FOSSASIA

This blog demonstrates how to use built-in mobile sensors in an Android application. This blog will mainly feature my work done in PSLab Android repository of making a Compass and Accelerometer instrument using built-in mobile sensors.

How to access built-in mobile sensors?

Android provides an abstract class called SensorManager which is able to communicate with the hardware i.e. here the sensors in the mobile. But the SensorManager can’t provide continuous data fetched by the sensor. For this, Android provides an interface known as SensorEventListener which receives notifications from SensorManager whenever there is a new sensor data.

How to implement the functionality of sensors in Android app?

Following is a step by step process on how to add support for different sensors in an Android app

• First, make a new class which extends SensorEventListener and override the default methods.

public class SensorActivity extends Activity implements SensorEventListener {

     public SensorActivity() {
        // Default Constructor      
     }

     @Override
     public void onAccuracyChanged(Sensor sensor, int accuracy) {
     }

     @Override
     public void onSensorChanged(SensorEvent event) {
     }
 }

Here, the SensorActivity() is the default constructor of the class and the onAccuracyChanged() and onSensorChanged() methods will be explained soon.

• Now declare the SensorManager and use the sensor needed in the app.

private final SensorManager mSensorManager;
private final Sensor mAccelerometer;

     public SensorActivity() {
         mSensorManager = (SensorManager)getSystemService(SENSOR_SERVICE);
         mAccelerometer =        mSensorManager.getDefaultSensor(Sensor.TYPE_ACCELEROMETER);
     }

Here, I have used Sensor.TYPE_ACCELEROMETER to use the built-in Accelerometer in the device. Some of the other options available are:

1. TYPE_LIGHT – To measure ambient light
2. TYPE_MAGNETOMETER – To measure magnetic field along different axis
3. TYPE_GYROSCOPE – To measure movements (sudden changes) in any particular direction

The list of all available sensors in Android can be found in [1].

• It is necessary to disable the sensors especially when the activity is paused. Failing to do so can drain the battery in just a few hours.

NOTE: The system will not disable sensors automatically when the screen turns off.

So, to save the battery and make the app efficient, we can use the registerListener method to notify the SensorManager to start fetching data from sensor and unregisterListener to notify it to stop.

@Override
protected void onResume() {
         super.onResume();
         mSensorManager.registerListener(this, mAccelerometer, SensorManager.SENSOR_DELAY_NORMAL);
     }

@Override
     protected void onPause() {
         super.onPause();
         mSensorManager.unregisterListener(this);
     }

The onResume() method activates when the app is resumed from a paused state and the onPause() method is called when the app is paused i.e. some other app draws over the current app.

• Now coming back to onAccuracyChanged() and onSensorChanged() methods, the onAccuracyChanged() method is used to set the accuracy of a sensor. For example, while using GeoLocation sensor, sometimes the position of the mobile isn’t very accurate and so we can define the accuracy level in this method so that the fetched data is used for calculations only if it is in the provided range. And the onSensorChanged() method is the main method where all the data is processed as soon as the new data is notified.

To get the latest value from the sensor, we can use

@Override
public void onSensorChanged(SensorEvent event) {
   data = Float.valueOf(event.values[0]);
   unRegisterListener();
}

Here, the event is an instance of the SensorEvent class which provides the updated data fetched from the sensor. Event.values is used to get the values for any of the three axis including the bias in their values. Following is the list of the index for which we can get a necessary value

values[0] = x_uncalib without bias compensation
values[1] = y_uncalib without bias compensation
values[2] = z_uncalib without bias compensation
values[3] = estimated x_bias
values[4] = estimated y_bias 
values[5] = estimated z_bias

So, in this way, we can add support for any built-in mobile sensor in our Android application.

Resources

• https://source.android.com/devices/sensors/sensor-types – List of available built-in sensors in Android