Re: Sliding window Viterbi
- From: porterboy76@xxxxxxxxx
- Date: 31 Aug 2005 00:38:33 -0700
> What has that to do with ISI? It seems to me to be a different issue.
I'm not 100% sure about this, but it appears that you can do Viterbi
equalisation, as well as Viterbi decoding. If a binary stream is input
to an ISI channel, the channel can be treated as a convolutional coder,
which means that a maximum likelihood estimate of the binary sequence
is returned by a Viterbi algorithm, matched to that "encoder". I am not
sure how this is done exactly, but I do know that the ISI channel must
be relatively short, since the number of states in the Viterbi decoder
(and hence memory requirements) grows exponentially with the "encoder"
length. I believe it works well for multipath channels with a small
number of paths. Maybe side information about the channel is also
needed. What I dont really understand is that a conventional
convolutional encoder works with modulo-2 adders, whereas a passive ISI
channel adds interference at each sample by regular addition. Maybe the
Viterbi sequence estimator must be modified to account for this.
Slainte
Porterboy
I woke up this morning and I got myself a beer
The future's uncertain and the end is always near
.
- References:
- Sliding window Viterbi
- From: john
- Re: Sliding window Viterbi
- From: Jerry Avins
- Sliding window Viterbi
- Prev by Date: Re: Mixing with a 3000Hz tone @ 8000 sampling rate
- Next by Date: Data width requirements for IFFT
- Previous by thread: Re: Sliding window Viterbi
- Next by thread: Re: Sliding window Viterbi
- Index(es):
Relevant Pages
|
|