METI has launched a national 5-year-project called Humanoid Robotics Project (HRP) since 1998 FY. Because humanoid robots have a human-like figure and can walk in a biped way and can take an action like a human being, we consider applying the humanoid robot platform in HRP to service fields of caring people such as the elderly, patients, etc. Taking the technical limitation of the current humanoid robot into account, we assume a situation where the robot serves such people. We regard the robot as users' avatar or another existence of users, and consider the class of users: a nurse, a patient, and a person in a remote site. This allows us to propose four types of user interfaces for the humanoid service robot system: an on/off-line user interface for a nurse who wants to control the robot as his avatar, a user interface for a patient who wants to control the robot as another existence, and a user interface for a person in a remote site who wishes to control the robot as his avatar. These user interfaces are described in this paper.

We are developing a complete digital Holter monitoring system (HMS). One important problem required with the digital HMS is the ability of huge data storage and processing. A conventional HMS uses an analog magnetic tape to record long-time data. Although recoding the huge data into a magnetic tape is the easiest way, it is not suitable for realizing a simple compact digital HMS. Our main objective is to realize a simple digital HMS. Therefore we have to recode the huge electrocardiogram (ECG) data into a semiconductor memory (IC memory card) instead of the magnetic tape. In order to record the long-time ECG signal into an IC memory card, it has to be compressed with a high ratio over 1/50. A noisy ECG signal with very low fluctuations cannot be compressed sufficiently. A noise cancelling method which is effective for the data compression, is proposed.>

This paper describes 3D vision techniques for an advanced robot system in terms of efficient hardware and practical software. Arm-controlled stereo cameras have been developed for efficient 3D sensing, whose interocular distance, yaw and tilt are under program control. Stereo vision programs are implemented with the aid of the stereo cameras and a high-speed vision processor. Knowledge-guided vision algorithms are realized in a high-level processor, where a priori knowledge regarding the environments, objects and procedure controls are utilized.