Learning Car-Following Behaviors for a Connected Automated Vehicle System: An Improved Sequence-to-Sequence Deep Learning Model

Abstract

Data-driven car-following modeling is of great significance to traffic behavior analysis and the development of connected automated vehicle (CAV) technology. The existing researches focus on reproducing the car-following process by capturing the behavior of the host vehicle using the information of its nearest preceding vehicle. While the other preceding vehicles may affect the host vehicle as well. To fill the gap above, this paper presents an improved sequence-tosequence deep learning-based (ISDL) car-following model for a CAV system. Firstly, the kinematics information considering the multiple preceding vehicles are organized as the input characteristics. Secondly, an improved sequence-to-sequence deep learning framework is proposed by integrating an encoder with the bidirectional gated recurrent unit (GRU) neural network and a decoder using an attention-based GRU neural network in an end-to-end fashion. Finally, the car-following data with multiple preceding vehicles captured from the NGSIM dataset are employed to train and calibrate the proposed model. Experimental results indicate that the deep learning-based models' performance in learning heterogeneous driving behavior can be enhanced by adding information about multiple preceding vehicles. In addition, the proposed ISDL model outperforms the benchmark car-following models in terms of the accuracy of the simulated speeds and simulated positions. Through tests on platoon simulation, the ISDL model is capable of reshaping the traffic oscillation phenomenon as well.INDEX TERMS Car-following modeling; improved sequence-to-sequence model; information flow topology; gated recurrent unit neural network; connected automated vehicle

Deep Reinforcement Learning Car-Following Control Based on Multivehicle Motion Prediction

“…The parameter θ typically originates from a neural network, known as the policy network, which approximates the policy function π(a|s). For a value-based method, the agent achieves the optimal policy by continuously modifying the policy according to updates from Equation (11), with π * = arg max Q π (s t , a t ). It is noteworthy that policy-based techniques, reliant on cyclic updates via the Monte Carlo approach, frequently manifest as algorithmic inefficiencies accompanied by pronounced variance.…”

“…Through the amalgamation of comprehensive field data and data-driven strategies, this innovative paradigm extends beyond the confines of traditional modeling techniques, embracing a multitude of variables to forge models of unparalleled depth and relevance. Particularly noteworthy is the adoption of artificial neural networks (ANNs) [6][7][8] and recurrent neural networks (RNNs) [9][10][11][12][13], which epitomize the synthesis of advanced pattern recognition and sequential data analysis. These methodologies not only refine the granularity of driver behavior predictions but also enhance the models' adaptability, transcending the limitations of their predecessors.…”

A High-Precision Car-Following Model with Automatic Parameter Optimization and Cross-Dataset Adaptability

“…Lu et al established a car-following model using an improved sequence-to-sequence deep learning framework, considering the kinematic information of multiple leading vehicles. This enhancement improves the model's ability to learn heterogeneous driving behaviors and reshape traffic oscillation [17]. Qin et al proposed a new car-following model combining CNN and LSTM, which can predict the speed of the following vehicle more accurately [18].…”