Fall Detection Algorithm for the Elderly Based on Trial Acceleration and Heart Rate

The paper introduces a method of judging the gesture according to triaxial acceleration data and heart rate data. The acceleration transducer is used to collect the triaxial acceleration data, and some experimental data is selected as training data. The collected data is classified according to KNN algorithm, and benchmark data is designed in each classification. The feature of falling data is achieved by comparing gesture data. For application, the matching algorithm should be designed, and the similarity degree with benchmark data should be computed. Lastly, the heart rate is used for judgment and is weighted to get the matching results.