AI-Based Cognitive Wireless Sensor Network for Dynamic Spectrum Access
DOI:
https://doi.org/10.71366/ijwos03042623038Keywords:
Cognitive Wireless Sensor Networks, Dynamic Spectrum Access, Random Forest, Machine Learning, Energy Efficiency, Spectrum Prediction, Underground Monitoring, Cognitive Radio, Resource-Aware Communication, Interference Mitigation.
Abstract
Wireless Sensor Networks (WSNs) deployed in harsh environments require efficient spectrum
utilization and low energy consumption. Conventional spectrum sensing methods, such as energy
detection, involve continuous monitoring, resulting in high power usage and unreliable
performance in dynamic conditions. This paper proposes an AI-based Cognitive Wireless Sensor
Network (CWSN) that predicts channel availability using a Random Forest classifier at the sensor
node level. A resource-aware retransmission strategy is incorporated, allowing retransmissions
only when prediction confidence and residual energy exceed predefined thresholds. The proposed
system improves spectrum efficiency, reduces interference, and enhances energy performance.
Simulation results demonstrate 96% prediction accuracy, an 80% reduction in collisions, and
approximately 20.5% energy savings, making it suitable for long-term deployment in challenging
environments.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
