摘要
针对城市复杂环境下车载组合导航系统定位精度不足及传统滤波算法模型参数固定、难以适应动态变化噪声的问题,文章提出了一种基于Sage-Husa算法的自适应无迹卡尔曼滤波(SH-AUKF)方法,并将其应用于GNSS/INS组合导航系统。该方法利用新息序列对过程与量测噪声的协方差矩阵进行实时估计和修正,旨在解决车辆在强非线性及动态变化场景下(如卫星信号丢失)的滤波精度下降甚至发散的问题。为系统性地验证该方法的有效性,本文设计了包含高速路、城市道路和隧道的典型驾驶场景,开展了实车对比试验,并将所提算法与标准的卡尔曼滤波(KF)、扩展卡尔曼滤波(EKF)及无迹卡尔曼滤波(UKF)进行了性能对比评估。结果表明:本文提出的SH-AUKF算法在卫星信号丢失的隧道路段及连续转弯的城市道路等关键场景下,展现出了最优的鲁棒性与定位精度;与标准UKF相比,其位置均方根误差(RMSE)显著降低。该方法为解决高动态环境下的车载组合导航问题提供了一种更有效且可靠的方案,对提升智能驾驶系统的环境适应性具有重要的应用价值。
Vehicle-mounted integrated navigation systems suffer from insufficient positioning accuracy in complex urban environments and traditional filtering algorithms with fixed model parameters struggle to adapt to dynamically changing noise.To address these challenges,this paper proposes an adaptive unscented Kalman filter(AUKF)method based on the Sage-Husa algorithm(SH-AUKF)and presents its application in a GNSS/INS integrated navigation system.The proposed method tackles the common problem of filtering accuracy degradation and even divergence under scenarios with strong non-linearity and dynamic changes,such as satellite signal loss,by utilizing the innovation sequence to perform real-time estimation and correction of the process and measurement noise covariance matrices.To systematically verify the effectiveness of this method,typical driving scenarios including expressways,urban roads,and tunnels were designed for real-vehicle comparative experiments.The proposed algorithm was evaluated against the standard Kalman filter(KF),extended Kalman filter(EKF),and unscented Kalman filter(UKF).Results show that the proposed SH-AUKF algorithm demonstrates optimal robustness and positioning accuracy in key scenarios such as tunnel sections with satellite signal loss and urban roads with continuous turns.Compared with the standard UKF,it significantly reduces root mean square errors(RMSE)in positioning.This method provides a more effective and reliable solution for addressing the challenges of vehicle integrated navigation in highly dynamic environments and holds significant practical value for enhancing environmental adaptability in intelligent driving systems.
作者
张琪
刘宁
朱丽璇
ZHANG Qi;LIU Ning;ZHU Lixuan(Beijing Key Laboratory of High Dynamic Navigation Technology,Beijing Information Science&Technology University,Beijing,100192,China)
出处
《控制与信息技术》
2025年第5期52-59,共8页
Control and Information Technology
基金
国家自然科学基金项目(62471048)。
关键词
智能驾驶
组合导航
自适应滤波
无迹卡尔曼滤波
定位精度
跑车试验
intelligent driving
integrated navigation
adaptive filtering
unscented Kalman filter
positioning accuracy
vehicle experiment
