Aiming at the problem that it is difficult for the following vehicle to quickly acquire the driving state of the preceding vehicle, the LSTM model of car-following is proposed in connected vehicle environment. First, the driving scene with connected vehicle environment is set up using the driving simulator. Both vehicle status data of the following vehicle and the preceding vehicle are collected. Second, two time series data mentioned above are taken as input, and the LSTM model of car-following is constructed to obtain the appropriate speed and acceleration on the basis of the car-following condition. Finally, the collected data is used to verify the proposed model, and the accuracy rate is 85.2%. The result shows that the car-following model based on LSTM network has a good effect in the following state of time series in connected vehicle environment. The result can provide a method supporting for car-following in connected vehicle environment.

CICTP