SMART HEALTH FOR ASSISTING LIVING
Description of the DVC THEME
Assisted living in smart homes can change the way millions of elderly people live, manage their conditions, and maintain their well-being. This could support the aging population to live longer independently and to enjoy comfort and quality of life in its private environments.
While current monitoring and assistive technologies are selectively deployed due to high cost, limited functionality, and interoperability issues, future smart homes could leverage cheap ubiquitous sensors, and interconnected smart objects, packaged with robust context interference and interaction techniques. The next generation of smart home technologies will be adaptive to fit versatile living environments, and interoperable for heterogeneous applications. In addition, a service-oriented end-to-end system architecture will support reconfiguration and modular design that is essential to empower care providers to customize solutions.
A monitoring system within a smart home checks the daily activities and decides whether the behavior of the residents is regular or irregular. The success of such a system depends on understanding the normal lifestyle and the degree to which the behavior of the elderly has deviated from what is defined as normal.
The main challenge in the problem of smart assisted living is to create an unobtrusive sensor network, using microphones and motion sensors that could handle computationally expensive algorithms, while keeping all the sensitive data on the device.
Sub-challenges composing this experiment:
This DVC THEME is composed of four main challenges:
Expected global results:
To create a data value chain that allows:
- To improve everyday life of elderly by providing a low-cost end-to end behavioural monitoring solution.
- To provide high accuracy on user’s profile monitoring through an optimized on-device running system
- To increase quality of behavioural assessment leveraging data from multiple sources, thus covering data loss and inaccuracy related risks
- To reduce visits to healthcare providers/hospitals could minimize risks of exposure to COVID-19
- To develop an accurate algorithm for recognizing the correct usage of an inhaler.
- To achieve an F1-Score higher than 95% of the four classes (actuation, button press, inhale and exhale).