Re: stds-802-16-mobile: OFDMA DL and UL pilot carrier allocation
Very well, Subhendu
My problem is "how can I used pilot carriers to set right the frequency domain
coefficients after I received preamble".
After I received preamble, becouse of channel variability, the frequency
domain coeffients can change: after I received some OFDM symbols,
found-by-preamble frequency domain coeffients assume slightly different
values whitch can be estimated (with a little error) by pilot carriers
(channel estimation).
I would like to understand how are this pilot carriers considered for channel
estimation: by hole OFDM symbol or by single subchannel.
If necessary, channel estimation techniques are well wanted.
Thanks.
Comments and suggestions are well wanted.
Cosimo Palazzo (Università degli Studi di Lecce - Italy)
Alle 18:12, venerdì 23 gennaio 2004, hai scritto:
> Hi,
> I have seen the UL Frame format for 16a and from that can answer your
> question. Observe that the UL Frame has a well defined preamble, which
> means that for each data subcarrier, the beginning of the frame will
> have a well defined preamble portion. Hence from that preamble you can
> find the frequency domain coeffecient for each subcarrier, as it is
> deterministic.
>
> Regards,
> Subhendu
> IIIT Hyderabad
>
> > Dear members,
> >
> > I'm a graduate student of the University of Lecce (Italy): I'm doing a
> > research about the TCP performance in the wireless environment.
> >
> > In this moment I'm making a Matlab implementation of the OFDMA physical
> > layer for 802.16e. My problem is the UL pilot carrier allocation.
> > In Std. 802.16a-2003 pilot carriers are handled as subcarriers of a
> > specific subchannel (for every subchannel), so (logical) pilot carriers
> > are
> > allocated using the described in 8.5.6.2 permutations (e.g. in
> > 8.5.6.2.1). In this way, in every OFDM symbol, physical pilot carriers
> > are allocated as
> > they are illustrated in ftp://193.204.78.98/pubblica/Mino/base.gif.
> > Otherwise DL, in this way UL physical pilot carriers are not distributed
> > uniformly in used spectra, but they are gathered together in particular
> > zones of spectra. I'm thinking this allocation is bad for correct channel
> > estimation operations !!!
> > If channel estimation operations are run "for subchannel" performance
> > does not change, because of great physical "frequency-distance" between
> > two consecutive pilot carriers compared as little logical
> > "frequency-distance" of Figure 128bb in Std.802.16a-2003.
> >
> > Could you tell me if I'm making a mistake or my observation is correct?
> >
> > Could you tell me how channel estimator should operate (for hole OFDM
> > symbol or for single subchannel)?
> >
> > What are the news for 802.16e?
> >
> > Comments and suggestions are well wanted.
> > Thanks in advance.
> >
> > Best regards.
> >
> > Cosimo Palazzo (Università degli Studi di Lecce - Italy)