A Consensus-Based Kalman Filter for Target Localization in Emergency Scenarios

In the last years, Unmanned Aerial Vehicles (UAVs) have been widely used for several types of missions, including aerial reconnaissance, search and rescue, and military operations. In first rescue missions, the ability to detect missing people after avalanche events in a short time is fundamental to increase the probability of saving them. The use of unmanned aerial vehicles in such scenarios can improve results with respect to multiple points of view: on the one hand, it can decrease the danger for rescuers, while on the other hand, it can speed up the search process. An effective solution can be the involvement of a formation of multiple drones to cover a greater research space and accelerate the process. However, an important challenge in deploying a formation of robots in emergency scenarios is the autonomy in terms of system scalability, since drones are usually teleoperated in a one-to-one ratio with operators, requiring a large crew of rescuers. In order to improve the situational awareness and distribute the communication burden, this paper deals with a decentralized Kalman filtering algorithm using sensor data from multiple drones to estimate a target position, besides UAVs state to support guidance and control algorithms. Such decentralized Kalman filtering algorithm combines the characteristics of Consensus on Information and Consensus on Measurement techniques. The proposed technique is preliminarily validated by means of numerical simulations on an example scenario.