We developed an embedded vision system that can accelerate the basic image processing functions for mobile robot navigation with compact hardware featuring low power consumption. The system is composed of a digital signal processor (DSP) and a dedicated LSI for low-level image processing, specifically for spatial filtering, feature extraction, and block matching operations. The image processing LSI has a dedicated systolic array processor consisting of 64 processing elements to accelerate basic block operations for image feature calculation and correlation-based image matching. The power consumption is only 10 W, about one-seventh that of a typical Pentium 4 personal computer, but the processing speed for correlation matching is roughly three times faster than such a system. With this vision system, we implemented a stereo-vision based navigation algorithm on our mobile service robot and performed a visual navigation experiment in a building hallway

This paper describes the development of applications for the building and home management service section of the "Humanoid Robotics Project (HRP)" supported by METI. The objective of this section is to develop a system that enables users to control the humanoid robots in their homes remotely via the Internet with a simple mobile terminal. To construct such a system, we focused on developing sensor-based autonomy sufficient to achieve task-level teleoperations. The system included view simulation based on CG technologies to help users perceive conditions at the robot's site and the client-server architecture necessary for remote task execution to reduce the load on the mobile terminal. Based on these technologies, we implemented a prototype system and conducted some experiments that illustrate that the system achieves the basic required functions for a building and home management service.

This paper describes the design and the implementation of a high-speed visual tracking system based on local correlation matching. The system consists of only two PC board modules, which conform to VME bus specifications. The video module is used to input and output video signals, while the tracking module is used to perform visual tracking with the local correlation method. The tracking module contains three frame memories, an image flow controller, a DMA controller for image data transfer, and a dedicated LSI to perform the correlation matching of 8-bit gray-scale images. The processing time is 110 microseconds for an 8/spl times/8 pixel template and 290 microseconds for a 16/spl times/16 pixel template. This paper also describes some of our experimental implementations, including human tracking, recognition of hand movements and signs, and traffic monitoring. Compared to conventional general-purpose image processing systems, our hardware system is compact and inexpensive. This high-performance system can be used for a wide variety of motion analysis applications.