An adaptive, energy-efficient and secure routing protocol for zone-related mobile Ad-hoc networks using reinforcement...

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Mobile Ad Hoc Networks (MANETs) face routing challenges due to dynamic environments and security threats like wormhole attacks.
• Traditional routing protocols struggle to maintain performance and security against evolving threats.
• Intelligent and adaptive routing is crucial for seamless communication in MANETs.

Purpose of the Study:

• To propose a novel secure routing protocol for MANETs using reinforcement learning.
• To enhance network stability, security, and energy efficiency in dynamic MANET environments.
• To address the limitations of traditional routing protocols against sophisticated attacks.

Main Methods:

• Developed the Reinforcement Learning-Based Secure Routing Protocol (RLSRP) utilizing adaptive k-hop clustering and deep Q-Networks (DQN).
• Implemented zone-related clustering for collaborative path optimization based on real-time network conditions.
• Evaluated network conditions by measuring latency variations and detecting anomalous nodes to identify threats.

Main Results:

• RLSRP demonstrated superior performance in large-scale simulations (up to 10 million nodes) using Dask and TensorFlow.
• Achieved a Packet Delivery Ratio exceeding 99%, significantly reduced latency, and improved energy efficiency compared to existing protocols.
• Outperformed FSSAM, Cluster-RL, and Reputation-based Q-learning in key performance metrics.

Conclusions:

• RLSRP offers a secure, scalable, and efficient routing solution for practical MANET applications.
• Deep reinforcement learning effectively enhances security and routing performance in zone-related MANETs.
• The proposed protocol ensures robust communication in large-scale, dynamic network environments.