Richard Mietz

SPITFIRE: toward a semantic web of things

The developed world is awash with sensors. However, they are typically locked into unimodal closed systems. To unleash their full potential, access to sensors should be opened such that their data and services can be integrated with data and services available in other information systems, facilitating novel applications and services that are based on the state of the real world. We describe our vision and architecture of a Semantic Web of Things: a service infrastructure that makes the deployment and use of semantic applications involving Internet-connected sensors almost as easy as building, searching, and reading a web page today.

Semantic Models for Scalable Search in the Internet of Things

The Internet of Things is anticipated to connect billions of embedded devices equipped with sensors to perceive their surroundings. Thereby, the state of the real world will be available online and in real-time and can be combined with other data and services in the Internet to realize novel applications such as Smart Cities, Smart Grids, or Smart Healthcare. This requires an open representation of sensor data and scalable search over data from diverse sources including sensors. In this paper we show how the Semantic Web technologies RDF (an open semantic data format) and SPARQL (a query language for RDF-encoded data) can be used to address those challenges. In particular, we describe how prediction models can be employed for scalable sensor search, how these prediction models can be encoded as RDF, and how the models can be queried by means of SPARQL.

Middleware for smart gateways connecting sensornets to the internet

There is an increasing trend to integrate sensor networks into the Internet, eventually resulting in an Internet of Things. Recent efforts of porting IPv6 to sensor networks turn sensor nodes into equitable Internet peers and RESTful Web Services on sensor nodes allow a distribution of the application logic among sensor nodes and more powerful Internet nodes. The touching point between a sensor network and the Internet is the gateway which translates between the link-layer protocols used in the Internet (Ethernet, Wi-Fi) and sensor networks (IEEE 802.15.4). So far, the functionality of those gateways was fixed and simple. We propose to turn these gateways into smart gateways by enabling them to execute application code. As only the gateway has full knowledge of and control over both the sensor network and the Internet, smart gateways can act as performance-enhancing proxies and intelligent caches to preserve the limited resources of the sensor network. Also, the smart gateway can perform application-specific protocol conversion between highly optimized but non-standard protocols in the sensor network and standardized, but less efficient protocols in the Internet. In this paper we present the design of a middleware for smart gateways that allows the execution of application code on the gateway by offering simplified interfaces to the sensor network and the Internet. We also report preliminary performance results for key functions of the middleware.

A P2P Semantic Query Framework for the Internet of Things

The Internet of Things (IoT) will connect billions of embedded computers that can sense and influence their environment. By integrating perception and control of the real world with data and services available on the Web, a wide range of novel applications can be realized, including Smart Cities, Smart Homes, or Smart Grids. A prerequisite for integrating sensor data with other data on the Web is a common data format that is not constrained to a specific domain, such that joint queries over diverse data sources can be efficiently performed. The SemanticWeb offers such a data format called RDF,which essentially consists of subject-predicate-object triples to formulate arbitrary facts, as well as a query language called SPARQL to pose queries over setsof such triples. Inorder to scale to thehugeamount of sensor data being produced in the IoT, RDF databases and SPARQL query engines need to be implemented in a distributed fashion, in particular using peer-to-peer (P2P) techniques. Existing solutions in that space offer only limited functionality andcannotbeeasily extended. Therefore, after surveying the state of the art, we propose a generic framework that fullysupportsSPARQL,butallowsplugging indifferentP2P systemsanddistribution strategies.Wealso presentandevaluateanovelprobabilisticdistributionstrategy that supports non-uniformly distributed RDF triples. *Dipl.-Inf. Richard Mietz: E-Mail: mietz@iti.uni-luebeck.de PDDr. rer.nat. habil SvenGroppe: E-Mail: groppe@ifis.uni-luebeck.de Oliver Kleine, M.Sc.: E-Mail: kleine@itm.uni-luebeck.de Daniel Bimschas,M.Sc.: E-Mail: bimschas@itm.uni-luebeck.de Prof. Dr. Stefan Fischer: E-Mail: fischer@itm.uni-luebeck.de Prof. Dr. Kay Römer: E-Mail: roemer@iti.uni-luebeck.de PD Dr.-Ing. habil. Dennis Pfisterer: E-Mail: pfisterer@itm.uni-luebeck.de

Self-Description and Protocol Conversion for a Web of Things

Many enterprise applications can benefit from the use of wireless sensors, for example for monitoring location and state of goods. While modern backend IT systems use elaborate Service Oriented Architectures based on Web Services, the resource constraints of sensor nodes require highly optimized embedded programming and proprietary networking protocols. This paper aims at bridging this gap by extending Web Service Oriented Architectures to wireless sensors in a standard compliant way while not exceeding the constrained resources of sensor nodes - thus contributing to the formation of a Web of Things. In particular, we enable self-description of embedded Web services through efficient compression of WSDL documents, we support discovery of embedded Web services by means of an efficient embedded discovery protocol, and we enable backend IT systems to invoke embedded Web services on sensor nodes and vice versa by providing an open framework for protocol conversion. We implement our proposal to obtain key performance figures.

High-level states with CoAP: Giving meaning to raw sensor values to support IoT applications

The number of sensors pervading our everyday life, e.g., in smartphones, cars, and buildings, is constantly increasing. These sensors, which are typically embedded into resource constrained devices such as sensor nodes or smartphones, allow measuring the state of the entities they observe or are attached to. If this information is accessible via the Internet, they contribute to the Internet of Things (IoT), where real-world objects have virtual representations. The standardization process to have a complete and efficient communication protocol stack for these constrained devices is in full swing. The Constrained Application Protocol (CoAP) on the application layer allows retrieving data from devices (e.g., metadata) and its sensors (e.g., sensor measurements). This information can be used in a variety of new real-time real-world applications. However, sometimes it is sufficient and desirable not to communicate raw sensor readings but abstractions, i.e., high-level states of the observed entities. Furthermore, as resource constrained devices will be accessible by everyone on the Internet, mechanism to reduce energy consumption play a key role. This paper presents a new option for CoAP which contributes to these two requirements as it allows the creation of high-level states from raw sensor readings. We show that the option can reduce the number of messages when observing a sensor resource which can substantially decrease energy consumption and thus increase device lifetime.

Work in Progress: Resourse-Aware Fault Localization in Large Sensor Networks

Sensor networks are exposed to hostile environments that may cause failures of single nodes and communication links which affect the whole network. Localizing the cause of the problem in space and time requires to collect diagnostic data from the network. Due to resource and energy constraints, however, it is not possible to continuously collect detailed diagnostic data from all nodes. We therefore propose an incremental approach where first data is logged to flash memory and later the user can pose a sequence of diagnostic queries with decreasing scope and increasing level of detail to pinpoint the cause of the problem.

Exploiting correlations for efficient content-based sensor search

Billions of sensor (e.g., in mobile phones or tablet pcs) will be connected to a future Internet of Things (IoT), offering online access to the current state of the real world. A fundamental service in the IoT is search for places and objects with a certain state (e.g., empty parking spots or quiet restaurants). We address the underlying problem of efficient search for sensors reading a given current state - exploiting the fact that the output of many sensors is highly correlated. We learn the correlation structure from past sensor data and model it as a Bayesian Network (BN). The BN allows to estimate the probability that a sensor currently outputs the sought state without knowing its current output. We show that this approach can substantially reduce remote sensor readouts.

Poster Abstract: iBAST—Instantaneous Bridge Assessment Based on Sensor Network Technology

This poster presents the iBAST project, the goal of which is the continuous monitoring of highway bridges with sensor network technology. Currently working on a first individual solution, the long term goal of the project, though, is to move from individual case studies towards a comprehensive monitoring of bridges, by developing a monitoring system which can be easily and cost-effectively customized for every bridge. This poster abstract presents the system architecture and implementation as well as first results.

LoCaF: Detecting Real-World States with Lousy Wireless Cameras

The Internet of Things (IoT) integrates wireless sensors to provide online and real-time access to the state of things and places. However, many interesting real-world states are difficult to detect with traditional scalar sensors. Tiny wireless camera sensor nodes are an interesting alternative as a single camera can observe a large area in great detail. However, low image resolution, poor image quality, and low frame rates as well as varying lighting conditions in outdoor scenarios make the detection of real-world states using these lousy cameras a challenging problem. In this paper we introduce a framework that addresses this problem by providing an end-to-end solution that includes energy-efficient image capture, image enhancement to mitigate low picture quality, object detection with low frame rates, inference of high-level states, and publishing of these states on the IoT. The framework can be flexibly configured by end-users without programming skills and supports a variety of different applications.