Efficient Localization Algorithms Using a Uniform Rectangular Array with Model Imperfections

Conventional localization algorithms, including maximum likelihood (ML) and multiple signal classification (MUSIC), are significantly influenced by array model imperfections. To address this problem, two generalized algorithms are proposed in this paper by extending the two conventional algorithms. The presented techniques are evaluated within the context of a passive uniform rectangular array (URA) where a rectangular subarray is assumed to be calibrated perfectly, while the remaining antennas incur array model errors. In this case, the performance of the conventional algorithms degrades seriously or even the operations fail. Nevertheless, the proposed algorithms can eliminate this issue by employing a separation technique. In specific, a separation strategy is applied to the introduced algorithms to automatically selects signals belonging to the calibrated subarray from the received signal matrix and eliminate the signals belonging to the uncalibrated antennas prior to conducting the azimuth-elevation-Doppler estimation. In order to reduce the computation complexity, the azimuth-elevation-Doppler estimation is divided into two sub-problems: the estimation process of the Doppler frequency and the estimation process of the azimuth-elevation angles. The simulation results are shown to demonstrate the efficiency and superiority of the proposed algorithms compared to the conventional algorithms.