Motion Strategies for People Tracking in Cluttered and Dynamic Environments

Abstract

The key contribution of this work is the construction of a robot tracking system that follows an unpredictable target (e.g people) in cluttered, unknown and dynamic environments. To do so, mobility constraints, sensor limitations, and uncertainty have been addressed at both the algorithmic and implementation levels. The work formulates target tracking as minimizing an objective function that models the risk of the target's escape and derives this risk function from basic principles. Various practical hardware limitations - e.g., sensor's field of view, safe navigation - are taken into account either as part of the risk function or as constraints during risk minimization. The resulting tracker demonstrates substantially improved performance, compared with other local online strategies. The tracker was able to track a person walking around in indoor office environments as well as in a highly dynamic and public environment like the school cafeteria successfully.