Weight-Prediction-Based Predictive Optimal Position and Force Controls of a Power Assist Robotic System for Object Manipulation

In the first step, a novel approach of inclusion of weight prediction in the dynamics of a power assist robotic system (PARS) for lifting objects is proposed. Then, a position control and two force control algorithms are derived using the weight-prediction-based dynamics so that the inclusion of weight prediction in the controls can mitigate the effects of the erroneous feedforward input force programming of the human. In the second step, a variable-inertia-based empirical predictive control strategy is proposed to augment the performance of the controls. The control algorithms are tested using a 1-DOF PARS for lifting lightweight objects. A novel scheme is proposed to evaluate and optimize the performance of the controls. The experimental results show that weight-prediction-based predictive controls help optimize human-robot interaction (HRI) and manipulation performance. The results show that the position control produces better HRI and performance than the force controls, and the force control designed with force feedback performs better than the force control designed with acceleration feedback. The results seem to be very useful to design predictive controls of power assist robots for handling heavy objects in industries.