Diagnosing health problems from gait patterns of elderly

A system for diagnosing health problems from gait patterns of elderly to support their independent living is proposed in this paper. Motion capture system, which consists of tags attached to the body and sensors situated in the apartment, is used to capture gait of elderly. Position of the tags is acquired by the sensors and the resulting time series of position coordinates are analyzed with machine learning algorithms in order to recognize the specific health problem. We propose novel features for training a machine learning classifier that classifies the user´s gait into four health problems and a normal health state. Results showed that decision tree classifier was able to reach 95% of classification accuracy using 7 tags and 5 mm standard deviation of noise. Neural network outperformed it with classification accuracy over 99% using 8 tags with 0–20 mm noise. Control panel prototype has been developed to provide explanation of the automatic diagnosis.