IDENTIFICATION AND CONTROL OF THE INVERTED PENDULUM BY USING ARTIFICIAL NEURAL NETWORKs

Abstract

Inverted pendulums are highly non-linear devices with strong tendencies to be unstable. Artificial neural networks being capable of describing non-linear phenomena can be used to simulate such systems. Using SIMULINK model, an inverted pendulum dynamics are simulated using resilient back propagation neural networks. The results show good match between predicted and actual outputs.