中文摘要：

对移动对象的不确定轨迹实时、高效的预测是智能交通系统研究热点之一.针对现有轨迹预测方法在动态环境中预测精度不高、实时效果差的问题,提出一种环境自适应车辆轨迹预测方法（EAVTP）,该方法主要步骤包括：首先运用历史数据构造虚拟参考点有效改进环境动态变化通讯信号不稳定情况下车辆位置不准确信息;其次利用高斯混合模型对虚拟参考点数据与历史轨迹数据集训练实现环境自适应功能;最后利用虚拟参考点和历史轨迹数据集对车辆轨迹实时预测.最后对所提方法模拟仿真,结果表明EAVTP方法具有一定的环境自适应性,且预测精度和实时性比现有其它方法有所提高.

英文摘要：

Real-time and efficient prediction to uncertain trajectory of moving object is one of the research hot-spots in intelligent trans- portation system. In this paper, we put forward an Environment Adaptive Vehicle Trajectory Prediction method（EAVTP） to solve the problem of low prediction accuracy and poor real time in the existing trajectories prediction methods under dynamic environment. The main steps included：First, using historical data to construct virtual reference point to valid improve the inaccuracy information of vehi- cle location caused by the unstable communication signal in dynamic environment;Secondly, training virtual reference point and history data set by Gaussian mixture models to achieve the environment adaptive function. ;Finally, using virtual reference point and historical data set to predict vehicle trajectory in real-time. The results of simulation shows that the proposed method is adaptive to environment, and is better than the existing methods in the prediction precision and real-time performance.