I am attempting to use location accuracy of kCLLocationAccuracyHundredMeters for my standard location updates in my application, which seem to work relatively well for a bit, but every once in a while I will receive a 10 minute gap between updates (while driving around multiple miles in different directions) which seems out of character for the setting. Here are the settings on my location manager. Also I am only using foreground location updates so for these tests I had the screen on the entire time and app always in foreground.
self.locManager = [[CLLocationManager alloc] init];
self.locManager.delegate = self;
self.locManager.desiredAccuracy = kCLLocationAccuracyHundredMeters;
self.locManager.distanceFilter = 15;
I tried out kCLLocationAccuracyBest with the same other settings and I get perfect measurements throughout my drive, but I don't need that accurate of a location. I assumed the hundred meter accuracy would still give relatively consistent updates with less accuracy, which is what we are looking for.
Is this just a side effect of working with the kCLLocationAccuracyHundredMeters setting vs kCLLocationAccuracyBest?
Tested on an iPhone 6s : 11.0.2 and iPhone 7Plus : 11.0.3
Any help is appreciated, thanks!
EDIT:
I should clarify I did setup logging in the CLLocationManager delegate methods listed below, but never received any callbacks / logs.
- (void)locationManager:(CLLocationManager *)manager didFailWithError:(NSError *)error;
- (void)locationManager:(CLLocationManager *)manager didFinishDeferredUpdatesWithError:(nullable NSError *)error;
- (void)locationManagerDidPauseLocationUpdates:(CLLocationManager *)manager;
- (void)locationManagerDidResumeLocationUpdates:(CLLocationManager *)manager;
When you ask for kCLLocationAccuracyHundredMeters iPhone does not use GPS at all and relies on cell towers and wifi access points for triangulation. That means iPhone updates location with the frequency of wifi scans, which is far from continuous. If you are getting away from areas with wifi coverage, cell towers become your only source and you might start getting location updates only when switching between cell towers.
To cut it short:
setting accuracy tells iOS which hardware to use,
distance filter tells iOS how far to move before waking up CPU with new data
Can somebody help me how to solve the issue.
My current beacon setup :
Beacon device - Estimote.
UUID - same for the two beacons.
I have placed each beacon in two rooms, rooms are 20-30 Meters away from each other
I'm hitting webservice and updating something (when the proximity is near or immediate) when i enter a specific room (checking if major and minor of this room actually belongs to this room)
3.It also hits another webservice and updating something (when the proximity is far and unknown).
Problem with the current algorithm:
1. When i enter room 1 and if it is near or immediate, it hits webservice1. When this happens, room2 proximity will be far or unknown, then again webservice2 is hitting and also there are lot of fluctuation. Sometimes when i'm in room1, room2 beacon proximity shows as near. I don't this to happen
New Approach and Questions
I would like to Monitor for specific region , say room1 when i'm near say 5 meters away, and start ranging the beacon only when proximity is near or immediate. and stop ranging when proximity is far or unknown or i'm more than 5 meters away
When i move near to room2 (5 meters away) and if the proximity of room2 is near or immediate, it should start ranging room2 and stop ranging room1
I would like to have clear approach or way to write algorithm to achieve the above. Calculation can be using the proximity or using some other idea (like combination of RSSI and proximity or something else)
It would be great if you can explain me in detail with example on how to achieve the new approach or you can tell me if there is any other better approach, but please explain with code.
My foremost concern is, it should't range room 2 beacon when i'm in room1 or just 5 meter away from room1 and vice versa.
Have you thought about putting each beacon in its own region? So room 1 would be a region, and room 2 would be another region.
After that I would make use of locationManager:didEnterRegion: and locationManager:didExitRegion:. Within these, you can start and stop the ranging of your two beacon regions depending on which room you are in.
Please note, the Apple Docs state that:
When testing your region monitoring code in iOS Simulator or on a device, realize that region events may not happen immediately after a region boundary is crossed. To prevent spurious notifications, iOS doesn’t deliver region notifications until certain threshold conditions are met. Specifically, the user’s location must cross the region boundary, move away from the boundary by a minimum distance, and remain at that minimum distance for at least 20 seconds before the notifications are reported.
The specific threshold distances are determined by the hardware and
the location technologies that are currently available. For example,
if Wi-Fi is disabled, region monitoring is significantly less
accurate. However, for testing purposes, you can assume that the
minimum distance is approximately 200 meters.
EXAMPLE:
- (void)locationManager:(CLLocationManager *)manager didEnterRegion:(CLRegion *)region
{
if (region.identifier == YOUR ROOM 1 IDENTIFIER)
{
//Start ranging room 1
}
else
{
//Start ranging room 2
}
}
- (void)locationManager:(CLLocationManager *)manager
didExitRegion:(CLRegion *)region
{
if (region.identifier == YOUR ROOM 1 IDENTIFIER)
{
//Stop ranging room 1
}
else
{
//Stop ranging room 2
}
}
The core problem is sending info to the server at inappropriate times. It really has nothing to do with starting ranging.
You cannot start ranging when proximity is relatively close -- you have to range in order to determine proximity.
Rather than an algorithm that starts and stops ranging, I suggest you simply leave it on all the time, and build a filter that decides when to make the server calls. Here's a starting point for the logic I would suggest:
Call the server if all the following are true:
beacon.accuracy < 3
all other beacons detected have beacon.accuracy > 3
no server call has been made for this beacon recently
EDIT: You can reduce the fluctuation on the accuracy reading by setting your beacon transmitter power as high as possible to maximize the signal to noise ratio.
I am currently trying to get my app to monitor particular regions using CoreLocation however I am finding that it does not seem to work as expected, it seems to me that it cannot work with small a small radius set for each location i.e. 10m.
I've also put together a little test app which plots the circle radius on a map so I can visually see what is happening.
The code I am using for monitoring locations is as follows:
self.locationManager = [[CLLocationManager alloc] init];
self.locationManager.delegate = self;
self.locationManager.desiredAccuracy = kCLLocationAccuracyBest;
// Set-up a region
CLLocationDegrees latitude = 52.64915;
CLLocationDegrees longitude = -1.1506367;
CLLocationCoordinate2D centerCoordinate = CLLocationCoordinate2DMake(latitude, longitude);
CLCircularRegion *region = [[CLCircularRegion alloc] initWithCenter:centerCoordinate
radius:10 // Metres
identifier:#"testLocation"];
[self.locationManager startMonitoringForRegion:region];
I have not put up the code here for DidEnter region etc as I know that works when I go over 100m away from the monitored region.
Here is a screen shot of the app when I am well over 10 meters away from the purple location on the map, the did exit region events do not fire, however if I switch my location to London it fires and also when I set my location back to where the blue location is currently it also fires.
Does anyone know if there is a limitation with the minimum region radius at all or perhaps I am doing something wrong.
Thanks Aaron
I don't think region monitoring will work well for such a small radius.
The best accuracy with the GPS chip and kCLLocationAccuracyBestForNavigation is often just 10 meters.
Apple says (in the Location & Maps PG) that the minimum distance for regions should be assumed to be 200m
I've heard that region monitoring is using WiFi to get it's position (which makes sense for power savings). WiFi accuracy is more like 20m-100m. I'm not sure how having another app using background location (i.e. using GPS) would affect this. Probably, the location manager would share information to make the accuracy better.
Region monitoring can take 30 seconds to fire once inside a region, and a couple of minutes to fire after leaving a region (to prevent location glitches from triggering it).
When region-monitoring was first introduced, they said that it would only work with 100m regions and anything smaller would be bumped up. This probably still happens.
There's a deprecated method startMonitoringForRegion:desiredAccuracy: which allowed you to specify the distance past the region border to start generating notifications. Presumably this feature has been rolled into startMonitoringForRegion: but is still there. A 10m region might end up with a 10m buffer.
If you want to do this, specify a larger region around where you want to monitor, and when the device wakes up in that region, start background location updates (GPS) and use CLCircularRegion's -containsCoordinate: to trigger when the device is within 10m manually. This method is officially sanctioned by Apple (see WWDC 2013 Session 307).
From the CLCircularRegion docs:
Remember that the location manager does not generate notifications immediately upon crossing a region boundary. Instead, it applies time and distance criteria to ensure that the crossing was intended and should genuinely trigger a notification. So choose a center point and radius that are appropriate and give you enough time to alert the user.
From the Location & Maps PG:
Region events may not happen immediately after a region boundary is crossed. To prevent spurious notifications, iOS doesn’t deliver region notifications until certain threshold conditions are met. Specifically, the user’s location must cross the region boundary, move away from the boundary by a minimum distance, and remain at that minimum distance for at least 20 seconds before the notifications are reported.
The specific threshold distances are determined by the hardware and the location technologies that are currently available. For example, if Wi-Fi is disabled, region monitoring is significantly less accurate. However, for testing purposes, you can assume that the minimum distance is approximately 200 meters.
There's further inside scoop from this post by Kevin McMahon, who asked the Core Location engineers about region monitoring at a lab at WWDC 2012. This info will have changed in the meantime, but the part about region categories is interesting. Here's an edit:
Fine Region (0 - 150m)
- With the floor of 100m this category's range is effectively 100-150m.
- For regions this size performance is heavily dependent on the location-related hardware
- The amount of time that it takes Core Location to detect and call the appropriate delegate method is roughly 2-3 minutes on average after the region boundary has been crossed.
- Some developers have figured out independently that smaller regions would see quicker callbacks and would cluster smaller regions to cover one large area to improve region crossing notifications.
This seems to be a bug in CLLocationManager. I've done extensive testing using various region radius configurations and locationManager:didExitRegion does not fire in an expected way. This seems to be either a rather nasty bug or region monitoring does not happen at all like the documentation suggests. I have the test harness available to anyone who wants it:
http://www.mediafire.com/download/x863zkttltyalk6/LocationTest.zip
Run it in the simulator and start the test by by selecting Debug -> Location -> Freeway Drive in the iOS simulator menu. The number you see is the distance from the center of the monitored region. The background color will be green while the device is within the monitored region and red when outside the region. The text below the distance are event logs.
After running the app, you should see locationManager:didExitRegion fire at 5319 meters from the monitored region. The route will loop every 37 minutes and you'll see the device exiting the region always at 5319 meters.
I've submitted a radar with Apple (17064346). I'll update this answer once I hear back from them. At least then we'll have some input from the canonical source.
Here's the detailed text sent to Apple:
Using a test app on the iOS simulator as well as on an iPhone 5S the
CLLocationManager doesn't seem to fire didExitRegion callbacks in an
expected way. Regardless of the radius of the circular region being
monitored, the callback won't happen until a threshold of around 5000
meters is hit.
Steps to Reproduce:
1. Run the attached app
2. Start region tracking by selecting Debug -> Location -> Freeway Drive in the iOS simulator
3. Monitor the app. The large # indicates the distance from the center of the watched region.
4. After about 190 seconds and 5300 meters didExitRegion will finally fire.
Ths issue does not seem to be related at all to the size of the region. According to the Apple docs, even small regions are supported:
In iOS 6, regions with a radius between 1 and 400 meters work better
on iPhone 4S or later devices. (In iOS 5, regions with a radius
between 1 and 150 meters work better on iPhone 4S and later devices.)
On these devices, an app can expect to receive the appropriate region
entered or region exited notification within 3 to 5 minutes on
average, if not sooner.
Although region events don't happen instantaneously, they should happen fairly quickly. From the Apple docs:
Region events may not happen immediately after a region boundary is
crossed. To prevent spurious notifications, iOS doesn’t deliver region
notifications until certain threshold conditions are met.
Specifically, the user’s location must cross the region boundary, move
away from the boundary by a minimum distance, and remain at that
minimum distance for at least 20 seconds before the notifications are
reported.
This is not at all what I am seeing in the test harness. On the simulator the device will always be 5000+ meters away from the region before a locationManager:didExitRegion event occurs.
I like the answers by both Michael and Nevan. I would like to add more information from my personal experience/opinion in developing Location Based iOS Application with Region Monitoring and also highlight some important points:-
Be realistic on Region Monitoring
Region Monitoring is using Global Positioning System (GPS), Wifi and other technologies to determine if the device is inside or outside the monitored region. Don't forget that our earth is 510 square kilometers and about 30% are land (149 million km2). It is a huge area. Remember the recent MH370 missing case? Our current most advance technology could not even pinpoint an estimated region of that missing plane.
If you want to monitor for a small region with only 10 meter radius. It could possibly work inside a highly dense city with a lot of cell towers and wifi connected areas. But at the same time, the signal might be blocked by high rise towers which might cause the signal loss for a few seconds/minutes which caused the delay in delivering the notification.
So, you really have to consider the above information before deciding how big is the region that you want to monitor. Personally I think 10 meter radius is too small.
Be Realistic on the Number of Monitored Regions
The current Core Location technology can only monitors up to maximum 20 regions on a single app. Make sure that the monitored regions are not too close to each other as well.
I personally have tested 3 regions that are about 100 meters in radius which are about 200 meters aways from each others. Sometimes I can get notifications from all these 3 regions when I am driving through them, but sometimes, I can only get the notification from the First region only. What could be the reason? I could not know. The regions might be too close to each other. Or the cell towers decide that my device does not actually inside the monitored region.
There was one person on StackOverFlow who wants to monitor 1800 points on our Earth. Don't be like him as he is quite unrealistic and probably does not understand the limitation of current Core Location technology. Link: Check if the user location is near of some points
Fine Tune The LocationManager
If your app needs to monitor a small area or needs the location update frequently. Here are the potential properties of your location Manager.
self.locationManager.desiredAccuracy = kCLLocationAccuracyBestForNavigation;
self.locationManager.distanceFilter = kCLDistanceFilterNone;
self.locationManager.activityType = CLActivityTypeAutomotiveNavigation;
kCLLocationAccuracyBestForNavigation will consume more battery compare with kCLLocationAccuracyBest. But, it will be more accurate.
I found a glitch in region monitoring in iOS 7 when there are multiple notifications triggered at the same time in different monitored regions. I have found a solution to filter out this glitch. For more info, please visit: Region Monitoring Glitch on iOS 7 - Multiple Notifications at the same time
Don't Be Over Ambitious
You might have used some apps which can monitor a small region and are very accurate and able to notify your the same second you step into the region. And you have the inspiration to develop the exact same app to compete with them. But do you understand what happens behind scene? What additional technologies that they are using? And what partners that they are collaborating with?
I have done some research on that and found out that some of the technologies that they use are not available publicly. Some of those companies are heavily funded and could pay a premium to the telecommunication companies in order to get the best location accuracy for the best user experience. I do not understand the details on how it works. I believe most of the location determination is actually on the server end (back end), not the mobile (front end).
So, the apps that are developed by those companies not only can pinpoint the best accurate location but also does not consume a lot of battery.
NOTE: I just want to share my 2 cents. The above information consists of my experience and personal opinion. It might not be 100% accurate as I am still learning Core Location and Region Monitoring.
I do agree with Michael G. Emmons, and want to share my experience too:
I tested my code with three regions as shown in the image below:
Explaining the behaviour:
My current location is Region-1, and i start monitoring for the above
three regions, and call to requestStateForRegion, to determine, if there is any region inside, where i am currenly standing.
Then i get "Enter" notifications, for first two region (region-1, and region 2), but it should only detect the region-1.
Now when i enter in region-2, i get the Enter notification for region-3. but i should get the notification for region-2 here.
Now when i enter in region-1 again, i get the Exit event fired for the region-3, and this continues.
but i don't get any Enter/Exit events for first two regions, until i move at-least more than 7Km-10Km far from first two regions.
Expected Behaviour:
- Enter/Exit event should be triggered only when i am crossing the boundary of regions, or inside the regions, not before 500 meter from the region.
My Assumption:
What i have noticed after all the experiment, that when i call
"requestStateForRegion" for all three regions,
it detects all regions inside region of radius 5000m, thats why it
detects first two regions at the same time (region-1 create a circle
of 5000m radius, and region-2 comes in its range, thats why region -2
is also getting detected).
and when user moves far more than 10Km, their Exit events will be called and when user comes back in these regions, their Enter event will be fired. Its the same case as explained by Aaron Wardle above.
Region-3 is getting detected, because,when user enters in region-1, ie. 8-9km far from the region-3, so Exit event is fired for this, and when user is on the route for region-2, here even when region-3 is 5000 meters far, still it detects the region-3 and fire, Enter event for region-3.
So i think that all the regions inside 5000 meters are being detected, and as user moves away 10 km from detected region, its Exit event will be fired. otherwise if user is inside the 5Km range, it will never call it Enter/Exit events again.
Please update me on, if anyone has fixed this issue, or Apple documents anywhere about this issue.
Based on #Nevan's answer, which indicated some sort of coverage in WWDC 2013 307 (which didn't directly address this), I came up with a reasonable solution to getting < 10m accuracy for the arrival to a location, though I have a feeling that implementing -(void)locationManager:didVisit: might make this more battery-conservative, but would provide less frequent updates.
First, have some regions with 0..150m radius, and start monitoring. Doesn't really matter, as the system seems to trigger these at around 150~200m:
_locationManager = [[CLLocationManager alloc] init];
_locationManager.delegate = self;
CLCircularRegion *region = [[CLCircularRegion alloc] initWithCenter:CLLocationCoordinate2DMake(location.lat, location.lng) radius:50 identifier:location.name];
[_locationManager startMonitoringForRegion:region];
Then, implement
-(void)locationManager:(CLLocationManager *)manager didEnterRegion:(CLRegion *)region {
for (CLCircularRegion *enteredRegion in _locationManager.monitoredRegions.allObjects) {
if ([enteredRegion.identifier isEqualToString:region.identifier]) {
self.locationManager.activityType = CLActivityTypeFitness;
self.locationManager.distanceFilter = 5;
[self.locationManager startUpdatingLocation];
break;
}
}
}
The system will start monitoring and reporting to your delegate a stream of locations, even if your app is suspended (need UIBackgroundModes to include location array element).
To check if one of those locations is within the centre of one of your regions, Implement:
-(void)locationManager:(CLLocationManager *)manager didUpdateLocations:(NSArray<CLLocation *> *)locations {
CLLocation *firstLocation = [locations firstObject];
CGFloat const DESIRED_RADIUS = 10.0;
CLCircularRegion *circularRegion = [[CLCircularRegion alloc] initWithCenter:firstLocation.coordinate radius:DESIRED_RADIUS identifier:#"radiusCheck"];
for (CLCircularRegion *enteredRegion in _locationManager.monitoredRegions.allObjects) {
if ([circularRegion containsCoordinate:enteredRegion.center]) {
[_locationManager stopUpdatingLocation];
NSLog(#"You are within %# of %#, #(DESIRED_RADIUS), enteredRegion.identifier);
break;
} else if ([enteredRegion containsCoordinate:circularRegion.center]) {
NSLog(#"You are within the region, but not yet %#m from %#", #(DESIRED_RADIUS), enteredRegion.identifier);
}
}
}
You'll also want to implement:
-(void)locationManager:(CLLocationManager *)manager didExitRegion:(CLRegion *)region {
[_locationManager stopUpdatingLocation];
}
This is more like an important comment. From Region Monitoring and iBeacon
Testing an iOS App’s Region Monitoring Support
When testing your region monitoring code in iOS Simulator or on a
device, realize that region events may not happen immediately after a
region boundary is crossed. To prevent spurious notifications, iOS
doesn’t deliver region notifications until certain threshold
conditions are met. Specifically, the user’s location must cross the
region boundary, move away from the boundary by a minimum distance,
and remain at that minimum distance for at least 20 seconds before the
notifications are reported.
The specific threshold distances are determined by the hardware and
the location technologies that are currently available. For example,
if Wi-Fi is disabled, region monitoring is significantly less
accurate. However, for testing purposes, you can assume that the
minimum distance is approximately 200 meters.
Sounds like even 1 meter should work (and work better on iPhone 4S+ devices):
startMonitoringForRegion:
(...)
In iOS 6, regions with a radius between 1 and 400 meters work better on iPhone 4S or later devices. (In iOS 5, regions with a radius between 1 and 150 meters work better on iPhone 4S and later devices.) On these devices, an app can expect to receive the appropriate region entered or region exited notification within 3 to 5 minutes on average, if not sooner.
In the past few days iv'e been testing a geofencing feature on my iOS 8.1 device (iPhone 5S) for an app iv'e developed.
The app is registering few regions to the iOS gefence service. The app's logic needs that each geofence radius is between 40 to 80 meters.
I'm seeing so far that in areas with larger number of cell towers and Wifi hot-spots, the geofence detection is good enough on entering regions. That is, in down town areas, business areas etc' the geofence detection is working fine.
Unfortunately, the opposite occurs in areas with few cell towers & wifi networks. My neighborhood, for example, is about 1000 meters width and 500 height (1KM x 0.5KM), and there are no cell towers in it. There are few cell towers thought, on the perimeter that surrounds the neighborhood. Unfortunately In the perimeter of the neighborhood the geofence service detects nothing.
Needless to say that i'm testing with Wifi enabled on the device.
When i test my app on Android: the geofencing service on android 4.3, 4.4 & 5.1 works much better than on iOS. The Android's geofencing service does not detect 100% of region transitions, however it detects 50%-90% of the region transitions.
I conclude the following: If there would have been more cell towers & Wifi hot-spots & if Apple would have improved the geofence service then the detection on iOS devices would have been as good as in Android's.
Geofencing works by detecting a user moving from one cell network tower to another cell network tower.
Therefore smallest area you can define is dictated by how close together the cell towers are.
Inside a shopping mall or sports stadium, it might be able to do 10 metres — cell towers are often extremely close together. In a regional area anything smaller than 100km can fail.
If you need smaller areas, you need to use bluetooth instead of cell towers (iBeacons). If there is a bluetooth low energy device in the target area you can set the range to very short (centimetres) or reasonably large (up to 30 metres or so). Note this all depends on the quality of the iBeacon hardware, some are better than others.
Unfortunately bluetooth (version 4.0 or newer) and cell network towers are the only way to monitor locations without significantly draining battery. Keeping the GPS active to check for a 10 metre boundary would drain the battery from full to completely flat in about 2 hours even with the screen switched off.
i use CLLocation for a app that record people's trace in map view when they are running or walking ,but i found when my device is still (not moved) ,the
- (void)locationManager:(CLLocationManager *)manager didUpdateLocations:(NSArray *)locations is also get called frequently ?
currently ,my locationmanager's desiredAccuracy is 10 meters and distanceFilter is 10.
how to deal with this situation? I have tried use big distancefilter value(like 150) ,but I found if i do this, then i can't record exactly when people is running or walking
GPS is not exact. You can move around a few feet and get the same location. Or you can sit still and get told you have moved a few feet.
You might be able to combine measurements from the accelerometer to determine if you have really moved but this would only work if the device was sitting on a table not moving.
Have you called stopUpdatingLocation after the initial startUpdatingLocation? It will keep updating location if you do not call it.
How does system know whether or not you have actually moved? It MUST fetch your location to find out. The more accurate your desired accuracy is, the more vigorously and frequently it would look. By vigorous I mean if it's suppose to use cell-tower information then it would look into more cell-towers to better triangulate. Or simply put it, the interval between its fetches would be smaller. Concluding: OS would fetch data even if you don't move.
Additionally to triangulate your position the OS (depending on your desiredAccuracy and previous movements) would use a mix of GPS, wifi, Cell-tower. And because someone may all of sudden turn off/on their wifi, or satellite that you were using to get your location has moved a bit or the satellites have changed, or a cell-tower signal may become more or less accurate due to its bandwidth limitations then your calculated location may change which triggers a callback if it's more than your distanceFilter. ( I don't believe you get callbacks for less than your distanceFilter, but I may be wrong) This likely means your distanceFilter is set to very small number which depending on your business requirements may or may not be a good choice. Concluding: your location is never ever 100% accurate
The result of periodical fetching, possibility of error and small distanceFilters lead to possible incorrect locationChanges.