[STDS-802-16] [PREAMBLE] - Comments on requirements
Hi all Preamble folks,
I would like to have some discussion on the requirements before the CC due
to the limited
time of the CC and the problem of on line interaction in a such large group.
I have a few comments on some of the requirements:
1) I do not see any reason why we should focus on simulations on the
proposed BW and FFT sizes, as
they will be almost the same (defining the same CS), this does not relates
to the large possibilities
of the BW/FFT size we already defined, as well as being relevant to only the
synchronization (maybe
tracking) and the estimation for the first 1/2 symbols.
2) I do not see the relevancy of the preamble to the MIMO modes as the first
symbols always start with
the PUSC transmission for all the users (especially the non STC/MIMO ones).
As we already agreed that
the preamble will hold for 1/2 symbols (for estimation) it has no relevancy
to the MIMO modes.
3) Complexity should be clarified, I guess people have in mind that some
preamble form in the time
domain has some correlation properties, which they claim exceeds the current
preamble (let us put it on the table),
the 1:2 preamble format (every second carrier modulated). I would like to
comment on the issue:
- In no way this is inherited from the 16d draft (out of scope I believe,
unless the mechanism is broken). From my perspective,
"scaled down" as defined in the PAR, means that the 2K FFT is the basic
structure, and all other FFTs should be
scaled down from it, meaning taking the basic building block.
- using time correlation properties to achieve frequency accuracy is, at
best, incorrect. We have to consider
the fact that in a mobile/multi-cell environment the preambles from
several BS's will be received by the user
and the correlation will give a weighted some of all the received
preambles (which will have no meaning as
it will include different delays and different frequencies due to
movement in different directions - towards the BS's).
Therefore one can not relay on this mechanism to achieve anything
(maximum a correlation value indicating
there is some reception from somewhere).
- Time locking on the preamble is also not accurate, due to the fact that
you will lock in time on the same weighted preambles
results which will not give you any information at best (and when trying
to stay locked on a specific BS which is
received poorly compared to other BS's - like in HO - the mechanism will
by default lock you to the one you do not want).
4) Overhead - I see no reason of using more then 1 preamble, as the
throughput punishment is severe.
5) Multi-Mode - I don't see how these new feature relates to mobility in any
way, and the fact that we already have
mechanism to switch between zones in the current definitions.
Best Regards,
Yossi Segal
yossis@runcom.co.il
CTO, Runcom Technologies LTD.
Broadband Wireless Technologies DVB-RCT, IEEE802.16
http://www.runcom.co.il/
Tel: 972-3-9528440 ext. 106 Fax: 972-3-9528805
2 Hachoma St. 75655, Rishon Lezion, Israel