Travel prediction-based data forwarding using realistic traffic model in vehicular networks

Vehicular Ad Hoc Networks (VANET) have become one of the most important research areas for the successful and deployment of intelligent transportation system and the unmanned vehicle technologies. In VANET, Travel Prediction-based Data Forwarding (TPD) was proposed as one of novel data forwarding schemes based on road traffic statistics. However, the TPD could not guarantee good performance in the environment with the traffic light because of the increasing deviation of the road traffic statistics. Since urban road networks have individual traffic lights at each intersection, it is necessary to solve the above problem. In this paper, we devise a method to optimize the TPD in order to make it operate so well in the realistic traffic model. Our performance optimizing method is carried out with a statistical approach and an algorithmic approach. In the statistical approach, we consider a way to control the rise of the deviation in statistics which is the main cause of the significant performance degradation of the TPD. In the algorithmic approach, we first improve the TPD with multi-hop delivery, and next we improve the probability calculating function. Our realistic simulation shows that our statistical and algorithmic optimization method improves the performance of the TPD on a realistic traffic model dramatically.