Falls are a massive public health problem, which results in many injuries worldwide and is of particular concern for vulnerable populations like older adults. The literature shows that falls on icy surfaces during the winter play a major role. We hypothesize the use of new types of slip-resistant footwear can dramatically reduce the risk of slips and falls. There have been some important recent advances in footwear evaluation methods such as the development of the Maximum Achievable Angle test, in which participants walk up and down progressively steeper icy slopes until they begin to slip. However, this and all existing methods are limited to the laboratory and it remains unclear whether better performance in the lab-based test reflects better performance in the real-world.
Therefore the objective of this project is to develop and evaluate a wearable slip detection system that can be used in real outdoor environments. The first version of this system will include an array of sensors including the body’s specific forces, angular rate, high-frequency vibration, and audio recording. Machine learning approaches will be used to train our detection system using combinations of the data collected from different sensors in a simulated winter environment where a Vicon motion capture system will be used to collect ground truth data. The trained system will then be evaluated with participants walking in real outdoor winter environments to determine the proportion of slips that are detected compared to the number self-reported by participants.
The results of this work will be used to focus the recommendations given to the public for selecting winter footwear to reduce the risks of slipping on icy surfaces.