Building a human-like car-following model that can accurately simulate drivers’ car-following behaviors is helpful to the development of driving assistance systems and autonomous driving. Recent studies have shown the advantages of applying reinforcement learning methods in car-following modeling. However, a problem has remained where it is difficult to manually determine the reward function. This paper proposes a novel car-following model based on generative adversarial imitation learning. The proposed model can learn the strategy from drivers’ demonstrations without specifying the reward. Gated recurrent units was incorporated in the actor-critic network to enable the model to use historical information. Drivers’ car-following data collected by a test vehicle equipped with a millimeter-wave radar and controller area network acquisition card was used. The participants were divided into two driving styles by K-means with time-headway and time-headway when braking used as input features. Adopting five-fold cross-validation for model evaluation, the results show that the proposed model can reproduce drivers’ car-following trajectories and driving styles more accurately than the intelligent driver model and the recurrent neural network-based model, with the lowest average spacing error (19.40%) and speed validation error (5.57%), as well as the lowest Kullback-Leibler divergences of the two indicators used for driving style clustering.

Objective: Maintaining a lower speed is recommended for curve negotiation and it has been shown that cognitive distraction may impair driving performance. This study examines the effects of different levels of cognitive distraction on the speed control of drivers negotiating a curve. Methods: Experiments were conducted on a 6 degrees of freedom driving simulator with 28 participants. A comparison of the speed and acceleration from 300 m before to 100 m after the curve was performed for baseline driving (without distraction) and 3 levels of cognitive distractions using n-back tasks. Results: The speed was significantly higher at the highest level of cognitive distraction (8%) than in baseline driving from the beginning of the curve to 50 m after it and the ratio of the highest level to the baseline was even greater throughout the range. The average acceleration was significantly higher than the baseline at the highest and the medium levels (80 and 70%, respectively) from 250 m before the curve to the one-quarter curve and from 250 to 150 m before the curve, respectively. It was also found that the point of deceleration was significantly delayed at the highest level of cognitive distraction and occurred only after the middle of the curve, whereas deceleration in baseline driving occurred just before entering the curve. Conclusions: The impairment due to cognitive distraction was confirmed in this study but was only significant at a high level of cognitive distraction. The highly distracted drivers failed to perceive the curve in advance, resulting in a slower response to changes in the roadway. The findings indicate that acceleration may be an indicator of cognitive distraction while negotiating curves. A driver is prone to cognitive distraction and the driving performance is affected when driving requires excessive attention such as curve negotiation.