optimize module: Levenberg-Marquardt / Gauss-Newton via autodiff
Nobody has claimed this yet.
- Dominant language
- C
- Stars
- 23
- Forks
- 2
- PR merge metrics
- No merged PRs in 30d
Description
On-device nonlinear least squares for sensor calibration. What makes this a real differentiator rather than a reimplementation: it can be built directly on numx's own autodiff module (forward-mode dual numbers, reverse-mode static tape) instead of requiring hand-derived Jacobians, which is normally the hard part of shipping Levenberg-Marquardt on a microcontroller.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading numx's autodiff module, including its forward-mode dual numbers and reverse-mode static tape, along with the existing numerical and linear-algebra entry points. Clarify the intended Levenberg-Marquardt or Gauss-Newton API, supported sensor-calibration inputs, and validation criteria before implementation; done should include an on-device-compatible solver without dynamic allocation or external dependencies.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- c
- Domain
- embedded-iot
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 30/100