Strategies for Dense Optical CDMA Communication Systems

资源类型:pdf 资源大小:677.00KB 文档分类:工业技术 上传者:蔡雪峰





【刊名】The Journal of China Universities of Posts and Telecommunications


1 Introduction Recently, considerable interest has been focused on Opti cal Code Division Multiple Access (OCDMA)[1~15], be cause of the advantages of flexible bandwidth management,asynchronous access, and the potential network scalabilityetc. In an OCDMA system, many users can simultaneouslyshare the same channel through the allocation of specific ad dress codes. All users signals are transmitted by the sameoptical fiber channel. At the receiver end, the matched de coder can recover the desired information. The signals fromundesired users are named as Multiple Access Interference(MAI). As a multiplexing technique, OCDMA has been exploredfor future capacity enhancement. In order to satisfy theDense Optical CDMA ( DOCDMA) communications, thecommon approach employed in the pure Direct Sequence(DS) or Frequency Hopping (FH) is to increase the codelength. It is evident that the longer code length is adverse tothe design of encoder/decoder and economy. As for DS CD MA, it applies the signature to the time domain, and needsvery long code sequences with good auto and cross correla tion properties in order to support many simultaneous users.The narrowness of pulse width is inevitably introduced inrepresenting the signature. The narrowness of pulse widthmakes the system to link impairment and may lead to lowspectral efficiency. All DS CDMA approaches have the spec tral efficiency problem, if the code length exceeds ten[1]. Ina FH OCDMA system, code sequences will be mapped inthe frequency domain. It mitigates the stringent requirementson the narrowness of pulse due to the relatively long chip du ration of the signature code. However, some adverse factorssuch as the impact of group velocity, etc. are also intro duced[2]. In this paper, we have formulated a strategy that the lim ited available code sequences in pure system can be reusedto realize DOCDMA communication: the strategy of novelhybrid DS/FH system. In which, a DS spread spectrum sig nature sequence is frequency hopped according to a specificFH patterns again. The case that there are n users employ ing the same FH pattern but different DS code patterns isconsidered. Based on Optical Orthogonal Code (OOC)[3]and the stationary random process theory, the upper boundprobability of error is evaluated on the condition that the im pact of channel noises is neglected. The results show thatthe hybrid system is suitable for DOCDMA communication,thereby, the problem that there are less available code se quences for multi users in the pure system is mitigated.Moreover, the problems such as the link impairment, dis persion of group velocity, etc. in the pure (DS or FH) sys tem can be solved effectively. This paper is organized as follows: Section 2 describes thedense OCDMA system model. The system performance anal ysis is carried out in Section 3. A comparison is given inSection 4, and then a conclusion is given in Section 5.2 The DOCDMA System Model2 1 The DOCDMA Scheme The schematic configuration of hybrid CDMA for densecommunication is shown in Fig.1. In the figure, the data bitis firstly encoded by the DS spreader, and then into the FHhopper to be encoded according to a specific FH patterns a gain. As a result, the signal sequences from the hybrid en coder is pseudorandom optical signal, and is sent in a formof FH hopping patterns with the DS spreading information.In the decoding process, the received signal is firstly de hopped by the matched dehopper, next, through thematched DS spreader the proper signal can be obtained.Fig.1 Block diagram of the hybrid DS/FH for dense communication system It is noted that the total K users in Fig.1 have been di vided into several groups (1, 2, …i…), and each groupuses the same FH but different DS code patterns. There isno the same FH code pattern being reused in differentgroup. However, the same DS code pattern can be reused indifferent groups.2 2 The Theoretical Model Without loss of generality, it is assumed that there are nusers employing different DS code sequences but the sameFH code sequence. For the sake of simplicity, the intensityof the signal is normalized equal to one. The effects of shotand thermal noise on the system performance are neglected,and the performance degradation is only caused by the Multi ple Access Interference (MAI). In our system, the spreadgain of the DS spreader isN, the code weight iswd(1≤wd≤N); the number of available frequencies of the FH hop per is q, and the code weight is wf(1≤wf≤q). Based onthe above scheme and OOC theory[3], the received signalcan be simply expressed asr( t)=∑Kk=1sk( t-τk) (1)sk( t)=bk( t)DCk( t)FCk( t,f) (2)Here,bk( t)= ∑∞l=-∞bk,lPT( t-lT), bk,l∈{0,1}is the lth data bit of user k; and T associates with the bitinformation duration.DCk( t)=∑Nm=1ck,mP( t-mTcD), ck,m∈{0,1}for 1≤m≤N, is the mth chip value of the kth user s codein DS process, and TcD =T/N is the chip duration,DCk( t) serv

1 2 3