Experimental Validation of Multi-UAV Applications: Formation Flight and Decentralized Target Estimation

This paper describes the experimental validation of two multi-UAV applications in an indoor environment. In particular, a formation flight task and a decentralized estimation procedure are analyzed for quad-copter platforms. First, the experimental setup is discussed, focusing on the hardware chosen for the navigation, control, and guidance layers of the UAVs. For the first application, a flocking protocol previously designed by the authors is implemented, proving satisfactory performance in terms of inter-agent distance and transient behavior. For the second application, a decentralized Kalman Filter is employed to collaboratively estimate the position of an ArUco marker. The experimental results show a successful information fusion by sharing the UAVs' measurements performed by onboard cameras.