Measuring position and velocity using 6-axis Accelerometer / Gyro

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Measuring position and velocity using 6-axis Accelerometer / Gyro

1,941 Views
mukulchandwani
Contributor I

I came across a post on the freescale forum wherein a document on "implementing position algorithms using accelerometers" was shared. In that document, there was statement on the 3-axis accel, claiming that "additional processing would be required to null the earth's gravity effect". I was wondering if you had a document outlining the algorithm measuring position and velocity using a 3 axis accel, reducing the possibility of skewed data. Also, since gyros help in establishing an orientation, does it help in improving the accuracy of the algorithm output?

Labels (1)
0 Kudos
3 Replies

1,023 Views
michaelestanley
NXP Employee
NXP Employee

Mukul,

I found the document you mentioned, which was written back in 2007.  We've learned a lot since then.  I generally try to discourage people from attempting a double integration for anything beyond a few seconds duration (which is what that app note does for a mouse).  Integrated noise is mathematically guaranteed to explode over time if you don't have some other means (like GPS or radio beacons) to periodically rebaseline your position.

And yes, for the general 3 dimensional space, you cannot rely on just a 3-axis accelerometer.  Generally full 9-axis sensor fusion is required because you have to map acceleration measured in the sensor's frame of reference back to the global frame.

Regards,

Mike

0 Kudos

1,023 Views
mukulchandwani
Contributor I

Thanks for the prompt response Michael.

Could you shed some more light on the use of a 3 axis gyroscope along with a 3 axis accel to accurately determine position and velocity. Could a gyroscope help periodically rebaseline the device's position.

0 Kudos

1,023 Views
michaelestanley
NXP Employee
NXP Employee

A gyro provides no position information at all.  What is does provide is an independent way to measure rotation, which is necessary to separate linear acceleration from gravity (an accelerometer measures both).  In fact, the earth's gravity is used to calibrate accelerometers because it's easy and free.  That's why almost all accelerometers provide outputs in terms of "gravities".

0 Kudos