›

RMS filter for C-type sensor

mikesvmikesv
in Android
Hi,

Is the FitnessTracker app only works for R-type sensors on Android? Seems like C-type sensor support was added later for IOS only.



I have a C-type sensor and would like to implement RMS filter for activity tracking, what sample app should I use?

 Thank you,
Michael

Comments

  • EricEric
    We do not have a fitness tracker app that uses the latest API.  The old API used by the app greatly differs from the new API so updating the project is not as simple as switching to the latest API.

    What you can do is implement the filter chain using the latest api. It is much simpler that the code outlined in the blog post.
  • mikesvmikesv
    Thank you Eric. Can you tell me why I don't get any output please?

    private Bmi160Accelerometer accModule = null;
    // on ready
    accModule = mwBoard.getModule(Bmi160Accelerometer.class);
    accModule.setOutputDataRate(5.f);
    accModule.setAxisSamplingRange(4);

    //((Bmi160Accelerometer) accModule).configureAxisSampling()
    //        .setFullScaleRange(  Bmi160Accelerometer.AccRange.AR_4G )
    //        .commit();

    Log.i("tutorial", "mvOnReady");

    // on clicked start button
    accModule
            .routeData()
            .fromAxes()
            .process(new Rms())
            .stream("acc_stream")
            .commit()
            .onComplete(new AsyncOperation.CompletionHandler<RouteManager>() {
                @Override
                public void success(RouteManager result) {
                    result.subscribe("acc_stream", new RouteManager.MessageHandler() {
                        @Override
                        public void process(Message msg) {
                            Log.i("tutorial", msg.getData(Rms.class).toString());
                        }
                    });
    Thanks!
  • EricEric
    You have to start the accelerometer and enable axis sampling.  See:

    Also, rms data is interpreted as a float.  The Rms class is not a valid data type.
  • mikesvmikesv
    That worked! Thanks a lot!

    If I may to ask another question here.

    I need to upload RMS average over 15 minutes but the "average" module only accept 128 sample points.
    If I use 12.5 sampling rate for 15 min * 60 = 900 seconds my average over 900*12.5= 11,250 times.
    I can do that on the phone but I thought I may save the battery life  if I only connect and get  data every 15 minutes.
    .process(new Rms())
    .process(new Average((byte) 128))// 12.5 times per second ->900*12.5= 11,250 times

    .process(new Time(Time.OutputMode.ABSOLUTE,9000000)) // 900 sec, getting only last 128?
    What would be the best way to accomplish this and what is the best practices to prolong the battery life?
    Thank you!
    Michael
  • mikesvmikesv
    edited May 2016

  • EricEric
    The average is a running average, not a true average over N samples.  If you need a true average, you could sum up the rms values over a 15 min period, record that value and reset the accumulator.
  • mikesvmikesv
    Thank you Eric for the suggestion but I am getting a very strange results now.
    RMS was in 0.5-0.6 range.
    When I totaled it over 3 seconds I was expecting to get 3*12.5*0.5= 18.75, I am seeing 2-3
    and numbers are changing without me touching the sensor.
    .process(new Rms())
    .process("accum_proc", new Accumulator())
    .process(new Time(Time.OutputMode.ABSOLUTE,3000)) // 3 sec
    .process(new Maths(Maths.Operation.DIVIDE, 1.0)) // must be 3*12.5=37.5
    .monitor(new DataSignal.ActivityHandler() {
        @Override
        public void onSignalActive(Map<String, DataProcessor> processors,
                                   DataSignal.DataToken token) {
            processors.get("accum_proc").setState(new Accumulator.State(0));
                       }
    })
    05-20 17:37:51.766  { "d": {  "ts": "2016-05-20 17:37:51","rms":"3.6166992"}}
    05-20 17:37:54.836  { "d": {  "ts": "2016-05-20 17:37:54","rms":"2.375"}}
    05-20 17:37:57.956  { "d": {  "ts": "2016-05-20 17:37:57","rms":"3.6193237"}}
    05-20 17:38:00.976  { "d": {  "ts": "2016-05-20 17:38:00","rms":"1.75177"}}
    05-20 17:38:04.046  { "d": {  "ts": "2016-05-20 17:38:04","rms":"3.6175537"}}
    05-20 17:38:07.116  { "d": {  "ts": "2016-05-20 17:38:07","rms":"2.3759155"}}
    05-20 17:38:10.286  { "d": {  "ts": "2016-05-20 17:38:10","rms":"3.616516"}}
    05-20 17:38:13.356  { "d": {  "ts": "2016-05-20 17:38:13","rms":"2.3738403"}}
    05-20 17:38:16.426  { "d": {  "ts": "2016-05-20 17:38:16","rms":"3.6172485"}}
    05-20 17:38:19.406  { "d": {  "ts": "2016-05-20 17:38:19","rms":"1.7537231"}}
    05-20 17:38:22.476  { "d": {  "ts": "2016-05-20 17:38:22","rms":"3.6200562"}}
    05-20 17:38:25.596  { "d": {  "ts": "2016-05-20 17:38:25","rms":"2.9978638"}}
    05-20 17:38:28.616  { "d": {  "ts": "2016-05-20 17:38:28","rms":"2.3742065"}}
    05-20 17:38:31.686  { "d": {  "ts": "2016-05-20 17:38:31","rms":"3.614441"}}
    05-20 17:38:34.756  { "d": {  "ts": "2016-05-20 17:38:34","rms":"2.371643"}}
    thank you,
    Michael
  • EricEric
    Increase the output size of the accumulator.
  • mikesvmikesv

    Worked! 

    Thank you Eric.

  • rajmishrarajmishra
    How did you get readable time format.I am getting raw data, here is my code. Please guide
    public void success(RouteManager result) {
    result.subscribe(STREAM_KEY, new RouteManager.MessageHandler() {
    public void process(Message msg) {
    Log.i(TAG, String.format("Accelerometer: %s", msg.getData())); sensorMsg(String.format(msg.toString()), "accel");
  • mikesvmikesv
    you mean this:float fRMS = msg.getData(Float.class);
  • mikesvmikesv
    you mean this:
    float fRMS = msg.getData(Float.class);
  • mikesvmikesv
    if you need date/time : java.sql.Date date = new java.sql.Date(msg.getTimestamp().getTimeInMillis());
  • rajmishrarajmishra
    Thanks Mikesv, I got the desired results
     

  • lgleasainlgleasain
    FYI,  the fitness app has been updated,  but it uses the build in step counter instead of the RMS,  but it is possible to chain things.
This discussion has been closed.