Color Recognition for NAO Robot Using Sugeno Fuzzy System and Evidence

Abstract

This paper examines the problem of recognition of colored objects for the NAO humanoid robot using cameras. Based on the ability of fuzzy systems to approximate any mapping, the input-output behavior between the triplet {H, S, V} and the linguistic output colors is represented by a Sugeno fuzzy system. In order to improve the recognition performance of the robot, more cameras are used. In this case, a decision mechanism based on the Dempster-Shafer (DS) theory is employed to resolve the conflict between cameras. The feasibility of the proposed methodology has been experimentally validated with a real-time implementation.