I've been doing some reading on this subject. It seems that both GPS and tower triangulation methods suck. For GPS, the problems are signal acquisition and penetration in urban environments, especially with non-dedicated handheld devices. For tower triangulation, the problem appears to be areas with poor signal coverage where you might only be able to barely see one tower, and where TDoA, AoA, and EOTD aren't going to do you any good.
In either case, simply keeping the last known signal lock may very well be one of the worst things you could do.
It seems to me that we need to use both technologies in order to get any real hope of reasonably sustainable accuracy, either for E911 or any other location-aware technology. And I'm not convinced even that's enough.
So, anyone want to place any bets on what's really going to happen?
In either case, simply keeping the last known signal lock may
very well be one of the worst things you could do.
Depends on what you want to do with the location info. If you
want to immediately dispatch a vehicle, then you have to realize
that you may be sending one to the edge of the cell tower's range
when the caller is many miles away. Or that you might be sending
one to the east side of the downtown highrise district when the
caller has moved on to the west side of the downtown highrise
district.
On the other hand, maybe all you want to do is to route the
call to the right E911 center. In that case, as long as you
are in the right county you are probably OK.
In any case, no solution to E911 and VoIP is likely to meet
100% of its requirements, but if you can improve the situation
significantly, then it is still worth doing.
--Michael Dillon
I have never seen any real study by the emergency response services
on how many problems they actually have other than isolated worst-
cases and a lot of political rah-rah. In the end I expect that any
technically feasible improvement to the cell phone position accuracy
is miniscule to the actual effort and expenditures it requires. So
my guess is that the real drivers are the law enforcement agencies
wanting to get better tracking abilities. Whether they get out of
deal what they are hoping for remains to be seen. Not that they will
tell us anyway.
Actually, the FBI has been at least somewhat open with their disappointment of carrier support for position-enhanced information and compliance with CALEA. I've seen recent articles in the press that has made this obvious. It seems that the problem is that too many people are holding onto their old phones, and the networks which selected GPS as their solution aren't getting enough uptake fast enough on the new position-enhanced models.
Contrary to my previous post, Nextel appears to be one of the carriers that selected GPS, and Verizon and Sprint appear to have done the same. AT&T, Cingular, and T-Mobile appear to have gone the tower triangulation route.
But as far as E911 is concerned, the problem appears that many of the emergency services providers still aren't equipped with the necessary equipment -- the article at <http://tinyurl.com/dmzdq> is a little old, but the situation was so bad at the time that I can't imagine the entire world has been turned around since.
The whole CALEA/E911 issue was known a long time back. Even _Wired_ picked up on it in early 1998, in the article at <http://wired-vig.wired.com/news/technology/0,1282,9502,00.html>, or <http://www.danbrown.com/secrets/digital_fortress/cell_phones.html>. And the interaction between VOIP, E911, and CALEA is still getting some traction (see <http://www.vonmag.com/issue/2005/jul/features/you_will_conform.htm>).
Andre Oppermann wrote:
I have never seen any real study by the emergency response services
on how many problems they actually have other than isolated worst-
cases and a lot of political rah-rah. In the end I expect that any
technically feasible improvement to the cell phone position accuracy
is miniscule to the actual effort and expenditures it requires.
(putting on my firefighting helmet for a moment)
I don't have any studies, per se, but we get enough "the house next to XXXXX Any Street" calls as it is that the "technically feasible improvement" is an improvement.
In San Antonio, people "give directions" by intersections, and leave it up to the recipient to actually figure out where the destination is. "281 and Bitters" represents a ~10 square mile area to most locals, and similar scenarios pop up all over. I-10 runs "east and west" from El Paso to Houston through downtown. I-35 runs "north and south" from Dallas to Laredo. Loop 410 is a 54-mile loop around the city. Loop 1604 is a 110-mile loop around the city. Getting cell phone calls from a tourist (of which we have plenty) reporting a wreck at I-10&410 narrows it down, but still leaves ~27 miles of inaccuracy. Add in 4-8 exit ramps, 10 square miles of surrounding area, and a language barrier, and you end up with some combination of delayed response, parallel response to each possibility, and increased risk to rescuers and innocent citizens.
So yeah, I'd like better locations whenever possible.
pt
This is actually more important then it sounds. Not long ago I was
driving around in Northern New Hampshire on I93 and saw a situation I
believed should be reported to the police. I used my cell phone to dial
"*SP" (which I saw on many signs in Massachusetts claiming it was how
you called the State Police).
Well, someone answers "State Police" and I begin to describe where I am.
Much confusion results until I realize that I am talking to the
MASSACHUSETTS State Police even when in Northern New Hampshire!
-Jeff