Fly Catcher is an open-source device that can detect aircraft spoofing by monitoring for malicious ADS-B signals in the 1090MHz frequency.
Angelina Tsuboi, the developer of Fly Catcher, is an enthusiastic pilot, cybersecurity researcher, and tinkerer. She was driven to embark on a project that merged these three distinct interests to address a significant issue in aviation radar systems.
Intrigued by the ADS-B system, initially used for basic plane spotting and tracking, Tsuboi looked more into its cybersecurity aspects. Surprisingly, she discovered a lack of security measures in place. Recognizing the extensive use of ADS-B by pilots and ground control, the developer was motivated to utilize their maker skills to raise awareness about this security gap and to create a solution. With a spare Raspberry Pi, an SDR, and a 1090 MHz antenna at hand, they began the development of Fly Catcher, a device aimed at addressing these security concerns.
“Throughout this project, I realized that finding which AI model to detect for a spoofed aircraft depended greatly on the situation. For instance, I learned that a Neura Network was the optimal model for detecting for obvious “script kiddie” aircraft. However, more research had to go into optimizing the model to classify more covert spoofing cases. After tinkering around with various optimizations, I found that RSSI Fingerprinting, which involves classifying spoofed aircraft by analyzing the signal strength of the aircraft ADS-B Out transmitter, was the most accurate, Tsuboi told Help Net Security.
Fly Catcher features
“Fly Catcher stands out against other spoof detection counterparts because it leverages affordable components with extensive documentation, such as the Raspberry Pi and the FlightAware SDR platform, making assembly much cheaper. Moreover, the design and research behind Fly Cather are completely open-source, and anyone can build the device, given they have all the materials. Other spoof detection devices are often inaccessible to the public, are very expensive, and do not have extensive documentation, presenting a barrier to entry for people interested in exploring the intersection of aviation and cybersecurity,” Tsuboi explained.
List of features:
- Detecting spoofed ADS-B messages
- Logging messages on the 1090 MHz frequency
- Mapping and visualizing ADS-B messages
- A portable Raspberry-Pi based device
- An accurate neural network classifier
- 3D printable case with small form factor
- Compatible with the FlightAware SDR
Future plans
Considering the long-term prospects of the project, the developer aims to enhance the core functionalities of the device to boost its accuracy in detecting spoofs.
At present, the device’s training is based on fabricated data, due to the developer’s limited access to actual spoofed signals. Tsuboi is keen on collaborating with signals intelligence experts in the aviation sector to acquire authentic spoofed ADS-B data from real-world situations. This would significantly improve the AI model powering the device. Additionally, the developer is interested in exploring various AI models and paradigms, including RNNs and LSTMs, to refine the model’s accuracy.
Go to GitHub to learn in detail how to build this device yourself.