基于 KITTI 数据集的无人车单目惯性 SLAM 算法评估

Translated title of the contribution: Evaluation of monocular inertia SLAM algorithms for unmanned vehicles based on KITTI dataset

Research output: Contribution to journalArticlepeer-review

2 Citations (Scopus)

Abstract

The evaluation of simultaneous localisation and mapping (SLAM) algorithms based on the KITTI dataset is carried out to address the localisation failure of driverless vehicles in special environments such as the absence of satellite signals. The visual SLAM algorithm VINS-Fusion is used as the evaluation object, absolute positional error (APE) and running time are used as evaluation metrics to realise the testing of localisation accuracy as well as algorithm efficiency in multi-sensor fusion mode, and the error results are analysed to provide a reference for the introduction and application of visual SLAM technology in the unmanned field. At the same time, the problems of visual SLAM technology and its application in the field of unmanned vehicles are summarised and prospected. The experimental results show that the visual SLAM technology based on the VINS-Fusion algorithm can achieve absolute positioning accuracy within 0.2~15 m.

Translated title of the contributionEvaluation of monocular inertia SLAM algorithms for unmanned vehicles based on KITTI dataset
Original languageChinese (Traditional)
Pages (from-to)50-55 and 72
JournalExperimental Technology and Management
Volume39
Issue number2
DOIs
Publication statusPublished - Feb 2022
Externally publishedYes

Fingerprint

Dive into the research topics of 'Evaluation of monocular inertia SLAM algorithms for unmanned vehicles based on KITTI dataset'. Together they form a unique fingerprint.

Cite this