rtk_mapper is a ROS 2 package which provides a GUI interface to help with viewing and recording landmark positions using a RTK-enabled GNSS receiver (such as). The specific use-case for this application is Formula Student AI.

The package contains two ROS nodes:

1. mapper: a GUI utility to help map tracks
2. viewer: a GUI utility to quickly view tracks stored in a CSV file

This package requires ROS 2 Galactic and on the following third-party Python packages:

1. pygame
2. numpy
3. pandas
4. matplotlib
5. utm

First, clone this repository into your ROS workspace:

$ mkdir -p ~/dev_ws/src
$ cd ~/dev_ws/src
$ git clone [email protected]:siliconlad/rtk_mapper.git

Next, define the $EUFS_MASTER environment variable to point to your ROS workspace:

$ export EUFS_MASTER=$HOME/dev_ws

Finally, we need a map directory in $EUFS_MASTER. This is where mapper saves all maps:

$ mkdir -p $EUFS_MASTER/map

mapper works by reacting to different keyboard inputs to create a map composed of markers. Markers can be one of the following types and represent different concepts:

1. BLUE: represents a blue cone
2. YELLOW: represents a yellow cone
3. ORANGE: represents a orange cone
4. BIG_ORANGE: represents a big orange cone
5. CAR_START: represents the rough starting position of the car

The first 4 marker types are self-explanatory. They allow the captured location to be associated with a cone color. CAR_START is used to help transform the GPS coordinates of each map into a 'normalized map'. In a normalized map, the GPS coordinates are first transformed into UTM coordinates. Then each marker is translated so that CAR_START is at (0,0). Finally, a rotation is applied so that the BIG_ORANGE cones (specifically the average position of the 4 closest BIG_ORANGE cones to CAR_START) indicating the start/finish line lies along the x-axis. This format is useful because coordinates are in meters and the car starts pointing along the x-direction (assuming a straight between CAR_START and the start/finish line, which is reasonable I think).

mapper operates in two modes:

1. Normal mode: Normal mode is the mode you will be using most of the time. With this mode, you can add new markers to the map and delete the last recorded marker to correct a mistake.
2. Update mode: Update mode is used to modify a map. In this mode you can select a cone by pressing TAB. You can then delete the selected cone, or update its position. You cannot add new markers in this mode.

When the program is quit, two CSV files are saved to $EUFS_MASTER/map (see CSV File Formats section).

Assuming you have the necessary packages installed, to running mapper is very simple. Simply build, source and run:

$ cd $EUFS_MASTER
$ colcon build
$ . install/setup.bash
$ ros2 run rtk_mapper mapper

At this point, a GUI should show up with an empty matplotlib figure.

The complete list of keys and their meaning in the two different modes are as follows:

The mapper node subscribes to the /fix topic. It must have type sensor_msgs/msg/NavSatFix.

mapper has the following parameters:

viewer is a simple program which takes a path to a CSV file and plots it in the GUI. No operations can be performed on the map. This is a read only mode of the map.

viewer has the following parameters:

The two nodes in this package operate on the following CSV file formats.

The first is a raw, unprocessed CSV file of the following structure:

The second is a CSV file of the 'normalized map' described in the introduction. The format is what is used in eufs_sim, and hence can be loaded directly into it.