CAREER: InteractiveRF: Fully-Adaptive, Physics-Aware RF-Enabled Cyber-Physical Human Systems

点击翻译按钮获取中文摘要

As technology advances and an increasing number of devices enter our homes and workplace, humans have become an integral component of cyber-physical systems (CPS). One of the grand challenges of cyber-physical human systems (CPHS) is how to design autonomous systems where human-system collaboration is optimized through improved understanding of human behavior. A new frontier within this landscape is afforded by the advent of low-cost, low-power millimeter-wave radio frequency (RF) transceivers, which can be exploited almost anywhere as part of the Internet-of-Things, smart environments, and personal devices. RF sensors provide a unique, information rich dataset of high-resolution measurements of distance, direction-of-arrival, and micro-Doppler signature in a non-contact, non-intrusive fashion in most weather conditions and in the dark. This CAREER project aims to pave the way for new and innovative RF-enabled CPHS applications in service of society and a better quality-of-life by transforming current fixed-transmission RF sensors into intelligent devices that can autonomously respond to human and environmental dynamics to optimize CPHS performance. Due to the burgeoning commercial sector utilizing radar across a variety of fields, such as transportation, health and human-computer interaction, this project features integrated academic preparation for multi-disciplinary, convergence research at both undergraduate and graduate levels to educate a new generation of engineers with experience in RF sensing, machine learning, signal processing and CPHS applications. Through K-12 outreach activities and recruiting at local historically black colleges and universities (HBCUs), this project will enrich and motivate students to study STEM fields, laying the foundations for a diverse and globally competitive STEM workforce for the future.This CAREER project simultaneously addresses critical challenges currently limiting effective exploitation of RF sensors in CPHS, such as the problem of RF data scarcity for training deep models, the wide range and continuity of possible human movements, the presence of other people and obstacles, and the dynamic nature of real-world scenes. Specific contributions include the development of 1) physics-aware ML techniques that leverage the domain knowledge embodied in models with data-driven deep learning; 2) spatio-temporal parsing techniques to extract and recognize human signal components from RF data streams to improve robustness of RF-CPHS under real-world conditions; and 3) a new task-cognizant, fully-adaptive RF sensing framework to improve performance and robustness of RF-CPHS for varying tasks in dynamic real-world environments. The proposed fully-adaptive RF framework also paves the way for collaborative, multi-modal RF-CPHS by exploiting information learned from RF and other sensor modalities in its decision process.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.