Abstract
Recent developments in electronic countermeasures have heightened the demand for direction of arrival (DOA) estimation, and it has become more urgent to improve the resolution of closely spaced sources for conformal arrays. In previous studies, distributed sources incident on linear or planar arrays have been the point of focus, and their performance for closely spaced sources deteriorates sharply. To address this issue, we design a multi-head attention (MHA) mechanism-based cascaded estimator, which is composed of a signal-to-noise ratio (SNR) classification network and DOA estimation networks. The former network utilizes fully connected neural networks for SNR classification, and its output can activate the corresponding estimating subnetworks. In the estimating subnetworks, we adopt multiple self-attention mechanisms for the first time to extract more implied and representative information from original features for effective estimation. Numerical results validate that the proposed estimator not only exhibits better performance in terms of resolution and recovery accuracy even at lower SNR, but also provides great robustness with respect to angle intervals and snapshots, which lays the groundwork for further DOA research for closely spaced sources.
Funding Information
  • National Science Foundation for Young Scientists of China (61801308)
  • Liaoning Revitalization Talents Program (XLYC1907195)
  • Aeronautical Science Foundation (2020Z017054001)

This publication has 34 references indexed in Scilit: