Michel Hidalgo

Funnily enough, the `likelihood_field_prob` model in `nav2_amcl` voices [the same concerns about dynamic obstacles](https://github.com/ros-planning/navigation2/blob/9b56556dfd9c50f4046d253317cd0949a4f34583/nav2_amcl/src/sensors/laser/likelihood_field_model_prob.cpp#L81-L83) and ~uses [log-odds representations](https://github.com/ros-planning/navigation2/blob/9b56556dfd9c50f4046d253317cd0949a4f34583/nav2_amcl/src/sensors/laser/likelihood_field_model_prob.cpp#L182-L191)~ not quite log-odds, only a log transform, misread the code.

I will eventually get back to this and try out [Log-PF](https://www.hindawi.com/journals/jece/2018/5763461/) ideas. Money on a numerical issue with likelihood functions and floating point representations.

> I will eventually get back to this and try out [Log-PF](https://www.hindawi.com/journals/jece/2018/5763461/) ideas. Money on a numerical issue with likelihood functions and floating point representations. FYI I finally got around...

> at first sight it looks very different from NAV2's amcl node. They gave up :smile:

This has been in the backlog for a while. I wonder if there is path (and interest) for the techniques in https://github.com/fmrico/mh_amcl to be recreated in Beluga. I also wonder...

Not that I know of. That said, the idea of clustering has been floating around for a long while e.g. https://arxiv.org/pdf/cs/0204044.pdf.

Addressed as of #275.

Considering we are moving away from mixins, I'm inclined to close this ticket. @glpuga please do if you agree.

Closing this due to lack of activity.

A substantial amount of work was pursued by @serraramiro1 already. I'll push it past the finish line.