Combining wearable physiological and inertial sensors with indoor user localization network to enhance activity recognition

Thanks to the pervasiveness of smart technologies, researchers could aggregate data and investigate user's activities thus to deliver personalized home-care services. Activity recognition system have been widely developed, however some challenges still need to be addressed. This paper presents a system where information on body movement, vital signs and user indoor location are aggregated to improve the activity recognition. The system was tested in a realistic environment with a total of 3279 instances acquired from ten healthy users. These results encouraged the use of this approach in activity recognition applications, indeed, the overall accuracy values are satisfactory increased.