Radar Signal Reconstruction in Severe Interference via Robust Tensor Completion

The sensing of the multi-function phased array radar (MFPAR) is fundamental to modern electronic reconnaissance systems, but current approaches to sensing and estimating MFPAR mainly rely on direct measurements of pulse descriptor word, which results in the issues of robustness deficiency and inflexible adaptability in complex electromagnetic and sporadic interference environments. Traditional compressive sensing and matrix-based signal reconstruction method fail to address the high dimensionality and impulsive noise in MFPAR signals, while existing tensor completion approaches generally lack support for the complex-valued data and complex electromagnetic environments inherent in electronic reconnaissance. Accordingly, this paper presents a robust reconstruction framework tailored for electronic reconnaissance, which incorporates complex correntropy and complex conjugate gradient (CCG) optimization into complex-valued tensor completion. This approach significantly enhances robustness against impulsive outliers, offering a promising foundation for downstream tasks such as signal sorting, parameter estimation, and radar recognition. Specifically, the robust signal reconstruction objective function is formulated under the maximum complex correntropy criterion to model the reconstruction task. Based on the resulting robust optimization framework, a half-quadratic optimization combined with the CCG algorithm is employed, leading to the proposedmaximum complex correntropy criterion with conjugate gradient signal reconstruction(MCCC-CGSR) algorithm. Extensive numerical simulations demonstrate that the proposed MCCC-CGSR algorithm outperforms several state-of-the-art counterparts in terms of reconstruction accuracy and computational efficiency