Precise position control of stencilling robot manipulator using neural network

In this paper, accurate position control of a stencilling robot manipulator is designed. The stencilling robot mounted on a truck is required to draw lines and characters in the Cartesian space. One of the goals of the stencilling robot design is to make the controller as simple as possible such that a simple PID controller can be used. Since the robot is huge and heavy, the inertia of the robot is expected to play a major role in the tracking performance. The conventional PID controller works within certain accuracy, but does not guarantee the accurate trading performance as desired. Here we propose a neural network control scheme for a computed-torque like controller for the stencilling robot. The stability of the closed loop system is analyzed based on the Lyapunov theory. Simulation studies with the stencilling robot are carried out to test the performances.