Localization and tracking of multiple near-field sources using randomly distributed sensors

The performance of the maximum likelihood (ML) estimator using fast and adaptive simulated annealing methods has been investigated for near-field localization of multiple narrowband sources. The estimation results are found to be close to the Cramer-Rao lower bound (CRLB). The effect of unknown parameters such as the source frequency, amplitudes and path loss factors is studied. The performance is found to improve with the increase in the signal-to-noise ratio (SNR) and snapshots. It is observed that nonlinearly quantized received samples can be successfully used in the ML estimator with little compromise in localization performance. An increase in the number of sensors, beyond a certain threshold number of sensors located closer to the sources, is found to provide only a marginal improvement in location estimation performance.