Senior Design - DOMINANCE

For my final year of undergraduate study for my aerospace engineering degree I had to pick a project for senior design. Fortunately, I was awarded my first pick and became part of the DOMINANCE project, sponsored by Lockheed Martin. The DOMINANCE project stood for “Drone Mine Obstacle Avoidance.” The intent of this project was for three teams to design a drone capable of navigating an obstacle course while trying to detect and avoid at least one autonomous mine, designed by another DOMINANCE team. This was all to be accomplished in a GPS denied environment, while competing against the other teams. I was lucky enough to be the team lead for one of the drone teams and had the chance to lead a diverse team of aerospace, mechanical, electrical, and computer engineers.

Along with leading this team, I handled the administration duties, procurement, and oversaw the flight controller and optical flow sensor. Not only did I have to pick the parts, but I handled configuration and PID tuning of our controller.

To fulfill the requirements for this project, my team designed, and 3D printed our own quadcopter frame. This became an advantage for our group, as 3D printing allowed for quick and inexpensive changes as part count increased, or unexpected problems were encountered. This quick-change capability was accomplished by designing a modular frame, which also allowed for fast repair work when a piece broke.

To navigate the course and detect the mine obstacle, computer vision was utilized with an Intel Real Sense camera and a Jetson Nano. To process the images, we trained an SSD-Mobile Net V2 algorithm, with our own collected dataset, utilizing TensorFlow and CUDA on an Ubuntu 18.04 LTS desktop I have. From there, a combination of ROS and Python communicated with our flight controller to issue navigational commands.

For a flight controller we went with a Pixhawk 1 running ArudPilot copter and outfitted with an optical flow sensor. This combination was able to give us basic height and airspeed data, along with being able to communicate with external devices, such as our companion computer and a laptop base station.

While the drone had a set of other sensors and equipment, I rather show off our final design. This was a project I was proud of, even with the struggles of having COVID and lockdowns near the end of the project necessitating the cancelation of the competition.