An implementation of a knowledge-based system method to an active force control robotic scheme

Abstract

This thesis presents a method to improve the response of a robotic control system in order to obtain a more robust system performance using an active force control with a knowledge-based system method called the Active Force Control and Knowledge-Based System (AFCKBS). The focus ofthe study is the implementation of a mechanism to refme the trajectory error generated by an Active Force Control (AFC) scheme applied to a two-link robotic planar arm. This is accomplished by utilizing features ofthe signal patterns produced by the input and output functions. Based on the previous research that employed a number ofrobotic control schemes, knowledge ofthe trajectory track error pattern was obtained when a specific trajectory tracking command was forced to the system. The correlation between the generated track error pattern and the input function of the robot scheme provides the essential knowledge to be utilized via the proposed technique. A simulation and experimental study was performed and the result obtained was consequently analyzed and compared to show the performance enhancement in the proposed method.