Weight-Perception-Based Novel Control of a Power-Assist Robot for the Cooperative Lifting of Light-Weight Objects

Abstract

We developed a 1-DOF power assist robot system for lifting objects by two humans cooperatively. We hypothesized that weight perception due to inertia might be different from that due to gravity when lifting an object with power-assist because the perceived weight differs from the actual weight. The system was simulated and two humans cooperatively lifted objects with it. We analyzed human features such as weight perception, load forces, motions etc. We found that the robot reduced the perceived weights to 25% of the actual weights, and the load forces were 8 times larger than the actual requirements. The excessive load forces resulted in excessive accelerations that jeopardized the performances. We then implemented a novel control based on the human features, which was such that a virtual mass exponentially declined from a large value to a small one when subjects lifted objects with the robot and the command velocity exceeded a threshold. The novel control reduced excessive load forces and accelerations and thus enhanced performances in terms of maneuverability, safety etc. The findings may be used to develop power assist robots for manipulating heavy objects in industries that may augment human's abilities and skills and may improve interactions between robots and users.