1
$\begingroup$

It seems like there is no shortage of sensor fusion libraries that can fuse basic common odometric sensors like IMUs and GPS, but none that can fuse from an arbitrary number of sources, especially with different update rates.

For example, let's say that I have:

  • 3 IMUs, for redundancy, but also different models with slightly different capabilities, with varying update rates between 500 and 1500 Hz
  • 2 3-axis magnetometers with update rates of 300 Hz
  • 1 GPS module at 10 Hz, but with RTK corrections at 5 Hz
  • 2 low res cameras doing stereo visual odometry at 120 Hz
  • 1 high res camera doing visual odometry at 45-60 Hz

Are there any fusion algorithms or libraries that gracefully handle a wide variety of sensors with different capabilities and update rates, without having to downgrade the fusion output to the lowest data rate? Would it be more effective to upsample via extrapolation or predictive methods before feeding to the fusion algorithm?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
0
$\begingroup$

I suggest to check the open source drone code PX4. You will need a fusion time horizon to solve this problem. Hope this will be helpful.

https://docs.px4.io/main/en/advanced_config/tuning_the_ecl_ekf.html

https://github.com/PX4/PX4-Autopilot

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.