Micro-routing using accurate traffic predictions

This study presents a new way of routing vehicles through a network. Current routing algorithms base their advice on measurements of probe vehicles and sometimes do predictions based on historical data. However, if all vehicles would follow up that advice, those predictions would not be valid anymore. As a result, oscillations could arise, in turn causing traffic jams. The micro-routing algorithm presented in this study takes previous advice into account for the next advice and is able to interact with traffic light control programmes. In simulation, a reduction of travel time up to 30% and a reduction in the number of stops up to 46% were achieved without any oscillations. This shows high potential for solving congestion and reducing CO₂ emissions. Furthermore, a brief review of literature on route choice behaviour and driver response to route advice gives the reader some insight into driver compliance factors. As the major requirements of high reliability, predicted information and prescriptive information are met, also from a compliance perspective the micro-routing algorithm is expected to be effective.