Frequency tracking technology for frequency-hopping signals using particle fiters

李世鹏,赵正予,黄硕,陈罡

DOI: https://doi.org/10.13245/j.hust.2011.09.009

2011-01-01

Abstract:To estimate the frequency of frequency-hopping signals in real-time without the knowledge of its hopping pattern or transition time, a phase particle filter algorithm was put forward for frequen- cy tracking. The stochastic model of frequency-hopping signalsr phase was proposed. The estimation of posterior probability density of signal phase was realized with particle filter algorithm. Finally the minimum mean square error （MMSE） estimation of hopping frequency was obtained with the posterior probability density of signal phase. The particle weights were recursively computed using sequential importance sampling （SIS）. Systematic resampling was employed for the reduction Of particle weights degeneration. Simulation results demonstrate the robust frequency tracking ability of the proposed method. In addition, compared with the reassigned smoothed pseudo Wigner-Ville distribution （RSPWVD） algorithm or the auxiliary spectrogram particle filter （ASPECT-PF） algorithm, the pro- posed method has a lower estimation error with high signal-to-noise ratio.