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. 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.

In order to develop such advisory systems, 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.

It is quite difficult to measure the traffic variables with the existing infrastructure. Hence it is important to develop a technique which can estimate the current state of traffic 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 behavior and the causes of lane changes on highways. The extent to which the drivers plan their path ahead and the time horizon they are taking into account are important in understanding the dynamics of the system.

Finally, these estimation and prediction models are integrated to develop a control model which will generate suitable advices to be given to the road users at the appropriate times. Advices can be in the form of what speeds to maintain, which lanes to use, when to change lanes and when not to etc. The approach to the lane specific control is shown in the figure.