Research On Radio Resource Management In Cognitive Radio With Fiber-Connected Distributed Antennas

With the rapid development of wireless communication, there are no adequate high quality spectrum bands. However, the utilization efficiency of the authorized spectrum bands is not satisfied. Thus cognitive radio (CR) is proposed to cope with this problem. In CR, in order to enhance the utilization efficiency of the authorized spectrum bands, secondary user (SU) could share the spectrum band with primary user (PU) under the constraint that SU can not disturb the normal transmission of PU. Thus, Radio Resource Management takes a considerable part in CR, not only because it needs to maintain the normal operation of PU, but also because it need optimize the performance of SU. Accordingly, study on radio resource management in CR has a significant meaning.In conventional CR, SU might not satisfy its demand of communication, as it has to protect the normal operation of PU. In order to solve this problem, this paper establishes CR with fiber-connected distributed antennas. In this novel architecture, massively distributed antennas cover the service area densely, and connect with secondary access point via fiber. The signal sent by SU is received by distributed antenna and transmit to the secondary access point with radio over fiber technology. Since the proposed system has a sophisticated architecture, the radio resource management in traditional CR might not be suitable for it. Therefore, in this paper, we discuss the radio resource management in CR with fiber-connected distributed antennas. The main work and contributions are as follows.1. We propose a multi-channel cooperative sensing scheme in CR. This scheme not only takes different sensing values of different channels into account, such as bandwidth, probability of being idle and frequency selective fading, but also considers that SUs might have heterogeneous sensing abilities, such as sensing accuracy. This cooperative sensing problem is formulated as a weapon target assignment (WTA) problem, which preferentially assigns SUs with more powerful sensing ability to more valuable channels. And then, the WTA problem is transformed to be an integer generalized network flow problem with convex flow costs in order to obtain the lower bound solution, and then solved by the branch and bound algorithm. Simulation results illustrate that the proposed scheme is superior to the existing scheme in the quantity of sensed idle channels and potential throughput supported by these idle channels.2. We propose a spectrum access scheme in CR with fiber-connected distributed antennas. This scheme allows multiple SUs to access one PU spectrum band via multi-antennas technology, where the link estimation from SU to PU is supposed to be inexact. This spectrum access problem is modeled as a stochastic knapsack problem with random weight and then relaxed to be a deterministic second order cone programming problem. With the deduced upper bound, this problem is solved by the branch and bound algorithm. Simulation results illustrate that the proposed scheme not only maintains the interference overflow probability below the given threshold, but also achieve more total throughput of accessed SUs, compared with the existing scheme.3. We propose subcarrier allocation schemes and power control schemes in CR with fiber connected distributed antennas. Firstly, considering different features of flat fading channel and frequency selective fading channel, we respectively propose the efficient subcarrier allocation (ESA) scheme to maximize the total throughput of SUs, and the efficient and fair subcarrier allocation (EFSA) scheme to achieve the tradeoff between efficiency and fairness. Moreover, based on the EFSA scheme, we propose the efficient power control (EPC) scheme, fair power control (FPC) scheme and efficient and fair power control (EFPC) scheme. Furthermore, it is proved that power control scheme hardly affects the total throughput of SUs when SINR in SU receiver is relatively high. Since different power control schemes have similar efficiency, FPC scheme is the best one. Especially, the CR with fiber-connected distributed antennas might satisfy the condition of high SINR easily, because of multi-antennas technology and maximal ratio combining technology. Finally, according to the simulation results, the following conclusions can be observed. Compared with the existing subcarrier allocation scheme, the ESA scheme can achieve higher throughput, but the fairness is not improved. The EFSA scheme compromise total throughput and fairness that are both superior to the existing scheme. Additionally, when the SU receiver has low SINR, EFPC scheme is the ideal one, while the FPC scheme is the best choice when the SINR is high.4. We propose an optimal energy-efficient access point (AP) selection scheme with indexability property in Cognitive WLAN over fiber. This scheme not only considers the influence of automatic rate fallback (ARF) over distributed coordination function (DCF), but also takes states of AP into account, such as the quantity of STAs associated with AP, quality of links and ISM interference. In this scheme, each STA selects its appropriate AP in distributed manner, in order to maximize the throughput based on DCF and ARF, and to minimize the energy consumption. The AP selection problem is formulated as a restless bandit problem and solved by the primal-dual index heuristic algorithm based on first order relaxation. The AP selection process is divided into offline computation and online selection, where the main computation is finished in offline stage. In online stage, according to the state of each AP, STA lookups the index table to find out the corresponding index. The APs that have the smallest indices are selected for this STA. Simulation results illustrate that compared with the existing scheme the proposed scheme not only enhances the throughput based on DCF and ARF, but also reduces the energy consumption.