A Simultaneous Localization and Mapping Algorithm in Complex Environments: SLASEM

Rongchuan Sun,Shugen Ma,Bin Li,Minghui Wang,Yuechao Wang

DOI: https://doi.org/10.1163/016918611x563373

IF: 2.057

2011-01-01

Advanced Robotics

Abstract:In this paper we present an algorithm for the application of simultaneous localization and mapping in complex environments. Instead of building a grid map or building a feature map with a small number of the obstacles' geometric parameters, the proposed algorithm builds a sampled environment map (SEM) to represent a complex environment with a set of environment samples. To overcome the lack of one-toone correspondence between environment samples and raw observations, the signed orthogonal distance function is proposed to be used as the observation model. A method considering geometric constraints is presented to remove redundant environment samples from the SEM. We also present a method to improve the SEM's topological consistency by using corner constraints. The proposed algorithm has been verified in a simulation and an indoor experiment. The results show that the algorithm can localize the robot and build a complex map effectively.

robotics