DAWN: A density adaptive routing algorithm for vehicular delay tolerant sensor networks

Vehicular Sensor Networks use moving vehicles to provide sensing coverage of the city. By adopting Delay Tolerant Network (DTN) techniques, the system enjoys very low deployment and operational costs. Therefore, the Vehicular Delay Tolerant Sensor Network (VDTSN) has become a promising solution for the urban sensing applications. However, routing in the VDTSN is a great challenge due to uneven and fluctuant node density caused by the vehicle mobility pattern, and the high node density in urban areas makes the local channel capacity an impacting factor to network performance. In this paper, we propose a VDTSN framework with the consideration of the channel capacity, and then a utility function reflecting the network performance in terms of node density is established within the framework. Based on the framework, we propose a routing algorithm for VDTSN, namely DAWN (Density Adaptive routing With Node awareness). DAWN enables the vehicle node's awareness of its neighbor nodes density, and the transmission behaviors are fine-tuned according to the local channel capacity and the node density to maximize the data delivery probability within the delay constraint. Through simulations over the random walk on the Manhattan Grids and the real GPS traces of about 5000 taxi cabs for 30 days in the Beijing city, DAWN is proved to outperform other classical routing schemes.