Abstract:
To tackle the challenge where existing estimation algorithms exhibit performance deterioration or complete failure in polar environments due to impulsive noise, this paper presents a robust orthogonal matching pursuit (OMP) algorithm. Firstly, accurate selection of atom bases is achieved by introducing the maximum correntropy criterion (MCC). Secondly, the
L_1
-norm is utilized to reconstruct the loss function, mitigating the impact of impulse noise on parameter estimation. Simultaneously, the alternating direction method of multipliers (ADMM) is employed to efficiently obtain the global optimal solution. Numerical simulations and the processing of experimental data collected from the 9th Chinese National Arctic Research Expedition have shown that the proposed method exhibits significant performance improvements compared to classical algorithms. Specifically, it exhibits higher estimation accuracy and stronger robustness under impulsive noise conditions.