MAVR: Code Reuse Stealthy Attacks and Mitigation on Unmanned Aerial Vehicles

As embedded systems have increased in performance and reliability, their applications have expanded into new domains such as automated drone-based delivery mechanisms. Security of these drones, also referred to as unmanned aerial vehicles (UAVs), is crucial due to their use in many different domains. In this paper, we present a stealthy attack strategy that allows the attacker to change sensor values and modify the UAV navigation path. As the attack is stealthy, the system will continue to execute normally and thus the ground station or other monitoring entities and systems will not be able to detect that an attack is undergoing. With respect to defense, we propose a strategy that combines software and hardware techniques. At software level, we propose a fine grained randomization based approach that modifies the layout of the executable code and hinders code-reuse attack. To strengthen the security of our defense, we leverage a custom hardware platform designed and built by us. The platform isolates the code binary and randomized binary in such a way that the actual code being executed is never exposed for an attacker to analyze. We have implemented a prototype of this defense technique and present results to demonstrate the effectiveness and efficiency of this defense strategy.