Uniting model-based and model-free methods for robotic manipulator direction control

Model-based and model-free control methods each have associated strengths and weaknesses, This paper discusses the positive effect in performance obtained by uniting the two approaches within a hybrid controller. In the domain of robotic manipulator direction control, the authors describe and compare: (a) the performance of resolved motion rate control, a well-known model-based control method; with (b) the performance of machine nervous systems for motor control, a novel model-free control method; and (c) the performance of a hybrid control scheme which incorporates the functionality of the other two methods. Their results show that the hybrid control possesses all the strengths of its constituent methods and none of their weaknesses