Dynamic Model Based Malicious Collaborator Detection in Cooperative Tracking
Dynamic Model Based Malicious Collaborator Detection in Cooperative Tracking
复制标题
DOI：
10.1109/wcnc45663.2020.9120552
复制
发表时间：
2020-05
期刊：
影响因子：
--
通讯作者：
Wang Pi;Pengtao Yang;Dongliang Duan;Chen Chen-Chen;Xiang Cheng;Liuqing Yang
中科院分区：
文献类型：
--
作者：
Wang Pi;Pengtao Yang;Dongliang Duan;Chen Chen-Chen;Xiang Cheng;Liuqing Yang
车辆的移动状态在自动驾驶汽车（AV）和智能交通系统（ITS）的大多数任务中起着至关重要的作用。为了安全运行，精确、稳定和强大的移动跟踪系统至关重要。与仅依赖于车载传感器（例如全球定位系统（GPS）、惯性测量单元（IMU）和摄像头）的移动性观测的自跟踪相比，协作跟踪通过V2X通信整合来自路边单元和附近车辆的观测，大大提高了移动性数据的精度和可靠性。然而，如果网络中有恶意的合作者发送虚假的观察结果，合作跟踪可能非常脆弱。本文提出了一种动态序列检测算法，基于动态模型的平均状态检测（DMMSD），以排除虚假的移动数据。仿真结果验证了该算法的有效性和鲁棒性，与现有的方法相比。
The mobility status of vehicles play a crucial role in most tasks of Autonomous Vehicles (AVs) and Intelligent Transportation System (ITS). To operate securely, a precise, stable and robust mobility tracking system is essential. Compared with self-tracking that relies only on mobility observations from on-board sensors (e.g. Global Positioning System (GPS), Inertial Measurement Unit (IMU) and camera), cooperative tracking increases the precision and reliability of mobility data greatly by integrating observations from road side units and nearby vehicles through V2X communications. Nevertheless, cooperative tracking can be quite vulnerable if there are malicious collaborators sending bogus observations in the network. In this paper, we present a dynamic sequential detection algorithm, dynamic model based mean state detection (DMMSD), to exclude bogus mobility data. Simulations validate the effectiveness and robustness of the proposed algorithm as compared with existing approaches.