Bernoulli filter-based direct tracking of an emitter under unified far-field and near-field situations

This paper considers the problem of single-emitter tracking based on multi-receiver, each with a sensor array, which is traditionally processed sequentially through detection, parameter extraction, and tracking. However, in such a framework, the tracking performance is seriously affected by the accuracy of the preprocessing steps, and cannot be applied to low signal-to-noise ratio (SNR) scenarios. Besides, in the far field and near field cases, the signal processing methods are different. To solve such a problem, we develop an algorithm for direct tracking based on the received signal applicable for both near and far field cases. The sensor arrays are equipped to estimate the state of a target directly. In this paper, a direct tracking problem is raised under the framework of the random finite set (RFS). A new likelihood model is proposed by evaluating the covariance of the measurement signal and it is distributed as a complex Wishart. The proposed algorithm is implemented by a Bernoulli particle filter, and the simulation results show performance improvements at low SNR.