Semantic MediaWiki for programming and teaching context-aware Ambient Assisted Living (AAL) agents
|SMWCon Fall 2016|
|Semantic MediaWiki for programming and teaching context-aware Ambient Assisted Living (AAL) agents|
|Description:||This talk presents how Semantic MediaWiki can be used to program a software-agent for Ambient Assisted Living (AAL).|
|Event start:||2016/09/29 15:35:00|
|Event finish:||2016/09/29 16:00:00|
|Keywords:||SWRL, SPARQL, SPIN, AAL|
Ambient Assisted Living (AAL) aims at helping elderly and imapaired people in their daily life because the use of devices in the living environment can mean a barrier for elderly. For this reason, different software-agents provide services for different tasks in order to assist elderly people in an adequate and customized way. These software-agents require to be implemented by application developers and to be integrated on-the-fly into an AAL environment. Mostly, agents interact with IoT devices in order get sensed data and to control devices. One problem is that AAL environments comprise heteregeneous IoT devices which are using different APIs and protocols so that an on-the-fly integration of agents can be problematic for application developers as they have to implement this heterogeneous APIs. In this talk, I want to show by using the example of an software-agent (Sherlock) how this agent is programmed and integrated by Semantic MediaWiki. The Sherlock agent programs are represented by rule-based formal languages (such as SWRL/SPARQL/SPIN) and are embedded and annotated in Semantic MediaWiki. The objective is to simplify the programming process and integration of context-aware AAL agents into an AAL environment in order to allow these agents to assist the user appropriate by controling heterogeneous IoT devices in the living environement. In our approach the IoT devices as well as the AAL environment are described in Semantic MediaWiki to make their capabilities and characteristics available. Furthermore, the rules provide a basis for AAL agents to learn by machine learning techniques, user preferences und to solve the problem of deciding in uncertain context. The objective here is to improve the assistance of an agent by increasing its experience.
The following aspects will be considered in this talk:
- How can developers program via Semantic MediaWiki and Rule based languages an adaptive software-agent?
- How can the rules be used as basis to learn the preferences of the user in order to overcome uncertainty?
- The application of the Sherlock agent.