期刊名称：International Journal of Grid and Distributed Computing
印刷版ISSN：2005-4262
出版年度：2016
卷号：9
期号：12
页码：143-158
出版社：SERSC
摘要：Compute Unified Device Architecture (CUDA) is a mature parallel computing architecture, which can significantly accelerate performance of the computation intensive algorithm. In this paper, FastSLAM algorithm based on the probability model is further studied and the resampling algorithm for the path estimation is improved. In the resampling phase, resampling rules are redesigned and the previous data limitations are broken for the purpose of parallelization. We propose the FastSLAM algorithm based on CUDA, which accelerates robot localization and mapping. The experiment results show that FastSLAM_CUDA can achieve a significant speedup over the FastSLAM with many particles.
