INTERNATIONAL JOURNAL OF ELECTRONICS AND COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET)

Transcription

1 INTERNATIONAL JOURNAL OF ELECTRONICS AND COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET) International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME ISSN (Print) ISSN (Online) Volume 3, Issue 3, October- December (2012), pp IAEME: Journal Impact Factor (2012): (Calculated by GISI) IJECET I A E M E SOURCE AND ADAPTIVE CHANNEL CODING TECHNIQUES FOR WIRELESS COMMUNICATION SUDHA.P.N1 Dr U.ERANNA 2 1 Assistant Professor & H.O.D. Department of Telecommunication Engg, K.S.Institute of Technology 2 Principal, Ballari Institute of Technology, BELLARY. pnsudha@gmail.com ABSTRACT The services offered by the mobile communication system are enormous and hence number of mobile subscribers are increasing at a larger rate. To accommodate all these users we need to redesign the cellular system which is a tedious process. This drawback can be overcome by the proposed algorithm. The proposed method uses low delay celp to compress the speech signal. The LD-CELP [1] has low processing delay compared to the CELP used in GSM systems. The source coded data is further error protected by using the adaptive channel coding technique. The type of channel coder chosen depends on the effect of the noise in the channel. This paper aims at designing an algorithm for memory and memory channels with high S/N ratio. Here the noise effect of the channel is predicted by transmitting a hand shake packet between the transmitter and the receiver. If the channel is erroneous then we use complex channel coding technique and the obtained data transmitted by using full rate packet coding[2,3] else we use simple channel coding technique and transmit the data by using half rate packet coding. Keywords: LD-CELP, CELP, QAM, Cyclic coder, Convolutional coder I.INTRODUCTION The analog I/P speech is converted into the digital data by using the A/D converter with the sampling rate of 8khz.The output digital data rate obtained is 64kb/sec and the bandwidth required to transmit the same will be very high. To overcome this draw back we are using the speech compression technique [4]. The speech compression technique used is LD- 314

2 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME CELP. By using this technique the speech is compressed at the rate less than 16kb/sec. The compressed data is further encoded using the adaptive channel coding techniques. The channel coding techniques chosen mainly depends on the quality (S/N) of the handshake packet received. If the S/N ratio of the packet is high then we use cyclic coding technique, where number of redundancy bits added for error correction and detection will be less else we use complex convolution coding technique[12,13] with large number of redundancy bits added to protect the data transmitted. The channel encoded data is transmitted using the QAM technique. The flow of the paper is as given below II.SPEECHCOMPRESSION TECHNIQUE Speech compression is one of the important domains in digital communication. As we have seen there are different algorithms for compressing the speech.the important factor that separates algorithm is its complexity. Coder complexity is a function of the signal processing involved.for a complex coding algorithm the processing time will be maximum hence the processing delay will be maximum. The different types of delays that can be encountered while processing the speech are o Buffering delay at the encoder: This delay is caused by the number of samples that are to be gathered before the processing begins. o Processing delay at the encoder: The time required to processes all the buffered samples. For the o complex algorithm the processing delay will be very high. o Transmission delay: The time required to transmit the processed data from source to the destination. o Buffering delay: The speech signal from the transmitter is transmitted in the form of different o parameters like LPC, pitch, gain and so on.to start the synthesis at the receiver side a full knowledge of these parameters is required.usually it is assumed that the buffering delay at the transmitter side is similar to the receiver side. o Processing delay at the receiver: The time required by the decoder to get back the original information that was transmitted from the transmitter using the parameters obtained from the decoder. The total delay can be estimated as 3.5 times the coding delay. The algorithm with least coding delay will have less total processing delay. Such delay can have important ramifications in aspects of speech communications such as echo control in long distance communication [12]. For any communication system the path with least path delay are chosen to avoid the use of echo cancellers. The conventional CELP has got high processing delay as it uses forward adaptive technique over frame size 20ms samples rather than LD-CELP [5, 6, 7] which backward adaptive technique with the frame size of five samples.hence for speech compression least processing delay algorithm plays a vital role. One such algorithm is Low Delay CELP [LD-CELP].This has processing delay less than ms. 315

3 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME Outline of LD-CELP algorithm: In conventional CELP [22,23]coder the predictor parameters along with the excitation signals are transmitted. The predictor coefficients are updated by performing a LPC analysis on the previously quantized speech. In case of LD-CELP it is backward adaptive version of CELP[8,9].LD-CELP coder works using the analysis by synthesis approach to code book search. This coder uses backward adaptation of predictors and gain to achieve an algorithmic delay less than ms. Here only the index to the excitation codebook is transmitted. The predictor coefficients are updated through the LPC analysis of previously quantized speech. The block size of the excitation vector chosen is of five samples [26].The long term predictor of the CELP[20] are replaced by the high order STP predictors. The coefficients of STP are updated once every four excitation vector by performing the LPC analysis[10] on previously quantized speech. The excitation gain is updated once every five samples [vector] using 10th order adaptive linear predictor in the logarithmic domain. The coefficients of the log gain predictor are updated once every four excitation vectors by performing a LPC analysis on previously quantized and scaled excitation vectors. A 10 th order perceptual filter is updated once every four excitation vectors by performing the LPC analysis on the original speech samples at the encoder. The five speech samples are quantized and represented by 10 bits. The excitation code book is made up of a 3-bit gain and 7-bit shape codebook [24]. In a backward adaptive configuration, the parameters of the synthesis filter are not derived from the original speech signal but are computed by backward adaptation method.the information is extracted only from the reconstructed signal based on the transmitted excitation information. Since both the encoder and decoder have access to the past reconstructed signal, side information is no longer needed for the synthesis filter and hence the low processing delay requirement is met [25].In Abs-LPC coding system a closed loop optimization procedure used to determine the excitation signal which when used to excite the model filter produces a perceptually optimum synthesis speech signal. The basic idea behind Abs is as follows. First it is assumed that the signal can be observed & represented in some form i.e. time or frequency domain. The model has a number of parameters which can be varied to produce different range of the observable signals. To derive the representation of the model a trial & error procedure is applied. In the LD-CELP only the excitation signal is transmitted. The predictor coefficients are updated by performing a LPC analysis on the previously Basic LD-CELP Encoder and Decoder: LD-CELP Encoder: The speech signal is sampled and is partitioned into blocks of five consecutive input signal samples. Each block is called a vector and these vectors are stored in the vector buffer. For each input block the encoder passes each of 1024 codebook vectors through the gain scaling unit and a synthesis filter. From the obtained resulting 1024 candidate quantized signal vector the encoder identifies the one which minimizes a frequency weighted mean square error 316

4 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME when the quantized signal vector is compared with the input signal vector. The best candidate quantized signal vector is represented by a 10 bit codebook index[11] is transmitted to the decoder. The best code vector is passed through the gain scaling unit and the synthesis filter to establish the correct filter memory in preparation for the encoding of the next signal vector. The gain coefficients and the synthesis filter coefficients are up dated periodically in a backward adaptive manner based on the previously quantized signal vector. The basic simplified encoder block diagram is as shown in the figure1.1 I/p voiced data Convert to uniform PCM Vector buffer VQ index Excitation VQ codebook Gain + Synthesis + - filter Perceptual weighting filter 16kb/s output Minimum MSE Backward gain adaptation Backward predictor adaptation Figure 1.1 Basic LD-CELP Encoder LD-CELP decoder: The decoder on receiving the 10 bit index performs the table look -up to extract the corresponding code vector from the excitation code book. The extracted code vector is passes through a gain scaling unit and a synthesis filter to produce the current decoded signal vector. The gain and the synthesis filter coefficients are up dated in the similar way as in the encoder. The decoded signal vector is then passed through an adaptive post filter to enhance the perceptual quality. The post filter coefficients are updated periodically using the information available at the decoder. The five samples of the post filter signal vector are converted into the PCM output samples. The basic decoder block diagram is as shown in the figure 1.2 III: CHANNEL CODING TECHNIQUES 3.1 Introduction: The compressed data from the speech coder [16,17] is further given to the adaptive channel encoder. Channel coding [27, 28] is a viable method to reduce information rate through the channel and increase reliability. This goal is achieved by adding redundancy to the information 317

5 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME symbol vector. The resulting symbol vector is a longer coded vector that are distinguishable at the output of the channel. The purpose of forward error correction (FEC) is to improve the capacity of a channel by adding carefully designed redundant information to the data being transmitted through the channel. The process of adding the redundant information is known as channel coding. In this paper we are using adaptive channel coding technique. Depending on the quality of the channel the type of the channel coder will be decided. The channel coders used in this paper are cyclic coder and the convolutional coder[34]. A brief note about the channel coder used is given in the following sections. 3.2 Introduction to cyclic coder: Let c be the codeword in a linear block code of length n. The cyclic code over GF(q) are class of linear codes of block length n if the code obtained after right shift or left shift is also a codeword. Therefore the linear code c is cyclic precisely when it is invariant under all cyclic shift [31].These cyclic codes are based on the Galois field and their structure is strongly related to Galois field because of this encoding and decoding algorithms are computationally more efficient. The cyclic codes form an important the subclass of linear block codes. These codes are attractive for two reasons as mentioned below: VQ Index 16kb/s Excitation VQ codebook Gain Synthesis filter Post filter Convert to PCM 64kb/s o/p Backward gain adaptation Backward predictor adaptation Figure 1.2 Basic LD-CELP decoder The encoding and syndrome calculation procedures can be easily implemented. The constructional methodology is simple and hence makes it possible to design codes with useful error correcting properties.in cyclic code the theory of Galois field has been used to spot a good code. Cyclic codes arevery important because of their underlying Galois field description that leads to encoding and decoding procedures that are algorithmic and computationally efficient. The implementation of the cyclic decoders is easy compared to the non cyclic linear binary codes. To perform the correction the decoder has to determine a correctable error pattern e(x) from the syndrome s(x)[28].the syndrome value obtained indicates the presence of error,if the syndrome value is zero then received vector v(x) will be similar to the transmitted vector c(x).if the syndrome is non zero then v(x) c(x) hence indicating the presence of error. Once the syndrome value is calculated then can find the position of the error with the equation below 318

6 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME e( x) = [q( x) + m(x)]g (x) + s( x) Where m(x) is the message polynomial, g(x) is the generator polynomial s(x) is the syndrome polynomial q(x) is the quotient polynomial obtained by dividing v(x)/g(x). After finding the value of e(x) using the cyclic decoder we can get the corrected codeword using the following equation given below. c(x) = v(x) + e(x) The cyclic decoders are less expensive and less complex. This channel coder basically used in this thesis to correct single or double error pattern. These cyclic codes are very efficient to correct single and double error pattern but not preferred to correct multiple error corrections as complexity of the algorithm increases. As the complexity of the algorithm increases the processing time required will also increase. Hence to overcome this drawback we use cyclic coding technique for less erroneous channels or for errorless channels. 3.2: Introduction to Convolutional coder: In case of block codes the block of n digits generated by the encoder mainly depends only on the k message within a particular time slot. In case of convolutional codes the n number of coded bits generated not only depends on the k message bit with in particular time slot but also on the previous blocks of message bits. These codes can be designed to detect and correct the errors.encoding in case of convolutional codes can be accomplished using simple shift register and decoding can be done using many practical procedures. Recent studies have shown that convolutional codes perform better than that of block code in many error control applications [28].The convolutional codes mainly used to achieve reliable data transfer in different applications like digital video, mobile communication and satellite communication. The convolutional codes can be both systematic codes and non systematic codes. Systematic codes are those where in the message bits are in a separate subblock and the check bit are in a separate subblock. The systematic convolutional codes are preferred as the implementation is easier than that of non-systematic codes. The very important advantage of systematic code is that it prevents the propagation of error. The error correction property of both the coding methodology is the same. A convolutional encoder takes sequence of message and generates sequence of code digits. At any time unit k message bits are fed into the encoder and the encoder generates a code block consisting of n code digits.the code generated by this encoder is called an (n,k) convolutional code of constraint length nn digits,where n is the size of the codeword generated and N is the total number of message blocks. The rate efficiency of the code can be written as k/n. The convolutional 319

7 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME encoders are mainly designed using shift registers and modulo-2 adders. The convolutional codes can be generated by the following techniques as mentioned below. Time domain approach Transfer domain approach The communication line between the transmitter and the receiver is set to check the noise quality of the channel.the quality of the channel is checked by transmitting the handshake packet which is known by both the transmitter and the receiver. If the S/N ratio of the received packet is more than the threshold then the cyclic coding technique will be used to generate the redundancy bits.the packet size at the output of the cyclic coder will be bits. Hence cyclic coder generates the half rate speech coded data when the channel is error less or less erroneous. By this methodology we can accommodate two users within one full rate traffic channel.hence the channel efficiency is improved. If the S/N ratio of the hand shake packet received is less than the threshold then convolution coding technique will be used. In convolution coding the number of error correcting bits added will be more and it can detect and correcting maximum number of errors. This in turn improves the signal quality through erroneous channel by improving the S/N. The size of the packet at the o/p of the convolution coder is This is further transmitted by using one full rate traffic channel by improving the S/N ratio. At the receiver the size of the packet is checked. If the packet size is equal to that of the half rate speech coder then the channel decoder used will be cyclic coder else convolution coder will be used. For the convolutional decoder mainly uses viterbi decoding algorithm to decode the transmitted data. IV.MODULATION TECHNIQUE The channel coded data is modulated using QAM [9] technique. Quadrature amplitude modulation (QAM) is a modulation scheme in which two sinusoidal carriers, one exactly 90 degrees out of phase with respect to the other, are used to transmit data over a given physical channel. Because the orthogonal carriers occupy the same frequency band and differ by a 90 degree phase shift, each can be modulated independently, transmitted over the same frequency band, and separated by demodulation at the receiver. For a given available bandwidth, QAM enables data transmission at twice the rate of standard pulse amplitude modulation (PAM) without any degradation in the bit error rate (BER). QAM and its derivatives are used in both mobile radio and satellite communication systems. V.RESULTS OBTAINED This graph gives the information about the energy of the signal to the bit error rate when the message signal is passed through the Gaussian channel and also through the fading channel when the message signal is sampled at the rate of 6kb/sec and 8kb/sec. 320

8 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME Figure:1.3 Signal to BER for signal sampled and compressed Figure:1.4 Input speech and decompressed speech In the figure 1.3 the red plot is w.r.t. Gaussian channel and the green plot is w.r.t fading channel. The second graph in the figure1.4 gives the comparative results of input speech at the transmitter and decompressed speech at the receiver. I/P SPEECH SAMPLE RATE RATE OF COMPRESSED O/P SPEECH PROCESSING TIME REQUIRED 48Kb/sec 12Kb/sec sec 64Kb/sec 16Kb/sec sec 321

9 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME VI.CONCLUSION The above presented speech compression technique gives toll quality speech signal at 16Kb/sec with low computation delay. The above presented technique can generate either the full rate traffic data or half rate traffic data depending on the noise level of the channel. If the channel is less erroneous then the coder uses cyclic coder and generates half rate traffic channel hence accommodating more number of user with high S/N ratio.if the channel is noisy then it uses convolution code and generates full rate traffic channel with high S/N ratio with minimum computational delay. ACKNOWLEDGEMENT I want to thank one and all who directly or indirectly helped me to complete this paper. REFERENCES [1]. Juin-HweyChen, Richard V. Cox, Yen-Chun Lin, Nikil Jayant, and Melvin J. Melchne, A low-delay celp coder for the ccitt 16 kb/s speech coding standard, IEEE journal on selected areas in communications, vol. 10, No. 5, June 1992,pp [2]. Gary J. Mullett, Introduction to wireless telecommunications systems and networks [3]. Theodore S. Rappaport, Wireless Communication 2 nd edition, Prentice-Hall, New Jersey, 1996 [4]. Philip B. Gieseier and John B.O Neal, Jr. Speech bandwidth reduction November 1980, FCC Office of plane and policy. [5]. Kamilo Feher, Wireless Digital Communication, IEEE Tr. Information Theory, Vol.53, no.1, Jan [6]. Alexis Pascal Bernard Source-Channel coding of speech Master of Science, 1998 [7]. V.lyengar, P. Kabal, A low delay 16 Kbits/sec speech coder -Telecommunications Universite du Quibec Montreal, Quebec H3A 2A7 Verdun, Quebec H3E 1H6 [8]. Vladimir Cuperman, Low delay speech coding, Communications Science Laboratory, School of Engineering Science Simon Fraser University, Burnaby, B.C., Canada V5A 1S6. [9]. Juin-Hwey Chen, Toll_quality16kb/s CELP speech coding with very low complexity Speech coding Research Department, AT&T Bell Laboratories. [10]. Lawrence Y.L.Wong, AMIEE.M.D A practical object oriented programming approach for implementing real-time speech compression algorithm. Hong Kong Design Centre, System Technology Group, National Semiconductor, Tsimshatsui, HONG KONG. [11]. J.Srinonchat Comparison of the efficiency of ordered and disordered codebook techniques in speech coding, Department of Electronic and Telecommunication, IEEE journal, ICICS [12]. Wilfred &Blanc, Christine Liu, and Vishnu Viswanathan An enhanced full rate speech coder for digital cellular applications Acoustic, Speech & Signal processing, IEEE Journal [13]. Fan Zhai, Zhu Li, and Aggelos K. Katsaggelos Joint Source coding and data rate adaptation for multi-user wireless video transmission, IEEE transaction

10 International Journal of Electronics and Communication Engineering & Technology (IJECET), ISSN (Print), ISSN (Online) Volume 3, Issue 3, October- December (2012), IAEME [14]. Atsushi Fukasawa, Takuro Sato, Tatumasa Yoshida and Manabu Kawabe Adaptive error control scheme for high speed data transmission through a fading channel, IEEE [15]. N.Seshadri, A.R. Calderbank and G.J.Pottie Channel coding for co-channel interference suppression in wireless communications, IEEE [16]. Douglas O Shaughnessy, Speech Communication, 2 nd edition, University press Ltd [17]. Thomas F. Quatieri, Speech Signal Processing, First impression 2006 [18]. A.M.Kondaz, Digital Speech coding for low bit rate Communication Systems, John Wiley & sons [19]. Audio, speech, and music processing, EURASIP Journal on Volume 2009, Article ID , 11 pages,doi: /2009/ [20]. Introduction to CELP coding. Douglas O Shaughnessy, Speech Communications, 2nd edition, Universities Press Limited, [21]. Coding of speech at 16kb/s using low delay code excited linear prediction Recommendation G.728 Geneva, 1992, CCITT the International (09/92) Telegraph and Telephone consultative committee. [22]. B.Atal. Efficient coding of LPC parameters by temporal decomposition Proceedings of ICASSP1983, pp 81-85, [23]. Low delay CELP coding at 8khps using classified voiced and unvoiced excitation codebook International Symposium on Speech, April 13-16, [24]. W.C.Y. Lee, Mobile communication Engineering McGraw-Hill, [25]. [26]. K.Sam Shanmugam, Digital and Analog communication systems [27]. Shu Lin & Daniel.J. Costello, Error Control Coding [28]. waves.blogspot.in/2008/05/types-of-channel-codes.html [29]. K.Giridhar, Information theory and coding, 1st edition. [30]. [31]. Richard E. Blahut, Theory and Practice of error control codes [32]. Salvatore Gravano, Introduction to error control codes [33]. Dr.P.S. Sathyanarayana, Concepts of Information theory and coding [34]. Sunaina Sharma, Combining Cryptography With Channel Coding To Reduce Complicity International journal of Electronics and Communication Engineering &Technology (IJECET), Volume3, Issue2, 2012, pp , Published by IAEME [35]. Pooja Prajesh and Dr. R.K.Singh, Investigation Of Outdoor Path Loss Models For Wireless Communication In Bhuj International journal of Electronics and Communication Engineering &Technology (IJECET), Volume3, Issue2, 2012, pp , Published by IAEME [36]. Dr. V. Murali Krishna, Karimella Vikram and Prof. Narasimha, Broadband Wireless Communication International journal of Electronics and Communication Engineering &Technology (IJECET), Volume3, Issue2, 2012, pp , Published by IAEME 323