Destination Prediction with Decision Tree An Intelligent Driver Assistant for Volvo Cars: Destination Prediction, Traffic Notifying and Route Advising

Abstract

In this thesis, a destination prediction method based on decision tree model will be presented. The method will be used to support the Volvo Car Intelligent Driver Assistant System. Accurate destination prediction method is useful in many cases, for instances, it helps to free driver's hand of manually inputting destination address, provide traffic jam report before running into traffic jam and recommend the optimal route with the supporting of third parties' services. The mean accuracy of the destination prediction introduced in this thesis can achieve about 75%. By collaborating with the INRIX Traffic API, WirelessCar API and Volvo On Call app, the Intelligent Driver Assistant System achieves the goal to notifying Volvo Car drivers with traffic information and providing possible routes to the predicted destination.