TLDR: A new framework uses commercial sensors and personal interviews to create tailored models of older adults’ daily routines in their homes. It then formally verifies if observed activities match personalized expectations, identifying deviations that could signal health changes or safety concerns, thereby enhancing independent living.
As our global population ages, there’s a growing need to support older adults who wish to live independently in their own homes. Ensuring their safety, addressing care needs, and maintaining a good quality of life are paramount. This often involves understanding their daily routines, known as Activities of Daily Living (ADLs), which are essential self-care tasks like showering, eating, or taking medication. Unexpected changes in these routines can sometimes signal a decline in health or cognitive abilities, potentially leading to unsafe situations.
A new research paper introduces a groundbreaking framework designed to monitor these vital activities in a personalized and non-intrusive way. The framework integrates data from commercially available sensors installed in the home with rich contextual information gathered directly from the older adults themselves. This includes insights from semi-structured interviews about their daily habits, preferences, home layouts, and even sociological observations.
How Does This Framework Work?
The core of this innovative system lies in its ability to create highly personalized models for each participant. Here’s a simplified breakdown:
-
Smart Sensors at Home: The system uses unobtrusive sensors, such as contact sensors on doors, windows, and appliances (like a refrigerator or shower door), and motion sensors in living spaces. These sensors capture everyday events, like opening a cupboard or moving through a corridor. This data is collected and stored over time.
-
Understanding Personal Routines: Before sensors are even installed, researchers conduct in-depth interviews with participants. These conversations help to understand individual daily routines, specific needs, and even privacy preferences. For example, some participants might be comfortable with contact sensors in their bedroom but not motion sensors. This qualitative data is crucial for tailoring the monitoring to each person.
-
Building a Digital Model: The collected sensor data and personal insights are then used to build a ‘formal model’ of the participant’s home environment and their observed behavior. Think of it as a digital representation of their daily life within their home.
-
Defining Personalized Expectations: Based on the interviews, specific ‘properties’ or rules are defined for each individual. These properties are essentially formal statements about what their expected behavior should be. For instance, a property might state: ‘Participant A moves from the bedroom to the bathroom to take a shower in the morning,’ or ‘Participant B accesses their medication dispenser at a certain time.’
-
Verifying Behavior: A specialized tool, called a ‘model checker’ (NuSMV), then compares the observed behavior from the sensors against these personalized properties. If the observed behavior deviates from the expected property, the system generates a ‘counterexample’ – a sequence of events that shows exactly when and why the deviation occurred. This helps caregivers or family members understand what happened.
For example, in one case study, the system detected that a participant, who usually took a shower after waking up, returned to the corridor briefly before re-entering the bathroom and closing the shower door. While the activity was completed, this minor deviation was flagged, suggesting that the system could be refined to accommodate small variations in routines. In another instance, a participant’s medication intake was flagged as a violation because it occurred later than their defined time window, highlighting a potential health safety risk.
Also Read:
- Keeping AI Models Reliable: A New Approach to Monitoring Performance During Adaptation
- Advanced AI Framework Accelerates Rare Disease Diagnosis
The Impact
This framework represents a significant step forward in supporting older adults to age in place safely and independently. By combining real-world sensor data with deeply personalized contextual information, it offers a robust way to monitor daily activities, identify unexpected behaviors, and potentially facilitate timely interventions. The approach is also designed to be generalizable, meaning it can be adapted to a diverse range of individuals and their unique living situations.
For more detailed information, you can refer to the full research paper: A Personalised Formal Verification Framework for Monitoring Activities of Daily Living of Older Adults Living Independently in Their Homes.
