Implementation of Compound Knowledge in Learning Control System for Obstacle and Goal Consideration

摘要

A learning based control system often being limited to the environment where the learning occurs. Applying a number of control function for any operation's environment could provide flexibility in executing an operation. However, a method of utilizing all the applied control functions into an operation requires a method of implementing all the control knowledge into one applicable source at the required moment. Here, a method of Learning Control that utilizes multiple Control Functions by Compound Control Knowledge is proposed for a control system. The proposed method was applied for obstacle and goal consideration functions. The results shows that the system was able to utilize the Compound Control Knowledge in prioritizing the usage of the required control function for the condition assigned.