| Abstrak/Abstract |
The rapid growth of the Internet of Things (IoT) has accelerated the adoption of smart home technologies, offering convenience and automation in daily life. However, this interconnected environment increases the risk of cyber threats, making information security a pressing concern. To address this, the study presents the design and implementation of an adaptive Intrusion Detection System (IDS) based on machine learning, deployed on a Raspberry Pi platform as a low-cost, flexible, and energy-efficient solution for smart home security. Unlike traditional IDS approaches that rely on static, rule-based detection, the proposed system leverages adaptive learning algorithms to identify evolving attack patterns in real time. It integrates network traffic monitoring with carefully selected sensors and detection algorithms to improve responsiveness across various threat types from application-level exploits to network infrastructure attacks. System performance was evaluated through simulated attacks, including DDoS, brute force, and malware injection scenarios. Results show that the adaptive IDS significantly improves detection accuracy to 85%, surpassing the 65?curacy achieved by conventional methods. The response time was also reduced from 5 seconds to just 2 seconds, demonstrating the system’s suitability for real-time threat mitigation in resource-constrained environments. The Raspberry Pi acts as the IDS host and a firewall enhancement tool, supporting custom iptables rules, whitelist-based access control, and integration with the Elastic Stack for real-time logging and visualization. The system also supports continuous learning by updating its detection models based on new traffic patterns, making it scalable and resilient to future threats. This research contributes to IoT cybersecurity by demonstrating that an adaptive, machine learning-based IDS can be effectively implemented on lightweight hardware without sacrificing performance. It offers a cost-effective and scalable solution to secure smart home networks against increasingly sophisticated cyberattacks. |
