Environmental sensors have continuously improved by becoming smaller, cheaper, and more intelligent over the past years. As consequence of these technological advancements, sensors are increasingly deployed to monitor our environment. The large variety of available sensor types with often incompatible protocols complicates the integration of sensors into observing systems. The standardized Web service interfaces and data encodings defined within OGC’s Sensor Web Enablement (SWE) framework make sensors available over the Web and hide the heterogeneous sensor protocols from applications. So far, the SWE framework does not describe how to integrate sensors on-the-fly with minimal human intervention. The driver software which enables access to sensors has to be implemented and the measured sensor data has to be manually mapped to the SWE models. In this article we introduce a Sensor Plug & Play infrastructure for the Sensor Web by combining (1) semantic matchmaking functionality, (2) a publish/subscribe mechanism underlying the SensorWeb, as well as (3) a model for the declarative description of sensor interfaces which serves as a generic driver mechanism. We implement and evaluate our approach by applying it to an oil spill scenario. The matchmaking is realized using existing ontologies and reasoning engines and provides a strong case for the semantic integration capabilities provided by Semantic Web research.
展开▼
机译:在过去的几年中,环境传感器变得越来越小,越来越便宜,越来越智能,从而不断地对其进行改进。这些技术进步的结果是,传感器越来越多地用于监视我们的环境。带有不兼容协议的大量可用传感器类型使传感器集成到观测系统中变得复杂。 OGC的Sensor Web Enablement(SWE)框架中定义的标准化Web服务接口和数据编码使传感器可以在Web上使用,并向应用程序隐藏异构传感器协议。到目前为止,SWE框架尚未描述如何在最少的人工干预下实时集成传感器。必须实现能够访问传感器的驱动程序软件,并且必须将测得的传感器数据手动映射到SWE模型。在本文中,我们结合了(1)语义匹配功能,(2)SensorWeb的发布/订阅机制以及(3)传感器的声明性描述模型,介绍了适用于Sensor Web的Sensor即插即用基础结构。接口,用作通用驱动程序机制。我们通过将其应用于漏油场景来实施和评估我们的方法。匹配是使用现有本体和推理引擎实现的,并为语义Web研究提供的语义集成功能提供了有力的依据。
展开▼