Imperceptible Sensor Infrastructure for Rule-Based Active Safety Prevention in the Household

Aging of population and increasing complexity of domestic equipment justify the growth of scientific interest in surveillance of independent older adults at home. Our proposal focuses on imperceptibility of a sensing network which, besides detection of complex actions, analyses their irregularity and issues control commands to a safety-oriented active environment for living. The investigated design of mixed residential and wearable sensor network includes domestic equipment-embedded usage sensors and accelerometer-fitted footwear with a presence message broadcast. Synchronously detected events contribute to recognizing complex activities of daily living and to building personalized behavioral patterns. A short- or long-time variation of these patterns may be understood as a deterioration of psychophysiological status and, in context of particular device operation error, modifies the home automation algorithms accordingly. An implementation case study supported by results of volunteers-based experiment confirms the correctness of actions recognition and feasibility of the automatic system for user behavior-dependent safety prevention .