Toward Realizing a Safe, Secure Autonomous Driving Society

While the word "mobility" suggests more than just a means of transportation, the world is shifting toward autonomous driving with the aim of efficiently using time, energy, and mobility assets as well as reducing traffic accidents. The realization of autonomous driving requires advanced technologies for recognizing and analyzing all kinds of traffic situations quickly to enable vehicles to drive safely and smoothly. In this article, we explore mobility technology and business issues associated with the coming era of autonomous driving.
[Fujitsu Forum 2017 Conference Report]

Development of New Technology for a Society with Self-driving Cars Underway

The first half of the conference consisted of presentations by VMware, Denso Corporation, and Fujitsu Limited. They introduced their activities with a focus on the latest technologies being key to realizing a society with self-driving cars.

AI Realizes IoT Technology for the Connectivity Era

Shikou Kikuta
Corporate Executive Officer
Fujitsu Limited

It has been predicted that more than 10 million connected cars will be manufactured annually in the future (*). Autonomous driving can be divided into various levels, and by around 2025, it is predicted that fully self-driving cars making use of artificial intelligence (AI) will appear. To move forward along this roadmap and turn the predictions into reality, the various underlying technologies must be combined into a platform. We, Fujitsu Limited, plan to offer such a platform, including various services, as a Mobility IoT Solution.

We are working with car manufacturers and suppliers to further drive technical innovation in the following five areas: automobile big data, cloud AI, dynamic map management, OTA reprogramming (**), and In-Vehicle security management. Allow me to introduce some of Fujitsu's unique technologies.

Overview of the Mobility IoT Solution for achieving connectivity

With the increase in the number of connected cars, data traffic volume is expected to grow explosively. This will cause high network traffic loads and incur data center costs, which has led Fujitsu to develop a unique technology for compressing high-volume image data and images seen from cars while retaining image characteristics. Experiments are now underway to commercialize this technology, which achieves a compression rate ten times better than the H.26x compression standard.

As for cloud AI, we will offer a platform that supplies In-Vehicle AI with knowledge obtained through analysis and learning of images and sensor information from the car. We will employ one of the world's fastest deep learning platforms to support automatic generation of teaching data for learning and to improve learning efficiency.

Since connected cars have many functions tied to networks, they will become cyberattack targets from all directions. In addition to key management, which supports message authentication inside and outside the cars, we will use AI as the latest security technology to promptly protect cars from new attacks.

Fujitsu has already expanded our SPATIOWL service into the international market. SPATIOWL enables smooth provision of location information services, traffic analysis, and fleet management and sharing services. We alone cannot realize all solutions and services for the new type of mobility. Rather, we hope to work with many key players to build a society with self-driving cars. We believe this will be the result of a process in which new value is generated through co-creation.

*: An automobile equipped with ICT device functions
**: Over-the-air (OTA) program updates and reprogramming using wireless networks

Roles of Device Management in Efforts to Create Connected Cars

Chikara Furuichi
Vice President
APJ Global and Strategic Accounts
VMware, Inc.

VMware is working with Fujitsu in the IoT field, primarily by focusing on an OTA reprogramming solution for in-car software. As part of this collaboration, we have proposed VMware Pulse IoT Center as a solution that enables easy, centralized management of complete IoT environments. Our technical intelligence building upon our experience in mobile device management can be used in car management for the connected car era.

In OTA reprogramming, data goes through six points starting from edge devices, which are the entry-point IoT devices, through data center. The IoT gateway must be capable of supporting changes dynamically since it connects local systems with the data center.

Key points in end-to-end IoT connectivity

For example, software technology will be able to transform a 2WD car into a 4WD car through OTA reprogramming. By creating business logic packages comprised of compartmentalized software implementations, core packages for 2WD and 4WD can be managed and used as necessary.

VMware Pulse IoT Center maps IoT gateways on a world map to visualize which devices are to be updated. I believe that high quality edge device management through enterprise IoT management on a global scale will play an important role in future IoT.

Realizing a Society with Safe, Secure Mobility through AI-enabled Technology

Hajime Kumabe
Executive Director
Denso Corporation

We, Denso Corporation, aim to build an advanced safety system--a system that provides convenience and comfort while empowering cars to realize an enjoyable driving experience. This system builds on the "Everyday Confidence" guaranteed through cognition, decisions, and operation support during driving as well as "Extraordinary Safety," which avoids dangers and reduces damage due to accidents. To realize such an advanced safety system, we are working not only on fundamental research but also to comprehensively develop technologies that connect people, cars, society, and infrastructure. My presentation today focuses mainly on our efforts to improve image sensors.

Conventional image recognition technology converts the characteristics of people and cars into patterns artificially, compares them, and recognizes objects from the image data. To realize a society with self-driving cars, we are developing an AI-enabled image recognition technology called Deep Neural Network (DNN) as our next generation image recognition system. DNN can independently learn objects' characteristics and then recognize them. This enables the entire driving landscape to be grasped in real time.

Image recognized with conventional technology (top) and image recognized using DNN (bottom)

Autonomous driving will become possible once this image recognition system is combined with a function to recognize empty space where the car being driven can move into and a technology for predicting the upcoming movements of pedestrians and other cars. To develop DNN, we are collaborating with other companies and universities in the areas of algorithms, learned data, and an environment for analyzing learned data. By incorporating connected technologies, we aim to create a safe, secure, and comfortable mobility system that can amaze drivers.

Co-creation Is Key to Realizing a Society with Self-driving Cars

The second half of the conference was an active discussion. With Mr. Tetsushi Hayashi from Nikkei Business Publications, Inc. as the moderator, the speakers discussed issues regarding the realization of safe, secure autonomous driving as well as their expectations for a society that featured new types of mobility.

Co-creation with Other Business Fields Hastens Technical Development

Mr. Hayashi, the moderator, opened by asking each speaker what their company's strengths were for realizing a society with self-driving cars. Shikou Kikuta from Fujitsu Limited answered: "Based on the technology we have acquired through our experience in the semiconductor and mobile device industries, Fujitsu is now working with companies in the automobile industry to study and develop mobility technology for the connectivity era. Also, mobility IoT should benefit from the speed and reliability of both the cloud and parallel processing, which we achieved by developing the TSE stock trading system and our experience with the K supercomputer."

The generation of new value through co-creation among multiple players is important for realizing a society with self-driving cars. Hajime Kumabe from Denso Corporation re-emphasized the necessity of collaboration beyond industry boundaries by commenting: "Early realization of a safe, secure mobile society cannot be achieved by the automobile industry alone. We are therefore working with various companies. For example, IT suppliers have advanced software version management. We will be sure to appreciate and incorporate IT companies' knowledge while deepening our collaborative relationships."

Leading the World with Japanese Mobility IoT Technology

During the discussion, Mr. Hayashi, the moderator, also emphasized that security technology was indispensable for mobility IoT. Chikara Furuichi from VMware responded: "We must consider IoT devices, including cars, to be potential cyberattack targets. Recent security technology tends to be designed to localize damage when an attack occurs." Shikou Kikuta from Fujitsu Limited added: "Awareness of security has risen in the field of mobility IoT. In terms of security threats, we plan to offer a security library that supports OTA reprogramming as well as to create a framework that facilitates prompt updating as necessary."

The discussion touched upon global mobility IoT in addition to that in Japan. From the perspective of someone in charge of the Asian Pacific market, Chikara Furuichi from VMware stated: "Asian companies are carefully observing what kind of roles the Japanese automobile industry will play in realizing a society with self-driving cars. We should be the leader in Asia, and we should do so without falling behind Europe or the US. In this regard, I believe IT suppliers like us can help much more."

Finally, Shikou Kikuta from Fujitsu Limited concluded by stating: "Although Fujitsu does not manufacture cars, we will compete in the global market by offering a safe, secure platform. Although much news comes in from overseas in the area of mobility IoT, there are still many areas where Japanese companies can lead. We will work with automobile manufacturers and suppliers as well as various vendors to navigate the challenges."

Speakers
  • Shikou Kikuta Corporate Executive Officer
    Fujitsu Limited

  • Chikara Furuichi Vice President
    APJ Global and Strategic Accounts
    VMware, Inc.

  • Hajime Kumabe Executive Director
    Denso Corporation

  • Tetsushi Hayashi Chief Research Officer
    Cleantech Institute
    Nikkei Business Publications, Inc.