A CPD-based AOA estimation algorithm with vandermonde-constrained preprocessing approach

This paper investigates a tensor-decomposition-based algorithm for the arrival angle estimation. With the assistance of multiple invariance modeling and uniqueness property of Canonical Polyadic Decomposition (CPD), the proposed method is capable of performing well even in worse conditions. Furthermore, by exploiting the feature of Vandermonde structure, a modified preprocessing step has been incorporated into this algorithm which can achieve a better performance than the standard CPD and classic subspace methods. Numerical experiments have demonstrated the feasibility and effectiveness of the proposed method.