Given a gravity vector, compute the rotation from the IMU reference coordinate to this vector.

Given the static acceleration values, one can get rotation matrix from global coordinate to IMU coordinate.
If plot=True, you can get the 3D plot illustrate the relation between global coordinate and IMU coordinate.

from gram_schmidt_tools import *


if __name__ == '__main__':
    # Given the random reasonable values read from accelerometer
    accel = np.array([4.29, 5.34, 7.02])
    rot_g_to_imu = gram_schmidt_process(accel, plot=False)
    print('\nGlobal to IMU:')
    print(rot_g_to_imu.dot([1, 0, 0]))
    print(rot_g_to_imu.dot([0, 1, 0]))
    print(rot_g_to_imu.dot([0, 0, 1]))

Orthonormal basis in IMU coordinate:
X(e2): [ 0.89927159 -0.26480864 -0.34811923]
Y(e3): [-3.48731968e-17  7.95899952e-01 -6.05428168e-01]
Z(e1): [0.43739069 0.54444435 0.71573021]

Global to IMU:
[ 0.89927159 -0.26480864 -0.34811923]
[-3.48731968e-17  7.95899952e-01 -6.05428168e-01]
[0.43739069 0.54444435 0.71573021]

• Use cross product method to get the same result.

1. Gram-Schmidt Process - Wiki
2. Animation Of Gram-Schmidt Process - zhihu

Welcome to contact me for any further question, below is my gmail address: [email protected]