Study On Cognition And Online Measurement In Wireless Mesh Network

Wireless Mesh Network, which has the advantage of high bandwidth, high reliability and easy to rapid deployment, is considered to be a promising design paradigm for network convergence. Because of the time-variant characteristics of the wireless node and channel, topology control and dynamic resource assignment are utilized to optimize the network performance. The accurate and real-time measurement of network characteristic is essential to optimize performance in wireless mesh networks. This dissertation is supported by the national 863 project"Hierarchical topology control of the broadband wireless ad hoc network system". The dissertation's major work and innovations are shown as follows:(1)The concepts and overviews of wireless link measurement, spectrum sensing and data processing technologies were introduced, including active measurement, passive indirect measurement, passive direct measurement, cooperative spectrum sensing, data fusion and data aggregation. Then, the advantage and disadvantage were analyzed in the application of wireless network.(2)An online measurement architecture for wireless mesh networks was studied to cope with the measurement data processing problem and the corresponding measurement nodes' location selection problem. By drawing on the advantages of existing data processing methods in the wireless sensor networks and combining the application situation and design aims, a data aggregation mechanism based on double queues was exploited, which synchronized the data aggregation between the measurement nodes and data processing center, ensured the consistency of the data aggregation and reduced the communication overhead in the networks. In the application of the proposed mechanism, a data aggregation algorithm based on prediction was proposed, which guaranteed the accuracy of the measurement and decreased the data loss by combining the Grey model and Kalman filter.(3)In order to obtain the best measurement nodes'location, an ellipse secant algorithm was designed, combining the signal detection capability with the transmission capability. The optimization location of measurement nodes were selected in a series of ellipses, the two focus points of which adopt measured node and data center nodes respectively, with the objective of optimizing system performance measure. The simulation results show that the online measurement architecture achieves the optimization between signal detection capability and transmission capacity while providing the quality guarantees to online measurement and the ellipse secant algorithm gains lower computational complexity than random selection algorithm.