−feedback control design.
Unit 3: Transfer−Function Controller Design
Unit 2: Digital Effects
Unit 1: Design by Emulation
Laboratory Orientation
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Laboratory Topic
Lecture Topic 1 2 3 4 5 6 7 8
Week of Semester
9 10 11 12 13 14 15 16
Figure 4
Synchronization between lecture topics and lab topics.
ADDITIONAL STUDENT FEEDBACK
Additional student feedback indicates that real learning is
taking place. Some students become so involved in the lab
that they try experiments that are not required. Other com-
ments validate the belief that the lecture and lab courses are
well coordinated.
CONCLUSION
We have developed a multidisciplinary analog and digital
control-systems laboratory, focused on the ECP MagLev and
Gyro plants, using Matlab/ Simulink/ RTW and RTLT soft-
ware, on the RTLinux operating system. We have written
a lab manual for the undergraduate Digital Control Labora-
tory, which uses the MagLev device and is available on the
Internet. The lab schedule coordinates with the Digital Con-
trol Systems senior-elective class so that experiments com-
plement and illuminate the theory. We have also developed
a graduate Digital Flight Control class. Initial evaluation in-
dicates that the laboratory experience has significantly aided
learning of control-systems concepts.