Last Friday we had the PDR with the reviewers, who are professors, alumni and friends at UCLA. They provided many insights to the three software components of the project, especially the controls part. From this review we all got a better sense of the progress each team made. Overall, they suggested that we should get some of the details sorted and work more on the simulations before actually flying the drones. The simulation part, given its complexity, might play a much more important role in the project in the future.
The next steps for the software team are digesting the advice we received from PDR, adjusting some of the designs, and getting to design some APIs that interact with different parts of the project.
Over the past two weeks, our Mechanical Airframe team has been heavily involved with making adjustments to the octagonal frame and docking mechanism in preparation for the Preliminary Design Review which was held on 10/30.
After analyzing the difficulty in securely attaching a large Apriltag to the center of the octagonal frame, most likely via some fishing wire, we decided to rethink our design and go with a more conventional octocopter-like frame which consists of spokes protruding from a central body. This way, the Apriltag will have a solid surface to be attached to, and the payload will also be more easily secured. They will attach to the frame’s central plate, from which the 8 boom arms will extend. These arms will just be single 2.11 cm diameter carbon fiber tubes, as opposed to twin rods, since torsion is no longer a significant concern with the new frame design. A free surface simulation was performed, due to it being more computationally efficient than a complete body simulation, on the carbon fiber tube to validate its strength under full throttle of the drone. The results of the simulation can be seen below and it is evident the selected tube is sufficient as the deflection at the end of the tube is insignificant. We saw 4.13 cm of deflection at the end of a 115 cm boom.
The next modification we made was to the docking mechanism. The revised docking adapter boasts a 3D printed body that can be attached to the drone. It has two slots for electromagnets of opposite polarities such that the orientation can be locked in with permanent magnets on the docking joint which is attached to the frame. These electromagnets will be powered during the docking stage to properly align as well as during undocking in order to push away from the frame by reversing their polarity. The next important aspect of the docking adapter is an internal rack and pinion key that is controlled by motors. These keys will slide into the docking joint insertion points, which are flared to account for any translational discrepancies, in order to securely attach to the frame. As a result, the docking mechanism has been designed to be as robust and simplistic as possible.
Overall, these design changes led up to the Preliminary Design Review where we got some great feedback and insight on our design. For instance, it was recommended that we look into stronger electromagnets so that the precision docking can be aligned from further away. It was also suggested that we open the back of the insertion points on the docking joint so that any wind gusts will have a reduced impact. Lastly, it was reiterated that we talk about the canted motor design with Brett Lopez, and discuss the best way to implement that feature.