ABSTRACT
One of the most important public health problems worldwide is falls, the leading cause of injury-related hospitalizations and a major cause of disability and death among seniors. Conventional fall risk assessments (FRAs) depend heavily on laboratory or clinical assessments and are generally performed under supervision of a geriatrician, nurse, or therapist. While current FRA techniques have demonstrated moderately effective predictions, there remain significant limitations, including the lack of clinically relevant information on a day-to-day basis. With the emergence of ubiquitous computing and sensing technologies, big data is being produced by everyday objects around us. Our bodies are no exception, generating diverse types of data including mobility (e.g., walking speed) that can be recorded and transmitted by sensors and mobile devices for further analysis. We anticipate that technological advances in sensors, machine learning, signal processing, computer vision, and big data will lead to discovery of new solutions for fall prevention and prediction. These emerging technologies have the potential to collect data in everyday environments during free-living activities over long periods to address gaps in knowledge. This chapter examines current FRA practices, highlighting key limitations, and summarizes recent research in emerging sensor systems to improve quality and frequency of quantitative assessments. Afterward, a machine learning–based approach for automatic detection of compensatory balance reactions and environmental context in daily life under development at the University of Waterloo Neural and Rehabilitation Engineering Lab is discussed.
