Adrian Figueroa

I made my changes public here: https://gitlab.com/mic_public/bayes-estimate/-/merge_requests/1 Sorry for the formatting clutter.

Add likelihood calculation for use with IMM

Hi! Sorry for the delay. Yes, it works with no_std. I also tested 0.19.0 with success. Feel free to close this issue.

Hi! Sorry for the delay. Yes, it works with no_std. I also tested 0.19.0 with success.

Hello again! I did some more testing and I may have run into numerical issues. When choosing small values for alpha, the sigma point spread is very small. I ran the UKF on f32 number format, so that can cause issues. When alpha is bigger, it appears to be stable. On a new project with a totally different process and implementation, both 0.17.0 and 0.18.1 behave similarly (except if I choose alpha to be too small). Maybe one more thing: there is an updated version of nalgebra, in case you want to...

Hello again! I did some more testing and I may have run into numerical issues. When choosing small values for alpha, the sigma point spread is very small. I ran the UKF on f32 number format, so that can cause issues. When alpha is bigger, it appears to be stable. On a new project with a different totally implementation, both 0.17.0 and 0.18.1 behave similarly (except if I choose alpha to be too small). Maybe one more thing: there is an updated version of nalgebra, in case you want to move to it....

Yes, I can use a fork that I modified already, no problem. Can I somehow help with testing the unscented approach? By the way, the output of the predict step didn't change much with the new implementation, only observe.

Yes, I can use a fork that I modified already, no problem. Can I somehow help with testing the unscented approach? By the way, the predict step didn't change much with the new implementation, only observe.