Skip to Main Content
This paper addresses the problem of localization and tracking multiple non-cooperative objects using only passive bearing sensor data. The challenges in this context lie in an unknown number of objects, false alarms and clutter measurements. To avoid the time consuming data association and data storage, an iterative approach, which only considers the sensor data from the actual timestep for an update of every object state, is preferable. Our approach to perform this is a Monte Carlo realization of a probability hypothesis density filter. In this context we use bearing data gained from an antenna or optical camera mounted on an airborne observer. Tests on simulated and real world scenarios show that our approach leads to a stable localization and tracking of multiple targets, even in the presence of clutter and misleading bearing measurements.