.png){kind=icon}
.png){kind=icon}
Return
Online Kernel Learning Target Localization for Distributed MIMO Radar
.png){kind=icon}
PRE
AI
DOI：10.1109/taes.2026.3709260.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Conventional maximum likelihood estimation and model-based target localization methods utilizing prior information of measurements would not provide desired performance in harsh environments with low signal-to-noise ratio for the distributed multiple-input multiple-output radar. This article first presents an online data-driven kernel learning-based target localization method by using radar measurements. The measurement vector is transformed into the reproducing kernel Hilbert space, and target localization is transformed into a function approximation problem in such a feature space. Then, the kernel recursive least square target localization (KRLSTL) algorithm is proposed to solve the generated problem, and the fixed-dimensional dictionary update strategy is introduced to generate the fixed-budget KRLSTL algorithm to fix the size of the radial basis function network. Some simulations are performed to verify the effectiveness and superiority of the proposed method over some counterparts under inaccurate radar measurements and nonstationary environment.
Keywords:
Distributed multiple-input multiple-output (MIMO) radar
fixed-budget strategy
kernel recursive least squares (RLS)
online learning
target localization
Journal
IF:
5.7
Papers:
676
Citations:
2.4W