Frequency-Hopping Signal Parameters Estimation Based on Orthogonal Matching Pursuit and Sparse Linear Regression

This paper addresses the problem of estimating frequency hopping (FH) signals parameters with the single array. The existing optimization algorithms are with unrealistic computational burden. In order to reduce the computation burden, a novel method, based on orthogonal matching pursuit and sparse linear regression (OSLR), is proposed in this paper. The OSLR method consists of two steps. First, by segmenting the received signals into several measurements, orthogonal matching pursuit is used to detect whether the segments include hop timings or not. Second, sparse linear regression is used to estimate the spectrogram of FH signals for the segmentations with hop timings. Numerical simulations demonstrate that the OSLR method can achieve superior performance, and the OSLR method can be exploited to deal with underdetermined blind source separation problem of FH signals.