Predictive Mechanisms increase Efficiency in Robot-supported Assemblies: An Experimental Evaluation

Markus Huber,Claus Lenz,C. Wendt,B. Färber,A. Knoll,S. Glasauer
Abstract:The presented work focuses on investigating the influence of different hand-over timing strategies on the fluency and efficiency of a human-robot team in an assembly task. To this aim, four different timing strategies were experimentally performed with 37 volunteers: (I) a fixed time interval between two hand-overs, (II) a reactive behavior, where the robot is triggered by the human, (III) a fixed time intervals depending on the current component, and (IV) a predictive assembly duration estimation algorithm. During the experiment, the time-to-completion of the task and the waiting times for human and robot were measured as reciprocal indication for the efficiency and respectively the fluency of the team. The results indicate that the efficiency of the human-robot team is significantly increased using a predictive timing strategy, because it enables the robot to provide the needed component just-in-time. The decrease in waiting times for the human worker leads to improved fluency of the collaboration. In addition, the predictive strategy enables the robot to perform preliminary tasks if no assistance is needed. Therefore, a nearly full usage of both partners’ capacities can be reached.
What problem does this paper attempt to address?