Distributed Opportunistic Channel Access In Wireless Networks

With the rapid development of wireless communications and mobile Internet in nowadays, high speed data transmission has became one important direction of communications industry in the future. However, the available spectrum resource is so rare that the research of high speed data transmission faces huge challenge. Under this circumstance, one technique called opportunistic channel access appears, which can improve the spectrum resource utilization efficiently. Opportunistic channel access is one cross-layer channel access protocol which considers PHY layer and MAC layer jointly. During the channel access decision, system decides whether or not to access by the quality of the channel. In centralized network, central controller can acquire global state information, thus opportunistic channel access is especially suitable for it, and the research based on centralized network has been well investigated. However, for distributed network, central controller does not exist, thus the research of opportunistic channel access is still at preliminary stage. Based on the conclusion above, this dissertation decides to focus on opportunistic channel access in distributed network. The main contribution of this dissertation is as follows:For DF relay networks, distributed opportunistic channel access with continuous transmission rate is investigated. Here continuous transmission rate refers maximal channel rate that can be achieved. With the instruction of optimal stopping theory, the optimization objective is to maximize system average throughout. At first, opportunistic channel access strategy is proposed when two hops channel information in DF relay networks are acquired completely. Since the channel gain of DF relay networks is determined by the minimal one between the two hops, channel access strategy which only requires the first hop channel gain information is derived. Comparing strategy under complete channel information, this strategy can save channel probing time, thus, system performance can be improved. To utilizing time diversity more efficiently, relay waiting strategy is proposed, under which the transmission is finished within multiple channel coherence time. Considering the character of relay waiting strategy and channel access with incomplete channel information, a combined two-level channel access strategy which considers the advantages of the two strategies jointly is investigated. All of the four strategies above are pure thresholds strategies, thus can be implemented easily in the real applications, the thresholds are derived by iterative algorithms with fast convergence and low computation complexity. Simulation results analysis the performance gains among the strategies, and demonstrate the validity of the proposed strategies.Under fixed and non-fixed relay scenarios, distributed opportunistic channel access in DF relay networks with discrete transmission rate is investigated. Here discrete transmission rate means the specific discrete transmission rates are adopted. Owing to the feature of discrete transmission rate, the derivation of optimal channel access rules are different from that with continuous transmission rate. Furthermore, the research of opportunistic channel access with discrete transmission rate under non-fixed relay can provide some inspiration for the investigation with continuous case. For DF fixed relay networks, opportunistic channel access strategies under complete and incomplete information are proposed firstly. Subsequently, reinforced relay waiting strategy which can improve system performance further is investigated. Finally, reinforced and combined two level channel access strategy which considers channel access with incomplete information and reinforced relay waiting strategy jointly is proposed. For DF non-fixed relay networks, channel access strategy with complete or incomplete channel information, and reinforced relay waiting strategy are investigated. All of the proposed strategies are pure thresholds strategies with low computation complexity. Simulation results show the performance gains of all of the proposed strategies, and validate the reliability of the proposed algorithms.After the introduction of queue state information, distributed opportunistic channel access in Ad-hoc networks with queue stable constraints is investigated. First of all, channel access strategy under sufficient data is proposed. Subsequently, channel access strategy under real transmission data is studied. For guaranteeing the stability of queue system further, channel access strategy with Lyapunov algorithm is derived, during which, channel access considers both channel and queue information is formulated as a constrained optimization model. Using Lyapunov optimization theory, the constrained optimization can be reformulated as a unconstrained problem, and be solved by optimal stopping theory. Finally, channel access strategy with Lyapunov algorithm under real transmission data is proposed by introducing real transmission data to objective function, system average throughput and queue stability can be improved further. All of the strategies above follow pure thresholds rule. The effectiveness of the proposed strategies are validated by simulation results.