The complexity of the algorithm grows exponentially

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: /s ,s , · · · , s(k+D) ) P (s(k+D)) s(k−1) (k −1) (k ) (k −1+D ) p k (s ,s ,···,s (25) ) where pk (s(k) , s(k+1) , · · · , s(k+D) ) is the probability of state {s(k) , s(k+1) , · · · , s(k+D) } at time k . L is the system memory. The Figure 4 illustrates this time-recursive computation in the form of a trellis diagram. 2.5.2 Forward/Backward Symbol-by-Symbol MAP A disadvantage of forward-only symbol-by-symbol MAP is that the estimator of s(k) uses data only up to time k + D . The complexity of the algorithm grows exponentially with the delay parameter D . For example, there are M D states/stage in the trellis diagram. Thus D can not be selected too large. Given a block of data/symbols over time k = 1, 2, · · · , N , with N >> D, performance can be improved (especially for the early symbols) using a block processing symbol-bysymbol MAP algorithm which maximizes P (s(k) = Am /r1 , r2 , · · · , rN ) . (26) Kevin Buckley - 2007 stage k−1 stage k p(r k+D / s P (s =A 1 , .... , s k−1 (k) (k+D) (k+D−L) 13 , ....
View Full Document

This document was uploaded on 10/12/2009.

Ask a homework question - tutors are online