Moving obstacle path detection for mobile robot

Abstract

This paper presents a system that detects the path of a moving obstacle that appears in the path of a navigating robot. The objective of the system is to avoid collision between the mobile robot and moving obstacles. The system uses polar mapping to simplify the segmentation of the moving object from the background. In the transformed space, qualitative estimation of moving obstacles is obtained by detecting the vertical motion of edges extracted in a few specified directions. The moving obstacle area is computed by component labeling and matching components of the same class over the two images. From the moving area we present the method to compute the moving-angle which shows the path of moving obstacles related to the path of the robot. Finally we present the collision-angle which shows the collision path of the moving obstacle. Results of experiment with a mobile robot under realistic conditions are given and discussed.