基于车辆与车辆的车联网分布式协同感知定位

Vehicle to vehicle-based distributed cooperative sensing positioning for internet of vehicles

中文关键词  车载自组织网络  到达时间  车辆到车辆  协同定位  无迹卡尔曼滤波

英文关键词  vehicular ad-hoc networks  time of arrival  vehicle-to-vehicle  cooperative positioning  unscented Kalman filter

中文摘要

      车辆协同感知定位是车辆定位的热点技术. 针对车载自组织网络, 本文在GPS卫星导航和车辆自身航位推 算(DR)的基础上, 利用车辆之间的到达时间(TOA)观测和车辆到车辆(V2V)的实时通信来设计一种分布式协同定位 方法. 针对协同定位中TOA测量函数的非线性和辅助车辆真实位置未知的问题, 提出了一种基于改进无迹卡尔曼 滤波(UKF)的协同定位算法. 相对于传统的UKF协同定位使用GPS观测值作为辅助车辆位置, 本文提出的算法将辅 助车辆位置作为未知参数, 扩维到状态向量, 有效降低了辅助车辆位置误差对定位精度的影响. Monte Carlo仿真结 果表明, 利用TOA观测的协同定位精度明显优于仅利用GPS和DR的独立定位精度, 且改进的UKF协同定位算法相 比于传统UKF算法, 具有更高的定位性能.

英文摘要

      Vehicle cooperative sensing positioning is a hot technology of vehicle positioning. This paper explores the distribute cooperative positioning problem by incorporating the time of arrival (TOA) observations into the GPS satellite navigation and vehicle’s own dead reckoning (DR) information for the vehicular Ad-hoc Networks. In the proposed distribute cooperative positioning method, an improved unscented Kalman filtering (UKF) algorithm is designed to deal with the nonlinearity of the TOA measurement function as well as the inaccurate locations of the auxiliary vehicles. In comparison with the traditional UKF cooperative positioning method, which uses the GPS observation as the corresponding auxiliary vehicle location, the proposed algorithm takes the auxiliary vehicle locations as unknown parameters and extends the dimension of the state vector, therefore, the influence of the location errors in the auxiliary vehicles is effectively reduced. Monte Carlo simulation results show that the cooperative positioning accuracy by incorporating the TOA information is obviously better than that of the independent positioning using only GPS and DR, and the performance of the improved UKF cooperative positioning algorithm is considerably more accurate than that of the traditional UKF algorithm.