Improving two-dimensional intelligent driver models to overcome overly high deceleration in car-following

In the Two-Dimensional Intelligent Driver Model (2D-IDM), desired time gap changes over time. The model can reproduce concave growth pattern of traffic oscillations and fit the experimental results quantitatively well. However, overly high deceleration is observed in the model. To overcome the deficiency, this paper proposes that changing rate of desired time gap should be restricted. Incorporating this restriction enables the 2D-IDM to eliminate overly high deceleration, and still has the good performance to reproduce oscillation growth pattern. Furthermore, it is shown that the restriction also eliminates overly high deceleration in the 2D-IIDM, which is another improved version of 2D-IDM that is able to reproduce synchronized flow. (C) 2019 Elsevier B.V. All rights reserved.