An Integrated Fuzzy and Learning Approach to Performance Improvement of Model-Based Multi-Agent Robotic Control Systems

This paper presents an integrated approach to improving the performance of model-based control for multi-agent robotic systems (MARS). The fuzzy logic and learning techniques are compactly and efficiently integrated into the proposed approach to yield an improved formation controller for MARS while ensuring the stability obtained from model-based control systems. As a case study the proposed approach is applied to a leader-follower MARS where the robotic leader agent has its own target and the robotic follower agent is constrained by formation tasks. Simulation results are presented to demonstrate the effectiveness of the integrated fuzzy and learning approach.