IMU's values

[Roberto Foletti]

Hello,

thanks for releasing the IMUSIM project!

I'm using it to write my graduation thesis, but I found some troubles..
I'm using the ASF/AMC file in the Mocap's database because I want to use the IMU's values to test others orientaion algorithms in Matlab. Is there a way to save the simulated IMU's values and the quaternions of the joints?
I simulate the walk using the database source by an IdealIMU and I see some glitch in the tibia accelerometer's values. Is it right or there's something wrong?
Is the IMU's orientation unvaried for all the simulated IMUs of the joints?

Thank,
Roberto

[Kjartan Halvorsen]

Roberto,

yes, indeed you can save the data to matlab .mat files. Here is how I did it:

import scipy.io as sio # For writing .mat files

import itertools

# Assuming you have simulated IMUs in code above. In my case imu1 and imu2 

 

# First write to numpy array, then write array as matfile
shape_ = imu1.accelerometer.rawMeasurements.timestamps.shape
imu1.accelerometer.rawMeasurements.timestamps.shape = (shape_[0], 1)
imudta = np.hstack(( imu1.accelerometer.rawMeasurements.timestamps, 
                               imu1.accelerometer.rawMeasurements.values.T,
                               imu1.gyroscope.rawMeasurements.values.T,
                               imu2.accelerometer.rawMeasurements.values.T,
                               imu2.gyroscope.rawMeasurements.values.T))

# The true trajectories of the two imus
# Columns are qx, qy, qz, qw, dx1, dy1, dz1, vx1, vy1, vz1, dx2, dy2, dz2, vx2, vy2, vz2

truedta = np.zeros((imudta.shape[0], 16))
for imu, tdta in itertools.izip((imu1, imu2), (truedta[:,4:10], truedta[:,10:])):
   for t, i in itertools.izip(imu.accelerometer.rawMeasurements.timestamps, range(tdta.shape[0])):
     qt = imu.trajectory.rotation(t)
     dt = imu.trajectory.position(t)
     vt = imu.trajectory.velocity(t)

     truedta[i,0] = qt.x
     truedta[i,1] = qt.y
     truedta[i,2] = qt.z
     truedta[i,3] = qt.w
     tdta[i,:3] = dt.ravel()
     tdta[i,3:] = vt.ravel()

# Save in mat format
sio.savemat("twoimus.mat", {'imudta':imudta, 'truedta':truedta})

I can't help you with your second question, though.

Cheers,

Kjartan

[Roberto Foletti]

Thank you very much Kjartan!!!!

Roberto