Describe the bug
Some dimension mismatching was found in sunlineSEKF.c and resolved in PR #1350. Beyond these, there seem to be some more dimension mismatching that would be worth investigating but were beyond the scope of that PR. Issues spotted were in use of memmove() and mSetSubMatrix() usage. In particular on line 162, measMat and x are 32x5 and 5x1 but are claimed by SKF_N_STATES to be 32x6 and 6x1 meaning Hx gets garbage values.
To reproduce
This was not producing any errors, which is why it warrants a more in depth investigation to see what could have been silently breaking.
Expected behavior
Dimension specifications should be tied to module level definitions and not hardcoded. Additionally, matrix copying and multiplication needs to have well defined dimensions as to not write nonsensical values.
Desktop (please complete the following information):
- OS: Macos
- Version 26.3.1
- Python version 3.13.8
Describe the bug
Some dimension mismatching was found in sunlineSEKF.c and resolved in PR #1350. Beyond these, there seem to be some more dimension mismatching that would be worth investigating but were beyond the scope of that PR. Issues spotted were in use of memmove() and mSetSubMatrix() usage. In particular on line 162, measMat and x are 32x5 and 5x1 but are claimed by SKF_N_STATES to be 32x6 and 6x1 meaning Hx gets garbage values.
To reproduce
This was not producing any errors, which is why it warrants a more in depth investigation to see what could have been silently breaking.
Expected behavior
Dimension specifications should be tied to module level definitions and not hardcoded. Additionally, matrix copying and multiplication needs to have well defined dimensions as to not write nonsensical values.
Desktop (please complete the following information):