Tracking 3d gait from accelerometer data

Question

Giovanni Mastromatteo 2016-11-16

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/312534-tracking-3d-gait-from-accelerometer-data

回答： Guillaume 2020-6-17

Hi, I'm testing a LPMS b2-bluetooth-imu board and I get from the device the accelerometer, gyro, magnetometer, Euler angles and Quaternions data.

The idea is to realize something similar to this (https://www.youtube.com/watch?v=6ijArKE8vKU).

Now, after two integrations of the linear acceleration data for three directions, I have found the resultant displacement.

How could I rebuild the step gait combining the result for each axis?

Thanking you in advance for your help in this matter.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Sharah 2016-11-16

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/312534-tracking-3d-gait-from-accelerometer-data#answer_243520

What do you mean by "rebuild the step gait combining the result for each axis"?

Usually you would do double integration on all 3 axis x y z. You shouldn't combined them all together if you want the real movement in all 3 direction. On the other hand, if you only require resultant displacement, then you'll combined the data from all theree axis as in:

xyz = sqrt(x^2 + y^2 + z^2)

And if you want a graphical representation of the gait, you will plot the data from all three axis.

_

SAM

Human Robotics Group

Imperial College London

2 个评论
显示无隐藏无

Giovanni Mastromatteo 2016-11-16

Sorry, I didn't explain well. What i would to track is the gait from the obtained values of displacement in the three axes. I've integrated the arrays of linear acceleration in 3-axis, in the sintax shown below:

vX=cumtrapz(Time, aX);
sX=cumtrapz(Time, vX);
vY=cumtrapz(Time, aY);
sY=cumtrapz(Time, vY);
vZ=cumtrapz(Time, aZ);
sZ=cumtrapz(Time, vZ);

Plotting the resulting arrays for the displacement I obtain these graphs

I wish to track the gait as this example

Thanks for your help

Sharah 2016-11-16

1) Your integration value especially from Z direction looks like there's drift, you might want to remove that

2) the picture you're showing looks like the force measure by a force plate during walking, so not sure what you're trying to see here.

3) If you think about a movement, there will always be a 3D representation of a movement. if you are interested in only the vertical (up and down the movement), just take the value from only that axis (either only x, y or z) depending on your measurement.

请先登录，再进行评论。