Prediction of Broadband Spectrum Utilization Data for Spectrum Sharing

The 5 th generation mobile network system (5G) comprises the infrastructure of next generation communication services such as large-capacity video delivery services and real time entertainment services. It is extremely important to secure more spectrum resources to satisfy the 5G performance requirements. Solutions include spectrum sharing technology. This paper shows that the sequences consisting of the channel vacancy durations follow the exponential distribution by using the spectrum utilization data in an urban area where the spectrum is used at high density. The spectrum usage state could be described by using the channel vacancy periods. Furthermore, a prediction method is proposed of the future channel vacancy period with the frequent sequence mining technology by using short time period observation data. The result of the evaluation experiment is also reported.