A Robust Super-Resolution DoA Estimation Algorithm Using Vision Transformer
Hao Yu, Sheng Yi Wu, Liujie Lv, Yi Su · 2024
Direction of arrival (DoA) estimation is a fundamental problem in wireless communication and has received considerable attention. Most existing methods are lacking of the adaptation to some challenging observation conditions such as low signal-to-noise ratio (SNR), limited number of snapshots and antenna array imperfections. To address these challenges, this paper presents a novel deep learning (DL) algorithm for DoA estimation, using the vision Transformer (ViT) for stronger global feature extraction capability. Specifically, we design a training label relying on the Gaussian distribution and integrate it into ViT to achieve superior DoA estimation precision. Furthermore, we use 1-bit quantized signals for simulation to verify the universality of our proposed algorithm. As the simulation results demonstrate, the proposed algorithm provides higher estimation accuracy and robustness in these challenging scenarios.