Lane-Specific Driver Advice: driver persuasion and the role of driver workload.

Subject

Traffic jams are a major discomfort for the road user. On top of this they lead to societal costs estimated up to €1.8 billion annually in the Netherlands. Not all traffic jams need to be there: some occur while there is still spare road capacity left. One cause of these traffic jams is unbalanced lane use. More people may choose to drive on a specific lane than others, which creates a situation where small disturbances easily grow into traffic jams. The aim of the project is to optimise lane use with a lane-specific control application. By evening out traffic over the available lanes, we can drastically reduce traffic jams. In addition to this, a more balanced traffic system leads to increased safety and reduced emissions.

Scientific challenges

Balancing traffic lane usage will require advising a driver on what particular lane to drive on. A given advice will be beneficial to the traffic system as a whole, but may or may not be beneficial to the driver himself. One part of the research is about how a driver can be persuaded safely to comply with these types of altruistic advices.
The second part of the research is broader and focuses on how to safely interact with the driver using in-vehicle information systems (IVIS). Timing the interaction with the driver is critical for safety, and we theorize also for compliance. To time the messages we are working on driver workload prediction using driver physiology and performance measures using various methods, including machine/deep learning.

Societal relevance

The study focuses on the reduction of congestion, which will benefit drivers by reducing travel time, and will reduce pollution. The work on safe interactions between drivers and in-vehicle information systems aims to improve safety and reduce the distractions posed by in-car systems.

Paul van Gent

Start/end date: September 2015 – January 2020
Daily supervisor: Haneen Farah, Nicole van Nes (SWOV)
Promotor: Bart van Arem

/* */