Hi @abertier, thanks for sharing your experience.
Regarding using RTIMULib, did you have the code in any repo already available (or a fork)? I’d like to give it a try.
Hi @abertier, thanks for your quick reply. I get a 404 page with the link provided (Might it be private?) Thanks.
it was indeed private… happened by mistake I guess. I first cloned GitHub - RTIMULib/RTIMULib2 locally just as a reference, but I started editing in the repo. Afterwards I wanted to share so I changed origin and pushed to github. (probably I should change the upstream as well to point back to it’s real source). So functionally it’s a fork, but due to the intermediate ‘local clone’ it’s probably not registered as a fork. I’ll need to take a look and fix that.
2 Likes
Hei Andy.
Nice work. As far as I understand you are already working/Trying with the ICM20948 no? How it is going? Also, I guess that calibration will be need it , as TB_RTK recommend it.
Also, once that you have the GNSS+IMU data, How are you planning to do the post-processing? Just curious if you have any open-source recommendation ?
Thanks!
Hello, welcome to the forum.
Calibration is indeed required for good results. I did succeed in getting the data from the imu-chip and to fuse them together into the correct pose with the RTIMULib2 library.
Initially I was only interested in the pose, so I didn’t try to fuse it with GNSS, however if I knew a simple way, I would have tried it as well. So I guess I’m as interested as you are to see some recommendation for GNSS+IMU fusion.
It’s been a while I worked on the project (at least the GNSS-side)…
Hei Andy
Thanks for your quickly answer!
Okey, I will let you know how it goes. At the moment I I’m working with a livox 40 lidar and I’m asking me if make more sense to try to log the imu + gnss data for post processing or process the data of the lidar on the fly so I don’t need to take care of this. The main difference of course is the computational power requiered for this…
I will read a bit to see what can I found already available, specially in the topic of post processing.
Greetings!
I use icm20948 I added other drivers too, and improved performance/accuracy:
3 Likes
Great to see that there are now versions of RTIMULib2 with support for the ICM20948. This thread made me discover yet more forks of the library. One that we have used is GitHub - HongshiTan/RTIMULib2: 9-dof, 10-dof and 11-dof IMU fusion library for Linux systems (development version) because it has a C++ version of the ellipsoid fit (and we have now integrated it in the code, which we will publish as a PR soon).
Anyway, I was wondering how are people still managing to start additional code on the Reach devices given how much they clamp down things. Essentially we run a command as a normal user via ssh from an external computer that is in the same network as the GPS. However there is one robot where we can’t do that because we can’t connect to the GPS using WIFI (for a reason too long to explain here).
Any suggestions on how to run your own code on the GPS in recent versions of the firmware?
1 Like