A Compressive Sensing Based Computationally Efficient High-Resolution DOA Estimation of Wideband Signals Using Generalized Coprime Arrays

Most recent state-of-the-art wideband direction of arrival (DOA) estimation techniques achieve reasonable accuracy at the expenses of high computational complexity. In this paper, a new computationally efficient approach based on compressive sensing (CS) is introduced for high resolution wideband DOA estimation. The low software complexity is achieved utilizing CS with deterministic chaotic Chebyshev sensing matrices that allow reducing the measurement vector dimension, while the high-resolution DOA estimation is acquired utilizing an efficient generalized coprime array configuration. The effectiveness of the introduced approach in enhancing the DOA estimation precision and reducing the computational complexity is studied along with a detailed comparison between three state-of-the-art wideband DOA estimation techniques, namely incoherent signal subspace method (ISSM), focusing signal subspace (FSS), and modified test of orthogonality of projected subspaces (mTOPS) with and without applying the proposed CS technique. The performance is examined utilizing various assessment metrics such as the spatial spectrum, the computational time, and the root mean square error between estimated and actual DOAs when varying the signal-to-noise ratio and number of elements. Results reveal that applying the proposed CS technique to the three algorithms (ISSM, FSS, and mTOPS) provides significant reduction in the execution time needed for the DOA estimation without affecting the resolution accuracy under various set of parameters. This reveals the importance of the proposed approach in wideband wireless communication systems.