An adaptive robust compensation control scheme using ANN for a redundant robot manipulator in the task space

The author presents a dynamic control scheme using artificial neural networks (ANNs) for a redundant robot manipulator in the task space. It has adaptability for the unknown parameters of the system model by rising ANNs. Through resolving the redundancy this scheme can solve the problem of singular free and collision avoidance if the proper coefficients of the performance index and the set of constraints of the states of the joints are given. This scheme has good robustness for bounded disturbances, parameter mismatch, the constraints on the control signals, and the states of the joints. Therefore, this scheme can not only increase flexibility of the redundant robot manipulator but also increase its compatibility with the task. By using decentralized control this scheme can effectively reduce the quantity of computations online