Autonomous-Driving-Turtlebot3

Project Goal: The goal of the Autonomous-Driving Turtlebot3 project is to achieve autonomous driving for a ground differential robot by using perception-based techniques. The primary task is to detect and follow lanes to complete a mission. In this project, the robot needs to navigate through a circuit, which includes passing through a low-light tunnel.

Required Libraries and Packages to Start:

• ros-noetic-image-transport
• ros-noetic-cv-bridge
• ros-noetic-vision-opencv
• opencv
• libopencv-dev
• ros-noetic-image-proc
• Autorace package

Steps to Get Started:

1. Installing the Required Packages

• Follow the instructions provided on the emanual robotics website.
• Install the AutoRace 2020 meta package on the Remote PC.
• Install additional dependent packages on the Remote PC.

cd ~/catkin_ws/src/
git clone -b noetic-devel https://github.com/ROBOTIS-GIT/turtlebot3_autorace_2020.git
cd ~/catkin_ws && catkin_make
sudo apt install ros-noetic-image-transport ros-noetic-cv-bridge ros-noetic-vision-opencv python3-opencv libopencv-dev ros-noetic-image-proc

2. Connecting to Turtlebot3

• Connect to the Turtlebot over SSH with the password ’napelturbot'.
• It’s crucial to bring up the Turtlebot before running any nodes related to camera and robot control.

ssh ubuntu@192.168.0.200
roslaunch turtlebot3_bringup turtlebot3_robot.launch

3. Camera Calibration

• Camera calibration involves the following steps:
  • Launch roscore on the Remote PC.
  • Trigger the camera on SBC.
  • Perform intrinsic camera calibration.
  • Perform extrinsic camera calibration.

4. ROS Package Structure:

• The ROS package includes several launch files, such as extrinsic_calibration.launch, intrinsic_calibration.launch, controller.launch, and lane_detection.launch.
• These files are responsible for subscribing to camera topics, performing lane detection, and controlling the robot’s movements.

5. Lane Detection:

• Lane detection is achieved using binary thresholding of the projected image and creating a histogram.
• The robot decides to turn right, left, or move straight based on the values in the histogram.
• The project’s lane detection algorithm is inspired by “The Ultimate Guide to Real-Time Lane Detection Using OpenCV.”

6. Controller:

• The project uses a Proportional-Derivative (PD) controller.
• The controller class subscribes to lane detection topics to get the center point.
• It calculates the error and computes the rotational velocity to minimize the error.
• Linear velocity is fixed, and only the rotational velocity changes.

To Run the Code:

1. Run ROS master on the PC: roscore.
2. Connect to the Turtlebot: ssh ubuntu@192.168.0.200.
3. Bring up the Turtlebot: roslaunch turtlebot3_bringup turtlebot3_robot.launch.
4. Start capturing from the camera: roslaunch turtlebot3_autorace_camera raspberry_pi_camera_publish.launch.
5. On the PC, run the following files in separate terminals:
   • roslaunch group_4 extrinsic_calibration.launch
   • roslaunch group_4 intrinsic_calibration.launch
   • roslaunch group_4 lane_detection.launch
   • roslaunch group_4 controller.launch.
6. Use rqt_image_view in another terminal to view the results on various topics.