An Adaptive Mac Protocol (DQ-MAC) for Efficient Dynamic Spectrum Access in Cognitive Radio Networks

Cognitive Radio Networks (CRNs) require dynamic spectrum access as a way to maximize the use of inadequate spectrum resources with minimal interference from licensed primary users. MAC protocols of a traditional nature frequently fail to respond effectively in real-time to changing channel availability, resulting in poor spectrum utilization and high rates of collisions. This work presents a new Deep Q-Network (DQN)-based MAC protocol that learns and adapts to the shifting spectrum environment, allowing secondary users to make insightful, instantaneous channel access choices. The development and assessment of the protocol occurred across different environments—urban, rural, and indoor—representing unique ranges of spectrum availability and interference issues. The simulations ran on MATLAB, utilizing actual user mobility, Rayleigh fading, interference, and noise conditions in the real world. Results show that the DQN-based MAC protocol markedly outperforms traditional random channel selection across major performance assessments, realizing up to 71% higher throughput, 58% less collisions, and improved equity within user interactions. The results show improvement in the spectrum efficiency and user performance in real time.