Is it possible to get position data of an IMU 9D: Accelerometer / Gyroskop /Magnetometer without GPS ?

Right now I am using the Matlab Sensor Fusion and Tracking Toolbox. Therefore I am using the ahrsfilter-function to get the orientation data of an IMU. Right now I am asking myself if its possible to use the orientation data (as quarternions or rotation matrix) to substract the gravity vectors acceleration from my linear acceleration.

Answer 1

[Brian Fanous]

You can use the orientation data from the ahrsfilter to compute the gravity vector and derive a linear acceleration, however the filter models linear acceleration as zero-mean Gaussian noise. This is a common assumption when a GPS or similar is not included in the sensor setup. You should plan to operate the filter under those conditions.

However, in general, double integrating the accelerometer measurement (or the linear acceleration you intend to compute) is not a good way to compute position without the aid of some other sensor, like a GPS. MEMs accelerometers have some bias in their measurements. If you were to simply double integrate the measurement, then you would also be double integrating the accelerometer’s constant bias. The position you calculate would have an error that increases quadratically.

[Matthias Schmidt]

Thank you for you answer! The problem is, that I want to track the jaw motion of a human and for this application a GPS sensor would be imprecisely. Now I will try to see the jaw-motion as an oscillation and I will try to compensate the accelerometer's constant bias by substracting the accelerometer data with an interpolated accelerometer bias function.