Week 3

1: Effecting the vz through z

The method implemented last week is able to control the z position pretty well but not the vz (velocity in z). To solve this, a carrot model was implemented that uses the value provided by the student as vz and translates it into a corresponding position (that is constantly updated). This way, when the student provides vz = 0, the good position control is enabled and in all other cases, a position setpoint is calculated and provided to the PX4. This has been implemented directly into DroneWrapper.

2: ROSification of Follow Road

The earlier implementation of Follow Road not only used Ice but relied on pure python (Qt) for the GUI. Since rqt is the standard for GUIs in ROS, it makes sense to translate the GUIs to using rqt and creating them as complete packages. These packages are envisioned to be provided in the assets deb package. The final workflow would be something like:

roslaunch jderobot_drones follow_road.launch

And then the GUI is used to run the code.

Since I had never worked with rqt before, the most part of the week went in trying to understand it. After going through the tutorials and other rqt plugins that are available on github [1] [2] [3] I was able to figure out how to code the logic for the exercise. This, of course, is all reusable i.e. almost all the code written this week shall be used for almost all exercises with a few tweaks.

Accounting for the two days leave(monday, tuesday) for my advisor’s wedding in Greece over the weekend, I have been able to make the interface to the gui package which doubles up as the student’s solution. I shall complete this and expect the pull request to be submitted in two days.