Site: Advanced Telecommunications
Date Visited: 31 May 1999
WTEC Attendees: A. Ephremides (report author), R. Pickholtz, R. Rao, J. Maurice, B. Mooney
Hosts: Dr. Bokuji Komiyama, President
Dr. Nagao Ogino
Mr. Keizo Inagaki
ATR is an institute/laboratory established in 1986 as a joint organization. Each laboratory receives 70% of its support from Japan's Key Technology Center (Key Tec) and 30% from private companies, whose mission is to engage in high risk, basic research ventures in the broad field of communications.
Its mode of operation is fairly unique in that it is based on externally approved, seven-year projects that are manned with personnel mostly borrowed from participating companies.
The WTEC panel visited only one of the current four seven-year projects (referred to as laboratories within ATR). This laboratory, called the ATR Adaptive Communication Research Laboratories, was established in 1996 and is currently operating with a budget of ¥11.9 billion for the period 1996-2003.
The other three (established respectively in 1995, 1993, and 1992) are the ATR Media Integration and Communications Research Labs, the ATR Interpreting Telecommunications Research Labs, and the ATR Human Information Processing Research Labs.
When its seven-year cycle expires, a project may be continued if its goals are not yet fully accomplished or it may be discontinued if it has fully achieved these goals. The goals consist of patenting and commercializing the products of the research effort. The term used within the organization for the latter is managing the fruits of research. Currently, there are four such successfully completed projects.
The overall annual budget for all four operating laboratories is currently ¥7.4 billion.
The WTEC panel visit commenced with a round-table welcome and introduction by Dr. B. Komiyama, the President of the ATR Adaptive Communications Research Labs. As explained by Dr. Komiyama, the Labs consist, in turn, of four departments. There is one department on Architecture for Adaptive Communications, one on Design and Control of Adaptive Systems, one on Advanced Wireless Communications, and one on Advanced Communication Devices.
Dr. Komiyama had selected four projects to brief the panelists in detail. Two of these are in the department of Architecture for Adaptive Communications and focus on upper-layer issues and the other two are in the department of Advanced Wireless Communications and focus on lower-layer issues. Specifically, the former two include self-organizing wireless networks and adaptive QoS management, and the latter two include microwave photonics and intelligent antenna arrays.
The briefings on these projects were conducted by the respective directors or principal investigators and took place within the specific lab environment of each project. All were explained in detail and were accompanied by demonstrations.
This project focuses on the development of an ad-hoc, flat network with adaptive topology control. The ultimate objective is to test topology control algorithms on a real testbed and to include routing and media-access-control (MAC) provisions. In addition, the algorithms are envisioned to operate in distributed fashion. For the moment, the emphasis is on centralized algorithms and limited to logical connectivity control (clustering). The main idea is to adjust the clusters based on the changing traffic conditions. Presently, the investigators are using genetic algorithms as the basis for such topology control. The intended applications include stadium events or other similar gatherings of large numbers of users.
The objectives of this project are to develop adaptive ways in which application-level QoS is controlled in response to changing environments in multimedia applications. The current focus is on controlling the compression and transmission parameters of video streams (bit-rate, frame-rate, etc.) through negotiation among network agents that correspond to the different user streams. It is later expected that these negotiations will include allocation of additional network resources (such as buffer memory, bandwidth, etc.). The investigators do realize that end-to-end QoS performance depends on lower-layer (link) QoS parameters and intend to include this coupling into their methods. At the time of the site visit, there was focus on the use of motion-JPEG because of its scalability. Thus, other stream formats are converted to motion-JPEG before the adaptive negotiations are performed.
The goal of the project is an imaginative use of optical signal processing for adaptive antenna array applications. Unfortunately, the speed and complexity of the needed FFT computations limit most algorithms for beam forming. This project proposes to convert the microwave signals up to infrared optical bands, perform the FFTs through a lens (essentially instantaneously), and then convert the signal back down to its natural carrier frequency. The multitude of problems associated with this idea, laser frequency stability, non-reciprocal structures at transmitter and receiver for the up and down conversions, and the electro-optical integration, are currently being addressed. This project is an excellent instantiation of integration across layers in wireless networking. In this case, the hardware, the beam forming, and the signal processing are all coupled in a unified design approach.
This project is close to current mainstream research that is geared toward multi-user CDMA applications. It focuses on beam forming and tracking for both narrow-band and wideband cases for the simultaneous reception of several signals. Constant Modulus Algorithm (CMA) techniques, that are blind, as well as recursive tracking techniques (based on Kalman filtering) are used; and experiments are conducted with 44-tap transversal filters, 16 or 32 antenna elements, and simultaneous operation of 4 beams. The difficulty of multi-beam A-to-D conversion (in terms of needed processing bandwidth) is recognized and will be addressed.
True to the stated mission, these labs are focusing on long-term research. The technologies they are focusing on include the challenging one of ad-hoc networking and are illustrating a visionary integrated approach across layers. They also reflect perfectly the perceived need for spatial diversity, hardware/software integration, and coupling of QoS to lower layer networking issues.