Hello!
I'm experimenting with the Rowberg DMP script for the MPU6050 sensor. The goal is to determine the most exact angles possible during acceleration. Currently, the angles are strongly distorted by acceleration ongoing overtime. I also understand this physically: Keyword: complementary filter and Gyrodrift.
As a result of the change in the pitch angle, the corrected acceleration is also falsified.
In the appendix a measurement of acceleration and deceleration. The expected angles should be close to 0. Here you can watch the error very well.
Due to the angle correction, the corrected acceleration falsely slows down to zero and then reacts completely wrongly with an acceleration change.
In addition, I do not understand why the initial value of the corrected acceleration is 0.1 g smaller than the raw value, although the angles are close to zero.
The raw values are plausible, the speed is determined by GPS.
Any ideas? 