Update #7: Janurary 15, 2021
Hardware Design
Author: James Tseng
We at Unmanned Aerial Systems at UCLA welcome the new year, and are very excited looking
forward to pushing ahead with our projects!
In the mechanical airframe sub-team, we have most recently been
looking into sourcing our drone fleet. In our initial proposal and designs, we are using hobbyist hexacopters,
particularly the F550 model, which had used to be produced by DJI, and has since been taken over by 3rd parties
since DJI discontinued them. This has been one of the issues with finding a quality and reputable supply, as
supply is both rare and expensive that fits our specifications.
Since we are outfitting our drones with
additional equipment such as a Raspberry Pi and our custom docking mount, as well as being able to lift the
payload frame, we need a stronger propulsion system. In switching our onboard battery design to 4S (16V), the
drones can achieve satisfactory performance.
This has since been our main focus as both the controls and
docking sub-teams are nearly ready to begin hardware testing with cooperative control and computer vision
tracking, respectively. Another mechanical airframe subteam involvement with these endeavors has been the design
of a mounting adapter that connects two drones together through carbon fiber tubes to test the minimum
cooperative control setup.
Figure 1: Two different mount designs to connect two drones
with carbon fiber spars to test minimum cooperative control.
Our next goal is to finalize our drone fleet
design, order them, and assemble them. As a priority, we will first be figuring out shipping the current drones
we have on hand to the sub-teams to conduct their testing. These tests are preparing us for our CDR (Critical
Design Review) this quarter, before we begin integrating each sub-team's subsystem for our tech demo hopefully
in June!
Software Design
Author: Ryan Nemiroff
The software team has been hard at work writing code for each subsystem and testing via
simulation. We have successfully demonstrated distributed control as well as vision-guided flight using
simulated AprilTags. Mesh networking needs a little more tweaking before we can fully test it, so stay tuned for
when that happens.
Firstly, here is the distributed control simulation, demonstrating a leader drone
commanded to takeoff and land (left), and a follower drone copying its attitude & thrust commands (right).
Communication is implemented via ROS2.
And here is our docking simulation, with the 3D world shown on the left
and what the drone sees from its perspective on the right as it approaches the AprilTags.
Obviously, this is a long way from our final product. One limitation of
the available simulation tools is the inability to connect multiple vehicles in a rigid body (and it's already
very computationally intensive). Previously, we had simulated this using a simplified model which did not
incorporate the actual flight software. To put these things together, there is hardware on the way which will
allow the first real-world test of cooperative control. The temporary test structure using two drones is shown
in this week's hardware blog.
Potentially, the full rigid structure could be simulated along with the flight
software, but this would require significant modification of the already complex simulation tools. Because we
are able to conduct real-world tests, the benefit of simulation is not very appealing given the cost of
implementation.
We are in anticipation of the results of this first real-world controls test, which will
inform us if and how we should implement alternative control methods -- there are many potential ways to
accomplish cooperative control, and we want performance to guide our decision on which to use.
Acknowledgment
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.