Log-Likelihood Ratio to Improve Hard DecisionViterbi Algorithm

Abstract

Hard decision of Viterbi decoder suffer from the need of high Signal to Noise Ratio (SNR) to achieve reasonable Bit Error Rate (BER). For the purpose of improving the efficiency of its performance, it must increase the constraint length of the code, in this case highlights the problem of complexity in the structure. Several methods are used to solve this problem. In this paper the Log-Likelihood Ratio (LLR) with 3 bit soft decision and unquantized scheme has been implemented with simple transceiver using Convolutional codes. Results are achieved using the last version of Matlab (R2011b) illustrates that such scheme 2.7 dB over conventional system. In addition it has been examine such system with increasing the speed of data rate to double, the results of simulation confirm that it need 7 dB to achieve 10-6 BER.