Dynamic Channel Reservation in Cognitive Radio Networks Using AI Driven SDN Approach

In cognitive radio networks (CRNs), secondary users (SUs) are permitted transmission access by using licensed channels previously assigned to primary users (PUs). Dynamic Spectrum Access (DSA) and spectrum sharing contribute to enhancing cooperative communications for PUs. Considering the scarcity of spectrum resources, extensive research attention has been dedicated to dynamic bandwidth allocation in recent years. In CRNs, dynamic channel reservation (DCR) improves network efficiency by optimizing the number of reserved channels. This paper proposes a centralized approach to DCR using a software-defined network (SDN) integrated with an artificial intelligence (AI) model. The proposed AI model employs a neural network architecture trained on historical network data to predict and manage channel allocation in real time. Key inputs to the model include PU and SU arrival rates, channel failure, and repair rates, enabling it to adaptively allocate reserved (R-CRN) and non-reserved channels (N-CRN) to balance PU channel availability and retainability while maintaining reliable access for SUs. The AI-driven controller reduces operational complexity and dynamically optimizes resource allocation, even in the presence of varying network conditions and failures. Simulation results demonstrate that this approach reduces PU blocking probability by up to 15% and lowers operational costs by 10%, while improving Quality of Service (QoS) metrics such as throughput and channel utilization. These quantifiable improvements highlight the adaptability and efficiency of the proposed solution, positioning it as a robust method for future spectrum management in CRNs.