MULTILAYER NEURO PID CONTROLLER BASED ON BACK-PROPAGATION ALGORITHM

Abstract

The report covers the basic idea, mathematical features of conventional Neuro PID controller which is a technique to make any linear or nonlinear system unaffected to unpredictability of system's parameters and disturbances such as noise. Here we suggest a technique to apply neural networks for the tuning of the PID (proportional, integral and derivative) controller's gains in a way human tune the gains depending on the environmental and systems requirements. Error back-propagation method is used as the tuning method for the controller which is also known as BP method and this method works on the local minima algorithm. The Multilayer Neuro PID controller will be used for system such as Magnetic Levitation, Ball and Beam and Inverted Pendulum. These systems are non-linear system, so control action on these systems by conventional controllers is very difficult and time consuming. Optimization of controller is required to get the best result. In the last comparison between systems without controller, conventional PID and Multilayer Neuro PID controller will be done to obtain the best suited controller for that system. This comparison will be in terms of steady state and transient parameters of the system.