Recent research from Binghamton University, State University of New York, focuses on enhancing response times through a human action recognition (HAR) algorithm. This algorithm uses local computing devices to analyse sensor data and detect unusual movements without needing external processing centres. Developed by Professor Yu Chen and PhD student Han Sun from the Thomas J. Watson College of Engineering and Applied Science’s Department of Electrical and Computer Engineering, the Rapid Response Elderly Safety Monitoring (RESAM) system leverages advancements in edge computing.
Published in the IEEE Transactions on Neural Systems and Rehabilitation Engineering, their study demonstrates that RESAM achieves 99% accuracy with a response time of just 1.22 seconds when deployed on standard devices like smartphones, smartwatches, laptops, or desktop computers. This places RESAM among the most precise methods available today for detecting falls among older adults.
Chen emphasises the significance of this research for an often-overlooked demographic: senior citizens who may lack the resources to voice their technological needs. He contrasts typical high-tech innovations with the practical benefits of RESAM, which utilises familiar devices instead of requiring a complete “smart home” setup, empowering older adults without adopting new technology.
To address privacy concerns, RESAM employs a privacy-preserving approach by reducing monitored images to skeletal representations. This technique enables the system to analyse critical body points such as arms, legs, and torso to detect falls or other incidents that could lead to injuries while respecting users’ privacy preferences.
Recognising that bathrooms pose a high risk for falls but are sensitive areas for surveillance, Chen highlights the reluctance towards camera-based monitoring in such private spaces. Instead, he envisions RESAM as a foundational component of a broader initiative termed “Happy Home,” which integrates thermal or infrared cameras and additional sensors to remotely assess various aspects of a person’s environment and well-being.
Looking ahead, Chen and Associate Professor Shiqi Zhang from the Department of Computer Science explore expanding RESAM to include robotic companions, such as a robot dog. This “pet” could accompany individuals through daily routines, offering personalised monitoring and assistance. Zhang’s earlier work demonstrated how a robot dog could guide visually impaired individuals, underscoring the potential for interactive and adaptive monitoring systems in home environments.
Chen envisions these advancements not only as technical innovations but as supportive companions that enhance the safety and independence of older adults. Integrating intelligent sensors and robotic assistants into daily life, the “Happy Home” concept aims to proactively monitor health indicators and predict potential issues before they escalate, promoting a safer and more supportive living environment for older adults.
More information: Han Sun et al, A Rapid Response System for Elderly Safety Monitoring Using Progressive Hierarchical Action Recognition, IEEE Transactions on Neural Systems and Rehabilitation Engineering. DOI: 10.1109/TNSRE.2024.3409197
Journal information: IEEE Transactions on Neural Systems and Rehabilitation Engineering Provided by Binghamton University
