Interference detection and robust mitigation method based on tensor decomposition

With the purpose of improving the robustness of the interference mitigation, methods using tensor operations are explored in this paper. Specifically, a three-order tensor-based signal model is constructed to cater to the three-dimensional characteristics of the signals received by the electromagnetic (EM) vector-sensor array. The polarized diversity of the signals is considered, in addition to the temporal and spatial information. Furthermore, canonical polyadic decomposition (CPD) of tensors in the presence of noise is explored to decompose the tensor-based signal model. The three factor matrices of the tensor can be obtained by CPD, wherein the second factor matrix is the steering vector matrix of the interference signals. The direction of arrival (DoA) of the interference signals can be extracted from the second factor matrix by algebraic methods. Additionally, the second factor matrix can also be used to suppress the interference. In a highly dynamic environment, considering that the degradation of anti-interference performance caused by the presence of the random steering vector mismatches, we propose a robust interference mitigation method. Hereinto, the key innovative steps of the proposed method is that with the presence of mismatch conditions, we derive the decomposition process of the tensor-based signal model to obtain the factor matrices, and calculate the projection matrix of the second factor matrix for interference mitigation. Whereafter, an intermediate matrix construction approach is proposed to widen the notch of the DoA of the interference, thereby improving the robustness of interference mitigation. The proposed method requires a smaller number of snapshots to achieve a high level of performance, and avoid the solution process of the covariance matrix which is indispensable to the traditional method.