基于车云协同的露天矿山数字地图更新方法  

Updating Method of Open-pit Mine Digital Map Based on Vehicle-cloud Collaboration

作　　者：李浩然 张利 陆鑫 王晓伟 欧阳文杰[1] 秦洪懋 LI Haoran;ZHANG Li;LU Xin;WANG Xiaowei;OUYANG Wenjie;QIN Hongmao(College of Mechanical and Vehicle Engineering,Hunan University,Changsha,Hunan 410082,China;Dong Feng USharing Technology Co.,Ltd.,Wuhan,Hubei 430056,China;Wuxi Intelligent Control Research Institute of Hunan University,Wuxi,Jiangsu 214115,China)

机构地区：[1]湖南大学机械与运载工程学院,湖南长沙410082 [2]东风悦享科技有限公司,湖北武汉430056 [3]湖南大学无锡智能控制研究院,江苏无锡214115

出　　处：《控制与信息技术》2022年第5期44-52,共9页CONTROL AND INFORMATION TECHNOLOGY

基　　金：国家重点研发计划(2021YFB2501800);国家自然科学基金(52172384);汽车车身先进设计制造国家重点实验室(61775006)。

摘　　要：高精数字地图是自动驾驶技术的基础,其为自动驾驶车辆提供精确的位置信息;但随着环境的变化和时间的推移,非结构化道路地图的保真性会逐渐下降。为此,文章提出了一种基于车云协同的露天矿山数字地图更新方法,以有效提高地图更新的效率。其首先构建了一种三层地图模型结构,以保证露天矿山地图更新的完整性;然后,针对露天矿山多种场景,提出了一种基于轨迹的地图更新方法,并设计了一种基于车云协同的地图更新系统。实验结果表明,采用基于车云协同的露天矿山数字地图更新方法,可有效降低地图更新时延,相比单车模式,地图平均更新时延降低了75%,保证了地图的鲜度;同时降低了地图更新的相对误差,整体精度提升了10%,满足地图保真性的要求。High-definition digital map is the foundation of autonomous driving technology, which provides precise location information for autonomous vehicles. However, with the change of environment and the passage of time, the fidelity performance of unstructured road map will gradually decline. Therefore, this paper proposes a digital map updating method of open-pit mine based on vehicle-cloud collaboration to effectively improve the efficiency of map updating. Firstly, a three-layer map model structure is constructed to ensure the integrity of map updating of open-pit mine. Then, a track-based map updating method is proposed for many scenarios of open-pit mines. Meanwhile, a map updating framework based on vehicle-cloud collaboration is designed to improve the efficiency of automatic driving map updating in open-pit mines. Experimental results show that this method can effectively reduce the map updating delay. Compared with the single-vehicle mode, the average map updating delay is reduced by 75%. At the same time, the relative error of map updating is reduced, and the overall accuracy is increased by 10%, which meets the requirements of map fidelity.

关 键 词：车云协同 自动驾驶 地图更新 高精数字地图 露天矿山 

分 类 号：U491.262[交通运输工程—交通运输规划与管理]