A framework for adapting social conventions in a mobile robot motion in human-centered environment

Interestingly in different situations, human not only plans differently for approaching, accompanying, passing by and avoiding another person, but also smoothly maintains an appropriate distance. But for a mobile robot it is not trivial at all, while also maintaining its goal. In this paper we present a generic framework of mobile robot path planning for adapting social rules at different states of execution, which apart from assuring safety, also respects the comfort and expectations of human, and convey its intention to human well in advance. In our approach for treating human explicitly robot constructs different sets of regions around human and iteratively converges to a set of points (milestones), using social rules, human proximity rules and task oriented rules to generate a smooth path. We have compared our results with the case, when robot is purely reactive.