Research On TDE And DOA Estimation Methods In The Presence Of Impulsive Noise

Time delay and derection of arrival are two basic parameters in the radio communication, other correlative important imformations of the signals can also be acquired from these parameters. For a long time, time delay and derection of arrival are always the hot research topice in signal processing fields, these have great theoretical and practical significance.The traditional algorithms to estimation the time delay and the arrival angle of signal are based on the second order statistics or high order statistics, and the Gaussian signal model is utilized. This Gaussian signal assumption is reasonable and convenient for research under many circumstances, but there are a lot of non-Gaussian signals in the practical appliance. For instance, echo of radar, low-frequency atmospheric noises, underwater acoustic signals and noises produced by human, there signals have the notable pulse characteristic. If the traditional algorithms based on the Gaussian model are used in the impulsive noise environment, the performance of which will be declined significently. Alpha-stable distribution can primely depict stochastic signals with a remarkable impulsive characteristic. Besides, the Gaussian distribution is a special case of alpha-stable distribution, so alpha-stable distribution noise model has better applicability.For this reason, firstly, this thesis analysed the character of alpha-stable distribution and related fractional lower order moment, discussed the reasons for the performance degradation of the traditional algorithms, researched the singal path time delay estimation based on fractional lower order covariance. Through integrating non-linear transformation, weighting technique and the adaptive processing technique, the SCOT weighted adaptive time delay estimation algorithm baesd on the the Sigmoid non-linear transformation is proposed. The computer simalations shows that this algorithm can give accurate estimates even if the statistics prior knowledge of signal and noise are unkown.Then, based on the multipath model, this thesis analysed and researched the time delay algorithm based on the weigthed fractional lower order covariance and the weigthed non-linear transformation.,it is shown by computer simulation that the accuracy of the estimations was greatly improved.Finally, considering that the coherent signals will result in serious performance degradation of the DOA estimation algorithm based on the FLOM, an efficient DOA estimation algorithm in impulsive noise environment based on forward-backward spatial smoothing technology, matrix remodeling technology and integrating FLOM and MUSIC method is proposed. The experimental results expressed that the performances of the algorithm proposed by this thesis is very conspicuously superiority, and this algorithm has many advantages such as good stability, high accuracy and doing an efficient estimation under low signal to noise ratio.