search

dstl / Stone-Soup

A software project to provide the target tracking community with a framework for the development and testing of tracking algorithms.
https://stonesoup.rtfd.io
MIT License
384 stars 126 forks source link

Adding an example that shows linearised ODE as transition model #952

Closed A-acuto closed 2 months ago

A-acuto commented 4 months ago

This PR adds an example that shows how to create linearised transition models from non-linear time continuing dynamics using Van Loan methods.

A more detailed implementation of linearisation method will be available in a future PR coming from a different work package.

A-acuto commented 4 months ago

Hi @sdhiscocks , thanks for the comments. I checked the Q matrix and I think I got an index wrong, now the track is guided by the measurements (which are pretty noisy) but it is not following as before.

Regarding your comment on the self.dA : I am not convinced on how to approach this issue. I might try simpler scenarios (eg smaller scales) and see if self.dA creates any issue.

For the stone soup installation I followed a recent example PR #886 for the clean installation, however if you feel like something different I can amend the current example.

simonrmaskell commented 2 months ago

Would it be useful for this PR to be augmented with another example, eg involving the model in this paper: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6290578. The paper has a non-trivial analytic result that we could check against: Van Loan wouldn't be expected to get the same answer (since it is unaware of the trigonometric identities used in the paper and uses a Taylor series in state to approximate the ODE and then solves the approximate ODE exactly)?

A-acuto commented 2 months ago

Closing this PR, while developing the new ODE example suggested by @simonrmaskell