-Table one and text on page 6?
Yes, in WirelessMobileData.pdf, but it goes back
to earlier presentations, (GRPS ass-u&me.. JohnG,
did you get the cultural translation??))
- Yes, simulation are just as good as how well one
knows and can model "mother nature", luckily one
learns a lot every time mother nature doesn't
agree with the model in the simulations.
-multipath, 200kHz vs 1-4Mhz, 5us vs 1..0.25us.
If one could telepathically know the multipath components
in advance, and additionally portion out the amount of
information sent on bandwidths acoording to how multipath
affects those sub-channels (to use multicarrier,OFDM-speak)
one could make the perfect system.
But unluckily it takes a lot of horsepower and battery
to track those fast changing multipath components
for a mobile termnial or a mobile surrounding.
Just to take two examples:
- to really constructively add up multipath echoes one
has to track the channel at that high frequency
(200khz for 5us, 1Mhz for 1us, 4 Mhz for 250ns,etc,
depends on carrier frequency, buildings,topology,etc))
To pick the three strongest echoes one has to first find
them, and then track (all of) them, running everything at a
high enough speed.
That is pretty expensive and battery draining to do when
no data or data at some 5-10kbps is transfered,
running a lot of assembly, ram accesses at 1-4Mbps,
charging gates and busses.
And CDMA does not still "portion out" the signal on
the "good frequencies, channels", avoiding those
which are "bad" (to jump to the optimum, perfect
solution)
GSM actually has OFDMA built-in, if the handset is in
a bad spot, the network can switch the (200kHz bandwidth)
carrier to a better one.
But just as one shouldn't update a 450Mhz pentium to
a 600Mhz but wait for a 1.3Ghz P4, new silicon generation
and then a mixture of TDMA-CDMA-OFDMA will make
tracking multipath more and more possible, but not quite
yet. (for power restricted handsets, basestations can
burn a lot of power)
That is, I would really enjoy asking the good doctor
what he thinks about the OFDMA aspect of GSM,
as well as the idea of starting the data sequence
with a small signal to measure the channel, like
GSM..(<vbg>)
But this, IMO, boils down to the next topic, politics and
building global standards through consensus, CDMA was,
IMO, the way to get USA to take a seat on the global
train..???
(luckily I do not know anything about this, just guessing)
- synchronizing the whole world (through GPS) would help
a lot, unluckily it cannot be done indoors or under ground,
so one just have to simulate the unsynchronized cases too.
(OFDMA will do a lot in this field)
But just as the old resonator, diode and capacitor is a very
good device for (non-coherent) FSK,etc detection,
present GSM is a very practical, good solution to everything
except multipath below 5us, but then one can
always switch to another carrier frequency, just like
the dream of shannon-optimal OFDMA would have to
do.
That is, splitting the WCDMA 5Mhz bandwidth into
1.25Mhz subcarriers is just as smart as splitting
5Mhz into 200kHz bands... the old question of
time, frequency or antenna diversity, then getting into
adaptive antenna arrays,etc..
But I think it will take some time before a regular
handset uses antenna diversity on carrier frequencies
below 500Mhz, and a lot of quantum DSP to do it
above 2-5Mhz (the task of calculating the
quarter wavelengths left as an excercise, I have
problems with feet and thumbs, meters and um)
- politics..
Better not get into that one, except that some gurus
claim that the goals as follows
- 1900 end of colonialization, universal suffrage, education
- 2000 to build a "common sense", "consensus",
"real politics", "center based", "representative"
multi-party system.
Finland, with one of the most modern constitutions in
the industrialized west, luckily has that already, but
it is (extremely) much more difficult to implement for a
larger population like EU..
Ilmarinen
P.S. Or like China, India or USA.
P.P.S. Synchronizing a fast moving handset is also
interesting.. |
