Attitude and position estimation for an UAV swarm using consensus Kalman filtering

This paper presents the application of a distributed attitude and position estimation algorithm to a swarm of cooperating UAVs with heterogeneous sensors on board. The algorithm, based on a Consensus Extended Kalman Filtering (CEKF) to account for nonlinearities, is implemented assuming kinematic relationships. Numerical simulations are presented on different flight scenarios to evaluate the benefits of dealing with prior and novel information in a separate way on the basis of recent theoretical results on CEKF. Inertial and vision sensors are supposed to be mounted on board of the aircraft. Realistic flight scenarios are analyzed in the light of possible time communication delays among the agents.