Research On Co-Synchronization Method And Channel Access Technology In Wireless Ad Hoc Networks

Wireless ad hoc is a distributed wireless multi-hop network with no-central node and distributed deployment, which has a rapit and dependent characteristic. The nodes in Ad hoc are equal and independent, without supporting from fixed equipments. As a result, they has a strong anti-damage ability and can satisfy the needs of many kinds of situations, which can make it has a broad application prospect.In wireless ad hoc network, the media access control(MAC) protocol controls the nodes how to obtain the channel, so it has a decisive impact on the performance of the network. In MAC layer, the protocols with time division mechanism have advantage in conflict-free transmission and controlling transmission delay. But how to ensure global time synchronization and how to increase the network throughput are still the two difficult problems in improving the network performance. In the distributed wireless network based on TDM A, the nodes can realize the key functions under the situation of slot synchronization. So it’s significant to study the rate of convergence. When appearing abnormal nodes, how can we find them and improving the anti-interference ability are also difficult problems. Besides, how can we increase the network throughput is another key problem in MAC protocol.Focusing on the characters of Ad Hoc, we researched the channel resource distribution and time synchronization on the MAC layer deeply. With the help of method of Complex Network, we take into account of the network topology impact on synchronization speed of network, also proposing synchroniziton abnormal-detection algorithm base on statistical correlation. Besides, in this paper we studied the througnout performance of multi-channel access protocol.The main works of this paper are summaried as follows:(1) The research on the complex networks found that the topological structure of network has a significant impact on network synchronization performance. Drawing on the experience of research methods of complex networks, analying the relationship between Ad hoc topological structure and synchronization performance, we had hign spot reviews on studying the effect of Betweenness centrality and network structure entropy in representing the synchronization speed.(2) If adopting the distributed synchronization algorithm in Ad Hoc, node’s next time state will be depend on its neighbours and itself, without obtaining the information of the whole network. In this case, the appearance of an exceptional node with chaotic or wrong state will lead to out of conbergence of the whole system. Considering this problem, we proposed a new algorithm based on the statistics correlation analysis to detect the exceptional node. This algorithm searches the exceptional node by computing the correlation coefficient between the nodes.The simulation results indicate that among the three parameters of length of buffer、node density and communication radius, the length of buffer is the most important. The algorithm can accuractely find the exceptional node if the length of buffer is long enough, no matter what the other parameters are. When the length of buffer is fixed, the larger the noed density is, the weaker the detection ability will be. And with the increasing of the communication radius, the performance of the algorithm will reduce first and rise finally.(3) According to the problem of the fixed value in collision window in HRMA protocol will reduce the throughput rate when the traffic load is heavy, we introduced the BEB algorithem to the HRMA to improve the performance of the system in high traffic load. Then we studied the influence of different parameters on the throughput performance in the case of multichannel model, and obtaining the theoretical expression.As shown in the simulation results, the throughput of the improved HRMA protocol can avoid reducing and keep the peak. Through the comparison of different value of the parameters, we find that the throughput of the system will be high when using the large average packet length, and the more the nodes the worse the performance will be when the nodes is less than the frequency.