Orientation 3-axis accelerometer from x y z only

Hi all,

i was wondering if it's possible to get an estimation of the orientation of the accelerometer, from x y z

I have a 3-axis accelerometer, with x(t), y(t) and z(t).

From just those 3 components, is it possible to determine the orientation of my accelerometer, eventually with respect to gravity?

Please find attached also a copy of x, y and z

Thanks a lot in advance for your time

Regards

[James Tursa]

What is the accelerometer attached to? Is it at rest or moving around?

[Francesco Lucarelli]

It is attached to a muscle under stimulation

[Mathieu NOE]

hello

I would suggest you to read about inertial measurement units

you would need normally 3 axis accel + 3 rotations (gyros) to be able to fully determine a position vs time

Inertial measurement unit - Wikipedia

[Francesco Lucarelli]

Hi @Mathieu NOEwhat about the orientation with respect to the gravity direction?

[Mathieu NOE]

if you take the mean of x,y,z data , you will have the projection of the gravity vector on your local X Y Z coordinate system.

from there you should be able to compute the Euler's angles

transformation matrix is :

you can start by computing theta as the 3,3 term of this matrix can be used in a straigthforward way

then, phi can be also computed by using term 3,1

etc...

[Francesco Lucarelli]

you mean form the left side i have the gravity vector [x',y',z'] (mean of x y and z) and from the right side i have my [x,y,z]

so [x',y',z']=A[x,y,z]

computing the math i can have the three angles. Is it okay?

Thanks again

Answer 1

[Prabhan Purwar]

Hi Francesco

Rough estimate of the Position and orientation of the device can be predicted by making use of Kinematic equations.

It's under the assumtion that there is no collision taking place and the sampling rate is high to assume a linear motion model.

% Motion Equation
s = @(dt,u,a) u*dt + 0.5*a*dt^2;
v = @(dt,u,a) u + a*dt;

Once velocity components are obtained make use of the following code to get orientation and visualization

velocityAngle = atan2d(velocityY,velocityX);

[u, v] = pol2cart(deg2rad(velocityAngle), 1); 
w = zeros(size(u,1),1);

% Visualization
figure
% Trajectory obtained using motion model
plot3(trajectoryX, trajectoryY, trajectoryZ);
hold on
% Orientation
quiver3(trajectoryX, trajectoryY, trajectoryZ,u,v,w,'LineWidth',0.75,'AutoScaleFactor',1)
view(0,90)
axis equal
hold off
title('Velocity plot')

Some manual adjustments will need to be made according to the setup condition assumed.

Hope it helps!!

Thanks