Visitor: Takashi Suehiro (TRC Active Intelligence Laboratory) Period: August 3 - 14, 1995 (12 days) Places: (Attended conferences): Intelligent Robots and Systems (IROS) '95, University of Pennsylvania, CMU, MIT
Major purpose of my business trip for this time was to participate in the IROS '95 held in Pittsburgh. Using this opportunity, I visited the Grasp Laboratory at the University of Pennsylvania before the conference, the Robotics Institute at the CMU during the conference period, and robot-related laboratories at the MIT in Boston after the conference to survey trends in the robot research there.
The IROS '95 consisted of four parallel sessions with 268 presented papers totally. The area-wise breakdown of the papers presented is 125 from the U.S.A. , 100 from Asia (Japan approximately 80) and 43 from Europe. The research level of Japan in this field is placed high and, therefore, I saw many Japanese researchers participated in such an international conference as this held even in overseas countries.
The subtitle for the conference was "Human Robot Interaction and Cooperative Robots" and I did confirm that the direction of our lab to pursue the robots that coexist with human beings was the worldwide stream. However, the session for the Human Robot Interaction and Cooperative Robots was only one and, in this session, the paper matched to the theme in its content was only one as well. My paper was classified in the session for the distribution system, as its content was implementation of the manipulator's servo through multi-agents yet. Though the coexistence with human beings is the major stream, I was realized once again that making an actual robot system requires enormous efforts.
A general chair (one of the two) for the conference was Professor Ikeuchi of the CMU. He used to be a researcher of the ETL and had been a host while I had been studying at the CMU for a year. One night of the conference period, I was invited by Professor Ikeuchi and promoted friendship with people of various nationalities enjoying his homemade wine (see a Note) at a large room of the hotel where the conference was held.
The Grasp Lab at the University of Pennsylvania where I visited ahead of the conference is the group led by Professor Bajcsy. The name of the Lab gives impression that they were focused in the research on hands, but actually they were not engaged in hand research, instead pursuing the mobile robots, the multi-robot cooperation, the visual sensation, the dynamic visual sensation using optical flow, the macro-micro cooperation for the observation control, the manipulation skills and other general robot affiliated issues extensively. As Professor Bajcsy has once visited the RWC Tsukuba Research Center and also has close contacts with the robot group members of the ETL, she invited us to her home and showed the bathroom that was equipped with an automatically washing system called "Washlet" Imported from Japan he proud of.
The director of the Robotics Institute at the CMU, where I visited during the conference period, was a Japanese professor named Kanade. Professor Ikeuchi also held his post in this Institute besides the Department of Computer Science. There were many Japanese visiting researchers in the Robotics Institute. I assume the Institute must be the world largest in its scale as the robot research group. The range of their research covers the entire robot-related issues. More specifically, they develop a moon surface investigation robot, a space station robot, an auto cruising vehicle, a dedicated device for the robot vision, and the chips, thus tackling a variety of things indeed. As to the auto cruising vehicle, they were carrying such advertisements as "Crossing the American continent by the automatic cruise more than 9?% (don't remember exact figures) " tied up with car manufacturers. I thought it was not a tall order because the American highways allow driving only with the auto cruise system and simple steering.
At the MIT where I visited after the conference, I called on three laboratories, the AI Lab, robot related labs of the Machine Engineering Department, and the Media Lab. I felt the impression that the AI Lab and robot related labs were not large labs in their scale, but were making steady academic researches. What particularly impressive were the cheap vision machine and the ant robots. The cheap vision machine is an inexpensive standard vision system developed at the AI Lab and commonly used in many laboratories. An ant robot measured approximately three centimeters in diameter and I was surprised hearing the only one student was developing it.
Professor Brook's lab, well known for a subsumption architecture, was developing not only the mobile robot, but also the humanoid, but what finished setting up was just the head and the trunk. I thought that we could cope with them well enough if we decide to develop a jumanoid system.
I have been told that in principle the Media Lab does not allow visitors to come in even if the people of the MIT, but a student who has once came to the ETL made arrangements for us and we could have had a chance to visit the Media Lab through students' volunteers. They were rather concentrated on multimedia research than the robot research. I felt the research was using enormous computer powers and money, but was interesting with such features as multimedia E-mail.
This is my brief report on the business trip to the U.S.A. Last but not least, I sincerely express my gratitude to the people at the TRC research planning section who took care of the preparations for the trip, the people at the ETL who made arrangements of worthwhile visits for me, and the people who accepted my visits.
Note: In Pennsylvania, wine brewing for private consumption is legally accepted up to 750 liters in a year per an adult.