Gridless Joint Multi-Band DOA Estimation for Spectrum Sensing

In spectrum sensing, the direction of arrival (DOA) information allows for a reduction in latency and capturing short-lived signals. However, challenges persist, including a limited size of antenna/sensor array, unknown signal waveforms, frequency spectrum overlapping and rapidly changing signals. To tackle these problems, a gridless joint multi-band (GJMB) DOA estimation method is proposed for spectrum sensing in dynamic environments. To jointly process the varying steering matrices from different frequencies, a common frequency grid is formed based on the greatest common divisor of multiple frequencies. Then, a truncated Hankel matrix is constructed to recast the problem as low-rank matrix recovery. An iterative solution is developed to solve the nonconvex optimization problem directly. Numerical results demonstrate the proposed method can robustly recover more sources than sensors of a uniform linear array with a small number of snapshots. The performance surpasses the convex relaxation using nuclear norm minimization and allows for downsizing the physical array in contrast to the sparse array.