School of Engineering, The University of Tokyo
(SE-UT) Mar. 8, 2006
[Excerption from T-time Vol.10 / 2005.12]
Topic 1: Researches on Humanoid Robots
We interviewed Professor Masayuki Inaba of the Graduate School of Information Science and Technology, the University of Tokyo. Professor Inaba is a leading expert of humanoid research, in particular, leaning abilities of humanoid. Recently, he made a presentation of a humanoid called “Kotaro” at Expo 2005 Aichi Japan.
 |
 |
Professor Inaba with
HRP-2W(left) and HRP-2JSK(right) |
HRP-2JSK using a
vacuum cleaner
|
Q: First of all, please tell us about your research.
A: We have been studying production of humanoid robots with action-perception cycle so that they can respond appropriately to various situations. If we can produce robots having abilities to judge when observing motions of people and to learn the differences from theirs, reprogramming will no longer necessary. Current robots are able to duplicate various actions, i.e. they merely execute a program. Therefore, our major theme at present is to make robots have the ability to imitate actions and make substantial evolution. Recently, we have been working on a project to make a robot handle a broom.
Q: How did you start studying robots?
A: I liked robots and used to make them when I was a child. In the sixth grade of elementary school, my friend and I tried to make a bipedal robot with task assignment. At that time, we could not obtain adequate motors so that we only make an upright robot. Those experiences of my childhood led me to enroll in the department of mechanical engineering, where I thought I may have opportunities to study robots. While in the department, however, I did not get involved in robot study but worked on other themes for my graduation study.
Q: What did you study in graduate school?
A: Just after I entered graduate school I started working on robot study because I was allowed to use a computer freely for my study. It was a real fan and I devoted myself in the study day and night, eating, breathing, and sleeping in the laboratory.
In the master’s course, I worked on a study to produce a robot with the ability to watch and grab an object. Because no experimental robot was available in those days, I made one myself by using cameras and machines functioning as arms and sending obtained information to a computer, with a help of my computer-literate colleagues.
I tried a method enabling the robot to have three-dimensional perspectives through one equipped camera and come up with a new idea of utilizing components of a single-lens reflex camera. After much trial and error, I finally succeeded in measuring with error margin of 1cm in 1m range. For my master’s thesis, I chose to produce a robot having an ability to put a string through a ring and tie it. It was the first study of robots being able to handle soft materials. The result was presented at international conferences, including Robotics Society of Japan, which gave me the first Best Paper Award.
Q: What is the objective in your future research?
A: I aim to produce humanoid robots with the ability to handle tools, for example, robots which learn how to sweep with a broom by observing people sweep and extracting the necessary “information” to sweep. I have been conducting an experiment to develop the mechanism enabling the robots to try and figure out a more suitable way to achieve their objectives, by comparing program-controlled behaviors with human activities. Recently, we have succeeded to develop a robot which can handle a vacuum cleaner and other home electric appliances.
Things are different now from my student days, and experimental robots for program testing are available. Graduate students are given the opportunity to actually make a robot work with the program they write, even in their first year of the graduate school. For the past few years, robots’ abilities have dramatically improved. Utilizing the robots of our laboratory, undergraduate and graduate students, and researchers have been addressing the increasing challenges for us to overcome.
Q: Could you give a message to students pursuing engineering career?
A: Please cherish what you want to do. You will certainly have an opportunity to realize your dream. Once you get such an opportunity, work hard. Keep your curiosity and imagination to achieve your goal.
The most important thing in study is to create a ripple effect. In other words, the variety and quality of further achievements measure success and failure of your study. While studying, you will recognize your ability. Once you know it, your objective becomes more specific. It is the time for you to choose and concentrate on what you can do. With that experience, you will be able to work more actively afterward. In that sense, I am glad that I have been studying robots.
Topic 2: Introduction of Akihabara Satellite of the Graduate School of Information Science and Technology
Creative Informatics Department, Graduate School of Information Science and Technology, was established in April 2005. The main office of the department is located in a modern high-rise building in Akihabara. We asked Professor Inaba about the Akihabara Satellite.
Q: Could you explain the background of establishing the new department, Creative Informatics Department, in Akihabara?
A: We already had five departments in the Graduate School of Information Science and Technology, namely, Computer Science, Mathematical Informatics, Information Physics and Computing, Information and Communication Engineering and Mecano-Informatics. In order to create inter-departmental linkages among these for more practical research, the Creative Informatics Department was established.
Q: What are the merits of studying robots at Creative Informatics Department?
A: The most significant merit for robot study is practical workability, because we can experiment as soon as we hit upon an idea, leading to rapid progress of the research. The availability of interdisciplinary and cross-departmental comments and discussion is also a great advantage for wide-range robot study.
Q: What will the role of the Akihabara satellite be?
A: Akihabara, where several railways intersect, is easily accessible from Hongo, Komaba and Kashiwa campuses. By having a robot research in Akihabara, we will be able to provide various information to many different locations. In addition, we expect the Satellite to boost exchanges of ideas among Hongo, Komaba and Kashiwa students. The Akihabara Satellite also has strong connection with society to promote active academic-industrial alliance. The Satellite will serve as a bridge between industrial firms and the university.
Q: How has the redevelopment of Akihabara, including the establishment of our Satellite, changed the town?
A: Akihabara used to be a town for geeks or “otaku”. However, with the establishment of satellites of industrial firms and universities as well as large commercial facilities, the “color” of the town has changed gradually from monotone to multicolor. The town will be developed further. Meanwhile, various exhibitions including ubiquitous and robotic exhibitions are held in Akihabara, which has been playing increasingly important role of offering information on new technologies.
Q: Do your students often use the “Denki-gai”, or electronics quarter, in Akihabara?
A: I often see our students go to electronics quarters to buy electronic parts for experiment. I sometimes have junior students go to buy various parts so that they can learn from real things instead of from textbooks.
Q: Could you give a message to the readers?
A: I believe that university is a door for the future, through which students can see the future. The Akihabara satellite is to fulfill that role. We are going to offer as much information as possible utilizing our robots.
<Back Numbers>
Contact：
Global Ware Project
School of Engineering, the University of Tokyo
e-mail:
y-obikane
＠
t-adm.t.u-tokyo.ac.jp
conference-gwp
＠
t-adm.t.u-tokyo.ac.jp
Tel: +81-3-5841-6044; +81-3-5841-7549
|
