Re: [STDS-802-16] [PREAMBLE] Aug 23 Preamble Ad-Hoc group CC Summary
Jeff,
If you like we can have discussion in Seoul.
First of all, I guess you are mixing CDMA cell planning with
IEEE802.16e PUSC. Our field test experience is 80%
of delay spread is within 6uS. In CDMA there is a
chance to have two SUI-5s from adjacent BSs. But this
is not the case in PUSC deployment. Same segment interference
happens from BSs that are two tiers away. Proper PUSC
cell planning with have different segments for neighboring BS.
So the same SUI-5 with equal strength in PUSC "is" rare.
Secondly, you are mixing CAZAC despreading operation with
channel estimation. You are correct in stating that
multpath CIRs overlap in my CAZAC despreading. It is not a
problem if the two multipaths happen in different segments.
If they are in the same segment with unequal strength, then
CAZAC operation locks on the strongest path and treat the
other paths as interferers.
With regard to channel estimation, CAZAC is only to
uncover the phase. There is no "overlapping of multipath"
issue here. Any of the time non-alighned entities are treated
as interference here. Overlapping of CIRs has no bearing
in channel estimation. If you mean channel estimation using
CIR reconstruction, then yes overlapping is an issue. But
with your GCL/PN, first of all you need to get quality CIR
(hardware complicated). Any false peaks from GCL/PN
crosscorrelation create large bias in channel estimation
even in noiseless environment. Also there is no good measure
to set the threshold of determining "peaks" especially in
large multi-path spread environment. Not to mentioned
the required complicated hardware to do the "channel
reconstruction" based on CIR.
Besides, your GCL and other PN proposals haven't mentioned how you
can reasonably do neighbor BS scanning during HO and sleep
and fast acquisition of TOA, CIR, channel quality, etc., with
minimal hardware complexity. We have waited for your simulations
during Ad-hoc but you haven't produce any. So far I believe
CAZAC preamble is the only one that is able to provide
HO performance with accurate UL TX timing with minimal hardware
complexity.
I plan to end this email discussion. If you want to
discuss further we can do it in Seoul.
Thanks,
Jason Hou