基于图优化的智能车辆多传感器融合定位方法  

Graph Optimization-Based Multi-Sensor Fusion Localization Method for Intelligent Vehicles

作　　者：张伟[1] 李旭东 曹伟 赵奉奎[2] ZHANG Wei;LI Xudong;CAO Wei;ZHAO Fengkui(Jiangsu Special Equipment Safety Supervision Inspection Institute Wujiang Branch,Suzhou 215200,China;College of Automobile and Traffic Engineering,Nanjing Forestry University,Nanjing 210037,China)

机构地区：[1]江苏省特种设备安全监督检验研究院吴江分院,江苏苏州215200 [2]南京林业大学汽车与交通工程学院,江苏南京210037

出　　处：《软件工程》2025年第4期53-56,72,共5页Software Engineering

基　　金：江苏省特种设备安全监督检验研究院科技计划项目(KJ(Y)2023042)。

摘　　要：为了提升车辆定位系统的精度和鲁棒性,针对单一传感器存在的局限性,提出了一种基于图优化的LIDAR(Light Detection and Ranging,LIDAR)、IMU(Inertial Measurement Unit)和GNSS-RTK(Global Navigation Satellite System-Real-Time Kinematic)的多传感器车辆定位方法。首先,使用IMU预积分模型,通过滑动窗口和扫描匹配的方法构建LIDAR里程计因子,加入GNSS-RTK绝对测量值以修正系统的长期漂移;其次,使用因子图优化框架将LIDAR、IMU和GNSS-RTK的测量数据进行融合,并加入回环检测因子,通过求解最大后验估计以获取最佳的定位结果。实验结果显示,所提出方法的相对平移误差低至0.34 m,具有较高的准确性和鲁棒性,弥补了单传感器的不足,提高了车辆定位系统的定位精度。To enhance the accuracy and robustness of vehicle localization systems and address the limitations of single-sensor approaches,this paper proposes a graph optimization-based multi-sensor fusion localization method integrating LIDAR(Light Detection and Ranging),IMU(Inertial Measurement Unit),and GNSS-RTK(Global Navigation Satellite System-Real-Time Kinematic).Firstly,an IMU pre-integration model is employed to construct LIDAR odometry factors through sliding window and scan matching techniques,while GNSS-RTK absolute measurements are incorporated to correct long-term system drift.Subsequently,a factor graph optimization framework is utilized to fuse measurements from LIDAR,IMU,and GNSS-RTK,augmented with loop closure detection factors.The optimal localization results are obtained by solving the maximum a posteriori estimation.Experimental results demonstrate that the proposed method achieves a relative translation error as low as 0.34 m,exhibiting high accuracy and robustness.This approach effectively compensates for the shortcomings of single-sensor systems and significantly improves the positioning precision of vehicle localization systems.

关 键 词：图优化 多传感器融合 智能车辆 定位 回环检测 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]