Speech-based prompting system to assist with activities of daily living: A feasibility study
Neeraja Murali Dharan MS*, Muhammad Raisul Alam PhD, Alex Mihailidis PhD PEng
Full text PDF 
( Download count: 500)
AbstractBackground: The onset of dementia can negatively affect an individual's ability to perform activities of daily living (ADL). Consequently, these individuals rely on formal or informal caregivers for ADL completion. Cognitive assistive technologies (CATs) can help adults with dementia during ADL performance and potentially alleviate caregiver burden. However, sensors that are commonly used by CATs (i.e., cameras, accelerometers, radio frequency identification tags) have some limitations; they can cause erroneous activity detection if not positioned properly, installing them can be a long and expensive process, and their usage can raise privacy concerns.
Objective: In this study, spoken dialogue was explored as an alternative method of providing input to CATs. The feasibility of using a speech-based CAT to assist with ADL performance was evaluated.
Method: A speech-based intelligent prompting system for ADL assistance was developed. This system was a modification of the Cognitive Orthosis for Assisting with aCtivities in the Home (COACH), a CAT that was designed to assist with the ADL of handwashing. The camera-based hand-tracking sensor in the COACH was replaced with a speech-based conversational agent. A study with 10 healthy adults was conducted to test the feasibility of using speech instead of camera input for the COACH. Results were compared to the outcomes of a previous study with the camera-based COACH.
Results: Results suggested that the speech-based COACH was able to identify completed task steps with 97.5% accuracy and true participant performance with 90% accuracy. In most cases, the performance of the COACH was not negatively affected by the use of speech in place of camera data.
Conclusion: This study is a promising first step in exploring the application of speech in CATs designed for ADL assistance. The results of this paper support the feasibility of using speech as input for CATs and highlight its potential in replacing sensors.Keywords: Dementia, cognitive assistive technology, human-centered computing, conversational interface, activities of daily living (ADL)
Neeraja Murali Dharan MS*, Muhammad Raisul Alam PhD, Alex Mihailidis PhD PEng (2021). Speech-based prompting system to assist with activities of daily living: A feasibility study. Gerontechnology, 20(2), 1-12
https://doi.org/10.4017/gt.2021.20.2.436.06