Efficient Low-cost Controllers for Constrained Manipulators with Uncertainties and Disturbances

This work presents an efficient control strategy for robot manipulators with constraints and affected by uncertainties and disturbances. The controller uses a combination of model predictive control (MPC) with an adaptive robust feedforward term. The predictive controller is based on interpolations between different simple solutions to guarantee the feasibility of the final solution applied to the manipulator. The proposed method improves the existing techniques in terms of robust capabilities. Feasibility is preserved with the MPC and applicability is also guaranteed as the computational load of the interpolation algorithm is low. The benefits of the strategy, compared with other existing controllers, are shown with simulation results obtained with a PUMA-560 manipulator.