Least square single-tone frequency estimator in non-Gaussian colored noise
Zhigang Wang · Journal of Communications · 2008
A least square frequency estimator of single-tone sinusoidal wave in non-Gaussian colored noise was proposed. Firstly, the spectral line of symbol-rate-sinusoid to band nearby zero was transferred by a mixer. Then the transferred signal was filtered and decimated to get a new one in Gaussian white noise. Finally, fine estimation of signal baud was got by applying least square criterion to the phase difference of extracted sinusoid. Differing from other FFT-based algo- rithms, the new one overcomes the shortcoming of finite resolving capability in frequency and can provide more accurate estimation for automatic modulation recognition and other application.