Soren Sabet Sarvestany
NΨ 1T7 + PEY
University of Toronto
Inertial Measurement Unit
In my dynamics course, while studying the physics of rotations, we were challenged with creating code for an inertial measurement. We were given a list of acceleration measurments from an IMU on a plane (in the plane's frame), and we had to transform this data to the world frame to plot the trajectory of the plane.
While my final results were off (significantly) due to oneconceptual misunderstanding, nevertheless I've included it here because it was a mistake, and I learn from my mistakes.
This project was a practical example of the physics of rotation in action. You can find the problem statement (question # 5), the data from the inertial measurement unit, and my MATLAB code here.
My calculated trajectory was this:
And the solution was this:
The error I made occured in calculating the rotations. I tried to use Poisson's kinemtical matrices to convert the acceleration data from the plane's frame to the world frame. However, I converted the angular rotations from each point individually, not summing them up to form the next instantaneous rotation matrix. This small misunderstanding resulted in huge variations. The lesson has been learned however, and I am now fairly confident in using rotation matrices.