以多頻段為基礎之合作式感知無線頻譜偵測與動態存取

Abstract

本計劃旨在探討分散式偵測理論與賽局理論在感知無線網路中對於多重頻段之偵測與配置問題。我們考慮一個實際且一般化之多重主要使用者與多重感知次要使用者並存的網路架構。在此環境下，所欲深入研究的問題可分為兩大層面，其一為實體層中之分散合作式多頻段頻譜感測課題，計畫以分散式偵測理論和最佳化理論深入探究偵測效能最佳化與衍生干擾最小化之系統設計與分析；另一為媒體存取層中對於多頻段之資源配置課題，計劃以貝氏賽局和聯盟賽局理論為基礎作一深入探究。 感知無線電技術為近年來蓬勃發展之研究領域。主要原因在於其能有效解決頻譜擁擠與效率低落之問題。然而由於感知無線電系統本質上需隨時感知的特點，再加上距離、遮蔽、干擾、和無線通道的時變性質的影響，感知無線電系統仍存在著許多技術上的挑戰尚待解決。本計劃的目的即在嘗試解決有關其中頻譜偵測與資源配置之問題。計劃分為三年執行，分述如下。第一年著重於實體層合作式多重頻段頻譜偵測之設計與分析。第二年與第三年則聚焦於媒體存取層內之多頻段資源分配問題，並嘗試探究跨層級設計之理論增益。細分之，第二年的研究課題在以貝氏賽局為基礎之頻譜拍賣與配置，分為靜態賽局和動態賽局來討論。第三年主要想要探究的內容為在考慮主要使用者回收頻段的情形下之頻譜拍賣與配置。此外，在考慮通道效應之跨層級設計，我們也會再第三年的計劃裡整合第一年和第二年之研究成果作整體跨層級之設計與分析。

In this project, we aim to thoroughly investigate the problems of dynamic multiband spectrum sensing and spectrum allocation in wireless cognitive radio networks, based on the decentralized detection theory and game theory. We consider a practical and general environment in which multiple primary users and multiple cognitive secondary users coexist and intend to transmit signals simultaneously. Under this framework, the intended research problems can be divided into two categories: (1) distributed cooperative multiband spectrum sensing problems in the PHY layer and (2) auction based multiband spectrum allocation problems in the MAC layer. In the first part, we plan to develop a cooperative sensing protocol that achieves minimum probability of missed detection and minimizes the interference possibly caused to the primary network. And, in the MAC layer issue, we plan to delve in the role of Baysian games and coalitional games to the complicated and dynamic resource allocation problems. Cognitive radio technology is a thriving research field in recent years, due to its effectiveness in dealing with the issue of spectrum inefficiency. However, due to the fundamental properties of cognitive radios and the underlying uncertainties of the channel environment, there remain a huge amount of technical challenges to be solved. In this project, we attempt to tackle some of the problems, particularly in the multiband spectrum sensing and dynamic spectrum allocation issues. We consider dealing with these issues in 3 years. In the frist year, we focus on identifying and detecting the multiple spectrum holes possibly existing in the cognitve network. In the 2nd and 3rd year, our focuses are laid on the issues in the MAC layer, consisted of auction based allocations of multiple spectrum holes under the framework of Bayesian games and coalitional games. Finally, in the 3rd year, we will integrate the research outcomes from the first two years and put them in a cross-layer perspective attempting to pave the way for possible future research issues.