IEEE札幌支部

講演会のご案内

人間型知的サービスロボットの開発

講師： Prof. Volker Graefe (ドイツ Bundeswehr Univ.)

After a brief overview over 25 years of Robotics research in the author's laboratory, the humanoid service robot HERMES is introduced.

HERMES has two "eyes", a microphone, tactile sensors, two arms, a bendable body and an omnidirectional undercarriage (22 degrees of freedom). He can see, hear, speak and feel as well as move about, localize himself, build maps and manipulate various objects. He can communicate with untrained humans in English, French and German. In his dialogues and other interactions with humans he appears intelligent, cooperative and friendly. Although he is not as competent as the robots we know from science fiction movies, the combination of all those characteristics makes him rather unique among today's real robots.

Moreover, they make him an excellent research tool for developing the technology necessary for realizing the personal service robots we hope to see in everybody’s home in the future.

The situation awareness, intelligence and cooperativeness of HERMES make it easy and pleasant for an operator to introduce the robot into a new environment. This not only improves the acceptance of the robot by novice users, but also reduces the likelihood of operator mistakes and, thus, contributes to the depend-ability of the robot.

Dependability is, in fact, an important issue in the context of intelligent experimental robots. Although dependability has recently begun to attract increasing attention among Robotics researchers, apparently many of them are still satisfied if their robots perform just once or twice, and refuse to worry about dependability. They argue that such matters will be dealt with by others, e.g., industry, in due cours. In contrast to this, we think that dependability is largely a consequence of fundamental design decisions and that it should be considered by anyone who builds an experimental robot. We will present a number of design principles that lead to an improved dependability of intelligent robots. Among them are:

• Learning from nature how to design reliable, robust and safe systems
• Providing natural and intuitive communication and interaction between the robot and its environment
• Designing for ease of maintenance
• Striving for a tidy appearance

These principles have guided us in the design and construction of our humanoid robot HERMES. They will be explained in greater detail, and experiences gained in a long-term test in a museum will be reported. In the special exhibition “Computer and Brain” in the Heinz Nixdorf Museums-Forum in Paderborn, the world's largest computer museum, HERMES was presented to the general public every day for 6 months. He chatted with visitors in natural language in German, English and French, answered questions and performed ser-vices as requested by the visitors.

We found it interesting to observe how HERMES, actually just a laboratory proto-type despite his designed-in dependability, mastered the daily hard work far away from his “fathers” where no easy access to repair and maintenance was available, and how he got along with strangers and even with presenters who did not know much about robot technology. In fact, he performed rather well, with only 3 failures: one motor controller, one drive motor and one audio amplifier, all commercially available and easily replaceable parts. According to the museum staff, HERMES is one of the few experimental robots that could regularly be demonstrated in action, and among them, he is considered the most intelligent and most dependable one.