search

nansencenter / DAPPER

Data Assimilation with Python: a Package for Experimental Research
https://nansencenter.github.io/DAPPER
MIT License
338 stars 119 forks source link

Unscented Kalman filter implementation? #64

Open cgrudz opened 3 years ago

cgrudz commented 3 years ago

I was wondering if there are any current plans to implement the unscented Kalman filter for comparison with other ensemble-based methods. I am planning on using DAPPER with stats, maths and engineering students in a DA class and I think the implementation of the UKF would be nice for the engineering students to see a method that is also familiar in a lot of mechanical and electrical engineering literature. Having a relevant comparison on standard benchmarks would also be nice in order to discuss the strengths and weaknesses of this approach versus other ensemble-based schemes like ETKF.

patnr commented 3 years ago

I have been wanting to do this for a long time, especially including comparisons. I could try to get to it this next week, at least some variant of it. I think I'd do comparisons for a Lorenz model, or possibly the KS model, seeing as that one has some engineering applications too. What do you think?

cgrudz commented 3 years ago

I think it would be great, there's no hurry though on my end, I don't foresee needing to go over this until September or October in my own lesson planning.

Thanks!

Colin

On 3/10/21 12:10 PM, Patrick N. Raanes wrote:

I have been wanting to do this for a long time, especially including comparisons. I could try to get to it this next week, at least some variant of it. I think I'd do comparisons for a Lorenz model, or possibly the KS model, seeing as that one has some engineering applications too. What do you think?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/nansencenter/DAPPER/issues/64#issuecomment-796018487, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABXIPFXUMOY7CASWGLFXEOLTC7G4LANCNFSM4Y6VEHUA.