A frequency estimation method of DFH signal based on periodogram and TLS

A method based on the combination of the periodogram and the ARMA modeling-total least squares (TLS) algorithm, is proposed for frequency estimation of differential frequency hopping (DFH) signals in Gaussian noise. The spectrum of the signal is first evaluated using the periodogram method. The frequency estimation can be directly obtained by peak-value-searching of the spectrum, if the frequency points are spaced far apart. When they are closely spaced, the singular value decomposition (SVD)-TLS algorithm is then employed for accurate estimation. Based on the above principles, this method enjoys the advantages of both, i.e., low computational complexity, simple implementation and high accuracy. This scheme can also avoid false frequency point estimation which might appear when only the SVD-TLS algorithm is employed. Simulation results verify the feasibility and effectiveness of the method.