Practical Aspects of Model-Based Collision Detection

Recently, with the increased number of robots entering numerous manufacturing fields, a considerable literature has grown up around the theme of physical \added[]{human-robot} interaction using data from proprioceptive sensors \added[comment={R2:3}]{(motor or/and load side encoders)}. Most of the studies have been taking for granted the accurate dynamic model of robot. In practice, however, model identification and observer design \replaced{precede}{proceeds}\added[comment={R2:4}]{} collision detection. To the best of our knowledge, no previous study has systematically investigated each aspect underlying physical human-robot interaction and the relationship between those aspects. In this paper, we bridge this gap by first \replaced{reviewing the literature on}{discussing theoretical side of} model identification, disturbance estimation and collision detection, \added[]{and discussing} the relationship between the three, then \replaced{ by examining the practical sides of model-based collision detection}{practical sides} on a case study conducted on UR10e. We show that model identification step is critical accurate collision detection, while the choice of the observer should be mostly based upon computation time, the simplicity and flexibility of tuning. It is hoped that this study can serve as a roadmap for equipping industrial robots with basic physical human-robot interaction capabilities.