Abstract
Recently it has become more important to monitor the daily human activities of the elderly and of children. In this paper, we propose a system for practical activity recognition using the Doppler effect in 24GHz microwaves. It extracts the features from the signals, selects the optimal features, and then classifies activities using a pattern matching technique. We can sense human activities simply with setting Doppler sensors on the wall or tables, without any body-attached sensors. As a result of performance evaluation, our system achieves over ninety percent in the classification of eight actions on average.
