Comparative Analysis of Gait Speed Estimation Using Wideband and Narrowband Radars, Thermal Camera, and Motion Tracking Suit Technologies

Abstract

Research has shown that cognitive and physical functioning of older adults can be reflected in indicators such as walking speed. While changes in cognition, mobility, or health cause changes in gait speed, often gradual variations in walking speed go undetected until severe problems arise. Discrete clinical assessments during clinical consultations often fail to detect changes in day-today walking speeds and do not reflect walking speeds in everyday environments, where most of the mobility issues happen. In this paper, we compare four walking speed measurement technologies to a GAITRite mat (gold standard): (1) an ultra wideband radar (covering the band from 3.3 GHz to 10 GHz), (2) a narrow band 24-GHz radar (with a bandwidth of 250 MHz), (3) a perception Neuron Motion Tracking suit, and (4) a thermal camera. Data were collected in parallel using all sensors at the same time for 10 healthy adults for normal and slow walking paces. A comparison of the sensors indicates better performance at lower gait speeds, with offsets (when compared to GAITRite) between 0.1 and 20% for the ultra wideband radar, 1.9 and 17% for the narrowband radar, 0.1 and 38% for the thermal camera, and 1.7 and 38% for the suit. This paper supports the potential of unobtrusive radar-based sensors and thermal camera technologies for ambient autonomous gait speed monitoring for contextual, privacy-preserving monitoring of participants in the community.

Review—Emerging Portable Technologies for Gait Analysis in Neurological Disorders

“…We see that there is a vicious circle to be escaped where as long as the technologies for simple and reliable gait analysis are not yet mature, the search for disease-specific biomarkers will be held up yielding skepticism toward the usefulness of these techniques. One future direction is the integration of several and novel sensor modalities (Buckley et al, 2019;Espay et al, 2019;Morita et al, 2020). Multiple sensors can provide redundant information and their fusion might reduce uncertainty, which can increase reliability in case of a sensor failure.…”

Understanding User Behavior Through the Use of Unsupervised Anomaly Detection: Proof of Concept Using Internet of Things Smart Home Thermostat Data for Improving Public Health Surveillance

“…Moreover, excessive time spent inside the house can represent the severity of detachment from the natural world and sunlight, which could have detrimental impacts such as respiratory problems or seasonal affective disorder [ 38 ]. This approach has the potential, after further validation through larger studies, to provide a nonobtrusive surveillance tool to assist public health officials and governments in policy development by reducing the public health care cost and improving the quality of services in the event of an emergency by measuring indoor behavior [ 13 - 17 , 39 , 40 ].…”

Artificial Intelligence in Medical Sensors for Clinical Decisions

“…The subject sits approximately 1–3 m away from a desk where the remote sensors are connected to a PC and/or a TV. The monitor displays vital signs data in real-time and can also monitor gait and detect falls. , …”