With our first design review two weeks away, the AVIATA team has progressed in designing and testing elements of the software system. We’ve conducted real-world tests to gain a practical understanding of our vision system’s ability to identify image targets. We are also progressing towards being able to simulate and benchmark various mesh networking algorithms for drone communication.
Although we have not written any flight code yet, we are busy designing and developing simulation techniques to ensure that flight will be as reliable as possible. We have developed a drone control & physics simulator in Python, that currently reflects a simplified model of the physics of drone flight. We are able to control the motors and calculate the resultant forces on the AVIATA structure, which so far has given us basic confidence that our current hardware selections are feasible. Next, we will be adding a simulated control system, which will give us the ability to analyze how the system might act in real flight. This will inform our decisions on how to allocate control to the motors, and tell us what scenarios the structure can safely operate in.
Simulated AVIATA structure. In this formation of eight hexacopters, each small box represents a propeller, with its color indicating the motor speed.
We are itching to begin writing the AVIATA flight code, but we can’t do it without preliminary testing of our ideas, which is why all these simulations are so important! Another preparation we’ve made is planning out the overall logic flow of the system, which is a tough design challenge due to all the physically-separated moving parts. Below is a taste of our logic diagrams, which we will refine as needed to support all modes of operation.
Since the last two weeks, our Mechanical Airframe have been focusing our attention on four major issues to resolve: the strength of the carbon fiber, configuring the Apriltag on the frame, the structure of the payload frame, and figuring out how to onboard the new members of our club onto this team.
We had decided on designing the octagon payload frame with twin rods of carbon fiber as the perimeter, as opposed to one thicker carbon fiber tube. This decision was made mostly due to the low price factor of the specific 0.548” carbon fiber tubes. I conducted a stress test by experimenting with weights hanging on one side of the tubes clamped to one side, creating an end loaded cantilever beam scenario. Multiple measurements were taken and the bulk modulus of elasticity was calculated to be 1.09 GPa for the carbon fiber tubes in parallel. This value was lower than we had anticipated, and thus led to the decision to increase the diameter of the carbon tubes of the frame.
Another concern we had was the effects of having a large, centralized Apriltag located in the middle of the frame and the aerodynamic effects this would cause, as well as the method of mounting it to the frame. One suggested idea was to place the large Apriltag on a screen mesh to allow air to flow through this large surface. Team member Bhrugu was able to test this idea with a smaller filter, and saw promising results.
However, the main issue arose with the consideration of mounting this Apriltag to the frame. To place the tag in the center of the frame without increasing weight significantly, we considered pulling taut fishing wires to support the tag. However, we would have to devise a similar solution for the payload attachment that would also be mounted in the center of the frame. From here, the team decided to revert our payload frame design to a more traditional spoke design which makes mounting both the Apriltag and the payload much simpler, at the possible expense of additional weight. We are considering conducting a trade study to determine which would be optimal.
Finally, as we have started a new quarter, we are excited to welcome our new students interested in UAS@UCLA. We are planning ways to incorporate tasks so that they can gain some new experience while not being overwhelmed by these unfamiliar and possibly challenging tasks. Part of the plan is to hold office hours for them to drop by and ask any questions or ask for help, or even just to socialize among us as a club. Being involved with a hands-on engineering project is tough in this current setting, but we hope to make progress the best we can.
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.