Optimized algorithm for rfid-based activity recognition of the elderly

Abstract

As the trend toward an increasingly aging global population accelerates, more attention is being focused on healthcare services for the elderly. Traditional healthcare services for the elderly require personal, meticulous care. With increasing numbers of the elderly and the 'inverted pyramid' of family structure, more and more elderly people are often unable to obtain the healthcare they need in their final years. This paper introduces a Densely Convolutional Network for optimizing the algorithm model of an RFID-based activity recognition system, leading to a more accurate analysis of the basic daily behaviors of the elderly.During the experiment, We also optimized the original recognition process from combination of a series object-use recognition to treating multiple object-use processes as a union process.Which improves the efficiency and speed of activity recognition.