On-line Automatic Detection of Human Activity in Home Using Wavelet and Hidden Markov Models Scilab Toolkits

In this work, a pilot study of on-line automatic detection of human activity in home using wavelet and hidden Markov models Scilab toolkits was carried out. The collected raw data are provided by a biaxial accelerometer ADXL202E attached to the person. Several activities were simulated by the researchers (walking slowly , walking quickly, sitting down-getting up, fall during walking, fall from a position upright, ...). The feature vectors of these data were then used to build different hidden Markov models of these activities with different persons. The built models were employed for online detection of these activities. The obtained results are very promising.