Title: A Probabilistic Approach to Driver Assistance for Delay Reduction at Congested Highway Lane Drops

Abstract: This paper proposes an onboard advance warning system based on a probabilistic prediction model that advises vehicles on when to change lanes for an upcoming lane drop. Using several traffic- and driver-related parameters such as the distribution of inter-vehicle headway distances, the prediction model calculates the likelihood of utilizing one or multiple lane changes to successfully reach a target position on the road. When approaching a lane drop, the onboard system projects current vehicle conditions into the future and uses the model to continuously estimate the success probability of changing lanes before reaching the lane-end, and advises the driver or autonomous vehicle to start a lane changing maneuver when that probability drops below a certain threshold. In a simulation case study, the proposed system was used on a segment of the I-81 interstate highway with two lane drops - transitioning from four lanes to two lanes - to advise vehicles on avoiding the lane drops. The results indicate that the proposed system can reduce average delay by up to 50% and maximum delay by up to 33%, depending on traffic flow and the ratio of vehicles equipped with the advance warning system.