Exploring gesture recognition with low-cost CW radar modules in comparison to FMCW architectures

Abstract

Radar-based hand gesture recognition is an area receiving a significant amount of interest in recent years due to the rapid increase in the availability of low-cost low-footprint RF sensors. The most common configuration is Frequency Modulated Continuous Wave (FMCW), whereas Continuous Wave (CW) radar is not receiving as much attention. In this paper we explore the use of extremely low cost CW radar modules for gesture recognition. In doing so a set of signal processing electronics is developed, implemented, and used to supply the resulting signal to PC audio input for recording. A dataset of gestures was recorded and gesture recognition accuracy was compared to FMCW recordings to show that CW systems can provide a high accuracy for gesture recognition at a very low cost.