Modelling the Effect of Lane-changing Mechanism in a Weaving Section

Weaving sections are subjected to complex lane-changing movements. In contrary to the other motorway sections where a driver selects a target lane and finds a suitable gap to change lanes, in weaving sections, the drivers’ choices can be significantly affected by the actions of the neighbouring drivers. For instance, if the leader vehicle is changing lanes in the same direction, the subject driver may be inclined to move as a platoon and accept smaller lead gaps to complete the lane change manoeuvre. Similarly, the acceptable gaps may be different if there is a weaving manoeuvre as opposed to a solo lane change with minimal interactions with the neighbouring drivers. Most of the existing lane-changing models however focus on solo lane changes and ignore the differences in behaviour due to the differences in lane-changing mechanisms. The current study, therefore, extends the state-of-the-art lane-changing models to explicitly account for the differences in behaviour under different lane-changing mechanisms using trajectory data collected from a weaving section of a Motorway in UK. The model parameters are calibrated using Maximum Likelihood Estimation technique and reveal significant differences in the parameters of the gap acceptance model under different lane-changing mechanisms. In particular, the results suggest that the relative speed to the current- and target-lane leaders have varying impacts on the gap acceptance behaviour. The developed models can have a significant impact in improving the fidelity of the micro-simulation outputs of the weaving sections.