Decentralized Delayed-State Information Filter (DDSIF): A new approach for cooperative decentralized tracking

This paper presents a decentralized data fusion approach to perform cooperative perception with data gathered from heterogeneous sensors, which can be static or carried by robots. In particular, a decentralized delayed-state information filter (DDSIF) is described, in which full-state trajectories (that is, delayed states) are considered to fuse the information. This approach allows obtaining an estimation equal to that provided by a centralized system and reduces the impact of communication delays and latency in the estimation. The sparseness of the information matrix maintains the communication overhead at a reasonable level. The method is applied to cooperative tracking, and some results in disaster management scenarios are shown. In this kind of scenario, the target might move in both open-field and indoor areas, so the fusion of data provided by heterogeneous sensors is beneficial. The paper also shows experimental results with real data and integrating several sources of information.

Research highlights
► Decentralized data fusion approach to perform cooperative perception with heterogeneous sensors.
► A Decentralized Delayed-State Information Filter (DDSIF) with full-state trajectories is considered to fuse the information.
► It allows obtaining an estimation equal to that provided by a centralized system and reduces the impact of communications delays and latency into the estimation.