Add generic pose sensor model

[hidmic]

Feature description

Lots of beacon-based tracking systems (ultrasonic, UWB, even GPS after some projection like UTM) yield position and orientation w.r.t. some frame in Euclidean space, with varying degrees of accuracy that may itself vary depending on where robot and beacons are.

Beluga can (and should) support this kind of measurement. A first order approximation would be a generic sensor model built around a gaussian pose distribution. In ROS terms, a sensor model that can take a geometry_msgs/msg/PoseWithCovariance measurement.

Implementation considerations

We need to define how to reconcile the global frame of the particle filter with that of the system that reports these poses. Should it be the same? Should it be a user-defined fixed transform? If so, how to would one go about calibrating it?