| With the rapid development of wireless networks and wireless mobile devices,the user demand for broadband wireless access(BWA)is increasing.The key mobile applications,i.e.,Mobile online payments,mobile e-commerce,etc.,are everywhere in wireless mobile networks.Wireless mobile channel has the characteristics of open propagation,complex receiving environment and random movement of the user.So,it is necessary to provide QoS guarantees for critical applications and multimedia applications in wireless mobile network.As a kind of broadband wireless access technology,WiMAX network has been widely concerned and favored by the industry.The contradiction between the increasing demand for the users and the scarcity of wireless resources presents a new challenge for the radio resources management;and the contradiction between the high speed access requirements of users and the time-varying characteristics of the wireless channel has very high requirement for providing the QoS guarantee algorithm.Therefore,it is an urgent problem to ensure the user's QoS requirement in mobile WiMAX network.The research is started from the cross-layer QoS support framework.After the verification and analysis of the mobile WiMAX network service flow with long-range dependence(LRD)characteristics,a self-siimilar call admission control algorithm is proposed;In order to exploit multi-user diversity gain of wireless channel,A channel aware scheduling algorithm is presented for real-time and non-real time traffic flow.Meanwhile,the stability of scheduling algorithms is analyzed.All in all,the effectiveness of key algorithm which ensures users' multi-service QoS is systemly solved in the mobile WiMAX network.The main results of this paper are in the following areas:(1)To study the characteristics of the traffic flow in mobile WiMAX system.First of all.after the methods which are used for analyzing the traffic flow characteristics are compared.that the wavelet estimation method has better accuracy is found.Through experiment and theoretical derivation,the wavelet vanishing moment is repeatedly choosed,and the range of scale parameter is optimized to decrease the effect of these factors on the results of test.Secondly,the ON/OFF model is used to simulate the five different types of traffic flow in mobile WiMAX network,and the wavelet estimation method is adopted to analyze the trace file generated by the model.The experimental results show that the LRD of the real-time traffic is small,but the LRD of non real time traffic flow is relatively large.With the increase of network load,the two LRDs will change slightly;the detection value of the LRD of aggregate traffic flow is basically consistent with the theoretical value.Modeling traffic flow with LRD can more accurately describe the essential characteristics of the traffic flow than the traditional modeling method.The LRD of traffic flow plays an important role in the design of the network call admission control algorithm,ensuring the quality of service(QoS),and improving the securety and performance of the network.(2)To design cross layer solution to ensure QoS.In order to enhance the effectiveness of the QoS support algorithm and exploit multi-user diversity gain,a cross layer QoS support scheduling scheme is proposed,and a call admission control algorithm and opportunity scheduling algorithm are designed accordingly.Because of the LRD of the mobile WiMAX traffic flow,a self-similar call admission control algorithm is proposed to control the number of connections in the system.To guarantee the delay requirements of the real-time traffic,uplink scheduling algorithm LWDF with prediction is applied;at the other hand,to meet the fairness of non-real-time traffic flow,the uplink scheduling algorithm PF with prediction is used.The simulation results show that the scheme can dynamicly adjust scheduling algorithm with the time-varying characteristic of wireless channel.So,the multi-user diversity gain of the wireless channel is taken advantage of to improve system throughput and reduce the call blocking rate.(3)To design Call admission control algorithm for mobile WiMAX system.Based on the LRD of traffic flow of the mobile WiMAX system,the M/Pareto model is used to model the network traffic.As for the statistical properties of the M/Pareto model and FBM model of traffic flow,the mapping relation between the parameters of the two models is established.The formula to calculate the effective bandwidth with the parameters of M/Pareto is derived,so that a self-similar call admission control algorithm(SS-CAC)is presented.The SS-CAC algorithm avoids using measuring method to obtain the model parameter values of the FBM traffic flow,and adapt to self-similarity of the traffic of the mobile WiMAX network.Consequently,the effective bandwidth calculation is more accurate and efficient,and the utilization of system resources is improved.The simulation results show that the proposed algorithm can improve the bandwidth utilization and reduce the call blocking rate of the system.(4)To design scheduling algorithm for resource management.In order to obtain multi-user diversity gain,using adaptive modulation coding techology,for the real-time traffic flow of the system,the modified largest weighted delay first scheduling algorithm(M-LWDF)is used to meet the delay requirements;for non-real-time traffic,the proportional fair scheduling algorithm PF is used to meet the users' fairness.To reduce the influence of the instantaneous rate in the scheduling process,the Markov prediction model is used to improve M-LWDF and PF through prediction of the transmission rate for the next time slot,so that the instantaneous rate is smoothed with predictive value.A predictive M-LWDF scheduling algorithm(Pre-LWDF)and PF scheduling algorithm(Pre-PF)are presented and implemented.The dimension of the transformation matrix of the prediction model is reduced by the sub-channel independence and the lumpability of the Markov chain.The simulation results show that the scheduling scheme guarantees the real-time traffic quality of service,improves the system throughput and fainess,and reduces the oscillation of the system;for the non-real-time traffic,the problem of "starvation" is sloved.(5)To analyze and prove that the scheduling algorithms Pre-LWDF and Pre-PF are stable.The packet level stability of the Pre-LWDF scheduling algorithm is proved by Lyapunov Drift stability theory.Markov chain is used to model a wireless time-varying channel,thus the M-WiMAX network in PMP mode is modeled as a M/G/1 queueing system.Based on the queue theory,the channel capacity and the stability region of the mobile WiMAX network are derived,and the Lyapunov Drift stability theory is used to analyze the packet level stability of the Pre-LWDF scheduling algorithm.Further in the NS2 simulation platform a mobile WiMAX network simulation environment is built to verify the packet level stability of the Pre-LWDF scheduling algorithm.At the same time,it also verifies the effectiveness of the call admission control algorithm and scheduling algorithm.On the other hand,as scheduling algorithm PF which has the flow level stability,the fluid limit is used to prove that Pre-PF aalgorithm has also the flow level stability.(6)To implement a cross layer guaranteed QoS framework by simulation.To verify and analyze the cross layer QoS framework,the call admission control algorithm and the scheduling algorithm which this paper introduces,the QoS mobile WiMAX module implants into the NIST mobile WiMAX simulation platform.In accordance with the cross layer guaranteed QoS framework,in the simulation system the five traffic flows with LRD are generated,self-similar call admission control algorithm and channel aware prediction scheduling algorithm is implemented.The simulation experiments indicate that cross layer guaranteed QoS framework and the related algorithms can be applied in the WiMAX simulation platform,improve the system throughput and delay performance,and has some practical value and research value. |
PDF Full Text Request