The Identification of Tool Suspension Impedance via the Measurement of Expert Motions and Its Application to Robot Teaching

Abstract

A new approach to the teaching of skills that human workers have acquired in performing specific tasks is presented. First, a skilled worker shows how to perform a given task, and his motions are measured. Specifically, the force exerted by the worker and the displacement of the tool manipulated by the worker are monitored and stored in a computer. The data are then analyzed in order to find the control law of the human expert. The functional relationship between the force and the displacement is derived from the data by using a curve fitting technique. The identified relationship is then used as a reference model for controlling a manipulator arm to replicate the expert motion. The method is also applied to the impedance control of a simple grinding robot. The impedance with which a human expert holds the grinding tool is identified. Then the identified impedance is transferred to the robot controller so that the system may follow the impedance model, which is a replica of the human expert.