SmartRL: A Context-Sensitive, Ontology-Based Rule Language for Assisted Living in Smart Environments
To automate assisted living tasks in smart environments, the contextual and temporal aspects associated with activities of daily life (ADL) can be exploited to (1) detect and act upon inconsistent context, i.e., when an activity occurs outside of its usua
- PDF / 729,581 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 73 Downloads / 142 Views
Abstract. To automate assisted living tasks in smart environments, the contextual and temporal aspects associated with activities of daily life (ADL) can be exploited to (1) detect and act upon inconsistent context, i.e., when an activity occurs outside of its usual context; and (2) guidance through ADL routines, by automatically executing or suggesting a next subtask at the correct context. This paper presents SmartRL, a context-sensitive rule language supporting task automation in smart environments, and applies it to an Assisted Ambient Living (AAL) use case. SmartRL realizes a number of key opportunities in this setting, such as linking the language to a domain ontology, and facilitating the detection and influencing of context; as well as considering the temporal nature of smart environment rules, the need to revert rule effects, and writing activity routines. Keywords: Smart environments Ambient Living (AAL)
Ontology-based Context-aware Assisted
1 Introduction Assisted living deals with the effects of cognitive decline, by assisting cognitively impaired people to perform activities of daily life (ADL). In particular, ambient assisted living (AAL) [1] relies on smart environments to automate assistive tasks; such as executing activities automatically (e.g., setting temperature), guiding people through ADL (e.g., via step-by-step instructions), or issuing alerts in case of unusual activities (e.g., falling, forgetting about cooking). To support task automation, one can exploit temporal and contextual aspects associated with activities; e.g., sleeping normally occurs at night in the patient’s bedroom; whereas cooking typically happens in the kitchen around mealtimes. Similarly, many ADL consist of atomic activities with clear temporal interrelations: e.g., after waking up, the patient needs to wash up in the bathroom for 15–20 min; then, the patient should have breakfast in the kitchen. To represent automated tasks in smart environments, we present a high-level, context-sensitive rule language called SmartRL, and apply it to an AAL use case. By linking SmartRL to a domain-specific ontology, we allow for high-level rule specification, reduce verboseness, and enable easy variation of rule specificity. In SmartRL rules, conditions refer to high-level context (location, activity and time), which is © Springer International Publishing Switzerland 2016 J.J. Alferes et al. (Eds.): RuleML 2016, LNCS 9718, pp. 341–349, 2016. DOI: 10.1007/978-3-319-42019-6_22
342
W. Van Woensel et al.
continuously inferred by a smart-environment middleware; whereas actions invoke smart services to perform tasks, raise alerts or instruct the user. Currently, the SmartRL parser translates (or “expands”) rules into RDF triple pattern-like expressions, which are fed into our rule engine. We note that this paper focuses on presenting the SmartRL rule language, and does not detail other aspects of our system; such as the smart-environment middleware, or its rule engine implementation. Section 2 presents our running AAL scenario. Sectio
Data Loading...
