|
On the basis of the conventional wireless access networks, the wireless Mesh network incorporates some advanced techniques from the mobile AdHoc technology, such as peer nodes relay, dynamic topology discovery, and real time routing calculation, providing a more efficient, reliable and convenient solution for civil wireless networks.In this thesis, we firstly introduce briefly the background and application prospects of the Mesh network. Then we discuss some technologies related to the 802.11 based wireless Mesh network, emphasizing on the study of the mechanisms of topology discovery and maintenance along with the routing algorithms. The description of the Mesh driving programs design follows, in which the whole system are divided into function modules of initialization, data transmission, data reception, topology discovery and maintenance, routing calculation, etc. The design and implementation of each module’s major data structure, realization flow chart and function structure are also given. Finally, a small scale test bed is established using the Mesh routers we developed. Aiming at the functionality characteristics of the wireless Mesh network, we design and execute some test cases for multi-hop transmission among Mesh nodes, data routing and forwarding, as well as the fault tolerance and self healing in the wireless Mesh network. The results show the usability of the Mesh modules we developed.In topology discovery and maintenance, a traffic charge calculation method which is using the radio signal loss and easy to be quantitatively analyzed, and a dynamic transmission power adjusting scheme according to the robustness of a mesh node, i.e., the number of its accessible neighbors, are adopted. According to various traffic types, different routing policies are adopted. For those upstream Internet traffics of which the destination address pointing to the network egress node, a active routing mechanism based on DV algorithm is adopted. For P2P traffics of which the destination address is randomly chosen, a simplified AODV algorithm is used.
|