Determining Movement Measures for Trust Assessment in Human-Robot Collaboration Using IMU-Based Motion Tracking

Close-proximity human-robot collaboration (HRC) requires an appropriate level of trust from the operator to the robot to maintain safety and efficiency. Maintaining an appropriate trust level during robot-aided production requires non-obstructive real-time human-robot trust assessment. To this end we performed an experiment with 20 participants performing two types of HRC tasks in close proximity to a Kuka KR 300 R2500 ultra robot. The two tasks involved collaborative transport of textiles and collaborative draping, respectively. During the experiment we performed full body motion tracking and administered human-robot trust questionnaires in order investigate the correlation between trust and operator movement patterns. From the initial per-session analyses we see the effects of task types on movement patterns, but the correlations with trust are weak overall. Further analysis at higher temporal resolution and with correction for participants' base movement patterns are required.