我需要有关如何检测 Android 设备移动量的建议。假设我将手机放在桌子或床上，然后如果有人敲击桌子或坐在床上或躺在床上，那么我想检测 android 设备的移动。
实际上我知道 android 有运动传感器 API，但我不知道使用哪种传感器以及哪种传感器类型最适合这种类型的运动检测。
如果有人可以分享一些基本的演示代码，我会很高兴。
【问题讨论】：
我一直在用类似的想法来测量手机的位移。我发现 LINEAR ACCELERATION（和 ACCELERATION）不够准确，无法正确测量位移。
这段代码应该会更好一些：
（初始化）
private SensorManager sensorManager;
private Sensor accelerometer;
double[] maxAccelerations = new double[3];
double[] position = new double[3];
long[] times = new long[3];
// time combined with maxAcceleration can approximate the change in position,
// with the formula Δpos = (maxAcceleration * time ^ 2) / 6
long currentTime;
(onCreate)
sensorManager = (SensorManager) getSystemService(SENSOR_SERVICE);
if (sensorManager.getDefaultSensor(Sensor.TYPE_LINEAR_ACCELERATION) != null) {
accelerometer = sensorManager.getDefaultSensor(Sensor.TYPE_LINEAR_ACCELERATION);
sensorManager.registerListener(this, accelerometer, sensorManager.SENSOR_DELAY_FASTEST);
}
currentTime = System.currentTimeMillis();
for(int i=0;i<3;i++){
times[i]=currentTime;
}
else{
throw "Error";
//Which will throw an error, if not the error that is expected. ?
}
（onSensorChanged 和 onAccuracyChanged）
@Override
public void onAccuracyChanged(Sensor ignore, int thisFunction) {
}
@Override
public void onSensorChanged(SensorEvent event) {
if (event.sensor.getType() == Sensor.TYPE_LINEAR_ACCELERATION) {
for(int i=0;i<3;i++){
if(Math.abs(event.values[i])<0.01){
// Note: this is to try to prevent accelerating time from being counted when the phone is stationary. 0.01 should be
// changed to an appropriate sensitivity level that can be calculated by finding an average noise level when the phone is stationary.
times[i]=System.currentTimeMillis();
}
if(event.values[i]>maxAccelerations[i] && maxAccelerations[i]>=0){
maxAccelerations[i]=event.values[i];
}
else if(event.values[i]<maxAccelerations[i] && maxAccelerations[i]<=0){
maxAccelerations[i]=event.values[i];
}
else if(event.values[i]>0 && maxAccelerations[i]<0){
currentTime = System.currentTimeMillis();
position[i]+=maxAccelerations[i] * (times[i]-currentTime)*(times[i]-currentTime) / 6;
times[i]=currentTime;
maxAccelerations[i]=event.values[i];
}
else if(event.values[i]<0 && maxAccelerations[i]>0){
currentTime = System.currentTimeMillis();
position[i]+=maxAccelerations[i] * (times[i]-currentTime)*(times[i]-currentTime) / 6;
times[i]=currentTime;
maxAccelerations[i]=event.values[i];
}
}
}
}
【讨论】：
如果您要查找手机的位移，则需要找到 作用在手机上的线性加速度，而不是重力加速度
android 有一个内置转换器可以找到作用在您手机上的 LINEAR ACCELERATION
https://github.com/yuvaramsingh94/AndroidSensorTestCode/tree/master
这是一段代码，您可以在其中看到如何获取 LINEAR ACCELERATION 的原始值
【讨论】：
此代码用于步行检测（由@anthropomo代码修改）
获得更平滑的值。
//初始化
private SensorManager sensorMan;
private Sensor accelerometer;
private float[] mGravity;
private double mAccel;
private double mAccelCurrent;
private double mAccelLast;
private boolean sensorRegistered = false;
// onCreate
sensorMan = (SensorManager) context.getSystemService(Context.SENSOR_SERVICE);
accelerometer = sensorMan.getDefaultSensor(Sensor.TYPE_ACCELEROMETER);
mAccel = 0.00f;
mAccelCurrent = SensorManager.GRAVITY_EARTH;
mAccelLast = SensorManager.GRAVITY_EARTH;
sensorMan.registerListener(this, accelerometer,
SensorManager.SENSOR_DELAY_NORMAL);
sensorRegistered = true;
// onSensorChanged
private int hitCount = 0;
private double hitSum = 0;
private double hitResult = 0;
private final int SAMPLE_SIZE = 50; // change this sample size as you want, higher is more precise but slow measure.
private final double THRESHOLD = 0.2; // change this threshold as you want, higher is more spike movement
@Override
public void onSensorChanged(SensorEvent event) {
if (event.sensor.getType() == Sensor.TYPE_ACCELEROMETER) {
mGravity = event.values.clone();
// Shake detection
double x = mGravity[0];
double y = mGravity[1];
double z = mGravity[2];
mAccelLast = mAccelCurrent;
mAccelCurrent = Math.sqrt(x * x + y * y + z * z);
double delta = mAccelCurrent - mAccelLast;
mAccel = mAccel * 0.9f + delta;
if (hitCount <= SAMPLE_SIZE) {
hitCount++;
hitSum += Math.abs(mAccel);
} else {
hitResult = hitSum / SAMPLE_SIZE;
Log.d(TAG, String.valueOf(hitResult));
if (hitResult > THRESHOLD) {
Log.d(TAG, "Walking");
} else {
Log.d(TAG, "Stop Walking");
}
hitCount = 0;
hitSum = 0;
hitResult = 0;
}
}
}
【讨论】：
我使用了以下类：
public class MovementDetector implements SensorEventListener {
protected final String TAG = getClass().getSimpleName();
private SensorManager sensorMan;
private Sensor accelerometer;
private MovementDetector() {
}
private static MovementDetector mInstance;
public static MovementDetector getInstance() {
if (mInstance == null) {
mInstance = new MovementDetector();
mInstance.init();
}
return mInstance;
}
//////////////////////
private HashSet<Listener> mListeners = new HashSet<MovementDetector.Listener>();
private void init() {
sensorMan = (SensorManager) GlobalData.getInstance().getContext().getSystemService(Context.SENSOR_SERVICE);
accelerometer = sensorMan.getDefaultSensor(Sensor.TYPE_LINEAR_ACCELERATION);
}
public void start() {
sensorMan.registerListener(this, accelerometer, SensorManager.SENSOR_DELAY_NORMAL);
}
public void stop() {
sensorMan.unregisterListener(this);
}
public void addListener(Listener listener) {
mListeners.add(listener);
}
/* (non-Javadoc)
* @see android.hardware.SensorEventListener#onSensorChanged(android.hardware.SensorEvent)
*/
@Override
public void onSensorChanged(SensorEvent event) {
if (event.sensor.getType() == Sensor.TYPE_LINEAR_ACCELERATION) {
float x = event.values[0];
float y = event.values[1];
float z = event.values[2];
float diff = (float) Math.sqrt(x * x + y * y + z * z);
if (diff > 0.5) // 0.5 is a threshold, you can test it and change it
Log.d(TAG,"Device motion detected!!!!");
for (Listener listener : mListeners) {
listener.onMotionDetected(event, diff);
}
}
}
/* (non-Javadoc)
* @see android.hardware.SensorEventListener#onAccuracyChanged(android.hardware.Sensor, int)
*/
@Override
public void onAccuracyChanged(Sensor sensor, int accuracy) {
// TODO Auto-generated method stub
}
public interface Listener {
void onMotionDetected(SensorEvent event, float acceleration);
}
}
用法：
关于我的活动onCrate()：
MovementDetector.getInstance().addListener(new MovementDetector.Listener() {
@Override
public void onMotionDetected(SensorEvent event, float acceleration) {
mMotionDetectionTextView.setText("Acceleration: ["+String.format("%.3f",event.values[0])+","+String.format("%.3f",event.values[1])+","+String.format("%.3f",event.values[2])+"] "+String.format("%.3f", acceleration));
if (acceleration > SettingsHelper.getInstance().getMotionDetectionThreshold()){
mMotionDetectionTextView.setTextColor(Color.RED);
} else {
mMotionDetectionTextView.setTextColor(Color.WHITE);
}
}
});
关于我的活动onResume()：
MovementDetector.getInstance().start();
关于我的活动onPause()：
MovementDetector.getInstance().stop();
【讨论】：
GlobalData 是什么？
一定要使用加速度计：
// Start with some variables
private SensorManager sensorMan;
private Sensor accelerometer;
private float[] mGravity;
private float mAccel;
private float mAccelCurrent;
private float mAccelLast;
// In onCreate method
sensorMan = (SensorManager)getSystemService(SENSOR_SERVICE);
accelerometer = sensorMan.getDefaultSensor(Sensor.TYPE_ACCELEROMETER);
mAccel = 0.00f;
mAccelCurrent = SensorManager.GRAVITY_EARTH;
mAccelLast = SensorManager.GRAVITY_EARTH;
// And these:
@Override
public void onResume() {
super.onResume();
sensorMan.registerListener(this, accelerometer,
SensorManager.SENSOR_DELAY_UI);
}
@Override
protected void onPause() {
super.onPause();
sensorMan.unregisterListener(this);
}
@Override
public void onSensorChanged(SensorEvent event) {
if (event.sensor.getType() == Sensor.TYPE_ACCELEROMETER){
mGravity = event.values.clone();
// Shake detection
float x = mGravity[0];
float y = mGravity[1];
float z = mGravity[2];
mAccelLast = mAccelCurrent;
mAccelCurrent = FloatMath.sqrt(x*x + y*y + z*z);
float delta = mAccelCurrent - mAccelLast;
mAccel = mAccel * 0.9f + delta;
// Make this higher or lower according to how much
// motion you want to detect
if(mAccel > 3){
// do something
}
}
}
@Override
public void onAccuracyChanged(Sensor sensor, int accuracy) {
// required method
}
【讨论】：
if(mAccel > 3)这一行。如果您将该数字减少到 1 或 0.5 或 0.25 或更低，您可以在较低的阈值下“做某事”。如果你想检测倾斜，谷歌“Android 检测倾斜”。它是相关的，但又不同。
(float)Math.sqrt(x*x + y*y + z*z); 来计算加速度。
虽然我没有演示代码（因为您不够具体），但这里有一个好的开始：http://developer.android.com/guide/topics/sensors/sensors_motion.html（以及左侧的其他项目）。
【讨论】：