Differences in Optimised Trajectories Under Selfish and collaborative Behaviour of multi-class Freeway Traffic: Fid-Bau Portal

Differences in Optimised Trajectories Under Selfish and collaborative Behaviour of multi-class Freeway Traffic

Wens, Maarten / Arman, Mohammad Ali / Abuamer, Ismail / Tampere, Chris / Vansteenwegen, Pieter

This paper proposes a Mixed Integer Linear Program (MILP) to optimise vehicle trajectories in two scenarios, user optimum (UO) and system optimum (SO), where the goal is to show the impact on the traffic in terms of speed, etc. The study provides insight into future control aspects of CAVs, and in the case if empirical data is available, this can help to impose some traffic regulations to improve the existing traffic conditions. This model assumes a central planner that schedules complete vehicle trajectories over a complex motorway weaving area. The model can handle the heterogeneity of a fleet of vehicles. In the short term, the optimised driving patterns of this model can be seen as a direction of improvement that guidelines and driving regulations should try to achieve. While on the longer term, it could help to guide autonomous vehicles. This paper will focus on the difference between optimised trajectories when drivers are assumed to be selfish, which corresponds to UO, versus collaborative behaviour, which corresponds to SO. One of the conclusions is that doing additional lane changes to open up lanes for fast vehicles benefits traffic as a whole in collaborative conditions.