Comparison of performance of neural network controller and DAC based adaptive controller for a robot arm

Disturbances in most of the systems makes their performance mitigating. So in order to comply with the situation, a controller should be developed which could adapt to these disturbances and get the best out of the system. This paper gives a comparison of the results of a disturbance accommodating control based adaptive controller (or better known as linear adaptive controller) and a neural networks model reference controller. The aim of the control algorithm is to control the movement of a robot arm, which is freely suspended. The input to the plant is the torque from the controller. The input to the controller is the difference between the reference angle and the output "angle" from the plant. In comparison, the results prove that the linear adaptive controller adapted in a much wider sense to the disturbances and changes in the system than the neural network model reference controller. The simulation figures support the results that were put forth.