CoMet: Modeling Group Cohesion for Socially Compliant Robot Navigation in Crowded Scenes

We present CoMet, a novel approach for computing group’s cohesion and using that to improve robot’s navigation in crowded scenes. Our uses cohesion-metric builds on prior work social psychology. compute this metric by utilizing various visual features of pedestrians from an RGB-D camera on-board robot. Specifically, we detect characteristics corresponding the proximity between people, their relative walking speeds, group size, interactions members. use our design scheme accounts different levels while robot moves through crowd. evaluate precision recall pedestrian datasets. highlight performance algorithm Turtlebot demonstrate its benefits terms multiple metrics: freezing rate (57% decrease), deviation (35.7% path length trajectory(23.2% decrease).