TinyML for Predictive Maintenance for Aircraft Ground Equipment; Navigating and Analyzing Internet-of-Things Security Risks

Technical Project Abstract
Aircraft Ground Equipment (AGE) is critical to the operations of the United States Air Force, but current maintenance processes are inefficient, leading to unexpected failures and increased costs. Tinker Air Force Base issued a Request for Proposal (RFP) with Booz Allen Hamilton for a 5G IoT Predictive Maintenance System to address these challenges. During my internship with Booz Allen, I contributed to developing this system using a Raspberry Pi equipped with a thermal camera, computer vision camera, and vibration sensor to detect anomalies in AGE. The system employed TinyML to process data locally on the edge device, with machine learning algorithms developed and quantized for optimal performance. These algorithms analyzed sensor data in real-time to predict potential failures and prevent further damage. To enhance the usability, we incorporated a digital twin of the equipment and the Air Force base in a 3D virtual reality interface that Booz Allen has created. We utilized this tool for interactive monitoring of the equipment and further developed it to display all of the information from the sensors on the Raspberry Pi and update in real time. The system demonstrated strong potential to reduce downtime, minimize costs, and improve operational efficiency. Future work includes adding more sensors, optimizing the models for smaller devices, enabling mobility across bases, and conducting large-scale field testing to validate performance across various types of equipment.

STS Project Abstract
The rapid expansion of the Internet of Things (IoT) and the adoption of edge computing technologies have transformed how people interact with their environments, from smart homes to healthcare to national infrastructure. While these technologies offer speed, automation, and connectivity, they also come with major security and privacy risks that are often overlooked during development. In this paper I investigate how security vulnerabilities in IoT devices and edge infrastructures affect user safety and trust, focusing on a case study of the Ring security camera. Ring is a widely used smart home device that has faced serious criticism over privacy violations, poor encryption practices, and a lack of strong authentication requirements—issues that reflect broader patterns in the IoT ecosystem. I researched each layer of the IoT architecture—examining the Perception Layer, Network Layer, and Application Layer— and then outline how each layer introduces unique vulnerabilities. Attacks such as credential stuffing, man-in-the-middle interception, and weak data protection illustrate that security is often deprioritized in favor of affordability and fast deployment. To better understand how these vulnerabilities are interpreted and addressed, I apply the Social Construction of Technology (SCOT) framework. SCOT emphasizes the role of relevant social groups in shaping technological development and reveals that IoT security is not just a technical problem—it’s a social one. Users, engineers, and regulators all have different stakes in IoT security, and their competing interests create tension around how security measures are designed, implemented, or neglected.
Through this analysis, I argue that widespread IoT insecurity stems from a lack of shared responsibility and enforceable standards. While Ring eventually implemented stronger protections in response to public backlash and federal scrutiny, most manufacturers have not followed suit. This paper calls for a more collective and proactive approach—one that includes mandatory regulations, transparency in data practices, and a shift toward secure-by-design development. Without these efforts, IoT systems will continue to put user data and safety at risk. Understanding the interplay between technology, social context, and regulation is essential for building a safer and more trustworthy IoT future.

Connection Between Technical and STS Projects
Both of my projects are connected through their focus on Internet of Things (IoT) technologies and the importance of security in their design. In my technical project, we developed an IoT-based predictive maintenance system for Aircraft Ground Equipment using a Raspberry Pi equipped with multiple sensors and machine learning models. This is a great example of how IoT is changing traditional processes—what used to rely on routine manual checks can now be handled automatically and more efficiently using connected devices. But while this system offers better performance and reduced downtime, it also introduces new security concerns. Anytime devices collect, process, and transmit data, especially in real-time and on the edge, there’s a risk if that data isn’t properly protected. This overlaps directly with my STS research, which analyzes how IoT devices like the Ring camera have been compromised due to weak default protections and lack of strong security design. My STS paper discusses how the rise of IoT in everyday life has made these devices more vulnerable to threats and how social groups like users, engineers, and regulators all have a role to play in making them safer. What I realized while working on both projects is that innovation in IoT—whether it’s in home security or military operations—requires more than just technical functionality. It needs to be developed with strong built-in security and an understanding of the larger impact these devices have on people’s lives.

School of Engineering and Applied Science
Bachelor of Science in Computer Science
Technical Advisor: Morrison
STS Advisor: Earle