Code and documentation for the task of tracking and localizing a robot on a grid track. (The only Computer Vision based implementation to perform the task.)
The goal of this project is to track the trajectory of a Lego Mindstorm NXT on a custom track and measure the errors in the final state over 20 rollouts.
To track the robot on the track, aruco markers with different IDs are used. Two markers are used, one on the front and the other on the back, to precisely know the orientation of the robot. (This can also be done using a single marker)
Markers of different IDs are added to the 4 edges of the map. The markers are aligned such that the first point of the marker that is detected directly coincides with the edge of the track.
Camera is used to take a video or a photo of the robot at its initial and the final positions.
The complete procedure and the setup is given here.
Images showing all initial and the final postions of the robot on the track.
