A parsimonious enhanced Newell’s model for accurate reproduction of driver and traffic dynamics

• Published in: Transportation research. Part C, Emerging technologies Emerging technologies, 2023-09, Vol.154, p.104276, Article 104276
• Main Authors: Zheng, Shi-Teng, Jiang, Rui, Jia, Bin, Tian, Junfang, Bouadi, Marouane, Makridis, Michail A., Kouvelas, Anastasios
• Format: Article
• Language: English
• Subjects: Car following; Newell’s car-following model; Stimulus-response behavior; Stochasticity; Traffic oscillation
• ISSN: 0968-090X; 1879-2359
• DOI: 10.1016/j.trc.2023.104276

•Empirically investigate the stimulus-response behavior between leader and follower in car-following.•Propose a parsimonious enhanced Newell’s car-following model for accurate reproduction of driver and traffic dynamics.•Calibrate and validate the proposed model with two field car-following experiments.•Test the model in the aspects of the spontaneous oscillation, the concave growth pattern of oscillation, the vehicle movement prediction and the linear speed-capacity relationship.•Compare the performance of the proposed model with a state-of-the-art model, namely, the improved Newell’s car-following model with stochastic wave travel time. This paper investigates the stimulus-response behavior between leader and follower in car-following, based on vehicle trajectories in the prevailing field experiments. The analysis result indicates that the follower’s reaction time is time-varying, which can change significantly or keep a constant value for some time; and the follower cannot accurately track the leader in car following, which results in a residual from the follower’s speed. Inspired by the findings, this paper proposes a parsimonious enhanced Newell’s car-following model incorporating the stochastic reaction time and the fluctuation around the vehicle’s desired speed subject to the mean reversion process. The numerical experiment is carried out. It is shown that the proposed model can qualitatively and quantitatively reproduce the following important field observations: (i) the spontaneous formation and evolution of traffic oscillations, (ii) the concave growth pattern of traffic oscillations, (iii) the oscillations’ amplitude and frequency, (iv) the stochastic reproduction of individual trajectories, and (v) the linear speed-capacity relationship. The robustness of the proposed model is demonstrated, compared with the state-of-the-art model. Finally, the sensitivity analysis is carried out to evaluate the effect of each parameter of the proposed model.