University of Minnesota


Blind PARAFAC Antenna Receivers for DS-CDMA Systems

This research project is lead by prof. Sidiropoulos at the University of Virginia and focuses on tri- and multi-linear models in signal processing for communications applications. We also explore multi-way analysis tools in Chemometrics (wherein research is headed by our collaborator, Rasmus Bro). The starting point is the trilinear decomposition, also known as Parallel Factor (PARAFAC) analysis. It has been first introduced as a data analysis tool in Psychometrics; to the best of our knowledge the first seed of PARAFAC ideas appeared as early as 1944, in a paper by R. B. Cattell in Psychometrika. PARAFAC has been used for "individual differences multidimensional scaling," but also in Phonetics, exploratory data analysis, statistics, and arithmetic complexity. Nowadays, most of the research in the area is conducted in the context of Chemometrics, where it is used for spectrophotometric, chromatographic, and flow injection analyses.

For our Signal Processing and Communications interests, PARAFAC offers a fundamental generalization of ESPRIT ideas - actually as a general principle underlying the most general formulation of ESPRIT to date. PARAFAC can also be seen as generalizing joint approximate diagonalization ideas. We look at novel applications of PARAFAC to sensor array processing, wireless multiuser communications (in particular, CDMA, TDMA, and OFDMA systems), and source separation. We also contribute to the theory of multi-way analysis, particularly towards uniqueness/identifiability results; and on the algorithmic side of things, we develop efficient fitting algorithms for the PARAFAC model.

Our work in this project and basic references can be found in the following papers:

  • R. Bro, N. Sidiropoulos, and G. B. Giannakis, "A Fast Least Squares Algorithm for Separating Trilinear Mixtures," in Proc. ICA99 Int. Workshop on Independent Component Analysis for Blind Signal Separation, pp. 289-294, Jan. 11-15, 1999.
  • R. Bro, N. D. Sidiropoulos, and G. B. Giannakis, "Optimal Joint Azimuth-Elevation and Signal-Array Response Estimation Using Parallel Factor Analysis," Proc. of 32nd Asilomar Conf. on Signals, Systems, and Computers, Pacific Grove, CA, Nov. 1-4, 1998.
  • N.D. Sidiropoulos, G.B. Giannakis, and R. Bro, "Deterministic Waveform-Preserving Blind Separation of DS-CDMA Signals Using an Antenna Array," Proc. of IEEE SP Workshop on Statist. Signal and Array Proc. (SSAP'98), pp. 304-307, Portland, Oregon, Sept. 14-16, 1998.


SPINCOM: (612) 624-9510 | 499 Walter Library, 117 Pleasant Street SE, Minneapolis, MN 55455