Cooperative Environment Recognition Utilizing UWB Waveforms and CNNs

Abstract

Cooperative navigation enhances localization performance and situational awareness in challenging conditions, such as in tactical and first responder operations. In this work we demonstrate how the waveform of the Ultra Wideband (UWB) signal used for ranging in cooperative navigation can also be used to detect the environment surrounding the user of the navigation system. Different environments affect the wave-form in different ways, and thus the received waveform contains features characteristic to the environment around the receiver. We show how the received UWB signal waveform can be used in a Convolutional Neural Network (CNN) in order to determine whether the user is outdoors, indoors or in a forest. The environment is recognized correctly more than 90% of the time.