The authors previously introduced the fuzzy learning controlalgorithm as a method of trajectory tracking control for roboticsystems, based on conventional kinematic control methods but utilizingfuzzy regression to eliminate the need for modelling the kinematicequations of the manipulator. In this paper, the authors extend thealgorithm to the case of redundant manipulators, i.e. manipulators whichhave more joints then degrees of freedom in which they can move. Theredundant case presents additional challenges not found in thenon-redundant case, namely, a kinematic equation which must be solvedusing least squares as opposed to matrix inversion, and theincorporation of subtask optimization for the utilization of theredundant degrees of freedom. The authors address these challenges byonce again determining the main task, i.e. the “ranksolution” using fuzzy regression to determine a fuzzy Jacobianmatrix, and by determining a fuzzy performance index for subtaskoptimization using fuzzy inferencing. A simulation study is performedusing a three joint planar manipulator, with singularity avoidance asthe subtask. The results show that the extended fuzzy learning controlalgorithm causes the manipulator to satisfactorily follow the desiredtrajectory while keeping its configuration far from singularities.Moreover, it does so without being supplied any explicit model of thekinematics or the performance index. The former it learns through fuzzyregression, while the latter it possesses in the form of a fuzzy rulebase. The disadvantages and areas of future investigation for thisextension of fuzzy learning control are also discussed in this paper
展开▼
