OREANDA-NEWS. Fujitsu Limited, Fujitsu Laboratories Ltd., and Transtron Inc. today announced that in the Fiscal 2016 Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology, they have received the Prize for Science and Technology in the Development Category for their "Development of Advanced Noise-Suppression Technology Using Microphone Arrays." In addition, Fujitsu Laboratories Ltd. and Fujitsu Laboratories of America, Inc. were honored in the Research Category for their "Research into High-Speed Data Transmission Circuits for Implementing High-Performance ICT Systems."
The Ministry of Education, Culture, Sports, Science and Technology (MEXT) commends individuals for their important achievements in science and technology R&D and their promotion of science and technology understanding. The awards are aimed at motivating researchers and helping to raise the level of Japan's science and technology.
Award Recipients and Technology Being Recognized
Prizes for Science and Technology, Development Category
This category recognizes people who have made extraordinary R&D achievements or inventions that have real-world applications and that improve Japan's society, economy, and the lives of its citizens.
Recipients:
- Naoshi Matsuo (Research Manager, Media Processing Laboratory, Fujitsu Laboratories)
- Shoji Hayakawa (Staff Researcher, Media Processing Laboratory, Fujitsu Laboratories)
- Chikako Matsumoto (Media Processing Laboratory, Fujitsu Laboratories)
- Junichi Watanabe (Department Manager, Vehicle Control Second Development Unit, Transtron)
- Hiroshi Katayama (Ubiquitous IoT Division, Fujitsu)
Project:
Development of Advanced Noise-Suppression Technology Using Microphone Arrays
Voice-quality degradation due to ambient noise is a problem for voice-recognition technology when using such recognition to control devices in moving cars, and for outdoor conversations on mobile phones. Arrays that use multiple microphones are an effective way to suppress such noise, but such noise reduction is difficult to achieve with the small units that can be installed in car interiors and mobile phones. This technology effectively suppresses the frequency components of noise that come from directions and distances other than that from a target speaker. The feature of this technology is to suppress only noise in bands likely to be perceptible by taking into account not only the properties of waveforms based on the direction and distance of the sound source, but also auditory properties used by voice-recognition and mobile phones.
Unlike conventional noise suppression technologies that require a great number of microphones for strict full-band processing, this approach, using only two microphones, achieves more than 20 dB (less than 1% relative power) of noise suppression in moving cars or on the street. This result contributes to safer and more certain device operation when driving, thanks to better voice-recognition accuracy in the noisy cabin environment via an in-car microphone array. This will also contribute to quicker responses to accidents, when used as microphones for in-car emergency-information systems in Europe. Finally, the microphone array named "Super Double Microphone" mounted on mobile phones such as "Raku Raku PHONE", contributes to improved mobile-phone usability, thanks to higher voice quality when calling on phones.
Prizes for Science and Technology, Research Category
This category recognizes people who have had highly creative research achievements or inventions with the potential to advance science and technology in Japan.
Recipients:
- Yoichi Koyanagi (Project Director, Computer Systems Laboratory, Fujitsu Laboratories)
- Takuji Yamamoto (Project Director, Computer Systems Laboratory, Fujitsu Laboratories)
- Yasuo Hidaka (Senior Researcher, Novel Computing Laboratory, Fujitsu Laboratories America)
Project:
Research into High-Speed Data Transmission Circuits for Implementing High-Performance ICT Systems
Modern servers and supercomputers use large-scale parallel architecture in which numerous processors are connected to process big data and compute massive scientific calculations. Maximizing the performance of these systems requires not only that individual processors run fast but also that the data be moved quickly between them. In the past, data transfer speeds between processors, either in the form of electrical or optical signals, have been limited to 20 Gbps, due to significant waveform distortions. Another problem has been that higher transmission speeds have required more power.
This research successfully developed an equalizer circuit that compensates for distortions of the electric signals in low frequency as well as high frequency, and also a control circuit that automatically adjusts the compensation for each of several thousand signal lines in a system. In addition, this research produced a circuit that drives optical modulators at high speed and in low power for optical transmission systems.
This research has enabled high-speed electrical and optical signal transmission at 25 Gbps or higher in low power. It also supports large-scale systems with numerous signal lines. These results will contribute to building higher-performance systems used for big data and scientific calculations, and their reduced power demands will help lighten environmental footprints, including reducing CO2 emissions.
Комментарии