Trajectory prediction for low-cost collision avoidance systems

In this paper, a novel algorithm for estimation, filtering and prediction of glider and light aircraft trajectories based on GPS measurements is introduced. The algorithm uses Interacting Multiple Model (IMM) filters to detect specific maneuvers such as turning, circling or straight flight. An integrated wind model allows for quick estimation of local wind fields and helps achieving consistent prediction quality in windy conditions. The algorithm is shown to perform well compared to algorithms currently used in the FLARM^reg collision avoidance system, particularly in windy conditions.