Intelligent Transport Systems (ITS) Allowing Creation of Smart Mobility Society

Innovative Methods for Observation and Control of Traffic Flow
ASAKURA Yasuo

ITS Technologies to Solve New Traffic Issues and Toshiba's Approach
OKI Yoshiaki / OZAKI Nobuyuki / NAKAMURA Junichi
The ongoing decline in the birthrate and aging of society in Japan have led to a deterioration in traffic issues including traffic congestion and accidents. An effective means of preventing age-related traffic issues is to provide information so as to avoid contingent events with sufficient time through the prediction of traffic congestion and accidents. There is consequently an increasing need for intelligent transport system (ITS) technologies with enhanced prediction simulation functions capable of predicting the occurrence of a variety of events in advance.
With the aim of realizing a smart mobility society achieving both smooth traffic flow and energy saving, Toshiba is continuously making efforts to improve the essential functions of ITS technologies. By promoting our "Spiral Life-Cycle Business" operations while accommodating customers' operation cycles, these ITS technologies are expected to contribute to the solution of social issues in cooperation with customers facing the daily changes that are taking place in the social environment.

Traffic Flow Estimation Using Image Probe System Based on Image Processing Technology
UENO Hideki / OZAKI Nobuyuki / ASAKURA Yasuo
Attention has been focused on methods of grasping road traffic conditions using a data processing function to process data gathered from roadside sensors and probe data from traveling vehicles based on global positioning system (GPS) information. However, the costs of device installation, operation, and maintenance have been pointed out as an issue. A further issue lies in understanding overall traffic conditions, including the traffic flow rate and traffic density, due to the lack of information on the positional relationship between a vehicle equipped with a GPS and other surrounding vehicles to complement the available information on the GPS-equipped vehicle's position and velocity.
In order to resolve these issues, Toshiba has been engaged in the development of an image probe system integrating a traffic condition estimation technology, which is being studied by the Tokyo Institute of Technology, and an onboard image processing technology to calculate the distances between vehicle heads as an input for the former technology. The image probe system makes it possible to estimate total traffic volumes on a targeted road and to construct an asset-light traffic system. We have conducted running tests using a prototype image probe system and confirmed the effectiveness of the system during driving on an actual traveling route.

System to Support EV Drivers Travelling on Expressways and Its Technology Using AI to Predict Power Consumption of EVs
SHIBATA Yasuhiro / FUKUSHIMA Arika / KANO Makoto
Toshiba has conducted simulations and confirmed the effectiveness of its energy management system (EMS) for expressways, which can control the electric power demand of an entire expressway including that for electric vehicles (EVs) and charging stations by systematically guiding each EV that requires battery charging to the appropriate charging station. We have subsequently been aiming to provide practical information to support EV drivers travelling on an expressway.
As part of this approach, we have developed a system with a function to recommend appropriate charging stations to EV drivers based on each predicted travel distance so as to avoid battery power shortages and shorten waiting times for charging. We are now evaluating it through field tests using a prototype system under various driving conditions. We also developed a technology to predict the power consumption of EV batteries with high accuracy using transfer learning, an artificial intelligence (AI) technology.

Efforts for Practical Realization of Method to Forecast Traffic Accident Occurrence on Expressways
SHIMOKAWA Yusuke / TANIMOTO Tomohiko / OHBA Yoshikazu
Rear-end collisions with vehicles that have stopped at times of traffic congestion account for many of the traffic accidents that occur on expressways. The provision of traffic information to drivers when there is a high probability of accidents occurring can therefore be expected to contribute to the prevention of such accidents.
In order to reduce traffic accidents, Toshiba has been engaged in research and development aimed at the practical realization of a method to forecast the occurrence of traffic accidents on expressways by means of a self-organizing map, which is a type of neural network, and to provide such information to traffic control personnel and drivers. We have now conducted studies on an actual case of applying this method, called the traffic accident occurrence forecast module, to a traffic control system, and have extracted issues and proposed measures for implementation of the module.

Basic Software for Central Processing Systems of ITS with Appropriate Performance for Expressways in Global Market
SHIMIZU Ryogo / YUHARA Takashi / YOSHIDA Hideaki
An intelligent transport system (ITS) for expressways, which consists of a traffic control system, facility monitoring system, and toll collection system, is comprehensively controlled by software installed in a central processing system. Accompanying the ongoing increase in traffic in various countries, particularly in developing countries, there is an urgent need to introduce the ITS into expressways in order to enhance safety and mitigate traffic congestion.
To meet the diverse requirements of the global market, Toshiba has developed the Global ITS Central Software as basic software for the central processing system of an ITS, incorporating the results of market research in each region together with its accumulated ITS technologies and operational know-how in Japan. This software has a compact configuration, which was achieved by narrowing down ITS functions to the necessary level appropriate for each region, thereby shortening the delivery time and reducing costs. It is also compliant with standard interface specifications, allowing external equipment produced by other companies to be easily connected. We have applied this software to the North-South Expressway Construction Project connecting Ho Chi Minh and Dau Giay in Vietnam, the Japan's first overseas ITS package project, for which we received the contract in March 2014. The system was successfully put into service in March 2017.

Vehicle Type Discrimination Technology for Large-Sized Vehicles Using Roadside Camera Images
AOKI Yasuhiro / SATO Toshio
A large number of roadside cameras have already been installed for the purpose of monitoring road and traffic conditions. In recent years, demand has been increasing for the discrimination of extra-large vehicles, including large-sized buses and semitrailers, using images captured by such roadside cameras in order to allow road administrators to estimate road maintenance requirements and the state of congestion of service areas and parking areas in advance. In order to discriminate different types of large-sized vehicles in detail, however, observation of the overall vehicles is required and it is difficult to discriminate large-sized vehicles that have different rear structures using only the front patterns captured by these cameras. Moreover, the side patterns of vehicles change in appearance according to the differences in camera locations.
To rectify this situation, Toshiba has developed a vehicle type discrimination technology for large-sized vehicles in which a pattern matching method is applied to a number of vehicle images decomposed into front, side, and top patterns based on the coordinate systems of the vehicle relative to the position of each roadside camera. We have confirmed the effectiveness of this technology through experiments on the discrimination of large-sized buses using recorded images of a traffic flow.

False Approach Detection System to Prevent Accidents at Entrances and Exits of Urban Expressways
KURATA Ryoichi / WATANABE Yasunari / WADA Akira
Traffic accidents may occur at the entrances and exits of urban expressways due to the false approach of a pedestrian, bicycle, or vehicle that is not permitted to run on expressways, including motor-assisted bicycles and motorcycles under 125 cc, because the entrances and exits of urban expressways are often connected to the main line by a road of relatively short length and it is sometimes difficult to distinguish between the ramp and public road.
In order to prevent such traffic accidents, Toshiba, in cooperation with Metropolitan Expressway Co., Ltd., has been engaged in the research and development of a false approach detection system consisting of the following subsystems: a detection subsystem to automatically detect an erroneous entry, an on-site warning subsystem to issue a warnings by voice and display, and an alarm receiving and popup display subsystem as well as a video recording subsystem installed at the traffic control center. We have confirmed the effectiveness of this system through experimental trials using an actual entrance and exit of an urban expressway.

Software Platform Capable of Lengthening Lifetime of ETC Systems
YAMAMOTO Masahiro / KUSANO Atsushi / YAHAGI Masayuki
The basic policy released by the Ministry of Land, Infrastructure, Transport and Tourism mandates the optimal utilization of roads in Japan, particularly expressways, through enhancement of their management and small-scale improvements. It is therefore necessary to implement strategic measures for the maintenance and renewal of expressways as a social infrastructure system. The lifetime of the equipment of an Electronic Toll Collection (ETC) system, a key element of an intelligent transport system (ITS), is expected to be become longer as a result.
Toshiba has been developing and supplying ETC systems in response to such customers' requirements. However, the life cycle of electronic parts in hardware is becoming progressively shorter in recent years. Furthermore, as embedded software that can perform billing processing in real time with high efficiency is closely related to the hardware used, time and labor are required to implement the software design change accompanying a change in hardware. The lifetime of an overall ETC system and its equipment is consequently affected by the life cycle of the hardware. With this as a background, we have now developed a software platform that can lengthen the lifetime of ETC systems by minimizing the effects of changes in hardware. This software platform, which is divided into core parts for common processes and noncore parts related to hardware changes, can flexibly and efficiently respond to changes in hardware as well as the introduction of automated verification.

Wireless Charging Technology for Expansion of Electric Bus Systems
SUZUKI Katsuyoshi / OBAYASHI Shuichi
In response to global environmental issues, Toshiba is promoting urban transportation solutions for the realization of a low-carbon society through the development of electric bus systems. A wireless charging system capable of easily charging secondary batteries mounted on electric buses will play a critical role in the expansion of electric bus systems. In the development of such a wireless charging system, it is necessary to take into consideration both safety and usability as well as interoperability for different types of electric buses.
We have been developing a magnetic resonance technology that can handle a wide margin of positional displacements between an onboard battery charger and a parking pad on the ground and relatively long transmission distances. Verification operation tests on public roads have confirmed that this technology is capable of meeting the requirements for practical use in an electric bus system.

85 kHz-Band Wireless Power Transfer Technologies toward Practical Realization of Contactless Charging of Electric Vehicles and Buses
OBAYASHI Shuichi / MATSUSHITA Akihisa / ISHIDA Masaaki
In order to simplify the process of charging secondary batteries mounted on electric vehicles (EVs), plug-in hybrid electric vehicles (PHEVs), and electric buses without the need to attach and detach a cable, attention is being increasingly focused on contactless charging through power transfer between an onboard battery charger and a parking pad on the ground.
Toshiba is making efforts to develop wireless power transfer technologies using frequencies in the 85 kHz band, one of the candidate frequency bands for the international standardization of contactless charging specifications. Based on these technologies, we have developed a 7 kW wireless charging system for EVs and conducted secondary battery charging tests for verification of a prototype system mounted on an EV. We have also developed a 44 kW wireless charging system for electric buses, which achieves a reduction in radiated electromagnetic emissions taking interference with peripheral devices into consideration while providing a charging speed almost equal to that of conductive quick chargers, and have conducted verification operation tests using electric buses running on public roads around Haneda Airport. The tests have verified that these wireless charging systems offer enhanced convenience and achieve the targeted power transmission efficiency.

BLE Wireless Receiver Using Digital PLL Regeneration Technology
SAI Akihide / OKUNI Hidenori / FURUTA Masanori
As a consequence of the expansion of battery-driven Internet of Things (IoT) devices, reduction of the power consumption of the wireless transceiver in such IoT devices is an important issue in order to lengthen the battery lifetime. In recent years, attention has therefore been increasingly focused on the Bluetooth^® Low Energy (BLE) standard, a technology for low-power applications including IoT devices, and efforts to realize power-saving technologies for BLE wireless transceivers at the time of data reception are being actively promoted. However, these technologies are vulnerable to radio wave interference.
Toshiba has now developed a new low-power BLE receiver based on a digital phase-locked loop (PLL) technology. The new receiver achieves higher interference tolerance and data demodulation using a novel single-path (in-phase channel) demodulation technology, which were made possible by incorporating a dual-loop structure consisting of analog and digital loops and by shifting the synthesizer frequency. Furthermore, the use of analog-to-digital converters (ADCs), which are necessary in conventional power-saving technologies, has been eliminated. Experiments on a prototype receiver chip have verified that it achieves more than 10 times higher interference tolerance and reduces power consumption by almost 10% compared with a receiver using the latest conventional technology.

Polyspector™ High-Speed Big Data Visualization Platform
LI Xinxiao / MATSUZAKI Hidenori
In line with the progress of the advanced information society, demand has recently been expanding for the collection and utilization of big data in various business fields. An effective means of rapidly and intuitively grasping trends and potential problems in such data is to visualize the data according to narrowed-down conditions.
In this context, Toshiba has developed Polyspector™, a big data visualization platform incorporating its proprietary database engine for dedicated visualization and an optimal aggregation algorithm. Polyspector™ makes it possible to perform big data visualization processing at high speed through both high-speed extraction of partial data necessary for processing by means of the database engine and reduction of the extracted data to the minimum level necessary for visualization by means of the optimal aggregation algorithm. We have conducted verification tests with more than a hundred million data processed by interactive operation using a personal computer, and confirmed that the response speed of the visualized display in even the worst-case performance is improved more than a hundredfold compared with that of conventional tools while reducing the response time required for the display updating operation to only 1 to 3 seconds.

Speech Recognition Engine with Detection and Removal of Filler Words Unnecessary for Understanding Spontaneous Speech
FUJIMURA Hiroshi / MASUKO Takashi / NAGAO Manabu
The performance of speech recognition has recently become increasingly sophisticated due to the application of deep learning technologies. However, the performance level in the case of recognition of spontaneous speech in daily communications is much lower than that of read speech including human-to-machine speech. Performance deterioration associated with filler words inserted into spontaneous speech, such as um and ah, is a serious technical issue that still remains to be solved. Furthermore, these filler words hinder understanding of the meaning of speech recognition results.
Toshiba has been actively focusing on the development of a speech recognition engine with enhanced robustness to spontaneous speech in order to provide users with easy-to-understand speech recognition results. We have now developed a decoding algorithm for speech recognition capable of detecting and removing filler words by applying an acoustic model that can simultaneously implement acoustic event detection and phoneme recognition, and have incorporated this algorithm into our speech recognition engine. We have conducted verification tests and confirmed the robustness of this speech recognition engine to spontaneous speech as well as high-performance filler word detection and removal.

Energy-Saving Heating Control Technology for Residences as Part of Lyon Smart Community Project in Lyon, France
YANO Toru / HUANG Chingchun / IMAHARA Shuichiro
Toshiba has been selected by New Energy and Industrial Technology Development Organization (NEDO) as a contractor for the Lyon Smart Community Project in the Lyon Confluence redevelopment district, France, and has introduced its latest home energy management system (HEMS), called OMOTENASHI HEMS™, into a newly constructed building named the Hikari Building.
OMOTENASHI HEMS™ incorporates a heating control technology effective for energy saving with the following functions: (1) a temperature calculation and setting function to ensure the comfort of residents and save energy, and (2) an automatic control function to change the room temperature by predicting the behavior of residents. We conducted verification tests in December 2016 and January 2017, and confirmed that these functions achieve an energy saving rate of 23.6% and 18.6%, respectively.

High-Power Type 10 Ah SCiB™ Lithium-Ion Battery Contributing to Reduction of Carbon Dioxide Emissions
MURASHI Yasuaki / YAJIMA Akira
The use of secondary batteries has recently been expanding in the automotive field, including vehicles with an idling stop and start (ISS) system, micro hybrid vehicles, and 48V mild hybrid vehicles, as well as in the fields of rolling stock and industrial instruments, due to the demand for efficient use of regenerative energy from the standpoints of improving fuel consumption and reducing the burden on the environment.
In response to these market needs, Toshiba has developed a high-power type 10 Ah SCiB™ lithium-ion battery that can be used in a broad array of applications requiring operation under high- and/or low-temperature conditions, and started to produce it in commercial quantities. Featuring an improved cathode material and electrolyte, the 10 Ah SCiB™ cell provides both superior durability under higher temperatures and superior input-output power characteristics under lower temperatures compared with the existing 2.9 Ah SCiB™ cell. We have confirmed the effectiveness of the 10 Ah SCiB™ cell from the results of a storage test assuming the conditions of a high-temperature region, in which the internal resistance after 15 years was shown to increase only to about 120%. This cell also has superior cranking performance under low temperatures. Furthermore, a cycle test at a temperature of -20°C has confirmed that the charge and discharge capacity is about 65% of the discharge capacity at a 1C discharge rate and a temperature of 25°C.