考虑通信时延的智能网联汽车协同定位方法  

Cooperative Localization Method for Connected and Autonomous Vehicles Considering Communication Delay

作　　者：闵海根 李尧 汪建球 张磊 李杰 童星 吴霞[1,2] 赵祥模[1,2] MIN Hai-gen;LI Yao;WANG Jian-qiu;ZHANG Lei;LI Jie;TONG Xing;WU Xia;ZHAO Xiang-mo(School of Information and Engineering,Chang'an University,Xi'an 710064,Shaanxi,China;The Joint Laboratory for Internet of Vehicles,Ministry of Education-China Mobile Communications Corporation,Chang'an University,Xi'an 710021,Shaanxi,China;China Mobile(Shanghai)Information and Communication Technology Co.Ltd.,Shanghai 201206,China;Shandong Hi-speed Information Group Co.Ltd.,Jinan 250014,Shandong,China)

机构地区：[1]长安大学信息工程学院,陕西西安710064 [2]长安大学“车联网”教育部-中国移动联合试验室,陕西西安710021 [3]中移(上海)信息通信科技有限公司,上海201206 [4]山东高速信息集团有限公司,山东济南250014

出　　处：《中国公路学报》2023年第6期220-234,共15页China Journal of Highway and Transport

基　　金：国家自然科学基金青年科学基金项目(61903046);陕西省创新能力支撑计划-创新人才推进计划项目(2023KJXX-020);陕西省重点研发计划项目(2021GY-290);陕西省自然科学青年基金项目(2022JQ-663);陕西省高校科协青年人才托举计划项目(20200106);中央高校基本科研业务费专项资金项目(300102243202,300102243501,300102243711,300102242103)。

摘　　要：为了满足网联环境下自动驾驶车辆安全行驶的需求,必须实现车辆全时空高精度定位。针对单车定位(Single Vehicle Localization,SVL)方法的不足,提出了一种基于双层滤波结构的智能网联汽车协同定位框架。首先,基于卡尔曼滤波对各车辆状态进行修正;然后设计基于联邦卡尔曼滤波的协同定位估计方法,通过构建一个主滤波器和多个局部滤波器,将本车状态与修正后的邻车状态进行融合;使用多种数据拟合方法,基于真实数据构建传输时延概率模型,基于高斯分布构建处理时延概率模型;此外,提出一种通信时延误差补偿方法,并融入协同定位框架;最后,设计了5组仿真试验,评估SVL、未进行通信时延误差补偿的协同定位方法(CLWC)和基于通信时延误差补偿的协同定位方法(CLC)的定位性能,并深入分析了速度和行驶方向对定位结果的影响。研究结果表明:在城市道路环境下,CLWC相较于SVL,精度提高了15%~23%;在空旷道路环境下,通信时延较小情况时,CLWC优于SVL,CLC在CLWC基础上将精度进一步提高了5%~13%。在长直道、弯道、隧道等场景,CLC能够保证定位轨迹平滑,精度明显高于SVL,同时进一步验证了存在通信时延情况下,车辆速度对协同定位的影响。所提方法不仅克服了SVL误差累积的缺陷,同时有效降低了通信时延的影响,可为车辆提供连续稳定的高精度位置。To satisfy the requirements for safe driving of autonomous vehicles in a networked environment,it is necessary to achieve high-precision localization of vehicles in both time and space.In response to the shortcomings of the single-vehicle localization(SVL)method,a cooperative localization framework for connected and autonomous vehicles based on a double-layer filtering structure was proposed.First,the state of each vehicle was corrected based on the Kalman filter.Then,a cooperative localization estimation method based on the federal Kalman filter was designed.A main filter and multiple local filters were constructed to fuse the state of the ego vehicle and the corrected state of the neighboring vehicle.Using a variety of data fitting methods,a transmission delay probability model was constructed based on real data,and a processing delay probability model was constructed based on a Gaussian distribution.In addition,a communication delay compensation method was proposed and integrated into a cooperative localization framework.Finally,five groups of simulation experiments and three groups of real-vehicle experiments were designed to evaluate the localization performance of SVL,the cooperative localization method without delay error compensation(CLWC)and the cooperative localization method based on delay error compensation(CLC).In addition,the influences of speed and driving direction on the localization results were analyzed in detail.The research results showed that in an urban road environment,the accuracy of CLWC was improved by 15%-23%compared with SVL.In an open-road environment,when the communication delay was small,CLWC was better than SVL,and CLC further improved the accuracy by 5%-13%based on CLWC.In scenarios such as long straight roads,curves,and tunnels,CLC could ensure that the localization trajectory was smooth,and its accuracy was significantly higher than that of SVL.Simultaneously,it further verified the influence of vehicle speed on cooperative localization in the presence of communication delay.Th

关 键 词：交通工程 协同定位 误差补偿 通信时延 信息融合 

分 类 号：U495[交通运输工程—交通运输规划与管理]