Pedometer / custom alghoritms

Hi, I'd like to ask you about Pedometer. Both single steps and walksteps function are completety inaccurate. They count new steps even when I'm sitting or driving a car. Anyway, my goal is to determine whether the user is walking (standing) or sitting still while e.g. driving. Is there anyone who could help me do it in AI? Please do not suggest putting gps speed into equation, it won't help.

Per poter avere un buon pedometro sullo smartphone; un'app ha bisogno di essere eseguita in backgroud. Ma purtroppo Mit App Inventor non lo permette; ma permette solo di poter impostare lo schermo in modalità sempre accesa.

It doesn't matter. The pedometer function works very badly even though the screen is active. I am asking for custom alghoritms using accelerometers to build a reasonable app which could deferentiate between walking and sitting still in a moving einviroment, like a car or train.

For example, how can I programme it in AI as described here:


These posts pertain to using the accelerometer as a pedometer with Blocks:

edit: Detecting a change in movement state after a fall is discussed here Lack of moov with accelerometre

Pedometer control source code: appinventor-sources/ at 83eb64a98cf2f7bb7860c9bd21169e2b22f59e33 · mit-cml/appinventor-sources · GitHub