Year: 2026
Type: article
Abstract: Unmanned aerial vehicles (UAVs) increasingly integrate autonomous navigation, wireless communication, artificial intelligence, and modular payloads, increasing the risk that faults or cyberattacks may propagate to flight-critical functions. This paper proposes a compartmentalized resilience architecture for autonomous, network-connected, mission-critical UAVs. It separates flight control, navigation, communication, mission autonomy, payload processing, data management, and safety monitoring into distinct trust domains. A default-deny policy gateway governs inter-compartment communication, while an adaptive trust mechanism restricts or quarantines abnormal components and activates predefined recovery modes. The architecture is assessed through a comparative analytical framework covering representative fault and attack scenarios. The assessment considers containment, flight safety, mission continuity, response time, and anticipated resource overhead. The paper also reviews major UAV applications and their challenges and outlines how compartmentalization can support a modular UAV platform with interchangeable subsystems. The approach aims to limit fault propagation and preserve safe operation while accounting for UAV real-time and resource constraints.
Institutions: Oakland University, University of Michigan–Dearborn
Language: English
Subfield: Safety, Risk, Reliability and Quality
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7223047734