Mmccgl-based tensor decomprosition for robust parameter estimation in bistatic MIMO radar

To address the performance degradation of bistatic MIMO radar parameter estimation in low-altitude environments with impulsive interference, this paper proposes a robust tensor-decomposition framework based on the maximum mixture complex correntropy criterion (MCCGL). The MCCGL function integrates Gaussian and Laplacian kernels to enhance robustness in PARAFAC decomposition. An adaptive mixture-weighting strategy is developed to suppress background noise and impulsive outliers without requiring prior noise statistics. A half-quadratic reformulation transforms the resulting optimization into weighted least-squares subproblems, enabling efficient closed-form updates within the TALS procedure. The MMCCGL-PARAFAC algorithm achieves joint estimation and automatic pairing of Doppler, DOD, and DOA parameters. Theoretical analyses establish the boundedness of the proposed cost function, mixed-noise suppression, and computational feasibility. Extensive simulations under SαS noise and other representative impulsive noise models demonstrate that the proposed method consistently outperforms conventional methods in estimation accuracy, robustness, parameter pairing, and runtime performance under low SNR and non-stationary conditions.