The Predictability of Driving in Typical Traffic Conditions

A reliable prediction of the vehicle velocity over the course of a defined route plays a vital role for the growing electric mobility and autonomous driving. The prediction of upcoming traffic, driving conditions and thus velocity trajectory helps to determine the range of electric vehicles with high accuracy and enables for energy-optimized driving of autonomous vehicles. In this paper, the requirements for a reliable prediction are discussed and an advanced approach of velocity estimation using various online APIs for data acquisition is presented. A prediction of a typical route profile is carried out regarding all available online information, vehicle and driver limitations and predictable influences of infrastructure and route topology. To validate the results, real driving patterns of the exact route are compared to the predictions. Hence, capabilities and limitations of the presented prediction method are shown.