Research On Heterogeneous Cognitive Network Coexistence Based On Markov Jump System

With the increasing demand of wireless spectrum resource,different unlicensed wireless communication devices and technologies have been applied in the television white space(TVWS).When coexisting in TVWS,the heterogeneous cognitive networks based on different communication standards and technologies can not cooperate with each other due to incompatible MAC/PHY designs of coexisting networks,moreover,mutual interference between networks is inevitable which heavily degrades the quality of service during coexistence between heterogeneous cognitive networks.Thus,it is vital to solve the issues of heterogeneous coexistenceIn this thesis,taking the activity of primary users into consideration,which leads to the available spectrum resources for secondary users stochastically variable,we formulate the heterogeneous coexistence problem over TVWS as a nonlinear Markov jump system based on the Lotka-Volterra(L-V)competition model.By designing the controller of the system,a spectrum allocation mechanism is constructed to ensure that the network resources are assigned to the primary users as expectationFirstly,the stochastic parameters are considered to build a nonlinear Markov jump system model based on L-V competition model of biological system,then a feedback controller is added to achieve the spectrum bandwidth that meets the requirements of the users.The nonlinear system is linearized by Taylor expansion formula at equilibrium point,then by constructing the quadratic Lyapunov function we find a sufficient condition that makes the network resources are assigned as expectation.The gain matrix of controller is obtained by using linear matrix inequality knowledge,which satisfies the diverse fairness requirement of users in the networks.Besides,we proposed an IEEE 802.19.1-compatible spectrum sharing algorithm which enables the nonlinear Markov jump system to converge to the assigned spectrum share between networks.However,the control gain of Markov jump system obtained above is on the condition that the transition probability matrix is completely known.Due to the high dynamic of the radio networks,there are many uncertainties of spectrum state in reality which leads to the transition probabilities between different modes are not easy to obtain.As a result,it is essential to propose the condition of unknown transition probabilities.The unknown transition probabilities are assumed varying between two known bounds in this thesis,and the controller is designed in this situation.We also employ spectrum sharing algorithm which enables the nonlinear Markov jump system to converge to the assigned spectrum share between networks and finally realize heterogeneous cognitive network coexistence.