Simulated annealing based control of an Inverted Pendulum System

Abstract

Inverted Pendulum System is a control engineering project as it is a highly nonlinear and unstable system. The aim of the study is to stabilize the inverted pendulum upright always during disturbances such that the position of the cart on the track is controlled quickly and accurately Thus this system is chosen as an ideal platform for verifying the effectiveness of many control algorithms. In this work, Linear Quadratic Regulator (LQR) and Artificial Neural Network (ANN) controllers are developed for controlling the Inverted Pendulum System. LQR controller is controlling the Linear Inverted Pendulum, whose weight matrices, Q and R are optimized using simulated annealing optimization technique. Nonlinear Inverted Pendulum System is controlled by ANN, whose weights and bias are optimized using Simulated Annealing optimization technique. Robustness of the controllers are also checked.