NavX MXP Accuracy

[KyloRen42]

Before we buy a NavX MXP, we’d like to know the following:

Is it reasonable for us to expect that using the NavX to determine the robot’s position relative to the field will remain reasonably accurate for a two and a half minute FRC round?

[cad321]

When you say reletive to the field, are you referring to heading, your xy coordinates, or both? If it is just your heading I would think you’d be ok, however if you’re trying to us the navX to track your xy coordinates aswell, you will most likely deal with a substantial amount of drift.

Disclaimer: I have never used the navX board, this response is purely based on what I have read others post here on chief.

[electronicsdude]

@KyloRen42

Before we buy a NavX MXP, we’d like to know the following:

Is it reasonable for us to expect that using the NavX to determine the robot’s position relative to the field will remain reasonably accurate for a two and a half minute FRC round?

We used one positioned on to of your robot relatively centered to run our mechanum wheels in a field oriented drive. Just set up to recalibrate when the match starts we never had any issues with the rotation drifting. In my opinion plenty accurate for a 2-minute match.

[slibert]

@KyloRen42

Before we buy a NavX MXP, we’d like to know the following:

Is it reasonable for us to expect that using the NavX to determine the robot’s position relative to the field will remain reasonably accurate for a two and a half minute FRC round?

If you are referring to orientation (rotation of robot with respect to the field), absolutely yes. Hundreds of teams have used NavX-MXP for this.

[KyloRen42]

@slibert

If you are referring to orientation (rotation of robot with respect to the field), absolutely yes. Hundreds of teams have used NavX-MXP for this.

What about orientation and x,y co-ordinates?

EDIT: Never mind, I think I found my answer:
“Accelerometer data from the navX-MXP’s onboard MPU-9250 are double-integrated by the navX-MXP firmware to estimate displacement, and are accurate to approximately 1 meter of error during a 15 second period.”

Frequently-asked Questions

So it’s a definite no.

[slibert]

@KyloRen42

What about orientation and x,y co-ordinates?

EDIT: Never mind, I think I found my answer:
“Accelerometer data from the navX-MXP’s onboard MPU-9250 are double-integrated by the navX-MXP firmware to estimate displacement, and are accurate to approximately 1 meter of error during a 15 second period.”
Frequently-asked Questions | navX-MXP

So it’s a definite no.

There’s much discussion on chief Delphi on how to address this. It’s pretty well known that there is currently no single sensor that can track position to the required accuracy for FRC. Teams are having success by a solution that fuses multiple sensors:

• wheel encoders + navX-MXP
• vision processing w/camera w/navX-MXP
• a combination of all of the above
• using a “follower wheel” that has a wheel speed encoder (for distance traveled) and an angle encoder (for current direction of travel).

[Maxwell777]

@slibert

There’s much discussion on chief Delphi on how to address this. It’s pretty well known that there is currently no single sensor that can track position to the required accuracy for FRC. Teams are having success by a solution that fuses multiple sensors:

• wheel encoders + navX-MXP
• vision processing w/camera w/navX-MXP
• a combination of all of the above
• using a “follower wheel” that has a wheel speed encoder (for distance traveled) and an angle encoder (for current direction of travel).

If you want absolute position over the whole duration of a match, I don’t think encoder wheels will cut it this year, with all the bumps and various defenses, not without re-zero-ing repeatedly.