3D attitude estimation using various filters

Estimate the 3D orientation or attitude from 6 DOF IMU data (accelerometer and gyroscope) using only accelerometer data, only gyroscope data, a complimentary filter, a Madgwick filter, and an Unscented Kalman Filter. (Check the full problem statements here project1a and project1b)

Steps to run the code

• Install Numpy, Scipy, and Matplotlib libraries before running the code.
• To run on the first training data in the Wrapper.py file in the 'main' function set the variables as: IMU_filename = 'imuRaw1' and vicon_filename = 'viconRot1'
• For the other data change the variables accordingly and run the file.
• To generate 3D animations uncomment the specified lines in 'main' function.
• In Code folder:

  python Wrapper.py
  

Report

For detailed description of the math see the report here.

Plots and Animations

For the train data 1, plots and animation showing roll, pitch, and yaw for all the filters:

Remaining plots are present in the report and links to rest of the animations are train1, train2, train3, train4, train5, train6.

References

1. S. O. H. Madgwick, A. J. L. Harrison and R. Vaidyanathan, "Estimation of IMU and MARG orientation using a gradient descent algorithm," 2011 IEEE International Conference on Rehabilitation Robotics, Zurich, Switzerland, 2011, pp. 1-7, doi: 10.1109/ICORR.2011.5975346.
2. E. Kraft, "A quaternion-based unscented Kalman filter for orientation tracking," Sixth International Conference of Information Fusion, 2003. Proceedings of the, Cairns, QLD, Australia, 2003, pp. 47-54, doi: 10.1109/ICIF.2003.177425.

Collaborators

Chaitanya Sriram Gaddipati - [email protected]

Shiva Surya Lolla - [email protected]

Ankit Talele - [email protected]