Performance Analysis And Quality-of-Service Guarantee Techniques Of Wireless Networks

With the development of the wireless network and the increasing communication requirement of users, wireless networks need to support more and more users with various applications which have stringent requirement on the network. Under the limit frequency resource, how to guarantee the QoS requirement of the increasing demanding wireless applications is a big challenge encountered by the wireless networks. The basis of QoS guarantee is fully understanding how the network parameters and the adopted policies affect the QoS performance of the network. Then some efficient QoS guarantee algorithms can be designed. Therefore, the performance analysis is an important research topic of wireless networks.The signal propagation characteristics give rise to the unreliable transmission and time-varying capacity of wireless channels. Hence, the service rate of the wireless channel is time-varying. Meanwhile, the limited frequency resource and the various resource sharing ways make the service processes of users in wireless networks more complex. Thus, the service rate of the wireless user is difficult to be modeled and calculated. Furthermore, the types of applications in wireless networks are increasing. The merging and mixing of these applications induce the self-similarity of the traffic arrival process. Due to these properties of the traffic arrival process and the complexity of the service process, it is difficult to evaluate the QoS performance of wireless networks, and traditional methods cannot work.To address the shortage of the traditional method, we propose an analytical method to study the performance of the wireless network by combining the probability theory and the network calculus. Based on analytical results, we design algorithms to guarantee the QoS requirement of the traffic in different scenarios. The main contents of this dissertation are summarized as follows.1. We propose a method to analyze the performance of wireless networks based on the residual effective capacity (EC) function. For the deficiency of the available analysis method, which cannot model the properties of the traffic arrival process and service process simultaneously, we propose a theoretical framework named as the residual EC analytical model (RECAM) to analyze the performance of wireless networks. This framework is based on the effective bandwidth (EB) theory and the EC theory. According to this framework, the calculation of the performance is turned into calculating the EB function of the traffic arrival process and the EC function of the service process. Through proving the linear property of the EC function, the concept of the residual EC function is proposed, and the RECAM is proved to be feasible. To describe the property of the resource sharing way, the multiple access protocol is modeled by a two-tuple. The two-tuple makes the calculation of the EC function easier, which connects the queue performance of user and its complex service process. By using the proposed model, we calculate the delay performance of the system which adopts TDMA, FDMA, CSMA/CA, or ALOHA protocol. The comparisons between simulation results and analytical results indicate that the proposed method provides good estimations for delay performance in many different scenarios.2. We propose an analytical method to study the end-to-end delay distribution of a wireless multihop network, and design an admission control algorithm to guarantee the stochastic QoS requirement of the traffic. To model the influence of the hidden terminal problem caused by the signal propagation in multihop networks and the self-similarity of the traffic arrival process on the performance of the access protocol, we propose a revised analytical model for the IEEE 802.11 protocol. By combining the analytical results of the access protocol and the RECAM, we provide a method to calculate the end-to-end delay distribution of the multihop wireless network. By comparing the analytical results with the simulation results, the accuracy of our method is verified. Furthermore, the results under different network settings indicate that the unreliability of the wireless channel and the hidden terminal problem are important factors influencing delay performance. The result ignoring these two factors will overestimate the delay performance of the practical system. Then based on the analytical result, we design a stochastic admission control algorithm (SACA). Comparing with other admission control algorithms, the SACA can not only efficiently guarantee the QoS requirement of the traffic but also improve the efficiency of the network resource.3. We propose an analytical method to derive the performance of the cognitive radio network by using the EC function. Since present analytical methods cannot characterize the influence of the primary user’s behavior on the performance of the cognitive user when the states of the primary user do not have Markov property, we propose an analytical method to derive the performance of the cognitive user. First, to model the states of the primary user with non-Markov property, with the help of the method of supplementary variables, we propose a two-dimensional Markov chain to describe the transitions of the channel states. And then, through solving this Markov chain, we model the service process of the cognitive user as a Markov modulated process. By using the property of the Markov modulated process, we present a method to calculate the EC function of the service process, and obtain the relation between the parameters of the primary user and the EC function of the cognitive user. At last, we substitute the EC function into the RECAM to obtain the queue length, dropping probability and throughput of the cognitive user. The comparisons between the simulation results and the analytical results verify that the proposed method can provide good estimations for the performance of the cognitive user. Furthermore, the analytical results provide some guidelines for QoS guarantee policy, such as admission control and channel selection. From numerical results, it can be found that the algorithm based the analytical results can guarantee the QoS requirement of the traffic.