Airborne multisensor tracking for Autonomous Collision Avoidance

This paper presents the tracking algorithms developed for a multisensor anti-collision system for Unmanned Aerial Vehicles. This system will be developed by the Italian Aerospace Research Center (CIRA) within a research project named TECVOL, funded in the frame of the National Aerospace Research Program (PRO.R.A.) on UAV. The hardware setup is composed by a pulsed radar, two infrared cameras, and two visible cameras used as aiding sensors, thus the adoption of a fusion algorithm was mandatory to obtain the most accurate and reliable tracking estimate of obstacles. The paper describes the different modes and the relevant attainable performances of the developed tracking algorithm. The adopted data fusion technique for tracking is the Kalman filter. In particular, three different algorithms are compared in a typical collision scenario, namely Conventional Filter in Rectangular Coordinates, Conventional Filter in Spherical Coordinates, and Extended Filter in Rectangular Coordinates. Though all the three algorithms exhibited satisfying performances, the Extended Filter in Rectangular Coordinates resulted the most adequate for this airborne application.