Deterministic Blind Separation of Sources Having Different Symbol Rates Using Tensor-Based Parallel Deflation

Abstract

Abstract. In this work, we address the problem of blind separation of non-synchronous statistically independent sources from underdetermined mixtures. A deterministic tensor-based receiver exploiting symbol rate diversity by means of parallel deflation is proposed. By resorting to bank of samplers at each sensor output, a set of third-order tensors is built, each one associated with a different source symbol period. By applying multiple Canonical Decompositions (CanD) on these tensors, we can obtain parallel estimates of the related sources along with an estimate of the mixture matrix. Numerical results illustrate the bit-error-rate performance of the proposed approach for some system configurations.

Blind source separation of underdetermined mixtures of event-related sources

“…In the style of [15], we exploit the event-related nature of the signals of interest to construct a data tensor with dimensions space, event-synchronized window, and time from the 2-dimensional measurements for each of Q event-related sources. To this end, for the qth source, we identify L q event-synchronized windows of length T q of the corresponding time signal.…”

“…These algorithms, which exploit the cyclostationarity property resort to statistical tools. In [15], a parallel deflation procedure based on a deterministic tensor decomposition has been proposed to address the problem of underdetermined BSS in the cyclostationary context. The basic approach consists in constructing a tensor by synchronizing on the symbol rate of a certain source, and decomposing the tensor using the Canonical Polyadic (CP) decomposition [16] to extract the characteristics of the source.…”

“…In this paper, the method described in [15] has been adapted for the estimation of the mixing matrix, temporal patterns, and amplitudes of event-related sources. The method described in [15] fails to extract a source which has very little power compared to the other sources because the latter act as interferers with high amplitudes that can be considered as outliers and impede on the accurate tensor decomposition.…”

“…The method described in [15] fails to extract a source which has very little power compared to the other sources because the latter act as interferers with high amplitudes that can be considered as outliers and impede on the accurate tensor decomposition. To overcome this problem, we propose to apply a robust tensor decomposition.…”