What we see and What we don't see: Imputing Occluded Crowd Structures from Robot Sensing
This addresses a critical safety and efficiency issue in robot crowd navigation, though it is an incremental step as it builds on existing sensing and prediction methods.
The paper tackles the problem of inferring human occupancy in occluded areas around a robot in crowded environments, proposing the first solution to sample predictions of possible human presence based on sensed people and previous crowd observations.
We consider the navigation of mobile robots in crowded environments, for which onboard sensing of the crowd is typically limited by occlusions. We address the problem of inferring the human occupancy in the space around the robot, in blind spots, beyond the range of its sensing capabilities. This problem is rather unexplored in spite of the important impact it has on the robot crowd navigation efficiency and safety, which requires the estimation and the prediction of the crowd state around it. In this work, we propose the first solution to sample predictions of possible human presence based on the state of a fewer set of sensed people around the robot as well as previous observations of the crowd activity.