DOA Estimation Based On Matrix Spatial Prefiltering Approach

The spatial filtering method and the direction-of-arrival (DOA) estimation algorithm are both crucial research contents in the field of array signal processing, which have been applied widely in sonar, radar and other areas. A spatial filter is a processor that passes the signal from a specific direction which is preferred and attenuates those from the dispreferred directions, arousing wide attention in scholars. Both the spatial filtering and the DOA estimation algorithms are concerned as main topics in this thesis with research contents as follows.First of all, based on the array signal mathematical models, three criterions for spatial filtering design are studied with corresponding solutions. The simulation analysis show that the matrix spatial filter can pass the signal from preferred angular sector without distortion, and attenuate the interference and noise from other sectors, which lowers the resolution threshold. Besides, the filter performance would be influenced by the changes of parameters including array-number, center position of passband and the range of passband. Therefore, it is needed to choose these parameters of filter appropriately according to the real situation.Subsequently, regular DOA estimation algorithms have disadvantages of poor robustness and high signal noise ratio threshold. Therefore, the spatial matrix filter is adopted as a prefilter to DOA estimate in this thesis, improving DOA estimation performance of the system effectively. By analyzing simulation results, compared with regular DOA estimation algorithms, the DOA approach based on spatial prefiltering can efficiently attenuate the spatial interferences and provide low resolution threshold, which improves the estimation and resolution capability for the sources located inside the passband sector.At last, a matrix filtering based on non-uniform linear arrays is proposed in this thesis. This new filter is an optimal non-uniform linear array matrix filter based on the genetic algorithm, which can be used as the preprocessor in the DOA estimation. It is verified from the simulation results that the new filter owns decent passband response and can pass the signal from the preferred directions with no distortion, and the DOA method based on the non-uniform linear array matrix prefilter can distinguish source direction accurately. In a word, this DOA estimation algorithm can get its capability improved greatly compared with the regular DOA algorithm.