A New Neural - Active Force Control Architecture in High Precision SCARA System

摘要

This article highlights of the success of developing new hybrid control architecture for Selective Compliance Articulated/ Assembly Robot Arm (SCARA) manipulator. The hybrid control architecture is a combination of classical Proportional-Derivative scheme and the intelligent Active Force Control (AFC) coupled with Artificial Neural Networks (ANNs). The advantage of this architecture is its non-dependency to physical model like other techniques used in the literature. The performance of the proposed method was assessed viz. simulation in which utilized MATLAB/SIMULINK. Based on the simulation results, it is proven that the proposed scheme has outstanding performance compared to the classical control scheme. Using AFC strategy, the robot was able to achieve its desired performance excellently with average error of only 0.01%. Incorporating neural networks into the system tightens the errors and reduces the time taken for the robot to achieve its desired response.