Unformatted text preview: mum. That is, the paths pruned at any stage are guaranteed not to be optimum at any time in the future. In other words, at any stage K ≥ n, the optimum path (form the M K possible paths), is guaranteed to be one of the best paths into the states at time n. The key to optimality of the Viterbi algorithm is that all paths through the same branch incur the same incremental cost. In retrospect, this can be considered as the rationale for how the trellis diagram is set up. The states are deﬁned such that all of the path information required to compute an branch incremental cost is represented by the originating and terminating state. It should be noted that not all MLSE problems can be represented by a trellis in which all paths through a branch incur the same cost. In Subsection 3.4 we will examine such a problem, and discuss algorithms for addressing this issue. The principal advantage of the Viterbi algorithm is that the required computation and storage does not grow exponentially with symbol time n. For a trellis with s...
View Full Document
This document was uploaded on 10/12/2009.
- Spring '09