Study On HF TDOA Estimation And Target Localization

With the establishment of the third generation of HF communication standards and the development of modern communications and micro-electronics, HF communications has gradually become an important means of radio communication. Illegal HF signal monitoring and source localization have become one of the most important contents in radio monitoring.Traditional direction finding HF localization technology has many problems and limitations.It is important to develop a new TDOA-based technique in HF localization.Doppler frequency shift and spread exhibit modulating effect on HF signals when propagating from ionosphere, declining the performance of traditional time delay estimation(TDE) algorithm in dealing with HF signals. Meanwhile, HF TDOA equation is a highly complicated nonlinear equation under ionosphere and geodetic coordinates, traditional TDOA positioning algorithms are limited in dealing with this kind of TDOA equation. In this thesis,several HF TDOA estimation methods and particle filtering based HF TDOA positioning algorithm have been proposed with details as follows:Firstly, HF TDE model was established with Watterson HF channel model. Based on the analysis of channel gain function’s effect on basic correlation TDE algorithm, this paper proposed HF modulus basic correlation and wiener weighted correlation TDE algorithm.These two algorithm can solve HF TDE problems and have high anti-noise performance.Secondly, particle filtering TDE algorithm was proposed with HF autocorrelation function and modulus cross-correlation function. This algorithm has limitations in anti-noise performance but strongly applicable.Thirdly, based on correntropy theory, a complete mathematical proof of the generalized correntropy function(GCF) algorithm, and the equivalence of the maximum correntropy criterion and the minimuTn dispersion criterion have been derived under sine kernel function.Thus this thesis proposed modulus GCF algorithm to solve HF TDE problem under alpha-stable distribution noise. Modulus GCF algorithm is sensitive to channel gain function but robust to various noise and resents a verood erformanceFinally, by establishing HF propagation model and TDOA equation,the particle filtering HF TDOA positioning algorithm, Parzen PDF(Possibility Density Function) estimation based particle filter algorithm were proposed. These two algorithms can accurately estimate HF source location, and the real HF data test showed high localization accuracy and correctness.