search

JuliaRobotics / IncrementalInference.jl

Clique recycling non-Gaussian (multi-modal) factor graph solver; also see Caesar.jl.
MIT License
72 stars 21 forks source link

Prior standardization #1500

Open GearsAD opened 2 years ago

GearsAD commented 2 years ago

As a user I'd like priors to be standardized across the various dimensions, either one prior for all (like Prior?) or a set of priors that are named similarly.

dehann commented 2 years ago

Prior will not be able to be used as a standard. There is a newly created IIF.ManifoldPrior which would be better but is clunky. The preferred approach for ease of use is to use names (some might end up as aliases of ManifoldPrior):

GearsAD commented 2 years ago

Great, thanks - keep me in the loop if there change and I'll make sure we accommodate the new factors in the API. Aliases are great then we can deprecate the older ones over time.