Paul Furgale

I think you implemented it for the sparse QR: https://github.com/ethz-asl/aslam_optimizer/blob/master/aslam_backend/include/aslam/backend/implementation/Cholmod.hpp#L402-L442 Does it work the same for sparse Cholesky?

See here for a possibility of how we could implement M-Estimators. They have to return p(x), p'(x), p''(x). http://homes.cs.washington.edu/~sagarwal/ceres-solver/stable/modeling.html#LossFunction

Here is where the M-Estimator is applied: https://github.com/ethz-asl/aslam_optimizer/blob/master/aslam_backend/src/ErrorTerm.cpp#L20-L24

These also need attention: https://github.com/ethz-asl/aslam_optimizer/blob/master/aslam_backend/include/aslam/backend/implementation/ErrorTerm.hpp#L169-L192

And https://github.com/ethz-asl/aslam_optimizer/blob/master/aslam_backend/src/ErrorTermDs.cpp#L144-L167

Maybe other places as well. Thanks!

Would you keep the original curves underneath? If so, this is extremely easy to implement. Yes! If this is useful, we could add this. However, there would have to be...

I wouldn't make this a high priority.

No, but we emailed Frank requesting access to the bare metal gtsam. You should do the same.

We are probably just missing Euler Rodriguez. Also http://en.wikipedia.org/wiki/Cayley_transform Also #23