Open gauteh opened 2 years ago
Mmmh, unfortunately no, I do not have such data available - I have always taken for granted that these Kalman like filters are correctly implemented and "just worked", and done system level testing by comparing to commercial buoys on the "end result" ie SWH and Tp. I think it is a reasonable hypothesis, as a large number of projects use these filters, but agree it would be best to test.
A possibility could be that I generate some data from a VN100 IMU for you, that you can use as test data? I could generate data for accel, gyro, magnetometer, accel X, accel Y, accel D quaternion etc. Then this would be enough material to perform a validation against the VN100 Kalman filtering? :)
That would be very useful!
On Wed, Dec 1, 2021 at 12:04 PM JR @.***> wrote:
Mmmh, unfortunately no, I do not have such data available - I have always taken for granted that these Kalman like filters are correctly implemented and "just worked", and done system level testing by comparing to commercial buoys on the "end result" ie SWH and Tp. I think it is a reasonable hypothesis, as a large number of projects use these filters, but agree it would be best to test.
A possibility could be that I generate some data from a VN100 IMU for you, that you can use as test data? I could generate data for accel, gyro, magnetometer, accel X, accel Y, accel D quaternion etc. Then this would be enough material to perform a validation against the VN100 Kalman filtering? :)
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gauteh/ahrs-fusion/issues/1#issuecomment-983528620, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAN366N7FZM5FUU6UY3TW3UOX6LTANCNFSM5JEBX4QA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
I could probably test against madgwick and mahoney too. Those are ready impls for in Rust and should give not too far off for smooth enough data.
Sounds good. That could actually be a neat little technical note and become a benchmark test case :) .
About these tests:
If you already have 15 mins that's more than enough to smoke out any glaring bugs!
By the way, if you're looking for some rust to do maybe you want to try and the data for an integration test? 😁
ons. 1. des. 2021, 14:06 skrev JR @.***>:
About these tests:
- it is simplest for me to provide you for now with test files that are 15 minutes long; this is a bit of a limitation I guess, but I suppose that this is enough for doing a few "high level tests", right? :)
- I think I will be able to provide you with some test files that span 1 hour, maybe a couple of hours, too. Guess this would be enough for some "long term testing"?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gauteh/ahrs-fusion/issues/1#issuecomment-983621936, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAN36ZRRDMDN2VSJB5MLFLUOYMULANCNFSM5JEBX4QA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
Sounds good :) . If you think 15 minutes are enough, I should be able to create some data this weekend by going on a walk with a VN100 :) . Also, if you are ok looking at some "boring" data (not much signal there though...), there are some data there (this is as much a reminder to myself :) ): https://raw.githubusercontent.com/jerabaul29/LoggerWavesInIce/master/Logger_GPS_SD_VN_Binary_output_Parser/Example_parsed_data/F00379_B
That could be fun :) . To be honest, I have not written any rust in a couple of months, and even then it was only a bit of exercism exercises, but I need to start ( / start again) at some point :) .
Yeah, anything including boring is good.
ons. 1. des. 2021, 22:39 skrev JR @.***>:
That could be fun :) . To be honest, I have not written any rust in a couple of months, and even then it was only a bit of exercism exercises, but I need to start ( / start again) at some point :) .
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gauteh/ahrs-fusion/issues/1#issuecomment-984078222, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAN3666MT4MFPHUARLVNLLUO2I2DANCNFSM5JEBX4QA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
There are no tests in the original repo, at a minimum I would like to have:
@jerabaul29: do you by any chance have some meaningful input data with output data for the ahrs nxp filter? This would also be pretty useful for validation of the C/c++ lib since the code is unmodified apart from making it easier to bridge through C (without allocations).