Research On Robust Beamforming Algorithm Based On General Linear Combination

The basic purpose of the array signal processing is that sample and process the signal in spatial and temporal, exploit the information more fully; so that to suppress the interference effectively and improve the efficiency of the system. Adaptive beamforming technology is an important research direction of the array signal processing. Because of there are various errors and no ideal factors in the practical applications, such as signals and environment factors, system factors and the algorithm itself insufficient. These will affect the performance of system seriously. Therefore, improving the robustness of adaptive beamforming algorithm has been a hot spot of the research.As we all know, the diagonal loading algorithm is a simple and efficient method for adaptive beamforming. However, there are usually no clear guidelines on how to choose the diagonal loading factor reliably. In this thesis, we present an algorithm that can compute the diagonal loading factor from the receive data, which based on the general linear combination technique. The algorithm uses a priori knowledge, through revise the sample covariance matrix to obtain the value of diagonal loading factor. In order to avoid the computation of inverse covariance matrix, the weight vector is obtained by recursive method. Simulation results demonstrate that the proposed algorithm provides excellent robustness against signal steering vector mismatches, has a low computational complexity and makes the mean output SINR consistently close to the optimal one.In case there are signal steering vector mismatches and small training sample size, the performance of CLMS algorithm will decline seriously. To solve this problem, scholars have put forward worst-case performance optimization CLMS algorithm. Although the algorithm provides well robustness against signal steering vector mismatches, the computation of Lagrange multiplier is more complicated, so that increase the computational cost. At the same time, the sample covariance matrix is prone to mismatch in the practical application, serious affect the performance of the algorithm. Thus, we present the worst-case performance optimization CLMS algorithm which based on the GLC technique. We utilize the GLC technique to amend the sample covariance matrix, to decline the diffusivity of noise eigenvalue, so that to obtain the optimal diagonal loading factor. Establish the relationship of Lagrange multiplier and diagonal loading factor, and then obtain the optimal estimate of Lagrange multiplier. In order to reduce the algorithm complexity, we adopt the recursive method to get the weight vector; for the sample covariance matrix mismatches, we utilize the modified covariance matrix method based on GLC technique to solve the range of Lagrange multiplier and prove the uniqueness. Simulation results demonstrate that the proposed algorithm with fast convergence rate and stable performance; effectively reduce the computational complexity while ensure the robustness against signal steering vector mismatches and the covariance matrix mismatches, provides a nearly optimal performance.