Speed-accuracy characteristics of human-machine cooperative manipulation using virtual fixtures with variable admittance

Abstract

This work explores the effect of virtual fixture admittance on the performance, defined by error and time, of task execution with a human-machine cooperative system. A desired path is obtained using computer vision, and virtual fixtures for assistance in planar path following were implemented on an admittance-controlled robot. The admittance controller uses a velocity gain, so that the speed of the robot is proportional to the force applied by the operator. The level of virtual fixture guidance is determined by the admittance ratio, which is the ratio of the admittance gain of the force components orthogonal to the path to the gain of the force components parallel to the path. In Experiment 1, we found a linear relationship between admittance ratio and performance. In Experiment 2, we examined the effect of admittance ratio on the performance of three tasks: path following, off-path targeting, and obstacle avoidance. An algorithm was developed to select an appropriate admittance ratio based on the nature of the task. Automatic admittance ratio tuning is recommended for next-generation virtual fixtures. Actual or potential applications of this research include surgery, assembly, and manipulation at the macro and micro scales.