Performing Allan variance analysis on accelerometer data

I want to perform Allan variance analysis on accelerometer data for Noise Parameters Identification and was not clear how to implement it using this post: Inertial Sensor Noise Analysis Using Allan Variance - MATLAB & Simulink - MathWorks India post for my application, as it talks only about the gyroscope, which I was able to implement. Will the procedure remain same and do I also have to perform this step:

theta = cumsum(omega, 1)*t0;

Can someone please guide me. Thanks.

Answer 1

[Shubham]

Hey Yadnik

I understand that you want to perform Allan Variance analysis on accelerometer data for Noise Parameters Identification.

Allan variance is a method used to analyse the noise characteristics of a system. The analysis can be done on gyroscope and accelerometer data.

Although the mentioned post is specific to gyroscopes but the same procedure for Allan variance can be applied to accelerometer data. The only key difference is that the gyroscopes measures rate of rotation while accelerometer measures linear acceleration.

As for the code line:

theta = cumsum(omega, 1)*t0; 

This line computes the angle “θ” using the angular velocity “ω” and the time interval “t0”. In the case of an accelerometer, this step is not required as the analysis can be performed directly on the acceleration measurements.

Please have a look at the following file exchange post where the provided script calculates Allan deviation of a time domain signal and is not specific to gyroscope data:

https://www.mathworks.com/matlabcentral/fileexchange/13246-allan

Also refer to the following file exchange post which provides a graphical user interface for Allan variance: https://www.mathworks.com/matlabcentral/fileexchange/55765-avar

Hope this helps!!