LT碼廣播之傳輸控制研究

噴泉碼(或rateless codes)是一種可以適應於各種erasure rate環境的erasure codes，而LT碼是第一種可實際運用的噴泉碼，擁有噴泉碼可無限編碼的特性，且編/解碼所需的運算皆為低複雜度。但是LT碼擁有幾項缺點，當原始資料量接近無限大時，才能確保冗餘度為最小，且接收端在解碼過程中，原始資料的解碼率會長時間的維持在低檔，直到接收到一定比率的編譯封包後解碼率才會劇烈地上升。目前許多改良自LT碼的rateless codes已被提出，能夠改善LT碼在解碼前期或(以及)中期的表現，但是幾乎只能使用在一對一的傳輸環境。本研究旨在設計一種改良自LT碼的傳輸控制方法，且環境是設定於易掉封包的廣播傳輸環境，在我們的方法中，發送端模擬出大量接收端的解碼過程，以此預測接收端的解碼率分布，而根據解碼率的不同，發送端進而編碼對接收端最有幫助的編譯封包、再傳送給接收端，以此提高傳輸的效能。除此之外，我們分析了接收端的解碼過程，設計出工具以優化我們的結果。經由實驗結果可以得知，我們提出的方法的解碼率在解碼早期及中期的表現較LT碼佳，並能減少完全解碼所需的冗餘度。

關鍵字

噴泉碼； LT碼；廣播傳輸環境

並列摘要

Fountain (or rateless) codes are a family of erasure codes where the rate can be adjusted on the fly. LT codes are the first realization of fountain codes, having this beneficial property and low computational complexity in both encoding/decoding as well. However, some disadvantages of LT codes have been perceived, such as the redundancy is optimal only if the amount of source symbols, K, is approaching to infinity. Consequently in decoding, the source decoded rate, R, stays lower than 10% for a very long period, and increases drastically till receiving a certain amount of encoded packets. Currently, there are many strategies to improve the performance of LT codes in the early and/or intermediate stages for the unicasting transmission services. In this thesis, a transmission control strategy of LT codes for the error-prone broadcasting environment is proposed. Our method is simulated-based, which means the sender simulates the decoding status of a large amount of receivers to predicate the distribution of all decoded rates. To enhance the transmission efficiency in broadcasting, according to this estimated decoding status, the sender thus encodes the most beneficial packets and transmits to the receivers. In addition, some other tools to further improve the performance of decoding are given and discussed. The simulation results demonstrate that the proposed strategy promotes the decoded performance in both early and intermediate stages with less redundancy practically. Key words: fountain codes, rateless codes, LT codes, broadcasting environment, simulated-based.

並列關鍵字

rateless codes ； LT codes ； broadcasting environment ； simulated-based

參考文獻

[2] Michael Luby, “LT codes,” The 43rd Annual IEEE Symposium on Foundations of Computer Science, pp. 271–280, November 2002.

[3] Pei-Chuan Tsai, Chih-Ming Chen, and Ying-ping Chen, “Sparse degrees analysis for LT Codes Optimization,” IEEE Congress on Evolutionary Computation (CEC), 10-15 June 2012.

[4] Zao J.K., Hornansky M., and Pei-lun Diao, “Design of Optimal Short-Length LT Codes Using Evolution Strategies,” IEEE Congress on Evolutionary Computation (CEC), 10-15 June 2012.

[5] E. A. Bodine and M. K. Cheng, “Characterization of Luby transform codes with small message size for low-latency decoding,” IEEE International Conference on Communications (ICC‘08), pp.1195-1199, 19-23 May 2008.

[8] Ali Talari, and Nazanin Rahnavard, “On the Intermediate Symbol Recovery Rate of Rateless Codes,” IEEE Transactions on Communications, pp. 1237–1242, May 2012.

國際替代計量

LT碼廣播之傳輸控制研究

全文下載

主題瀏覽