Development of an artificial intelligence algorithm for the analysis of wheelchair movements
Abstract
Monitoring wheelchair user movement is an essential task for assessing a wheelchair user’s mobility and helping them maintain an active lifestyle. Research has shown that increased mobility leads to healthier overall lifestyles, and that people with disabilities are at an increased risk for sedentary lifestyles and the health problems associated with that lifestyle, including cardiovascular disease, obesity, and the development of pressure ulcers (WHO, 2014). Existing technology for analyzing wheelchair user mobility data requires the use of external sensors that must be purchased and maintained (Warms & Belza, 2004). To improve the ease by which mobility data is maintained and analyzed, a wheelchair user can utilize existing technology, such as smart mobile devices, to gather and analyze motion data. This study will focus on the development of a recurrent neural network (RNN) that is trained using wheelchair user data collected from smart devices attached to the wheelchair or wheelchair user. The benefit of collecting data this way is that it does not require the use of additional sensors or equipment, as most wheelchair users will already have access to a smart device capable of collecting movement data. The study found that it was feasible to meaningfully analyze data gathered from a smart device using an RNN. The raw data is analyzed with the RNN to gather information about the mobility of a wheelchair user. The final analysis includes the total time spent moving, number of bouts of movement, and the longest bout of movement. This resulting data could be used by a wheelchair user or healthcare professional to help assess healthy lifestyle habits.
Collections
- UCO - Graduate Theses [699]