|
Issue
110, September 1999
Internet
Control
by
Jacob Apkarian
Start
• Model Derivation
• Linearization • Control-System
Design • Simulation
•
Coding •
Configurations•
Implementation •
Tuning and Results •
Ready for Takeoff •
Software
and Sources •
Equations PDF
TUNING AND RESULTS
My system is implemented
as described in the third configuration. The server is
located in the engineer’s office, and the client
computer and helicopter are in the laboratory, one floor
above.
Note that the operator of
the server can’t even see the helicopter while running
the controller. He has to rely solely on the server’s
real-time plotting capabilities to monitor the system’s
performance.
I augmented the block diagram
to include the simulation of the system here. Therefore,
I needed to generate code that runs in real time and performs
two tasks—running the system and simulating it at
the same time.
Using this technique, I can
simultaneously monitor the simulation and the actual response,
which is not usually done because it consumes processor
power. However, I’m sampling only at 100 Hz, and
the computations required for the simulation won’t
lengthen the computation time significantly.
Figure 6a shows the results.
The simulation and actual results match quite closely,
indicating that the linear model is a good representation
of the system under these conditions.
a) |
b)
Figure 6a—This graph shows the results with the pitch limit
set to 15° and the travel rate limit set to
15°/s. The top plot is desired (blue), actual
(red), and simulated (green) elevation. The
second graph is actual pitch in degrees. The
third graph is desired (blue), actual (red),
and simulated (green) travel. Note how close
the actual and simulation results are. b—This
graph shows the results with pitch limit set
to 45° and travel rate limit set to 100°/s.
The top plot is desired (blue), actual (red),
and simulated (green) elevation. The second
graph is actual pitch in degrees, and the third
graph is desired (blue), actual (red), and simulated
(green) travel.
|
But note that the pitch command
is limited to 15° and that I limited the travel rate command
to 15°/s, operating close to the linear region. What happens
if you let the helicopter pitch to 45° and tell the system
to travel at 100°/s?
This question is easily answered
by changing the associated parameters in the relevant
block and running the simulation and real-time controller
again. The results are shown in Figure 6b. Note that although
the simulation and the actual system match relatively
closely this time, they diverge when the helicopter pitches
to 45°.
They diverge because the
simulation doesn’t take into account the fact that
a pitch reduces the effective vertical thrust. This behavior
shows that care must be taken when simulating complex
systems with nonlinearities.
Real-time tuning can be performed
by changing the values right off the diagram or by implementing
a WinCon control panel. This panel enables you to associate
controller parameters with sliders, knobs, or switches
and gives you access to these parameters independent of
the Simulink diagram.
|
