Neural net architectures to convert existing servo controllers into intelligent adaptive controllers

An artificial neural network was used to convert a servo motor controller based on conventional design techniques into an intelligent adaptive controller for better performance and accuracy in the presence of system nonlinearities, parameter variations over time, and uncertainties. Two learning algorithms are proposed to correct the motor inputs properly. The use of an existing controller guarantees coarse learning and provides better generalization and correction capabilities. Simulations show very encouraging results. The performance of the proposed controller is compared with a PID controller and a MRAC