Tools
Tools
Mariyanayagam, Dion (2025) Bio-inspired lightweight polymorphic security system for IoT devices. Doctoral thesis, London Metropolitan University.
Polymorphism, defined as the ability to dynamically alter form, has long been exploited by viruses and malware to evade traditional security mechanisms. This thesis proposes a novel application of polymorphic principles, inspired by biological immune systems, to engineer a lightweight, adaptive, and resilient security system for resource-constrained Internet of Things (IoT) devices. The Bio-Inspired Lightweight Polymorphic Security System introduces a comprehensive framework that detects, rejects, and neutralises unauthorised clients within a secure, encrypted client-server model.
Drawing parallels to innate and adaptive immunity, the system dynamically rotates encryption keys, session credentials, and network configurations in real-time, ensuring robust defences against intrusion and desynchronisation threats.
Furthermore, the research identifies a critical limitation in conventional IoT security architectures: the lack of integrated, adaptive energy management. Addressing this gap, the thesis introduces the Adaptive Amoeba Battery Curve Mapping Management System (AABCMS), a biologically inspired subsystem that predicts battery health trajectories and dynamically modulates operational states, ranging from active processing to ultra-low power sleep modes. The AABCMS mirrors biological neural energy management, adjusting system behaviour based on real-time energy availability to maximise device longevity without compromising security.
The entire system was implemented and validated on a custom ESP32-S3 development board and benchmarked against an ATMEGA328P microcontroller, encompassing extensive cycle, timing, power, and energy consumption analyses. Testing demonstrated the system's adaptive encryption selection, session integrity preservation during power fluctuations, desynchronisation recovery through honeypot redirection, and sustained security under energy-limited conditions.
The thesis concludes by situating this bio-inspired security architecture within broader technological trends, highlighting its potential synergy with machine learning, large language models (LLMs), and future quantum-resilient cryptographic methods. By uniting principles from immunology, embedded systems, cryptography, and adaptive energy management, this research contributes a pioneering interdisciplinary approach to the sustainable, secure, and autonomous evolution of IoT systems.
 |
View Item |