A fuzzy behavior-based architecture for decision control in autonomous vehicles

Abstract

A three layer behavior-based architecture using fuzzy logic is presented. A distinguishing feature is the similarity among its layers, facilitating architecture understanding and improving robot control designability. Other distinctive features are the simplicity in the way behaviors are modulated; how robustness is achieved, and how behaviors are fused. As a consequence, the whole architecture presents some interesting properties from a design perspective. Some experiments over a Khepera mobile robot are presented, attesting the interest in using such architecture in long runs.