Low-Complexity High-Accuracy 5G and LTE Multichannel Spectrum Analysis Aided by Unsupervised Machine Learning

Abstract

In this paper we propose a new method of occupied spectrum analysis for channel detection in a shared spectrum environment. Our approach is based on iterative multi-stage multi-resolution scanning using configurable Sliding Discrete Fourier Transform (SDFT) aided by an Unsupervised Machine Learning (UML) clustering method. The proposed low-complexity high-accuracy real-time spectrum scanning and channel detection is simulated for multiple Radio Access Networks (RAN) of Long-Term Evolution (LTE) & Fifth Generation (5G) channels in a shared frequency band. The results of the simulation show possible successful utilization of the proposed method as a sensing tool for spectrum sharing management and other applications where accurate channel detection occupancy is required.