ℓp-PARAFAC for joint DOD and DOA estimation in bistatic MIMO radar

In this paper, a new method to jointly estimate the direction-of-departures and direction-of-arrivals of bistatic multiple-input multiple-output radar in additive impulsive noise is proposed based on parallel factor analysis (PARAFAC). Since most of the existing method in PARAFAC model are based on the Frobenius norm, which are sensitive to outliers, we utilize the ℓP-norm to measure the residual error tensor, where 1 < p < 2, and transform it to an iterative ℓ2 minimization problem. We first construct the received data with the tensorial structure and then apply an alternative approach based on the iteratively reweighted least squares to recover the factor matrices. In the end, standard subspace techniques, i.e., MUSIC, is proposed for target estimation. Simulation results show that our proposed method outperforms the state-of-the-art methods in terms of mean angular error under α-stable noise.