Oak Ridge National Laboratory (ORNL) and our strategic partners are working together to build a comprehensive incident management system for the near-real-time detection, identification, and assessment of chemical, biological, radiological, nuclear, and explosive (CBRNE) threats. The goal of SensorNet is to bring together and coordinate all necessary knowledge and response information quickly and effectively. This will be done by providing a common data highway for the processing and dissemination of data from CBRNE, meteorological, video and other sensors in order to provide near-real-time information to emergency management decision makers and first responders.

SensorNet is a robust data architecture and infrastructure that supports plug-and-play sensors of various types, archival storage of sensor data, standards-based publication of sensor data, and sensor control services. SensorNet allows for the integration of many dissimilar sensor systems into one system of systems infrastructure. Sensitive sensor data is protected using a secure multi-level access control system. SensorNet provides high reliability of data through the use of self-organizing/self-healing network connectivity, distributed processing, distributed data storage, and backup power supplies. The SensorNet system is based on open standards to maximize flexibility, extensibility, maintainability, and commercial support.

SensorNet Components and Open Standards

The SensorNet system of systems is divided into four main areas of development: the SensorNet node hardware and software, web service interfaces, relational database, and client applications. The SensorNet node is a self-contained processor with software for communicating with the individual sensors and redundant communication capabilities for reporting data to a Regional Data Center (RDC) using standards-based web services. The node software gathers observation and measurement (O&M) data from the sensors using standard interfaces defined by IEEE 1451. The IEEE 1451 standard was developed to provide common plug-and-play interfaces between smart transducers and sensor control networks. SensorNet takes advantage of this existing standard to easily integrate a variety of dissimilar sensors. The node periodically sends O&M data to the RDC to be made available to SensorNet consumers. The node also monitors sensor data in real time and propagates any unusual condition alarms to the RDC such as sensor over/under threshold, out of range, and low battery. Each node has multiple network communication options. If the primary network connection fails, communication is automatically rerouted through an alternate channel. In addition to SensorNet owned nodes, SensorNet is also designed to accept data from external data sources. The combined information is made available from the RDC using a single, common interface.

The SensorNet database and web service servers reside at the Regional Data Center. The web service interfaces use Open GIS Consortium (OGC) XML schemas for publication of sensor data and alerts. OGC is an association developing open, vendor-independent computing standards for geographic information systems (GIS) and other related technologies. OGC specifies the usage of the Geography Markup Language (GML), an XML grammar written in XML Schema, for the encoding of geographic information. Although GML provides specifications for describing a variety of geographical information, the fundamental element of GML is a geographic "feature". According to the OGC description of GML 3.0, a geographic feature is "an abstraction of a real world phenomenon; it is a geographic feature if it is associated with a location relative to the Earth". Many features contain both spacial and temporal information. As such, almost all aspects of SensorNet can be treated as geographic features, including nodes, sensors, and O&M data.

To facilitate the communication of feature data between clients and servers, OGC has developed the Web Feature Service Interface Specification (WFS). The WFS specification describes the request and response documents for web services using HTTP as the distributed computing platform for creating or modifying feature instances and for querying features based on both spatial and non-spatial constraints. SensorNet takes advantage of these open standards and uses a WFS interface for all access to node, sensor, and O&M data. The SensorNet nodes use transactional WFS services to insert and update O&M data at the RDC. The transactional WFS services for processing data is also used by external data sources for making their data available to the SensorNet system. Using open standards makes SensorNet data easily available to a variety of clients without requiring custom development or proprietary software.

SensorNet alert handling is accomplished using services based on the draft OGC Web Notification Service (WNS) specification. This specification is still under development (currently at version 0.1.0) and SensorNet has helped to determine areas where the specification is lacking. OGC has acknowledged that the current WNS specifications do not meet the needs of systems like SensorNet and has begun development of a new standard Web Alerting Service (WAS). SensorNet plans to migrate alert handling to WAS when a draft specification is available. For external alert notifications, SensorNet may migrate to the Common Alerting Protocol (CAP). CAP is a non-OGC open specification designed to collect and relay all types of hazard warnings. Regardless of the communication standard used, applications and clients may register with SensorNet to receive notification of a variety of alert types. When an alert is received, the client application may then request additional information using the WFS services to determine the cause of the alert and take appropriate action.

The relational database model is designed to store as much information as possible about SensorNet nodes, sensors, and data, but still allow sufficient flexibility to incorporate data from both existing and unknown data sources. A hierarchical structure is used to allow very flexible high-level queries as well as lower-level inquiries. The data model hierarchy is headed by a generic "GIS Feature" table that allows all key elements of the database to be referenced as geometric features.

The end goal of SensorNet is to make available the data and information from a nation-wide sensor network to a variety of data consumers (end users such as emergency management operations centers). SensorNet allows for both on-demand information queries and automated responses to sensor alerts. Using ad hoc queries, researchers may use SensorNet as a rich data source for their research needs. SensorNet provides a wealth of information to researchers in a single location using standard interfaces. Other consumers and applications may register for various SensorNet events and alerts to perform their own custom calculations to meet a variety of needs. Many such applications are designed to run automatically when an alert is received. For example, once a CBRNE event is detected by chemical, biological, and/or radiological sensors, combined with weather, a modeling system can, in real time, produce a plume model to chart the geographical extent of the hazard, the downwind hazard and to estimated the number of people exposed and predict immediate and latent effects on the population. SensorNet allows for the inclusion of information from external data sources so that such modeling data may be archived and made available to CBRNE first responders.

A number of applications are being developed that showcase the use of SensorNet information by first responders. HPAC (Hazard Prediction and Assessment Capability) is one such application. The existing HPAC hazard and effect modeling software is being modified to automatically respond to a SensorNet alert message by query SensorNet for the information needed to calculate a plume model and then inserting the calculated data into SensorNet as an external data source. Integrated modeling software such as HPAC provides additional valuable information to emergency personnel in the event of a CBRNE threat. Other applications are being developed on top of Keyhole products to allow real-time and 3D mapping of node and sensor locations as well as viewing of measurement data from individual sensors. These applications are only examples of the many possibilities of SensorNet data usage.

Implementation Technologies and Deployment Plans

SensorNet uses a variety of technologies in its implementation. The SensorNet node operates with a customized Linux distribution and agent software for monitoring sensor data. All web service clients and servers are implemented using Java and XML technologies, including Sun's Java Web Services Developer Pack (JWSDP) and JDOM XML handling. A Java class library is being developed to allow consumer applications written in Java to communicate with SensorNet without having to write custom web services client code or fully understand the subtleties of the OGC schemas. Although this class library will be offered for convenience, the use of open standards makes SensorNet interfaces available to any number of non-Java clients and applications. The Regional Data Center uses an Oracle database.

ORNL has SensorNet testbeds in East Tennessee, Fort Bragg, North Carolina, New York City, and Washington, DC. Partnerships with the National Oceanic and Atmospheric Administration (NOAA), the U.S. Army, and the American Tower Corporation (ATC) allow ORNL to expand SensorNet deployment. NOAA installs measuring stations in urban areas to study various meteorological events, and ORNL is augmenting these deployments with SensorNet nodes and CBRNE sensors. ORNL is providing its partners, the U.S. Army's Communications and Electronics Command and the Stevens Institute, with a SensorNet database archival service for their ferry GPS locator trial in New York Harbor. ORNL has a deployment agreement with ATC that provides the lab access to approximately 10,000 cell phone towers that have power, security, and telecommunications, generally located in areas of significant population. SensorNet has the support of the Fort Bragg Public Safety Business Center, of Fort Bragg, NC, and officials in the Department of Homeland Security for continued funding for research and deployment.