Direction finding in MIMO radar with large antenna arrays and nonorthogonal waveforms

As an important research branch in multiple-input multiple-output (MIMO) radar, orthogonal waveforms are always desirable. In practice, however, the orthogonality may not be always guaranteed. In this paper, a computationally efficient parallel factor (PARAFAC) estimator is proposed, which is suitable for direction-of-arrival (DOA) estimation in colocated MIMO radar with large antenna arrays as well as nonorthogonal waveforms. Firstly, the spatially colored noise caused by the nonorthogonal waveforms is eliminated via temporal cross-correlation of the measurement. Thereafter, a third-order PARAFAC (or called trilinear decomposition) model is established by exploiting the multidimensional structure as well as the low rank property of the cross covariance matrix. The DOA estimation problem is then linked to PARAFAC decomposition and finally obtained via least squares technique. Compared with the existing matrix completion-based algorithm, the proposed estimator is attractive from the perspective of computational complexity and estimation accuracy, especially with large antenna arrays scenario. In addition, the stochastic Cramer-Rao bound on DOA estimation with waveform imperfectly is derived. Numerical simulations are provided to verify the effectiveness of the proposed estimator. (C) 2019 Elsevier Inc. All rights reserved.