Contains code for specific simulations and real world deployments

First, check https://github.com/smarc-project/smarc_simulations for how to set up the simulation environment.

When that works, we can start the pipe following scenario. Before starting, make sure to source your workspace in the .bashrc, similar to: source /path/to/your/catkin_ws/devel/setup.bash. The simulation environment is started using a tmux script. For information on how to launch it, check https://github.com/smarc-project/smarc_utils/blob/master/README.md#smarc_bringup .

Now that you know how to use and navigate tmux, let's start the session using:

rosrun smarc_bringup pipe_following.sh

In the fourth tmux tab, you can start the keyboard teleop window. Try to use this to steer the auv to follow the pipe. Instructions on how to steer the auv can be found in https://github.com/smarc-project/smarc_utils#smarc_keyboard_teleop .

The AUV has a camera that publishes on /small_smarc_auv/small_smarc_auv/camera/camera_image and laser scanners on small_smarc_auv/sss_left and small_smarc_auv/sss_right. All these topics can be visualized in rviz if we have set world as the base frame. We can also add a robot model with robot description small_smarc_auv/robot_description to see the robot in rviz. Add a tf to rviz to see the relative position of all the frames.

In addition, the auv publishes several topics that might be of interest, you can look at the topics:

/small_smarc_auv/current_velocity
/small_smarc_auv/fins/0/output
/small_smarc_auv/fins/1/output
/small_smarc_auv/fins/2/output
/small_smarc_auv/fins/3/output
/small_smarc_auv/imu
/small_smarc_auv/is_submerged
/small_smarc_auv/thrusters/0/is_on
/small_smarc_auv/thrusters/0/thrust

This is something that we need to discuss!

These are just some random hints that came to mind.

You will need to convert the sensor_msgs/Image to opencv images using cv_bridge, see http://wiki.ros.org/cv_bridge/Tutorials/ConvertingBetweenROSImagesAndOpenCVImagesPython .

To find the pipeline, you might want to calculate some gradients https://docs.opencv.org/trunk/d7/d4d/tutorial_py_thresholding.html , followed by thresholding https://docs.opencv.org/2.4/doc/tutorials/imgproc/threshold/threshold.html or just go for some thresholding based on the color directly.

Another idea might be to use the Hough transform to find the straight lines outlining the pipe: https://docs.opencv.org/3.0-beta/doc/py_tutorials/py_imgproc/py_houghlines/py_houghlines.html

You will want to read about the transform buffers in ROS: http://wiki.ros.org/tf/Tutorials . The coordinate system in your case is defined with respect to the world frame, which is fixed with respect to the simulated world. You can use the transform to the auv /smarc_auv/base_link frame.

Some simple P controller, https://en.wikipedia.org/wiki/Proportional_control , maybe? Or just something simpler. Do we need to add a depth sonar or a pressure sensor to be able to properly know the height? Uncomment these lines to get a sonar: https://github.com/smarc-project/smarc_simulations/blob/master/smarc_auvs/models/small_smarc_auv/urdf/small_smarc_auv_base.urdf.xacro#L227