Possibility of Dynamic Spectrum Sharing System by VHF-band Radio Sensor and Machine Learning

In this paper, we propose an outdoor location estimation scheme of a high-priority wireless system by using VHF-band radio sensors and a machine learning technique for dynamic spectrum sharing (DSS) systems. The location is estimated by machine learning of delay profiles measured in the VHF-band. By using the estimated location of the high-priority terminal, more precise protection area can be calculated. As a feasibility study, delay profiles were measured in a mountainous environment in Japan by the ARIB STD-T103 system operating in the VHF-band. The profiles and the location information at the measurement points are learned by the deep neural network (DNN). By using the trained DNN, the location cluster of the high-priority terminal can be predicted without the GPS by only measuring the delay profile of the high-priority terminal. In the evaluation, the total correct localization rate of the proposed scheme is up to 80.0 %.