Outline Sensor manager vs. sensor scheduler Information based - - PDF document

▶

Aug 08, 2023 194 likes •333 views

Information Based Sensor 5/22/2006 Management Information Based Sensor Management Ken Hintz Department of Electrical and Computer Engineering Center of Excellence in C 4 I Control/Tracking # 000000000 Item or Rev # 00000 Copy # 0000

SLIDE 1

Information Based Sensor Management 5/22/2006 1

5/22/2006 Information Based Sensor Manager 1/23

DOCUMENT DATE: 5/22/2006

Information Based Sensor Management

Ken Hintz

Department of Electrical and Computer Engineering Center of Excellence in C4I

Page ___ of ___ Control/Tracking # 000000000 Item or Rev # 00000 Copy # 0000

5/22/2006 Information Based Sensor Manager 2/23

DOCUMENT DATE: 5/22/2006

Outline

Sensor manager vs. sensor scheduler Information based sensor management GMU SMS simulation/visualization demo Summary

SLIDE 2

Information Based Sensor Management 5/22/2006 2

5/22/2006 Information Based Sensor Manager 3/23

DOCUMENT DATE: 5/22/2006

Motivation

Who else is

ut there?

Is anyone else

ut there?

What is the best way to find out? Where should I look next? Is anyone a threat to me?

5/22/2006 Information Based Sensor Manager 4/23

DOCUMENT DATE: 5/22/2006

Sensor Management vs. Scheduling

Manager

Determine which observations sensors should make

in order to best meet mission objectives

Scheduler

Determine the sequence of measurements to make

within the constraints of sensor and platform capabilities

SLIDE 3

Information Based Sensor Management 5/22/2006 3

5/22/2006 Information Based Sensor Manager 5/23

DOCUMENT DATE: 5/22/2006

Need for a Sensor Manager

Cannot sense from all directions at the same time

Sensors are constrained in measurement or

computation space

Tradeoff between accuracy and timeliness of

measurements

Single sensor does not have “big picture”

Need to use integrated world model to direct individual

sensor actions

Context of measurements defines contribution to

mission

5/22/2006 Information Based Sensor Manager 6/23

DOCUMENT DATE: 5/22/2006

Predominant Sensor Management Model

Rule/doctrine/expert system approach

Rules are predicated on experience from last conflict Rules are based on what experts think the next

conflict will be like

Sensors and rules are based on sensor capabilities

and the expected load

Rules are hierarchical and dogmatic

Global optimization is considered computationally intractable

SLIDE 4

Information Based Sensor Management 5/22/2006 4

5/22/2006 Information Based Sensor Manager 7/23

DOCUMENT DATE: 5/22/2006

Information Based Sensor Management

Shannon information

Information rate through a signal-to-noise and

bandwidth limited channel based on coding

Used to develop methods for encoding data with no

regard for content or meaning of data

Information based Sensor Management

Maximizes information gain to minimize valued

uncertainty of platform’s world model by choosing

What to measure, The sensor action to utilize, while Leaving detailed sensor scheduling to individual sensors

5/22/2006 Information Based Sensor Manager 8/23

DOCUMENT DATE: 5/22/2006

Why Information?

Information, not data, is the raison d’etre for sensing Need a common reference system within which to evaluate alternative sensing actions

Many performance measures for sensor systems are

noncommensurate, e.g., Pd, Pkill, Plost_track, etc.

All sensing actions can be formulated as entropy

changes, hence there is a computable information gain that can be associated with each

The amount of information (bits or nats) can be

calculated independently of the sensor type, its characteristics, or which random variable one is interested in observing

SLIDE 5

Information Based Sensor Management 5/22/2006 5

5/22/2006 Information Based Sensor Manager 9/23

DOCUMENT DATE: 5/22/2006

Non-Shannon Information

Five types of sensor-based computable information about a process

Kinematic state Search probability mass function Target Identification Cuer Information Situation

All of these “informations” are based on the fundamental definition of information as being a measure of the change in uncertainty (entropy) about a random variable Note that information is continually being lost about a random process and that information is a dynamic quantity

5/22/2006 Information Based Sensor Manager 10/23

DOCUMENT DATE: 5/22/2006

Types of Sensor “Informations”

Kinematic information How much does a norm of error covariance matrix change as a

result of a sensing action

Search information How much does the probability mass function describing the

location of targets change as a result of a sensing action

Identification Information How much does the probability of a target being of a particular type

change as a result of a sensing action

Cuer Information How much does the detection of a target with one sensor type/mode

change our ability to gain future information utilizing other sensors

r modes

Situation Information How much does our uncertainty about the intention or possible

actions of a target change as a result of a sensing action

SLIDE 6

Information Based Sensor Management 5/22/2006 6

5/22/2006 Information Based Sensor Manager 11/23

DOCUMENT DATE: 5/22/2006

Situation Information

Purpose of a sensing system is to gather information in

rder to infer the intent of processes in its environment

Situation assessment is crucial to the sensor management paradigm since it allows us to decide what information we need A modification of influence diagrams allows us to compute the expected gain in situation information which will result if we select one of several possible informations to request

5/22/2006 Information Based Sensor Manager 12/23

DOCUMENT DATE: 5/22/2006

SIEV Network

Modify influence diagram to produce situation information expected value (SIEV) network Influence diagrams assign value to a decision based on expected value of utility function, essentially, expected monetary value Need probabilistic method of predicting the amount of situation information which can be gained by taking one of several decisions to acquire specific information independent of how we decide to get that information Interpret chance nodes in non- standard manner Compute utility based on topmost goals weighting of information gain associated

nly with situation chance

nodes

SLIDE 7

Information Based Sensor Management 5/22/2006 7

5/22/2006 Information Based Sensor Manager 13/23

DOCUMENT DATE: 5/22/2006

SIEV Net Nodes

Chance nodes are subdivided into

Non-managed nodes Sources of probabilistic data over which we have no control,

e.g., air order of battle, electronic order of battle, are we being attacked?, etc.

Situation nodes Hypotheses about our situation, e.g., hostile/friendly, target

identification, target kinematics, etc.

Change in probability and/or error covariance of these nodes is

situation information gain

Dynamically instantiated when target detected (Sensor Manager) Managed nodes Parameters associated with random variables whose values

can be improved as a result of a sensor action

5/22/2006 Information Based Sensor Manager 14/23

DOCUMENT DATE: 5/22/2006

SIEV Net Nodes

Decision Node Mutually exclusive decisions each of which equates to the

launching of an information request

Utility node Computes the gain in situation information associated with

each possible information request

Based on changes in only the situation evidence nodes

The information request is launched which produces the maximum SIEV

SLIDE 8

Information Based Sensor Management 5/22/2006 8

5/22/2006 Information Based Sensor Manager 15/23

DOCUMENT DATE: 5/22/2006

Annotated SIEV Net

non-managed evidence nodes situation chance nodes managed evidence nodes Decision Node Utility Node Topmost Goal Values From GL

5/22/2006 Information Based Sensor Manager 16/23

DOCUMENT DATE: 5/22/2006

Information Is Not Enough

We know how to

Compute the amount of sensor information Compute the amount of situation information

Need to incorporate dynamic mission goals into

ur valuation of a potential sensing action as

expressed in a Goal Lattice

Topmost values related to situation information Bottommost values related to value of sensing actions

SLIDE 9

Information Based Sensor Management 5/22/2006 9

5/22/2006 Information Based Sensor Manager 17/23

DOCUMENT DATE: 5/22/2006

Mission-Based GL

Dynamic Goal Classes

Topmost Goals

Protect Self
Protect Friendlies
Conserve Power
Penetrate

Defended Area

Collaborate

5/22/2006 Information Based Sensor Manager 18/23

DOCUMENT DATE: 5/22/2006

Valued Information

For determining which situation information to request

Weight situation evidence nodes in the SIEV Net by topmost

values of mission goal lattice

Produces mission-valued situation information expected value

(SIEV)

For determining which particular sensing action to request to fulfill the information request

Evaluate applicable sensing functions as the product of the

amount of information, the bottommost goal values, and the probability of obtaining that information to yield expected sensor information value (EIV)

Use EIV to determine which of several possible sensing actions

to instantiate

SLIDE 10

Information Based Sensor Management 5/22/2006 10

5/22/2006 Information Based Sensor Manager 19/23

DOCUMENT DATE: 5/22/2006

Fused tracks

Real World

Off-board info req

Off-board Platforms Comms Mgr

SA DB Track Fusion Pilot GL GUI Interface

Observation Req

Information Instantiator

Platform Sensor Manager

Goal Lattice SIEV Net Sensors

GMU SMS Real-time Simulation Structure

5/22/2006 Information Based Sensor Manager 20/23

DOCUMENT DATE: 5/22/2006

Simulation

Primarily implemented in Matlab

Multiple independent modules interacting via TCP/IP

interconnects

Runs in real time Physical environment

6-DOF target dynamics of ownship and multiple, simultaneous

targets using open source FlightGear

DEM contour data, 30 meter resolution, IAD DCA Overhead photo “draped” over contour, 15 meter resolution

Ownship has multiple, heterogeneous managed sensors

SLIDE 11

Information Based Sensor Management 5/22/2006 11

5/22/2006 Information Based Sensor Manager 21/23

DOCUMENT DATE: 5/22/2006

PVI Demo over IAD @ 35 k ft

< 1000 ft > 1000 ft DCA

5/22/2006 Information Based Sensor Manager 22/23

DOCUMENT DATE: 5/22/2006

PVI Movie Demo

19flightgears_indio51.avi

52 MB compressed video Low resolution version of actual Matlab visualization demo

which runs in real-time at 30 Hz frame rate

Pilot’s visualization interface (PVI) and all sensor manager software running on one computer

Full 120 Hz flight dynamics of ownship plus 18

aircraft

Flight paths flown by autopilot utilizing same flight plan, DCAIAD DCA etc.

SLIDE 12

Information Based Sensor Management 5/22/2006 12

5/22/2006 Information Based Sensor Manager 23/23

DOCUMENT DATE: 5/22/2006

Summary

Developed new approach to managing sensors

Based on Maximizing sensor information Maximizing situation information Closed loop, indirect control via dynamic mission goals Readily extendable to collaborative environments Common sensor manager for all types of platforms

Developing real-time, distributed, 3D simulation

f realistic environment to evaluate information