Fast Performance Evaluation of Turbo Codes Using Importance Sampling

• Main Authors: Bo-shen Lin, 林柏伸
• Other Authors: Shou-sheu Lin
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/71437747477234474696

碩士 === 國立高雄第一科技大學 === 電腦與通訊工程研究所 === 100 === Turbo codes is a channel coding technique with very powerful error correcting capability and widely used in various communication systems. Turbo codes have adopted by the fourth generation mobile communication – the Long Term Evolution（LTE）standard. Owing to the system complexity and it astonishing performance, analytical evaluation of turbo codes performance is quite difficult. So, our thesis use the LTE turbo codes as the research topic and evaluate its bit error rate（BER）. Owing to use the traditionally Monte Carlo（MC）simulation approach simulate bit error rate performance often requires a very large sample runs achieve accurate low BER estimates for LTE turbo codes, we propose a novel fast performance evaluation of turbo codes approach using importance sampling integrate with distance spectrum. The aim of this method is to make the decoder occur decoding error events more frequently, so that the fewer simulation samples can be achieve a given precision. As a result, when under the same simulation parameter, the result presents that the proposed method requires much smaller simulation samples and shorter simulation time to achieve the same accuracy required by the MC simulation approach. CRC stopping criterion in LTE protocol when block length is short, the turbo codes performance will be much worse. As a result, we have to use another stopping criterion or modify the protocol to solve this problem.