[tea1990]

Hello Crackberry users,

I have a very weird problem with my Blackberry Z10 (STL100-1) running 10.3.1. I came to first notice the problem when my home WLAN was unstable, so the phone keep switching back and forth from Wifi to Mobile Data connection. This is where my phone's GPS starts behaving abnormally.

Normally, when using Wifi, I would assume that GPS location would be greatly improved, at least that's what my experience with iOS and Android told me. But the problem now, my GPS would show the correct location if use mobile data, but as soon as I turn on Wifi, it would show my office's location. This doesn't happen with my other phone (Android). Well I am not sure what happened but on daily routine, I only get Wifi connection at home or at the office, if that helps somehow to figure out the real problem here. I tried different apps to see whether it makes any different, like Blackberry Maps, Google Maps, Foursquare as well as Compass, and every app shows the exact same symptom - as soon as Wifi is on, it would go back to my office doesn't matter where I am at the moment (I'm testing this at home now).

Everything works the way it should on mobile data however. I'm really puzzled. I already tried to reset Map settings and restarting the phone etc, but the problem persist. Is there anyone who knows what happened here? Any help would be much appreciated.

Cheers.

[sk8er_tor]

With mobile network, your phone is able to use the mobile network co-ordinates to assist in aided GPS. When you're on Wi-Fi, I assume this is not possible and therefore you would need a GPS lock and cannot use the aided GPS.
iOS and Android store your Wi-Fi's location in their database... recall the big fiasco a few years ago with the uproar of some iOS users because Apple was doing this without a user's consent. I'm quite sure BlackBerry doesn't do this, which is why you're having a problem when on Wi-Fi. I would bet that if you just went outside or near a window to obtain a GPS lock when you're on Wi-Fi, you'd be fine.

[scrannel]

You should be able to use wifi navigation just fine with your Z10. I have no issues. Possibly your phone has cached your location and can't "clear it's head". I had this problem when I came back from overseas once and my wifi/network (non real GPS) kept telling me I was still in Qatar.

Try loading an app called "GPS Status" (can be found on BBerry World). Select more (three dots lower right) > tools > Manage a GPS State, Then press "reset" to clear cached GPS data. See if that helps.

[tea1990]

sk8er_tor: You sir are a genius! That was spot on. I wonder why I didn't think of doing that. Yes correct, it works just as you said it would. Thank you. Problem solved.

scrannel: I haven't tried that. But getting a clear view of the sky solved the problem. Thank you.


P.S. I made sure I am still connected to Wifi while getting outside.

[scrannel]

By using GPS + wifi you probably cleared your cache. You do not need real GPS to navigate by cell tower (data) nor wifi (difficult anyway as you would need tro move from router to router), but in marginal areas, doing what you did is sometimes needed. As I'm sure you know, you need cell data or wifi for maps to download to your phone as you move. GPS cannot download maps. It can only tell your phone where it is in coordinates. (Latitude and longitude). In other words, you should not need real GPS to always help your phone.

[Richard Buckley]

As sk8t_tor alluded to Wi-Fi and cell site location use the "known" location of the cell tower or hotspot and assume you are at that location. This is OK if the location is correct and you are on Wi-Fi or in an urban location where cell towers are densely packed. If the location is wrong or you are in the country the position can be very wrong.

Your experience with iOS and Android won't transfer to BlackBerry in this case. All Apple mobile devices and many Android devices have very poor GPS hardware. That is why iOS begs us to leave Wi-Fi on for navigation and why pilots who use iPads for navigation usually buy an external GPS.

But sometimes the GPS signal can be blocked by surroundings. That is why I wrote HelloGPS. Check it out, it is free from BlackBerry World.

You can find it at http://appworld.blackberry.com/webstore/content/128442




Z10STL100-3/10.3.1.2708 SR 10.3.1.1865

[scrannel]

R Buckley, would you explain something else. I presume that if one uses wifi for navigation, then one is using a "triangulation" between all "seen" wifi locations regardless of whether or not one is able to logon to any of them? Otherwise do not see how navigating with wifi makes any sense in the real world.

[Richard Buckley]

R Buckley, would you explain something else. I presume that if one uses wifi for navigation, then one is using a "triangulation" between all "seen" wifi locations regardless of whether or not one is able to logon to any of them? Otherwise do not see how navigating with wifi makes any sense in the real world.

Happy to. The short answer is no. The TL;DR is that there is really no practical way to determin distance between a transmitter and receiver unless the radio protocol connecting them is specifically designed for the purpose. GNSS signals are so designed, but even GPS and GLONASS have difficulties in many situations.

This does not prevent people from attempting it. Back in the original analog cell phone days (AMPS) before small GPS chips were practical phone companies tried to do this. It can give better results than just assuming the phone is co-located with the cell tower, but not to the extent that TV and movies would have you believe. Current cellphone signalling measures two parameters that sound like they could be use to triangulate a position: received signal strength and signal propigation time from the cell tower to the handset. These are actually used by the protocol for quality of service and throughput, more on that later. To use either of these values for location one must assume that the radio path between the tower and handset is fixed over some period of time, and straight. In actual use this almost never happens. Cellphones move, and even when stationary things around them move. The radio signal almost never travels a straight path but is reflecting, refracting or diffracting around some obstacles and penetrating others. This actually causes the received signal strength and propagation time to vary in time frames on the order of milliseconds. Interestingly this is why we have HSPA and HSPA+.

The 3G protocol UMTS is throughput limited by the amount of noise in the cell. Even though it uses code division multiple access W-CDMA and more than one transmitter can use the frequency at a time, in practice having too many transmitting with high power increases the noise and reduces the range of the cell tower. As more phones use the cell the size of the cell actually decreases. This is known as cell breathing. This happens becuase UMTS was designed to give constant quality of service at the millisecond level. At the time it was designed voice calls were still the dominant product offered by carriers and protecting the quality of the call was seen to be more important. So the cell is constantly adjusting its power, and commanding phones to adjust their power to keep everyone as quiet as possible. This also makes triangulation more difficult since the phone never knows how much power the cell is using at any given time. Over time though, engineers noticed that cells could never reach the theoretical maximum throughput. They had wasted, unused capacity. Before they could reach that level so many phones would be using it that cell breathing would cause the range of the cell to drop and the more distant phones would switch to neighboring cells. At this time carriers also noted that data was a far more lucrative product, and with data aggregate bandwidth over a period of seconds is more important than consistent quality over milliseconds. HSPA uses the same physical technology as UTMS but operates it differently.

With HSPA rather than trying to service all the phones continuously the cell determines which phones have a good signal path and which don't. The cell then puts all its effort into getting as much data to the phones with a good path, even if it has to raise power, shouting in effect. In a few milliseconds a different set of phones will have a good path and the cell will "shout" at them for a while. This gives consumers higher bandwidth, and allows carriers to run their cells at near theoretical maximum throughput. Everyone is happier.

Finally Wi-Fi and LTE use similar technology that is different again from UMTS but no more friendly to location determination. These system use OFDM almost always in combination - certainly in the case of LTE towers and any modern Wi-Fi router or cellphone - with MIMO. MIMO uses multiple antennas to direct the radio signal in the best direction. With OFDM these multiple antenns can also be used to combine signals from multiple directions or received over different times into one stronger signal before it is demodulated. All of this accounts for the better throughput and longer range we see, but makes using the signals for location all but impossible. Before you think that MIMO could be used to locate phones because the signal is being aimed at them, this is not very often the case. Think of trying to find a friend (or enemy) in a maze of mirrors. The direction you see them in might actually be the opposite direction they actually are. The same happens with MIMO. The transmitter doesn't really know or care which direction the receiver is in, only which direction to send the most energy to get the signal to them.

Its all done with very sophisticated maths.

[scrannel]

Thanks RB, but was primarily interested in routers being used in navigation -- alone. I recall (but now cannot locate) a paper describing urban navigation using router positioning. In other words, in an urban environment, there are a large number of routers (most of which one cannot access) but IF each one's position could be fixed, would allow for a "GPS LIKE" navigation. Of course, one would have to use off-line maps to actually navigate. The idea being if one finds oneself in a foreign country where data downloads are expensive, one could use router locations for navigation.

Now, this was an "exercise". Naturally one would normally use off-line maps with genuine GPS.

On the other hand here's an interesting paper that hits on WPS (wifi positioning systems) (and MESH). Plus, love the Mall navigation system.

[Richard Buckley]

Thanks RB, but was primarily interested in routers being used in navigation -- alone. I recall (but now cannot locate) a paper describing urban navigation using router positioning. In other words, in an urban environment, there are a large number of routers (most of which one cannot access) but IF each one's position could be fixed, would allow for a "GPS LIKE" navigation. Of course, one would have to use off-line maps to actually navigate. The idea being if one finds oneself in a foreign country where data downloads are expensive, one could use router locations for navigation.

Now, this was an "exercise". Naturally one would normally use off-line maps with genuine GPS.

On the other hand here's an interesting paper that hits on WPS (wifi positioning systems) (and MESH). Plus, love the Mall navigation system.

Well they can't be used as an independent navigation system. First some system has to geolocate each router and make that information available. Ignoring for the moment how one could build any assurance agains malicious manipulation, the hotspot universe is organically maliable. Any system would have to allow for routers to come and go, change BSSID, change MAC, radiation pattern (businesses often have their hotspots shapped to limit the ability of neighbors leaching from them). So the set of routers by itself can't be used for navigation, some other system has to overlay some sort of data set that, if nothing else, provides location data on each router. We have such datasets now, but they generally require network access to use. When faced with the problem of accuracy verification, sparse coverage or combinations of all a common solution is multi-sensor integration. To show how powerful this is, there was a recent paper published that describes using accelerometer data to determing phone location by comparing how the phone is moving with know subway train operating methods and routes. My car GPS navigator can determine if I've taken an exit before the two (or more) routes I could have taken diverge enough for reliable descrimination by GPS by monitoring direction of travel and accelleration. The problem is not much different than fitting a curve to data point you may have done in highscool math or science class.

If you find the paper again I would be interested in reading it, but you will probably find, perhaps by reading between the lines, that they are either talking about quite course positioning, or they are layering on some other sensor data or pre-analysis data set.