7/30/2019 NCTT2003 - 01188314

1/5

4* National Conference on TelecommunicationTechnology Proceedings, Shah Alam, Malaysia

Error Correction ForCDMA Code In Mobile Communication

Tiong Sieh Gong , Tan Chang Ha?, Joseph Ding, Maham od Ismai141~2~3C011egefEngineenng,Universiti Tenaga Malaysia (UNITEN)

km 7, Kajang-Puchong Road, 43009 Kajang, S elangor Darul Ehsan. MalaysiaFaculfy OfEngineeringUniversiti Kebangsaan M alaysia43600 UKMBangi ,Selangor Darul Ehson. MALAYSIA

Abstraet - Error correction coding is one of the topicsthat obtained a lot of studies lately to improve thereliability of covunication on digital channels.Thus, this research is carried out in order to find abetter and effective way to be included in CDMAsystem to protect, detect and correct the error bitswhile information data is transmitted through variouskinds of environments. A new encoding algorithm,which is named as logic symbol repetition isintroduced to replace the normal symbol repetition inCDMA forward traffic channel. The new logic symbolrepetition is a memorizing encoding scheme thatable to correct the error bits, not only presentlyreceived bits but also errors in the previous receivedbits in case of block noise occurrence. A simulationprogram is developed to study the performance oferror correction by studying the BER of the receivedinformation bits.

1. IntroductionWhenever digital data is transmitted over a noisychannel, there is always a chance that the receiveddata will contain errors. Thus, error correction codingis used to reduce errors to a level that is lower than therequired bit error threshold level. In recent years theuse of error correction coding for solving this type ofproblem has become widespread.The utility of coding was demonstrated by the work ofShannon [I]. In 1948, he proved that if the datasource rate is less than a quantity known as thechannel capacity, communication over a noisy channelwith an error probability as small as desired ispossible with proper encoding and decoding.Essentially, Shannons work states that signal power,channel noise, and available bandwidth only set a limiton communication rate and not on accuracy. FromShannons research work, it is cleared that real limiton communication rate is set not hy channel capacitybut by the cost of implementation of coding schemes.

0-7803-7773-7/03/$17.00 0 2003 IEEE. 110

Cost constraints force communication at ratessubstantiallybelow capacity.

In recent years numerous researches have beendirected to find efficient and practical coding schemesfor various types of noisy channels. Most of theprogress toward finding practical schemes has come inthe last f ifteen years.Now, the error correction codingcan provide significant performance improvements inmany applications. There have been a number ofapplications where coding equipment has been builtand used successfully. The increasing practicality ofcoding is due to new developments within the field oferror correcting codes and the dramatic reductions incost and s ue of solid-state electronic devices.Although individual coding schemes take on manydifferent forms and have their roots in diversemathematical disciplines, they all have two commoningredients. One is the use of redundancy. Codeddigital messages always contain extra or redundantsymbols. These symbols are used to accentuate theuniqueness of each message. They are always chosenso as to make it very unlikely that the channeldisturbance will corrupt enough of the symbols in amessage to destroy its uniqueness. The secondingredient is noise averaging. This averaging effect isobtained by making the redundant symbols depend ona span of several information symbols. Some valuableinsight into the coding process can be obtained byexamining each of these ingredients separately [I].A simulation program is developed in Visual C+t@ tostudy the performance of error correction. Thesimulation s o h a r e provides simple buttons anddialog boxes on the monitor screen to let the userssetting for the simulation parameters and monitoringthe simulation.

7/30/2019 NCTT2003 - 01188314

2/5

2. Er ro r Protection in Forward Traffic ChannelIn this research, error correction coding scheme inforward traffic channel (FE) is studied. As shown inFig. 1, there are three channel encoding blocks inFTC; convolutional encoder, symbol repetition andblock interleaver. These three channel encodingprocesses protect the information hits whentransmitted through medium and able to detect andcorrect the error hits if the error rate is not too severe.

IPilo lPN

L o n p d i m a L g# lql lmlorI l r . ~ n d o n o t,p8ckm.rBNI i.nuMw I~ . ZXSPS

Fig. 1 :Forward traffic channel for Rate Set 1Fig. 1shows the forward traffic channel for Rate Set1. For this rate set, the vocoder is capable of varyingits output data rate in response to speech activities.Commonly, there are four different data rates used;9.6khps. 4.8khps, 2.4kbps and 1.2 khps [2].The baseband data from vocoder is convolntionallyencoded for error protection. Convolutional codes arebased on a linear mapping of a set of informationwords to a set of code words. Conceptually,information and code words are of infinite length, andthey are mostly referred to as information and codesequences. The existence of a maximum likelihooddecoding precedure that can he implemented withreasonable complexity is the reason for theirwidespread usage. An important fact of theconvolution is that soft input can he used and softoutput can he created when decoding convolutionalcodes [3]. For convolutional codes, the encoded hitsare functions of information bits and functions of theconstraint length. Specifically, every encoded bit is alinear combination of some previous information bits.In CDMA IS-95 system, the forward link uses a rate fiand constraint length K =9 convolutional code.After convolutional encoding, the data undergoessymhol repetition, which repeats the symbols whenlower rate data are produced by the vocoder. Thereason for repeating symbols is to reduce overall

interference power at a given time when lower ratedata are transmitted [4]. The motivation for thesymbol repetition scheme is to decrease the power perrepeated symbol when the vocoder is running at lowerrate. The scheme is effectively a way of takingadvantage of the voice activity factor in hardwareimplementation.After symbol repetition, the data is interleaved tocombat fading. Most error correcting codes performwell in correcting random errors hut during periods ofdeep fades, long streams of successive or burst errorsmay render the error correcting function useless.Interleaving is a technique for randomizing the bits ina message stream so that burst errors introduced bythe channel can be converted to random errors. InCDMA IS-95 system, the interleaver for the FTC nsesa matrix of 24 rows by 16 c o l m s the full rate [2].In this research, normal symbol repetition in CDMAsystem is replaced hy a new symbol repetitionencoding scheme, which is named as logic symbolrepetition. It is a memorizing scheme' that able tocorrect the present error bits based on the previousreceived hits in case ofblock noise occurrence.

3. Logic Symbol RepetitionSimulation is carried out for the forward trafficchannel with the logic symbol repetition scheme.Logic symbol repetition scheme consists of two parts,sending (Base station) and receiving terminals(Mobile station). The flowcharts for both sending andreceiving terminals are shown in Fig. 2 and Fig. 3.

I Get input from interleaver II Logical produeed output hit I

Repeat output hit based on thevocoder rate1Send the data for modulation+

EndFig. 2 Flowchart for logic symbol repetition

111

7/30/2019 NCTT2003 - 01188314

3/5

De-logic symbol

Receive data fromdemodulation

majority of bit repetitionI II

I

1 ISend to interleaver

Fig. 3 Flowchart for de-logic symbol repetitionLogic symbol controller at sending terminal is asimple controller which consists of 2 operation blocks;AND and OR operations. Data from interleaverwill be seut to logic symbol controller before beingrepeated certain times. Assume that the binary datasequence afler interleaver is 1011101001 and thevocoder is operating at 1200bps. Assume also theinitial condition for logic controller bits are zero. Thelogic symbol repetition will then operate as shown infig. 4 below.

Final data to be repeated+ 1111111(xS for 1200bps)1 [ 6

Step (2)9- x8oooooooor 12OObps)o r a oL 0m z g i n a l ata frominterleaver

Step(3) 9+ 1111111(x8 or 1200bps)

I* I 0

-Original datafrom interleaverStep(4) q- 0000000(x8 for 1200hps)

Orieinal data fromu -interleaverFig. 4 Logic symbol repetition at sending terminal

The data bit from interleaver is AND and O Roperated,with the logic control bits and the result fromboth logical operations is Exclusive-Owed (EXOR)to produce the final data bit. The final data of EXORwill then be repeated to a certain number of timesdepending of vocoder rate before sending formodulation. While the data is sent for modulation, theoriginal data bit from interleaver is shifted into theAND bit in the logic controller and the next data bitfmm interleaver is AND and O R operated with thenew condition of logic control bits. The processrepeats until all data from the interleaver is finish. Inthe example shown above, the data output that will besent for repetition is 1010011010 and the outputafler repetition will be;111111110000000011111lll0000000000000000111111111111111100000000111ll11100000000.

112

7/30/2019 NCTT2003 - 01188314

4/5

At the receiving terminal, the received informationfrom antenna will first be demodulated anddetermined the actual data bit based on the majority ofbits occurrence in binary information. The logicsymbol repetition processes of receiving terminal areshown below.Data determinedaccording tomajority ofbitP ccurrenceA

Data determinedacc,or$ng to ,majonty of bit+ occurrenceData to be sent to

logic symbol repetition.Data to be sentto de-interleave1 Thus, step (3a) is changed to step (3b) as below.

Data determinedaccording tomajority ofbitBuffer to keep result ofde-logic symbol de-interleaverData to be sent to

Actual data

Data determinedaccording tomajority of bitoccurrence logic symbol repetition.Buffer to k x p result of de-

Data to be sentto de-interleave10 is not accepted in thisproblem as step 3(a)

Y Wrong data is determinedbased on majority of bit'r '8-J occurrenceBuffer to keep resultof de-logic symbol

. . In this case thereare two possibilities

Data determinedaccording tomajority of bitP ccurrenceData to be sentto de-interleave1

Buffer to keep result ofde-logic symbolData determined

1 according to majorityDatae-interleavero be sent to lf i0 is'l' P f bit occurrencedataData to be sent tode-interleaver

Buffer to keep result of de-logic symbol repetition.

de-logic symbol repetition. T

113

7/30/2019 NCTT2003 - 01188314

5/5

Data determined according to- majority ofbi t occurrence

Data to be sentto de-interleaver

Buffer to keep result of de-logic symbol repetition.Data determined according to1 majority of bit occurrence

ModifiedData to be sent tode-interleaver

logic symbol repetition.El0 s no t accepted, as it issame result as step 7(a)

Fig 5 De-logicsymbol repetition at receiving terminalFor the receiving part, the received modulated data areactually the bits that being sent by the sendingterminal. The data bit to be decoded by logic symbolrepetition is determined by the majority of bitoccurrences in the demodulated binary string. Thedata bit is then decoded by Exclusive OR function,A and OR functions to retrieve back theoriginal data bits.In certain cases, there might be more than onepossible logical outputs as shown in the fig. 5 step (2).Therefore, either one of the data can be randomlychosen and the chosen data will be shifted to the lesssignificant bit (AND bit) of the logic control. Thedecoding scheme in the logic symbol repetition is ableto recover back the original data even though a wrongdata is being chosen at the first place. Example, whena wrong data bit is chosen as shown in Fig. 5step (2),the system will soon realize when come to thefollowing steps of the process as shown in Fig. 5 step(4a). This is because the bit that being chosen will not

be valid in the condition of the logical control and thiswill not allow the system to proceed to next step. It ishow the logic symbol repetition has the characteristicof memory n recovering the error bits.The system will also able to recover back the originaldata if the received data is loss during propagationthrough interfered medium by using the same method.In the cases where interference in propagation bascaused wrong determination of majority binary bit asshown in Fig 5 step (6). The system will realise atfollowing step as shown in step (sa). Then, the systemcorrects the produced output as shown in step 7(b).

4. ResultBy implementing logic symbol repetition in theforward traffic channel, the performance of the wholesystem can be improved. The hit ermr rate @ER) Willdecrease due to the improvement of recovering backthe error bits by logic symbol repetition scheme. Thegrade of improvement is significantly valuableespecially the infarmation data was transmittedthrough high interference environment. The detail ofthe simulation result will be presented in next paper.

5. SummaryAD efficient method of error correction in the forwardtraffic channel is urgently needed in todays era as thenumbers of the mobile phone are increasingdramatically. The increasing numbers of mobile usershave proportionally increased the BER in mobile linktraffic. Thus, in order to maintain or upgrade thequality of service provided, a better solution of errorcorrection is needed. Hence, implementing logicsymbol repetition in the conventional channelencoding is an alternate method to improve the BERperformance in CDMA system.

6. References[ I ] C.C. George Jr. and J. B. Cain, Error CorrectionCoding For DigiiaI Communications, PlenumPress,NewYork, 1981.[2] C. Y. Samuel, CDMARF Sysiem Engineering,Artech House, INC, 1998.[3] B. Vucetic & J. Yuan, Turbo Codes Principlesand Application, Kluwer Academic Publishers,

2000.Communications and Nehvork Security, PrenticeHall PTR, 1998.

[4] M. Y. Rhee, CDicz4Cellular Mobile

114