On the Limits of Collision Detection Performance of a Sense-and-Avoid System for Non-Cooperative Air Traffic

Abstract

Unmanned Aircraft Systems (UAS) face limitations on their utilization in civil airspace because they do not have the ability to Sense and Avoid (SAA) other air traffic. In recent years, there has been growing interest to provide an effective SAA solution for UAS operations. An effective SAA solution must address both cooperative as well as non-cooperative air traffic. A number of different sensor solutions are being evaluated for SAA pertaining to non-cooperative traffic. Examples of such sensors include electro-optical (EO), on-board radar, passive acoustics, laser radar and ground radar. Using one or more such sensing modalities, it is possible to track a non-cooperative aircraft in the vicinity of the own aircraft otherwise known as an intruder. However, an intruder’s future trajectory is never perfectly known and a SAA system’s performance will always be limited by these uncertainties. One of the components of an SAA system is the logic to decide whether a certain aircraft is on a collision course with the own craft. It, therefore, follows that the performance of this collision detection component will also be limited by the uncertainties in the future trajectory of the intruder. In this paper, we investigate the problem of what is the best that a SAA collision detection system can perform in spite of the future uncertainties in the intruder trajectory.