Anomalous MMRL Quaternions

jecjec
edited July 8 in Hardware

Hi

Device MMRL Firmware 1.7.3

Setup:
We have our MMRL devices configured to output Quaternions.
The mode is NDOF
Sensor fusion ACC range is set to 16G
Gyro range is set to 2000DPS
Calibration state is reported as 3 for Acc, gyro and magnetometer

Occasionally we receive quaternions where axes flip polarity.
Any ideas as to the cause / remedy?
The examples below were captured using the Java API on an Android device but this has been observed using the Python API running on a pc as well.

1657196745400,2022-07-07T14:25:45.400,72.690,0.003,-0.499,-0.867,0.016
1657196745410,2022-07-07T14:25:45.410,72.700,0.002,0.499,0.867,-0.006

1657196792750,2022-07-07T14:26:32.750,120.040,0.000,0.437,0.880,-0.188
1657196792760,2022-07-07T14:26:32.760,120.050,0.000,-0.437,-0.880,0.188

1657196792890,2022-07-07T14:26:32.890,120.180,0.000,-0.439,-0.879,0.186
1657196792900,2022-07-07T14:26:32.900,120.190,0.000,0.439,0.879,-0.186

1657196792910,2022-07-07T14:26:32.910,120.200,0.000,0.439,0.879,-0.186
1657196792920,2022-07-07T14:26:32.920,120.210,0.000,-0.439,-0.879,0.185

TIA
Best regards
jec

Comments

  • jecjec
    edited July 11

    Hi

    An amendment to the above is that in our recordings we never see any negative 'w' values which we find strange.

    Using other manufacturers' sensors we find that performing similar movements we would tend to see negative 'w' values. These are necessary as they are fundamental in preventing discontinuities in movement tracking

    TIA
    Best regards
    Jeremy

Sign In or Register to comment.