Multi-Robot Cooperative Object Tracking Based on Particle Filters

摘要

This article presents a cooperative approach forrntracking a moving object by a team of mobile robots equippedrnwith sensors, in a highly dynamic environment. The tracker’srncore is a particle filter, modified to handle, within a single unifiedrnframework, the problem of complete or partial occlusion for somernof the involved mobile sensors, as well as inconsistent estimates inrnthe global frame among sensors, due to observation errors and/orrnself-localization uncertainty. We present results supporting ourrnapproach by applying it to a team of real soccer robots trackingrna soccer ball.