Abstract:
This paper presents a relocalization method which provides functions of odometry cumulative error correction, vehicle stopping position difference detection and localizat...Show MoreMetadata
Abstract:
This paper presents a relocalization method which provides functions of odometry cumulative error correction, vehicle stopping position difference detection and localization algorithm system error correction for the under-vehicle inspection robot equipped with 3D solid-state LiDAR. This method is based on a novel point cloud uniaxial registration algorithm. The proposed algorithm utilizes 3D grid occupancy to construct an efficient point cloud similarity measure function and uses translation space search to avoid falling into local optimum, providing accurate transformation estimation for relocalization. According to experiments on real-world datasets in metro depot, the proposed algorithm can provide better point cloud uniaxial registration performance compared to Iterative Closest Point (ICP algorithm) and the combined use of Normal Distributions Transform (NDT algorithm) and ICP algorithm, which verify the effectiveness and robustness of this method.
Published in: 2022 China Automation Congress (CAC)
Date of Conference: 25-27 November 2022
Date Added to IEEE Xplore: 13 March 2023
ISBN Information: