Shelby Hall Graduate Research Forum Posters

Download

Download Full Text (340 KB)

Description

This research proposes a new manner of implementing machine learning models such that, when applied on a drone, it will be able to accurately identify and maintain the authenticity of the entity sending the control data to the drone. To begin with, the drone will, for a pre-determined amount of signals received per unit time, determine the average signal strength (RSSI) of them and use that average to determine the approximate distance between the drone and the source of those signals. This single data point will be fed into a custom implementation of the SCluStream algorithm (a real-time clustering machine learning algorithm) onboard the drone to assign that point to the cluster that best fits it. As the drone is flying, and as time goes forward, a series of clusters will naturally emerge that showcase the natural change in distances between the drone and the sources of the control signal, and due to the nature of the SCluStream algorithm, these clusters themselves can gradually change to reflect a new natural pattern of distances. If, however, any of these data points being fed into the SCluStream algorithm don’t fit into one of the established clusters (or there isn’t enough room to make a new one), then this indicates to the encompassing framework that the source of the signal is either too close/far away to/from the drone to be the established authentic signal source. This fact will then be communicated to the drone’s main flight control unit, at which point the drone can take appropriate action against this probable threat.

Publication Date

3-2025

Department

Computer Science

City

Mobile

Disciplines

Cybersecurity | OS and Networks | Other Computer Sciences | Programming Languages and Compilers | Systems Architecture

Using Machine Learning Models to Improve the Cyber Physical Security of Drones

Share

COinS