Classification of transmission environment in UWB communication using a support vector machine

Abstract

Impulse-radio(IR)-based Ultra-wide band (UWB) technology has great potential in the fields of positioning, target detection and data transfer due to its significant advantages. In this paper, the ability of UWB to perceive different transmission environment is discussed. Four scenarios are simulated using finite-difference time-domain (FDTD) method. There are three NLOS(obstacle are concrete wall, glass wall, wood wall respectively)scenarios and one LOS scenario. Representative parameters are extracted from the UWB's channel models, and are sent into a support vector machine(SVM) to classify those scenarios. The results show that these scenarios can be classified with proper SVM parameters.