When the COVID-19 pandemic hit the US and stay-at-home orders began, we at UAS@UCLA learned that we wouldn’t get the chance to fly our brand-new UAV at the annual AUVSI-SUAS competition in Maryland this year. This left us with a lot of time on our hands and not much to show for our work this school year.
Fortunately, we were informed of a little-known NASA research grant for undergraduate students, called the University Student Research Challenge (USRC). The onus is on the students to come up with a research topic that aligns with NASA’s mission, and to then draft a research proposal for it. Our topic? AVIATA - the Autonomous Vehicle Infinite Flight Time Apparatus. We plan to build a swarm of drones that’s able to collaborate together to carry heavy items for an infinite amount of time by innovatively swapping charged and discharged drones in and out of the swarm.
Fast-forward to July 2020. NASA reached out to us and let us know we had been selected to receive grant money to fund our project! Riding on the heels of this fantastic news, we launched a successful crowdfunding campaign as required by the USRC, which unlocked access to more of NASA’s pledged grant money, and today we are in the initial stages of project design and planning.
Firstly, it was important to determine what the software team would be responsible for. While we have a hardware team designing the mechanical aspects of AVIATA, the software team would be in charge of creating the control systems. This means figuring out how to make several drones fly together, how to communicate with the drone swarm, and how to autonomously dock and undock from the structure while in flight.
In the design stage, we are having high-level discussions on how these processes will work. Currently, this involves a lot of flow-charting and “what if” conversations to make sure we handle all possible scenarios. For example, what if a drone flies to the structure to replace another drone, but it’s unable to find it’s docking location?
Another software-oriented goal is to develop simulation capabilities. One of our engineers, Ryan, has successfully created a Python program that can model how the drone swarm would react if a drone were to leave. This is important because it will allow us to predict the minimum number and configuration of drones we need based on the current weight we’re carrying. So far, the results are very promising, and we’ll continue to develop this simulator.
The goal for the upcoming weeks is to fully flesh out what AVIATA will look like from a software perspective. We have a preliminary design review planned for mid-October, where we hope to get constructive feedback on our current design. Afterwards, we’ll be ready to start writing code and programming the drones.
As you might imagine, developing a collaborative drone swarm for lifting objects involves a whole host of technical challenges, both mechanical and in software, that need to be dealt with to produce a robust final product. The hardware team’s job is to design the fixtures and mechanisms that help this type of interdrone collaboration run smoothly, and to make the software team’s job easier by reducing the mechanical uncertainties that make communication and control more difficult.
The hardware team focuses on three main areas of development: a frame to which the payload is secured, the docking mechanism that facilitates the drones’ attachment to and detachment from the frame, and modifications to existing drones so that they function as desired within our system. Though Covid-19 related restrictions make it difficult for team members to meet in person, we’ve developed a plan for distributing manpower and resources between each of these areas so that we can make steady progress even without in person meetings for the foreseeable future.
The hardware team is already hard at work researching, prototyping, and designing the components to make our system run smoothly. For development of the frame, one of our engineers, Willy, has determined the weight budget for the frame and payload by calculating the thrust produced by only three drones flying in suboptimal conditions. Another engineer, James, is researching ways to reduce oscillation and vibration of the payload, by damping or other means, to reduce the corrections that need to be made during flight in software.
For the docking mechanism, I’ve been testing a series of 3D printed fixtures that attach to either the frame or a swarm drone to help magnetically align the drones in the proper orientation for attachment. Progress on all fronts has been smooth and continual so far.
Over the next few weeks the hardware team plans to finalize our preliminary design and test subassemblies individually before moving to the next stage of development. We hope to test the docking mechanism using a human-piloted hexacopter before our preliminary design review in mid-October, where UCLA faculty and industry experts will give us feedback on our developments up until that point. We’re very excited about this project and our progress, and hope you’ll stick around with us on this journey.
These results are based upon work supported by the NASA Aeronautics Research Mission Directorate under award number 80NSSC20K1452. This material is based upon a proposal tentatively selected by NASA for a grant award of $10,811, subject to successful crowdfunding. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of NASA.