An Approach to Robot Motion Behaviour Representation

Abstract

We propose a set of algorithms for robot motion behaviour representation based on our previous research, fuzzy qualitative trigonometry and fuzzy qualitative robot kinematics. It aims at developing a fuzzy qualitative framework for the connection of qualitative and quantitative representations. Firstly, we revisit fuzzy qualitative robot kinematics and derive fuzzy qualitative robot kinematics in terms of D-H robotic structure, fuzzy qualitative trajectories of a PUMA robot have been provided to prove the proposed method. Then robot behaviour extraction has been studied using three algorithms which are an adopted ordered weighted averaging operators, k-means classifier and EM clustering. Finally, simulation results of the PUMA robot using the three algorithms have been demonstrated and discussed with concluding remarks.