Dynamic Acceleration

Hi There!

I need to get the value of dynamic acceleration using ADIS16210.

Will it be possible to obtain dynamic acceleration by omitting the value of static acceleration from total acceleration value from ADIS16210?


ADIS16210 is linked to PIC16F877;

I have float point calculations  & sqrt  on board.


  • 0
    •  Analog Employees 
    on Apr 24, 2015 4:55 PM

    If you know the orientation of the ADIS16210, with respect to gravity during the dynamic acceleration, you can use that to calculate and remove the gravity component of the acceleration.  In most systems that I have seen, this requires a gyroscope, which can help estimate changes to orientation during the dynamic acceleration, while having very minimal response to the dynamic acceleration itself. 

  • Thanks for fast response!

    Yes, You are right! I've tried to get dynamic inclination by omitting static value from total acceleration...

    But it didn't work. So the only way is to take gyroscope on board.

    Please correct me if I'll be wrong in my step-by-step solution:

    1. Static position:

    1.1. Turn on Gyroscope & Inclinometer/Accelerometer as Inclinometer;

    1.2. Store received static inclination values;

    1.3. Switch to Accelerometer mode;

    1.4. Store received gravity acceleration values;

    2. Dynamic mode:

    2.2 Calculating dynamic inclination from Gyroscope inclination with static inclination values;

    2.3 Calculating dynamic inclination by omitting static acceleration values from measured dynamic ones

    2.4 Dynamic acceleration = SQRT( DIFF_ACCL_X^2+DIFF_ACCL_Y^2+DIFF_ACCL_Z^2)


  • 0
    •  Analog Employees 
    on Apr 24, 2015 6:26 PM

    That might work for some situations but we typically see the gravity removal function implemented in more continuous estimation processes, such as a complementary filter or extended Kalman filter. 

  • Thanks a lot!

    It was very helpful 8o)

    Just a last question:

    If it's possible, please provide any link to Kalman filter.


  • 0
    •  Analog Employees 
    on Apr 24, 2015 7:07 PM

    No problem. At this time, we don't have a library of Kalman filter routines, as many of our customers are prone to keep them to themselves, as their own IP.  From what I understand, this paper is considered to be a good starting point:


    I would also suggest going through some searches like "MEMS gyro kalman filter" or add "code" to it to see what comes up.  There are a lot of examples for this on the web but understanding how applicable they are to your application will require your perspective.

    I hope that this helps.  If we have earned this, can you mark the most appropriate post with Correct Answer, which is available under the Actions tab on each post.  If we have not earned this yet, we look forward to your feedback on what else we can do to help.  Thanks again for posting!