基于ESKF的数据融合车辆定位系统设计  

Vehicle localization and trajectory inspection system design based on ESKF

作　　者：张伟[1] 张健 赵奉奎[2] ZHANG Wei;ZHANG Jian;ZHAO Fengkui(Jiangsu Special Equipment Safety Supervision and Inspection Institute Wujiang Branch,Suzhou 215200,Jiangsu,China;College of Automotive and Transportation 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,南京210037

出　　处：《智能计算机与应用》2025年第3期192-197,共6页Intelligent Computer and Applications

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：自动驾驶车辆需要实时获取自身准确的定位结果进行轨迹规划和导航。为了提高定位精度,提出了一种基于误差状态卡尔曼滤波器(Error State Kalman Filter,ESKF)的多传感器数据融合定位算法,并进行了系统开发。系统由全球导航卫星系统(Global Navigation Satellite System,GNSS)、惯性测量单元(Inertial Measurement Unit,IMU)及处理器构成。基于ESKF设计数据融合算法,对IMU数据积分后得到系统名义状态,根据状态向量各变量的误差和零偏,对预测的状态向量进行校正,给出误差后验高斯分布,更新状态向量,迭代运算后,获取更加准确的定位和行驶轨迹。分别采用仿真数据和实车实验对本算法进行了验证,结果表明,本算法能够有效提高车辆的定位结果,准确记录车辆行驶轨迹。本算法对于智能车辆的定位功能开发及定位功能的检验具有重要的意义。Autonomous vehicles require real-time acquisition of accurate positioning results for trajectory planning and navigation.To improve positioning accuracy,a multi-sensor data fusion positioning algorithm based on Error State Kalman Filter(ESKF)has been proposed.The system is developed,which consists of Global Navigation Satellite System(GNSS),Inertial Measurement Unit(IMU),and processors.Using the ESKF-based data fusion algorithm,the IMU data is integrated to obtain the nominal system state.By considering errors and biases in the state vector variables,the predicted state vector is corrected,and the error posterior Gaussian distribution is determined.This process updates the state vector through iterative computations,resulting in more accurate vehicle positioning and trajectory information.This algorithm is validated using both simulation data and real vehicle experiments.The results indicate that the algorithm can effectively improve the vehicle′s positioning accuracy and accurately record the vehicle's trajectory.This algorithm holds significant importance for developing positioning functions in autonomous vehicles and the testing of intelligent vehicle positioning capabilities.

关 键 词：车辆定位 误差状态卡尔曼滤波 GNSS IMU 数据融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]