Internship - Reliable Robotics

My primary project has been to develop what is being called a “Low Lift” Video DAQ system for recording flight test footage. Photos not included at the moment as I await approval.

Problem: Currently, when flight test videos are recorded, individual GoPros are placed around the aircraft on the morning of the flight. When all other procedures are complete and the plane is about to depart, the GoPros are turned on and started. Upon return of the aircraft, we pray the battery lasted long enough for the entire flight to be recorded. Usually, it dies sometime during the flight. In either case, we then have to go around the aircraft and remove every camera, recharge the batteries, and then someone has to sit and upload videos from each GoPro SD card, ultimately taking about 2+ hours just to have individual files uploaded to Google Drive. Once there, someone else will download the footage, manually sync them using visual cues, and put it into a 4-panel view, then reupload the final video for purposes of flight reviews, marketing, etc. This takes a ton of time, is unreliable, and can be improved.

Solution: Enter, my project, the Video DAQ. Inside the aircraft will now sit a small pelican case. It connects to the main power bus as all other systems do, and has a power switch to turn it on. When you open the lid of the case, you are greeted with only 4 LED Indicators, 4 Switches, and a removable USB Stick. Each switch is associated to a simple action (Record on/off, download videos, Wipe SD, and for now the last is actually a blank one for a future utility), and the LEDs let you know what process is taking place. When you click any switch, a command is sent simultaneously to every GoPro connected to the VDAQ.

To start the system, you simply turn the power switch on. The internal computer boots up, and all GoPros turn on. To start recording, turn the record switch to high, and all GoPros simultaneously begin recording. The flight can now begin. Upon return of the aircraft, simply set the switch to low, and the GoPros stop recording. The plane can now be powered down and returned to the hangar. Once back inside, we can repower the system and press the download button. At this time, every GoPro video that was just taken is downloaded directly to the USB stick, with naming conventions that a later program understands. You can walk away, and after about 45 minutes, all of the videos are on this thumb stick.

Return the stick to HQ, where you can plug it into a computer here. Upon running the script, it automatically extracts the GoPro videos. Using FFmpeg, it parses the embedded GPMF data to validate timestamps in the video. This is used to align each of the camera videos in a time-sync fashion. Once running, you can leave to do other things until it finishes processing in about 40 minutes. Once completed, you have a single video file (along with all individual files) that has a multi-panel view (Typically 4) to show a side view from each wing, the nose wheel, and the tail wheels, with every video synced to within 50 ms. This gets passed onto the marketing team for any potential use, to Ops and flight test for flight reviews, and later to GNC and our internal tools team for additional reviews.

Under the control panel of the Pelican Case is a full power and data distribution system I designed that allows us to control GoPros in a wired fashion from over 40 feet away (The distance needed to get the GoPros out to the wing). Due to time restrictions of the internship, we use COTS hardware that will be replaced in the future. For the time being, I designed water-tight housings to hold plugs in place and seal them off from any environmental factors. In every possible instance I could, I follow military and aviation standards for hardware selection, which in this case is mostly in connectors and power busses.

Each GoPro is now encased in a custom aluminum housing to again seal and protect COTS connectors that are attached to the GoPro, which allows me to control and stream video from the cameras. I performed modal and bolt loading analyses to confirm that these would not be a cause for concern when mounted on the aircraft.

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.