ユークリッド変換に不変な特徴量を用いた二次元大域スキャンマッチング方式

Abstract

This paper presents a new scan matching method for mobile robot localization and mapping. The proposed method utilizes Euclidean invariant features in order to match an input scan with a reference scan using no initial alignment. This method is capable of global scan matching in environments having curved objects. Experimental results show that our method successfully performed map building and global scan matching in a large environment (corridor) and a complex-shape environment (room) .