Artificial Intelligence Laboratory · Hosei University
We focus on developing smart support systems for individuals in need, particularly older adults, people with hearing and visual impairments, and those experiencing bullying.
Support requires a chain from safe sensing to accountable action.
Direction 01 / 05
A millimetre-wave radar system recognises everyday activities and detects falls, without a camera in the home.
A practical activity-recognition pipeline for bathroom-like rooms: radar point clouds, DBSCAN and KDE denoising, temporal modelling, and an alert for review.
Can the same radar follow heart-related signals without touching the body? In a pilot the median heart-rate error was about 6.8 bpm, and heart-rate variability cannot yet be recovered reliably. We report this openly and keep working on it.
Reminders and support that adapt to the person, not only to the schedule.
Direction 02 / 05
Non-contact millimetre-wave sensing, boundary extraction and a Bi-LSTM with attention turn gestures into text.
Service robots that recognise white-cane and walking-stick users and guide them in hospitals and large shops.
Direction 03 / 05
60 GHz FMCW radar for in-cabin sensing: detecting a child left in a vehicle, an urgent, time-critical scenario.
Abusive-language recognition from speech, violence detection from activity, and emotional-state analysis, combined into a risk estimate for early intervention.
Millimetre-wave radar and audio recognise fighting behaviour so that people can step in early.
The aim is not constant surveillance. The aim is context-sensitive protection, minimal necessary sensing, human oversight, and clear accountability.
Direction 04 / 05
The four examinations — inspection, listening and smelling, inquiry, palpation — supported by tongue and pulse analysis and by diagnosis agents that use retrieval-augmented generation.
Simulated sessions that help students practise therapeutic conversation.
Direction 05 / 05
A model of objects, their relations, and what they afford, so a robot can reason about a real home.
An augmented scene graph of the home, LLM planning grounded in that model, and a robot that asks for help when it is uncertain.
Recognising indoor environments and planning safe paths through them.
Best Student Paper
2023 IEEE Smart World Congress
Study of Feature Fusion Models for Human Activity Recognition on Point Clouds from mmWave Radar
Proceedings of the 2026 9th International Conference on Software Engineering and Information Management
2025 IEEE Smart World Congress (SWC)
2024 16th IIAI International Congress on Advanced Applied Informatics (IIAI-AAI)
2023 IEEE Smart World Congress (SWC)Best Student Paper
IEEE Internet of Things Journal
2022 IEEE 21st International Conference on Cognitive Informatics & Cognitive Computing (ICCI*CC)
Future Generation Computer Systems
2024 Joint 13th International Conference on Soft Computing and Intelligent Systems and 25th International Symposium on Advanced Intelligent Systems (SCIS&ISIS)
IEEE Vehicular Technology Magazine
2026 9th International Conference on Artificial Intelligence and Big Data (ICAIBD)
Proceedings of the 2026 9th International Conference on Software Engineering and Information Management
2025 IEEE Cyber Science and Technology Congress (CyberSciTech)
2025 IEEE Conference on Cloud and Big Data Computing (CBDCom)
2025 IEEE Cyber Science and Technology Congress (CyberSciTech)
12th International Conference on Behavioural and Social Computing (BESC 2025), Lecture Notes in Computer Science, vol. 16432
IEEE Transactions on Computational Social Systems
Proceedings of the 2026 9th International Conference on Software Engineering and Information Management
2026 20th IIAI International Congress on Advanced Applied Informatics (IIAI-AAI)
Full lists of publications:Google ScholarDBLP