Research On Key Technology And Application For Blind Adaptive Beamforming

Nowadays, increasing demands for communications have placed higher requirements for data rates, system capacity and transmission reliability. Antenna arrays which can distinguish signals with spatial difference are becoming more and more popular. Adaptive beamforming technique is one of the key techniques of array signal processing, which allows to transmit or receive signals from a specific direction in the presence of interference or when the system environment varies slowly. However, when it comes to the large-scale antenna array, the traditional full-rank algorithm suffers from a slow convergence speed as well as high computational complexity. Besides, the conventional algorithms are extremely sensitive to errors, the performance degrades a lot in the presence of steering vector mismatches. In this paper, we focus on the reduced-rank technique for large-scale antenna array and robust algorithms specific for mismatches. Further, we extend the spatial filter design to a space-time adaptive processing in the GPS systems. Based on the alternate updating idea, we propose an alternating low-rank decomposition algorithm. The main contributions are summarized as the following items:1) We propose a blind adaptive CCM reduced-rank beamforming algorithm for large-scale antenna arrays, where the transformation matrix is constructed based on the Krylov subspace. The widely linear (WL) processing which can fully exploits information of the complementary correla-tion matrix is utilized to further improve performance. Considering the block conjugate symmetry property of the augmented correlation matrix, we design a structured algorithm which reduces the computational operations by almost 50 percent compared with the direct WL algorithm. Moreover, we carry out detailed analysis in terms of the convergence of reduced-rank filter, the steady-state MSE as well as the computational complexity. Simulation results verify the correctness of our analysis and further demonstrate that our proposed algorithm yields better performance compared with the other analyzed algorithms.2) We propose a robust adaptive beamforming algorithm based on the worst-case optimization and WL processing for the model mismatch scenario. Firstly, we clarify the impact of steering vector mismatch and distortion of signal constellation on the augmented steering vector in WL case. Then, we construct a worst-case optimization cost function. Based on the alternating updation idea, we design a modified conjugate gradient algorithm as well as a steepest gradient algorithm to alternately update the filter coefficients and the diagonal loading factor, respectively. Numerical results are carried out to verify the performance of the proposed algorithm.3) We extend the simple spatial filter technique to the space-time adaptive processing, and novelly propose an alternating low-rank decomposition (ALRD) algorithm for interference cancel-lation in the GPS systems. In construction of the transformation matrix, we restrict the number of non-zero elements of each projection vector, so as to force it sparse. Then, the proposed algorithm alternatively update D low dimensional projection vectors plus one reduce-rank filter. We provide detailed derivation and computational complexity analysis for the proposed algorithm. Simula-tion results indicate that our proposed algorithm outperforms other analyzed algorithms with faster convergence speed and lower computational complexity simultaneously. Beampattern related to spatial-time response demonstrates the advantage of the ST processing compared with the simple spatial filter.