A Research On Access Mechanism Of Cognitive Networks Based On Cross-layer Awareness

The rapid increasment of wireless standards and rich network applications have expanded greatly the applications of wireless networks.New modes of access are popping up,such as WLAN,WiMax,UMTS and Bluetooth,which compose the unified and complex heterogeneous wireless network environment.Although networking mode with multiple access methods coexisting can effectively make users have the ability to access networks at anytime and anywhere,the contradictions between the relatively fixed access mode and the users' flexible and changeable demands for wireless network services protrude out.In the current complex heterogeneous wireless network environment,how to take the factors of network characteristics,user business requirements and terminal states into consideration,and the best access network for users which works steadily,and ensure the efficient allocation of network resources and the overall performance of the network,has become one of the most pressing issues.Accordingly,this paper aims at the access selection mechanism of wireless networks,which bases on cross-layer awareness for wireless network,in the complex heterogeneous cognitive network environment.During the research,it takes cognitive computing as basic,introduces a variety of theories such as multi-scale entropy,fuzzy logic,game thery,Lyapunov theory and so on,and combines existing research results,and finally realizes the goal to provide effective and reliable access selection and cross-layer resource optimization method with analysis of characteristics of user business under the cross-layer awareness framework model.The research can provide references for the promotion and improvement of cognitive radio networks technologies.The main research contributions of this dissertation are listted as follows:In view of the dynamics and complexity of the network environment,a cross-layer awareness framework model of cognitive network based on MDE-K is presented to solve the problem of the dynamic adaptation between network and environment.This makes the network can work with integrity,hierarchy and adaptability,and provides the users with high capacity,high speed,and reliable end-to-end services and efficient and intelligent network management.PEPA models is used for formal representation then complete the validation of the model,experimental results show that the model has good stability,to enhance the efficiency of cross-layer model.Under the guidance of this cross-layer model,expand the cognitive selection mechanism study on three methods of network access.A mobile terminal business flow analysis method of cognitive network based on multi-scale entropy is promoted to solve the problem that terminal business characteristics are difficult to analysis directly with the time-variant of mobile terminal and diversity of communication business.The method consists of business flow model with time and space scales,traffic classification with k-nearest neighbors,and multi-regression analysis.The method of mobile terminal business flow classification and the ability to achieve on-line prediction of network traffic,cognitive network access mechanisms for access to information provides the research basis.An access selection method based on fuzzy comprehensive evaluation is designed to solve the following problem existing in the traditional access selection method which cannot meet the future demands of network,merely rely on link capacity and transmission power,or some other single network performance parameters.In addition,the paper adopts the Quantum Genetic Algorithm(QGA)to optimize not only the performance factors of access network but also the evaluation of weights of the main parameters.The method is better able to meet the needs of users,and to protect the network's overall performance.Comparing with the traditional methods of performance parameters,fully demonstrates the superiority of the proposed algorithm compared with similar access selection algorithms with the representative,also has certain advantages.For cognitive users the uncertainties of data arrivals,channel occupancies and so on,a distributed transmission control method based on stochastic network optimization is proposed,which can ensure the ability to optimize the transmission processes effectively.Using the method of game theory,according to maximum transfer rate based on user needs,authorized and unauthorized channel is selected.If select unauthorized channels,you can directly transfer,and if select authorized channel,you will need to build support for the primary user crash-tolerant system model.First,a system model,allowing to set up different collision tolerable levels depending on data types of the terminal users,is established,which could be beneficial for improving system throughput.Then,a transmission control algorithm based on cross-layer optimization is proposed.Using stochastic network optimization theory,we have designed a cross-layer scheme that integrates access rates control and channel-resources allocation,and reaches the approximately optimal throughput by a constant approximate way.