< Terug naar vorige pagina

Publicatie

Daily Living Activities Recognition and Segmentation in Older Adults with Sensor Technology and Machine Learning Algorithms

Boek - Dissertatie

In Europe as well as in China the fraction of older persons continues to rise compared to the fraction of the active population. This growing group of older people requires an individualised and adaptive care plan as their health declines. The level of self-reliance is one of the strongest contributing factors in the care plan. Due to the development of wearable sensors and internet-of-things (IOT) sensors, it becomes possible to continuously monitor health and health related parameters over long periods of time, outside the lab or doctor's office. This in turn enables health care providers to gain insights into the behaviour and lifestyle of healthcare consumers. Identification of the changes of self-reliance of older individuals who are at risk of becoming frail will allow for early adaptation of the care plan before major health problems arise. These preventive measure will allow elderly person to maintain their independence for a longer period of time. Such an identification of elderly at risk for frailty combined with appropriate subsequent evaluation and intervention thus is a cornerstone of geriatric medicine and quality care. The self-reliance is typically measured with a pen-and-paper interRAI screening tool which takes an 1h30min to complete by a professional caretaker. This scale is typically filled in after grave events (falls) or at least 3 months apart. The aim of the PhD project is therefore to monitor self-reliance of older persons based on data from wearables and IOT sensors hence partially replacing the interRAI scale (especially the items related to instrumental activities of daily living). The systems will be able to continuously monitors changes in the self-reliance hence allowing immediate adaptation of the care plan. Key technological challenges of this project are: the heterogeneous nature of the sensor data and the inter-person variability of the data.
Jaar van publicatie:2022
Toegankelijkheid:Open