Design: Map Server

The map server (package cpm_lab_map_server, executable cpm_lab_map_server_node) loads one Lanelet2 map and publishes it to the lab. The maps of the lab are in the package cpm_lab_maps; the map format is described in Lanelet2 Maps.

Source: src/maps/cpm_lab_map_server/ — src/MapServer.cpp (loading and projection), component/map_server_node.cpp (ROS interface), component/MapServerParameterRos.hpp (parameters).

Loading

The control flow starts the map server with the absolute path of the selected map in map_path: a file of $CPM_LAB_HOME/maps/ or of cpm_lab_maps/maps/. Started by hand, map_path may also be relative to the share directory of cpm_lab_maps:

ros2 run cpm_lab_map_server cpm_lab_map_server_node --ros-args -r __ns:=/lab -p map_path:=$HOME/cpm_lab/maps/mylab.osm

The map must contain a node with the tag origin (latitude and longitude of the lab origin) and may contain a node orientation. The map server projects the geographic coordinates into metric lab coordinates with these reference nodes, so the map coordinates match the IPS, whose origin is the outer corner of the calibration board (see IPS Camera Calibration). Loading errors stop the node with an error message.

Interface

Topic	Type	Notes
/lab/map	cpm_lab_map_msgs/Lanelet2MapBin	the binary Lanelet2 map with its name; reliable, transient local, published once
/lab/system_trigger	cpm_lab_lab_msgs/SystemTrigger	a MAP_REQUEST republishes the map
/tf_static	map under world	frame_id and parent_frame_id are parameters

Because the map is transient local, every late subscriber receives it: the bridge forwards it to the planner domain, the control flow reads the spawn points of the simulated vehicles from it, and the recorder stores it with the run.

Parameters (config/default.yaml): map_path, frame_id (map) and parent_frame_id (world).

Using the Map

• C++: cpm_lab_lanelet2_extension_cpp converts the message back into a lanelet::LaneletMap (lanelet::utils::conversion::fromBinMsg) and provides the projector.

• Python: cpm_lab_lanelet2_extension_py deserializes the message for lanelet2_python, as the example planner and cpm_eval do.

• RViz: cpm_lab_map_server_visualization turns the map into a MarkerArray on /lab/map_markers, scaled by map_scale (18.04) for the display. The colors and layer heights are parameters in its config/default.yaml.

Adding a Map

For a lab, put the .osm file into ~/cpm_lab/maps/; no rebuild is needed, see Your Own Maps and IPS Configurations. Maps for all labs go into cpm_lab_maps/maps/ with a changelog entry. Check a new map with the tools in Lanelet2 Maps before using it with vehicles.