Vincent Huang

Ontology paper: The SSN ontology of the W3C semantic sensor network incubator group

The W3C Semantic Sensor Network Incubator group (the SSN-XG) produced an OWL 2 ontology to describe sensors and observations - the SSN ontology, available at http://purl.oclc.org/NET/ssnx/ssn. The SSN ontology can describe sensors in terms of capabilities, measurement processes, observations and deployments. This article describes the SSN ontology. It further gives an example and describes the use of the ontology in recent research projects.

Semantic Sensor Information Description and Processing

Wireless sensor networks (WSN) generate large volumes of raw data which possess natural heterogeneity. WSNs are normally application specific with no sharing or reusability of sensor data among applications. In order for applications and services to be developed independently of particular WSNs, sensor data need to be enriched with semantic information. In this paper, we propose a semantic Web architecture for sensor networks (SWASN). This information oriented architecture allows the sensor data to be understood and processed in a meaningful way by a variety of applications with different purposes. We develop ontologies for sensor data and use the Jena API for processing which includes querying and inference over sensor data. By studying a building fire emergency scenario, we show that semantic Web technologies can provide high level information extraction and inference of sensor data.

Modeling of sensor data and context for the Real World Internet

The Internet is expanding to reach the real world, integrating the physical world into the digital world in what is called the Real World Internet (RWI). Sensor and actuator networks deployed all over the Internet will play the role of collecting sensor data and context information from the physical world and integrating it into the future RWI. In this paper we present the SENSEI architecture approach for the RWI; a layered architecture composed of one or several context frameworks on top of a sensor framework, which allows the collection of sensor data as well as context information from the real world. We focus our discussion on how the modeling of information is done for different levels (sensor and context data), present a multi-layered information model, its representation and the mapping between its layers.

A resource model for the real world internet

Integrating Wireless Sensor & Actuator Networks (WS&AN) on a large scale allows horizontal applications to access real-world information in real-time and changing the state of the real world, thus providing the basis for the Real World Internet. In this paper we present a resource model that semantically describes sensor, actuator and processing resources for the Real World Internet, this model is designed to efficiently support information queries and execute actuation requests in a large scale system.

Privacy Preserving Data Publishing for Recommender System

Driven by mutual benefits, exchange and publication of data among various parties is an inevitable trend. However, released data often contains sensitive user information thus direct publication violates individual privacy. Among many privacy models, k-anonymity framework is popular and well-studied, it protects information by constructing groups of anonymous records such that each record in the table released is covered by no fewer than k-1 other records. In this paper, we first investigate different privacy preserving technologies and then focus on achieving k-anonymity for large scale and sparse databases, especially recommender systems. We present a general process for anonymization of large scale database. A preprocessing phase strategically extracts preference matrix from original data by Singular Value Decomposition (SVD) and eliminates the high dimensionality and sparsity problem. We developed a new clustering based k-anonymity heuristic named Bisecting K-Gather (BKG) and it is proven to be efficient and accurate. To support customized user privacy assignments, we also proposed a new concept called customized k-anonymity along with a corresponding algorithm (BOKG). We use MovieLens database to assess our algorithms. The results show that we can efficiently release anonymized data without compromising the utility of data.

Relation discovery of mobile network alarms with sequential pattern mining

In telecommunication network systems, there are a large number of interconnected components which also contain many subcomponents. Heavy rain, thunder or other factors can cause mal-function of the components or disconnections between the components which trigger alarms. Because of the interconnection of elements, triggered alarms may propagate to other components. This creates harsh challenges to network operators when it comes to root cause analysis. We address this issue by proposing a method on utilizing network alarms for automatic relation discovery between network nodes. By understanding how network elements or network problems are related to each other, a network operator can easily correlate the alarm events and treat clustered groups of alarms instead of specific events. In this study, we use the temporal and spatial aspects of alarm events to cluster network elements. Our results demonstrate that by analyzing the network alarms, a relationship graph showing the connections between different network elements and network problems can be automatically generated. Such relationship graphs can help network operators mining node dependencies and discovering insights within their network.

Sensor Information Decay Process Modeling

Interest in Context Awareness is gaining ground in both Academic and Industrial communities. Most context information is provided by sensor networks. However, there is always a delay between the collection of the data and the usage of the information. In this paper, we developed a generic decay model for sensor information. The model contains an information-specific function and several parameters. In the evaluation, we identified the function and parameters for user location information and the model describes the decay process very well. With this model, the applications can be provided with the quality of information which is a necessary requirement of usage of sensor information.

A Semantic Web Based System for Context Metadata Management

With the increasing usage of embedded systems and sensors in our surroundings, a new type of information systems – context aware systems – are gaining importance. These user-centric systems acquire context information which describes the state of the user and the user environment, and offer adaptable and personalized services based on the user context information. The central part of a context aware system is the context model used for describing user context information. The context information originates from a multitude of heterogeneous sources, such as personal calendars, sensors attached to the users or to the user’s environment and Web based sources, such as social networking sites. The information from these sources is typically on different abstraction levels and is organized according to different data models. This work proposes a Semantic Web based context metadata management system. The first part of the work develops an ontology model for user context. The user context model integrates information from multiple and heterogeneous sources which are modeled largely by reusing existing well accepted ontologies. The second part of the work proposes a method to infer and reason about additional user context information based on the available context information using rules and ontologies. We instantiate and evaluate the proposed system by performing a social networking case study called meetFriends. In this application, information is collected from various sources such as sensors attached to the users and public web sources - YellowPages. The moods of the users are inferred from a set of rules. A meeting between two users can be set up based on the moods, locations and preferences of the users. The results indicate that Semantic Web technologies are well suited for integrating various data sources, processing of user context information, and enabling adaptable and personalized services.

System and interfaces for water quality monitoring and control in aquaculture

In this paper, we describe an environment monitoring and control system for fish farming applications. This system can receive sensor data in proprietary format and map it to a standard data model. All sensor data and control commands can be accessed via well defined HTTP RESTful interfaces. With this interface, historical data and statistics can be easily accessed and application development time can be greatly shortened.

Mapping the Physical World into the Virtual World: A Com2monSense Approach

Networks of the future will contain a vast array of sensors and actuators which users will wish to interact with. Orchestrating this task effectively and efficiently will be one of the major challenges facing the deployment of these future networks. We propose a three-tier architecture which provides a middleware to resolve high-level service requests from applications to low-level sensor or actuator actions. Our architecture is designed around the core principle of horizontalization, that is, breaking apart vertically integrated silos into their component parts, thus allowing flexible recombination and reuse of data & services.