Design and analysis of Clustering Algorithm for a Redundant Manipulator

摘要

This paper proposes a novel approach for trajectory problem of redundant manipulator with minimum singularity using non-linear Sugeno technique with cluster classification. The new algorithm includes system identification as well structure identification of fuzzy inference system for a MIMO system. Comparison with ANFIS and 'numerical method with optimization' shows that the proposed approach can produce the fuzzy model with higher accuracy, smaller number of rules and less computation time due to the absence of iterations than those which were previously achieved in other works like FCM. The proposed algorithm may be extended to solve other system identification problems also.