Mouse with scraper, to scrape off sweat spots.

ChairmanOfTheBored hath wroth:

Now we know your overload threshold. About 200 lines of careful analysis, detailed refutations, and personal suggestions is required. Predictably, you retreat to a safe topic which you allegedly know something about. Amazingly, you offer no accusations of incompetence or ineptitude, so I'll presume my status has been elevated from "idiot" to something more acceptable. I'll be expecting your apology shortly.

You're absolutely correct about cellular tracking as far as it goes. Given sufficient technology, motivation, and funding, such things are possible. In most cases, the location needs to be sufficient to perform an apprehension. That's often difficult with some of the existing technology.

For example, handset based GPS and AGPS (augmented GPS) rely on the ability of the GPS engine to supply sufficient satellite delays for the location services provider to calculate a location. The GPS data is augmented by latency data from the cell towers. Under ideal and FCC test conditions, it works as advertised. Under an ever growing number of conditions, it sucks. GPS barely works indoors and in some "shielded" vehicles. Acquisition time is decreasing with newer and smarter chipsets, but still requires more operating time than the typical cell phone battery can comfortably provide. Even the PSAP (public safety answering points) is hesitant to MANUALLY request a position report due to the power consumption problem. Tracking someone around town with regular will certainly show up as a major battery drain and unusual transmission rate, which is easily detected by a personal RF detector, or one of those antennas with the LED inside.

The indoor GPS problem is bad enough to support a startup.

Rosum uses TV transmitters to triangulate a location in areas where GPS sucks.

The network based location technologies are even worse. Most use some variation on TDOA (time difference of arrival) technology, where the latency between the handset and more than one cell tower yields the location. Two towers will give an ambiguous location. Three towers will provide a non-ambiguous fix. The problem is finding a location where you have a simultaneous non-reflective path to 3 cell towers. That's possible in open flat areas but a rather dubious proposition in crowded cities.

Reverse engineering how James Kim was found should offer some clue as to capabilities and limitations.

There's also a problem with cell site capacity. Presumably, one of the three sites is being used for communications, while the other two are listening on the same channel (or time slot) for just location data. That's fine for following perhaps one or two cell phones around the city, but could overload the channel capacity of the cell site if used extensively. Certainly, the cellular companies are not particularly interested in installing extra hardware at their cell sites that do not generate revenue. So, they sell LBS (location based services) to pay for the extra hardware:

The accuracy of such systems are not particularly good. AGPS handset based systems must locate handsets within 150 meters 95 percent of the time, and within 50 meters about 67 percent of the time. Network based location systems are only expected to locate handsets within 300 meters 95 percent of the time, and within 150 meters 67 percent of time. Even that is problematic as the FCC had instigated a review of location accuracy and testing methods. It seems that the network providers decided that everyone with a handset must obviously be inside a vehicle and driving down a road. By using a database of known road locations, they simply rounded off the calculated position to the nearest roadway. The probably is adequate for most E911 applications, but can cause some rather interesting looking plots if one were trying to follow someone around town. Incidentally, the FCC also wants to extend the location requirements to VoIP providers.

I won't go into detail on countermeasures for fear of hitting your 200 line posting limit. Suffice to say that shielding the GPS antenna in the cell phone for handset based systems, and using a directional cellular antenna for network based systems, is sufficient to break either.

For the truly paranoid, there's also a problem with temporarily turning off the cell phone. For handset based systems, I've noticed that the initial handshake is now taking much longer than with earlier non-GPS handsets. My guess(tm) is that the handset is sending the last set of satellite delays to the cellular provider in order to reduce computation times. However, this is admittedly speculation.

Drivel: Cell Sites of Santa Cruz County:

I haven't done anything with this site since about 2001. Since then, the number of cell sites in the county has roughly doubled. Since I now have a new camera (Canon S5-IS), I'll probably expand the web site. Anyway, the plan was to provide coverage estimates and maps. For example (at bottom of page):

His head would explode, and it would take the Hazmat crew weeks to clean up the place. :(

F*ck off, s*****ad. Your bullshit E-1 grade assessments are invalid.

Nice job of ignoring the fact that you did the same thing too, you retarded f*ck.

You're a goddamned idiot.

Dumbfuck. Cellphones keep their last good reading and that gets reported if a polled reading is not being reported by the phone.

Bullshit. The CELL PHONE takes a reading, and that is what gets reported. ACTUAL COORDINATES, not some delay degraded data stream, you retarded twit.

As if a retarded f*ck like you even knows.

All vehicles are "shielded" to a certain degree, idiot. It all depends on where one places one's GPS antenna.

You're an idiot. It is a constant operation of the phone, and has a specific update period.

When are you going to stop making s*it up? Cell phones have had GPS capacity for over three years now, and they have most certainly tackled any power consumption issues.

You're an idiot. ANY cellphone that is turned on is CONSTANTLY in contact with the cell node tower that it is currently in communication with. Cell phones get date and time updates, and they transmit specific data from the phone to the tower on CONSTANT basis.

You're an idiot.

Not where... WHEN, and the WHEN is not as often as you would like to claim, dingledorf. If one can get a call indoors, one can be tracked to their location indoors. The GPS poll will merely report the last valid reading taken, and the cell tower triangulation will verify it, or correct for any difference from it.

ChairmanOfTheBored hath wroth:

Amazing. I guess it would then be trivial to display the Lat-Long on the cell phone without the assistance of a for-pay location service. Too bad there are only a few phones capeable of doing this. It seems that to get location information, you have to subscribe to a server based service. For example, Qualcomm has gpsOne, which has OPTIONAL locally calculated position reports.

There are 4 different modes that it can operate. See:

To the best of my knowledge, only the "Mobile-Assisted" mode has been implimented in deliverable handsets. Incidentally, Snaptrack is now owned by Qualcomm.

Nope. See:

"There are three basic types of data that the assistance server provides to the GPS receiver: precise GPS satellite orbit and clock information; initial position and time estimate; and for AGPS-only receivers, satellite selection, range, and range-rate information. The assistance server is also able to compute position solutions, leaving the GPS receiver with the sole job of collecting range measurements." "Collecting range measurements" means satellite delays.

A bit on a location server:

I've seen some measurements with various cellular handsets for GPS signal strength and quality under various conditions. Sitting on your belt or purse, while inside a vehicle, is actually fairly usable if you want to take long samples. Indoors, in a forest, or in an urban canyon full of reflections, GPS signal quality is really awful. Fortunately, the typical operating position and user location for talking to the cell tower, is also usually an adequate location for obtaining a location fix. With the aquisition time reduction provided by A-GPS, it can be done without draining the battery.

Nope. For normal E911 only operation, the receiver is powered off until the PSAP requests a fix. The location server provides the necessary to initialization data, the handset provides the satellite delays, the location server chews on everything and provides the PSAP with location data. It then turns off the GPS section until the next request.

SiRF GPS chips currently are down to about 250mw of power consumption. It takes about 1-2 seconds for an A-GPS initialization and an average of about 15 seconds to obtain a fix sufficient to meet FCC's accuracy specs. That won't kill a handset battery if used sparingly, but will rapidly deplete the typical handset battery if used for continuous tracking. Older implementations are not so efficient. Some sucked as much as 1 watt.

Sure, mostly because it's not being used much and then only if polled. However, you're correct that current GPS chips are far less power hungry and have a faster TTFF (time to first fix) than older handsets. As long as these modes of operation are being used, there's no power problem. However, if you're planning on tracking someone continuously, or polling excessively, as in advertiser supported LBS (location based services), you'll probably have a signifigant effect on battery life.

Amazing. How does a single cell tower triangulate a handset position? I thought it takes 3 to tango err... triangulate.

Incidentally, I live in a mountain forest, where GPS satellite visibility is rather poor. I just put my Samsung SCH-i700 into service mode to display the GPS data. Zilch. No birds in view. My Verizon XV6700 showed blank fields for GPS data, which means either no satellites or I screwed up somewhere (i.e. forgot to enable GPS receiver). I'll try it again when I go to my palatial office in a few days. Anyway, if you have a cell phone handy (does not need to be activated), with a GPS, I can probably find the magic test mode code so you can see for yourself what is being sent and whether it will work indoors at your location.

Jeff Liebermann hath wroth:

This might give you a clue as to what GPS data is available to send in the xv6700. This is the output of FIELDTEST.EXE. The GPS data fields are 4 pages long:

I haven't figured out how to enable the GPS receiver yet, so all the fields are currently blank. I'll see if I can find which fields are received by the Location Server, and which ones are returned for a calculated location solution.