Concurrent Learning for CSI-Based Applications in Smart Environments

Abstract

Due to the omnipresence of radio frequency signals, the Channel State Information (CSI) can offer an alternate source to image, video, and other high-dimensional streams in a great many Internet-of-Things (IoT) applications. As a result, an ever increasing number of researchers are advocating for the use of passive CSI data for ranging, tracking, perception and automation across many domains such as robotics, healthcare, and surveillance. Specifically in indoor environments where movements cause classifiable effects, the CSI can be leveraged to provide a high-dimensional signal source for a broad set of applications including activity, gesture, pose, location, and orientation recognition. This task however remains a challenge on two accounts. On the one hand, the radio frequency channel is highly susceptible to environment changes and artifacts. On the other hand, distributed models need to strike a balance between efficiency and performance. In this work, we focus on tackling these issues by proposing a novel scalable architecture that is robust against environment variations and achieves high accuracy across multiple tasks in practical settings using light weight neural networks.