Android车辆运动轨迹平滑移动(高仿滴滴打车)最佳实践

github源码下载地址：https://github.com/geduo83/android-amap-movecar

关于车辆运动的相关文章一共写过两篇，一篇为Android车辆运动轨迹大数据采集最佳实践，另外一篇是Android车辆运动轨迹数据采集服务保活的探索与发现，
一直打算写一篇车辆运动轨迹平滑移动的文章，年后由于工作项目太忙也就没时间写，工作的事情忙完了，紧接着就是忙自己的另外两个开源项目，一个是Android客户端框架FlyTour
、和另外一个SpringCloud微服务框架FlyCloud，上个周终于告一段落。

数据也能采集了，而且采集服务保活也做了，有效的避免了数据采集服务在后台被系统杀死的可能，接着就需要把采集到的数据在移动设备上实时的展示出来，这两天我也特意在网上搜索了一下，关于Android车辆运动轨迹平滑移动的文章，还真没有，大部分都是提问的多，问怎么实现的？都是一些只言片语很零碎的一些回答，在实战项目当中没有太大的实用价值。

关于车辆运动，在我们在日常生活中见到最多的就是滴滴打车，想必这款app大家都使用过，当你在app的叫车页面，输入完毕你的目的地，点击叫车，如果有司机接单了，你就清楚的看到车辆会平滑的移动到你所在的位置去接你。今天我就带领大家一步步的实现一个车辆平滑移动的功能。

这个功能是基于高德地图开发的，因此我特意去高德官网查阅了一下，高德的确提供了官方轨迹移动api，我们暂且不用官方API.先用自己的方法去实现一个最简单的功能
我们先来一睹为快：

单次轨迹回放

已知有一段轨迹数据，点击回放按钮，小车沿着路线自动的往前运动，播放完毕也就结束了

 public class MoveSingleThread extends Thread{private List<LatLng> mLatLngList;private Marker mCarMarker;public MoveSingleThread(List<LatLng> latLngs, Marker marker) {super();mLatLngList = latLngs;mCarMarker = marker;}@Overridepublic void run() {super.run();}public void moveTrack(){// 第一个for循环用来计算走了多少部int step = 0;for (int i = 0; i < mLatLngList.size() - 1; i++) {LatLng startPoint = mLatLngList.get(i);LatLng endPoint = mLatLngList.get(i + 1);double slope = getSlope(startPoint, endPoint);// 是不是正向的标示（向上设为正向）boolean isReverse = (startPoint.latitude > endPoint.latitude);double xMoveDistance = isReverse ? getXMoveDistance(slope) : -1 * getXMoveDistance(slope);// 应该对经纬度同时处理for (double j = startPoint.latitude; !((j >= endPoint.latitude) ^ isReverse); j =j - xMoveDistance) {step++;}}// 通过距离,计算轨迹动画时间间隔double mTimeInterval = 0;// 轨迹回放时间戳if (!TextUtils.isEmpty(mDistance)) {float totalDistance = Float.parseFloat(mDistance) * 1000;if (totalDistance <= 500) {mTimeInterval = 1000.0 / step;} else if (totalDistance > 500 && totalDistance <= 7500) {mTimeInterval = 2.0 * totalDistance / step;} else {mTimeInterval = 15000.0 / step;}}// while (true) {for (int i = 0; i < mLatLngList.size() - 1; i++) {if (stopFlag) {stopFlag = false;break;}mIsCarMoveing = true;LatLng startPoint = mLatLngList.get(i);LatLng endPoint = mLatLngList.get(i + 1);mCarMarker.setPosition(startPoint);mCarMarker.setRotateAngle((float) getAngle(startPoint, endPoint));double slope = getSlope(startPoint, endPoint);// 是不是正向的标示（向上设为正向）boolean isReverse = (startPoint.latitude > endPoint.latitude);double intercept = getInterception(slope, startPoint);double xMoveDistance = isReverse ? getXMoveDistance(slope) : -1 * getXMoveDistance(slope);// 应该对经纬度同时处理double mSleep = 0;for (double j = startPoint.latitude; !((j >= endPoint.latitude) ^ isReverse); j =j - xMoveDistance) {LatLng latLng = null;if (slope != Double.MAX_VALUE) {latLng = new LatLng(j, (j - intercept) / slope);// latLng = new LatLng(j, k);} else {latLng = new LatLng(j, startPoint.longitude);}mCarMarker.setPosition(latLng);// 如果间隔时间小于1毫秒,则略过当前休眠,累加直到休眠时间到1毫秒:会损失精度if (mTimeInterval < 1) {mSleep += mTimeInterval;if (mSleep >= 1) {SystemClock.sleep((long) mSleep);mSleep = 0;}} elseSystemClock.sleep((long) mTimeInterval);}}}
}

实时轨迹数据排队问题

如果要显示实时轨迹怎么办，上面的代码就有问题了，Thread.start()方法调用后，就会立马执行他的run方法，run方法执行完毕，线程也就结束了，也是说上面的代码只能跑一次轨迹数据，如果每间隔五秒从后台取一次轨迹数据，就需要一数据队列来存储这些数据，每跑完一次数据，就从数据队列里面去取，如果有就取来接着跑，如果没有就处于等待状态。我们创建异步消息处理线程，这一问题就可以迎刃而解，来数据了我们就可以通过handler把数据post给我们的子线程，Handler自带数据队列，它处于排队状态，如果有数据了就开始跑轨迹，如果没有数据就处于等待状态，直到有数据的到来，如果对异步消息处理线程不熟悉，请查看我的另外一篇文章Android实战开发Handler机制深度解析https://menxindiaolong.blog.csdn.net/article/details/86560330

一个标准的异步消息处理线程应该怎么写？
方法1：

class LooperThread extends Thread {public Handler mHandler;public void run() {Looper.prepare();mHandler = new Handler() {public void handleMessage(Message msg) {// process incoming messages here}};Looper.loop();}
}

方法2：

  // Step 1: 创建并启动HandlerThread线程，内部包含LooperHandlerThread handlerThread = new HandlerThread("gityuan.com");handlerThread.start();// Step 2: 创建HandlerHandler handler = new Handler(handlerThread.getLooper()) {public void handleMessage(Message msg) {// process incoming messages here}};// Step 3: 发送消息handler.post(new Runnable() {@Overridepublic void run() {System.out.println("thread id="+Thread.currentThread().getId());}});

上面就是Android系统中异步消息处理线程的通用写法

运动轨迹的暂停、继续问题

由于运动轨迹是在子线程里面完成的，我们自然而然会想到线程的等待、唤醒，也就是wait、notify的问题了
因此我们在运动过程加上就如下代码就可以了

if (pause) {try {lock.wait();} catch (InterruptedException e) {e.printStackTrace();}}
}

怎么让他恢复运动呢？notify一下即可

public void reStartMove() {synchronized (lock) {pause = false;lock.notify();}
}

完整的代码如下：

/*** Description: <MoveCarCustomThread><br>* Author:      mxdl<br>* Date:        2019/7/10<br>* Version:     V1.0.0<br>* Update:     <br>*/
public class MoveCarCustomThread extends Thread {public static final String TAG = MoveCarCustomThread.class.getSimpleName();private Handler moveCarHandler;//发送数据的异步消息处理器private Object lock = new Object();//线程锁private boolean moveing = false;//是否线程正在移动private boolean pause = false;//暂停状态，为true则暂停private boolean stop = false;//停止状态，为true则停止移动private WeakReference<MainActivity> mActivityWeakReference;//防止内存Activity导致的内容泄漏private MOVE_STATE currMoveState = MOVE_STATE.START_STATUS;public void setCurrMoveState(MOVE_STATE currMoveState) {this.currMoveState = currMoveState;}public MOVE_STATE getCurrMoveState() {return currMoveState;}public MoveCarCustomThread(MainActivity activity) {mActivityWeakReference = new WeakReference<>(activity);}//暂停移动public void pauseMove() {pause = true;}//设置暂停之后，再次移动调用它public void reStartMove() {synchronized (lock) {pause = false;lock.notify();}}public void stopMove() {stop = true;if(moveCarHandler != null){moveCarHandler.removeCallbacksAndMessages(null);}if(mActivityWeakReference.get() != null){mActivityWeakReference.get().mLatLngList.clear();mActivityWeakReference.get().mMainHandler.removeCallbacksAndMessages(null);}}public Handler getMoveCarHandler() {return moveCarHandler;}public boolean isMoveing() {return moveing;}@Overridepublic void run() {super.run();//设置该线程为loop线程Looper.prepare();moveCarHandler = new Handler() {@Overridepublic void handleMessage(Message msg) {super.handleMessage(msg);//通过锁保证发过来的数据同步入列synchronized (lock) {if (msg.obj != null && msg.obj instanceof List) {List<LatLng> latLngList = (List<LatLng>) msg.obj;moveCoarseTrack(latLngList);}}}};//启动loop线程Looper.loop();}private void moveCoarseTrack(List<LatLng> latLngList) {if (latLngList == null || latLngList.size() == 0 || latLngList.size() == 1) {return;}Log.v(TAG, "moveCoarseTrack start.........................................................");long startTime = System.currentTimeMillis();Log.v(TAG, "startTime:" + startTime);int step = TrackMoveUtil.getStep(latLngList);// 通过距离,计算轨迹动画运动步数Log.v(TAG, "move step:" + step);float distance = TrackMoveUtil.getDistance(latLngList);Log.v(TAG, "move distance:" + distance);double mTimeInterval = TrackMoveUtil.getMoveTime(distance, step);// 通过距离,计算轨迹动画时间间隔mTimeInterval = 10;// 每走一步停止10毫秒Log.v(TAG, "move mTimeInterval:" + mTimeInterval);moveing = true;for (int i = 0; i < latLngList.size() - 1; i++) {// 暂停状态，线程停止了if (pause) {movePause();}if (stop) {break;}moveing = true;LatLng startLatLng = latLngList.get(i);LatLng endLatLng = latLngList.get(i + 1);MainActivity mainActivity = mActivityWeakReference.get();moveCar(startLatLng, endLatLng, mainActivity);moveLine(startLatLng, mainActivity);moveCamera(startLatLng, mainActivity);double slope = TrackMoveUtil.getSlope(startLatLng, endLatLng);// 计算两点间的斜率double intercept = TrackMoveUtil.getInterception(slope, startLatLng);// 根据点和斜率算取截距boolean isReverse = (startLatLng.latitude > endLatLng.latitude);// 是不是正向的标示（向上设为正向）double xMoveDistance = isReverse ? TrackMoveUtil.getXMoveDistance(slope) : -1 * TrackMoveUtil.getXMoveDistance(slope);// 应该对经纬度同时处理double sleep = 0;int flag = 0;for (double j = startLatLng.latitude; !((j >= endLatLng.latitude) ^ isReverse); j = j - xMoveDistance) {// 非暂停状态地图才进行跟随移动if (pause) {movePause();}if (stop) {break;}moveing = true;flag++;if (slope != Double.MAX_VALUE) {startLatLng = new LatLng(j, (j - intercept) / slope);} else {startLatLng = new LatLng(j, startLatLng.longitude);}moveCar(startLatLng, mainActivity);moveLine(startLatLng, mainActivity);if (flag % 100 == 0) {moveCamera(startLatLng, mainActivity);}// 如果间隔时间小于1毫秒,则略过当前休眠,累加直到休眠时间到1毫秒:会损失精度if (mTimeInterval < 1) {sleep += mTimeInterval;if (sleep >= 1) {Log.v(TAG, "sleep:" + sleep);SystemClock.sleep((long) sleep);sleep = 0;}} else {SystemClock.sleep((long) mTimeInterval);}}}long endTime = System.currentTimeMillis();moveing = false;Log.v(TAG, "endTime:" + endTime);Log.v(TAG, "run mTimeInterval:" + (endTime - startTime));Log.v(TAG, "moveCoarseTrack end.........................................................");}private void moveLine(LatLng startLatLng, MainActivity mainActivity) {mainActivity.mLatLngList.add(startLatLng);// 向轨迹集合增加轨迹点mainActivity.mMovePolyline.setPoints(mainActivity.mLatLngList);// 轨迹画线开始}private void moveCar(LatLng startLatLng, LatLng endLatLng, MainActivity mainActivity) {moveCar(startLatLng,mainActivity);if (mainActivity.mCarMarker != null) {mainActivity.mCarMarker.setRotateAngle((float) TrackMoveUtil.getAngle(startLatLng, endLatLng));// 设置小车车头的方向}}private void moveCar(LatLng startLatLng,MainActivity mainActivity) {if (mainActivity.mCarMarker != null) {mainActivity.mCarMarker.setPosition(startLatLng);// 小车移动}}private void movePause() {try {lock.wait();} catch (InterruptedException e) {e.printStackTrace();}}private void moveCamera(LatLng startLatLng, MainActivity mainActivity) {Message message = Message.obtain();message.what = MainActivity.EventType.MapMove;message.obj = startLatLng;mainActivity.mMainHandler.sendMessage(message);}}

核心算法工具类

 * <h1>轨迹平滑所需要的工具方法</h1> Date: 2016-10-27 Created by mxdl*/
public class TrackMoveUtil {private static double DISTANCE = 0.0001;/*** 根据两点算斜率*/public static double getSlope(LatLng fromPoint, LatLng toPoint) {if (fromPoint == null || toPoint == null) {return 0;}if (toPoint.longitude == fromPoint.longitude) {return Double.MAX_VALUE;}double slope =((toPoint.latitude - fromPoint.latitude) / (toPoint.longitude - fromPoint.longitude));return slope;}/*** 根据两点算取图标转的角度*/public static double getAngle(LatLng fromPoint, LatLng toPoint) {if (fromPoint == null || toPoint == null) {return 0;}double slope = getSlope(fromPoint, toPoint);if (slope == Double.MAX_VALUE) {if (toPoint.latitude > fromPoint.latitude) {return 0;} else {return 180;}}float deltAngle = 0;if ((toPoint.latitude - fromPoint.latitude) * slope < 0) {deltAngle = 180;}double radio = Math.atan(slope);double angle = 180 * (radio / Math.PI) + deltAngle - 90;return angle;}/*** 根据点和斜率算取截距*/public static double getInterception(double slope, LatLng point) {if (point == null) {return 0;}return point.latitude - slope * point.longitude;}/*** 计算x方向每次移动的距离*/public static double getXMoveDistance(double slope) {if (slope == Double.MAX_VALUE) {return DISTANCE;}return Math.abs((DISTANCE * slope) / Math.sqrt(1 + slope * slope));}/*** 根据轨迹线段计算小车走了多少步* * @param latLngList* @return*/public static int getStep(List<LatLng> latLngList) {int step = 0;if (latLngList != null && latLngList.size() > 1) {for (int i = 0; i < latLngList.size() - 1; i++) {try {LatLng startPoint = latLngList.get(i);LatLng endPoint = latLngList.get(i + 1);double slope = getSlope(startPoint, endPoint);// 是不是正向的标示（向上设为正向）boolean isReverse = (startPoint.latitude > endPoint.latitude);double xMoveDistance = isReverse ? getXMoveDistance(slope) : -1 * getXMoveDistance(slope);// 应该对经纬度同时处理for (double j = startPoint.latitude; !((j >= endPoint.latitude) ^ isReverse); j =j - xMoveDistance) {step++;}} catch (Exception e) {e.printStackTrace();}}}return step;}/*** 根据总距离和步数计算运动时间* * @param distance* @param step* @return*/public static double getMoveTime(float distance, int step) {double timeInterval = 0;if (distance > 0) {float totalDistance = distance * 1000;if (totalDistance <= 500) {timeInterval = 1000.0 / step;} else if (totalDistance > 500 && totalDistance <= 7500) {timeInterval = 2.0 * totalDistance / step;} else {timeInterval = 15000.0 / step;}}return timeInterval;}/*** 根据轨迹点集合计算总距离* * @param latLngList* @return*/public static float getDistance(List<LatLng> latLngList) {float distance = 0;if (latLngList != null && latLngList.size() > 1) {for (int i = 0; i < latLngList.size() - 1; i++) {try {distance += AMapUtils.calculateLineDistance(latLngList.get(i), latLngList.get(i + 1));} catch (Exception e) {e.printStackTrace();}}}return distance;}// latitude - 地点的纬度，在-90 与90 之间的double 型数值。// longitude - 地点的经度，在-180 与180 之间的double 型数值。/*** 根据一个经纬度字符串求一个经纬度集合a|b|c|d;* * @param latlonStr* @return*/public static List<LatLng> getListLatLng(String latlonStr) {if (!TextUtils.isEmpty(latlonStr)) {String[] trackArr = latlonStr.split("\\|");if (trackArr != null && trackArr.length > 0) {List<LatLng> latLngList = new ArrayList<LatLng>();for (int i = 0; i < trackArr.length - 1; i = i + 2) {try {String lat = trackArr[i + 1];String lng = trackArr[i];// Logger.v(TAG,"trackArr index:" + i);// Logger.v(TAG,"trackArr lat:" + lat);// Logger.v(TAG,"trackArr lng:" + lng);if (!TextUtils.isEmpty(lat) && !TextUtils.isEmpty(lng)) {Double dLat = Double.valueOf(lat);Double dLng = Double.valueOf(lng);if (dLat >= -90 && dLat <= 90 && dLng >= -180 && dLng <= 180&& !(dLat == 0 && dLng == 0)) {LatLng latLng = new LatLng(dLat, dLng);latLngList.add(latLng);}}} catch (Exception e) {e.printStackTrace();}}return latLngList;}}return null;}
}

高德API实现的实时运动轨迹

/*** Description: <MoveCarCustomThread><br>* Author:      mxdl<br>* Date:        2019/7/10<br>* Version:     V1.0.0<br>* Update:     <br>*/
public class MoveCarSmoothThread implements IMoveCar {public static final String TAG = MoveCarSmoothThread.class.getSimpleName();private MovingPointOverlay mMovingPointOverlay;private WeakReference<MainActivity> mActivityWeakReference;private boolean isfirst = true;private MOVE_STATE currMoveState = MOVE_STATE.START_STATUS;public void setCurrMoveState(MOVE_STATE currMoveState) {this.currMoveState = currMoveState;}public MOVE_STATE getCurrMoveState() {return currMoveState;}public MoveCarSmoothThread(MainActivity activity) {mActivityWeakReference = new WeakReference<>(activity);}@Overridepublic void startMove(List<LatLng> latLngs) {if (latLngs == null || latLngs.size() == 0) {return;}Log.v("MYTAG","startMove start:"+Thread.currentThread().getName());Log.v(TAG, "moveCoarseTrack start.........................................................");long startTime = System.currentTimeMillis();Log.v(TAG, "startTime:" + startTime);final MainActivity mainActivity = mActivityWeakReference.get();if (mMovingPointOverlay == null) {mMovingPointOverlay = new MovingPointOverlay(mainActivity.mAMap, mainActivity.mCarMarker);mMovingPointOverlay.setTotalDuration(5);mMovingPointOverlay.setMoveListener(new MovingPointOverlay.MoveListener() {@Overridepublic void move(double v) {if(isfirst){isfirst = false;Log.v("MYTAG","MoveCarSmoolthThread move start:"+Thread.currentThread().getName());}LatLng position = mMovingPointOverlay.getPosition();mainActivity.mLatLngList.add(position);// 向轨迹集合增加轨迹点mainActivity.mMovePolyline.setPoints(mainActivity.mLatLngList);// 轨迹画线开始Message message = Message.obtain();message.what = MainActivity.EventType.MapMove;message.obj = position;message.arg1 = (int)v;mainActivity.mMainHandler.sendMessage(message);}});}mMovingPointOverlay.setPoints(latLngs);mMovingPointOverlay.startSmoothMove();long endTime = System.currentTimeMillis();Log.v(TAG, "endTime:" + endTime);Log.v(TAG, "moveCoarseTrack end.........................................................");}@Overridepublic void reStartMove() {if(mMovingPointOverlay != null){mMovingPointOverlay.startSmoothMove();}}@Overridepublic void pauseMove(){if(mMovingPointOverlay != null){mMovingPointOverlay.stopMove();}}@Overridepublic void stopMove(){if(mMovingPointOverlay != null){mMovingPointOverlay.destroy();mMovingPointOverlay = null;}if(mActivityWeakReference.get() != null){mActivityWeakReference.get().mLatLngList.clear();}}}

最后我把整个项目的的完整代码传到GitHub上了https://github.com/geduo83/android-amap-movecar