The presentation discusses the development of a system to test corner cases in autonomous vehicles and drones by spoofing GNSS signals.
- The system developed by the company involves hardware and software components to spoof GNSS signals for testing purposes.
- The company owns an autonomous vehicle fleet and uses a Lincoln MKZ outfitted with an autonomy kit for testing.
- Testing results focused on position translation, velocity manipulation, and halting the GNSS system.
- The system developed is legal as it complies with FCC regulations.
- An anecdote is provided about the issues faced with GNSS sensors in drones and the need for testing corner cases.
- Tags: AI, Cybersecurity, DevOps, Autonomous Vehicles, Drones, GNSS Spoofing, Testing
The speaker shares their experience with drones and the issues faced with GNSS sensors, which were the second most common issue after mechanical issues. The GNSS would often move to the right by 200 meters or the altitude would jump by 50 meters, causing the system to respond in different ways, sometimes appropriately and sometimes not. This highlights the need for testing corner cases to improve the reliability of autonomous vehicles and drones.