感知雲端網路實現控制頻道之研究

Abstract

在感知無線電網絡（CRN）中，CR用戶可以通過頻譜感知來使用頻帶以增加頻帶利用率，但同時要避免干擾授權用戶的使用。因此關鍵是精確的頻譜感測與有效地避免干擾。在我們的設計中，雲端會維護頻譜資料庫，以供CR AP查詢頻帶的可用性。此外，CR AP會負責協調CR用戶的資料傳輸。為了實現這兩個目標，所考慮的是CRN必須有一個控制頻道以供CR AP與所管理的CR用戶來交換所需的控制訊息。但是TVB的可用性會隨時間動態地改變，因此在CRN中使用TVB來當作控制頻帶是不可行的。由於上述原因，我們提出了一種新的設計，透過IEEE 802.11 PCF模式來達成CR控制通道的實現。在PCF期間，CR AP可以收集感測裝置的感測資料，收集CR用戶的資源配置請求，以及通知資源分配結果。通過正確的設置DCF和PCF模式的週期，我們可以即時的更新頻譜資料庫，增加TVB的頻道利用率。然而，這種設計的主要挑戰是如何保證PCF的週期性。在這篇論文中，我們提出了兩種機制來解決這一設計問題。另一方面，雖然更長的PCF週期時間會顯著改善頻譜資料庫的準確性與TVB利用率，但也減少DCF的週期時間並降低傳統WiFi用戶的效能。為了研究這個問題的權衡，我們探討此異質網絡，並分別推導出用於收集感測回報與資源請求的適當時間。此外，探討在給定WiFi用戶數和效能降低門檻，我們可以服務多少CR用戶。並且通過簡單的TVB頻譜配置演算法和考慮頻譜估測準確度，來探討TVB系統的效能。

In cognitive radio networks (CRN), CR users can use the available channels through spectrum sensing to increase the bandwidth utilization while avoiding the interference to licensed users at the same time. Accurate spectrum sensing is critical to effectively avoid interference. In our design, there is a spectrum database maintained in the Cloud for CR Access Points (AP) to query channel availability. Besides, CR APs are responsible to coordinate data transmissions of CR users. To achieve these two goals, the considered CRN must have a control channel for the CR AP and managed CR users to exchange required control messages. The availability of TV Band channels may dynamically change, it is not feasible to utilize a TVB channel as the control channel of the CRN. Due to the above described reason, we propose a novel design of CR control channel implementation by enabling the PCF mode of IEEE 802.11. In a PCF period, a CR AP is able to collect sensing reports sent by sensing devices, collect resource requests sent by CR users, as well as announcing the resource allocation results. Through proper setting the cycle of DCF and PCF modes, we can update the spectrum database in time and increase the TVB channel utilization. However, the major challenge of this design is how to guarantee the periodicity of PCF cycle. In this thesis, we propose two mechanisms to tackle this design issue. On the other hand, though a longer PCF cycle time will significantly improve the accuracy of the spectrum database and the TVB utilization, it also decreases the DCF cycle time, as well as the throughput performance of legacy WiFi users. To investigate this tradeoff, we model this heterogeneous network, derive the proper time slots for sensing report collection and resource requests, respectively. Besides, we investigate that, upon given he number of active WiFi users and throughput degradation threshold, how many CR users can be served. Furthermore, by integrating a simple TVB channel scheduling algorithm and taking the sensing accuracy into consideration, we study the throughput performance of the TVB system.