Prediction of 90-degree turn using kinematic parameters for implementation of home robot

To adequately adapt to aging and lifestyle diversification, this paper proposes a Biofied building that provides a safe, secure and comfortable space for residents using a small sensor-mounted robot. The proposed robot acquires the positions and activities of residents in real-time. Moreover, the prediction of occupant behavior is effective for realizing more natural robot control, such as predicting the risk of falling or choosing a moving path that does not disturb the occupant. This paper proposes a method for predicting the 90-degree turn, which is a non-steady motion of activities of daily living using kinematic parameters.