Research On Wireless Network Communication Based On Cross-layer Design

Cross-layer design is a joint optimization of crossing all the layers or severallayers of the protocol stack, in the case of a given resource constraints. As a keytheoretical innovation of the next generation wireless communication systems, theidea of cross-layer design is to regard the originally fragmented network layers as aunified whole to design, analysis, optimization and control, taking full advantage ofthe strong correlation between the layers to optimize the whole wireless networkprotocols. With the continuous expansion of the wireless network applications,distributed states, topology changes, resource constrains, high transmission quality,and other problems of the wireless network is becoming increasingly fierce. And theseissues will lead to the occurrence of the phenomenon of the wireless networkcongestion, in the wireless network transmission process. The reason for wirelessnetwork congestion has a very high complexity, making the traditional networkcongestion handling mechanism do not handle the phenomena of the wireless networkcongestion. The propose of cross-layer design, combined with the concept of thewireless network collaborative communication, provides a new direction to solve theproblem of the wireless network congestion.This article, based on the concept of the cross-layer design in wireless networks,combined with the idea of the wireless network collaborative communication,analyzed the wireless network congestion problem, improved the congestion handlingmechanism of the wireless network. In this article, firstly, proposed the congestiondegree dividing mechanism, breaking the traditional congestion handling system,using the method of cross-layer design to solve the problem of diversity of wirelessnetwork congestion reasons, improved the dynamic adaptive ability of the wirelessnetwork. Then, specific to the mechanism of the wireless network congestion degreedivision, according to the principles of the collaborative communication, proposed theconcept of directed cooperative path network. Finally, to solve the blindness ofcross-layer node collaboration, proposed a model of average throughput forecastingbased on multivariate time series. Through the combination of cross-layer design and node collaborative communication, effectively improved the congestion handlingmechanism of the wireless network.In this paper, specific to the complex and divers characteristics of the congestionreasons of the wireless networks, proposed a congestion handling mechanism tohandle the wireless network congestion. Through the analysis of the broadcast natureof wireless links, link interference, the dynamic changes of the network, terminalmobility, high bit error during the transmission, and other reasons causing congestion,which are existing in wireless networks, demonstrates the necessity of the congestiondivision mechanism. Through the cross-layer design method, monitoring the nodeinformation of every layer, formulated the mechanism of the congestion division ofthe wireless network nodes. The proposal the mechanism of congestion degreedivision makes wireless network congestion handling targeted, and improves thedynamic adaptive mechanism of the wireless network.On the basis of the idea of congestion degree division, this article proposed theconcept of directed cooperative path network based on cross-layer collaborationcommunication. During the process of forming directed cooperative path network,through the method of cross-layer design, monitor the parameters of every layer of thecongested nodes, calculate load pressure of the nodes and the links, and the channelinterference level, then, select communication nodes according to the congestiondegree of the wireless network nodes. According to the establishment of the directedcooperative path network, effectively conduct load shunt and load pressure protectionfor the congested nodes, to avoid the occurrence of catastrophic congestion. To furtherenhance the fairness of packet send competition and channel competition of the node,based on the game theory and on the basis of the idea of fairness, this article proposedthe mechanism of QRD and the mechanism of QPCG channel competition, reducingthe virtual collision phenomenon when the queues send packets, reducing the delaytime of packet sending and channel competing. Through desalinating the priorityweights during frequent collisions, achieve an idea if non-fully collaborativecommunication. Through the application of game idea, improve the fairness ofwireless network competition, achieving the purpose of wireless network resourceoptimization.Due to the time-varying characteristic of the wireless networks, this articleapplied time series to forecast the average throughput of the wireless network nodes, effectively provided a supplement of congestion degree division idea, conducting thenode to make a pretreatment before congested. Through the application of the ARIMAmodel, singly predict the average throughput of the nodes within short time, althoughit is with a certain accuracy, but for the complexity of the impact reasons of thetransmission, it is too far single, and impacts the prediction effect. In order toeffectively predict the average throughput of the node, this article also proposed aaverage throughput prediction model based on multivariate time series, through themethod of cross-layer design, monitor the changes of average throughput and packettransmit queue length during a period of time, then use the multivariate time series toforecast, providing prediction data for the node collaborative communication and theinter-node congestion handling.In this paper, the methods and models, proposed above, through mathematicalmodeling and simulation, in a certain wireless state, improve the resource sharingdegree, strengthen fare competition, optimize collaborative communication methods,and enhance the dynamically adaptive capability of the wireless network.