等價于MAP的SOVA譯碼方法
MAP Decoding Methods Derived from SOVA
-
摘要: 不同于MAP(Maximum A Posteriori)算法，SOVA(Soft-Output Viterbi Algorithm)算法的軟輸出不是真正意義上的后驗概率，很少有文獻給出SOVA算法的完整數(shù)學(xué)解釋。該文給出了一種完整的SOVA的數(shù)學(xué)表達形式，并從SOVA的數(shù)學(xué)表達出發(fā)推導(dǎo)出了兩種等價于MAP，具有SOVA形式的譯碼方法，一種是Li等人(1995)給出的適用于連續(xù)傳輸?shù)淖罴衍涊敵鏊惴?Optimal Soft output Algorithm, OSA)；后一種是對OSA算法的改進，后者可以得到與前者等價的軟輸出，但是降低了運算復(fù)雜度。
-
關(guān)鍵詞:
- 軟入軟出譯碼;軟輸出維特比譯碼;最大后驗概率譯碼;最優(yōu)軟輸出譯碼;最大似然
Abstract: Different from MAP, the soft output of SOVA is not the exact a posteriori probability, and few literatures describe SOVA in mathematical form. This paper gives a new comprehensive mathematical expression of SOVA, and deduces two MAP-equivalent SISO decoding methods in the form of SOVA. The first method is just OSA algorithm given by Li et al, (1995)which is suitable for continuous decoding. The other one is the improved version of OSA，whose output is equivalent to that of OSA with the decrease of complexity. -
計量
- 文章訪問數(shù): 2508
- HTML全文瀏覽量: 129
- PDF下載量: 1330
- 被引次數(shù): 0
下載:
