Android Geofencing with Google Maps

A geofence is a virtual perimeter of interest that can be set up to fire notifications when it is entered or exited, or both. For example, a geofencing app can alert us that our kid has left a previously specified area, or send us a coupon (e.g. the “Present this SMS an get 20% off” offer type) when we happen to walk or drive in the proximity of a movie theater.

Now, with the new Location APIs, Google’s location algorithm has been rewritten to be more accurate and use significantly less battery life. There is just enough documentation  plus sample code and a downloadable sample app (GeofenceDetection) to help us get started creating geofencing apps. Prerequisites are:

1. Download  Google Play Services (via Android’s SDK Manager) and set it up as a library
2. Get a Google Maps v2 API key, and maybe run the sample app. This short Quick Start guide might help
3. Download (again via SDK Manager)  the Support Library to cater to older Android versions.

For practical purposes, let’s just start where the sample geofencing app (GeofenceDetection) stops, and introduce a few enhancements to make the app semi-decent and show a sample of possibilities with the new Location API.

1. Zoom and Camera Position

First, a little taste of Google Maps API v2. Let’s choose zoom level and camera angle:

view sourceprint?

01.import com.google.android.gms.maps.GoogleMap;

02.import com.google.android.gms.maps.CameraUpdateFactory;

03.import com.google.android.gms.maps.model.CameraPosition;

04.import com.google.android.gms.maps.model.LatLng;

05.//...

06.// inside class, for a given lat/lon

07.CameraPosition INIT =

08.new CameraPosition.Builder()

09..target(new LatLng(lat, lon))

10..zoom( 17.5F )

11..bearing( 300F) // orientation

12..tilt( 50F) // viewing angle

13..build();

14.// use GooggleMap mMap to move camera into position

15.mMap.animateCamera( CameraUpdateFactory.newCameraPosition(INIT) );

The code above has a zoom level allowing the viewing of buildings in 3D. Google Maps v2 uses OpenGL for Embedded Systems (OpenGL ES v2) to render 2D and 3D computer graphics.

2. Options menu

Even if we are not big fans of an Options menu, it might be adequate in this case, since we would not want to clutter the map with too much “touch” functionality (we will  have plenty of that shortly).

We can toggle between “normal” and satellite view:

view sourceprint?

01./**

02.* Toggle View Satellite-Normal

03.*/

04.public static void toggleView(){

05.mMap.setMapType( mMap.getMapType() ==

06.GoogleMap.MAP_TYPE_NORMAL ?

07.GoogleMap.MAP_TYPE_SATELLITE :

08.GoogleMap.MAP_TYPE_NORMAL);

09.}

We can also provide a “Flight Mode”, where we let the camera scroll away. Not tremendously useful, but kind of cool nonetheless:

view sourceprint?

01.import com.google.android.gms.maps.GoogleMap.CancelableCallback;

02.//...

03.private static CancelableCallback callback = new CancelableCallback() {

04.@Override

05.public void onFinish() {

06.scroll();

07.}

08.@Override

09.public void onCancel() {}

10.};

11. 

12.public static void scroll() {

13.// we don't want to scroll too fast since

14.// loading new areas in map takes time

15.mMap.animateCamera( CameraUpdateFactory.scrollBy(10, -10),

16.callback ); // 10 pix

17.}

3. Geocoding/Reverse Geocoding
The sample is here to demonstrate features and makes heavy use of latitude/longitude coordinates. But we need to provide a more user-friendly way to interface with locations on the map, like a street address. We can use geocoding/reverse geocoding  to transform a  street address to coordinates and vice-versa using Android’s Geocoder.

4. Adding geofences
OK, now on to geofences. Thanks to geocoding, we can request an actual physical address from the user instead of coordinates. We will  just change that address to a latitude/longitude pair internally to process user input. Notice how we use transparent UIs as much as possible to enhance what some might call the user experience. Notice also that we provide a spinner so that the user can choose between predefined values. That saves the user some typing and it saves us from validating coordinates values each time.

Still, if we want to be even more user-friendly, we can give our users the possibility to pre-fill the address field by long-pressing a point on the map. We will then use reverse geocoding to translate the coordinates to a physical address for display (screen below on the right):

Processing long-presses is pretty straightforward:

view sourceprint?

01.import com.google.android.gms.maps.GoogleMap.OnMapLongClickListener;

02.//...

03.public class MainActivity extends FragmentActivity

04.implements OnMapLongClickListener  {

05.//...

06.mMap.setOnMapLongClickListener(this);

07.//...

08.@Override

09.public void onMapLongClick(LatLng point) {

10.// reverse geocode point

11.}

12.}

Adding /Removing geofences is pretty much covered in the sample app (by the GeofenceRequesterand GeofenceRemover classes). The thing to remember is that the process of adding/removing fences is as follows :

1. A connection to Google’s Location Services is requested by our app.
2. Once/if the connection is available, the request to add/remove a fence is done using  aPendingIntent.
3. If a similar request made by our app is still underway, the operation fails.
4. Although the method we call (e.g. addGeofences()) returns at once, we won’t know if the request was successful until Location Services calls back into our app (e.g. OnAddGeofencesResultListener‘s onAddGeofencesResult() ) with a success status code.
5. Finally, the preceding method will use a Broadcast Intent to notify other components of our app of  success/failure.

Needless to say, we need to code defensively at almost every step of the way. Now, once a geofence is added, we can add a marker (the default or a customized one) and choose between different shapes (circle, polygon, etc.) to delimit the geofence. For instance we can write this code to add the default marker and circle the fence within a specified radius:

view sourceprint?

01.import com.google.android.gms.maps.model.Circle;

02.import com.google.android.gms.maps.model.CircleOptions;

03.import com.google.android.gms.maps.model.MarkerOptions;

04.//...

05.public static void addMarkerForFence(SimpleGeofence fence){

06.if(fence == null){

07.// display en error message and return

08.return;

09.}

10.mMap.addMarker( new MarkerOptions()

11..position( new LatLng(fence.getLatitude(), fence.getLongitude()) )

12..title("Fence " + fence.getId())

13..snippet("Radius: " + fence.getRadius()) ).showInfoWindow();

14. 

15.//Instantiates a new CircleOptions object +  center/radius

16.CircleOptions circleOptions = new CircleOptions()

17..center( new LatLng(fence.getLatitude(), fence.getLongitude()) )

18..radius( fence.getRadius() )

19..fillColor(0x40ff0000)

20..strokeColor(Color.TRANSPARENT)

21..strokeWidth(2);

22. 

23.// Get back the mutable Circle

24.Circle circle = mMap.addCircle(circleOptions);

25.// more operations on the circle...

26. 

27.}

Here are the resulting screens, including the one we get once we “Touch to Edit” the Info Window:

The sample app has all we need to fire notifications once the circled area above is entered or exited. Notice how we set up the marker’s Info window to allow editing the geofence radius, or removing the geofence altogether. To implement a clickable custom Info window, all we need is to create our ownInfoWindowAdapter and OnInfoWindowClickListener.

As for the notifications themselves, this is how they look like in the sample app:

We can of course change a notification’s appearance and functionality, and… that would be the subject of another article.  Hopefully, this one gave a glimpse of what is possible with the new Location API.  Have fun with Android geofences.