研究目的
To propose a framework for detailed daily activity recognition, distinguishing between static and dynamic activities and their intense counterparts using wearable and smartphone-embedded sensors.
研究成果
The proposed framework achieves more than 94% accuracy in recognizing detailed daily activities, including their intense counterparts, using a combination of smartphone-embedded accelerometer and wearable heart rate sensor data.
研究不足
The framework's performance is validated on a limited number of subjects (four healthy individuals), which may not represent the diversity of user behaviors and physical conditions.
1：Experimental Design and Method Selection:
The framework uses an ensemble of classifiers with weighted majority voting for classification.
2：Sample Selection and Data Sources:
Data was collected from four healthy subjects performing twelve detailed activities.
3：List of Experimental Equipment and Materials:
Smartphone-embedded accelerometer and wearable heart rate sensor.
4：Experimental Procedures and Operational Workflow:
Data collection, preprocessing, feature extraction, feature selection, and learning and sensor fusion.
5：Data Analysis Methods:
The ensemble classifier's performance was evaluated using accuracy and error metrics.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
