6-DoF EKF SLAM with global planar features in artificial environments

Ryota Ozaki, Yoji Kuroda

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper presents online SLAM for self-localization of mobile robots in artificial environments. This proposed method exploits global planar features as landmarks in extended Kalman filter (EKF). The main purpose of using global planer features is reducing accumulative error of the estimation. Planes are extracted from point cloud of 3-D LiDAR as normals. These normals are projected onto 'depth-Gaussian sphere' Those points from each plane are concentrated in one place on the sphere since planes has many normals which are almost same directions. These concentrations of points are deemed as planar features. The state vector of EKF has states of both a robot and landmarks. Prediction steps compute integration of angular velocity from gyroscope and linear velocity from wheel encoders, respectively. Observation steps update the states by associating observed planes and landmark planes. Area of landmark planes are also expanded with every association. Observed planes which are not associated with any landmarks become new landmarks, and the state vector is augmented. Similar landmarks are merged as necessary. To avoid mismatching between observations and landmarks, candidates for the association are selected by some conditions. To evaluate the proposed method, an experiment with an actual robot was performed. It shows the method suppresses accumulative error of self-localization by using the landmarks.

Original languageEnglish
Title of host publicationProceedings of the 2020 IEEE/SICE International Symposium on System Integration, SII 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages531-535
Number of pages5
ISBN (Electronic)9781728166674
DOIs
Publication statusPublished - Jan 2020
Event2020 IEEE/SICE International Symposium on System Integration, SII 2020 - Honolulu, United States
Duration: 12 Jan 202015 Jan 2020

Publication series

NameProceedings of the 2020 IEEE/SICE International Symposium on System Integration, SII 2020

Conference

Conference2020 IEEE/SICE International Symposium on System Integration, SII 2020
CountryUnited States
CityHonolulu
Period12/01/2015/01/20

    Fingerprint

Cite this

Ozaki, R., & Kuroda, Y. (2020). 6-DoF EKF SLAM with global planar features in artificial environments. In Proceedings of the 2020 IEEE/SICE International Symposium on System Integration, SII 2020 (pp. 531-535). [9026222] (Proceedings of the 2020 IEEE/SICE International Symposium on System Integration, SII 2020). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SII46433.2020.9026222