L2M NSERC - UWB Radar-Based Indoor Human Event Monitoring System

点击翻译按钮获取中文摘要

Owing to the wide spreading of the highly contagious coronavirus (COVID-19), the whole world has suffered a shortage of healthcare resources in the last three years. However, the majority of infected patients are found to have light or moderate symptoms and thus are suggested to rest and recover at home. As reported by the Public Health Agency of Canada, 97.6% of COVID-19 deaths in Quebec occurred in the age group of 60 years and older. Moreover, a significant number of COVID-19 deaths at home have been reported in resting state (sleep) or fall down situations, caused mainly by a sudden deterioration of breathing along with other related symptoms. On the other hand, monitoring of human activities concerns several issues, including the feasibility and reliability of the technology used, the cost of sensing devices and the privacy of the subject. Currently, video monitoring and wearable sensor have been developed for smart sensing. Although straightforward, they have two limitations: privacy issues and the presence of blind spots to the camera. Meanwhile, wearable sensors are usually not user-friendly and are easily forgotten by elders. Developing an intelligent indoor surveillance system that can accurately detect abnormal events like falls as well as vital status in a way that protects people's privacy and allows easy access to data is a potential challenge. Motivated by the above observation, in this project, we will develop a non-contact indoor sensing system, exploiting the impulse radio ultra-wide-band (IR-UWB) radar and microphone array technology for human locations, behaviors (especially fall down and cough) and vital signs. The technical objectives of this project are to (1) augment existing sensing systems by developing and optimizing vital sign (breathing, heart rate, and heart sound) detection algorithms; (2) innovate recognition algorithms based on UWB radar and microphone array, which are capable of real-time identification of events and activities (e.g. walking, falling, eating, drinking, sleeping, coughing) as well as early prediction of disorders (such as Alzheimer's, Parkinson's, or COVID); and (3) develop a compact ambient sensor capable of detecting moving object through the cloud without calibration and risk of safety.