Binary Classification for Network Intrusion Detection In Drones
Grade Level at Time of Presentation
Sophomore
Institution
Western Kentucky University
KY House District #
76
KY Senate District #
13
Faculty Advisor/ Mentor
Ismail Abumuhfouz
Department
Computer Science Dept.
Abstract
The technological growth of drones and other Unmanned Aerial Vehicles (UAVs) is inevitable. They are currently used in a variety of fields, but this usage is susceptible to cyberattacks; which may endanger privacy, national security, and even human life. Network Intrusion Detection Systems (NIDS) are used to monitor a network or systems for malicious activity or policy violations. The current NIDS are ineffective in protecting drones against novel attacks. This project seeks to create a binary classification NIDS to protect drones from WIFI attacks. We combine the NIDS with Artificial Intelligence (AI) to increase detection of previously unidentifiable attacks. Several AI-enabled NIDS already exist in different fields, but to the best of our knowledge, this is the first NIDS that is tailored specifically for drones. We evaluate the performance of the AI algorithms of random forests, stochastic gradient boosting, and K-Means for binary classification on the CICIDS2017 dataset.
Binary Classification for Network Intrusion Detection In Drones
The technological growth of drones and other Unmanned Aerial Vehicles (UAVs) is inevitable. They are currently used in a variety of fields, but this usage is susceptible to cyberattacks; which may endanger privacy, national security, and even human life. Network Intrusion Detection Systems (NIDS) are used to monitor a network or systems for malicious activity or policy violations. The current NIDS are ineffective in protecting drones against novel attacks. This project seeks to create a binary classification NIDS to protect drones from WIFI attacks. We combine the NIDS with Artificial Intelligence (AI) to increase detection of previously unidentifiable attacks. Several AI-enabled NIDS already exist in different fields, but to the best of our knowledge, this is the first NIDS that is tailored specifically for drones. We evaluate the performance of the AI algorithms of random forests, stochastic gradient boosting, and K-Means for binary classification on the CICIDS2017 dataset.