SLAM-based Aircraft Positioning in Mixed LOS/NLOS with Prior Airport Map Information

Aircraft localization by wireless local area networks in airport areas is a crucial methodology for air traffic surveillance; however, it is challenged by signal reflections from buildings and obstacles such as runways, aprons and hangars. While multipath is typically considered an impairment in conventional Multilateration (MLAT), in this paper we propose to favorably exploit these reflections through a virtual representation of the physical anchors that receive reflected localization signals. The locations of the virtual anchors are to be estimated and depend on the environmental reflectors, i.e., on the wireless channel response. We consider first-order reflections, designing a localization algorithm that operates on available Time Difference of Arrival (TDOA) measurements, without necessitating the availability of the full multipath channel response. By estimating the Line of Sight (LOS) or Non Line of Sight (NLOS) conditions associated with each measurement, leveraging a belief propagation methodology, we enhance aircraft positioning by exploiting the locations of the Virtual Ground Stations (VGSs) associated to NLOS links and using them as anchors in the MLAT problem. Simulations in a 3D airport scenario with ray-tracing are used for validation, showing improvements against a conventional MLAT approach.