Virtual data mules for data collection in road-side sensor networks

Abstract

It has been proven that the use of mobile nodes as data mules in wireless sensor networks can improve the information collection process, increase the sensors' lifetime, and reduce the probability that sensors experience buffer overflow. To that end, data mules are typically required to follow precise trajectories. However, controlling the movement of mobile nodes is not always possible. In this paper, we introduce the concept of virtual data mule (VDM), which, by leveraging opportunistic cooperation between mobile nodes, abstracts the data collection function from specific physical mules. We apply the VDM concept to the case of data retrieval from road-side sensors through vehicular nodes. We propose two VDM algorithms targeting such a scenario and, by means of simulation in realistic mobility settings, we show that VDMs can fulfill data collection requirements that are simply unfeasible when traditional data mules are employed.