CubeSats have become a transformative platform for space exploration and education, providing cost-effective access to orbit for universities, startups, and small organisations. Yet, mission statistics consistently reveal reliability challenges, particularly during the commissioning phase when subsystem interactions are first exercised under real conditions. The cost and complexity of environmental testing constrain small teams from performing exhaustive validation, underscoring the need for accessible, realistic methods to assess system behaviour under faulted conditions.
This thesis presents a state-aware, real-time fault-injection framework for a modular 1U CubeSat, developed as a tabletop testbed that unifies hardware and software fault emulation under a single orchestration interface. The framework implements a two-layer architecture: a communication-layer (Layer 1) injector that emulates protocol disturbances such as timeouts and corruptions, and a hardware-abstraction-layer (Layer 2) injector that introduces controlled sensor and actuator degradations. Fault activation is governed by the satellite’s operational state, ensuring that injections occur only within admissible mission
contexts and preserving realism and safety.
The system is implemented around the Satbus Commons library—a platformgnostic middleware that standardises communication, fault scheduling, and state management across heterogeneous microcontrol-
lers (STM32 and ESP32). A comprehensive fault catalogue spanning all major subsystems—Electrical Power System (EPS), Communications (COMMS), Attitude Determination and Control System (ADCS), and On-Board Computer (OBC)—enables consistent testing of both software and hardware failure modes without modifying nominal application layer code.
Experimental campaigns validate the framework through three representative fault scenarios, demonstrating deterministic activation timing, precise recovery behaviour. The results confirm that the hybrid approach achieves realistic, reproducible fault emulation suitable for early development-stage verification. Beyond technical validation, the same framework also serves as an educational platform, allowing students to explore fault propagation, system interactions, and recovery logic in a controlled environment.
Overall, this work contributes the first integrated, state-aware fault-injection framework for CubeSats, bridging research and education. By enabling early, realistic, and safe validation of subsystem interactions, it advances the goal of improving small-satellite mission reliability while fostering the next generation of satellite engineers.