Efficient Off-grid DOA Estimation based on Modified MUSIC for Arbitrary Linear Arrays

Abstract

The off-grid direction-of-arrival (DOA) estimation performance degrades due to finite discretized sampling grid points. Meanwhile, most of the off-grid estimation algorithms restricted to uniform linear arrays (ULAs), such as root-MUSIC, ESPRIT and Fourier coefficients interpolation based methods, can not be employed to nonuniform linear arrays (NLAs). To solve these two problems, we propose an efficient off-grid DOA estimation technique based on modified multiple signal classification (MUSIC) for arbitrary linear arrays in this work. The on-grid DOA estimates are obtained using MUSIC in the first iteration. Subsequently, we apply the first-order Taylor expansion of the steering vector to MUSIC for calculating the closed-form formula of the angle offset. Finally, the angle biases of off-grid sources are corrected iteratively based on the angle offset formula. Simulation results demonstrate the effectiveness and efficiency of the proposed method for off-grid DOA estimation using arbitrary linear arrays.