### Project 4 - University of Cincinnati

```Advantages of Using Fuzzy Logic in Dynamic Feedback Control Systems
Adam Katterheinrich ASE, Nicholas Nielsen ASE, Tyler Parcell CompE
Dr. Kelly Cohen (Faculty Mentor), Dr. Jeff Kastner (Sub-Faculty Mentor), and Wei Wei (Graduate Assistant)
College of Engineering and Applied Science, University of Cincinnati, Cincinnati, OH 45221
University of
Cincinnati
Introduction
•
•
•
Dynamic control systems require quick,
accurate and robust responses
PID Stability Envelope
PID Envelope (Sensor Noise)
PID Envelope (Force Noise)
Avg. Settling
Time = 4.69 sec
Control of an inverted pendulum system and
autonomous rotorcraft is the main application
focus
Consistent Stability Envelope Increase using Fuzzy PID
2
Fuzzy PID Stability Envelope
Expand stability envelope using Fuzzy Logic PID controller
Fuzzy PID Envelope (Sensor Noise)
Fuzzy PID Envelope (Force Noise)
Provide active gain updating for noise and varying initial conditions
Procedures
4
PID Stability Envelope Subjected to Internal and External Noise
Fuzzy logic control can provide a new edge to
conventional Proportional, Integral, Derivative
(PID) techniques
Objectives
•
•
Inverted Pendulum System Results
1
Avg. Settling
Time = 4.81 sec
3
• Dual PID control loop using the cart position and
pendulum angle feedback errors for comparison
Stability Envelope Performance Comparison
Settling Time
Stability Envelope
• Dual FUZZY PID control
loop using the same
setup but replacing each
controller with fuzzy
inference systems
•
•
FUZZY inference system outputs
PID gains based on angle and
cart position errors
Each gain has ± 10% deviation on
bounds
± 10% based off original tuned PID
gain values
Sensory
Noise
Force
Noise
PID
2.91 s
8.07 s
3.08 s
2.66 s
334 pts
Fuzzy PID
4.79 s
4.82 s
4.81 s
0.03 s
+44%
Fuzzy
compared
to PID
+64%
-40%
+56%
No
Noise
Sensory
Noise
Force
Noise
PID
204 pts
172 pts
232 pts
Fuzzy PID
312 pts
274 pts
Fuzzy
compared
to PID
•
No
Noise
Average change
in settling time
with Noise
+53%
+59%
Conclusions
Stability: Applying a Fuzzy Logic rule base to PID, expanded the stability envelope by at least 44%
High Performance: Average change in settling time decreased from 2.66sec (PID) to 0.03sec (Fuzzy
PID)