電通大の国際交流
国際教育センターホーム arrow 電通大の国際交流 arrow 国際交流事業 arrow 組織的な若手研究者等海外派遣プログラム arrow 各派遣者の報告 arrow AKHTAR Muhammad Tahir

AKHTAR Muhammad Tahir先端領域教育研究センター　特任助教

派遣先：アメリカ合衆国 カリフォルニア大学サンディエゴ校
派遣期間：平成22年11月1日～平成23年3月25日

研究テーマ：Novel Signal Processing Techniques for Brain Machine Interface (BMI) Applications

活動の概要

The Institute of Neural Computation (INC) at UCSD is one of the leading institute famous for pioneering research in the area of neuroscience and related disciplines. The main objectives of our research visit to INC can be summarized as:

  1. Learn cutting-edge information about research in the area of neuroscience and its applications
  2. Investigate novel signal processing techniques, well suited for real-time BMI applications
  3. Develop strong collaboration with researchers at INC for future research at UEC

In order to achieve the first objective, I attended many seminars and lectures from the leading researchers. Furthermore, I participated in "EEGLab Workshop", held at INC, UCSD, Nov., 18-22, 2010.

For our research activities in the area of signal processing for BMI applications, we mainly focussed on two problems.

  1. Developing efficient recursive algorithms for online Independent Component Analysis (ICA):
    It is well know that the natural gradient-based ICA algorithms can be used for online ICA, however, the convergence speed is very slow. We investigated least-square based ICA algorithm well suited for real-time blind source separation of EEG signals.
  2. Blind source separation of long term EEG recordings:
    The main problem is to maintain continuity of separated sources, which is due to inherent permutation ambiguity of ICA. In order to address this issue, we employed Wavelet analysis and cross-correlation between wavelet basis in the overlapped window.

In order to establish strong collaboration for future research activities, I gave a seminar in ChalkTalk Series of INC, UCSD. The title of talk was "From Active Noise Control to Biomedical Signal Processing: Research Background and an Outlook on Future Activities", and the main objective was to present an overview of our research activities to find directions for future research collaboration.

Besides working on above-mentioned research topics, we continued our work on active noise control (ANC) and completed work on four papers which include one paper accepted for publication in IEEE Transactions on Audio Speech and Language processing, and one chapter for a book project on Adaptive Filtering to be published by InTech publishers.

研究成果概要

As described in "6. Summary of research activities", the main outcomes of the research visit can be summarized as follows:

  1. Got new information for stat-of-art research in the area of signal processing for neuroscinece and bioengineering applications.
  2. Worked on developing new algorithms for online processing of EEG signals for real-time applications, and blind source separation of long term EEG recordings. Now we are preparing conference and journal papers for publication of these latest findings [i].
  3. Completed our work on various papers on ANC, including one paper with IEEE Transactions Audio Speech Language Processing [ii], one book chapter [iii], and two papers with Applied Acoustics [iv], [v],
  4. Last but not least, we established a strong link with our host professor, Prof. Gert Cauwenburgs, IEEE Fellow and Professor of Bioengineering, for possible research collaboration in future research activities at UEC.

References:

  1. M. T. Akhtar and G. Cauwenburghs. Recursive independent component analysis for online blind source separation. Twenty-Fifth Annual Conference on Neural Information Processing Systems, Granada, Spain, December 13-15, 2011 (in preparation)
  2. M. T. Akhtar and W. Mitsuhashi. Improving performance of hybrid active noise control systems for uncorrelated narrowband disturbances,” IEEE Trans. Audio Speech Language Processing. (In Press)
  3. M. T. Akhtar, W. Mitsuhashi, and A. Nishihara. Applications of adaptive filtering: Recent advancements in active noise control. in Adaptive Filtering - 2 (Ed.), InTech Publishers, 2011. (In Press)
  4. M. T. Akhtar and W. Mitsuhashi. Improving robustness of filtered-x least mean p-power algorithm for active attenuation of standard symmetric-alpha-stable impulsive noise. Applied Acoustics. (in Press)
  5. M. T. Akhtar and W. Mitsuhashi. Variable step-size based method for acoustic feedback modeling and neutralization in active noise control systems. Applied Acoustics, vol. 72, no. 5, Pages 297-304, April 2011.

国際化に関する所感及び提言

This research visit provided me with a unique opportunity to work at the Institute for Neural Computation (INC) at UCSD, California. It enabled me to interact with and learn from the leading researchers and experts in the area of neuroscience. It has helped me to establish a strong collaboration link for my future research activities at UEC. Beside learning new scientific things and getting research experience at prestigious international institute, I have personally learned many new things.

I believe that these types of research visits give us great exposure to interact with the international research community, and enable us to better contribute in serving the humanity by generating new and exciting ideas. Thus these visits are very important for our professional growth as a researcher as well as personal growth as a human being.

作成日：2011年3月25日 / 更新日：2011年11月18日