Building a Sky Hub UAP Tracker
A UAP data collection solution using the NVidia Jetson platform
Introduction
Unexplained Aerial Phenomena (UAPs) are now widely acknowledged. However, reliable, publicly available data on this enigmatic phenomenon is difficult to find. Sky Hub is taking a citizen-science approach, applying an open-source observational science platform (check out the recent article by founder Steve McDaniel).
The Sky Hub “Tracker” is the centerpiece of the Sky Hub system. The Tracker can be built using readily available hardware. It leverages the NVidia Jetson platform and uses recent advancements in AI and Machine Learning to intelligently process and record real-time data from cameras and sensors monitoring the environment.
I’ve been following the Sky Hub project with great interest for several months now and with Version 1.0 about to be released I thought it’d be useful for others to see how I built my Tracker.
You can easily join this novel scientific approach to monitor our skies and help in the global effort of investigating one of the most thought provoking puzzles of our time.
Building the Tracker
Equipment used
For my build I used the equipment recommended on the Sky Hub website and the Sky Hub wiki. I received advice on selecting the items in their active online community. Everything was purchased on Amazon with the exception of the camera.
- NVidia Jetson Nano with 5v fan and 32GB SD Card
- Dahua 12 MP Fisheye Camera
- Adafruit Ultimate GPS module
- USB3 SSD
- ATX PSU
- Ethernet switch and cables
It’s worth noting that much thought and discussion has gone into the recommendation of these items by experts in the Sky Hub team.
Power Supply
First and foremost the tracker requires a stable power source to supply to all the above hardware. I chose the Corsair VS350. Sky Hub plans to support additional sensors and embedded devices in order to gain a deeper understanding of what may be happening around the tracker during UAP events.
The 350W Corsair provides more than enough wattage to cover everything for this base level build and for the future. The added benefit of using a PSU is that it can be used to exhaust air from a tracker enclosure —neat design thinking by the Sky Hub team.
Out of the box the PSU needed some modifications. I cut the wires from the ATX connector and used screw terminal barrel jacks for the 5V Jetson, 12V camera and 5V Ethernet switch. I also added a SPST toggle switch for a master power control.
Setting up the Jetson Nano
The Nano is the brains of this system and will be doing lots of number crunching on the CPUs and GPU so it’s important to regulate heat. I used a Noctua 5V PWM fan which will keep things cool by moving hot air away from the board.
I followed NVidia’s instructions and flashed Ubuntu v18.04 on a 32GB SD card using the freely available Etcher. This is a standard approach to preparing the operating system on a card for the Nano, which takes about 10–15 mins to complete.
After that, I connected the SSD, and GPS to the ports of the Nano, connected up the Ethernet switch, plugged in the 5V barrel jack and powered on the Jetson Nano for the first time. After setting up Ubuntu with a user account I was ready to install Sky Hub!
The developers have made this step really straightforward. It is literally just a one line command in a terminal window.
What about the camera?
Once the install had completed, the final piece to connect was the camera. Sky Hub support a range of camera types. Their main requirements at the time of writing are that the camera is IP based and supports the RTSP protocol, and can be configured to H265 standard. They recommend a fisheye camera with a reasonable resolution as a starting camera.
From the image below you will see that the higher the megapixels for the camera, the further the system can see, and therefore track a UAP. For example, a 4MP camera can only detect a roughly 100m sized object at a 5km max range, but a 12MP camera has a range of 10km.
Version 1 of Sky Hub is designed to support a fisheye camera. In future versions, Sky Hub plans on supporting other camera types such a Pan-Tilt-Zoom which can be quickly directed to focus on any objects detected by the fisheye.
For my build I selected a 12 MP Dahua IPC-EBW81230. This camera has a 180° hemispherical view of the sky. It uses a highly sensitive STARVIS™CMOS sensor and is one of a range of Dahua cameras currently recommended due to their high lux sensitivity (0.001) and ability to withstand the elements when mounted upside-down (IP67 rated).
Testing the Tracker
With everything now connected and my tracker up and running, I opened up the main user interface for Sky Hub which displays information about recorded events. Videos started popping up in the event log immediately, and even placed outside my window for the last couple of days it has been tracking movement of the various biodiversity and aircraft.
I could see I had a GPS lock — the little Adafruit unit works really well and and had a fix within a couple of minutes. The role of the GPS is to provide an accurate location, and stable time for tagging all of the data recorded by the tracker.
Next steps
Now that I have my tracker hardware up and running, I plan on building an enclosure and mounting it in a more optimal position where the camera can see as much of the sky as possible (i.e. my roof). I can’t wait to see what it records!
I’m excited to see how this project progresses as the Sky Hub community continues to grow and more trackers are deployed around the globe. It will be interesting to see what the data says about this great mystery.