Abstract: The movement of vehicles is regular and controlled by the consciousness of people. Through analyzing vehicles’ trajectory data and establishing the historical trajectory model the movement of vehicles in the future can be predicted. Firstly, a large number of real vehicle trajectory data is collected, the information entropy is calculated and the movement regularity which provides the basis for predicting vehicle trajectory is analyzed. Secondly, according to the special application scenarios of VANETs and the vehicle movement regularity, the state transition matrix is constructed using the history vehicle trajectory data and the vehicle trajectory prediction method based on markov chains is put forward. Simulation results show that the method can achieve the effective prediction of vehicle trajectory. Finally, some factors that influence the accuracy of the prediction are also analyzed.

Key words: vehicle movement regularity, trajectory prediction, information entropy, Markov chains

摘要： 车辆的移动是受人意识控制的，有规律的移动，通过对车辆已知轨迹数据的分析，建立历史轨迹模型，可以实现对其未来轨迹的预测。采集了大量真实的车辆轨迹数据，通过计算信息熵分析车辆运动的规律性，为预测车辆轨迹提供依据。根据车载自组织网络的特殊应用场景以及车辆移动的规律性，利用车辆轨迹的历史数据构建状态转移矩阵，提出了基于马尔科夫链的车辆轨迹预测方法，仿真结果表明，该方法可以实现对车辆轨迹的有效预测。对影响预测精度的一些因素进行了对比分析。

关键词: 车辆移动规律性, 轨迹预测, 信息熵, 马尔科夫链