A Study on High-Efficiency Energy Detection-Based Spectrum Measurements

Abstract

Statistical information in terms of spectrum occupancy is useful for the efficient and smart dynamic spectrum sharing, and it can be obtained by long-term, broadband, and wide-area spectrum measurements. In this paper, we investigate an energy detection (ED)-based spectrum measurements, in which the noise floor (NF) estimation is a key functionality for the appropriate ED threshold setting. Typically, the NF has the slowly time- varying property and frequency-dependency, and several NF estimation algorithms, including forward consecutive mean excision (FCME) algorithm-based method, have been proposed. However, these methods did not deeply consider the slowly time varying property of the NF and is computationally inefficient. Accordingly, we propose a computational complexity reduction algorithm based on NF level change detection. This algorithm is computationally efficient, since it skips the NF estimation process when the NF does not change. In numerical evaluations, we show the efficiency and the validity of the proposed algorithm.