Optical interconnect technologies have been drawing much attention for realizing ExaFLOPS scale high-performance computing systems. In this research, polymer optical waveguide devices for high-bandwidth-density interconnects are designed and fabricated in order to demonstrate their outstanding performances.

PEACH is a switching fabrics which build a supercomputer with a low latency direct communication between GPUs through PCIe.

Three main access paths to the Web are search engine, book mark and hyper-links. Each of these accompanies eco-system of information providers and consumers. WIX (Web Index) is meant to provide the fourth access path to the Web and we would like to exhibit the new eco-system based on WIX.

We will demonstrate RMX system, which extends the syntax of Email address to cooperate with SQL to achieve higher functionalities.

The evaluation of human emotions has been a multi-disciplinary area of research interest. Although there are several methods for such evaluation, such as subjective evaluation and behavioral taxonomy, direct evaluation from the human brain is more reliable. Electroencephalograph (EEG) signal analysis is particularly widely used because of its simplicity and convenience. In the site, we show the MANGA recommend system using the EEG.

In the Software Defined Transport Network which makes carriers’ large-scale networks dynamically manageable, a single management device controls all networks across multi-carrier. From the point of view of scalability, cooperation between management devices of each carrier is needed in the future. Therefore, we propose a secure cooperation method between management devices in the multi-carrier environment.

Recently, datacenter network architecture using optical switches is studied to reduce the power consumption. On the conventional method, the number of traffic flows is limited due to the low switching speed of mechanical switches. Therefore, we propose a novel datacenter network architecture using ultra high-speed optical switches to increase the number of flows that are able to pass through optical lines.

Network Function Virtualization (NFV) makes it possible to create communication services by using a virtualization of network node functions. We demonstrate “Atomic NFV” that realizes super-distributed virtual network devices.

Conventional optical network requires multiple optical edges which implement dedicated communication processing function for each network service such as IP and Ethernet. In our research project, single programmable optical edge can accommodate multiple network services using reconfigurable communication processing modules composed of FPGA/LSI/CPU fabric.

An integrated processing environment for a dedicated mark-up language, LMML, is introduced, which is intended to describe and visually analyze forensic big data towards the digitization of criminal investigation and court.

This booth exhibits IoT environmental sensing system which measures the weather, radiation or PM2.5 information using low power sensor and small solar cell. The data transmits through Wi-Fi or sensor network to collect or record by cloud system. IoT technologies visualize the data and analyze by tablet mobile device.

Imai laboratory studies on Interactive Intelligence to achieve smooth Human machine interaction. We investigate the mechanism of human cognition and design and adaptive system based on the findings of the human cognition. We prepare an interactive robot, and an autonomous mobile wheelchair and explain our technical motivation in this exhibition.

Arrayed-waveguide grating routers for newly developing waveband (T-band, 1000-1260 nm) are fabricated. It realizes low-cost, ultra large capacity transmission systems for data center networks.

We are researching vehicle-to-vehicle communication that vehicles communicate each other directly on the roads. For example, it is used for acquire the information such as accidents, traffic jam and so on faster and change the route. Moreover, we evaluate the application performance using network simulator and traffic simulator.

4K/8K broadcast is scheduled to be started towards the Tokyo Olympics in 2020. We have developed the world’s fastest plastic optical fiber that enables a real-time transmission of the staggering amount of information. Moreover, our photonics polymers bring innovation to the liquid crystal display structure which is getting bigger and more complicated.

System development using general-purpose robots takes a lot of time. If the intelligent software is incorporated, cost (human, monetary, time) will be even higher. Therefore, in order to facilitate the development of the intelligent system using a robot, we perform research and development for building tools and intellectual software.

We present our recent research on 3D object recognition and augmented visualization by 3D image sensing, which is based on computer vision technologies for understanding scenes from images captured with cameras. A live visualization system for visualizing hidden scene will also be presented as a demonstration of the augmented visualization.

This works aims at coexistence of real-time and IP communication in wide area LAN by synchronized TDMA. Periodic real-time communication occupies its communication path and transmission delay can be bounded. Additionally, throughput deterioration of IP communication is prevented by transmitting pause frame from switch.

Modulated lighting during the exposure of a photography superimposes the temporal information in motion blurs. The following image processing well designed for the modulation can extract the optical flow velocity and even acceleration from a single image.

We aim at realizing a content distribution mechanism called ZINK with efficient bandwidth sharing on a new generation network architecture called ZNA. ZINK enables content retrieval with location-independent content name for high content availability and enables in-network caching for efficient bandwidth sharing. ZNA provides ZINK with the following features: lower latency, broader bandwidth, and higher fault-tolerance than those in the current Internet.

We develop KANVAS, which collects various network information and provides it as well-systemized open data. KANVAS realizes network management with network knowledge, such as network monitoring and inference of the cause of network failure.

In the future content service, we believe that the size and number of contents will grow much larger. We demonstrate two services for the future content service in this exhibition.

We need to handle data more delicately and highly, while developing of AI or appearance of IoT. Although we treated computing and networking separately, these will be integrated IT platform by next generation’s IT platform so that users will focus on service.

We present a system that facilitates asymmetric collaboration among users with two different viewpoints in the design of living or working spaces, such as in the preparation of architectural schematics and floor plans. The system supports a pair of interaction techniques that facilitate communication between these two user viewpoints.

We present a bare hand grasp observation system called Wrap & Sense. We built a piece of band sensing equipment composed of infrared distance sensors placed in a line. The sensor band is attached to a target object with all sensors directed along the object surface and detects the hand edge partially. Type of grasp (e.g.,“power grasp”) is used in combination with the 3D shape of the object to determine the whole hand posture.

We proposed a stretchable sensor for measurement of body expansion. This sensor consist of a photo reflective sensor and spring. The photo sensor is embedded in the spring. When the spring is not stretched, the emitted IR light will reflect back to the phototransistor. However, when the spring is stretched, some of the emitted IR light can flow out through the aperture of the spring, reducing the light reflecting back to the transistor. Therefore, the more the spring is stretched, the less the amount of light would be detected. This method allows user to construct a stretchable sensor that is simple, low-cost and customizable.

AffectiveWear is a smart eyewear that can recognize a wearer’s facial expressions in daily life. The eyewear can classify eight facial expressions using photo reflective sensors and machine learning method. AffectiveWear can help understand a wearer's mind. The information can be used for a marketing research and a support system for older adults.