Distance at Closest Point of Approach for Airborne Collision Avoidance

This paper focuses on algorithm description and test results from an Obstacle Tracking algorithm based on a Particle Filter developed in Spherical coordinates, aimed at demonstrating Unmanned Aerial systems (UAS) autonomous collision detection capability in terms of reliable estimation of Distance at Closest Point of Approach (DCPA). In fact, in the framework of UAS Sense and Avoid problem, the estimate of DCPA constitutes a fundamental requirement for the collision risk assessment. Since assesses techniques, such as Extended Kalman Filter (EKF), can cause some loss of accuracy in obstacle tracking performance in case of non-linearities in obstacle dynamics, innovative methodologies are expected to provide more accurate estimates of tracking performance. In particular, Particle Filter technique is likely to improve the estimate of DCPA and potentially reduce the delay in collision detection. The developed software has been tested in off-line simulations based on flight data gathered during a test campaign conducted with a very light aircraft in the framework of TECVOL project. In particular, the algorithm performances have been evaluated in radar-only configuration since the main interest was addressed to the analysis of the impact of innovative technologies on tracking capabilities.