Maximum Likelihood Direct Position Determination of Multiple Sources with Channel State information

Abstract

This work presents a maximum likelihood direct position determination (DPD) technique for multiple stationary radio sources emitting or reflecting identical waveforms. Conventional importance sampling (IS)-based DPD approaches are designed under the assumption of large distances between the sources, thus suffering from accuracy degradation when some of the sources are closely spaced. To this end, using channel state information (CSI), we propose a novel IS-based DPD algorithm by introducing a source classification step before generating position realizations. A more appropriate importance function is also developed. Specifically, rather than processing all sources independently, the new importance function assumes that the closely-located radio sources are uniformly distributed around a central point. The simulation results validate the robustness of the developed importance function. Furthermore, the results clearly show that the proposed algorithm can achieve a better estimation accuracy than state-of-the-art DPD techniques.