【正文】 技術(shù)簡(jiǎn)介及其應(yīng)用背景．廣東通信技術(shù)．20091月．[2] 曲嘉杰，龍紫薇．北京：中國(guó)移動(dòng)通信研究院北京 100053．[3] 吳承承，白煒，桑林．郵電設(shè)計(jì)技術(shù)．20086月．[4] 白煒，張欣，楊大成．北京郵電大學(xué)無線理論與技術(shù)中心，2008122．[5] 我國(guó)研究與發(fā)展3GPP LTE的建議．．[6] 李高廣，呂廷杰．LTE發(fā)展現(xiàn)狀和前景分析．移動(dòng)通信．20085月．3235[7] LTE Technical White Paper．MOTOROLA．2007．[8] Carlos U180。 beda Castellanos, Francesco D. Calabrese, Klaus I. Pedersen and Claudio Rosa．Uplink Interference Control in UTRAN LTE Based on the Overload Indicator．IEEE，9781424417223/08．[9] Malek Boussif , Nestor Quintero , Francesco D. Calabrese , Claudio Rosa , and Jeroen Wigard．Interference Based Power Control Performance in LTE Uplink．IEEE，9781424424894/08[10] Arne Simonsson and Anders Furuskar．Uplink Power Control in LTE –Overview and Performance．IEEE，9781424417223/08[11] 沈嘉，索士強(qiáng)，全海洋，趙訓(xùn)威，胡海靜，姜怡華等．3GPP長(zhǎng)期演進(jìn)(LTE)技術(shù)原理與系統(tǒng)設(shè)計(jì)．人民郵電出版社，2008．191201論文附件 英文原文：Cognitive OFDM network sensing:a free probability approachAbstract—In this paper, a practical power detection scheme for OFDM terminals, based on recent free probability tools, is proposed. The objective is for the receiving terminal to determine the transmission power and the number of the surrounding basestations in the network. However, the system dimensions of the network model turn energy detection into an underdetermined problem. The focus of this paper is then twofold: (i) discuss the maximum amount of information that an OFDM terminal cangather from the surrounding base stations in the network, (ii) propose a practical solution for blind cell detection using the free deconvolution tool. The efficiency of this solution is measured through simulations, which show better performance than the classical power detection methods.I. INTRODUCTION The ever increasing demand of high data rate has pushed system designers to exploit the wireless channel medium to the smallest granularity. In this respect, the orthogonal frequency division multiplexing (OFDM) modulation has been chosen asthe next mon standard in most wireless munication systems, . WiMax [5], 3GPPLTE [4]. OFDM converts a frequency selective fading channel into a set of flat fading channels [20], therefore providing a high flexibility in terms of power and rate allocation. Future wireless networks therefore tend to be based on highly loaded OFDM , in multiple cell environments, intercell interferenceis still the bottleneck factor which considerably reduces the networkwide capacity. Cooperation between base stationsare envisioned to reach the capacity performance of the socalledbroadcast channel [13], but many problems (essentially of power allocation and user scheduling) prevent those solutions to appear soon in practical standards. Therefore,it is essential for mobile terminals to be able to determine which neighboring cell provides the best quality of service,so that the terminal quickly hands over this best performance base station. Classically, only scarce and narrowband pilot sequences allow the terminals to estimate the transmission power of the main surrounding base stations, . in 3GPPLTE,two sequences of the MHz band are available every 5 ms. Those synchronization sequences are usually affected by fast channel fading and overlap data from other base stations。as a consequence numerous occurrences of those pilots need be accumulated to achieve a satisfying estimation of the base stations transmission power. The classical alternative to the pilotaided (also referred to as dataaided) power detection is to perform a blind estimation from the ining interfering signals. This raises the fundamental cognitive radio question [23], [12], which will be an important topic of the present work: “how much information can a cognitive receiver recover from the ining signals?”. The response to this question answers two classical concerns of engineers and system designers: (i) is the additional information brought by blind detection worth the putational effort?, (ii) is some given blind detector solution far from providing all the accessible information?. It is clear in particular, from an information theoretic viewpoint, that the information received on the N OFDM subcarriers must ideally not be filtered in order to provide as much information as possible on the problem at hand, . any filtering process diminishes the available information in the Shannon’s sense[1]. Therefore, if as many as L consecutive OFDM symbols are received, the available information is contained in thereceived N L matrix Y, with N typically large. As a consequence, since L cannot be taken infinitely large, N/L is non trivial. This leads to the study of large random matrices problems, which is currently a hot topic in the wireless munication munity [8]. This is in sharp contrast with classical power detection methods [10], [11] which are only asymptotically unbiased, . these methods assume that one of the system dimensions is large with respect to the others and this condition is necessary to ensure the convergence of the underlying algorithms. Our purpose is to retrieve relevant information on the base station transmission powers. It will be shown hereafter that, depending on the a priori knowledge of the receiver, the essential part of the power information is, in most practical situations, contained in the eigenvalue distribution of the matrixThis naturally leads to the consideration of recent research on random matrix theory (RMT) [8] and more specifically on free deconvolution [16]. In particular, in [7], a similar study of terminal power detection in code division multiple access (CDMA) networks is derived from these tools. However, the model in [7] only considers flat fading channels and dodges the difficulty of multipath channels。 moreover the structure of the CDMA encoding matrix allows to easily recover the transmitted signal variances, which is not the case of multicell OFDM in which multiple streams overlap withno dedicated code to separate them. We propose here first to discuss