Fuzzy Logic PD Control of a Non-Linear Inverted Flexible Pendulum
Kou A. Kong · California State University ScholarWorks (system-wide DSpace) · 2009
ABSTRACT FUZZY LOGIC PD CONTROL OF A NON-LINEAR INVERTED FLEXIBLE PENDULUM by Kou A. Kong Master of Science in Electrical and Computer Engineering Electronic Engineering Option California State University, Chico Summer 2009 In automatic control systems, an inverted pendulum is a classical system to be controlled. The inverted pendulum is more complex and complicated to control when flexibility is present in the pendulum. Flexibility in the system can cause a system to be unstable. In this project, an intelligent fuzzy proportional-derivative (PD) control is proposed to control the flexible inverted pendulum. The flexible inverted pendulum is modeled in non-linearity using Euler-Lagrange energy method. The fuzzy proportionalderivative (PD) control is compared with the classical PD control in this system. Even though the classical PID control is widely used in industrial control systems, it will break down in most non-linear systems. The simulations of the flexible inverted pendulum are done in Simulink and Matlab. The linear and nonlinear flexible inverted pendulums are simulated for six different cases. The simulation results demonstrate the performance differences between the classical PD control and fuzzy PD control. The classical PD control works for the linear system in terms of oscillations and settling time for all step inputs and cases. On the nonlinear system, the classical PD control has more oscillations in all three modes of vibrations for the second and third step inputs. The fuzzy PD controller shows better system responses and steady-state performances for the nonlinear system. The second and third step inputs have fewer oscillations for all three modes of vibrations. The fuzzy PD controller has a better overall performance because it has the ability to adjust the gain values according to the system output response. On the other hand, the conventional PD controls have a constant gain value for different system output responses. These results are integrated in Chapter VI. It can also be seen by running the programs on the CD included with this report. The instructions to simulate the programs on the CD are in the Read Me Instructions in Appendix E.