Force feedback in remote tele-manipulation

Abstract

It is becoming increasingly necessary to carry out manual operations in environments which are hazardous to humans - using remote manipulator systems that can extend the operators reach. However, manual dexterity can become severely impaired due to the complex relationship that exists between the operator, the remote manipulator system and the task. Under such circumstances, the introduction of force feedback is considered a desirable feature, and is particularly important when attempting to carry out complex assembly operations. The dynamic interaction in the manmachine system can significantly influence performance, and in the past evaluation has been largely by comparative assessment. In this study, an experimental remote manipulator system, or tele-manipulator system, has been developed which consists of three electrically linked planar manipulator arms, each with three degrees of freedom. An articulated 'master' arm is used to control an identical 'slave' arm, and independently, a second kinematically and dynamically dissimilar slave arm. Fully resolved Generalized Control has been demonstrated using a high speed computer to carry out the necessary position and force transformations between dissimilar master and slave arms in realtime. Simulation of a one degree of freedom master-slave system has also been carried out, which includes a simple model of the human operator and a task based upon a rigid stop. The results show good agreement with parallel experimental tests, and have provided a firm foundation for developing a fully resolved position/position control scheme, and a unique way of backdriving the master arm. Preliminary tests were based on a peg-in-hole transfer task, and have identified the effect on performance of force reflection ratio. More recently a novel crank-turning task has been developed to investigate the interaction of system parameters on overall performance. The results obtained from these experimental studies, backed up by simulation, demonstrate the potential of computer augmented control of remote manipulator systems. The directions for future work include development of real-time control of tele-robotic systems and research into the overall man-machine interaction.