06-12-2018 08:21
- Mark as New
- Bookmark
- Subscribe
- Permalink
- Report this post

06-12-2018 08:21
- Mark as New
- Bookmark
- Subscribe
- Permalink
- Report this post
I wrote a simple app following the guides in the dev section of the web site. It takes the accelerometer data and outputs them to the screen. The problem I'm having is that when I lay the watch flat on the table I get the standard 9.8 along the z-axis that I'd expected, but what also accompanies it is extra acceleration along the other axises as well. I was wondering if others are experiencing this, if I got a bad ionic Fitbit, or there's another problem I haven't foreseen.
Answered! Go to the Best Answer.
Accepted Solutions
08-30-2018 11:14
- Mark as New
- Bookmark
- Subscribe
- Permalink
- Report this post

08-30-2018 11:14
- Mark as New
- Bookmark
- Subscribe
- Permalink
- Report this post
Okay. So since no one has replied to this topic, I will.
Sensors like the one in the Fitbit and cell phones, are prone to variation/errors in the data that they send. Only much bigger accelerometers will send data that has very lower number of variation, which making use of them in a watch would make it much too big to put on someones wrist, hence the use of the smaller accelerometer. Depending on the errors produced by the accelermeter (or other sensors), it might be required to use a DC offset to correct some alignment issues, a filter to help estimate the correct data (ex. a Kalman filter), a merging or sensor output, or even use all three. Now I'm definitely not an expert, there may be more ways to go about this. As for now, this hopefully will help those who didn't know about this issue and help them get on the right track to making their app a great one.

08-30-2018 11:14
- Mark as New
- Bookmark
- Subscribe
- Permalink
- Report this post

08-30-2018 11:14
- Mark as New
- Bookmark
- Subscribe
- Permalink
- Report this post
Okay. So since no one has replied to this topic, I will.
Sensors like the one in the Fitbit and cell phones, are prone to variation/errors in the data that they send. Only much bigger accelerometers will send data that has very lower number of variation, which making use of them in a watch would make it much too big to put on someones wrist, hence the use of the smaller accelerometer. Depending on the errors produced by the accelermeter (or other sensors), it might be required to use a DC offset to correct some alignment issues, a filter to help estimate the correct data (ex. a Kalman filter), a merging or sensor output, or even use all three. Now I'm definitely not an expert, there may be more ways to go about this. As for now, this hopefully will help those who didn't know about this issue and help them get on the right track to making their app a great one.

