JERAJournal of Electronic Research and Application2208-35022208-3510Bio-Byword Scientific Publishing Pty. Ltd.10.26689/JERA.v10i9.15926ArticleResearch on Production Method of High‑Definition Map for Autonomous Driving Sweeping Robotshttps://artdesignp.com/journal/JERA/10/9/10.26689/JERA.v10i9.15926LiuJian20261092026-10-30Sweeping robots require centimeter‑level high‑definition maps to realize edge‑hugging cleaning and narrow‑road passage. Passenger‑vehicle‑oriented solutions suffer from point‑cloud smearing caused by unsynchronized sensor timestamps, lack of sweeping‑related semantics such as curbs and trash cans, and map deformation induced by odometer drift during long‑hour operation. This paper redesigns a full‑link solution: GPS‑Pulse plus PTP for sub‑millisecond time synchronization; improved BevFusion combined with bilateral filtering for depth completion in semantic processing; LOAM plus FPFH loop‑closure detection and graph optimization for pose estimation; the editing module supports point‑cloud import, curb delineation, cleaning‑path generation and topology network construction. A verification test covering approximately 5 km² in Changsha achieves synchronization error < 1 ms, semantic segmentation accuracy > 95 cm, positioning error < 5 cm, coverage rate 98 %, and an efficiency improvement of around 30 %.Autonomous driving, Sweeping robot, High‑definition map, Multi‑sensor fusion, Pose optimization

[1]       Ministry of Housing and Urban-Rural Development, 2021, China Urban-Rural Construction Statistical Yearbook (2020). Beijing: China Statistics Press. [2]       Duan X, Zhou Y, Tian D, et al., 2021, Review of Deep-Learning Applications in Autonomous Driving. Unmanned Systems Technology, 4(6): 27. [3]       Xiu X, Wang K, Wu J, 2025, Design and Implementation of Optimized PTP Synchronization Based on MAC. Journal of Time and Frequency, 48(2): 101–108. [4]       Du M, 2022, Research on Image Semantic Segmentation of Road Scenes Based on Deep Learning, thesis, Henan Polytechnic University. [5]       Ruder S, 2016, An Overview of Gradient Descent Optimization Algorithms. arXiv. [6]       Zhang J, Huang Y, Zhang R, 2021, CNN-Based Object Detection via Point-Cloud-Image Fusion. Opto-Electronic Engineering. [7]       Yang W, 2026, Research on Multi-Sensor-Fusion-Based SLAM Algorithm for Mobile Robots, thesis, University of Electronic Science and Technology of China. [8]       Wang M, Wei G, Cai J, et al., 2024, Robust Pose-Graph-Optimization Algorithm for 3D SLAM Under Heavy Noise. Control Engineering of China, 31(7): 1305–1313. [9]       Kong T, Gao H, Chen X, 2024, Full-Coverage Path Planning for Sweeping Robots Based on Improved RRT. Computer Engineering and Applications, 60(13): 311–318. [10]     Xia X, Wang D, Cao J, et al., 2022, LiDAR Point-Cloud Semantic-Segmentation Method for Vehicles Based on Sparse Convolutional Neural Networks. Automotive Engineering, 44(1): 10. [11]     Li B, Guo Y, Zhou J, et al., 2024, Development and Prospect of High-Definition Maps for Intelligent Driving. Geomatics and Information Science of Wuhan University, 49(4): 491–505. [12]     Xu J, Hou F, Cao G, 2022, Production Methods and Key Technologies for High-Definition Road Maps. Bulletin of Surveying and Mapping, 2022(1): 129–133. [13]     Yang M, Jiang K, Wen T, et al., 2023, Status and Challenges of Crowdsourced Updating Technology for Autonomous-Driving High-Definition Maps. China Journal of Highway and Transport, 36(5): 244–259. [14]     Chen L, Liu K, Zhou B, et al., 2021, Research on Key Technologies for Multi-Agent Collaborative High-Definition-Map Construction. Acta Geodaetica et Cartographica Sinica, 50(11): 10. [15]     Shomer H, Xu J, 2025, Automated Label Placement on Maps via Large Language Models, arXiv preprint arXiv:2507.22952.