Target localization in passive and active systems : performance bonds

The main goal of this dissertation is to improve the understanding and to develop ways to predict the performance of localization techniques as a function of signal-to-noise ratio (SNR) and of system parameters. To this end, lower bounds on the maximum likelihood estimator (MLE) performance are studied. The Cramer-Rao lower bound (CRLB) for coherent passive localization of a near-field source is derived. It is shown through the Cramer-Rao bound that, the coherent localization systems can provide high accuracies in localization, to the order of carrier frequency of the observed signal. High accuracies come to a price of having a highly multimodal estimation metric which can lead to sidelobes competing with the mainlobe and engendering ambiguity in the selection of the correct peak. The effect of the sidelobes over the estimator performance at different SNR levels is analyzed and predicted with the use of Ziv-Zakai lower bound (ZZB). Through simulations it is shown that ZZB is tight to the MLEs performance over the whole SNR range. Moreover, the ZZB is a convenient tool to assess the coherent localization performance as a function of various system parameters. The ZZB was also used to derive a lower bound on the MSE of estimating the range and the range rate of a target in active systems. From the expression of the derived lower bound it was noted that, the ZZB is determined by SNR and by the ambiguity function (AF). Thus, the ZZB can serve as an alternative to the ambiguity function (AF) as a tool for radar design. Furthermore, the derivation is extended to the problem of estimating target’s location and velocity in a distributed multiple input multiple output (MIMO) radar system. The derived bound is determined by SNR, by the product between the number of transmitting antennas and the number of receiving antennas from the radar system, and by all the ambiguity functions and the cross-ambiguity functions corresponding to all pairs transmitter-target-receiver. Similar to the coherent localization, the ZZB can be applied to study the performance of the estimator as a function of different system parameters. Comparison between the ZZB and the MSE of the MLE obtained through simulations demonstrate that the bound is tight in all SNR regions