Secure and Energy-Aware Data Dissemination in Wireless Sensor Networks Using Federated Deep Reinforcement Learning and Quantum-Inspired Optimization
Year: 2026
Type: article
Abstract: Wireless Sensor Networks (WSNs) have gained significant application in sensitive applications like smart cities, military, and healthcare because they are self-forming and thus can overcome data collection and transmission delay. Nevertheless, energy shortage, delay in transmitting data and exposure to security attacks are some of the problems that WSNs encounter that are deemed as critical. The schemes of conventional Centralized Deep Reinforcement Learning (DRL) have high communication overhead, privacy threats, and drawbacks of scale. The study will suggest a safe and power-efficient routing mechanism using the Federated Deep Reinforcement Learning (FDRL) technique combined with the Quantum-Inspired Optimization (QIO). The principal advancement is in developing a decentralized routing system founded on learning in an equal balance of the energy associated with routing, delay, and security. The introduced approach has less data exposure because of the federated updates and improves routes search with QIO. Simulations demonstrate the consistent Packet Delivery Ratio (PDR) of 100% on all nodes-density levels whereas the energy consumption increases corresponding to the number of nodes yet it is highly effective. The compromised value is maintained as less than 10% and Q-values in FDRL are stable as compared to centralized DRL. Scalability of the system is also high and the computational overhead involved is low and the system proved to be attack immune thus the system is very appropriate in dynamic WSN environments.
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7224361876
