Chapter 7. Conclusion and Future Scope

Transcription

1 Chapter 7 Conclusion and Future Scope

2 CHAPTER 7 CONCLUSION AND FUTURE SCOPE This chapter starts presenting the prominent results and conclusion obtained from this research. The digital communication system and its components were studied and experimental work was performed on MATLAB SIMULINK The work presented in this thesis demonstrates the detailed study of designing of linear block code and convolutional code. The work in this thesis mainly focused on error control code and signal power consumption by communication system. By this thesis work it was tried to minimize transmit signal power consumption by use of these error control codes. Section 7.1 of this chapter is giving summary and conclusion of this thesis. In section 7.2 potential future work and enhancement in field of error control code are being discussed Conclusion This subsection outlines a series of conclusions stemming from each of the result section of this thesis are as follows: In this thesis digital communication system and error control codes were examined. Different performance measurement parameters were studied and among them communication overhead/redundancy was taken for this thesis work and was tried to minimized this. The minimized error control code was used in digital communication and tried to minimize the transmit signal power consumption. The FEC technique is particularly suited for AWGN channel which is realistic noise model have been implemented for proposed block code, and convolutional code. In digital communication system mainly two types of error control codes namely block codes and convolutional codes are used. Block codes have larger message block size compared convolutional codes. Block codes are 107

3 capable in correcting limited number of errors and convolutional code can correct burst errors. Channel encoder designer uses different strategies for calculating error control bits; due to this overhead vary for different codes. After analyzing overhead in chapter 2, it was found that for same error correction capability codes have different numbers of parity bits. It was concluded that these parity bits can be reduced by doing modification in arrangements of message bits and in calculation of parity bits. Code rates are not same for same error correction capability code due to different strategy of parity bits calculation. Therefore communication overhead comparison was performed according to single bit and double bit error correction capability. Minimum communication overhead 5.88% was obtained for single error correction code and 11.11% was obtained for double bit errors. A communication overhead gain of 0.57% for single bit and 1.43% for double bit was obtained. The proposed block code has larger block size comparatively existing code and when block size was increased communication overhead got decreased and error correction capability remains same. For large block size say 256 or greater the encoding decoding delay and memory buffer size got increased and when more than two errors occurs system performance degraded. Without error correction code system requires E b N 0 range from 10 to15db for accepted BER rate while coded system requires 8 to 13dB. For calculating signal power gain; required power for coded and un-coded system at the same BER level was measured. The coded system requires approximately 2dB less power comparatively un-coded system. At lower BER level high power gain was obtained. Table 7.1 shows the power gain obtained at various BER levels by proposed block code. Convolutional code performance depends on code rate, constraint length, and generator polynomials. When code rate decreases system performance increases but up-to a limit after that constraint length has to be increased. 108

4 Very small code rate increases communication overhead therefore in this thesis only code rate 1/2, 1/3, and 1/4 were considered. In convolutional encoder; the code rate are taken according to the system requirements. When constraint length was increased for same code rate BER performance of system got increased but up-to a limit after that code rate has to be decreased. For code rate 1/2 suitable constraint lengths are 2, 3, and 4. Performance for these constraint lengths got varied but after constraint length 4 system performance remained constant. Now for increasing performance code rate has to be decreased. Again for code rate 1/3; constraint lengths are 3, 4, and 5 was taken and it was observed that after constraint length 5 system performance remained constant so again code rate has to be decreased for upgrading system performance. Generator polynomials affect the system performance. Good generator polynomials produce large Hamming distance between codes. The high Hamming distance code (distinct code) has high error correction capability. For selecting good generator polynomials BER performance were measured for each possible set of polynomials and final list was shown in Table This table shows the final designed convolutional code for code rate 1/2, 1/3, and 1/4. For designed convolutional code in Table 6.19 BER performance was measured and compared with un-coded system. The BER performance shows that coded system requires less power than un-coded system. The highest power gain was achieved by code rate 1/4 and lowest by code rate 1/2. When code rate was decreased power gain got increased and for lower BER level high power gain was obtained. The power gains for various BER levels and code rate are shown in Table 7.2. The highest power gain was 8.8dB for code rate 1/4 at BER level 10-5 and lowest was 3.4dB for code rate 1/2 at 10-7 BER level. 109

5 Table 7.1 Signal Power Gain by Proposed Block Code S. No. Bit Error Ratio (BER) Un-coded System Proposed Code Signal Power Gain (db) E E E E Table 7.2 Signal Power Gain by Proposed Convolutional Code S. No. Code Rate Bit Error Ratio (BER) Un-coded System Coded System Signal Power Gain(dB) E E /2 1.00E E E E E /3 1.00E E E

6 E E /4 1.00E E E Future Scope Error control coding applications have grown rapidly in the past several years in various field of communication and information storage mechanism. There are number of techniques of error correction based on applied mathematics which correct various types of errors. These codes have some limitations in mathematical or practical considerations or in other ways. It is impossible to correct all the errors but these errors can be minimized. Still no error correcting code is available which can correct all the random errors and burst errors. When number of errors was increased designed codes turn out to be inefficient. Small error correction codes with desired correction capabilities can be easily developed but with large error correction capability; developing a code is real practical problem. Now future work can be done on such error correcting code which would be capable to correct the errors with high probability. Study of this thesis provides many recommended processes that can be added in future work as a next step of system development which are briefed as follows: Considerations other performance evaluation parameters which are given in chapter 4 in continuation work. 111

7 Considering high block size and low code rate for block codes elaborated in chapter 5. Increasing the constraint length and developing a new low code rate convolutional code which leads to less BER and more ability to correct burst error elaborated in chapter 6. Generalization of Convolutional and Block codes in other important fields. Exploring the possibility of combining two block codes and their further studies. Searching for more perfect and large Hamming distance code. Study for higher generation cellular services for improving quality by low code rate containing high data rates. Searching for more quality real time data in space communication. 112