Pioneering advanced security solutions for reinforcement learning-based adaptive key rotation in Zigbee networks

Area of Science:

• Computer Science
• Cybersecurity
• Artificial Intelligence

Background:

• Zigbee networks are crucial for Internet of Things (IoT) communication but face security vulnerabilities.
• Key management and resilience against distributed denial of service (DDoS) attacks are significant challenges.
• Traditional key rotation methods lack dynamic adaptation to changing network conditions.

Purpose of the Study:

• To propose and evaluate a novel reinforcement learning (RL) model for adaptive key rotation in Zigbee networks.
• To enhance network security, efficiency, and resilience against cyber attacks.
• To compare the RL model's performance against traditional key rotation strategies.

Main Methods:

• Implemented a reinforcement learning (RL) model for adaptive key rotation.
• Tested the RL model in a simulated Zigbee network environment.
• Evaluated performance over 30 days using metrics like network efficiency, DDoS response, resilience, latency, and packet loss.

Main Results:

• The RL model significantly outperformed traditional methods (periodic, anomaly detection, heuristic, static).
• Demonstrated improved network efficiency, higher intrusion detection rates, and faster response times to DDoS attacks.
• Showcased superior resource management and network resilience under various simulated attacks and fluctuating traffic.

Conclusions:

• AI-driven adaptive strategies, specifically RL, offer a robust solution for enhancing Zigbee network security.
• The proposed RL model provides a more intelligent and effective approach to key management in IoT environments.
• This research paves the way for more secure and resilient wireless communication in the evolving IoT landscape.