Abstract:

In order to improve the performance of hop timing estimation when the frequencies of frequency hopping (FH) signals are not in the known frequency set, a novel method is developed based on the sparse Bayesian inference (SBI). The proposed method first sets the frequency bias in the signal model. The updating rules of the parameters are calculated based on the Dirichlet process and SBI, and then the frequency bias is utilized to correct the dictionary matrix. Finally, the sparse matrices and the hop timing estimation can be obtained. Simulation results demonstrate that compared to the existing methods, the proposed method can obtain higher correct ratios of the hop timing estimation and a better performance of the spectrum estimation.

Key words: frequency hopping, Dirichlet process, sparse Bayesian inference, sparse matrices