Cross-Layer Optimization In Cognitive And Cooperative Wireless Networks Based On Game Theory

As rapid development of information science and technologies, it is expected to achieve the objective of personal communications, i.e.,5A (anyone, anytime, anywhere, anydevice, anything), by deploying the next generation of broadband wireless mobile communication networks (BWMCNs). However, there are still great challenges both in network architecture and wireless access technologies as described in the following.Communication protocols:It is a common practice to design communication protocols in a layered and separate way in existing wireless networks. Although the layered architecture is simple, it is certainly suboptimal. For example, based on the separate architecture each layer in a protocol stack is usually designed based on the worst-case network condition and does not adapt to the time-varying wireless networks. Moreover, output of a layer is not necessarily a strictly monotonic function of input of the layer. Therefore, optimization within the layer might be suboptimal with respect to the whole network. Cross-layer design (CLD) has been evolved to address these problems by enabling more interactions among different layers in a protocol stack or combining multiple layers that have the same or similar functionality into a single new layer to simplify the whole protocol stack.Radio spectrum resource:As rapid development of wireless communications, radio spectrum is becoming more and more scarce due to several reasons. On one hand, based on the traditional spectrum management, each wireless communication system was assigned with a dedicated spectrum band, and so far almost all available spectrum bands have already been occupied. On the other hand, most of the assigned spectrum bands are considerably underutilized both in space and time dimensions implying a significant waste of precious spectrum resources. To alleviate the shortage in radio spectrum, cognitive radio (CR) was proposed by Joseph Mitola in 1999 to increase spectrum efficiency by allowing secondary users (SUs) to access spectrum bands assigned to the primary users (PUs) without causing notable decrease in quality of service (QoS) of the PUs.Connectivity of wireless links:Compared to existing wireless communication systems, the 4G requires that wireless network should support much higher data rate with a peak of 1 Gbits/s. However, given a single wireless link with certain level of signal-to-noise-plus-interference (SINR), the achieved rate based on present communi-cations techniques is already very close to the limit of Shannon capacity and hence it is very hard to considerably improve the rate any more. Instead, more efforts should be made to increase the SINR of a wireless link such that connectivity of the links and cov-erage of the whole network can be improved. Cooperative communication techniques have been proposed to allow a mobile device equipped with only single transceiver antenna to use other devices to form a virtual multiple-input single-output (V-MISO) wireless link. This significantly alleviates the strict restricting condition on the physical antenna spacing in multi-antenna techniques to achieve the spatial diversity.Based on the above discussion, we study the cross-layer optimization in cognitive and cooperative wireless communication networks, with the objective of developing a theoretical framework for designing new wireless networks, evaluating performance of a newly developed wireless network, and finally designing practical communication protocols and algorithms. Three parts of research are incorporated in the thesis, i.e., the co-existing scheme of multiple wireless sub-systems in a cognitive radio system, cross-layer optimization for cooperative wireless networks, and distributed radio resource management in cognitive and cooperative wireless networks. We use utility function to build a bridge between radio resource at lower protocol layers (e.g., physical and MAC layers) and quality of service (QoS) at upper layers (e.g., application layer) such that different layers can be jointly optimized. Moreover, we design utility functions for users and then based on mathematical analysis about property (e.g., convexity, continuity) of the utility functions we derive practical distributed resource management algorithms. Research in this thesis was supported by the National Natural Science Key Foundation of China "Research on Cognitive and Cooperative Communications" (No.60832008) and National Natural Science Foundation of China "Cross-Layer Design in Wireless Communications" (No.60672036). Main contributions of the thesis are as follows: 1. We propose a novel co-opetition strategy for spectrum management in cognitive radio systems. The co-opetition strategy is designed based on the notion of "KSBS" in cooperative game theory. Central idea of the co-opetition strategy is to let users with high competing ability in a resource allocation stop competing for more resources such that QoS of those users with low competing ability can be improved. Numerical results indicate that the proposed co-opetition strategy can result in more satisfied users while keeping fairness among unsatisfied users. Moreover, through the proposed strategy it becomes easier for different wireless subsystems to coexist in a cognitive radio system. This contribution corresponds to Chapter 2 and paper 1,2,3 and 4 listed at the end of the thesis.2. We study cross layer optimization for multiuser video streaming in cooperative wireless networks. The problem of joint optimization of video encoding rate, transmission power and relay selection is formulated as a mixed integer non-linear non-convex optimization problem. Since the problem is NP-HARD, we propose a globally optimal and centralized algorithm. The algorithm is designed based on a combination of branch and bound framework and convex relaxation of non-convex problems. It is shown that the proposed algorithm always converges to the globally optimal solution and in practice has very low computational complexity. The pro-posed algorithm can be used to provide a performance benchmark for distributed algorithms studying the same or similar problems in this field. This contribution corresponds to Chapter 3 and paper 5 and 7 listed at the end of the thesis.3. We study distributed radio resource management in cognitive and cooperative wire-less networks. The problem of joint spectrum management and relay selection is decomposed into two subproblems, i.e., spectrum management with given relay selection strategy and relay selection with given spectrum profile. We design dis-tributed algorithm for each individual problem based on the notion of game the-ory and variational inequality (VI) theory. It can be proven that the proposed algorithm converges to VI solution of each subproblem, which is also a Nash equi-librium. Through numerical results, it is shown that the proposed algorithm can achieve a performance very close to the globally optimal solution. Algorithms pro- posed in this chapter can be used for practical protocol design in future cognitive and cooperative wireless networks. This contribution corresponds to Chapter 4 and paper 6 and 7 listed at the end of the thesis.