Joint frequency and DOA estimation of sub-Nyquist sampling multi-band sources with unfolded coprime arrays

Nowadays, as the working frequency band gradually increases, the efficient joint frequency and DOA estimation of multi-band sources has been an urgent task in cognitive radio and future communications (such as beyond 5G and 6G etc.). This task relies on twofold: enlarge array aperture and reduce the sampling costs in spatial domain and temporal domain. Therefore, we propose a sub-Nyquist receiver architecture with two perpendicular unfolded coprime arrays (UCPAs) based modulated wideband converter. On one hand, the proposed architecture completely excludes the dense distribution of sensors, thus yielding a large aperture. On the other hand, both the frequency and DOA can be estimated analytically by performing root-MUSIC on the pseudo noise subspaces constructed from the collected sub-Nyquist samples. Particularly, we also develop a beamforming based correlation matching method to solve the intractable paring problem among multiple targets, which facilitates the spectrum recovery of the individual source. Due to the large array aperture of UCPA, our proposed joint estimator acquires a considerable performance improvement over the existing ULA based architectures. Numerical results show that, the proposed estimator concurrently possesses high accuracy and high resolvability for densely distributed sources.