Adaptive Algorithms for Multilinear System Identification Using Tensor Modelling 

(AdAMSysTeM)

Algoritmi adaptivi pentru identificarea sistemelor multiliniare folosind modelarea tensorială

 

Program name and purpose: Postdoctoral Research Projects (PD) - Supporting young researchers, PhDs, who want to develop their independent professional research career in research institutions in Romania, in order to stimulate scientific excellence in Romanian research.

Funding Institution: Executive Unit for Higher Education, Research, Development, and Innovation Funding (UEFISCDI)

Project code: PN-III-P1-1.1-PD-2019-0340

Budget: 246.950,00 lei

 

Last update: 12.2022

 

Research team:

Project manager:         Ș.l.dr.ing. Laura-Maria Dogariu - Google Scholar profile

Mentor:                       Prof.dr.ing. Constantin Paleologu - Google Scholar profile

 

Project summary:

In many applications related to system identification, the unknown system can be modeled as a finite impulse response filter with a large number of coefficients, which raises challenges for the adaptive filter used for this purpose. In this context, the system can be regarded as a multidimensional structure, mathematically modeled by a tensor, which may be decomposed into shorter, easier to identify impulse responses. In addition, specific properties of the systems, such as the sparsity, could be used to simplify the problem. However, solutions have been proposed so far only for second and third order tensors. Motivated by the good performance of these solutions (developed during the PhD studies of the project leader), the project aims to provide a general framework for multilinear system identification, by using tensor modelling and decompositions. The solution will be applied first in the context of the Wiener filter, then for the main categories of adaptive filters, i.e., the least-mean-square (LMS) algorithm, the affine projection algorithm (APA), and a Kalman filter (KF) will also be developed. Convergence analysis and low-complexity versions of these algorithms will also be derived, as well as combinations of adaptive filters (e.g., combination scheme of LMS and Kalman-based algorithms, a combination scheme of APAs using different projection orders), to be used in certain applications. Software simulations will be conducted to test the performance of all the proposed algorithms, and, based on the results, the best performing solutions will be implemented on fixed-point DSP or FPGA, targeted for specific applications. The results of the project will be disseminated in top ISI journals and conferences.

 

Project objective

 

Dissemination results:

·       Journal papers:

  1. A. C. Becker, E. V. Kuhn, M. V. Matsuo, J. Benesty, C. Paleologu, L.-M. Dogariu, and S. Ciochină, “Stochastic model for the LMS algorithm with symmetric/antisymmetric properties,” Symmetry, vol. 14(9), p.1908 (11 pages), Sept. 2022 – ISI Q2.
  2. L.-M. Dogariu, J. Benesty, C. Paleologu, and S. Ciochină, “Identification of room acoustic impulse responses via Kronecker product decompositions,” IEEE-ACM Transactions on Audio, Speech, And Language Processing, vol. 30, pp. 2828-2841, 2022, WOS:000852841100004 – ISI Q1.
  3. I. D. Fîciu, J. Benesty, L.-M. Dogariu, C. Paleologu, S. Ciochină, “Efficient algorithms for linear system identification with particular symmetric filters,” Applied Sciences, vol. 12(9), p. 4263 (26 pages), April 2022, WOS:000794561600001 - ISI Q2.

4.     L.-M. Dogariu, J. Benesty, C. Paleologu, and S. Ciochină, “An insightful overview of the Wiener filter for system identification,” Applied Sciences, vol. 11, p. 7774, August 2021, WOS:000694165800001 - ISI Q2.

5.     J. Benesty, C. Paleologu, L.-M. Dogariu, and S. Ciochină, “Identification of linear and bilinear systems: a unified study,” Electronics, vol. 10, p. 1790, July 2021, WOS:000681959500001 – ISI Q3.

6.     L.-M. Dogariu, C. Paleologu, J. Benesty, C.-L. Stanciu, C.-C. Oprea, and S. Ciochină, “A Kalman filter for multilinear forms and its connection with tensorial adaptive filters,” Sensors, vol. 21(10), p. 3555, May 2021, WOS:000662590100001 - ISI Q1.

7.     L.-M. Dogariu, C.-L. Stanciu, C. Elisei-Iliescu, C. Paleologu, J. Benesty, and S. Ciochină, “Tensor-based adaptive filtering algorithms,” Symmetry, vol. 13, no. 3, p. 481, March 2021, DOI: 10.3390/sym13030481, WOS:000634166700001 - ISI Q2.

 

·       Conference papers:

  1. I. D. Ficiu, L.-M. Dogariu, R.-L. Costea, C. Paleologu, J. Benesty, and S. Ciochină, “A Kalman filter for the identification of impulse responses with particular symmetry features,” in Proc. IEEE International Symposium on Electronics and Telecommunications (ISETC), Timișoara, Romania, Nov. 2022 – ISI.
  2. J. Benesty, L.-M. Dogariu, C. Paleologu, and S. Ciochină, “An iterative Wiener filter for stereophonic acoustic echo cancellation,” accepted for publication in the Asilomar Conference on Signals, Systems, and Computers, online, Oct.-Nov. 2022 - ISI.
  3. L.-M. Dogariu, J. Benesty, C. Paleologu, and S. Ciochină, “An iterative Wiener filter for the identification of impulse responses with particular symmetric properties,” accepted for publication in the 11th edition of the International Conference "Advanced Topics in Optoelectronics, Microelectronics and Nanotechnologies" (ATOM), Constanta, Romania, August 2022 - ISI.
  4. A.-G. Rusu, L.-M. Dogariu, R.-L. Costea, C. Paleologu, J. Benesty, and S. Ciochină, “A variable step-size affine projection algorithm based on data reuse,” accepted for publication in the IEEE 45th International Conference on Telecommunications and Signal Processing (TSP), online, July 2022 - ISI. *Best Paper Award*
  5. L.-M. Dogariu, C. Paleologu, J. Benesty, S. Ciochină, E. V. Kuhn, A. C. Becker, and M. V. Matsuo, “An affine projection algorithm for the identification of impulse responses with symmetric/antisymmetric characteristics,” accepted for publication in the IEEE 45th International Conference on Telecommunications and Signal Processing (TSP), online, July 2022 - ISI.
  6. L.-M. Dogariu, C. Paleologu, J. Benesty, and S. Ciochină, “On the performance of a data-reuse fast RLS algorithm for acoustic echo cancellation,” accepted for publication in the IEEE 10th International Black Sea Conference on Communications and Networking (BlackSeaCom), Sofia, Bulgaria, June 2022 - ISI.
  7. J. Benesty, L.-M. Dogariu, C. Paleologu, and S. Ciochină, “Efficient Identification of Acoustic Linear Systems,” in Proc. the IEEE 14th International Conference on Communications (COMM), online, June 2022 - ISI.
  8. I.-D. Ficiu, C.-L. Stanciu, L.-M. Dogariu, C. Elisei-Iliescu, and C. Paleologu, “Tensor-based recursive least-squares algorithm with low computational complexity,” in Proc. the 15th International Conference on Sensor Technologies and Applications (SENSORCOMM), Athens, Greece, November 2021. *Best Paper Award*
  9. L.-M. Dogariu, C. Paleologu, J. Benesty, and S. Ciochină, “Improved affine projection algorithm for the identification of multilinear forms,” in Proc. IEEE the 27th International Symposium for Design and Technology in Electronic Packaging (SIITME), online, October 2021 - ISI.
  10. A.-G. Rusu, L.-M. Dogariu, S. Ciochină, and C. Paleologu, “A data-reuse approach for an optimized LMS algorithm,” in Proc. IEEE the 11th Conference on Speech Technology and Human-Computer Dialogue (SpeD), online, October 2021 - ISI.
  11. L.-M. Dogariu, C. Paleologu, J. Benesty, and S. Ciochină, “Identification of multilinear forms with the tensorial Kalman filter,” in Proc. IEEE the 44th International Conference on Telecommunications and Signal Processing (TSP), online, July 2021 - ISI.
  12. L.-M. Dogariu, C. Elisei-Iliescu, C. Paleologu, J. Benesty, and S. Ciochină, “A tensorial affine projection algorithm,” in Proc. IEEE the 15-th International Symposium on Signals, Circuits and Systems (ISSCS), Iasi, Romania, July 2021 - ISI.
  13. I.-D. Ficiu, L.-M. Dogariu, C.-L. Stanciu, and C. Paleologu, “Identification of multilinear forms using combinations of adaptive algorithms,” in Proc. the Sixth International Conference on Advances in Signal, Image and Video Processing (SIGNAL), Valencia, Spain, June 2021.

 

·       Book chapters:

 

1.     L.-M. Dogariu, C. Paleologu, J. Benesty, S. Ciochină, “Identification of Multilinear Systems: A Brief Overview”, IntechOpen, 2022, ISBN: 978-1-80355-766-3.