Research On Neighbor Discovery In Wireless Ad Hoc Networks

Neighbor discovery is one of the important steps in the process of self-organization of wireless ad hoc networks, which has significant impact on the network performance. With the introduction of directional antennas to ad hoc networks, new challenges come up in neighbor discovery.We design and implement some novel neighbor discovery schemes for static wireless ad hoc networks based on different types of antenna. Our work mainly makes the following contributions:(1) We propose a novel channel access strategy for neighbor discovery of ad hoc networks. Most randomized neighbor discovery algorithms proposed before improved the efficiency of neighbor discovery by adjusting the probability of transmission and reception of wireless nodes. However, the ratio of the channel usage cannot be effectively raised in this way. The dual channel competition scheme we design could solve this problem, which can use the wireless channel more effectively and then improve the efficiency of neighbor discovery.(2) We propose some neighbor discovery algorithms with auxiliary tones. Basing on the dual channel competition scheme, we design some busy-tone aided algorithms for neighbor discovery. When using directional antennas, with the help of omni-directional tones, the wireless nodes know the right direction to which their active antenna should point. In this way, the communication efficiency can be improved. In ad hoc networks with omni-directional antennas, the busy-tone aided algorithm also outperforms traditional randomized neighbor discovery schemes.(3) We proposed a centralized neighbor discovery scheme for ad hoc networks with hybrid uni/omni-directional, through which we can solve the neighbor discovery problem when directional and omni-directional antennas exist simultaneously. In the meanwhile, the neighborhood information of the whole network can be gathered to the centralized node. Our research compensates for the inadequacy of the relevant research domestically. The results of our research can be used by protocols in upper layers of ad hoc networks. Analysis and simulation show that, the neighbor discovery algorithms we propose are effective and efficient.