Radio-frequency (RF) sensing for deep awareness of human physical status - part II

Abstract

With our society becoming more reliant on technology, it is undeniable that we have seen more areas of our lives being improved with the aid of technologies. Especially in the area of healthcare and lifestyle, more health-related devices such as health-trackers, oximeters and health related applications have appeared exponentially in the consumer market.

Chronic respiratory diseases (CDRs) are one of the many main disorders that is growing in numbers in many developed countries. Threatening their health and welfare of their population, hence there is a need to create a system for preventing this category of diseases. Therefore, with the fame of deeplearning methods in this decade, more solutions are being found and implemented using these methods.

However, in this field of early detection of CDRs, there may be many similar systems, which utilise other form of sensors such as Infrared and Sound, but many of these systems do not integrate with deeplearning methods and are solely use for monitoring only.

For this project, a RF sensor is used to monitor the compression and expansion of the patient’s lungs during respiration. This data collected would be processed and implemented to a system to conduct machine learning methods. Then, through continuous tuning and adjustments, the best fit and values of the parameters for this neural network would be made known. Finally, additional methods and improvements would be discussed in the last part of the report to further optimise and made this system implementable for use in real life.