Structured Mobile Activity Recognition and Sensor Based Healthcare

Sozo Inoue,
2nd Meeting on Interaction with Smart Artefacts
(Not Available)
(Not Available)
Schloss Dagstuhl, Germany
In this talk, we address a method for structuring high-level activity recognitions using mobile sensors. We use unsupervised multi-level segmentation, which could be used for a pre-process of activity recognition, and could construct a multi-level activity recognition using accelerometers on mobile phone. We extend single-level segmentation to multi-level by sweeping the temporal parameter. To confirm the validity of our approach. we pursued the experiment of gathering accelerometer data of real nursing in a hospital.

We also address large-scale experiments for sensor-based preventive healthcare. After an experiment of 3 month with 50 diabetic patients, we found effective aspects of motivating users and preventing deterioration. This project is being extended to in-hospital settings and to more demanded fields such as developing countries.

Data Files