Resilient Navigation for UAVs: GNSS Attack Detection and Seamless Transition to Visual-Inertial Localization

Modern aerial surveillance and autonomous drone operations rely heavily on Global Navigation Satellite System (GNSS) technology for precise navigation and mission execution. However, this dependency exposes critical systems to significant vulnerabilities from signal interference, including malicious spoofing and jamming attacks, which can compromise operational integrity. This paper presents a resilient navigation framework designed to ensure mission continuity in GNSS-compromised environments by integrating real-time attack detection with a seamless transition to alternative navigation methodologies. We first demonstrate these critical vulnerabilities by conducting controlled signal injection attacks on commercial navigation modules using portable software-defined radio (SDR) platforms. Addressing these risks, we propose and validate the core components of an autonomous framework that intelligently detects signal anomalies and seamlessly transitions the navigation state from GNSS-dependent to a map-aided visual-inertial or any other situational awareness-based localization mode. This system is designed to enable robust autonomous navigation in GNSS-denied or degraded environments, ensuring resilience for a wide range of mission scenarios. Experimental validation confirms the system's ability to detect malicious interference in real-time and execute a rapid, reliable handover to sensor-based localization, ensuring operational continuity and safety even under adverse GNSS signal condition.