A Neutral Emulation of the PID Controller for the Positioning Inverted Pendulum System

Abstract

A neural network is one of the artificial intelligence solutions whose ability to clone transfer functions can be used to solve complex real life problems that can not be modeled using mathematical equations. This project is about a neural clone of a simple non-optimal proportional integral-derivative (PID) controller that is used to control a position-based inverted pendulum system. The neuro-controller is allowed to adapt the characteristics of the PID controller and it will be used as an alternative controller for the inverted pendulum system. Results showed that the neural net was able to clone the non optimal PID controller and exhibit a more stable characteristic than the latter. The neuro-PID controller was able to
control the inverted pendulum system over prolonged period of operation