The Research On The Doa Estimation And Adaptive Beamforming For The Polarization Sensitive Array

Polarization sensitive array(PSA) can obtain the polarization state of the incident electromagnetic wave, thus it possesses many performance advantages than the traditional scalar sensor array. PSA has become one of the hot research topic in recent years. However, the processing methods for PSA are mostly the direct expansion of the methods for scalar sensor array without too much utilizing of the inherent element structure and the performance of the PSA remains to be further improved. Therefore, the focus of this dissertation is on the methods of parameter estimation and beamforming for PSA. The inherent orthogonal structure of the crossed electric dipole pair is considered and a newly processing method based on the spatially non-collocated PSA is presented.For the parameter estimation of PSA, coherent signals and non-coherent signals are discussed simultaneously. The common Lv-Music and ESPRIT algorithms are presented for the non-coherent signals and the quaternion Music is also analyzed based on the orthogonal structure of the array element. Computer simulation results show that the performance of the quaternion signal model is better than that of the long vector signal model. For the coherent signals, the algorithms of spatial smoothing and polarization smoothing as well as the improved algorithms are both discussed in this dissertation. Simulation results indicate that the improved algorithm has lower root mean square error and better resolution than the regular algorithm. Also, the improved polarization smoothing algorithm is suitable for different forms of noise.For the beamforming of PSA, the principle and criterion of the joint space-polarization filtering are introduced and the different beamforming methods including regular long vector method and the quaternion method are both presented. The optimal weight vector under derivation constraints and zero constraints are derived by using the eigenspace-based method. And the influences between the beam pattern and the filtering performance are given through computer simulations. The quaternion based beamforming methods, including Q-MVDR and Q-APES algorithms, are derived theoretically by using the inherent orthogonal structure of sensor elements and the quaternion optimal weight vector is obtained following. Simulation results show Q-APES algorithm can achieve a stable beam pattern under the circumstances of high desired signal power, low sampling snapshots and coherent interference signals. At the same time, it is robust to the pointing errors of PSA.In order to alleviate the mutual coupling effect among the collocated elements of each sensor and reduce the hardware cost and the complexity of the algorithm, the spatially non-collocated array geometry is discussed in this dissertation. Based on the non-collocated polarization sensitive array, the newly methods of parameter estimation and beamforming are derived by distributing the elements of each sensor. The performance of the non-collocated array is compared with that of the conventional collocated array and the scalar sensor array.