Unmanned Aerial Vehicles (UAVs) find increased use in many military and commercial applications such as remote weapons, surveillance, good delivery, filmmaking, and agriculture, among others. With the proliferation of Internet-of-Things (IoT) and the use of off-the-shelf components, UAVs have now become targets for cyber-attacks, raising serious privacy and security concerns. A successful attack can allow complete control of UAVs, corrupt collected data, damage software, or disrupt availability. Adversaries may also monitor and collect information for later purposes. Our research focuses on detecting malicious actions in a UAV swarm. In this work, we study the “expected” behavior of a UAV swarm in specific scenarios. We use PX4 and Gazebo to simulate malicious behavior in UAV fleets and set up system parameters that could allow us to detect a malicious UAV within the fleet. Our focus is on identifying how system resources (CPU, memory, battery usage) are being affected by malicious attacks. We developed a proof-of-concept UAV testbed using a Raspberry Pi 4 Model B with a Navio2 Autopilot HAT that will be incorporated into a quadcopter to collect data from a physical system. Our ongoing work explores the scalability of our implementation. We intend to simulate a swarm of UAVs using a PX4 simulation running through Gazebo. This study will be used for developing a distributed intrusion detection system that aims to detect malicious UAVs in a UAV swarm.