Target localization with a distributed Kalman filter over a network of UAVs

Abstract. Unmanned Aerial Vehicles (UAVs) have gained significant usage in various kinds of missions, including reconnaissance, search and rescue, and military operations. In rescue missions, timely detection of missing persons after avalanches is crucial for increasing the chances of saving lives. Using UAVs in such scenarios offers benefits such as reducing risks for rescuers and accelerating search efforts. Employing a formation of multiple drones can effectively cover a larger area and expedite the process. However, the challenge lies in achieving autonomous and scalable systems, as drones are typically operated on a one-to-one basis, requiring a large team of rescuers. To enhance situational awareness and distribute communication load, this paper proposes a decentralized Kalman filtering algorithm that exploits sensor data from multiple drones to estimate target positions and support guidance and control algorithms. The algorithm combines Consensus on Information and Consensus on Measurements techniques. Preliminary validation is conducted through numerical simulations in a sample scenario.