Site: Fujitsu Laboratories
Date Visited: 3 June 1999
WTEC Attendees: N. Moayeri (report author), A. Ephremides, B. Mooney, H. Morishita, R. Rao, W. Stark
Hosts: Jifeng Li, Researcher, Wireless Communication
Systems Laboratory, Network Systems Laboratories
Tamio Saito, Senior Researcher, Wireless Communication Systems Laboratory, Network Systems Laboratories
Yukio Takeda, Director, Wireless Communication Systems Laboratory, Network Systems Laboratories
Dr. Takeda started the meeting at 9:35 a.m. and went over the agenda for the meeting. Dr. Ephremides then outlined the WTEC panel goals for the visit.
Dr. Takeda gave the first Fujitsu talk on the organization of Fujitsu Laboratories, Ltd., which was established in 1968, has about 1,500 employees, and has capital of ¥5 billion. He discussed Fujitsu, Ltd., which was founded in 1935, has over 180,000 employees, and had consolidated revenues of $44 billion in the fiscal year, which ended in March 1999.
Dr. Takeda discussed the Wireless Communication Systems Laboratory, which consists of 25 employees and which he is in charge of. The laboratory is concentrating on 3G and 4G wireless systems and Intelligent Transportation Systems (ITS). In 3G and 4G the researchers are investigating interference cancellation using multiuser detection and adaptive array antennas for W-CDMA base stations. He showed a picture of two hardware racks researchers have developed to study the two approaches mentioned above. These were developed in collaboration with FTRC. In the ITS area researchers are developing a millimeter wave radar.
Dr. Saito presented a talk on interference cancellation and mobile multimedia communications. He talked about the drastic increase in the number of wireless communication users in Japan. Two data services are currently offered in Japan, a 28.8 kbps service for PDC users and a 64 kbps service for PHS users. The interference cancellation scheme is a two-stage system, followed by a RAKE receiver. This system results in 1.6 Db performance improvement at a BER of 10(-3). This results in an almost doubling of system capacity. However, in practice in a multi-cell system, the capacity improvement is only 1.3. In the practical system the interference cancellation scheme is applied to high rate data users, because they transmit at higher powers. Multiuser detection is also employed. The system can cancel a maximum of 32 interferers.
In the second part of his talk, Dr. Saito described the Multimedia Mobile Access Communications Promotion Council (MMAC-PC) and its collaborations with IEEE 802.11B and ETSI BRAN. Fujitsu is considering two systems, one with a bit rate of 6-10 Mbps (with a maximum of 25 Mbps) in the 25/40/60 GHz band and another with a bit rate of 156 Mbps in the 60 GHz band. These systems are for stationary or slow moving users. He gave some details on the physical link aspects of these two systems and mentioned that the work on the MAC layer is underway.
Next Dr. Li described turbo coding for W-CDMA, which Fujitsu is proposing to standardization bodies such as ARIB and ETSI and recently within 3GPP. Turbo coding is used for higher data rate modes of W-CDMA and possibly CDMA2000. He went over a Fujitsu puncturing method for turbo codes and compared it with another one proposed by the Korean company LOGIC. The Fujitsu approach was slightly better at higher signal-to-noise ratios. In addition, it was simpler to implement.
He then described a multidimensional turbo-coding scheme, which used three decoders in each iteration. He showed the performance of this coding scheme when used in conjunction with W-CDMA over the FPLMTS Vehicular B channel model. It shows improvements over the conventional turbo coding method.
He then briefly talked about interleaver design for turbo coding, specifically a modified symmetric S-interleaver. In the discussion following Li's presentation, the issue of the tradeoff between degree of spreading and the coding rate for the turbo code came up. This is an open question for low bit rate users.
The panel heard about Fujitsu's activities on 3G wireless systems and beyond. The hosts went over an interference cancellation method they have studied and experimented with. The researchers have built a hardware test-bed for use in a laboratory environment (as opposed to field tests) for their experiments, a trend that was observed in many other Japanese companies. WTEC panelists also heard about Fujitsu's activities on multimedia mobile communications within the framework of MMAC-PC. Technical specifications are now available for MMAC.
Fujitsu is collaborating with other industry partners to build a prototype for MMAC-type systems. It has a project on turbo coding with an emphasis on reduced complexity decoding algorithms. In particular, their researchers have looked at interleaver design and multidimensional turbo coding. Fujitsu has some research activities with a 1-2 year timeline, such as the work on 3G wireless systems, and some with a 5-7 year horizon, such as the work on 4G wireless systems.