Decentralized sensor fusion with distributed particle filters

摘要: This paper presents a scalable Bayesian technique for decentralized state estimation from multiple platforms in dynamic environments. As has long been recognized, centralized architectures impose severe scaling limitations distributed systems due to the enormous communication overheads. We propose strictly approach which only nearby exchange information. They do so through an interactive protocol aimed at maximizing information flow. Our is evaluated context of surveillance scenario that arises robotic system playing game laser tag. results, both simulation and using physical robots, illustrate unprecedented capability large teams vehicles.