Smart Sensing and Computing for Dementia Care
Primary Investigator (PI) Name
Zongxing Xie
Department
CCSE – Computer Science
Abstract
This project investigates the use of radar sensing and signal processing to monitor gait characteristics in elderly individuals with dementia. The central goal is to develop a contactless, automated system capable of detecting changes in walking patterns that may reflect early signs of cognitive or physical decline. Our work builds on emerging research in radar-based gait monitoring, which offers an alternative to traditional wearable devices that can be inconvenient or impractical for older adults.
Using Texas Instruments radar modules, we collect raw radar data from individuals walking in indoor environments. A series of signal processing techniques are then applied to isolate the motion of the subject from background clutter, including range Fast Fourier Transforms, static clutter removal, and advanced beamforming algorithms. Detected points are clustered and tracked over time using density-based algorithms to consistently locate the subject’s torso. From this tracked motion, we extract key gait features such as step timing, step length, and torso velocity.
To improve clinical relevance, we are currently developing a post-processing stage that will automatically isolate the stable walking phase—excluding periods of acceleration and deceleration that could distort analysis. This is intended to provide a more accurate representation of habitual walking behavior, which is critical in evaluating gait consistency and health changes over time.
We are also working toward a fully contactless, low-cost radar system capable of detecting detailed gait parameters in real-world environments without requiring users to wear sensors or follow scripted procedures. Once complete, the system is intended to support long-term, passive monitoring in care facilities or homes and may offer early detection of mobility decline or cognitive issues. This research contributes to the development of practical, technology-based tools for enhancing dementia care and promoting aging in place.
Smart Sensing and Computing for Dementia Care
This project investigates the use of radar sensing and signal processing to monitor gait characteristics in elderly individuals with dementia. The central goal is to develop a contactless, automated system capable of detecting changes in walking patterns that may reflect early signs of cognitive or physical decline. Our work builds on emerging research in radar-based gait monitoring, which offers an alternative to traditional wearable devices that can be inconvenient or impractical for older adults.
Using Texas Instruments radar modules, we collect raw radar data from individuals walking in indoor environments. A series of signal processing techniques are then applied to isolate the motion of the subject from background clutter, including range Fast Fourier Transforms, static clutter removal, and advanced beamforming algorithms. Detected points are clustered and tracked over time using density-based algorithms to consistently locate the subject’s torso. From this tracked motion, we extract key gait features such as step timing, step length, and torso velocity.
To improve clinical relevance, we are currently developing a post-processing stage that will automatically isolate the stable walking phase—excluding periods of acceleration and deceleration that could distort analysis. This is intended to provide a more accurate representation of habitual walking behavior, which is critical in evaluating gait consistency and health changes over time.
We are also working toward a fully contactless, low-cost radar system capable of detecting detailed gait parameters in real-world environments without requiring users to wear sensors or follow scripted procedures. Once complete, the system is intended to support long-term, passive monitoring in care facilities or homes and may offer early detection of mobility decline or cognitive issues. This research contributes to the development of practical, technology-based tools for enhancing dementia care and promoting aging in place.