Novel Off-Grid Estimation Method for Direction-of-Arrival Estimation Based on Deep Learning
Jie Zhang, Yanjun Zhang, Jun Tao, Jiang Zhu, Zhanye Chen, Yan Huang · 2024
Direction of arrival (DOA) estimation has always been an important and widely studied topic. In practical scenarios, achieving higher resolution and timely response speed is always a priority. Incorporating deep learning (DL) techniques into DOA estimation confers a multitude of advantages, such as efficient computation time. Nonetheless, similar to traditional algorithms, they encounter challenges with grid resolution limitations. This paper introduces a novel deep learning framework for accurately estimating DOA of off-grid signals through regression task, effectively addressing grid mismatch issues. Experimental results confirm the enhanced efficacy of the proposed method under various scenarios, surpassing traditional and other DL methods.