Forums

Viterbi algoritms for STTC

Started by ibto75 August 18, 2009
Hi all,
Can anybody please tell me if there is any different between the viterbi
algorithm for Convoluntional code and that for Space time trellis codes.
I have tried using the decoding approach for convoluntional and trellis
coded modulation (TCM) to decode my 2 states BPSK  space time trellis code
but the ber that  is producing is quite embarasing.
Can anybody please help me out.
I am writing my code in C++.
It will be aprreciated if someone can lend a helping hand here.
On Aug 18, 10:53 am, "ibto75" <ibt...@gmail.com> wrote:
> Hi all, > Can anybody please tell me if there is any different between the viterbi > algorithm for Convoluntional code and that for Space time trellis codes. > I have tried using the decoding approach for convoluntional and trellis > coded modulation (TCM) to decode my 2 states BPSK space time trellis code > but the ber that is producing is quite embarasing. > Can anybody please help me out. > I am writing my code in C++. > It will be aprreciated if someone can lend a helping hand here.
Well, if you are trying to replicate the original Tarokh et al. paper, you need to include the channel in the trellis representation, and consider vectors instead of scalar values for each symbol. Or, are you expecting somebody to give you C++ code for this? I'm not clear what you are asking for, since I have no idea what exactly you have tried. First of all, does your trellis decoder work for a one-antenna system? Julius
>On Aug 18, 10:53 am, "ibto75" <ibt...@gmail.com> wrote: >> Hi all, >> Can anybody please tell me if there is any different between the
viterbi
>> algorithm for Convoluntional code and that for Space time trellis
codes.
>> I have tried using the decoding approach for convoluntional and
trellis
>> coded modulation (TCM) to decode my 2 states BPSK space time trellis
code
>> but the ber that is producing is quite embarasing. >> Can anybody please help me out. >> I am writing my code in C++. >> It will be aprreciated if someone can lend a helping hand here. > >Well, if you are trying to replicate the original Tarokh et al. paper, >you need to include the channel in the trellis representation, >and consider vectors instead of scalar values for each symbol. > >Or, are you expecting somebody to give you C++ code for this? > >I'm not clear what you are asking for, since I have no idea what >exactly you have tried. First of all, does your trellis decoder >work for a one-antenna system? > >Julius >
Hi Julius, Thanks for your reply. I have tried something and it seem is working for bit error rate . I am getting the bit error rate for the taroch paper for 4 state QPSK which i am trying to reproduce. But as i increase the state there is no performnace improvement as expected. Another problem i am facing is getting the frame error rate. Please if you will not mind i can send the written code to you may be you can find where i am missing it. I used the generator matrix notation for the code and have simulated using 2 transmit and one recieve antenna. Will be glad to rea from you.
ibto75 <ibto75@gmail.com> wrote:

>I have tried something and it seem is working for bit error rate . >I am getting the bit error rate for the taroch paper for 4 state QPSK >which i am trying to reproduce. But as i increase the state there is no >performnace improvement as expected.
Why would you expect the performance to improve when the state is increased? In most turbo and turbo-like codes, there is no improvement in using constituent codes with larger states. IS there a published result claiming this for STTC? Steve
>ibto75 <ibto75@gmail.com> wrote: > >>I have tried something and it seem is working for bit error rate . >>I am getting the bit error rate for the taroch paper for 4 state QPSK >>which i am trying to reproduce. But as i increase the state there is
no
>>performnace improvement as expected. > >Why would you expect the performance to improve when the >state is increased? In most turbo and turbo-like codes, there is no >improvement in using constituent codes with larger states. > >IS there a published result claiming this for STTC? > >Steve >
Hi steve. As reported by the athour i.e Taroch in the original paper, as the state increases there is performance improvement though with increase in decoding complexity and this was shown in the paper and other authors that have published something in that area too