Target object classification based on a fusion of LIDAR range and intensity data

Abstract

This paper describes a new target identification algorithm adapted for the rules of the Tsukuba Challenge 2013. According to the rules of the Tsukuba Challenge 2013, an autonomous mobile robot has to identify a specific standing signboard while navigating a prescribed course area. With the aim of accomplishing this goal, light detection and ranging (LIDAR) is used to detect both the range and reflection intensity profiles around a mobile robot. The combination of the range and reflection intensity profiles enables the stable and robust identification of a specific standing signboard. To find the standing signboard, we focus on its shape and reflection intensity pattern. The range profile is used to detect the line-like shape, and the reflection intensity is used to detect the reflection pattern on the standing signboard. The proposed target identification algorithm is implemented to find an actual standing signboard and carried out in the actual Tsukuba Challenge 2013 environment. The validity of the proposed algorithm is confirmed through an actual mobile robot experiment.