Demonstrating significant passive radar performance increase through using known communication signal format

In order to determine if passive radar performance can be good enough to allow passive radars to take the place of active radars, we seek the ultimate limits of passive radar performance. The focus is on the case where the radars are employed for estimating the position and velocity of a target. Further, the passive radar employs several dispersed existing communication transmitters, which could be cellular base stations. We compare the Cramer-Rao bound (CRB) for an active radar with the CRB for a passive radar under various amounts of prior information concerning the transmitted signals. The mean-square-error of maximum likelihood estimation is presented to provide the threshold at which the estimation performance approaches the CRB. One important result shows that if we know the employed transmitted signals are of known form, for example those coming from GSM base stations, but contain unknown bits, then under reasonable models the passive radar can estimate the bits and ultimately achieve target estimation performance that is very close to that of the active radar. Estimation performance for a number of other interesting cases with less information shows considerable performance loss.