One of the projects I worked on while at the Biorobotics Lab was for the DARPA SubT challenge. This DARPA program was trying to develop autonomous robotic systems capable of independently exploring environments too dangerous for humans, such as caves, tunnels, and abandoned or destroyed urban environments. In the process, the robots would search for specific items as specified by DARPA and use SLAM to create a map of the explored terrain.

One of the SubT Ground Vehicles

I was responsible for helping with some of the communications systems. In order to extend our communication range to our robots, the team decided to drop Wi-Fi nodes to create a loose mesh network back to the base station. I worked on machining the 8020 for the node droppers, and this was one of my first introductions to manual machining. I also took over making corrections to the system after the undergraduate team found out we had been given inaccurate drawings and the staff engineers were away on break.

The Node Dropping Device

I also was able to work on the electrical side of the hardware going inside the nodes. The original nodes were using a magnetic switch, but it was also getting mechanically actuated when dropped from the robot. This led to the board power-cycling and delaying the robot. As a stopgap measure, we installed additional capacitors to maintain the voltage when the node was dropped. Eventually the switches were exchanged for hall effect sensors that would stay on.

Oscilloscope Data From Nodes Losing Power

Modified Node with Capacitor