Question about the AHRS2 and AHRS3

[Neve]

Hi all.

I have a question while reading the mavlink messages for AHRS2 and AHRS3.
I searched imformation on the forum and found the results of roll, yaw, pitch in AHRS2 and AHRS3 show a slightly difference.
Can anyone explain why there is such a difference? What is the reason for this result?

@patrickelectric

32: AHRS {omegaIx : 0.0389807932079, omegaIy : -0.0134896915406, omegaIz : 0.0616827979684, accel_weight : 0.0, renorm_val : 0.0, error_rp : 0.00437930645421, error_yaw : 0.15694104135}
31: AHRS2 {roll : -1.58072090149, pitch : -0.0987562090158, yaw : -0.525687396526, altitude : 0.0, lat : 0, lng : 0}
31: AHRS3 {roll : -1.47963392735, pitch : -0.104524955153, yaw : 2.29476451874, altitude : 0.0, lat : 0, lng : 0, v1 : 0.0, v2 : 0.0, v3 : 0.0, v4 : 0.0}

Thanks for your help!

[EliotBR (Eliot)]

Hi @Neve,

In short, different AHRS instances are run using different IMUs and/or different algorithms, and the ‘most trusted’ (by its data consistency) is used by the flight controller (the results of which appear in the ATTITUDE message). They aren’t expected to be exactly matched, and they may have quite significant differences.

I’m not actually sure on the specifics for which AHRS is using what (since it depends on the flight controller board and the firmware in use), but I’m aware that

• the Pixhawk has two separate IMUs
• there are multiple sensor fusion algorithms that have been used throughout the development of ArduPilot