Humanoid robot hierarchical navigation using Petri nets and fuzzy logic

Abstract

This paper presents a supervisory control system for humanoid robot motion planning. The proposed system is a supervisory structure formed by two hierarchical levels of a discrete event system. The high level system is represented by a Petri net. This Petri net behaves as a supervisor that indicates the sequence of motions that the robot has to make. A robot walking in a closed space forms the lower level. The robot decisions are modeled by a fuzzy logic configuration using a Fuzzy Inference System (FIS).