On-line affective cognitive learning and decision-making for autonomous navigation of mobile robots

Abstract

A new autonomous navigation control framework is presented for mobile robots by integrating affective cognitive learning and decision making (ACLDM) model with behavior-based robot system. Cognitive states for work environment of mobile robot are gotten from a pattern classifier based on Adaptive Resonance Theory-2 (ART-2) network. Then rational strategies for behaviors coordination are developed by on-line affective cognitive learning. The behaviors of robot navigation are designed by dynamic system approach. This control strategy can make the mobile robot navigate autonomously and safely in unknown environment. Simulation studies have demonstrated that the integration of the affective system with cognitive system can speed up the learning process, and the proposed strategy can effectively improve the capability of robot's autonomous navigation in unknown environment.