DSIRNS: Decision Support for Integrated Real …proceedings.ndia.org/JSEM2004/GeoBase/Kreyer.pdfFrank Varcolik -Science Applications International Corporation ... What is DSIRNS? Decision

DSIRNS:Decision Support for Integrated Real-time Network of Sensors

DSIRNS:Decision Support for Integrated Real-time Network of Sensors

Steve Kreyer, Bill Dyer, Eric Tollar, Mark Vandersluis,Frank Varcolik - Science Applications International Corporation

Glenn Waldron - ITspatial, LLCMajor Mark Sward - USAF/DTRA

Steve Kreyer, Bill Dyer, Eric Tollar, Mark Vandersluis,Frank Varcolik - Science Applications International Corporation

Glenn Waldron - ITspatial, LLCMajor Mark Sward - USAF/DTRA

Presented at the 2004 GeoBase Compass ConferencePresented at the 2004 GeoBase Compass Conference

Presentation Outline

What is DSIRNS?Problem being AddressedRole of GeoBaseRole of VisualizationDSIRNS Components and OperationFuture DirectionsConclusion

What is DSIRNS?Decision Support for Integrated Real-time Network of Sensors

Initial implementation of a strategy for Radiological and Nuclear Countermeasures (RNC) using sensors and decision support techniquesProvides improved sensor location and prioritization decisions (Joint Program Guardian Installation Protection Program)Ability to dynamically adjust alarms and direct mobile sensor to potential threatsEasy visualization and integration of all regional sensor and mitigation asset informationProvides alert to silent release scenario

Problem: Improving Radiological/Nuclear Sensor System Technology and Utilization

Typically sensors are used in a stand-alone approachBoth DoD and DHS propose a “layered defense” approach with system of linked sensorsWithout DSIRNS, n independent decisions based on fixed alarms in time and space

Detection Event

Threshold Exceeded

Alarm/ Stop

PASS

Detection Event

Threshold Exceeded

Alarm/ Stop

PASS

Detection Event

Threshold Exceeded

Alarm/ Stop

PASS

Detection Event

Threshold Exceeded

Alarm/ Stop

PASS

n events

DSIRNS seeks to improve this with an Integrated Bayesian Influence Network Approach

Real-time Visualization Value Added

Understanding Coverage/Situational AwarenessCommon Operational Picture among multiple usersGeospatially Accurate Databases with Drilldown CapabilityDynamic Links between 2D and 3D GIS Systems

Attribution Database

Drill-down Capability

Relevant Infrastructure

Layers

Discrete Event

Simulation Output

GeoBase Data was used to Create 3D Model of MacDill AFB

DSIRNS relies on Geospatial Database for OperationData for MacDill AFB was obtained through GeoBase under the DTRA MIDAS ProgramITspatial InterSCOPETM* used as Real-time Common Operational Picture

* ITspatial InterSCOPE is an Interactive Site, Communications, Operations and Planning Environment that combines 3D geospatial visualization with existing 2D GIS and facilities systems and existing analytical models

Decision Support ToolsGeospatial Visualization

Sensors

DSIRNS Analysis System

Data Server

Data FlowSensor

Data

Data Repository

Decision Support Algorithm

Visualization Software

Decision System

Other Systems or Response Teams

1. Sensor Readings

2. Readings + Other Data

3. Decisions

4. Update (Tracking, Analysis)

5. Alter Sensor Thresholds

Value of Advanced Decision SupportBayesian Network Technology

Bayesian Decision Net uses integrated information to categorize vehicles: Inputs

sensor dataalert level informationother available information

Provides decisions on trackingProvides estimates on radiation sourceNetwork can learn automatically from previous experience to generate improved decisions

What does Correlated Data add to the Identification Process?

By correlation of data across sensors :

Tracking (individual detections correlated to a moving vehicle)

Direction(movement towards high value target)

Multi-data fusion(integrate other data sources such as video)

Cumulated data(city-wide radiation patterns)

SAIC Developed Scenario Generator

Used to easily create Case Study SimulationsThree entities types can be placed in the scenario1. Threat Radiator (Bad guy

with RDD)2. Benign Radiator (Good guy

with Radiological/medical)3. RDD sensor (stationary or

mobile)

Entities can move in specified path or remain stationary

DSIRNS Operation 1. Combine sensors

portalsareamobile

2. Allow multiple thresholdsstoptrack

3. Integrated decision supportsituational sensor informationintegrate sensors to track vehicles

4. 2D and 3D Visualization

Situational AwarenessMonitor & Tracking

Bayesian NetworkInfluence Decision Net

Snapshot of a DSIRNS Scenario in NYC:System of Fixed and Mobile Sensors with Moving Real and Benign Threats

KEYGREEN: Fixed or Mobile DetectorRED: Threat Emitter (Mobile)PURPLE: Benign Emitter (Mobile)

Table shows status of sensors

and tracking functions

Potential Expansions

This Initial Implementation is IR&D looking for full-implementation opportunities Implementation strategy allows for easy integration of other sensor types

chemical, biological sensors could be easily integrated

Visualization architecture supports easy integration of other tools and algorithms

integration of products readily added– Consequence Assessment Tool Set (CATS), HPAC, etc. for WMD

effects and consequence management readily incorporated

Conclusion

Integration of additional information beyond single sensors will result in improved RDD detection

Increased DetectionImproved InterdictionDecreased False PositivesImproved Procedures

Integration with visualization assists real-time decision process

Reduces time in assessment of geospatial situationsAbility to focus and drill down to specific locations immediately

Contact Info:

Steve [email protected]

William H. [email protected]

www.SAIC.com