Intelligent Environment & Wireless Sensing

An intelligent environment is a physical space equipped with information and communication technology, sensor systems, and embedded computing so that it can understand human states and behavior and provide support naturally. In our lab, we pursue spaces that behave intelligently through wireless sensing, portable 3D LiDAR, privacy-preserving spatial sharing, and physiological data analysis.

A major feature of intelligent environments is that they can understand context and provide timely support without requiring explicit user commands. At the same time, we address many practical challenges, including avoiding excessive intervention, collecting data with privacy in mind, and realizing sensing systems that minimize battery replacement and deployment burden.

Wireless Sensing and Backscatter Tags

We study wireless sensing technologies that use existing radio signals such as Wi-Fi and Bluetooth to understand human activities and surrounding environments without physical contact. By capturing different reflection characteristics caused by material properties and spatial configurations, we connect radio observations to identification of people and objects, activity recognition, and scene understanding.

We also develop tags based on ultra-low-power backscatter communication. By controlling how surrounding radio waves are affected, we separate the influence of multiple targets, infer nearby situations such as conversation or TV watching, and pursue monitoring systems that reduce the burden of battery replacement.

Portable 3D LiDAR Device "Hitonavi-µ"

We developed "Hitonavi-µ," a portable sensing device based on micro-sized 3D LiDAR sensors, and study activity recognition and anomaly detection for elderly care, monitoring, and daily-life support. A key feature is that the device can capture human motion and posture three-dimensionally while remaining compact enough to be carried and deployed easily.

See the project page for Hitonavi-µ here.

Privacy-Preserving Spatial Sharing and the Metaverse

In remote spatial sharing and metaverse systems, convenience must be balanced with visual privacy and careful handling of spatial information. We study privacy awareness in immersive spatial-sharing systems as well as per-user control of visible regions and information presentation.

See the spatial-sharing project here.

Behavior Understanding with Smartphone Sensing

We also study methods that capture subtle changes in daily life by using mobility and usage patterns obtained from smartphone sensors. As an application in education and welfare, we investigate ways to gently detect early signs of school attendance issues.

Physiological Data Analysis and Core Body Temperature Estimation

We use physiological data obtained from wearable sensors and thermography to study core body temperature estimation, early heatstroke detection, stress-level estimation, and related tasks. We emphasize methods that can be measured in practical settings such as hospitals, sports, and universities, with a view toward real-world deployment.