Enhanced fall prevention: A real-time hybrid analysis with smart walking stick & Edge-based IoMT

Human fall poses a significant risk for the elderly. A fall can result in hospitalization or, tragically, death. The primary causes of falls among the elderly are often linked to a loss of balance or insufficient limb support. While fall prevention is difficult, fall risk analysis can help predict and prevent future falls. The present work proposes a real-time 3-way hybrid fall risk factor analysis methodology, , employing a smart walking stick. It uses an Edge-based IoMT (Internet of Medical Things) architecture that is extendable to the Cloud. The smart walking stick has a 10 kg load cell (YZC-133) with HX711 (mounted on the grip) and a MPU 6050 kinematic sensor paired with an ESP8266 WiFi Micro Controller (MCU) to transfer the accelerometer, gyroscope and load cell data to the processing unit. A Raspberry Pi-based edge device evaluates pressure (support) on the grip and walking patterns using an accelerometer, gyroscope, and load sensor signals connected with ESP 8266. The system could perform fine grain fall analysis by classifying the subjects into the risk categories of High/Medium/Low/None. The system used the clinically established parameters and tests for its multimodal analysis at low cost based on Timed Up and Go (TUG), Force test and Gait analysis modules. The individual results from each module were then combined to predict the final risk category. The long-term analysis is done on a cloud server and the system could predict falls with a maximum of 92% accuracy.