The Modeling Of Link Connectivity Inmobile Ad Hoc Network Based On Simulation Experiment

Mobile Ad hoc network( MANET) is a kind of self-organized wireless communication network with non fixed access point, flexible networking, and broad application prospects. The communication path between nodes in mobile Ad hoc network is constituted by a series of wireless links. The performance of communication links will affect the communication quality of network directly. And link connectivity is the most basic characteristics of the link. The mobility of node makes link connectivity changes frequently, and the topology of network and communication routing changes with it. Therefore, the researches of link connectivity research shows great significance for network topology control, routing protocol analysis and so on.Link connectivity is mainly controlled by wireless channel propagation characteristics and mobility models. According to propagation loss character of the radio electromagnetic wave in air, three kinds of fading model of signal power is derived: path loss, shadow fading and multipath fading. The model of node transmission range is established based on the features of signal power fading. Simulation experiments are performed to analyze the impacts of signal loss characteristics on the effective transmission range of the node. Also, contrast and analysis is performed for three kinds of node movement model: Random Walk mobility model, Gauss Markov mobility model and Semi-Markov Smooth mobility model. And Semi-Markov mobility model, which is more close to the actual movement rule, is selected as node movement model for the follow-up study. Baseed on this, adjacency matrix in graph theory and link connectivity probability vector is used to express link connectivity. And combined with the Markov chain theory, the time-varying first-order Markov link connectivity model is established. In order to improve the accuracy of the model, the first-order Markov link connectivity model is extended to high order. Link connection state at different time is obtained through simulation, and a statistical link connectivity model parameter, transition probability matrix, is obtained using the statistical method. One of characteristic parameter of the wireless network, the link life time, is evaluated by comparing the results of Monte Carlo simulation and multi state first-order Markov link connectivity model experiment. And the relationship between link life time error and Markov chain order is studied.Following conclusions are reached according the simulation experiment. First, high-order Markov model of link connectivity can effectively describe the time variation of the link connectivity. In addition, the accuracy of link life time improves with the increase of Markov chain order. The accuracy of link life time generated by Markov chain whose order is higher than 4 improves inconspicuously. Finally, compared to the first-order Markov link connectivity model, link life time error of the fourth-order Markov model is declined by 68%.