U.S.-Ireland R&D Partnership:CNS:Small:SWEET: Hardware and Software for Sustainable Wearable Edge Intelligence

实时远程监测生理指标和早期干预可以挽救生命。这些关键服务需要具有强大预测能力的可穿戴技术、快速网络和快速服务器来从收集的数据中提取见解。不幸的是，数亿人无法使用这些技术组件，特别是那些生活在宽带连接有限和投资本地计算和通信基础设施有限的地区的人。在这个项目中，我们为可持续和高效的可穿戴边缘智能开发硬件和软件。我们为服务不足的社区解决可穿戴健康监测设备和人工智能服务的基本可及性和可持续性挑战。我们的研究，教育和推广计划是锚定的可持续性和可访问性为重点的计算机系统研究的观点。基于机器学习的医疗服务严重依赖于庞大的数据存储、快速的网络和云服务器群，而世界上大部分人口都无法访问这些服务。这项工作的智力挑战在于如何改变硬件和软件设计，为无法依赖无线或云服务提供商的弱势用户提供先进的机器学习服务。这一挑战的基础是具体的智力挑战(i)延长执行生物医学信号采集和处理的可穿戴设备的使用寿命，同时试图扩展其计算和处理能力；（ii）在云之外的个人和边缘计算设备上执行更高效、稳健和值得信赖的机器学习；（iii）在没有云数据中心的健壮性和可用性保证的情况下，为分布式机器学习服务寻找可扩展和可持续的开发和部署模型。该项目汇集了四个研究团队，他们在可穿戴传感器和硬件、软件、系统和算法方面具有示范和互补的专业知识。我们最近在降低边缘传感器功耗、透明计算、无服务器计算和网络系统优化方面的研究为本研究奠定了基础和起点。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Real-time remote monitoring of physiological indicators and early intervention can save lives. These critical services require wearable technologies with strong predictive abilities, fast networks, and fast servers to extract insights from the collected data. Unfortunately, these technology components are inaccessible to hundreds of millions of people, specifically those living in areas with limited broadband connectivity and limited means to invest in local computing and communication infrastructure. We develop hardware and software for sustainable and efficient wearable edge intelligence in this project. We address fundamental accessibility and sustainability challenges of both wearable health monitoring devices and artificial intelligence services for under-served communities.Our research, education, and outreach plans are anchored on a sustainability- and accessibility-focused view of computer systems research. Health services based on machine learning lean heavily on vast data stores, fast networks, and farms of Cloud servers, which are inaccessible to large parts of the world’s population. This effort's intellectual challenges lie in how to change hardware and software design to bring advanced machine learning services to unprivileged users who cannot depend on wireless or Cloud service providers for their well-being. Underlying this challenge are specific intellectual challenges in (i) lengthening the lifetime of wearable devices that perform biomedical signal acquisition and processing while trying to expand their computational and processing capabilities; (ii) performing more efficient, robust, and trustworthy machine learning in personal and edge computing devices outside the Cloud; and (iii) finding scalable and sustainable development and deployment models for distributed machine learning services, without the robustness and availability guarantees of Cloud datacenters. The project brings together four research teams with demonstrated and complementary expertise in wearable sensors and hardware, software, systems, and algorithms. Our recent research on reducing power consumption of edge sensors, transprecise computing, serverless computing, and network systems optimization lays the foundation and serves as a starting point for this research.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.