Abstract
A real-time activity monitoring system within an Android based smartphone is proposed and evaluated. Motion and motionless postures may be classified using principles of kinematical theory, which underpins hierarchical rule-based algorithms, based on accelerometer and orientation data. Falls detection was implemented by analyzing whether the postures classified as ‘lying’ or ‘sit-tilted’ posture are deemed normal or abnormal, based on the analysis of time, users’ current position and posture transition. Experimental results demonstrate that the approach can detect various types of falls efficiently (i.e., in real-time within a smart phone processor) and also correctly (95 % and 93 % true positives for falls ending with ‘lying’ and ‘sit-tilted’ respectively). The approach is reliable for different subjects and different situations, since it is not only based on empirical thresholds and subject-based training models, but in addition it is underpinned by theory.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
