This approach will strictly distinguish between impersonalized knowledge during context generation and personalized knowledge despite for anomalous behavior discovery.Changes in behavior concerning ADLs can be seen as an early indicator of autonomy loss [6]. One of the earliest assessments of human behavior related to the ability to live independently was conducted by Katz et al. [7]. A significant drawback of questionnaires such as those used in that study is that elderly people tend to lie about their difficulties (due either to fear of the consequences or to shame).Later, Lawton [8] constructed a hierarchical taxonomy of behavioral competence. He presents five major categories, from simple to complex: health, functional health, cognition, time use and social behavior.
The hierarchy demonstrates that the complex tasks rely heavily on the simple ones. He states that residential behavior is closely related to cognitive competence.Telemonitoring products concentrate on well-being at the health [8] level (e.g., 24-hour ECG monitoring). We tackle the automatic detection of behavioral competence at the, functional health level. Basically, we focus on the physical (sometimes called basic) and instrumental aspects of ADLs. So far, we have defined the following ADLs: going to the toilet, transferring, waking/sleeping, dressing, eating, washing, bathing, combing and napping.To a certain degree, people reveal their current context (activity, location, identity and time) [9] by interacting with their environment (e.g., by opening the doors of a cupboard).
Assuming additional semantic information is provided, it is possible to draw inferences about any activity currently being performed simply by augmenting their surroundings with the appropriate sensors.When considering an individual’s daily habits, some repetitive patterns can easily be observed, starting with the person’s wake-up routine [10]. After collecting enough context information about a specific person to create a temporal relationship model of his/her daily activities, we can assess to what extent the current day’s pattern of activities matches those of previously observed days.The system described in this article is intended as an aid to the caregiver. Therefore, one highly influential concern in its design is the awareness that in practice the caregiver will have very limited time to interact with an assessment tool.
To this end, we have created a novel human autonomy assessment system (HAAS) Brefeldin_A at different levels: during the specification phase, a clear and easy-to-understand language had to be chosen so the developer can profit from the domain knowledge of the caregiver. When it comes to visualization, we had to find a highly condensed behavioral information representation sellectchem of the client’s daily living skills.We use the term client to refer to the person the caregiver looks after.