Practical location-based routing in vehicular ad hoc networks

Rapid advancement in wireless communication has made it possible to develop vehicular ad hoc networks, in which a vehicle can communicate with other vehicles via a wireless, multi-hope fashion. A variety of appealing real-world applications can be enabled by VANETs, such as driving safety and urban monitoring. Many location based routing algorithms have been proposed for data delivery in VANETs. Most of them assume that accurate location information is available when needed. In practice, however, such assumption is unrealistic. It incurs considerable cost to retrieve location information. In addition, a vehicle is on the fast move over time, and a location previously obtained may become invalid after certain time. This paper proposes a routing algorithm that is based on a practical location information model. To solve the problem of location inaccuracy and vehicle mobility, we devise a location predictor which estimates the possible location of a vehicle by using history information. Based on the greedy forwarding strategy, the proposed routing differentiates packets in terms of closeness to destination and jump distance. We evaluate the performance of the proposed algorithms with a large real trace of taxi motion in Shanghai. Trace-driven simulation results demonstrate that data delivery performance is improved.