A particle filtering approach to FM-band passive radar tracking and automatic target recognition

We present two stochastic filters for an FM-band passive air surveillance radar. The first system uses an extended Kalman filter and delay-Doppler measurements to track targets. The second system uses a particle filter to simultaneously track and classify targets. Automatic target recognition is made possible by the inclusion of radar cross section (RCS) in the measurement vector. The extended Kalman filter cannot take advantage of radar cross section measurements because there is no closed-form relationship between the state elements which determine target aspect and the resulting RCS measurement. We believe that this is the first work to propose the use of RCS for the purpose of target recognition within a passive radar system. We also present many simulation results for a challenging 2-target 3-sensor task involving trajectories which nearly coincide for a portion of their track length.