Tri Axis Accelerometer Data Help

[Lysworld]

Hi

I am studying Sports Science and researching the use of an accelerometer during specific wieght training type movements. I have a MEMS device (unfortunately I only have mininal details regarding it) which can measure upto 24G and 400Hz and comes with basic software that transfers the data into an excel file.

My questions are as follows:

1. How do I convert the acceleration from G to m/s/s? (I am aware that 1G is technically 9.81, however, I would like to know the 'real-time acceleration' i.e. as the bar is moving - in that as it is stationary this value would be 0m/s/s)

2. How can I integrate this data into velocity and then displacement? (I am aware this may cause an exaggeration of error if done in the time domain - FFT? - and that data needs to be converted to the frequency domain. Is this correct?)

I am looking for any advice/direction to necessary reading that will help me and if possible any example excel files.

I have seen the following thread before: https://www.physicsforums.com/archive/index.php/t-123128.html although I am a novice to this and only have a vague idea of how to do any of the instructions given

1. Remove the mean from your sample (now have zero-mean sample)
2. Integrate once to get velocity using some rule (trapezoidal, etc.)
3. Remove the mean from the velocity
4. Integrate again to get displacement.
5. Remove the mean. Note, if you plot this, you will see drift over time.
6. To eliminate (some to most) of the drift (trend), use a least squares fit (high degree depending on data) to determine polynomial coefficients.
7. Remove the least squares polynomial function from your data.

Also, am I correct in thinking the trapezoidal rule is not appropriate for non-constant acceleration data?

3. Is it essential to run a filter on the data before these transformations? (Would this be a low-pass or Kalman - again, any advice or direction would be hugely appreciated)

Thanks for reading

 

[privitmj]

Hi, what type of motion do you expect to see? Sinusoidal or Random? I assume random? Also, what domain is your data recorded in? Time or frequency?

 

[Lysworld]

Thanks for the reply. It is random (although a general patern emerges for the type of action being performed). All the data is recorded in the time domain.

I could forward you an example file (excel) if that would help, or perhaps upload it to this post?

 

[privitmj]

Just a sample (with units) will work. Also, I don't know much about how you are recording this data. Do you have any info on that. Also, what type of accelerometer is it (ICP or charge)?

 

[Lysworld]

Hi

I have attached an excel file with one movement and a graph, units are g and s. The graph shows two peaks (with some noise?) which can vary in magnitude and time to reach this based on the weight lifted. I was unsure whether the measurements were correct at 12+ g on the second spike, especially when converted into m/s/s.

The accelerometer (all I know is that it is a MEMS device attached to a AA battery) is housed in a foam casing that attaches to a barbell so that the sensor doesn’t pick up vibrations from the battery or transmitter/aerial. We can record at either: 100/200/400Hz.

Let me know if you need any more information will help and I can answer any questions.

Thanks again