Conversational AI agents for Elder Care
MIT Media Lab
Conversational AI agents for Emotional Wellness and Medication Adherence
The emotional wellness of individuals emphasizes the importance of a positive outlook towards life circumstances, the capability to cope with stress and ability to maintain fulfilling relationships with others. However, the digital intervention to improve the emotional wellness of older adults is still an underexplored area [1].
In this project, we develop a system that integrates Fitbit data with a conversational agent to support elderly's emotional wellness. This framework is to help caregivers of elderly and our aim is to enhance the user experience by providing quantitative emotional wellness-related data (e.g., sleep, mood, stress, etc). Fitbit data could help users better reflect on their current behaviors and the effectiveness of the intervention [2].
Research Topics
Fitbit
In our study, we use Fitbit Charge5 as an interface to collect user data and use Python implemented Fitbit API to utilize and feed the data into the conversational model (chatbot). From creating Fitbit account to API authorization, please refer to here.
Once the set-up process is done, we can now refer to here to utilize all the sensor data gathered from user activity and below is an example user sleep data printed in the terminal:
Conversational model and Dataset
TBA
References
[1] Warraich, M. U., Rauf, I., & Sell, A. (2018). Co-creation Model to Design Wearables for Emotional Wellness of Elderly. In Bled eConference (p. 6).
[2] https://github.com/wmerians/FitBit-Data-Collection-and-Visualization-System