Making Robotics Accessibile

Andy Zhang bio photo By Andy Zhang

I’m finally done with summer research!

I worked for 12 weeks on Harvey Mudd College as a part of the Research Experience for Undergraduates Program. It was definitely a rich and valuable experience, and now I feel I have a good grasp of both robotics and computer vision.

For the first half of the project, we focused on accessible robotics. As many of you may know, robotics can be an expensive endeavor. We started with a cheap RC car purchased from RadioShack (I know, I’m also surprised that those still exist). Our team tried to use an Arduino board and send our own radio signals, following this tutorial. Sadly, it failed to work.

Using a piece of hardware called the HackRF, we did manage to record commands from the remote control and replay them, in effect controlling the RC car. See the GitHub repo for the code.

However, as you can see, RC cars are not great robots, as they lack the all-too-important ability to pivot about a point. Instead, they must execute multiple-point turns, and inaccurately at that, too.

Instead, we used an iCreate, connected through serial to a Raspberry Pi running Raspbian. The Raspberry Pi, in turn, was connected to a camera module and a WiFi adapter. Using Python sockets to send bytes through WiFi, and mjpg-streamer to read images from the camera over WiFi. Thus, we had a completely wireless robot that could send camera images to a remote computer and receive commands from that same computer.

Here’s a video of our robot in action.

Check back in later for the next part of my research, the actual computing and computer vision!