DOA Estimation Based On Fourth-Order Cumulants And Sparse Decomposition

Array signal processing is a very important issue of signal processing. It is widely applied in many fields such as radar, communication, sonar, earthquake, chronometer etc. As one of the main research field of array signal processing, spatial spectrum estimation has obtained a rapid development in theory. Due to its limited application system practice, spatial spectrum estimation becomes the main research subject at present.Sparse decomposition of signals is emerging as a new-born signal processing method which has many excellent features. Until currently, some references have already applied sparse decomposition of signals to solve the problem of direction-of-arrival (DOA) successfully. This paper applies sparse decomposition in array signal processing, and researches the method of DOA estimation algorithm. The main work and contribution of this dissertation are as follows:1. The principle of sparse decomposition is introduced. The Matching Pursuit (MP) algorithm and its detailed implementation process are analyzed. Then the application of sparse decomposition in DOA estimation is studied.2. Current existing DOV estimation algorithms based on sparse decomposition all have a problem. That is when its source signal form changes, its over-completed atom data base have to be rebuilt according to the source form of the signal. Actually, this increases the complexity of the algorithm. To solve this problem, we have proposed a DOA estimation algorithm based on MP array signal sparse decomposition with the thought of ESPRIT. This algorithm improved the building method of the over-completed atom data base as it doesn' t have to rebuilt according to the source signal form. So the complexity of the algorithm is reduced. Simulation results show that the estimation performance of the proposed algorithm under the zero mean additional Gaussian noise is better than the traditional MUSIC and the ESPRIT.3. If the mean of the additional Gaussian noise is not zero, the DOA estimation algorithm based on sparse decomposition performance decreases greatly. To solve this problem, we proposed a union high-order accumulation and sparse decomposition method for the DOA estimation. In order to restrain any Gaussian noise, this algorithm finds out the fourth-order accumulated matrix of the received array signal first. Then, it realizes signal's DOV estimation under nonzero environment based on the MP array signal sparse decomposition DOV algorithm. Simulation results show that the estimation performance of the proposed algorithm under nonzero mean additional Gaussian noise is better than the traditional MP algorithm.