Goal-Based Semantic Queries for Dynamic Processes in the Internet of Things

Abstract

Internet of Things-aware process execution imposes new requirements on process modeling that are outside the scope of current modeling languages. Internet of Things (IoT) devices may vanish, appear or stay unknown during process execution, which renders process resource allocation at design time infeasible. Devices’ capabilities are often only available in a particular real-world context at runtime. This is not considered by current approaches that use services for encapsulating device functionality. We propose a novel approach to enable both service discovery and invocation for IoT-aware processes based on users’ goals that are defined as part of a process. We apply the Tropos goal modeling methodology to represent the dependencies between these goals and IoT device capabilities. Furthermore, we present a Semantic Access Layer (SAL) to transform these goals into service invocations using generated SPARQL queries. The SAL executes the queries on a knowledge base representing runtime domain knowledge about IoT services, their capabilities, and context. As a result, it invokes the identified IoT services and transfers the responses back to the process engine. The evaluation of our approach within several Smart Home scenarios shows an increase of flexibility and separation of concerns for scalable, IoT-aware process execution.