On-Line Identification of Line Segments Using Hypothesis Testing for Robot Mapping

Abstract

A novel features identification method is proposed for mobile robot indoor environment perception. In this paper, firstly, multi-dimensional data space is building based on laser scanner data, in which corner point is located by hypothesis testing theory. Secondly, line segments are obtained and localized by fitting method, which are previously estimated using probability model. At last, experiments verify the efficiency of proposed algorithm in indoor environment.