This preview shows page 1. Sign up to view the full content.
Unformatted text preview: provides hard decisions. Motivated by this desire for soft outputs, soft Viterbi Algorithms (SOVA’s) have been developed over the past decade. SOVA is a more complex but rather straightforward extension of the Viterbi algorithm. SOVA is, however, is suboptimum, and will not be covered in this subsection. . 2.5.1 Forward-Only Symbol-by-Symbol MAP For this Subsection we will use notation from Subsection 5.1.5 of the Course Text. Consider estimating the symbol at time k , denoted s(k) . This symbol can take on values Am ; m = 1, 2, · · · , M . This symbol is to be estimated based on the data rk ; k = 1, 2, · · · , k + D . Note that ”future data” is being used. (What this really means is that at time k + D we are to estimate s(k) , i.e. the estimate is delayed by D symbol durations.) We use future data because in general future data is a function of the present symbol, so that it can be used to better estimate the present symbol. D is called the delay parameter. Kevin Buckley - 2007 12 Le...
View Full Document
- Spring '09