Skip to content

Latest commit

 

History

History
37 lines (26 loc) · 1.9 KB

README.md

File metadata and controls

37 lines (26 loc) · 1.9 KB
Raw

Line Tracking Race

Simulated line tracking robot developed for the final exam of the Robotics course at UniPi.

Usage

You can use the provided full.launch launch file to launch the simulation along with all nodes:

roslaunch dapozzo_line_tracking full.launch

Remember that you can pass launch arguments with the name:=value syntax. Consult the section below to learn about available arguments.

Nodes

Planner Node

The planner node extracts track information from raw camera images. A launch file (planner.launch) is provided. Available parameters:

  • viz (True | False): whether to enable live visualization of planning data with OpenCV;

  • strategy (centerline | centroid): the first planning strategy chooses a point from the track's centerline as waypoint, while the second strategy uses the centroid computed on the detected track shape;

  • error_type (offset | angle): the first approach uses the x-axis offset between the waypoint and the center of the image as error, while the second one uses the angle between the line connecting the bottom-center pixel and the waypoint and the vertical line passing through the image's center. Both errors are remapped into the [-1, 1] range.

Control Node

The control node uses a PID controller in order to accelerate and steer the car based on the provided waypoint error. A launch file (control.launch) is provided. Available parameters (PID parameters are read from config/pid_params.yaml):

  • k_p: PID proportional gain;
  • k_i: PID integral gain;
  • k_d: PID derivative gain.
  • duration: how long this node should run for (in seconds). Useful for performing multiple runs of the same length. The car automatically stops when the duration reaches 0.

Manual Control Node

The manual control node allows for basic keyboard control of the car using the arrow keys.

Logs

Logs in .csv format can be found in the logs directory.