Parameterized Experience Exchange in Expert - Fellow Swarm Robotic System, Controller Performance Context

摘要

This paper provides a novel approach to enhance the swarm agents' performance throughout sharing a new experience, acquired by an expert robotic agent, among the swarm group via a communication broadcast. This work is an extension to the swarm robotics cooperation in certain mission scenario. Few of the basic swarm group drills - like grouping about a target and following a leader - is realized. This is implemented by the swarm agents distributed intelligence. As a further step in this research, an Expert - Fellow swarm model is proposed where an expert agent acquires a certain skill that targets improving the performance of the control response, task execution time and power consumption. This work plots the swarm group performance before and after sharing the new experience that contribute to enhancing the agent's performance and accordingly the swarm group overall.