MPC to optimise performance in power‐assisted manipulation of industrial objects

Abstract

A 1-DOF power-assist robotic system (PARS) is developed using a linear actuator for vertical manipulation of heavy objects in collaboration with its human user. Human's differential perception of inertia and gravity is considered in the manipulation dynamics, and an admittance-type feedback position control scheme is designed to control the motion of the system. A human manipulates a heavy object with the robot in harmonic motion (the object is repeatedly lifted up and lowered down). A common performance index for the human and the robot for the collaborative task is derived in terms of manipulation velocity and precision. An event-triggered model predictive control (MPC) scheme is designed to maintain optimal performance of the robot and the human through maintaining optimum robot velocity and precision at the events when low performance is observed. Results show that the weight-perception-based control helps produce satisfactory collaborative performance, and the MPC is proven effective to maintain optimal performance and improved human–robot interaction. The findings are useful to develop predictive control strategies for human-friendly PARSs for manipulating heavy objects in industries in particular, and for other variable-speed electric drive systems for industrial applications, in general.