31 March 2009 MMI OntDev 1

Autonomous Mission Operations for Sensor WebsAl Underbrink, Sentar, Inc.

31 March 2009 MMI OntDev 2

• Using Intelligent Agents to Form a Sensor Web for Autonomous Mission Operations (SWAMO)

– PI: Ken Witt, West Virginia High Technology Consortium Foundation

– Co-I: Al Underbrink, Sentar, Inc.

– Co-I: Dan Mandl, NASA/GSFC

• Objective

– Enable dynamic, composable interoperability of sensor web products and services

Project Overview

31 March 2009 MMI OntDev 3

• Consistent with OGC SWE standards and specifications – Enables compatibility with broad base of sensing systems

– SensorML, SOS, SPS, CWS

• Semantics of the SWE-compatible sensor systems– Physical and logical structure and descriptions

– Platforms, sensing devices, observations, detector performance characteristics, etc.

– Services, processes, workflows, required parameters, process interfaces, etc.

• Semantics of SWAMO control systems – Describes autonomous agents for system-wide resource sharing,

distributed decision making, and autonomic operations

– Intelligent agent capabilities, experiment schedules and tasks, platform workloads, etc.

Semantics for Autonomous Operations

31 March 2009 MMI OntDev 4

Technology Infusion Path

• cFE• GMSEC

• ST-5 • MidSTAR-1

• SensorML• SOS, • SPS, CWS

• MidSTAR-1

• Models• Semantics • Automated Decision Making

31 March 2009 MMI OntDev 5

Levels of Interoperability

Bits and Bytes

Common data and protocol formats for information exchange

Shared meaning

Each system is aware of methods and procedures for other systems

Each system takes into accountstate changes of other systemsover time

Maximum abstraction,implementation independent

31 March 2009 MMI OntDev 6

SensorML

SensorML Concepts– Everything viewed as a

hierarchical “process”– Models physical systems

and logical processes– Each process has common

representation

31 March 2009 MMI OntDev 7

SensorML ProcessesProcess Concepts

– Each process has inputs, outputs, metadata, and parameters (similar to IDEF0)

– Inputs, outputs, and metadata provide lots of flexibility for legacy descriptions

– UCUM may be replaced

31 March 2009 MMI OntDev 8

Sensor PlatformsSystem Platforms

– Spacecraft, aircraft, ground systems, subsurface

– Mobile and fixed location

– Passive or actively controlled sensing

System Components – Actuators, sensors,

detectors, and agents– Physical, functional,

logical capabilities and behaviors

– Performance characteristics

31 March 2009 MMI OntDev 9

Sensor Observations

31 March 2009 MMI OntDev 10

SWAMO Agents

31 March 2009 MMI OntDev 11

SWAMO Agents

Agent Performance Concepts– Capabilities, responsibilities,

and goals– Service registry and discovery – Time, tasks, plans and

schedules – Resource capacities and

availabilities

31 March 2009 MMI OntDev 12

SWAMO Rules

• Actually, none!

• Ontology used as prototyping method – Knowledge representation – Decision rules implemented directly

• More agreeable to flight hardware and software systems (MidSTAR-1, Global Hawk, etc.)

31 March 2009 MMI OntDev 13

Instances: MidSTAR-1

• Representation of MidSTAR-1 instances

– Nano ChemSensor Unit and Variable Emmissivity Film sensors/detectors

– Onboard computer and bus subsystems

– Battery as “actuator” that can be controlled

31 March 2009 MMI OntDev 14

• Supports Semantic and System-Of-Systems Interoperability– Able to consume, use, and produce SWE-standard descriptions

– Compatible with existing and future Sensor Web systems

– Amenable to Sensor Web Services interfaces

• Targets Dynamic and Composable Interoperability – Enables automated reasoning and decision support systems

– Sufficient detail for autonomic functions

– Support for failure recovery and adaptive operation

– Enables capability to dynamically discover, combine, and adapt chains of sensors distributed over multiple sensing platforms

Benefits

31 March 2009 MMI OntDev 15

The Way Ahead

• Delay Tolerant Networking • Global Hawk• Benefit science users • AIST 2008

– SWAMO plus – Deeper ontology

development