Measuring Human Handheld Robot Collaborative Performance Using the Steering Law in Realms Exceeding Their Individual Limitations

Abstract

The overall goal of this work is to demonstrate quantitative methods of measuring the performance of humans using handheld robots to complete difficult tasks. Handheld robots build on the existing dexterity and capabilities of the human hand to enable high precision and high speed performance at the tooltip. Existing handheld robots enhance accuracy through error compensation. This work presents active collaboration through task delegation as a new paradigm in handheld robotics. The motivational use case for such high-speed, high-precision collaborative handheld robots is timely surgical treatment of sub-millimeter tissue spread over large distances. In this work, I developed a collaborative handheld robot with sub-millimeter closed-loop position servoing accuracy and a bandwidth of over 250Hz, which is 40 times the bandwidth of the human hand (5 Hz). Moreover, while Steering Law is an established framework used in the Human-Computer Interaction field for measuring and comparing individual device performance, there is no generalized framework for measuring collaborative performance particularly in tasks outside the capabilities of the collaborating agents. Through a pilot study, I established the feasibility of an experimental setup using the Steering Law framework to measure the collaborative performance of the human and handheld robot team when tracking mock lesions, and showed a 5 times improvement in the Steering Law Index of Performance. Finally, I demonstrated through a statistical study that the Steering Law framework holds in collaborative straight line steering tasks exceeding both human and robot limitations, and also showed a significant improvement in performance. This contribution will allow future researchers to quantify, characterize, and compare handheld robots, and match them to task needs.