Activity of Daily Living-aware Healthcare for Elderly in Pervasive Environment
Date Issued
2015
Date
2015
Author(s)
Chen, Ya-Hung
Abstract
The high development of medicine causes the world’s population aging quickly. To resolve the problem with limited medical resources, constant monitoring of elders’ activity of daily living is important. We propose an activity recognition system for smart home, so elders can live alone and their children can monitor their parents’ living activity to achieve the concept of “Aging in Place”. The model proposed for monitor-ing the living activities is highly effective to recognize meaningful activities by using both ambient and wearable sensors. Given that the mentioned model is a non-parametric learning model, it should be feasible to deploy the resuting monitoring system in our real living environment. It turns out that it takes elders limited effort to do activity labeling in the part of model training, and later on the model may even have chance to find some special activities that have been overlooked by the elders. Besides, the proposed activity recognition system can discover new activity which does not appear in the training stage. Since case-based reasoning can find the most similar known activity and provide the same service to the user, we use mechanism to achieve immediate service provision and adaptive learning once a new activity is encountered. If the elderly user falls into a serious abnormal situation, the system could, say, notify the caregiver immediately by the case-based reasoning mechanism, followed by the system’s request for confirmation of this new activity from the user and then re-training of the online model. Subsequently, if the activity occurs again, the system can provide service without asking confirmation as before. To validate our proposed recognition system, we have invited several users to test the system, and the average precision of activity recognition is up to 97.67%, which is promising for real deploy-ment of this ADL-aware healthcare system in the future.
Subjects
Agin in Place
Activity Recognition
Adaptive Learning Model
Internet of Things
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-104-R02922073-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):4a7b1198ae07b085accf6a8a1d4dbfd8