The Dynamical Prediction of V2V Link Duration in Urban VANETs

Link duration prediction is an important standard which determines many performance of network in VANETs. Existing analytical methods about link duration based on mobility of nodes in VANETs have no function to predict link duration between any two nodes in the future, so it is not practical for these methods to predict link duration between two vehicles. We propose a dynamical prediction model which considers the distribution of relative velocity, inter-vehicle distance, traffic density change and traffic light to estimate the expected link duration between any pair of connected vehicles, because these factors change continuously in the process of link connection. By taking into account the relative velocity distribution, the model is able to adjust the principle in real time to adapt variation of vehicle speed. By automatically adjusting computing method of the relative distance between two vehicles, DPLD(dynamically predict link duration) model can automatically adapt to the change of relative distance between two vehicles. Therefore, DPLD model can effectively predict the link duration between the two vehicles. Such model is implemented on each vehicle along with parameters estimation methods of relative velocity distribution, exponential moving average method processes speed exception and considering the impact of the traffic light on link duration. Simulation results show that this model predict link duration for urban scenario has the high accuracy.