0

I'm developing a positional tracking system and found references to Madgwick's 2010 paper, which included an open-source method for obtaining Absolute Orientation.

However, upon executing the filter, the gyro data I'm getting seems to be distorted somehow, by a magnitude of ~100 (smaller than I would expect).

I did a bit of searching and found this post talking about what seems to be an alternate form of the original Madgwick filter, however comparing it to the "original 2010 version" showed different values for the Gradient Descent algorithm---as well as most folks talking about their Yaw-values "flipping".

I'm not really sure what version of the Madgwick AHRS I should be using anymore, nor am I certain if anyone has been experiencing the same magnitude/scaling problem I have been with this "older version". Have other folks used the 2010 MARG in C? If so, have you had the same issue of "scaling" in Gyro Heading values?

Community
  • 1
  • 1
Thomas Solley
  • 3
  • 4

1 Answers1

0

In IMU, some registers are usually dedicated to the sensor range. Check it, it may be the source of your distortion.

Note that the sensors must be calibrated, especially the magnetometer. Have a look at this page, it's a working implementation of the Sebastian Madgwick's AHRS algorithm.

Fifi
  • 3,360
  • 2
  • 27
  • 53