Main content area

LISA: Lightweight context-aware IoT service architecture

Gochhayat, Sarada Prasad, Kaliyar, Pallavi, Conti, Mauro, Tiwari, Prayag, Prasath, V.B.S., Gupta, Deepak, Khanna, Ashish
Journal of cleaner production 2019 v.212 pp. 1345-1356
Internet, decision making, models, processing time, tourists, value added
Internet-of-Things (IoT) promises to provide services to the end users by connecting physical things around them through Internet. The conventional services build for web are primarily based on the pull technology, where the user actively engages with system to get the services. However, in IoT environment, the services are based on push-based, where information and value added services will be pushed towards the user. Unless, these push-based services are properly managed they would overwhelm the user with unnecessary information, thus, it will soon start annoying the user.In this paper, we propose a lightweight context-aware IoT service architecture namely LISA to support IoT push services in an efficient manner. In particular, LISA filter and forward the most important and relevant services to the users by understanding their context. To achieve its goals, LISA formulates a user model to resolve local decision making by using agents and available web services paradigm. The proposed user model describes the user in an abstract way by considering the context and profile information of the user. For evaluation, we simulate LISA by considering an IoT tourist guide system as a use case scenario, and we show the performance of the our user model concerning precision and recall metrics. The results of our preliminary experiments confirm that LISA successfully reduces the information provided to the user by selecting only the most relevant among those. The evaluation shows that LISA can extract services for a user by selecting from 15000 services with precision upto 0.3 and recall upto 0.8, and it can be further optimized by tuning the user-specific design settings. Additionally, our approach shows improvements in query processing time which also includes the query generation time.