Extended Neighbor Discovery In Wireless Opportunistic Networks

With the development of information industry and the popularity of pocket devices, our life is closely tied to Internet services. However, due to the limitations of current wireless access technologies, people living in complex Environment such as Marine, Remote rural areas still cannot enjoy the Internet services conveniently and inexpensively. Under such circumstance, opportunistic Networks arise at the historic moment. Opportunistic networks are such special kinds of ad hoc networks who employ nodes’ mobility to deliver messages between source nodes and destination nodes, even if there does not ever exist any complete continuous path, which is considered as a significant advantage in extreme environment compared to the forwarding mechanism in traditional ad hoc networks.In opportunistic network, Communications between hosts rely on nodes’ mobility. To successfully route messages in such networks, it is essential to efficiently discover neighbors and collect their information. However,so far,there are few papers studying this problem.In addition, nodes’ energies are limited by battery capacities. Even worse, messages are supposed to be stored in nodes for a long time before forwarding to the next successor, which dramatically intensifies energy consumptions. Therefore, we present three principles for designing effective neighbor discovering methods in the opportunistic network: accuracy, energy efficiency and utilization of neighbor information. Based on these ideas, we propose a novel algorithm which combines an “ack” mechanism and the power regulation algorithm to improve the efficiency of neighbor discovering.Then we establish the proposed algorithm and analyze its performance on the state-of-art network simulator- EXata. From the first experiment, we learn the linear relationship between neighbor number and transmission distance. Inspired by this experiment, we set nodes’ transmitting power levels according to nodes’ distances. The second experiment shows that our method Outperforms the classic algorithm in energy consumptions by aproximately cutting 30%. From the Third and the last experiment, we can see that the efficiency of our algorithm decreases as the velocities of nodes accelerate. In addition, the result shows that our algorithm cuts power consumption by 20% while maintains a high delivery probability with the rate of 80%.Our Neighbor discovering algorithm performs well in such environments whose nodes move slowly, such as marine scenarios.