With the continuous development of the Internet of Vehicles, safety monitoring and assisted driving during vehicle operation has become the core of vehicle intelligence. In vehicle-assisted driving, state perception and trajectory prediction of surrounding vehicles is critical. Mathematical models mainly realize the traditional behavior and trajectory prediction of adjacent vehicles, and it is difficult to make reasonable use of the motion time series information, and it is difficult to predict the development trend further. In particular, traditional methods usually have some limitations, which limit the degree of real-time response and accuracy. This paper proposes an LSTM-based approach for predicting the behavior trajectory of adjacent vehicles. The network can autonomously learn the environmental relationship by constructing an adjacent vehicle interaction pool and digitizing the surrounding information. At the same time, the attention mechanism is comprehensively used to improve the accuracy and reliability of the prediction. It is expected that the research in this paper can help the research and optimization of vehicle intelligent assisted driving.
Behavior trajectory prediction research of adjacent vehicles based on LSTM
Eighth International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023) ; 2023 ; Hangzhou, China
Proc. SPIE ; 12790 ; 127904V
2023-09-07
Conference paper
Electronic Resource
English
Behavior trajectory prediction research of adjacent vehicles based on LSTM
British Library Conference Proceedings | 2023
|LSTM-Based Trajectory Tracking Control for Autonomous Vehicles
British Library Conference Proceedings | 2022
|Pedstrian Trajectory Prediction Based on LSTM
IEEE | 2023
|Research on Vehicle Trajectory Prediction Based on Improved LSTM Model
Springer Verlag | 2024
|