Cyber-Physical Fuzzing Framework for Extracting Butterfly Effect in UAV Systems

Today, commercial Unmanned Aerial Vehicles (UAVs) are being utilized in various industrial fields due to the emergence and development of various models. However, cyber threats targeting UAVs that exploit this convenience of use are also increasing. Cyber threats such as controlling commercial UAVs are causing human casualties by being used for criminal activities such as war, terrorism, and unauthorized surveillance. Therefore, a cyber threat response plan for UAVs is necessary. In this regard, we propose a new fuzzing framework for UAV systems based on the butterfly effect, where a small cause leads to a large-scale incident in the future. In the proposed fuzzing framework, flight behavior control-related factors that can affect the occurrence of a neutralizing situation in a UAV system are analyzed and collected. Then, the factors (i.e. impact vectors) for finding the impact chain, which is a UAV neutralizing flow that can occur due to a combination of various factors, are identified. In this paper, we delineate the design configuration of the framework, outline the implementation methodology, and introduce a testbed for UAV simulation and flight status data acquisition within the framework. This framework can be utilized in the future to secure cyber security of UAV systems, ensure security, and establish offensive security strategies to respond to cyber threats.