Flight - Mode - Based Aircraft Conflict Detection using a Residual - Mean Interacting Multiple Model Algorithm ∗

Based on the trajectory prediction error model proposed by Paielli and Erzberger, we propose nominal and probabilistic conflict detection algorithms using flight mode estimates as well as the aircraft current state estimates. This is different from previous conflict detection algorithms which use current state estimates only. Our algorithms are therefore based on hybrid models of aircraft, which allow for both continuous dynamics and discrete mode switching. To obtain accurate state and mode estimates, we propose a modified version of the Interacting Multiple Model (IMM) algorithm designed by Bar-Shalom et al. called the Residual-Mean Interacting Multiple Model (RMIMM) method. RMIMM is a multiple-model-based estimation algorithm based on a new likelihood function which uses the mean of the residual produced by each mode matched filter (usually Kalman filter), producing better mode estimates, and therefore better state estimates, than in the IMM case. We demonstrate our algorithm on multiple aircraft scenarios, and in the latter part of the paper, the probabilistic conflict detection algorithm is combined with the protocol-based conflict resolution algorithm, designed by the authors in earlier work.