CAREER: Safeguarding Smart Homes: A Foundation for Building User Privacy-Enhancing IoT Systems

人们越来越多地部署物联网（IoT）设备，以实现智能家居自动化。不幸的是，这些设备生成的网络流量数据可能包含敏感信息，这些信息可能会通过基于人工智能（AI）的恶意隐私攻击而暴露。目前的防御方法没有充分考虑这些新兴的“智能”攻击。该项目旨在通过一个数据驱动的、隐私友好的物联网管理框架来应对理解和防御这些新的基于人工智能的攻击的挑战，该框架使人们能够理解和管理家庭信息的泄漏。该项目将研究增强智能家居用户隐私的新技术。具体而言，该项目将（1）开发新的基于人工智能的隐私攻击模型，以更好地了解隐私漏洞;（2）构建新颖、低成本和分布式的流量重塑方法，以保护用户隐私;（3）设计新的基于人工智能的优化方法，以适应重塑配置文件，以反映真实世界的用户行为，进一步保护隐私。该项目的成果将作为构建物联网系统的基础，使人们能够更好地控制数据收集和共享。这三项技术创新将通过模拟智能家居的真实实验以及行业合作者的实验室进行原型设计和验证。研究的具体成果将包括一系列算法、机制和原型，这些算法、机制和原型将推动智能家居物联网安全和隐私的最新发展，而这一领域目前还没有得到足够的关注。尽管智能家居市场增长迅速。该项目还将为科罗拉多公立中学/高中教师开发物联网网络安全夏季研讨会，重点招聘黑人，棕色人种和有色人种的土著居民。在物联网系统，网络安全和人工智能方面开发的实践课程模块将用于培训K-12，本科生和研究生。该项目还将在当地的小学、初中和高中开展计算机科学推广活动，以指导下一代物联网网络安全和人工智能研究人员。该奖项反映了NSF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。

People are increasingly deploying Internet of Things (IoT) devices to enable smart home automation. Unfortunately, network traffic data generated by these devices can contain sensitive information that may be exposed through malicious artificial intelligence (AI)-based privacy attacks. Current defense approaches do not adequately consider these emerging “smart” attacks. This project aims to address challenges in understanding and defending against these new AI-based attacks through a data-driven, privacy-friendly IoT management framework that enables people to understand and manage the leakage of in-home information. This project will investigate new techniques for enhancing smart home user privacy. Specifically, this project will (1) develop new AI-based privacy attack models to better understand privacy vulnerabilities; (2) build novel, low-cost, and distributed traffic reshaping approaches to safeguard user privacy; and (3) design new AI-based optimization methods for adapting the reshaping profiles to mirror real-world user behavior, further safeguarding privacy. The outcomes of this project will serve as a foundation for building IoT systems that enable people to have greater control over data collection and sharing. The three technical innovations being pursued will be prototyped and validated through real-world experiments in mock smart homes, as well as within industry collaborators’ laboratories.The specific outcomes of the research will include a family of algorithms, mechanisms, and prototypes that will advance the state-of-the-art in strengthening IoT security and privacy for smart homes, an area that has received relatively little attention, despite the rapidly growing smart home market. This project will also develop IoT cybersecurity summer workshops for Colorado public middle/high school teachers, with a recruitment emphasis on districts with significant populations of Black, Brown, and Indigenous People of Color. Hands-on course modules developed on IoT systems, cybersecurity, and AI will be used to train K-12, undergraduate, and graduate students. This project will also conduct computer science outreach events at local elementary, middle, and high schools to guide the next generation of IoT cybersecurity and AI researchers.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.