Re: [Discuss-gnuradio] Question on gr-trellis PS and PI matricies in FSm

[Eric Blossom]

On Thu, Feb 01, 2007 at 12:51:03PM -0500, Achilleas Anastasopoulos wrote:
> Dear Tim,,
> 
> indeed the current implementation supports only this kind of "regular" 
> trellises.
> I believe the intruduction of non-regular trellises (ie, ones having 
> exactly I outgoing edges from each state, but different incoming edges 
> to each state) is a good idea.
> 
> They way I think about it, I have to substitute the PS (and PI) 
> internally generated matrix which is now of size SxI with a vector of 
> size S where each element is a  vector of size equal to the number of 
> incoming edges in the particular state. Then a couple of minor 
> modifications to the VA, and it is done...
> 
> I am not sure how familiar you are with the code I developed in 
> gr-trellis, but I am willing to help you make the modifications.
> 
> Best
> Achilleas

Thanks to both of you!
Looking forward to having more great stuff in GNU Radio.

Eric