My student Ainara Picot Arizmendi disserted his Telematics BSc thesis

Created in July 22, 2025 by Néstor Lucas Martínez

2025   ·   bachelor-thesis   ai-cybersecurity   iot-security   uav   machine-learning   honors   ·   news

🎓 Outstanding Bachelor’s Thesis Defense

On Tuesday, July 22, 2025, my student Ainara Picot Arizmendi successfully defended her Bachelor’s Thesis titled “Artificial Intelligence-powered Cybersecurity for Secure IoT Communications”. Her work received the highest possible grade—Outstanding (10/10)—from the evaluation committee, along with a nomination for Honors Distinction (Matrícula de Honor).

This project addresses an increasingly critical issue: how to defend autonomous aerial vehicles (UAVs) from cyberattacks that target their wireless communications. Ainara developed a full-stack intelligent system capable of detecting cyberattacks in real-time, using a combination of machine learning, network simulation, and embedded deployment techniques.

Her solution features a hybrid AI architecture, combining:

  • CBLOF for unsupervised anomaly detection,
  • Autoencoders for non-linear feature compression,
  • and XGBoost for high-precision supervised classification.

The system was trained and validated using the ECU-IoFT dataset, which contains realistic traffic from UAV telemetry under attack. The solution reached over 95% F1-score for binary classification and performed strongly in multiclass attack detection, identifying threats such as:

  • Wi-Fi deauthentication floods,
  • WPA2 key cracking attempts, and
  • malicious API misuse.

Ainara’s work stands out not only for its technical quality, but also for its deployability. The system is:

  • fully modular,
  • lightweight enough to run offline on embedded devices,
  • and validated in real time using tools like tcpdump and aireplay-ng.

She also integrated her system with the Damn Vulnerable Drone simulation environment to test cyberattack scenarios under controlled conditions.

The project adheres to principles of ethical AI, computational efficiency, and aligns with several UN Sustainable Development Goals, particularly around infrastructure resilience and digital security for autonomous systems.

This result is a reflection of Ainara’s strong analytical capabilities, technical excellence, and dedication to real-world problem solving—a truly outstanding final project and a solid contribution to the field of AI-based cybersecurity for IoT and UAV systems.