An Ontology-Based Approach for IoT Data Processing Using Semantic Rules

摘要

Internet of Things (IoT) applications rely on a network of heterogeneous devices including sensors and gateways. These devices are endowed with the capacity to continuously sense the environment and collect data, which can be further transfered through gateway devices to the cloud. The generated data by IoT systems is often massive. Therefore, the communication gateways might become a bottleneck affecting the system performance due to their resources constraints. This is further exacerbated in the case of bandwidth limitation. The huge amount of data generated increases also the cost associated with data storage and processing at the cloud level. Edge computing, which is a recent IoT trend can contribute to addressing these issues by delegating data processing task to the edges (e.g. gateway devices). In this paper, we propose an approach, which aims at supporting the data processing and minimizes the size of the transferred data to the cloud side. To this end, our approach is based on the notion of rules used to filter the collected data. In order to support the principle of sharing and reusing the rules and the domain knowledge, we propose a Platform Independent Model (PIM) to specify this knowledge independently from the used platform (gateway node). In particular, we define a rule meta-model to support the creation of the model that captures the domain rules. Furthermore, we use Web semantic techniques to represent the knowledge at the semantic level. This representation facilitates the instantiation of these rules and domain knowledge to obtain the Platform Specific Model (PSM) at the gateway level to process and filter the data.