Deep reinforcement learning based neurosense scheme for energy-efficient and secure wireless sensor networks.

Area of Science:

• Computer Science
• Network Engineering
• Artificial Intelligence

Background:

• Wireless Sensor Networks (WSN) face a critical challenge in balancing energy efficiency with robust security measures.
• Existing strategies often struggle to dynamically adapt to evolving network conditions and security threats.
• Optimizing data transmission is key to prolonging network lifespan and maintaining data integrity.

Purpose of the Study:

• To introduce NeuroSense, a novel strategy leveraging deep reinforcement learning (DRL) for WSNs.
• To achieve an optimal trade-off between energy efficiency and network security.
• To enhance real-time decision-making for data transmission paths.

Main Methods:

• NeuroSense employs deep reinforcement learning (DRL) models to learn and optimize data transmission paths dynamically.
• The strategy integrates security-state information into the DRL decision-making process.
• Routing and protection decisions are adjusted in response to simulated attack scenarios.

Main Results:

• NeuroSense demonstrated superior performance compared to five baseline strategies in simulations.
• Key improvements were observed in total energy consumption and communication delay.
• Enhanced resilience against security threats and more even energy depletion across nodes were noted.

Conclusions:

• NeuroSense effectively optimizes energy efficiency and security in WSNs through intelligent, DRL-driven routing.
• The strategy contributes to extended network operation and improved resilience.
• NeuroSense offers a promising approach for secure and energy-aware WSN management.