Measuring the Effectiveness of Readability for Mobile Robot Locomotion

The inclusion of social and human-like behaviors is recently emphasized within mobile robot locomotion. These behaviors are required to support the seamless integration of mobile robots into environments which they share with humans. This work demonstrates a distinct benefit of these behaviors, which goes beyond positive apperception. It is shown that human-like robot locomotion reduces the planning effort for all agents within an environment. This effect is revealed in an experiment that compares human locomotion during avoidance of an oncoming human or wheeled robot. In order to evaluate recorded data, a framework for the analysis of human trajectories is proposed. Confidence intervals based on a spline regression model are used to account for variance in the data. This qualitative method is complemented by a comparative analysis, that quantifies differences and analogies within the data. Thus, the framework allows for a statistically feasible qualitative and quantitative analysis of trajectories. Results show, that extra planning effort for the avoidance is prevented by readable human-like robot locomotion. The study indicates that locomotion planning requires less effort from subjects if the mutual trajectory prediction is facilitated by robots that externalize intentions and comply with human-like behaviors.