WEB MAGAZINE
menu
logo_UTokyo
logo_UTokyo

TAGS

Frontiers of Science

A New Hero’s Journey, Showcasing Operating Systems for Autonomous Driving

Shinpei Kato

Associate Professor, Department of Information Science, Graduate School of Information Science and Technology

October 13, 2021

research01

A Young Man Inspired by Two Heroes

In the 1990s, two heroes emerged in the world of operating systems (OSs) for personal computers. 

One was Bill Gates, who, needless to say, was the creator of Windows. The OS that he pioneered spread rapidly from its first iteration, Windows 95, and since then the Windows OS has become the de facto standard across the globe. It’s an OS that everyone has encountered at least once in their lifetime. It is the key that unlocks digital activities for many people, and it was Gates who created it. 

The other hero was Linus Torvalds. As a university student in Finland, Torvalds developed a new OS he named Linux and released its source code publicly on the Web, so that anyone with the requisite skills could make improvements. This collaborative, sharing approach is called open source. Skilled programmers from all over the world then began to modify and enhance the Linux OS. Over time, Linux’s performance grew to rival that of  Windows and MacOS, which were created by teams of professionals. The development of Linux through the open source approach was one of the revolutions we have to thank Torvalds for. 

Inspired by the success stories of these two heroes, a young Japanese student dived headlong into researching OSs at university. He developed Autoware, an operating system for autonomous driving, releasing it as Open-Source Software like his hero Linus Torvalds, and at the same time, began a commercialization drive that rivaled Bill Gates’ feats. Inspired by these OSs as a youth, what kind of future does the Associate Professor Shinpei Kato envision as an adult? 

An OS for Autonomous Driving That Can Be Used Out of the Box

“I entered university in 2000, so I’m a member of the personal computer and Internet generation. Since my twenties, I’ve had a keen interest in programming and operating systems. The feats of Bill Gates and Linus Torvalds were success stories in themselves, and I was blown away by the achievements of the programmers that came before me. I was outgoing, enjoyed social interaction, and had a lifestyle revolving around popular culture, so I was dead-set on pursuing research on a nerdy topic. Working in computer science, I sunk my efforts into research on OSs, rather than on applications. I didn’t tell anyone at the time, but I was aiming to be a good-looking geek (laughs).” 

Lighthearted talk from the start. A tale from his younger days that is typical of the energetic Associate Professor Kato. 

“In Bill Gates’ day, OSs were all about personal computers, but in the last 10 years, it’s also been smartphones. When I was a student, I thought that robots would be next in line after computers and cell phones, so I focused on OSs for robots. However, I began to realize that autonomous driving would become a reality sooner than the arrival of humanoid robots, so I switched my focus to autonomous driving. That was about 10 years ago.” 

In line with this trend, the profile of artificial intelligence (AI)  has risen in the community. In the future, AI will likely find implementation in both robots and autonomous driving. However, for Kato, AI is not something to be considered in isolation, but rather thought of as an extension of the OSs of computers and cell phones (with AI functions) for robots and autonomous driving. Personal computer and cell phone OSs are just tools for people to use, but robots and Aautonomous Ddriving cars are different in that they ‘work autonomously’.  It’s the ‘autonomous’ component that everyone thinks of as being AI.

Kato continued his research, completing and releasing an autonomous driving OS, dubbed Autoware, in Open-Source Software mode in 2015. He then established a venture company, Tier IV, in order to take Autoware to the global market and develop his autonomous driving business.

Just as there are multiple OSs including Windows and MacOS for personal computers, and iOS and Android for smartphones, there are several OSs currently used for autonomous driving. Compared to other autonomous driving OSs, Autoware is unique in that it can be used immediately after installation. What are the implications of this? 

“It means you have to minimize  the unique components and eliminate special features as far as possible. The biggest appeal of an Open-Source Software OS is that anyone can tweak and customize it to their own specifications. To that end, we need to make the system as plain and simple as possible, so that it works immediately after installation.” 

Aha! So that’s the idea behind Open-Source Software . Now I’ve seen the light! 

『Autoware』an open source autonomous driving OS.

Redefining the OS  

Why has Associate Professor Kato been able to devote himself to OS research for so long? Where in his research does he feel the academic interest lies?

 “There’s something like a law in the world of computing that if you improve one thing,  something else always becomes worse: it’s called a trade-off. For example, if you make a computer faster, it uses a lot of power; if you make it smaller, it becomes less powerful; if you make it smaller but make it more powerful, it becomes expensive, and so on. Because you can’t make improvements in all areas, the key to computer science is to find the best trade-offs. That’s my fundamental interest in doing research. And as for my interest in research into OSs for autonomous driving, I’m determined to change the concept of an OS or rather, redefine what we mean by an OS.”

Strictly speaking, the OS of a personal computer is required to control only a closed system inside the computer itself. Similarly, smartphone OSs are also only designed to control the phone. When they’re connected to the outside world via the Internet, they are controlled by applications that are outside the reach of the OS. However, the OSs of robots and autonomous driving act independently of  the external environment. In the case of a personal computer, only the disk, memory, and processor need to be controlled, but when it comes to autonomous driving, it’s necessary to respond to the external environment—that includes humans, objects, and natural phenomena—when the vehicle is in motion or at rest.

For example, if a person runs out onto the road in front of an autonomous vehicle, the OS must avoid a collision by prioritizing human life above all else. The OS will autonomously determine the correct course of action and take it. In short, the OS must incorporate the ethics inherent in the traffic setting into its rules of action. 

“Such an OS that can make decisions in response to factors in the external environment is fundamentally different from the conventional OSs in personal computers and smartphones. We need to change the concept of the OS from the ground up. That’s why I want to redefine the conventional image of the OS with a totally new concept.” 

A bold endeavor, indeed! Since ancient times, hasn’t it been the hero’s task to achieve such a change in paradigm? 

Logical Thinking: The Key to Combining Research and Management. 

Associate Professor Kato is currently both a researcher and the manager of a venture company. In  general terms, a researcher is someone who pursues his or her research interests to the limit, while a manager is someone who seeks profits through commercial activities and grows their business. I feel that it requires a very different mindset to balance these two occupations.

“That’s the same thing for me. I think there are two types of researchers. One is not interested in money and just wants to pursue the truth. The other is one who seeks substantial funds to conduct large-scale research that is approaching social implementation and leads to innovation. I think I’m one of the latter type, and I believe that they and managers have the same mindset. The latter type of researcher has to manage an organization from the moment they form a research team. You have to explain the importance of the research and obtain sufficient funds to run the organization, just as a manager pitches their business to investors to raise funds.” 

The bigger the project, the more research comes to resemble management. 

“When you examine their skills, tasks, and outputs separately, you find that both managers and researchers have the same mindset and skills. However, the tasks they perform are quite different, and so the outputs are also different. One key skill is logical thinking. If you can think logically, you can combine both research and running a company. Another important skill is to be able to imagine what success looks like. If you can visualize success, then you can set that as your goal and think logically about how to get there. That’s why I do the same as a researcher and as a manager of a venture company.” 

Creating a Massive Ecosystem Around the Venture Company Tier IV

Tell us about your venture company, Tier IV. What kind of company is it, and what is its appeal? 

“In a nutshell, Tier IV is ‘a technology group that seeks to create and grow an ecosystem with an Open-Source Software approach.’ We assume that our competitors are GAFA [Editor’s note: a collective term for the four IT giants in the U.S., coined by combining the initials of Google, Amazon, Facebook, and Apple], Chinese IT majors, and other huge companies. Tier IV is a venture company, so it’s very small. Logical thinking leads us to the conclusion that we’ll never be able to beat those giants. So our weapon of choice is the open source approach.” 

When you make your software available as Open-Source Software, you attract a lot of like-minded people who want to use it too. Countless companies with similar philosophies have now gathered around Tier IV. In the beginning, many were small startups, but now they include 1 trillion yen and 10 trillion yen companies such as semiconductor majors and smartphone manufacturers. Considering the size of the economic zone surrounding Tier IV, the scale of the ecosystem has already grown to surpass that of Google. Kato sees the ecosystem as having the might of a global coalition of forces, attempting to create new value by destroying existing values such as ‘monopolies like GAFA’ and ‘computers as tools for human use.’ 

An epoch-making platform research on domain-specific computer

Heroes Create Stories for the Common People. 

When I listen to Associate Professor Kato speak, I don’t sense much motivation to build wealth, the sine qua non of the corporate sector. He seems to be trying to take Autoware to the world through an ecosystem centered on Tier IV, creating social common capital, or in other words, a massive infrastructure. Is this an approach supported by a spirit of public-mindedness? 

“It’s true that my motivation is not my own self-interest. Do you know what a ‘Hero’s Journey’ is?  It’s a form of storytelling articulated by the mythologist Joseph Campbell: the main character or hero goes on a journey in which they overcome various difficulties, grow, finally achieve victory over a major opponent, and return home as a hero. I have a habit of likening my life to this Hero’s Journey. I tend to think of myself as the main character, and as the main character of the story I’m a hero, so it wouldn’t be cool if I acted in my own self-interest (laughs). My story is about creating an ecosystem, a global coalition of forces using open source as its weapon and defeating the existing giants, and that’s for the benefit of the people, right? Anyway, I want to finish this game, or rather, I want to complete the story. Perhaps you could say that I have a strong desire for self-actualization.” 

The young man who admired two heroes of OS development when he was in his twenties is now an adult, trying to write his own hero’s journey. 

We’ve heard a tale spun by a unique Associate Professor in the University of Tokyo’s School of Science. “What happens next?,” you ask. I don’t know, but it’s sure to give you an adrenaline rush. 

laboratory: https://www.pf.is.s.u-tokyo.ac.jp
personal: https://shinpeikato.com

Interview/Text: Osamu Shimizu [ ACADEMIC GROOVE ]
​Photography: Junichi Kaizuka
video: provided Associate Professor Shinpei Kato

Shinpei Kato
Associate Professor,Department of Information Science, Graduate School of Information Science and Technology
Graduated from the Graduate School of Information Science and Technology, Keio University in 2004, and completed his doctorate in Open and Environmental Sciences at the same university in 2008. D. (Engineering). D. in Engineering from Keio University in 2008, and was a researcher at Carnegie Mellon University and the University of California.
TAGS

image01

Visualizing space and planetary research at the School of Science

October 11, 2021

image01

Clues for solving the mystery of evolution

September 21, 2021

image01

Searching for the essence of life through physics

September 21, 2021