I started the week by completing the Kalman filter tutorials I was planning on doing. I finished watching all the Michel van Biezen Kalman Filter tutorial videos I mentioned last week (https://www.youtube.com/watch?v=CaCcOwJPytQ&list=PLX2gX-ftPVXU3oUFNATxGXY90AULiqnWT&index=1) and also went over the Kalman filter code on GitHub that Dan initially used as a template for the 3D accelerometer Kalman filter (https://github.com/balzer82/Kalman/blob/master/Kalman-Filter-CA.ipynb?create=1).
After expanding my understanding of Kalman filters, I spent time with Dan finishing the first attempt at the 3D accelerometer Kalman filter code, as planned. We have Arduino code that reads accelerometer data from one IMU using I2C and converts the raw data into m·s−2. The acceleration data is passed into a Kalman filter that attempts to determine new position and velocity (all in 3D).
This past Friday, Dan, Hana and I met with Professor Mukherjee to discuss Kalman filter challenges and next steps. Some topics which we discussed include accelerometer and gyroscope calibration as well as accelerometer and gyroscope Kalman filter integration.
After the meeting, I started doing research on Kalman filters that take in angular velocity readings to determine the new angle. This coming week, I’ll be working on code for a 3D gyroscope Kalman filter. Once the code works reasonably well, I will work with Dan to integrate the accelerometer and gyroscope Kalman filters.
0 Comments