Characterizing Web of Things Interactions with Existential Reasoning

Tracking #: 2075-3288

This paper is currently under review
Victor Charpenay
Sebastian Käbisch
Harald Kosch

Responsible editor: 
Guest Editors Sensors Observations 2018

Submission type: 
Full Paper
The Web of Things (WoT) is a collection of interlinked Web resources exposed by autonomous sensors and actuators that interact to perform complex automation tasks. This paper presents a method to characterize interactions on the Web of Things in terms of relations between these devices and the physical world entities that compose their environment. In particular, based on the recent standardization by the W3C of the Thing Description (TD) model and its alignment with other Web ontologies, devices expose logical assertions on themselves that form a knowledge graph from which a `graph of interactions' can be derived. The reasoning task of interest in WoT is query answering over ontologies that feature existential restrictions on the `things' WoT devices observe or act upon. Because no complete algorithm exists for this task, we present a tractable skolemization algorithm for the ELP fragment of Description Logics (DLs), at the intersection of EL++ and Datalog. We tested our approach on two use cases in different industry domains: Building Automation (BA) and Industrial Control Systems (ICS).
Full PDF Version: 
Under Review