Maps and observations compatibility matrices

There exists many kinds of metric maps and observations in MRPT, but some operations involving a map and an observation (for example “inserting” an observation in a map to update it) only make sense for a small subset of map-observation combinations.


The following tables summarize the valid combinations, as implemented so far:

Valid implementations of insertObservation()

List of observations and the metric maps in which mrpt::maps::CMetricMaps::insertObservation() is implemented for them:

Valid implementations of computeObservationLikelihood()

Next follows the list of observations and the metric maps in which mrpt::maps::CMetricMaps::computeObservationLikelihood() is implemented for them. That means that, for example, particle filter algorithms can use those observations to localize a robot in the corresponding map, fusing the information from many sources if several observation-map pairs are valid for a given robot.