In th?s part of the paper we would look at the rake reciever perfomanceanalysis and discuss it. especially in thischapter we go in details in every stage of our model to show how this modelwork, 1)Modulation and spreading; After we generate a local signal as explained inprevious chapter, we apply the out put to block of modulation and spreading,the result of modulation and spreading is shown Figure Fig.6 System ModelFig.7 System Model Results Infigure 7 the first graph shows the modulated signal, second one shows the outputof PN code, and the final one shows modulated signal after spreading, if thechip rate in WCDMA is 3.84 Mcps then the sample time should be 2.
6µsec, butthis value makes the simulation very small , so that it is multiplied as in allour parameters by 1000 ,hence it becomes 2.6e-3 sec .While the bit rate is 128Kbps should be 128 bps , this signal has the processing gain 15dB (10log3.84e6/128e3) , from the interference present in the CDMA system .
This effectis termed ‘processing gain’ and is a fundamental aspect of all CDMA systems. Itis important to understand that spreading/despreading by itself does notprovide any signal improvement for wireless applications. Indeed, theprocessing gain comes at the price of an increased transmission bandwidth. 2)Multipath Fading Channel and Noise; As mentioned in earlier chapters, theMultipath is implemented by delay blocks and combined together; the result ofthis process is shown in Figure (8) The first graph shows the output from thefirst path, the second one shows the second path and the third one shows thethird path. The fourth graph shows the combined of all these three paths beforeadding together, the output looks as one signal because all the three paths aredelayed with the same delay sample as shown below. The fifth one after addingand the last one indicates the signal before Multipath channel The last twofigure show the combination of three paths in case of different delay samplebefore adding together and after adding together respectively. Fig. 8 MultipathFading Channel Results Theeffect of adding noise to the signal is shown in Figure (9) The first figureshows the signal before fading channel and the second one is after fadingchannel, and the last one after adding noise The signal power needs to betypically a few decibels above the noise, the required power density isdesignated as Es/No ,where Es is the energy or power density and No is thenoise power density .
In this model Es/No = 10 dB Fig. 9 Signal WithAdded Noise 3)Rake Receiver; As we mentioned before , the RAKE receiver contains five stages,these stages are (1)bank of correlators ,(2) selector , (3)tracking loop ,(4)alignment ,despreading and combiner , and (5)demodulation .here we discussthe result of each stage below. I) Bank ofCorrelators Theresult is shown in Figure(10) Fig. 10 Bank Of Correlators Results II) Selector In thisstage we select the strongest signal from correlator, a test of this block isshown in Figure (11) Fig 11A Test Of Block As shown abovewhen we apply random input ,the selector selects the highest one and goes tonext one and so on ,it determines the number of highest input not the value,now we need to organize these outputs from the highest number of input to thelowest one, this process is shown in Figure (12) Fig 12 Selection Process III) Tracking Loop Once thecomplete received signal, we try to synchronize it. Within each correlationreceiver, we have to track the fast-changing phase and amplitude valuesoriginating from the fading process .this tracking process has to be very fast.
the result is shown in Figure (13) The first graph illustrates the PN codewhile the other one shows the code generated by the tracking loop. As we seethe local code synchronizes With PN code Fig. 13 The Tracking Loop Process IV) Alignment,despreading and combiner In thisstage all outputs signals are aligned as in step seen earlier, then it ismultiplied by the outputs of the tracking loops such that all these signal havethe same phase as shown in Fig.
14 After Spreading Fig. 15 After Despreading Figure (15) On the figure (15), the last diagram showsthe combined signal while the first ,2nd ,3rd and 4th diagram show the outputafter despreading (modulated signal) V) Demodulation Thereceived signal after demodulation is illustrated in second diagram in Figure(16) below where the transmitted signal is the above one. Fig. 16 Transmitted and Demodulated Signal IV.
ANALYSIS AND COMMENTS OF RESULTS From the resultsobtained after running the program and for several cases to see the influenceof the change on these result due to an increase in the number of paths andincrease the number of selectors, as shown in Figures (17-1, 17-2, 17-3, and 17-4)which show the rate of error change in each case. When we have two paths andthree selectors, the error range is about .0009 and reduces gradually when thenumber of selectors increases to four, to become .0005. The error then drops to.00029 when paths increase to three paths and selectors to three. The errordrops to .00025 when number of selectors becomes four, so an increase in thenumber of paths and figures in the rake receiver leads to improvement in thebit error rate.
Then the rake receiver is able to overcome the problem offading channel due to multipath, though the present multipath was quite ahandicap to the proper reception of CDMA, it has now become an advantage for abetter performance because the more paths that exit, the better the receptionof RAKE receiver and the lesser the error rate incurred. We note that from theseresults that the error ratios were constant. As the number of bits increases,the error ratio starts to rise such that its ratio is constant As the errorratio is in the shape of a straight line, this indicates that the receiverdeals with the signal according to various conditions in a good manner.
Theerror ratio drops only if the number of paths or the number of selectorsincreases. Fig.17-1 Result of two path, three selectors Fig. 17-2 Fig.17-3 Result of three path.three selectors Fig. 17-4 Result of three path, four selectors Effect of noise There are many ways in which digitalcommunication systems might be compared.
One of the most important comparisonsis based on how efficiency the system can utilize the available signal energyto transmit information. A useful measure of this efficiency is the energy perbit (or per symbol). Since all systems have noise in them, the energyutilization is defined as Es/No, where Es is the energy per symbol and No isthe one sided noise spectral density To employ the energy utilization incomparing our model, we simulate this model in four cases as discuss before, itis necessary to relate (Es/No) to the bit error, to show how the change inEs/No will effect on bit error rate, however, the change in fingers and paths.This result shown below; Fig. 18 log(BER) vs. Es/No The Figure (18) shows clearly the effect of bit errorrate, whenever the Es/No, for each case in the study. From it we can note theextent of the influence of the multipath.
The more paths there are, the lesserror rate we have. In the case of three paths and three selectors, the errorrate drops more than the case where these are two paths and four selectors,despite the increase in the number of selectors. In the second case the signaldid not improve while it improved when the paths increase from two to three. Inthe case of four selectors and two paths, and four selectors and three paths,bit error dropped in the second case at a greater rate There is a greatproximity between three selectors, three paths and four selectors, three pathsbecause the influence of multipath has been nullified in this case and only theeffect of the selector exits, which improves the signal slightly To compare thisstudy with previous one, and to a conventional receiver. Figure(19) below showsprevious study compare between RAKE and conventional receiver.
In this diagramthe red color shows RAKE receiver, where the green one shows a conventionalreceiver. The RAKE receiver will gives about 2dB of gain when compare to theconventional receiver. Fig. 19 It is clearly seen that a RAKE receiver performed muchbetter than conventional receiver.
V.CONCLUSION The block of fading channel in simulink program doesnot give acceptable results because it operates only in the basebandfrequencies, so we applied some delays multiplied by factors that lead toRayleigh fading. Other limitations we have never changed the code of any blocksin simulink, hence have not been able to control the results of our model. Wehave only changed the properties of the blocks to get as good results aspossible In future models, either the blocks codes are to be altered in orderto have better control or to device our own code without using simulinkprogram. Simulation was carried out to study the performance of RAKE receiverin a multipath channel. It was noticed that, the greater the number of paths,the better the signal received. Also the more fingers (selectors) we have, thebetter the received signal. Then AWGN channel model is fallowed to our model,to provide proper channel noise.
The result shows that increasing of Es/No,decrease the bit error rate. These results as discussed before are limited bysimulink program blocks. But we have adjusted them to obtain the best possibleresults. Appendix A Model of RAKE receiver Four selectors Fig. Model Of A Rake Reciever With Four Selectors Referrences (1)Rodger E. Ziemer & Rodger L.
Peterson “DIGITALCOMMUNICATION AND SPREAD SPECTRUM SYSTEMS” Macmillan Publishing Company, Inc,1985 (2) Peterson, Ziemer, Borth INTRODUCTION TO SPREADSPECTRUM COMMUNICATION Prentice-Hall, 1995 (3) B.Sklar DIGITAL COMMUNICATION FUNDAMENTAL ANDAPPLICATION Prentice-Hall, 1988 (4)Harri Holma and Antti Toskala “WCDMA FOR UMTS” JohnWiley & Sons, Ltd, 2000 (5)George R. Cooper & Clare D. McGillem “MODERNCOMMUNICATIONS AND SPREAD SPECTRUM” McGraw-Hill, Inc.
1986 (6) ECPE4654: DSP Implementation of CommunicationSystem RAKE RECIEVER IN ASYNCHRONOUS CDMA SYSTEM Project Report 10th May 2001 (7) http://kilyos.ee.bilkent.edu.tr/~signal/BCSP/szabo.pdf