拒绝服务攻击下基于事件触发无迹卡尔曼滤波的智能车位姿安全估计

In the complex networked environment where information attacks and noise interference both exist, the modern intelligent vehicles face great security risks. Hence, it is of great theoretical and practical significance to investigate the secure pose estimation of intelligent vehicles. Since some problems such as energy limit, lack of network bandwidth and computing resources are taken into consideration for intelligent vehicles, a method based on event-triggered UKF is proposed to estimate the pose of intelligent vehicles under DoS attacks. The main contents are presented as follows: (1) The nonlinear motion model which is suitable for different driving structures, and multi-sensor fusion measurement model are established in this project. (2) According to the characteristics of packet loss caused by DoS attacks, a model is established with data compensation measures. The multi-sensor fusion measurement model is then reconstructed on basis of the compensatory strategy. (3) The theoretical framework based on event-triggered UKF is constructed to securely estimate the pose of intelligent vehicles under DoS attacks. The theoretical results are verified, since the test platform of intelligent vehicles is built and the relationship between related design parameters and state estimation performance is explored. The completion of this project will enrich and develop the theory of secure pose estimation, which overcomes the problems of energy limit and insufficient resources, and has important guiding significance to the navigation and localization of intelligent vehicles in the complex networked environment.

在信息攻击和噪声干扰的复杂网络环境中，智能车位姿估计问题面临很大的安全风险。考虑智能车能源储备有限、网络带宽和计算资源不足等问题，本课题针对拒绝服务攻击环境中的智能车系统提出一类基于事件触发无迹卡尔曼滤波理论的位姿安全估计方法，具体研究内容包括：(1)提出多模驱动下智能车在三维空间的非线性运动模型和多传感器融合测量模型；(2)根据拒绝服务攻击引发的数据丢包特性，建立带有数据补偿措施的拒绝服务攻击模型，在此基础上重构智能车系统的多传感器融合测量模型；(3)构建拒绝服务攻击下基于事件触发无迹卡尔曼滤波的智能车位姿安全估计理论框架，搭建智能车系统测试平台，对理论结果进行验证，并探索相关设计参数与状态估计性能之间的关系。本课题的完成将会丰富和发展智能车的位姿安全估计理论，克服智能车能源有限、系统资源不足等问题，对复杂网络环境下的智能车导航定位技术具有重要指导意义。

在信息攻击和噪声干扰的复杂网络环境中，智能车位姿估计问题面临很大的安全风险。考虑智能车能源储备有限、网络带宽和计算资源不足等问题，本课题针对拒绝服务攻击环境中的智能车系统提出了一类基于事件触发无迹卡尔曼滤波理论的位姿安全估计方法，研究成果包括：.(1)提出了多模驱动下智能车在三维空间的非线性运动模型和多传感器融合测量模型；.(2)根据拒绝服务攻击引发的数据丢包特性，建立了带有数据补偿措施的拒绝服务攻击模型，在此基础上重构智能车系统的多传感器融合测量模型；.(3)构建了拒绝服务攻击下基于事件触发无迹卡尔曼滤波的智能车位姿安全估计理论框架，搭建了智能车系统测试平台，对理论结果进行验证，发表了相关高水平论文10余篇。.本课题的完成丰富和发展了智能车的位姿安全估计理论，克服了智能车能源有限、系统资源不足等问题，对复杂网络环境下的智能车导航定位技术具有重要指导意义。