Sensor fusion, stationary condition problem

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

Sensor fusion, stationary condition problem

Jump to solution
877 Views
youngguekim
Contributor III

This is Younggue Kim, a software staff at Doosan infracore, Korea

 

I am writing to ask some advice.

 

I am Using sensor fusion development kit (FRDM-FXS-MULT2-B and FRDM-K64F) and Toolbox for window with USB cable

And when the board is in stationary condition and selecting the Gyro, the board picture is moving and gyro sensor values keep increasing or decreasing.

 

The firmware file’ name is FSFK_K64F.bin in precompiled folder.

And the video was filmed in stationary condition.

 

Please help me. Thank you.

Labels (1)
Tags (2)
0 Kudos
1 Solution
582 Views
michaelestanley
NXP Employee
NXP Employee

Younggue,

You don't state in your posting whether you have "exercised the board" before setting it motionless.  The Gaming Handset movie is rotating more than I expect - unless you failed to rotate the board a few times previously.  That is required to provide the Kalman filter enough information to zero out gyro offset.  Try rotating the board a few times then sitting it down.  You'll still see a bit of rotation as time goes on, because the gyro will drift, and if motionless on a table, there is no information to cancel out yaw offset.  But it shouldn't be as bad as the video.  By the way, do you have another board you can try, just in case we have a bad board.

The gyro stablized movie looks just right to me.  It works better because that algorithm includes a magnetometer to help alignment to magnetic north.

The Rotation video is using just a gyro, and I would actually expect it to have the worst drift performance of the three, as it doesn't even have an accel to help compute gyro offsets.

One other thing you should do, download version 5.00 of the fusion library.  It does a better job of managing gyro drift.  I can tell from the movies you are still on the older release.

Regards,

Mike

View solution in original post

0 Kudos
1 Reply
583 Views
michaelestanley
NXP Employee
NXP Employee

Younggue,

You don't state in your posting whether you have "exercised the board" before setting it motionless.  The Gaming Handset movie is rotating more than I expect - unless you failed to rotate the board a few times previously.  That is required to provide the Kalman filter enough information to zero out gyro offset.  Try rotating the board a few times then sitting it down.  You'll still see a bit of rotation as time goes on, because the gyro will drift, and if motionless on a table, there is no information to cancel out yaw offset.  But it shouldn't be as bad as the video.  By the way, do you have another board you can try, just in case we have a bad board.

The gyro stablized movie looks just right to me.  It works better because that algorithm includes a magnetometer to help alignment to magnetic north.

The Rotation video is using just a gyro, and I would actually expect it to have the worst drift performance of the three, as it doesn't even have an accel to help compute gyro offsets.

One other thing you should do, download version 5.00 of the fusion library.  It does a better job of managing gyro drift.  I can tell from the movies you are still on the older release.

Regards,

Mike

0 Kudos