Inconsistent gravity magnitude when orientation changes

[Francisco-Aguilera-xd]

Hi!

I'm really new to using the UM7 sensor, and I'm experiencing an issue with it. Using the sensor’s acceleration measurements, I computed the magnitude of the total acceleration:

sqrt(a_x^2 + a_y^2 + a_z^2)

When the sensor is stationary at 0° roll, 0° pitch, 0° yaw, the result is approximately 9.81 m/s². The problem arises when I start rotating the IMU. For example:

• At 0° roll, 0° pitch, 75° yaw, the magnitude becomes 9.84 m/s².
• At 180° roll, 0° pitch, 0° yaw, it reaches 10 m/s² approximately.
• At 90° roll, 0° pitch, 0° yaw, it measures 9.9 m/s².
• In some configurations, it drops to 9.76 m/s².

In all of these experiments, the IMU was completely stationary.

I don't really understand why this happens. I need to remove the gravity component from the IMU’s measurements to obtain the actual acceleration it is experiencing. Any ideas would be greatly appreciated!

[tonybaltovski]

Sounds like there might be some sensor bias or cross-axis sensitivity. Have you tried to calibrate it? Here are the resources: https://www.pololu.com/product/2764/resources