The Key to Robot Control Lies in the Adept Movement of Living Things
In this Tomorrow's Pioneers, President Toshio HIRANO visited Professor Koichi OSUKA and Assistant Professor Kenjiro TADAKUMA at the Graduate School of Engineering and talked with them about the fun of robot research and its possibilities in the future. The two researchers took part in the development project of a small robot Minerva II 2 mounted on the asteroid probe Hayabusa 2 launched on December 3, 2014.
Development of a Small Robot that Can Operate on an Asteroid
HIRANO As shown in the examples of powered exoskeletons compensating for the lack of manpower in the field of nursing care, an artificial intelligence robot taking entrance examinations of The University of Tokyo, and talking with an android just like a human, public interest in robots is on the rise and they have become closer to us than ever. Before asking you about your research, first, Professor Osuka, please tell me how you were involved in the development project of Minerva II 2, a small robot for checking the movement on the surface of an asteroid and what role the Osaka University team played in that project.
OSUKA Three small robots are mounted on Hayabusa 2: two units of Minerva II1, the revised type of Minerva mounted on Hayabusa (launched in May 2003) and the unit of Minerva II 2, the unit with 4 types of driving mechanisms developed by Osaka University, Tokyo Denki University, Tohoku University, and Yamagata University, respectively. Hayabusa 2 is scheduled to arrive on asteroid 1999JU3, with a 920m diameter, in the summer of 2018. This small robot has a mission to send information while moving around on the surface of the planet. However, the situation of the surface of the asteroid and its gravity are unknown and the machine's operating environment will not be known until the robot gets to the asteroid. The development of a robot in such a setting was a challenge.
HIRANO Is the probe powered by a solar panel alone?
OSUKA A battery is used as well, but since electricity is used for operating observation equipment, the drive system was not allowed to consume a lot of electricity. There was also a limit in size: the body of the small robot was 15cm in diameter, 16cm in height, and about 1kg in total weight. The weight limit for the drive system was just 30g. In the beginning, our team thought of using a motor for operating the robot, but the motor consumed a lot of power and was bulky. Then, we came up with the idea of using a mechanism of a spring powered clock. Finally, we got rid of a mainspring and contrived ways to jump the machine with spring power by making use of the phenomenon called "snap-through buckling." In the process of pursuing reliable operation, we reached a simple mechanism for jumping with low power. We mounted two springs on the machine, but once they operated, their mission was complete. It took four years for the machine to arrive on the planet, but the machine's operation finished in a moment.
TADAKUMA In this case, it was important that the machine didn't work at all during the four years from its launch to arrival. We focused on giving assurance that no troubles occurred when it operated for the first in four years.
HIRANO You made a spring mechanism? Wow. The mechanism is more low-tech than I expected. (laughs) The newspaper article that I read the other day said old versions of IBM computers were used for Marsprobe in the U.S. in order to assure reliability. That applied to Hayabusa, didn't it?
Stunned by a Passive Dynamic Walking Machine Operated by Gravity
HIRANO I've heard that Professor Osuka conducts dynamic control research in order to pursue the essence of dynamics and control behind movement by using bipedal walking robots, quadrupedal walking robots, and snake-type robots. In recent years, you are conducting research targeting living things, aren't you?
OSUKA The field of control engineering has developed while pursuing how to manipulate artificial things. While researching how to control machine systems and robots, I got interested in how living things can move sophisticatedly in various environments. Speaking of control, there should be some kind of controller to control the target; however, living things are not made of controlling parts and parts to be controlled. They exist in their harmonious combination. In researching, we reached the questions of "Are nervous systems needed for walking?" and "Where is the source of intelligent behavior?" Then we became to understood that a body and a situation will work only if they are integrated.
HIRANO How do you verify that in experiment?
OSUKA A passive dynamic walking machine accounts for our theory. This machine does not have a control drive mechanism, but when placed on a slope, it starts to walk due to gravity. Its way of walking is much closer to humans' than that of bipedal walking robots operated by motor. The machine is walked by dynamic action on a place where the body touches the ground surface, which was expressed in an equation. I was stunned to see that. It was in 1990. Then we made bipedal walking machines, quadrupedal walking machines, and machines walking on 20 "feet". The body of the machine walking on 20 "feet" curves and moves like a snake. Contrary to "explicit control" that controls an object by applying power through computer and other means, controlling the object according to the environment is called "implicit control." I think the source of intelligence lies in this "implicit control."
HIRANO That means, if explicit control factors are added to an implicit control robot made by Professor Osuka instead of gravity, it becomes possible to reproduce the movement close to that of living things freely regardless of the direction of gravity?
OSUKA That's right. Intelligent movement will be available.
Developing Research While Focusing on Mechanisms
HIRANO I heard that Professor Tadakuma developed many mechanical robots such as omnidirectional mobile robots and drive mechanisms. In recent years, you've also been involved in medical engineering collaboration research to develop a device which handles cell sheets.
TADAKUMA I use the Chinese characters 機巧 [nifty device] for the word usually written as 機構 [mechanism]. In the past, the Chinese characters 機巧 were used for the word "gimmick." We pay attention to ingenious mechanisms in the times when there were no computers and try to develop our research into inventions. As a drive mechanism, there are wheels and crawlers. Conventional omnidirectional wheels and crawlers have this in common: the diameter is small for the mechanism as a whole, which makes it impossible for them to overcome steps. We tried making the whole mechanism a sphere, which overcame rough terrains with a step or gap, allowing it to move any arbitrary direction.
HIRANO You made the best use of your inventiveness, I see.
TADAKUMA Not only that, we cut a sphere in half and made a crawler with two-piece ball-shaped wheels. The mechanism of a sphere and that of a crawler look different, but the concept is the same. In order to use this mechanism for planet exploration, we conducted joint research with the Japan Aerospace Exploration Agency (JAXA).
Robots Were My Childhood Dream
HIRANO Both of you are really enjoying research, aren't you? You were interested in robots as children?
TADAKUMA I liked making goods from my childhood. When I was an elementary school student, I read manga The Secret of Robots , which led to my delving into robots. In that manga, there was a scene where people discussed why robots existed. They reached a conclusion that robots existed because they should help people. The manga ended in a way to encourage readers to think of robots were helpful in what ways specifically. Later, when I was a 1st year student at junior high school, I was excited to see MORI Mamoru flying to space on the space shuttle Endeavour. Around that time, I saw a Mars rover. I was sure that robot did help people. That determined my career.
OSUKA I liked animations such as Astro Boy , Tetsujin Nijuhachi-go , and Mazinger Z . I think these robot animations affected me in choosing my career. When I was a senior high school student, I had a classmate with a prosthetic hand. So one time I thought of going into research regarding prosthetic hands and legs.
After completing the master course, I worked at a research center of Toshiba Corporation. One day, my boss instructed me to make a demonstration of a robot beating a small drum. At first I thought it would be easy but found it difficult to let the robot make a sound. I struggled to think and, before the dawn of the demonstration day, a flash of sudden inspiration came to me like a vivid flash of lightning.
Until then, I tried to control the path of arm movement in a planned manner, but when I stopped controlling it and took the arms upward and let them down, the drumsticks bounced, making a sound. This experience made me think that the essence of movement does not lie in controlling, but in the body itself. We can move the object in response to a variety of situations not only by working mechanism, but by working both mechanism and dynamics. In that sense, Professor Tadakuma and I are alike.
HIRANO Some high school students say they don't know how to choose their course but both of you decided your direction in life when you were high school students, didn't you?
Flexible way of thinking is a driving force for promoting research
HIRANO At the end of this interview, I'd love to hear your dreams in the future.
OSUKA Previously, I was not interested in living things and didn't like them, but recently, I feel sick when I see even centipedes and cockroaches. (laughs) I became to wonder if we could research artificial objects unless we understand how sophisticated movement of living things are produced. I want not to physically understand living things but to be emotionally convinced. I'd like to continue to pursue that in terms of control.
TADAKUMA When I was a junior high school student, I wanted to land on Mars with a robot I made, but after I became a researcher, I felt like being lost because the existence of Mars is already known. One saying goes "I climb the mountain because it's there." However, in this case, the mountain is already discovered. In that sense, I want to discover the mountain itself that is not elucidated yet. And my ultimate goal is to work out a new theoretical system from which anyone can come up with an answer from the view point of mechanism.
Professor Osuka's laboratory has many toys and plastic models. I felt as if I were in my grandkid's room. (laughs) I imagined a more mechanic and cold world, but I see that you have flexible ideas and are conducting human research. I hope you will work hard with these ideas.
Thank you for your time today.
Concentration and Persistence Provides Inspiration -- Comments from President Hirano after the Talk
I felt that these two have a bug for research and have continued to hold the same innocent feelings as they did when they were children. I had a renewed sense of the importance of concentration as well. Some researchers and students come up with good ideas and achieve results, but others don't, even if they conduct experiments in the same way and have the same amount of data. When thinking of the difference between them, I had a renewed sense of how seriously one concentrated on thinking produces this difference.
Dr. Osaka completed the master course at the Osaka University Graduate School of Engineering Science in 1984. He received his doctorate degree in engineering in 1989. In 1984, he joined Toshiba Corporation and in 1986 he became an assistant at Osaka Prefecture University. He served as a lecturer and assistant professor at Osaka Prefecture University, an assistant professor at Kyoto University's Graduate School of Informatics, and a professor at the Kobe University School of Engineering. He took his current position as a professor at Osaka University's Graduate School of Engineering in 2009. His specialty is control engineering. His main research themes are advanced control theory and application; non-linear dynamics, chaos, and control; robotics, mechanical analysis, and synthesis.
Dr. Tadakuma completed the master course at Tokyo Institute of Technology's Graduate School of Engineering in 2004 and the doctoral course at the same school in 2007. He served as a researcher at Massachusetts Institute of Technology and Tohoku University and an assistant professor at The University of Electro-Communications. He took his current position as an assistant professor at the Graduate School of Engineering in 2009. He is involved in medical engineering collaboration research to develop device handling cell sheet. He won the Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology in 2011.
A graduate of Osaka University's Faculty of Medicine in 1972, President HIRANO studied at NIH (U.S.) from 1973 through 1976. He became an assistant professor at Kumamoto University in 1980. He became an assistant professor in 1984, a professor at Osaka University in 1989. Following that, he became a director at the Graduate School of Frontier Biosciences in 2004, a director at the Graduate School of Medicine, and a dean at the Faculty of Medicine in 2008. He assumed the the 17th presidency of Osaka University in August 2011. He served as Chairman, Japan Society for Immunology (2005~06). He is also a member Council for Science, Technology and Innovation, The Science Council of Japan. He has a doctoral degree of medicine. His awards include Sandoz Prize for Immunology, Osaka Science Prize, Academic Award of the Mochida Memorial Foundation, Medical Award, Fujiwara Prize, Crafoord Prize, Japan Prize, and Medal with Purple Ribbon by the Emperor of Japan.