inFuse

h1. Conventions

h2. Terminology

* A *pose* is the position of a given frame with respect to a reference frame. Whenever one writes or says the word "pose", one must say the "pose of which frame with respect to which frame".

* A *transform* is the 3D transformation from one frame to an other one. Whenever one writes or says the word "transform", one must say the "transform from which frame to which frame".

Note these two words are exactly the same concept, they denote the same data structure. They can be used indifferently, provided one use them according to the convention above.

h2. Definition of the various frames

 

Note: of course all Cartesian frames are right-handed (direct). The definition below are somewhat inspired by the "MER coordinate systems":https://pds-imaging.jpl.nasa.gov/data/mer/opportunity/mer1po_0xxx/document/coordinate_systems.pdf

 

h3. Terrain related frames

* The *World Frame* (aka the *absolute frame*) is a frame attached to the considered terrain, which never moves. The convention is the East-North-Up (ENU): the x axis points eastwards, the y axis points northward, and the z axis points up. Its origin is located at the surface that models the considered planet shape (WGS84 currently on Earth, which is close to the mean sea level).

* The *Site Frame* is a fixed frame associated to an area in which the rover evolves. It follows the ENU convention, with an altitude reference local to the site (note that often the site frame and world frame are considered to be the same). Over several days, there can be several sites defined.

* The *Mission Frame* is a horizontal frame defined locally for a given mission (e.g. "Reach a goal located _x m_ forward"). Its orientation with respect to a Site Frame may be non null (its x axis does not necessarily points eastward)

h3. Robot related frames

* The *Robot Body Frame* (aka the *RBF*) is a frame attached to the robot body. Its origin is the center of the robot, at the ground level when the robot lies on a flat ground. Its x axis points forward, its y axis points leftward, and z axis points upward

* The *Robot Redressed Frame* is a frame attached to the robot body at the same origin than the RBF (center of the robot), with its x and y axes being horizontal. This is typically the frame in which the Rover Map is expressed.

* Sensor Frames

To each sensor is associated a Reference Frame, in which the acquired data are expressed prior to any transform.

** Camera: the frame associated to a camera follows the "OpenCV standard":https://docs.opencv.org/2.4/modules/calib3d/doc/camera_calibration_and_3d_reconstruction.html 

** Stereovision: the frame associated to a stereovision bench is the frame of the left camera

** Lidar

h2. Definition of the various maps

* the "Rover Map" is the Digital Elevation Map (DEM) built from a single point cloud observation (be it stereo or LIDAR). It models the close environment of the robot for this one observation. 

* the "Fused Rover Map" is the DEM built from the integration of the Rover Map into the internal model. It models the close environment of the robot from all the observations taken of this environment.

* the "Fused Total Map" is the DEM built from the integration of all the rover maps over time. It models the entirety of the environment perceived by the robot during the mission.