Multiple Emitters Direct Tracking With Distributed Single-Sensors: Improved Particle Filter Based on Multiple-Frequency Fusion

Abstract

Tracking the trajectory of multiple moving wideband emitters with distributed single-sensors is addressed in this article. Traditional tracking methods typically begin by collecting the positions of emitters within each observation interval and subsequently derive the tracking trajectories through trajectory correlation. However, repeated localization and multitarget data association create tremendous complexity challenges for tracking. In this article, a direct tracking method dependent on the particle filter (PF) for multiple emitters within the same frequency band is proposed. The method integrates the received data from all sensors and utilizes data segmentation and multiband division to reduce computational complexity. Subsequently, a spectral function based on multiple-frequency fusion is constructed to replace the likelihood function of the PF, which enables effective utilization of all sensor data for calculating particle weights. Finally, the weights of particles are obtained by utilizing multiple frequency bands, enabling the tracking of multiple emitters based on multiple sets of particles. In comparison to state-of-the-art methods, the proposed method avoids trajectory correlation and demonstrates better tracking performance. The effectiveness of the proposed method is substantiated by both numerical and experimental results.