Vector-Sensor-Based Signal Parameter Estimation by Exploiting CPD of Tensors

An approach for parameter estimation of signals impinging on an array of electromagnetic vector sensors is presented by exploiting canonical polyadic decomposition of tensors. The uniqueness of the approach is that it can distinguish the signals with the same direction of arrival (DOA) and copolarized state: one function that has not been fulfilled by the up-to-date approaches such as the derivative approaches of multiple signal classification and estimation of signal parameters via rotational invariance techniques, which are the traditional tensor-based approaches. Regarding the reason of our superiority in this function, it lies in that the traditional approaches usually require that the matrices involved in the operations are full column rank. However, in this article, the parameters can be estimated by virtue of the factor matrices, as long as the matrices satisfy the requirement that at least one is full column rank. In other words, as long as the signals comply with any one of the spatial, temporal, and polarized diversities, the proposed approach can effectively estimate the parameters of these signals.