An Efficient Predictive Spectrum Allocation Algorithm for Multi-User Cognitive Radio IoT Devices Within GSM900 Band

IoT, fifth generation and multiple new wireless systems are growing up rapidly and spectrum resource is being congested continuously. Therefore, an efficient spectrum allocation strategy should be taken place to tackle this serious situation. This work presents a predictive spectrum allocation algorithm designed to enhance the transmission performance of cognitive radio Internet of Things (CR-IoT) devices operating within the GSM900 band. Our proposed system combines spectrum prediction and sharing to identify and allocate unused frequency channels efficiently. This dual-process approach optimizes channel access for cognitive users (CUs), maximizing transmission duration and bandwidth sharing while minimizing the need for spectrum handoffs (SHOs). The algorithm is evaluated using three key performance metrics: “Throughput in time-domain”, “Throughput in frequency-domain”, and “Spectrum Handover rate” (SHO) (for QoS enhancing and latency decreasing). Results indicate that the proposed allocation algorithm achieved a global TAR of 78.47%, a maximum allocation rate of more than 96.8% of available channels for 2158 simultaneously accessed CUs, a high interference avoidance rate (IAR) of 94,5% in frequency-domain, and a low SHO rate of approximately 0.19%, highlighting its effectiveness in supporting seamless and robust spectrum access for multi-user CR-IoT applications without interfering with licensed users.