With the aging population in many parts of the world, low-cost, easy-to-deploy health-care technologies, such as vital sign monitoring (e.g., respiration) and abnormal respiration detection, have attracted increasing attention. To overcome the limitations of traditional respiration rate monitors, various respiration monitoring systems have been proposed recently. Among these techniques, the RFID-based technique called AutoTag utilizes attached cheap RFID tags to monitor human breathing. Furthermore, considering the challenge and cost of collecting labeled training data from patients with breathing problems, a recurrent variational autoencoder, which is an unsupervised learning approach, has been adopted in the system for abnormal breathing detection. We present the design of the AutoTag system, its prototype implementation, and experimental results in this chapter. It provides a low-cost, effective solution to contact-free vital sign monitoring.
