.png){kind=icon}
.png){kind=icon}
Return
Multi-geometric distance method for clutter covariance matrix estimation
.png){kind=icon}
PRE
AI
DOI：10.1016/j.dsp.2025.105313.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In the background of airborne radar space-time adaptive processing (STAP), a clutter covariance matrix (CCM) estimation method is proposed, based on the first-order Taylor proximal gradient algorithm for multiple geometric distances (FTPG-MGD). This method aims to address the degradation in clutter suppression performance caused by CCM estimation with small sample sizes. The method combines Euclidean, log-Euclidean, and root-Euclidean distances to establish the weighted minimization problem. Subsequently, the approximation of the first-order Taylor expansion of the objective function is designed to transform the original nonlinear problem into a more tractable linear optimization problem. The problem is finally solved by employing a proximal gradient algorithm. Simulation and real-world data experiments indicate that the proposed method outperforms other similar algorithms in terms of CCM estimation accuracy and significantly enhances clutter suppression performance.
Keywords:
Clutter suppression
Covariance matrix estimation
Space-time adaptive processing (STAP)
Multi-geometric distance