Weighted maximum mixture complex correntropy-based past algorithm: a novel approach for low-altitude target parameter estimation

Aiming at the challenges of strong ground clutter interference, low signal-to-noise ratio, and near-field non-stationary signals in detecting low, slow, and small (LSS) targets in low-altitude environments, this paper proposes a joint parameter estimation method for low-altitude targets based on an improved weighted maximum mixture complex correntropy-based projection approximation subspace tracking (WMMCC-PAST) algorithm. Initially, an innovatively developed weighted mixture complex correntropy function integrates complex Gaussian kernels with a dynamic weighting mechanism. This approach effectively addresses the limitations of traditional correntropy methods in suppressing continuous similar-amplitude noise, overcoming fixed kernel bandwidth constraints, and limited adaptability of single-kernel approaches, thereby markedly enhancing the algorithm’s robustness in complex noise environments. Subsequently, a cost function based on the WMMCC criterion is constructed, integrating the weighted maximum mixture correntropy criterion with the PAST algorithm framework. By embedding the robust correntropy-based metric into the subspace tracking process, this formulation significantly improves the algorithm’s resilience to heavy-tailed noise commonly encountered in low-altitude radar scenarios. Next, this work theoretically derives and implements a robust WMMCC-PAST algorithm tailored for impulsive noise environments, which significantly improves tracking performance in low-altitude radar applications affected by non-Gaussian disturbances, particularly under heavy-tailed noise conditions. Furthermore, theoretical analysis establishes the boundedness of the proposed weighted mixture complex correntropy function, proves the convergence of the WMMCC-PAST algorithm, and further examines its robustness under α-stable noise. In addition, the computational complexity and its practical relevance to radar applications are analyzed to support the feasibility of real-time deployment. Experimental results demonstrate that the proposed algorithm has good estimation performance. This study provides an effective technical solution for LSS target detection in low-altitude security fields, holding significant theoretical value and promising engineering applications.