Abstract

Cognitive Networks was introduced to resolve issues by incorporating intelligence into the network functions. Hence, a hybrid algorithm was developed to improve spectrum resource allocation in a cognitive network for distributed computing. Five hundred thousand data sets were collected from Nigeria Communication Commission (NCC) repository. Artificial Neural Network was used to divide the dataset into 4 stages UHF, FM, GSM 900 and DCS 1800. The FM is for radio stations, UHF is for television stations, GSM 900 is for MTN, GLO, Airtel and DCS-1800 is for Etisalat networks. The hybrid algorithm was achieved using Hidden Markov Models (HMMs) which was combined with Markov-based Channel Prediction Algorithm (MCPA) in CN for dynamic spectrum allocation and higher efficiency of the spectrum holes, for primary users in order to improve the accuracy. The MCPA approach can prevent possible collisions, as the CN leaves the channel before data transmission is being initiated by the primary users. The combination of the mentioned two algorithms were simulated using MATLAB simulator 2019b for accuracy test. The result showed that the efficiency of the radio networks was found to be closed. The CN signal strength varies while the signal power and Signal to Noise Ratio (SNR) was monitored continuously. The efficiency of power showed increase from 76.66% to 86.82% for FM Broadcast, 76.91% to 86.82% for GSM-900, 78.19% to 89.04% for DCS-1800 and 78% to 88.55% for UHF TV. The lower the interference, the better the reception of signal on the secondary users. The results were compared with the existing technology at NCC. The higher the SNR, the better the interference reduction. The best signal response was at 12db SNR, the interference was able to reduce to 25%. The computation results showed a level of improvement in spectrum occupancy license distribution from 19.6 to 61.1%. In conclusion, it was demonstrated that a hybrid algorithm offered better solution in wireless networking by using an improved algorithm to provide a cognitive spectrum occupancy technique in a wireless network. It was recommended that the telecommunication companies should adopt the improved algorithm to further enhance CN in their operations.