A Multi-Target Track-Before-Detect Particle Filter Using Superpositional Data in Non-Gaussian Noise
share
In this paper, we propose a general and tractable approach to multi-target track-before-detect based on the particle filter. It is even applicable to general superpositional sensor signals, and/or in the presence of non-Gaussian observation noise. Superpositional sensor signals depend on the sum of general nonlinear target contributions, and arise in diverse domains, such as radio-frequency (RF) tomography, wireless communications, and array signal processing. Moreover, the proposed method realizes MTT for an unknown, time-varying number of targets, in an online manner, without knowing their initial states. We conducted a simulation involving superpositional sensor signals in the context of RF tomography. The proposed method was shown to outperform the state-of-the-art approximate cardinalized probability hypothesis density filter for superpositional sensor signals (Σ-CPHD) in terms of the optimal subpattern assignment (OSPA) metric by a factor of approximately two to five.
READ FULL TEXT