If positions are in the local frame of ENU (or ECEF - either works), how do you rotate that to the body frame of the device?
Logically, I thought that I could use the pitch, roll, and yaw angles of the body frame to form a rotation matrix to apply to ENU positions. However, when I apply the rotation matrix, the positions seems to grow cumulatively as shown in the figure. The across track variations grow over time with each sinusoid wave
I would like to be able to rotate the positions rather than the velocities in this instance because the positions are obtained from a mechanization process and filtered.