Low-complexity adaptive filtering-based direction of arrival (DOA) estimation algorithms have been proposed to avoid the computationally expensive steps of covariance matrix estimation and eigenvalue decomposition (EVD). However, their performance degrades significantly in impulsive noise scenarios. To address this issue, an adaptive filtering algorithm named complex total maximum versoria criterion (CTMVC) algorithm has been designed to perform robust DOA estimation under impulsive noise. Nonetheless, determining an appropriate step size remains challenging, yet it plays a vital role in accurate DOA estimation. This paper puts forward a variable step-size scheme for the CTMVC algorithm by incorporating cumulative gradient error and signal power estimation, thereby striking a compromise between the speed of convergence and the error of the steady-state. Furthermore, this article provides a detailed computational complexity analysis of the proposed algorithm. Extensive simulations across various metrics highlight the enhanced effectiveness of the VSS-CTMVC over current algorithms.
