Taking the Fast Lane: Lane-specific traffic flow control models

Subject

In this project, we aim to find the best lane to drive on! Should you take the right or the left lane in the traffic flow? Which lane will go faster? Which lane will be more comfortable; where do you need to accelerate and brake a lot? But also, In what lane can you facilitate this fast and comfortable trip for others, and how much do you need to give in?Underutilization of lanes on multi-lane motorways can lead to traffic jams setting in on heavily used lanes, while capacity is still available on other lanes. Currently, traffic control applications are limited to the roadway level. There is a need for lane specific traffic control measures which can make optimum use of all the lanes by influencing the “longitudinal” and “lateral” behaviour of road users. This project aims at developing a first of its kind lane specific motorway traffic control system to reduce the individual travel times, improve the overall traffic flow efficiency on multi-lane highways by optimizing the distribution of traffic over the lanes.Scientific

Challenges

In order to develop advisory systems for lane level traffic control, it is quite important to estimate the current state and predict the future state of traffic networks on a lane level. This will help in understanding when to generate the advices and send them to the road users. The traffic state is primarily characterized by the average speed and the number of vehicles flowing over a section of the road. Traffic states are not observed everywhere and observed measurements suffer from noises, reliability errors etc. Hence it is important to develop a technique which can estimate the current state of traffic accurately with the limited amount of information available and is also able to combine data collected from multiple sources. Whereas traffic state estimation describe the current state of traffic, the prediction methods try to describe the state of traffic sometime in the future. For a good prediction model, it is important to understand the individual driver behaviour and the causes of lane changes on highways which are difficult to model. The developed estimation and prediction models along with an understanding on lane change decisions and behaviour are integrated to develop a control model.

Societal Relevance

The developed control model is expected to generate suitable advices to be given to the road users (via in-car systems) at appropriate times with the aim to improve the overall traffic performance on multi-lane motorways and reduce delays to road users. Advices can be in the form of what speeds to maintain, which lanes to use, when to change lanes and when not to etc.