Malicious User Detection for Cooperative Mobility Tracking in Autonomous Driving
Malicious User Detection for Cooperative Mobility Tracking in Autonomous Driving
复制标题
自动驾驶中协同移动跟踪的恶意用户检测
DOI：
10.1109/jiot.2020.2973661
复制
发表时间：
2020-02
影响因子：
10.6
通讯作者：
Wang Pi;Pengtao Yang;Dongliang Duan;Chen Chen-Chen;Xiang Cheng;Liuqing Yang;Hang Li
中科院分区：
文献类型：
--
作者：
Wang Pi;Pengtao Yang;Dongliang Duan;Chen Chen-Chen;Xiang Cheng;Liuqing Yang;Hang Li
在自动驾驶（AD）和智能交通系统（ITS）中，自身和周围车辆的移动状态为各种任务提供了重要信息。因此，一个精确、稳定、健壮的机动性跟踪框架是必不可少的。与仅依赖车载传感器（如全球定位系统（GPS）、惯性测量单元（IMU）和摄像头）的移动观测数据的自跟踪相比，协同跟踪通过车联网（IoV）中的车联网（V2X）通信整合路边单元（rsu）和附近车辆的观测数据，显著提高了移动信息的精度和可靠性。然而，如果有恶意用户在合作网络中发送虚假观察结果，那么合作跟踪就会非常脆弱。在本文中，我们提出了一种恶意用户检测框架，该框架包括两种顺序检测算法和安全移动数据交换与融合模型，以检测和去除虚假移动信息，并将所提出的检测算法与先前的数据融合算法相结合，从而保证了AD、ITS中的协同移动跟踪。仿真结果验证了该框架在不同攻击类型下的有效性和鲁棒性。
The mobility status of self and surrounding vehicles provides important information to various tasks in autonomous driving (AD) and intelligent transportation system (ITS). Accordingly, a precise, stable, and robust mobility tracking framework is essential. Compared with self-tracking that relies only on mobility observations from onboard sensors [e.g., global positioning system (GPS), inertial measurement unit (IMU), and camera], cooperative tracking markedly increases the precision and reliability of the mobility information by integrating observations from roadside units (RSUs) and nearby vehicles through vehicle-to-everything (V2X) communications in the Internet of Vehicles (IoV). Nevertheless, cooperative tracking can be quite vulnerable if there are malicious users sending bogus observations in the cooperative network. In this article, we present a malicious user detection framework, which includes two sequential detection algorithms and a secure mobility data exchange and fusion model to detect and remove bogus mobility information and integrate proposed detection algorithms with previous data fusion algorithms, which secures the cooperative mobility tracking in AD, ITS. Simulations validate the effectiveness and robustness of the proposed framework under different types of attacks.