GPS L1 Receiver System with Machine Learning Based Spoofing Detection Capability

[Audio] Autonomous platforms, including civilian vehicles and military systems, depend heavily on global navigation satellite systems for accurate localization. Welcome to our presentation on developing a GPS L1 Receiver System equipped with machine learning-based spoofing detection capability..

[Audio] However, the open architecture of these signals exposes them to high-fidelity spoofing attacks. Unlike simple signal jamming, spoofing introduces counterfeit data that can silently reroute a vehicle or drone into dangerous environments, all while internal systems report nominal operation..

[Audio] This vulnerability exists because the civilian GPS L1 band relies on the unencrypted C/A code, providing only a single layer of open security. While military-grade L2 signals feature robust encryption, civilian receivers remain inherently open and unprotected against manipulation..

[Audio] Current detection methods generalize poorly against these zero-day threats. To address this critical vulnerability, we present a framework designed to reliably identify spoofing without relying on prior knowledge of specific attack signatures, ensuring safety-critical systems are never compromised. To demonstrate how this framework operates in real time, let's look at the system architecture and live performance metrics in our deployment demo..