Vital Sign and Sleep Monitoring Using Millimeter Wave

Abstract

Continuous monitoring of human’s breathing and heart rates is useful in maintaining better health and early detection of many health issues. Designing a technique that can enable contactless and ubiquitous vital sign monitoring is a challenging research problem. This article presents mmVital, a system that uses 60GHz millimeter wave (mmWave) signals for vital sign monitoring. We show that the mmWave signals can be directed to human’s body and the Received Signal Strength (RSS) of the reflections can be analyzed for accurate estimation of breathing and heart rates. We show how the directional beams of mmWave can be used to monitor multiple humans in an indoor space concurrently. mmVital also provides sleep monitoring with sleeping posture identification and detection of central apnea and hypopnea events. It relies on a novel human finding procedure where a human can be located within a room by reflection loss-based object/human classification. We evaluate mmVital using a 60GHz testbed in home and office environment and show that it provides the mean estimation error of 0.43 breaths per minute (Bpm; breathing rate) and 2.15 beats per minute (bpm; heart rate). Also, it can locate the human subject with 98.4% accuracy within 100ms of dwell time on reflection. We also demonstrate that mmVital is effective in monitoring multiple people in parallel and even behind a wall.

Sleep Stage Classification Using Bidirectional LSTM in Wearable Multi-sensor Systems

“…Rather than relying on EEG signals from PSG, recent researchers focus on using cardiorespiratory signals because they can also indicate the different sleep stages and can be captured in plenty of the low-cost simplified devices. This approach is easily adapted into the various scenarios of mobile and Internet of Things (IoT) healthcare, sleep health monitoring in smart home [2] and other accommodations [3]. As the deep learning methods are rapidly developing in recent years, some latest researches have demonstrated the effectiveness that applying Recurrent Neural Network (RNN) to capture the implicit patterns of the time-series cardiorespiratory signals [4], [5].…”

Millimeter Wave Sensing: A Review of Application Pipelines and Building Blocks

“…Suitable antenna types include, but are not limited to, horn antenna [64], [65], [68], [74], [102], [109], [116]- [119], [142]- [155], yagi antenna [149], lens antenna [59], [105], reflectarray antenna [85], [86], cassegrain antenna [88], [113], (on-chip integrated) patch antenna [49], [54], [82], [127], [128], [139], [151], [156], [157], and parabolic antenna [110], [112]. Suitable antenna designs include, but are not limited to, phased array antenna design [144], [145], [147], [158]- [162], frequency scanned antenna design [87], fan and pencil beam antenna design [119], gaussian beam antenna design [100], and omni-directional antenna design [146].…”

“…During this time, receivers make a variable measurement at one angle of arrival if the angle is not in line with the current beam angle. This process is repeated sequentially for a number of angles of arrival to form a measurement matrix [143], [144], [147], [148]. The authors in [160] use a robot that moves and rotates.…”

On the feasibility of estimating soluble sugar content using millimeter-wave

“…However, the reconfigurable or programmable options of these devices are not available to the public, and the measurements for RSS and raw wireless signals are not allowed. Recently, several 60 GHz testbeds [39,53,56,57] were built for developing better networking and sensing techniques. Since our objective in this work is to understand the feasibility of Brix estimation using 60 GHz reflections, we develop our own 60 GHz mmWave software radio testbed which acts as a bi-static (separate transmitter and receiver devices) probing system.…”

“…Beyond object recognition, 60 GHz is a promising technology for health monitoring. A vital sign monitoring system was designed in [57] to monitor breathing rates, heart rates, and sleep using the RSS of 60 GHz signals. Authors in [45] designed a blood glucose estimation system which exploits the signal penetration property of 60 GHz signals to measure the penetration on the thin skin between thumb and pointer finger.…”