Collaborative Research: CISE Core Small: CSR: Multimodal Sensing-Based Security Mechanisms for IoT Systems

Multimodal sensing refers to the integration of multiple sensor types to collect various environmental data. By leveraging the strengths of different modalities, it enables more accurate and comprehensive understanding of complex scenarios. As a result, multimodal sensing is integral to mission-critical Internet-of-Things (IoT) systems such as smart cities, intelligent transportation, and defense applications. Robust cybersecurity is essential to their reliable operation, ensuring core objectives like encryption, authentication, and access control. However, while IoT systems increasingly rely on multimodal sensing, cybersecurity approaches that exploit this capability remain largely unexplored. This project addresses this gap by developing multimodal sensing-based security mechanisms to enhance the protection of mission-critical IoT systems. It also lays the foundation for broader applications of multimodal sensing in security. In parallel, the research advances scientific knowledge at the intersection of wireless networking, mobile sensing, and AI. Educational materials will be made publicly available, and the research will be integrated into curriculum development, undergraduate research, and increased participation in computing.

This project pursues a challenging research agenda, termed MSS, focused on developing, prototyping, and evaluating innovative Multimodal Sensing-Based Security (MSS) mechanisms for mission-critical IoT systems with inherently heterogeneous sensing capabilities. The research is organized into three integrated thrusts. Thrust 1 develops MSS-Key, a novel framework for establishing ad hoc secure keys between IoT devices by leveraging indirectly correlated multimodal sensing data shaped by unforgeable physical-domain randomness. Thrust 2 focuses on MSS-Val, a framework designed to protect multimodal sensing-native IoT systems against manipulated sensor inputs. Thrust 3 introduces MSS-Aug, a framework that streamlines and automates the integration of new sensing modalities into existing MSS systems via autonomous labeling and cross-domain generative learning. To support and evaluate the proposed techniques, the project will also develop a dedicated MSS testbed equipped with a wide range of sensor modalities. This testbed not only enables rigorous experimental validation but also serves as a hands-on educational platform, engaging students and promoting broader public awareness of cybersecurity challenges in IoT systems.

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.

Award Number: 2523517
Principal Investigator: Dianqi Han
Funds Obligated: $200,000
State: TX