rosorg-logo1About a year ago I went through the beginner level tutorial for ROS, the open-source Robot Operating System. The beginner's track ends with a "Where next?" page with lots of options. At the time I did not have a specific robot or a specific project in mind, so there was no particular motivating factor to help me choose from the list of options. This has the unfortunate side effect of analysis paralysis and so that's where I stopped on the last go-around. At the time, laser cutting an acrylic enclosure for a network attached storage server seemed more interesting, and I got distracted.

Things have changed in the past year. The most important change is that I have access to multiple robot platforms. We have the SGVHAK Rover, my Sawppy the Rover, and Emily's Craglist acquisition Bart the robot. All of these hardware would benefit from a powerful and sophisticated software platform to elevate them from fancy remote control vehicles to something really interesting.

Another change was a realization as I got started thinking about deep neural networks and playing with AI. I started with the powerful laptop and got TensorFlow installed, but as I went through the tutorials I realized it's all very interesting in the abstract but I can see myself getting distracted again. I need a goal to work towards as I apply my AI education and experimentation, and this circled me back to the robot platforms we have running around.

ROS is the software infrastructure where these two would meet - the hardware robot platforms and the TensorFlow neural networks. So before I get started on another big complicated software framework like TensorFlow, I should finish what I started with ROS.

As a refresher, I went through all the beginner tutorials again. It made a lot more sense this time around given the context of what I've learned in the past year. Ranging from Python programming, to Linux fundamentals, to getting a better feel of why ROS features are desirable viewed through the lens of building Sawppy.

The progress was a lot faster this time... and now I'm at the Where Next page once again.

Hello, old friend. This time, let's keep the momentum moving.