Dynamic Spectrum Tracking of Multiple Targets With Time-Sparse Frequency-Hopping Signals

Time-sparse frequency-hopping (FH) signals detection and sequences identification present a significant challenge in spectrum tracking of multiple targets. The non-continuous observations that arise from their temporal sparsity complicates identification efforts in low signal-to-noise ratio (SNR) environments. In this letter, a dynamic temporal perception probability hypothesis density (DTP-PHD) filter for spectrum tracking of multiple targets with potential periodicity was proposed by leveraging the periodicity alignment likelihood ratio (PALR). The PALR enables the estimation of the time transition function of targets' FH signals, which also facilitates the extraction of the spectrum track by identifying each target using a joint posterior intensity. Moreover, a closed-form solution of DTP-PHD was derived under linear Gaussian assumptions. The validity of the periodicity estimation was established by implementing a particle version of the proposed algorithm, which demonstrated robust tracking performance in noisy environments.