An Empirical Study on Cyber Attack and Defense of Robot Operating System using Security Platform in UAV

摘要

Recently, UAVs (Unmanned Aerial Vehicles), which represent a key application of cyber -physical systems, have been used for malicious intent. Accordingly, research on the neutralization of UAV has become critical. The UAV neutralization process can be divided into three phases. The identification stage is aimed at distinguishing whether the UAVs are friendly or adversarial. Next, the actual mission of the UAVs is neutralized. Subsequently, post-processing is performed to guide the UAV to safe areas. Because the existing UAV neutralization studies did not consider phase 3, secondary damage such as that to people and property may occur after an attack. Therefore, we propose a UAV neutralization method in which phase 3 applied. In this paper, to implement functions such as autonomous driving, navigation, and collision avoidance, we investigate the vulnerability of the MAVROS environment commonly used in unmanned vehicles and clarify the method of attack that exploits the vulnerability. Additionally, we propose a MAVROS API to enhance the security to prevent these vulnerabilities. Finally, we verify the proposed attacks and security through empirical research by considering UAVs with MAVROS.