Site: Nippon Telegraph and Telephone Corporation
Wireless Systems Laboratory
1-1 Hikarino-oka, Yokosuka-shi
Date Visited: nbsp;2 June 1999
WTEC Attendees: J. Maurice (report author), T. Itoh, L. Katehi, J. Winters, D. Friday, M. Iskander, L. Young, B. Mooney, (also present was Ms. Fumiko HAZAMA, serving as interpreter when required)
Hosts: Dr. Hideki Mizuno, Executive Manager, Wireless
Systems Laboratory, NTT Network Innovation Laboratories
Mr. Masashi Shimizu, Deputy Executive Manager NTT Network Innovation Laboratories
Dr. Masayoshi Tanaka, Executive Research Engineer, NTT Science and Core Technology Laboratory Group
Dr. Shuji Kubota, Senior Research Engineer, Supervisor, Wireless Systems Laboratory, NTT Network Innovation Laboratories
Mr. Toru Otsu, Senior Research Engineer, Supervisor, Wireless Network Control Research Group, Wireless Systems Laboratory
Dr. Masahiro Muracguchi, Senior Research Engineer, Supervisor, Wireless Communication Electronics Research Group, Leader, Wireless Systems Laboratory, NTT Network Innovation Laboratories
Mr. Katsuhiko Araki, Senior Research Engineer, Supervisor, Satellite Communications Electronics Research Group, Leader, Wireless Systems Laboratory
Dr. Toshio Nojima, Senior Executive Research Engineer, Wireless Circuits Laboratory Wireless Laboratories
The visit to NTT R&D Laboratory, Yokosuka Site, was very hardware-focused, with presentations by every director of wireless communication activities except for one. Each department's focus and activities were briefed, and Dr. Mizuno, executive manager and the panel's key host, represented the single absent member. Representatives from the commercial side of NTT did not attend.
The technical presentations were followed by a variety of NTT-developed system demonstrations. These were Cyber Communication Lab Group focused, because cyber-communications (mass media products and technologies) is a Yokosuka site specialization. NTT's Yokosuka site houses both the Cyber Communications Laboratory Group and the Network Innovation Laboratories.
At NTT, R&D on advanced wireless communication technologies for transmission systems is grouped according to whether it is (1) software-defined radio (SR) or (2) local information systems. Regarding these larger categories, NTT explained in brief the following research areas (departments) within them:
These areas of research are reported upon individually in what follows, together with the introductory overview NTT presented.
NTT employs over 3,000 people in three facilities. The staff at the Yokosuka site visited numbers about 1,000.
As mentioned, R&D on advanced wireless communication technologies are grouped as follows at NTT:
Conventionally, radio technology is analog-specified, digital-circuit based, and fixed function. The distinguishing feature of SR, on the other hand, is that SR is general-purpose-programmable3/4a digital circuit, but with software. SR is thus a multi-standard, flexible wireless system. It operates to and from a multi-standard (PHS, PDC, AWA) terminals and base stations and features multi-mode operation, that is, a seamless operation in multiple systems, bridging different standard wireless systems. This is because it is based on adaptive transmission technologies, employing an adaptive antenna, adaptive modulation, coding, RF frequency, and bit rate.
NTT's local information system is called "P-POINTS." The first "P" in "P-POINTS" is for "positioning," and "points" indicates the high positioning accuracy that is attainable3/4within 10 m. The P-POINTS system is configured as follows: P-POINTS links, by radio, to a PDA (personal digital assistant) equipped with a PHS (personal handy system), which in turn links to a base station and, through an ISDN public network, a host node. The following are parameters of the P-POINTS system:
P-POINTS is a hybrid system employing SR. Its applications are for indoor and outdoor high accuracy, seamless positioning. Deployment of its service area is by self-multiplication via a wireless network. A schematic shown at NTT linked GPS satellite to PHS to very many P-POINTS (SR markers).
The purpose of the first and next-generation systems described below is to provide an interactive MM satellite network service. The main schematic shown featured three key subsystems: (1) high-speed Internet, (2) multicast communication, and (3) multicast transmission.
The first-generation MM satellite communication system is configured as follows: a central Earth station (CES) uplinks to a communication satellite, the satellite links to an Earth customer station, and the customer station back to the central station (CES) via either Internet or PSTN/ISDN (Public-Switched Telephone Network Integrated-Services Digital Network) links. The CES is the network control center. It consists of an MPEG encoder, multiplexer plus scrambler, an ATM router, a video server, and guest workstations. In the first-generation system, the satellite links to the user/customer with an antenna, PC, and TV at 30 Mbps.
First-generation satellite MM system parameters include the following:
The second-generation system is a portable MM satellite communication system. By means of it, NTT will expand the service area to where there is no available terrestrial link. The projected system configuration consists of a communication satellite, a portable user terminal, a CES with a forward-link access server. The user terminal will use the Ku band. It is portable and highlights the mobility of the user. (The Ku band antenna can be atop a car, etc., and a video demonstration was given.) The user terminal also will be capable of realizing bi-directional communication via satellite, and have flexible satellite backward high-speed link. The user will not have to point the antenna at the satellite. The uplink (forward link) is SS-FDMA 9.6 kbps (x 256) at 100 mW, and the backward link is TDM (8 Mbps).
The miniaturization of MMIC technology at NTT started 20 years ago. Ten years ago NTT started using copper in its MMIC designs and devices, and now researchers have progressed considerably towards 3D-interconnect technology that further improves performance, reduces chip size, and, hence, reduces cost. NTT moved from the standard microstrip MMIC device to a uniplanar (2D) one with 1/5th the area for a function and 1.5 times the 2D's depth. Under their new 3D MMIC technology, the same functionality is achieved on 1/20th the area of 2D MMICS, and on 3 times the thickness. The chip consists of (1) thin-film microstrips with narrow line widths (5-30 (m) and narrow line-width spacings, (2) it has grounding and shielding metal layers, (3) dielectric layers, and (4) ground wires3/4linking to logic, and all on a substrate. The U-band single-chip down-converter measures 1.78 mm x 1.28 mm. It has a conversion gain of 0 dB +/- 1.5 dB and an image rejection ratio greater than 15 dB (at 42.5 GHz to 47.5 GHz). The K-band, Si-based 3D MMIC has an 0.70 mm x 0.46 mm amplifier coupled with 0.46 mm x 0.42 mm mixer. The mixer is based on analog circuitry (it is not digital). In heterodyne-down converting, it averages a -10 dB conversion loss in the 10-30 GHz frequency band of the local oscillation signal.
Considerable cost savings are projected with the miniaturization of K-band Si 3D MMICS, in particular. NTT showed a comparison chart of fabrication cost-saving projections for its Si masterslice MMIC. The cost of Si masterslice MMIC is less than 1/10 the cost of conventional GaAs MMICs. (Additional information on the matter: the GaAs 3D MMIC has 1/2 the cost, the GaAs masterslice MMIC has 1/2 to 1/4 the cost, while the Si masterslice 6%-12% the cost.)
In summary, these are the key features briefed: The area for a function is 1/3 to 1/20 that of conventional 2D MMICs. The application frequency is improved up to 65 GHz. Using masterslice MMIC design methodology, the turnaround time from design to fabrication is less than 1/2 that of 2D MMICs. The simple and parasitic-free MMIC design is compatible with CAD. The process is 3D interconnect technology using 6 metal layers with polyimide as the insulator. Further regarding the structure, it is based on 4 layers, of 10-12.5 (m thickness each, containing both lumped and distributed elements. The logic devices are GaAs MESFETs and Si BJTs. The capability is (1) mixed-mode IC, (2) millimeter-wave Si MMIC, and (3) "radio on a chip."
The first item of discussion addressed aspects of the MMIC structure. Although there is vertical isolation, still problems still occur at high frequencies. Vias are used, but parasitics remain. NTT reported that it uses a groundline at every layer to equalize the potential. Next discussed was NTT's BFN MMIC technology, that is, MMIC technology that achieves beamforming. NTT has developed a 2.5 GHz-band BFN MMIC. The input signal is split into 32 components. All components of the 32 amplitudes and phases can be controlled individually. The dimensions of the MMIC are a mere 11 mm x 11 mm, and it can talk to a satellite. The last technology item discussed was an "intermodulation distortion controller for HPA." A new type of intermodulation distortion control circuit is proposed.
From the provided schematic: Under beam-forming and control technology, high-gain antennas link together to cover segments of a geographical area. Discussed was a large-scale antenna reflector with an active phased-array feed. The active phased-array feed consists of (1) a small and lightweight beam-forming network (BFN) that takes RF input and, with input from a beam controller (above), feeds (2) a high-efficiency power-amplifier array that drives (3) the array of primary radiators. The BFN is MMIC-based.
A 2.5 GHz-band BFN-MMIC has been developed. Its input signal is divided into 32 signals, and the amplitudes and phases of the 32 signals can be individually addressed and controlled. The dimensions of the MMIC are 11 mm x 11 mm.
A high power amplifier with even-order-distortion control has been developed.
NTT antennas for high-speed wireless access include (1) rod-type printed antennas, (2) a multi-sector monopole Yagi-Uda antenna, and the "smart antenna."
The rod-type are 6-sector, small-sized antennas for operation in the 25 GHz band. There are three types of rod antennas. In each, microstrip antennas are printed on panels within the rod, the beam radiating out from a cylindrical, disk-like radome atop it.
The multi-sector monopole Yagi-Uda antenna (MS-MPYA) consists of 2 very low profile 12- and 6-sector units (120 mm and 60 mm diameters respectively) for operation in the 19 GHz band. The 12-sector unit functions as the control module with a gain of 14 dB, and the 6-sector unit functions as the user module with a 10 dB gain. Experimentally its beam-forming performance maps closely to calculated predictions of the beam pattern.
The prototype smart antenna uses an adaptive technique for indoor high speed wireless communication systems. In indoor propagation environments, multiple-reflected waves significantly degrade the transmission quality in wideband wireless systems. However, NTT has demonstrated an adaptive array that completely suppresses the multi-path waves, regardless of the incoming waves' numbers in the indoor environment. By adaptively pattern controlling itself, the array suppresses the undesired long-delayed wave effects, thus lowering BER and achieving improved transmission quality over pencil-beam antenna systems. Also, compared with sector antennas, under same beam-width conditions, the Smart Antenna achieves lower BERs (from 10-2 to 10-6, and a 38 Mbps data rate). This is because the Smart Antenna can steer and adapt for not only the main beam direction but the null as well. Furthermore, by enhancing the desired signal by combining short-delayed waves compared with a sector antenna, the smart antenna's adaptive technique is effective in miniaturizing base station antennas.
At present NTT's smart antenna is a prototype, awaiting the appropriate A/D converter and signal-processing device to enable it as a commercial product.
(As an aside, Dr. Itoh, at this point, introduced the UCLA-developed quasi-Yagi-Uda antenna, which was reported during an antenna propagation conference in July 1999, and he briefly showed its data, etc.)
NTT has developed several deployable on-board antennas3/4solid, wire mesh, and inflatable. The current non-inflatable mesh-structure model has seven modules and a 10 m diameter and weighs about 80 kg. The target for the future has 14 modules, with a diameter of approximately 14-15 meters, and weight of 120 kg. The panels of this type fold out in a deployment sequence. This technology was successfully transferred to NASDA, Japan's space agency. An on-site anachoic test chamber is available for testing the structures. An anachoic chamber is available at another NTT facility, as well. The problem is the long focal length required.
Topics of discussion included the inflatable structure itself. To form a reflector plane for the inflatable structure, NTT sees an inflatable polymer-based film as the most promising for realization of this.
The Engineering Test Satellite (ETS) Experiment is a series of seven testing and experimentation satellites aimed at developing common-base communication technologies. The series is produced by NASDA, Japan's NASA organization. The series ETS-VIII will be launched in 2002.
ETS-VIII will establish and verify the world's largest geostationary satellite bus technology. It features 10-meter-plus deployable antennas with phased array feeds and is an advanced 3-ton-class spacecraft bus with the world's largest and most advanced large-scale deployable reflector. It will enable technology for mobile, multimedia audio/data communications with CD-level quality sound and image transmission from hand-held terminals. These will be similar to popular cellular phones.
The ETS series is a big national project with many large companies involved to cooperate and implement it. The relationship between NTT and NASDA is more than "just business." NTT participated in the ETS-V, ETS-VI and ETS-VII, and ETS-VIII projects; NTT's mission objectives are (1) to provide the beam forming network (BFN) component technology and (2) to assist in the analysis for deployable antenna technology. As mentioned in a previous section, NTT's technology for BFN uses monolithic microwave integrated circuits (MMICs). NTT also developed software that calculates precise deployment positioning motion taking into account the elasticity of the truss and the cable structure. Key NTT technologies for ETS-VIII can be summarized as (1) NTT designed the subsystem for ETS-VIII and (2) NTT provides the technology for BFN and large deployable antennas in Japan.
Advanced 3D MMICs & Smart Antenna technologies
The following were among those items demonstrated:
The demonstrations highlighted the focus of the particular facility visited at Yokosuka, that is, highlighted was the information-based research areas of the Cyber Communications Laboratory Group housed there. This group performs R&D of service foundations and solutions technologies and cooperates in joint projects focusing on various software servers, terminals, etc., related to business-oriented information sharing. R&D encompasses essential technologies as content creation, support technology, media technology, database technology, etc., which contribute to the advancement of business-oriented information sharing.
Other research areas of the Cyber Communications Laboratory Group include the following:
Also at the Yokosuka site are the Network Innovation Laboratories. As suggested, this organization does R&D on network systems based on cutting-edge telecommunications methods. They were represented at the site visit.