2022-08-12, 02:31 AM
I really dont know but it could be a sensor glitch. This meaning the sensors for odd reason (cosmic ray? heh) start outputing strange values for a while.
Keep us posted, and for ultimate reliability have two systems that can swap... If you can take note of the raw calibration values for compass and accel, and maybe even the gyro bias. Then compare these over time. The results could help make the program smarter.
Already the driver has special logic to reset the sensors in cases where the data seems very wrong and this usually fixed the problem (so I thought) which was somewhat rare and difficult to reproduce but I was able to (by plugging/unplugging i2c enough times randomly), So although this problem should not actually ever occur.. it does but may be a different problem, or maybe the same but just not quite bad enough to trigger the reset.
Keep us posted, and for ultimate reliability have two systems that can swap... If you can take note of the raw calibration values for compass and accel, and maybe even the gyro bias. Then compare these over time. The results could help make the program smarter.
Already the driver has special logic to reset the sensors in cases where the data seems very wrong and this usually fixed the problem (so I thought) which was somewhat rare and difficult to reproduce but I was able to (by plugging/unplugging i2c enough times randomly), So although this problem should not actually ever occur.. it does but may be a different problem, or maybe the same but just not quite bad enough to trigger the reset.