Research On Weak Underwater-target Detection Under Strong Interference

In this paper,the uniform linear array is used as the model to study the weak target detection method under strong interference.There are two main methods of strong interference suppression: matrix filtering and adaptive beamforming.This paper focuses on these two aspects as follows:Because the parameters of conventional matrix filter(independent of data)can not be adjusted online with the interference signal strength,the energy of strong interference signal may leak to the region of interest,which affects the detection of weak target signal.In this paper,by constructing the subspace of signal plus interference,matrix filtering is carried out in the subspace.Since the subspace transformation is equivalent to normalizing the power of each source,even if the interference signal is strong,it can be effectively filtered out.At the same time,the eigenvalue distribution of the filtered target signal and interference signal is analyzed,and the threshold of eigenvalue detection is designed to judge the filtered signal,which effectively improves the detection probability of weak target under strong interference.The adaptive beamforming method is very sensitive to the error of guidance vector and covariance matrix.This paper focuses on these two aspects to improve the robustness of adaptive beamforming.In view of the influence of steering vector error,a two-step optimization robust beamforming method of steering vector based on rank-1 decomposition is proposed.The subspace of the preset steering vector of the target signal is constructed by using the possible angle interval of the target signal,and the actual steering vector of the target signal is projected on the subspace.The first step is to estimate the projection vector of the target signal in the subspace,and the second step is to estimate the projection vector of the target signal in the subspace Based on the projection vector,the actual guidance vector is estimated,and the solution is transformed into a convex optimization problem which is easy to calculate by using the rank-1 decomposition theorem.The simulation results show that the method can correct the error of the guidance vector,reduce the phenomenon of signal self cancellation,ensure the ability of interference and noise suppression,effectively improve the output signal to interference ratio of the beamformer,and improve the ability of weak target detection.In view of the influence of covariance matrix error,this paper proposes an estimation method of interference plus noise covariance matrix based on generalized loading.The load matrix is constructed by the noise gain of the beamformer to reduce the proportion of the target signal in the overall covariance matrix.At the same time,with the optimal guidance vector estimation method based on the generalized sidelobe canceller,the closest one to the actual guidance vector is estimated Superior guidance vector.The simulation results show that this method is consistent with the diagonal loading method in improving the covariance matrix error performance,but it avoids the disadvantage of difficult selection of diagonal loading amount.At the same time,with the optimal guidance vector estimation,this method also has a strong correction ability for the guidance vector error,effectively improves the robustness of adaptive beamforming,improves the output signal-to-noise ratio,and enhances the ability of weak target detection under disturbance.