低密度奇偶校驗碼在頻道容量限制上有非常優異的能力,在錯誤更正碼方面的解碼能力非常接近沈濃界限(Shannon Limit)。現階段很多通訊系統的制定上都採用了低密度奇偶校驗碼像802.11n、802.16e…等等。在這篇論文中,我們研讀了近年有關低密度奇偶校驗碼在實做方面的論文。我們發現實現低密度奇偶校驗碼解碼器是非常困難的,因為硬體方面繞線的複雜度相高。若要實現多種編碼率的低密度奇偶校驗碼解碼器是更加的困難。我們將會採用每篇論文的優點設計低複雜度和良好解碼效率的低密度奇偶校驗碼解碼器。最適合的解碼演算法是階層信任延續演算法(Layered Belief Propagation Algorithm),因為在設計方面有較低的複雜度。我們去選擇班尼斯網路(Bense network)去解決硬體繞線困難的問題,因為班尼斯網的效果優於傳統的硬體繞線網路。我們將設計班尼斯網路的控制演算法,為了去節省記憶體使用和減少控制網路的複雜度。我們提出的架構可實現多種編碼率的低密度奇偶校驗碼解碼器而且不會消耗太多硬體的代價。我們也分析最後的結果為了使我們的設計能達到最好狀態。
Low-density parity-check codes have gained interest due to their excellent error-correction capacities, and performance is very close to the Shannon limit. There are also more and more communication systems adopt LDPC codes like 802.11n, 802.16e ....In this thesis, we survey the implement of LDPC decoder in recent years. We will find that it is very difficult to implement LDPC decoder because of high routing complexity. Implementing the LDPC decoder of multiple- code rates is more difficult. We will take every advantage of papers to design low-complex and good performance. The most suitable decoding algorithm “Layered Belief Propagation Algorithm” is chosen to implement because of lower complex. We will choose Benes Network to solve the problem of routing network because Bense network is better than the traditional routing network. We design the control algorithm of Benes Network in order to save memory and reduce the complexity of the control. Finally, our architecture of LDPC decoder can support multiple- code rates without consuming too much hardware cost. We also analyze the result in order to make our structure better.