No abstract

This paper presents a system that detects unexpected moving obstacles that appear in the path of a navigating robot, and estimates the relative motion of the object with respect to the robot. The system is designed for a robot navigating in a structured environment with a single wide-angle camera. The objective of the system is to detect moving obstacles in order to gradually stop the robot to avoid collision; maneuvering around the obstacle is not considered here. The system has been assembled using pieces of existing vision techniques with a strong emphasis on real-world applications and very fast processing with conventional hardware. The system uses polar mapping to simplify the segmentation of the moving object from the background. The polar mapping is performed with the focus of expansion (FOE) as the center. A vision-based algorithm that uses the vanishing points of segments extracted from a scene in a few threedimensional (3-D) orientations provides an accurate estimate of the robot orientation. This is used to maintain the motion of the robot along a purely translational path and also used to subtract the effects of any drifts from this path from each image acquired by robot. By doing so, the determination of the FOE is simplified. In the transformed space qualitative estimate of moving obstacles is obtained by detecting the vertical motion of edges extracted in a few specified directions. Relative motion information about the obstacle is then obtained by computing the time to impact between the obstacles and robot from the radial component of the optical flow. The system was implemented and tested on an indoor mobile robot at our laboratory. Results from the robot navigating in real environments are presented and analyzed here. The system is able to detect moving obstacles at 100 ms/frame and can be used as a cueing mechanism for moving obstacles.

No abstract

No abstract