Kyle M. Hart

He/Him. Part-time robotics PhD candidate. Full-time lab director. Sometimes has a spare moment.

2 minute read

It has been a bit since I last posted anything. Turns out, this semester is busier than I first expected. While I am only taking a single class, I will also be taking the qualifiers in May. This involves a research component and a knowledge component, so I have been brushing up on my lessons from prior classes to prepare. To stay up to date with it all, my non-working time is pretty stricly regimented. Monday: review practice problems, Tuesday: lecture, Wednesday: … you get the picture. The studying is going well, but it takes a lot of time. What little downtime I have left is spent with the family or just recharging. Consequently, I have had little time left to dive in to any new projects or interesting programming tools. Hence, the lack of new content.

That being said, I still have managed to start a new project! I am getting up to speed on ROS2. I use ROS1 extensively at work and have been aware of ROS2 for some time. I just haven’t had any opportunities to learn and use it. I have a few personal project concepts in mind that would benefit tremendously from ROS2 as well, so I have a personal motivation to learn it (as opposed to just free labor for my employer).

This project involves creating a package to control the turtlesim example via keyboard. This is definitely not an original idea. In fact, there is a package in the turtlesim examples that offers this exact functionality. However, turtlesim control offers an opportunity to learn all the fundamentals. In particular, I am planning to allow the following functionality:

  • Drive the turtle via keyboard control (topics).
  • Change the pen colors via key presets (services).
  • Instruct the turtle to rotate to absolute orientations (actions).
  • Change key mappings prior to running (parameters).
  • Include Docker files to allow development anywhere.
  • Include rigorous documentation.
  • Use a CI pipeline.
  • Explore some basic static code analysis tools.

I have only just started the project and only completed the first goal so far. The documentation is at https://kylerobots.github.io/turtlesim_teleop/ and the source code is at https://github.com/kylerobots/turtlesim_teleop. I will chip away at it when I have time. Once the qualifiers are done, progress should speed up.

Additionally, this project actually started because of a bigger project idea I had. The two knowledge areas under test for the qualifiers are dynamics and controls. I thought it would be interesting to recreate the cart-pole/inverted pendulum control problem in Gazebo/Ignition to help pull all this information together. This could also be an opportunity to explore some of my reinforcement learning knowledge as an alternate controller. However, that would be a lot to tackle in a single project. I figured it would be easier to learn the ROS2 fundamentals first with this keyboard control project, then progress to the cart-pole controller. There is no way I will finish before the qualifiers, but it is still an interesting project to tackle that encorporates a number of areas of interest.

Until then, though, I think I have another practice problem that I need to get to…

Updated: