※ 引述《datagram (...)》之銘言：
: 下面是網路上某人寫的，也剛好是我要問的問題
: 由於英文蠻淺的，所以我就不把它翻成中文：
: I have this problem where I need to use 8PSK and convolutional
: code rate 1/2. My problem is to do the soft-decision.
: For example：
略～
: But imagine for soft decision, how am I going to produce soft input
: out of each symbols? I mean I need to have 3 soft outputs from each
: symbol of this complex number, before it can go into the viterbi decoder?
: Thank you.
你說的沒錯～進入Soft Viterbi之前的確要先取得每個Bit的Soft值
然後再利用這些值做Soft Decoding....
取得的方法可參考很多最近關於Turbo Decoding的文獻...
總之就是一個Soft Demodulator.....
System Desciption r = h*x + n;
x = trasmit signal; h = fading channel; n = AWGN
ri = i-th bit in the 8PSK received symbol
Yi1 = the set of the modulation symbols which i-th bit is equal to 1
Yi0 = the set of the modulation symbols whick i-th bit is equal to 0
"he" is the estimation of channel
Soft Value:
p(xi=1) = sum p(ri=1 | he*Yi1) & p(xi=0) = sum(ri=0 | he*Yi0)
we need to normalize two above probabilities by p(xi=1)+p(xi=0)=1
大概就像降子，上面是自己憑感覺寫的，有錯請糾正囉～
--
現代人普遍的現象：
「小學而大遺」、「捨本而逐末」、「以偏而蓋全」、「因噎而廢食」
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.115.152.45
※ 編輯: lovewa 來自: 140.115.152.45 (04/17 02:25)