A higher order time-varying moment method for multicomponent chirp signals parameter estimation

Parameter estimation for nonstationary chirp (LFM) signal is a noticeable signal processing problem that shows up in many applications. A special time-varying higher order moment approach is proposed here to realize multicomponent chirp signals parameter estimation. Though chirp signal is nonstationary, its special time-varying fourth order moment in this paper is a function independent of time. Namely this special fourth order moment presents a stationary character. Then a cross fourth order moment MUSIC method is applied to do the following signal processing. Because this technique needs only one dimension search, it is superior to ML method and time-frequency distribution methods (which need two dimension search) in computer burden greatly. Even working in colored noise and relatively low SNR phenomena, the method here is very accurate, highly reliable, and can operate efficiently