Direction Finding For Wideband Signals Based On Sparsely Decomposition

The estimation methods of traditional array signal parameters were based on the assumption that the signals to be dealt with were stationary narrowband signals. Their performances were proved to be ineffective when they were applied to deal with wideband signals. Therefore, wide research interests have been aroused as soon as the theory of wideband signal's direction of arrival (DOA) was put forward.At present, the widely applied method in wideband signal DOA estimation is time frequency subspace method, which could realize the estimation of none-stationary signals' DOA by means of establishing spatial time frequency distribution (STFD) matrix and combining with the traditional array signal processing methods. However, this method will be affected by the selection of the correct time-frequency points and the impact of cross-term. Their estimation performance could turn to be acceptable only under the condition of higher signal-to-noise ratio. Furthermore, signals applicable with this method are too limited.The signal sparse decomposition is a new method for analyzing and processing signals. It is of many excellent characteristics. This paper has introduced the signal sparse decomposition into the field of wideband array signal processing, and put forward a new method of parameter estimation of wideband signals based on sparse decomposition. The main work and contribution of the paper are explained as follows:1. A typical time-frequency analysis method - Wigner-Villedistribution (WVD) is worked out. And the DOA estimation algorithm for wideband linear FM signals based on the time-frequency subspace method is discussed. Two methods for analyzing and processing signals are also presented: the traditional signal orthogonal decomposition method and sparse decomposition method. 2. A new method based on sparse decomposition is proposed to estimate the frequency and DOA of wideband linear FM signals. Firstly, it has established a over-complete atom dictionary according to the linear FM signals' frequency parameter and the array geometry, then a best matched atom can be found to obtain the DOA of sources by sparse decomposition of the array output. A faster algorithm is presented to solve the calculation problem of match pursuit (MP) algorithm. Simulation experiments have proved that the performance of the method based on match pursuit (MP) algorithm is much better than traditional STFD method.3. Having proposed a phase differential wideband signals DOA estimation algorithm. Simulation results have testified the effectiveness of the algorithm.