Microwave Radar for Health Risk Assessment

Microwave Radar Sensing for Motion Analysis in Health Risk Assessment (Cognitive Impairment, Fall risk, Apathy, etc.)

Microwave and millimeter-wave radar can measure daily human motions such as walking, standing up, and sitting down without physical contact. Doppler radar and micro-Doppler radar provide velocity information generated by the motion of the legs, arms, and trunk. These signals can be used to estimate gait speed, gait cycle, stride-related parameters, trunk motion, motion speed during sit-to-stand and stand-to-sit transitions, and posture-related movement features.

These daily motions contain information related to muscle strength, balance function, gait function, cognitive function, motivation, and activities of daily living. Therefore, radar-based non-contact measurement is a promising sensing approach for assessing fall risk, frailty, cognitive decline, mild cognitive impairment, dementia-related risk, apathy, and risks related to activities of daily living.

Compared with camera-based measurement, radar is advantageous for privacy-preserving monitoring because it does not capture visual images and is less affected by lighting conditions. It also does not require the subject to wear sensors, making it suitable for continuous health monitoring in homes, nursing-care facilities, hospitals, and community health studies.

Background

In aging societies, technologies for early detection of physical and cognitive decline are important for preventing falls, loss of independence, and decline in daily living function. Gait speed, stride parameters, sit-to-stand motion, the Timed Up and Go test, and the Sit-to-Stand test have been widely used in physical therapy, gerontology, and exercise epidemiology as indicators of physical function and health risks.

However, continuous assessment of these functions in daily environments is not easy. Cameras, motion capture systems, force plates, and wearable sensors are useful, but they have limitations related to installation conditions, privacy, sensor attachment burden, and measurement location.

Radar-based non-contact sensing can observe natural walking and sit-to-stand motions with relatively low burden on the subject. Micro-Doppler radar signals contain velocity changes generated by leg and trunk movements during walking, as well as by standing up from and sitting down on a chair. By analyzing these signals, it may be possible to perform gait and physical-function assessment in a more natural and unobtrusive manner.

Research Objectives

This research aims to estimate features related to physical function, cognitive function, and daily living function from velocity information obtained using Doppler radar and micro-Doppler radar during walking, standing-up, and sitting-down motions.

Main topics include:

Estimation of gait speed, gait cycle, and stride-related parameters
Extraction of gait features such as trunk motion, asymmetry, and leg velocity
Estimation of motion duration, velocity changes, and posture transitions during sit-to-stand and stand-to-sit motions
Analysis of features related to fall risk, frailty, and physical-function decline
Analysis of gait features related to cognitive function, MCI, and dementia-related risk
Analysis of relationships between motion features, apathy, motivation, and activities of daily living
Health risk assessment using radar signals and machine learning
Privacy-preserving healthcare monitoring in environments where cameras are difficult to use

Recent Work

Our previous work has addressed Doppler radar-based gait measurement, extraction of gait features from micro-Doppler signatures, analysis of relationships between gait features and cognitive functions in elderly adults, MCI screening, non-contact measurement of sit-to-stand and stand-to-sit motions, and analysis of motion features related to apathy and risks in activities of daily living.

In particular, our 2019 IEEE Access paper analyzed associations between gait features measured using a 24-GHz micro-Doppler radar and multiple cognitive-function domains in elderly adults. This study is positioned as one of the early representative works connecting short-range radar signal processing with gait-based cognitive-function assessment in gerontology, physical therapy, and exercise epidemiology.

Subsequent work has investigated Doppler radar-based MCI screening, health risk assessment using sit-to-stand and daily motions, and estimation of physical and daily living functions by combining radar signal processing and machine learning. These studies aim to assess age-related health risks from subtle motion features in daily activities, rather than simply measuring walking speed.

This research has been supported by projects including MIC SCOPE, AMED research programs on longevity science, JSPS KAKENHI, and the Nakajima Foundation. We also conduct collaborative research with AIST toward practical healthcare applications.

Research Perspective

A key feature of this research is the connection between radar signal processing and healthcare-related fields such as physical therapy, gerontology, exercise epidemiology, and healthcare engineering. From the radar engineering perspective, Doppler velocity, time-velocity distributions, micro-Doppler features, signal classification, and feature extraction are important. From the healthcare perspective, gait speed, standing ability, balance, frailty, fall risk, cognitive function, apathy, and activities of daily living are important. This research aims to bridge these areas by interpreting physical velocity information obtained from radar as indicators related to physical function, cognitive function, and daily living function.

Selected Publications

Jiale Ren, Hengyi Li, Aihui Wang, Kenshi Saho, and Lin Meng, "Radar-based gait analysis by Transformer-liked network for dementia diagnosis," Biomedical Signal Processing and Control (ELSEVIER), vol. 91, 105986, May 2024.
Kenshi Saho, Keitaro Shioiri, Shoma Kudo, and Masahiro Fujimoto, "Estimation of Gait Parameters from Trunk Movement Measured by Doppler Radar," IEEE Journal of Electromagnetics, RF, and Microwaves in Medicine and Biology, vol. 6, pp. 461-469, December 2022.
Kenshi Saho, Kouki Sugano, Momoka Kita, Kazuki Uemura, and Michito Matsumoto,"Classification of Health Literacy and Cognitive Impairments using Higher-Order Kinematic Parameters of the Sit-to-Stand Movement from a Monostatic Doppler Radar," IEEE Sensors Journal, vol. 21, pp. 10183-10192, April 2021.
Kenshi Saho, Keitaro Shioiri, Masahiro Fujimoto, and Yoshiyuki Kobayashi, "Micro-Doppler Radar Gait Measurement to Detect Age- and Fall Risk-Related Differences in Gait: A Simulation Study on Comparison of Deep Learning and Gait Parameter-Based Approaches," IEEE Access, vol. 9, pp. 18518-18526, February 2021.
Kenshi Saho, Kazuki Uemura, Kouki Sugano, and Michito Matsumoto, "Using Micro-Doppler Radar to Measure Gait Features Associated with Cognitive Functions in Elderly Adults," IEEE Access, vol. 7, pp. 24122-24131 , March 2019.
Summary: This study presented experimental data connecting radar sensing with health risk assessment at the boundary of gerontology, physical therapy, and exercise epidemiology. It clarified associations between radar-based gait parameters and multiple cognitive-function domains.

Keywords

micro-Doppler radar, Doppler radar, millimeter-wave radar, microwave radar, gait analysis, gait measurement, sit-to-stand, stand-to-sit, Sit-to-Stand test, Timed Up and Go, physical function assessment, cognitive function, MCI, dementia-related risk, fall-risk assessment, frailty, apathy, activities of daily living, ADL, IADL, non-contact sensing, human sensing, healthcare monitoring, exercise epidemiology, physical therapy, gerontology, radar signal processing

Google Sites

Report abuse