Is it possible getting x coördinates from an accelerometer measurement?

Hello, I have a dataset from a measurement with an accelerometer. The figure below displays the resultant vector of the X, Y and Z axis.

I would like to know if it is possible getting the x coördinates (Time in seconds) from the beginning when the signal starts getting bigger until the moment it gets back to normal. So this means the first x coördinate would be around 270 sec en the second 570 sec and so on until the end of the signal.

Until now I have done it visually because I have had only two test subjects. But it would be of great help if there is some way to do this automatically.

[Jan]

I cannot resist to mention, that the word is spelled "coordinates". Sorry, I know this does not help you to solve the problem.

[Star Strider]

@Jan —

Shawn Imming is correct. The diaeresis (not umlaut) in English is placed over a second consecutive vowel to indicate that the second vowel is pronounced. The same applies, for instance, in the word naïve. (This is an obscure point that few people discover until they are corrected on their first-year undergraduate English assignments.)

Answer 1

[Massimo Zanetti]

It seems you need to threshold the values of XYZ (that I think is a vector). Fore example, fix a thereshold to T=1 (or some that you believe is reasonable) and select only the values of XYZ that are greater than T, for those elements, get the time coordinate:

T=1;
extracted_coord = x(XYZ>T)

Answer 2

[Jan]

A simple threshold might be triggered by the noisy signal for one or two frames accidently. Therefore it would be helpful to analyse the nature of the noise at first to apply a low-pass filter. Afterwards thresholding is safer.