Multi-robot Joint Localization and Target Tracking with Local Sensing and Communication

In this paper, we study the problem of target state estimation using a mobile robot network, where a team of robots jointly localize themselves and track multiple targets with onboard sensor measurements. We consider a general case, where 1) Absolute measurements might be accessible intermittently; 2) Any robot might detect (or be detected by) multiple robots synchronously; 3) There exists a time-varying communication topology between robots; 4) The robots not directly sensing the targets might change with time. A fully Distributed Hybrid Extended Information Fusion(DHEIF) algorithm is introduced. Unlike most existing works, we do not assume that each robot has good knowledge of its own state. Hence, each robot establishes a filter bank to estimate its own state (localization) and the target states (tracking). To avoid the measurements being used by one robot more than once, the Interleaved Update(IU) technique is adopted in the localization part to compute a consistent estimate. Through communicating with its neighbors, each robot maintains consistent state estimates of the targets even if they are outside the visibility range. The performance of our algorithm is tested in simulations.