Multi-sensor system for detection and classification of human activities

Unsupervised discovery of activities of daily living characterized by their periodicity and variability

“…As a consequence there is an increasing trend to propose integrated frameworks which include various machine learning tools (Chen et al, 2010;Cao et al, 2012;Ugolotti et al, 2013). Moreover, the combination of machine learning tools with domain knowledge is also an interesting perspective to increase the systems performance (Luštrek et al, 2012;Simonin et al, 2015).…”

Non-Parametric Bayesian Human Motion Recognition Using a Single MEMS Tri-Axial Accelerometer

“…Methodologies investigated to this point follow heuristic classifiers, Gaussian mixture models (GMM) [16], support vector machines [12], and hidden Markov model (HMM) [17,18]. However, those approaches consider a fixed number of human motions.…”

New methods in the application of inertial and magnetic sensors in online pattern recognition problems

“…The most widely used activities were standing and walking, which can be found in almost all works. Besides standing, other stationary activities can also be found in the literature, such as lying [34], sitting [25,[41][42], or both [23,35,[38][39]. Different transitional movements were also parts of the activity classes in some works, e.g.…”

“…Different transitional movements were also parts of the activity classes in some works, e.g. sit-to-stand and stand-to-sit [35,38,41], lie-to-sit and sit-to-lie [35], lie-to-stand and stand-to-lie [41], or stopping after walking [36].…”

“…Special complex activities were also parts of the constructed databases, e.g. falling [24,41], jumping [23,33,42], writing [22], brushing teeth [40,44], eating and drinking [44], sweeping the floor, lifting a box onto a table, bouncing a ball [43], driving [34], cycling [23,44], etc.…”