Spread of Autonomous Driving Creates Growing Large Market - Knowing Technology to Capture Business Opportunities [Part 2]

Development of automatic driving systems is now advancing rapidly worldwide. Many features have been developed, including automatic braking, cruise control (an automatic speed control system that maintains a set speed without using the accelerator pedal), and obstacle avoidance; however, some major challenges remain, such as security issues. We asked Shinpei Kato, general producer of the program and associate professor at the University of Tokyo as well as a developer of open-source autonomous driving systems, how to overcome these challenges.

Barriers to Overcome to Achieve Level 3

Many companies and research institutes are developing autonomous driving systems. However, their development remains at "Level 1," where just one of the functions of acceleration, braking, or steering has been automated, or at "Level 2," where two or more of these functions have been automated. "Level 3" automation means that the vehicle can take over all Level-2 driving functions and the human driver need only take control in an emergency. "To reach such a level, we must overcome many difficulties," Kato said.

Shinpei Kato, Associate Professor, University of Tokyo

Kato Lecturing at the TierIV Academy

"Self-driving capabilities require human drivers to take control only in emergencies. Thus, the driver's role will shift from conventional 'driving' to 'monitoring.' Even so, the driver must always understand the outside situation and state of the driving system, but doing so is difficult. Since first of all it is automatic driving, the driver should be able to concentrate on other tasks, and it is unrealistic to expect the driver to instantly judge the situation and control steering in response to an emergency request from the car."

Detecting Danger by Incorporating Gamification into Monitoring Behavior

Demonstration of the RoV monitoring system

Use of virtual reality (VR) is said to be one possible solution. Kato has developed a monitoring system using RoV, an advanced form of VR, in order to discover clues leading toward a solution. RoV is the abbreviation of Real-oriented Virtuality. With RoV technology, users can feel virtual textures and see virtual objects in a real landscape by wearing a head-mounted display. For example, the technology provides a mechanism to display a wall on-screen to prevent the user from proceeding when the traffic light is red, and it displays holes on the ground to alert the driver when there are pedestrians in the street. "Going forward, a mechanism to improve drivers' and passengers' monitoring awareness must be developed by incorporating elements of gamification," said Kato.

Participants carrying out an exercise

"While promoting the development of Autoware, which is open-source autonomous driving software, I also focus on developing services and software based on these systems. To tell the truth, we do not aim to install systems we ourselves have developed onto vehicles sold all over the world. Because Autoware is an open-source software product, we cannot profit from it as a business. It would also be difficult for us to become a dominant player in the market for autonomous driving systems by directly competing with companies such as Google and Microsoft, which have strong technical capabilities and capital bases."

Safety Assurance Necessary for the Development of Autonomous Driving

A sensor installed on an autonomous vehicle

What should autonomous driving aim to achieve in the future? For example, Android, an open-source OS installed on smartphones, acquired its market share by enabling users to freely develop applications. Kato says that, in the same way as for Android, the spread of autonomous driving will likely accelerate if applications that run on autonomous driving systems can be vigorously developed.

"One social challenge that can be addressed by systems employing autonomous driving is the maximization of transportation efficiency. Taxis traveling without passengers and package redeliveries due to the recipient's absence are very inefficient. It is said that 70 to 80% of transportation costs are personnel expenses, and we believe that realization of autonomous driving can contribute not only to personnel expense reduction but also to solving social problems related to road transport, such as exhaust gases and traffic congestion. Such system development is the essence of autonomous driving, which we strive to achieve."

However, he points out that, "Before developing application technologies based on autonomous driving, another barrier must be overcome, namely safety assurance."

"Currently, each company individually sets its own safety standards, which thus vary among companies. For example, in the manner of Google, adopting a method that ensures accident-free travel except for 1 mile out of 10 million is one way. Ensuring accident-free driving by test driving on autonomous driving courses and assuming all possible environments is another. Assuming that autonomous driving systems will be installed in various vehicles going forward, major problems will occur unless strict industry safety standards are established, though this is the most difficult, time consuming task."

We daily endeavor to develop autonomous driving systems to achieve our high goals while tackling technical barriers. In the future, cooperation among businesses will be a key driver for realizing a society with autonomous driving systems.

Shinpei Kato
Associate Professor, Graduate School of Information Science and Technology, the University of Tokyo
Visiting Associate Professor, Institutes of Innovation for Future Society, Nagoya University
Director and CTO, TierIV, Inc.
Graduated from the Faculty of Science and Technology, Keio University in 2002. Completed his doctoral course at the School of Science for Open and Environmental Systems, Graduate School of Science and Technology, Keio University in 2008. Ph.D. in Engineering. Visiting Research Fellow at Carnegie Mellon University from 2009 to 2011 and the University of California from 2011 to 2012, and Associate Professor at the Graduate School of Information Science, Nagoya University from 2012 to 2016.
He is currently engaged in research on operating systems, parallel distributed systems, and cyber physical systems as an Associate Professor at the Graduate School of Information Science and Technology, the University of Tokyo, and concurrently serves as a Visiting Associate Professor at the Institutes of Innovation for Future Society, Nagoya University.