Adaptive Beamforming Algorithm Based On Fourth-order Cumulants

Adaptive beamforming technology is one of the main research fields of array signal processing.Its applications include radar,communication,navigation,seismic exploration and biomedicine.In practical application,the traditional adaptive beamforming algorithm is not good enough in signal enhancement and interference suppression,and its performance will be greatly reduced when dealing with coherent signals and non-Gaussian signals.In this paper,based on the traditional beamforming algorithm,the higher-order cumulant combined with adaptive beamforming,join the coherent processing,make the algorithm has better filtering performance and can deal with coherent signals and non-gaussian,and aiming at the existence of higher order cumulant of high computational complexity problem,put forward a kind of fourth-order cumulant based on beam space beamforming algorithm.The algorithm focuses on the three aspects of decoherence,reduction of computational complexity and concrete implementation.The main contents are as follows:Firstly,in view of the failure of traditional beamforming algorithm under the condition of coherent source,the received signal is decoherence processed and the signal covariance matrix is reconstructed.Based on the traditional Toeplitz,only the cross-correlation function of the sampled data is used for reconstruction.The first row of the data is used as the cross-correlation function with the vector elements of other rows,and the obtained matrix is conjugate rearranged.The weighted average matrix of the two matrices is taken as the covariance matrix of reconstruction.This decoherence algorithm can not only effectively remove the coherence between signals,but also greatly reduce the computational complexity of reconstruction.Aiming at the problems of high computational complexity and poor applicability in practical application of high-order cumulant adaptive beamforming technology,a fourth-order cumulant LCMV algorithm based on beamspace was proposed.On the basis of the decoherence algorithm,the data received by the array element is transformed into the beam domain,and the dimension of the processing matrix is reduced without affecting the data correctness.Fourth-order cumulant correlation matrix is constructed in the beam space,and the construction process is further deduced to reduce the complexity of construction,and the formula for solving the optimal weight vector is derived.The matrix inversion module adopts the in situ substitution algorithm based on the principal component exchange,and introduces the reduced coefficient matrix.On the principle of in situ substitution,the inverse operation is completed through the principal component exchange and the diagonal correction operation,and the whole matrix inversion has a high degree of parallelism.Through simulation,the advantages of the algorithm in enhancing the desired signal,weakening the interference signal and expanding the array are verified.By analyzing the computation amount of inversion operation and matrix multiplication operation,it is shown that the algorithm can effectively reduce the computational complexity,reduce the computation amount,and improve the practicability and parallelism.Finally,the correctness and effectiveness of the algorithm are verified on the beamforming system platform.A hardware experiment platform based on FPGA suitable for algorithm realization is built.The composition of each module of the experiment platform is introduced,and the specific realization method of data processing module is expounded emphatically.Finally,the overall algorithm is simulated and verified,and the desired signal is obtained,which shows that the proposed algorithm can effectively suppress the coherent interference and complete the adaptive filtering.