These projects are running Robot Operating System (ROS) with the Turtlebot3 utilizing the raspberry pi camera, motors, and lidar sensor. In the lab, we worked on SLAM techniques such as Kalman Filter and Bayesian Localization as well as P, PI, PID and bang-bang controllers, and pose-to-pose control using odometry.
In this lab my partner and I implemented the kalman filter which uses odometry measurements to predict where the robot is and then updates this prediction using a measurement model from the data received using the lidar sensor (cylinder as reference) on the robot. In the video you see the turtlebot stopping on the markings, this is accomplished using the kalman filter to predict its pose. The robot uses path following to stay on track.
The node which is running is subscribed to another node that takes in the raspberry pi camera data and outputs an index of where the black line is. Our program uses this index to control the necessary response of the robot's angular velocity to follow the line. This is done using a PID controller whose output is the angular velocity. To optimize the speed at which the robot goes through the path we change the linear speed of the robot depending on the size of the correction from the PID controller.