Modular adaptive neurocontrol of robotic manipulators

In this paper a modular neuroadaptive controller for robotic manipulators is proposed. First, a modular model-based adaptive scheme and a modular neurocontroller based on the Newton-Euler formulation are presented. The adaptation technique is then utilized for the on-line direct adaptive training of the modular neurocontroller. Both the controller and the training scheme are arranged in a modular form, resulting in the proposed modular neuroadaptive controller, which can be applied to any N-link serial manipulator. Simulation results of the proposed controllers for a 3 DOF manipulator are presented