Research On Robust Adaptive Beamforming And Space-Time Adaptive Processing Algorithms

This dissertation is focused on the research of robust adaptive beamforming and space-time adaptive processing (STAP) algorithms against the steering vector error and the covariance matrix error. The key contributions are:1. For the robust beamformer under the norm constraint of the weight vector, the equality constraint is proposed, and its solution is given. Since the performance is determined by the constraint parameter, and its choice is analysed and the selecting bound is given. It is deduced that the norm constraint parameter is selected in the allowable bound, the performances vary unobvious respectively, but for the same parameter, the equality constraint has the better performance than the inequality constraint, namely the optimal negative loading has the best improvement2. For the robust beamformer under the uncertainty set constraint of the steering vector, its efficient solution is given. For the loading expression, the optimal loading level can be computed efficiently. The selection of the uncertainty set constraint parameter becomes simple, and the improvement can reach the optimization. It is deduced that the negative loading can obtain the optimal performance improvement, and the constraint parameter is selected as larger, the performance will be close to the optimization.3. For the STAP covariance matrix estimation under nonhomogeneous condition, the diagonal loading generalized inner products algorithm is proposed for the sample selection method, and the weight coefficient is given for the data weighting method. Their performances are analyzed particularly. Moreover, the unscented transformation is used to obtain the approximative covariance matrix in the nonhomogeneous condition. The nonhomogeneous impact is reduced effectively. Simulation results indicate that the proposed algorithm has the favourable performance for moving targets detection.4. Diagonal loading is used as a means to improve the robustness to the case of covariance matrix mismatch for STAP. The selecting method of the loading level is given, and the particular performance analysis indicates that the rational diagonal loading can improve the detection probability and the output signal-to-noise ratio.5. The robust STAP algorithm is proposed that using the worse-case performance optimization to overcome the steering vector and covariance matrix mismatches. Therein, the original algorithm is converted to the loading sample matrix inversion (LSMI) STAP algorithm equivalently, the optimal weight vector is obtained in the closed form, and the exact loading level of LSMI STAP is given. Simulation results attest the correctness and the validity of the proposed algorithm.6. By analyzing the FRACTA algorithm and its enhancements deeply, a new improved algorithm: FAGATA is proposed. It reduces the reiterative censoring to once via multi-global outlier censoring algorithms, and also reduces the detection levels from three to two. Therefore, the computational efficiency is improved greatly. It not only detects the targets and estimates the parameters accurately, but also has higher detection efficiency and robustness. The simulation attests its correctness and effectiveness.