Vehicle lane changing track real-time prediction method based on deep neural network

The invention discloses a vehicle lane changing track real-time prediction method based on a deep neural network. The method comprises the steps of: obtaining running parameters and vehicle trajectories of a lane changing vehicle and a following vehicle; establishing an HMM lane changing decision model, and training a driver lane changing decision stage; establishing a lane changing execution model, and training a driver lane changing execution stage; judging whether a driver is in a lane changing state or not by using the trained HMM model according to the collected vehicle real-time data; and if the vehicle state is identified as a lane changing state, predicting the track of the lane changing vehicle by using the trained deep neural network. The method improves the prediction precision of the lane changing track of the vehicle, facilitates the analysis of collision risks, and improves the driving safety. 本发明公开了一种基于深度神经网络的换道过程轨迹预测方法获取换道车辆和跟驰车辆运行参数和车辆轨迹；建立HMM换道决策模型，进行驾驶员换道决策阶段的训练；建立换道执行模型，进行驾驶员换道执行阶段的训练；对采集到