Comprehensive Mine and Sensor Simulation Functional Overview

1
Comprehensive Mine and Sensor SimulationFunctiona
l Overview

• Mid Self
• Deputy Director, Modeling Simulation
• CECOM RDEC
• Night Vision Electronic Sensors Directorate
• mid.self_at_nvl.army.mil
• 703-704-1285

2
Intelligent Munitions SystemConcept
Economy of Force
40 - 50 Km
Intelligent Munitions System
Troop Cdr
Tactical UGS
ARES
15 Km
Precision Strike
UA CDR
AREMS

• Remote deployment
• 15 50 Km
• Rocket, Mortar, Helo, AVN
• Extended communications
• Air and/or ground relays
• Networked, smart engagement strategy

3
Smart UGS Cluster

• 4 multi-mode sensors working as an integrated
  cluster
• 3 non-imaging acoustic/seismic nodes
• Each node as 3 microphone array
• Cluster gateway node with imaging non-imaging
  sensors
• Cluster computes target classification, range and
  line of bearing estimates based on
  acoustic/seismic sensor response
• When target range lt 500m, cluster cues IR sensor
  to LOB and captures an image
• Image and target report are sent to Human in the
  loop controller

Acoustic footprint from ABFA
4
Smart UGS Components
12 cm dia

• Gateway with imaging IR sensors
• 8 lbs
• 1 per M87A1 volcano canister
• Cluster
• 1 Gateway
• 3 Pointers
• Equivalent to 155 mm M718 RAAM payload
• Non-imaging pointer node
• 2.5 lbs
• 3 per M87A1 volcano canister

36 cm
96 cm
Stowed
Deployed
Non-imaging Sensors Comm module Processor Power
source
12 cm
12 cm
5
Generic IMS Model

• Smart UGS Field
• Support the munitions field
• Target recognition, ID, and BDA
• Target location tracking for supporting IDF
• Feeds the FCS C4ISR
• C2 FCS Battle Command System
• Universal Controller
• Intelligent Munitions
• Wide Area Top Attack Munition (WATAM)-AT/AV
• Smart Munition (SM)-AT/AV/AP

Comm Relay
200 m
SUGS Cluster Coverage Area
500 m
200 m
Integrated suite of sensors, C2 system munitions
6
Intelligent Munitions Systems

• Wide area top attack munitions
• 3 microphone acoustic array
• Seismic sensor
• Target classification
• Range LOB estimation
• Engage when closest point of approach lt 100 m
• Smart AT Munitions
• Single microphone acoustic array
• Seismic sensor
• Target classification
• Range LOB estimation
• Engage when target range lt 5 m

7
UGS/IMS Control

• All FCS vehicles have common C2 system
• UGS/IMS share common sensor architecture and
  communications
• UGS/IMS controller is a SW module that runs on
  the common C2 system
• UGS/IMS gateway module communicates via standard
  FCS combat net radio
• LOS range approx 8 km
• UGS/IMS communicate using a standard message and
  data structure
• Sensor Interface and Access Management System
• Any controller can initialize and assume control
  of an UGS/IMS cluster
• Control can be passed from one controller to
  another

8
UGS Reporting to the Network

• UGS cluster generates target reports each time
  they detect a target
• Gateway node in the field filters these reports
  before sending them to a controller to prevent
  the field from constantly chattering (and to
  reduce simulation network load)
• Gateway node uses three criteria
• Target reports are re-transmitted after a
  configurable timeout (default timeout is 1minute)
• Reports are transmitted once the target moves a
  configurable distance (default distance is 500m)
• Report is sent immediately if the acquisition
  level is upgraded

9
UGS Controller Model

• Human in the loop controller receives the target
  spot reports sent by UGS clusters
• Controller maintains a database of targets
  reports
• When a target spot report is received, the report
  is fused into the target database using the
  following algorithm
• Quad-tree lookup is performed for existing
  targets using a configurable region of interest
• ROI is scaled according to reported velocity, so
  that fast-moving targets are fused properly
• For targets close enough to report location,
  target types are compared
• If existing target with compatible type is within
  query region, the targets are considered to be
  same
• Existing or previous targets location and
  velocity are updated
• If the spot report provides more detailed
  information than the database had on the target,
  then the target type and acquisition level are
  upgraded
• If no target with a compatible type is within the
  query region, or no targets are within the query
  region, a new target is generated
• Targets, which are not updated for a configurable
  amount of time, have their acquisition certainty
  downgraded
• Once the certainty reaches zero, the target is
  removed form the database

10
Comprehensive Mine and Sensor Server (CMS2)

• Redesign of CERDEC NVESD/ ARDEC FSAC
  Comprehensive Mine Simulator
• Operates as a server to primary force-on-force
  simulation engine (OneSAF Testbed, JCATS, etc)
• Allows large scale simulation of mines or
  distributed sensors with minimal network burden
• Scaleable to the simulation environment and task
• Models may run on multiple processors or machines
• Low to high fidelity
• Physics-based sensor models
• NVESD Acquire IR search and target acquisition
  model
• ARL Acoustic Battlefield Aid
• ERDC CRREL Seismic Rule-of-Thumb model
• System models
• Composable UGS model (can vary sensor
  configurations)
• Conventional and smart AT and AP
• Dispersion patterns for a variety of deployment
  mechanisms

11
CMS2 Data Flow Architecture

• OTB
• 1 Publishes environment variables
• 1 Publishes target state data (truth)
• 1 Publishes deployment event for creation of
  UGS/IMS field
• Calculates target damage state
• 1 Sends target damage to network

• CMS2
• 1 Instantiates UGS/IMS field
• 1 Publishes location status of individual
  mines/UGS
• Computes in-field go/no-go message completion
  delay
• UGS/IMS go dormant until interaction with target
  entity
• 1 Monitors target locations/states
• Sends detonation event to SAF
• OTB Target entity enters UGS/IMS ROI
• Calculates Pd/Pc/Pr
• Calculates estimated target range velocity
• Calculates target track
• 2 Sends/updates target spot report from field
• 2 Sends detonation event to controller

1
2
2 / 3
2 / 3
1 Ground truth 2 Perceived truth w/out comm
effects 3 Perceived truth with comm effects
2 / 3
3

• CES
• Simulates tactical network comms effects

• MITL Controller
• Monitors field activity
• Sends commands to field (arm / detonate /
  neutralize / etc)

12
CMS2 Architecture
CMS2 Federation
CMS2 Core
IMS Platform (System) Model
Infrastructure / Terrain Library / Map GUI / RTI
Interface / etc
Command Control Logic
Sensor Fusion Model
Target Tracking Model
Other as required
Attack Logic
Sub-system Models
Contractor Provided Models or Data
Subsystem Model Data or Model Services
Contractor Provided Models or Data
Image Server
Comms
Munitions Model/Data
Acoustic Model/Data
Seismic Model/Data
Other as required
13
0th Order Multi-mode Sensor
Acoustic/Seismic/Magnetic Ground Sensor Model
Look up tables derived from higher fidelity model
Radial Ranges (km)
P
0.95 _at_ 0.6 km
P
0.85 _at_ 1 km
det
det
Probability of Detection (Pdet)
Heavy
Light
Heavy
Light
Tracked
Tracked
Wheeled
Wheeled
2.00
0.70
1.00
0.50
2.00
0.70
1.00
0.50
PDet0.70
PDet0.85
1.00
0.35
0.50
0.25
1.00
0.35
0.50
0.25
PDet0.95
0.60
0.25
0.35
0.15
0.60
0.25
0.35
0.15
Probability of
Probability of
Classification
Given
Classification
Given
(P
)
(P
)
Detection
Detection
Class

Det
Class

Det
PclassDet0.85
gt1.00
gt0.35
gt0.50
gt0.25
P
0.70 _at_ 2 km
det
UGS Parameters for Heavy Tracked Vehicle
1.00
0.35
0.50
0.25
PclassDet0.95
14
1st Order Acoustic Model

• Rule-of-thumb look up tables generated by the
  Acoustic Battlefield Aid
• Variable target, terrain, and environmental
  parameters
• Based on field derived data
• CMS2 implementation
• 10 specific targets
• 4 generic targets
• Low / medium / high speed
• Day vs night
• Light vs moderate-heavy wind
• Open grassy vs forested terrain
• Gentle rolling vs mountainous terrain

15
1st Order Seismic Model

• Rule-of-thumb look up tables generated by ERDC
  CRREL seismic model
• Modular algorithms based on hi-fidelity
  simulation
• Field derived target and environment
  approximations
• Low computational burden
• CMS2 implementation
• 3 generic targets (tracked / wheeled / human)
• Variable speed
• APG homogeneous costal ("normal" silty sands)
• YPG homogeneous desert aluvium (strong sandy
  attenuation)
• CRTC unconsolidated glacial till (highest
  attenuation)

16
2nd Order Acoustic/Seismic Model

• Currently evaluating two approaches for a
  dynamic, real-time implementation of AFBA and
  Seismic ROT models
• Background model to generate on-the-fly look up
  tables
• Specific to sensor, terrain location, environment
• Periodic updates to accommodate environment
  changes
• Incorporate algorithm modules directly into CMS2
• Scalability
• Requires synthetic target signature generators
  for both spectra
• Each block below represents a run-time code
  module or library input

Target and Environment
Sensor System
Target (Type, speed, direction)

• Detection
• Information
• Bearing
• Range
• Track
• Class
• (w/Statistical variations based on field
  observations

Signature (Sum of Harmonics)
Signal Level (Propagation)
Sensor Thresholds
S
Background Noise (Empirical)
Receiver Directivity Index/ Geophone Transfer
Function
Environment Parameters
Input Output
17
Acoustic/Seismic Sensor Fusion for Target Location

• Acoustic model (ABFA) outputs a single target
  location estimate with an error estimate
  (circular ellipse)
• Seismic model outputs n sample estimates of
  target location
• Generally an ellipse with better range than
  azimuth accuracy
• NVESD computes weighted center of mass and error
  estimate of the n samples
• We then compute a weighted center of mass of the
  intersection of the 2 ellipses

Location error output from Acoustic model
Area of intersection
Northing
Location error output from seismic model
UGS center of mass
Easting
18
Example Acoustic Detection Models
Day / Flat / Forest
Day / Flat / Grass
Night / Flat / Forest
Night / Flat / Grass
NSfOF UGS Cluster light wind
19
Example Target Location Models
Tracked Vehicle Coastal Region at 400mLine of
Bearing STD 5 deg Range STD 46
mAverage Location Error 53 m
Tracked Vehicle Coastal Region at 800mLine of
Bearing STD 10 deg Range STD 71
mAverage Location Error 125 m

• NSfOF UGS Cluster seismic sensor components
• 1024 samples per second
• 2 second integration time

20
Traditional Software Design Approaches

• Top-Down Design
• Advantages Cohesive system architecture due to
  higher-level abstractions
• Disadvantages Minimal code reuse, potentially
  poor performance
• Bottom-Up Design
• Advantages Efficient, reusable code
• Disadvantages Hard to foresee how low-level
  pieces will serve overall architecture

21
Software Development Approach

• Bi-Directional ('Sandwich') Design
• Top-Down Application-level components based on
  Architectural Design Patterns
• Bottom-Up Simulation Libraries, each written to
  satisfy a domain-specific requirement
• Top-Down portion is relatively simple because we
  use well-known solutions
• Most of our time is spent developing
  domain-specific libraries

22
Software Hierarchy
Applications
UC
Snap Server
CMS2
GEC
Simulation Libraries
Interface Libs ALCES, DaVinci, DIS 2.0.4, JVB,
SEM
GUI Libs GLCanvas, GnomeUtils, GTK2Scheme,
MapGUI, MapRenderer, Overlays, Symbols
Miscellaneous Libs GeomUtils, Coordinates,
Joysticks, NITF, ATM, XMLUtils
Roam Libs RoamCore, RoamContext,
RoamPluginManager, BasePlugins, SimPlugins,
ShapeFile
Sim Libs CommEffects, Detection, Entity,
Munitions, Sensors, TargetDatabase
Invoke Open-Source Libraries
Scheme Libs scheme-access, config-system,
command-line, data-streams, scheme-utility,
shader-lib
Base Libsutility, multicast, threadlib,
data_streams, sim_utility, sim_math
23
CMS2 Design

• CMS2 represents mines and sensors internally as
  individual, high-fidelity 'field entities'.
• Field entities are assembled from component
  objects such as target sensors and warheads.
• New field entities may be assembled from existing
  components by modifying data files. No
  programming effort is required as long as the
  necessary components exist.
• Field entity behavior can be modified without
  re-compilation by modifying scripts and data
  files. This can even be done at run-time.

24
CMS2 Design

• Component objects encapsulate all data and logic
  need to model themselves. Existing components
  include
• Target sensors
• Tripwire, Pressure fuse, Tiltrod, Acoustic (ARL
  model), Seismic, Magnetic, Passive IR
• Munitions
• AP warhead, AT warhead, Shape charge, Top-attack
  fly-out, grenades
• Mine Casings
• Metallic, Plastic, Wooden
• Miscellaneous
• Transmitter, Receiver, Antenna, CPU, Battery

25
CMS2 Design

• Field entities are grouped into fields.
• Fields are a convenient, familiar concept that
  allow the user to manipulate large numbers of
  entities as a whole.
• Fields reduce network load. CMS2 publishes one
  message that describes an entire field instead of
  a separate message for each entity within that
  field.

26
CMS2 Design

• Scalability
• CMS2 can represent very large numbers of field
  entities without reducing modeling fidelity.
• Geometric algorithms are used to eliminate all
  out-of-range target-sensor pairings.
• Performance data is pre-calculated whenever
  possible.
• Example ARL statistical detection tables.
• Efficient coding practices streamline the target
  detection process.
• A single CMS2 workstation has modeled 140,000
  mines while tracking 25,000 target entities.

27
CMS2 Design

• Current work
• Extract the CMS2 simulation module into a
  separate back-end process.
• Implement a controller process which manages
  multiple back-ends running on separate processors
  or workstations.
• Modify the CMS2 GUI to communicate with the
  controller process.
• When complete, a user will be able to create and
  simulate millions of mines and sensors through a
  single CMS2 GUI at the existing level of
  fidelity.

28
Simulation Libraries

• Define domain-specific vocabularies
  ('interfaces')
• Usable (and reusable) at the binary level
• Encapsulate and establish resource management
  policies
• Encapsulate algorithms

29
Library Advantages

• Flexibility/Adaptability
• Can be used in the native environment of any
  experiment or system
• Examples UACEP (DIS), LSI Capstone (JVB), C4ISR
  OTM (DaVinci)
• Can be composed in different ways, depending on
  requirements (scalability, simplicity,
  reliability)
• Doesn't require us to predict future environments
• Reusability
• Approximately 85 of the code written for CMS2,
  UC, Snap Server, and other applications is in the
  simulation libraries
• Changes to a library are reflected in all
  applications that use it
• Scalability
• Not limited by the scalability of communication
  infrastructure
• Programs that use libraries are as scalable as
  the libraries themselves
• Libraries scale up and down

30
Implementation Details

• Algorithms are encapsulated for easy upgrading
• Examples GLCanvas, SymbolRenderer
• Each library sets resource management policies
• Examples LibDetection
• Libraries make no assumptions about the process
  environment
• Examples LibDIS

31
IMS Platform/System Simulation

• Utilize CMS2-Armament Server Federation as the
  system simulation of IMS
• UA / FCS warfighter-in-the-loop simulations
  (Generic IMS)
• LSI SoSIL integration and interoperability
  testing (contractor specific designs)
• IMS PAT support and augmentation (contractor
  specific designs)
• Utilize CMS2 as surrogate T-UGS to provide Layer
  1 sensor information to IMS
• PM-RUS is funding use of CMS2 to support FCS
  T-UGS development
• Utilize Government-LSI defined data/messaging
  interface (Sensor Data Link) between the IMS
  field and the FCS network
• Sensor Data Link (SDL) data and message framework
  under development by PM-NV/RSTA and NVESD
• SDL will be the standard interface from T-UGS
  to FCS C2 network
• CMS2 readily supports implementation of tactical
  messaging interface to support IMS in the SoSIL
• Migrate to MATREX simulation architecture when
  that environment becomes mature and available
• Armament server is core component of MATREX
• PM-FCS has previously funded integration of CMS2
  into JVB
• CMS2 architecture readily supports decoupling
  of embedded sensor services IAW the proposed
  MATREX architecture
• NVESD is tracking the migration of JVB to MATREX

32
IMS HW/SW-in-the-Loop Simulation (Proposed)
IPS 12
CMS2 Federation
Contractor Provided Models or Data
CMS2 Core
Image Server
Comms
Munitions Model/Data
Acoustic Model/Data
Seismic Model/Data
Target Emulator
Other as required
IMS Innerfield Simulation Net
IPS 3
Tactical Message XLATOR
Engage Mgr Gateway Prototype
Munitions / Sensor Prototype
Tactical CS / UC Surrogate
HWIL Tactical IMS Net
Tactical C2 Net
Tactical C2
33
Other Features Planned Upgrades

• DIS message interface
• Spotted PDUs
• Signal PDUs
• HLA interface
• JVB FOM
• MATREX FOM migration
• Tactical message interface (XML)
• Heartbeat (periodic entity status and SA update)
• Contact report (target acquisition)
• Target track database
• Correlation algorithm to minimize multiple
  reports on same target
• 2nd Order sensor algorithm implementation
• Sensor Data Link message interface
• Improved acoustic/seismic fusion algorithm
• ATR implementation for UGS imagers
• Additional sensor types (magnetic, impulse radar)
• Embedded intra-field communications network
  model
• Integration with external communications effects
  server

34
FCS Interoperability

• PM-NV/RSTA is funding the definition and
  development of Sensor Data Link as a proposed new
  standard
• Migration of SLP messages to Joint Variable
  Message Format transport protocol
• LSI reviewing for incorporation into FCS
  architecture
• NVESD is funding Sensor Interface and Access
  Management System (SIAMS)
• Provides the data structure and messaging formats
  for the NSfOF ATD
• Provides a flexible prototyping methodology to
  develop new message/data requirements and data
  management techniques
• FCS LSI is responsible for developing a sensor
  data management architecture for FCS
• PM CCS is responsible for developing a command
  and control, and information architecture for
  integrating IMS with FCS and T-UGS
• SDL/SIAMS provides a common framework for the
  basis of the interface from IMS T-UGS to FCS
• CMS2 architecture supports the implementation of
  tactical messaging (SDL) to support or stimulate
  the development and testing of tactical or sensor
  command and control systems

35
Recommended FCS Interoperability Approach
FCS Information Management Layer
Cluster IV Systems
Common Data Message Interface (Sensor Data Link)
Target Msg. Status Msg. Self-Position Msg.
C2 (local)
Tasking
Attack Guidance
Sensor Performance data
Target Msg. Status Msg. Self-Position Msg.
Target Reports Status Msg. Self-Position Msg.
UGS/IMS Control
API
Tasking
Tasking
FCS NET2
Sensor Comm. interface
Sensor Fusion
Target Msg. Status Msg. Self-Position Msg.
Fused data
Tasking
Raw data
Target Msg. Status Msg. Self-Position Msg.
NET1
Tasking
36
Development Methodology

• Spiral development of PM-NV/RSTAs proposed Sensor
  Data Link standard
• Requirements development and prototyping
  demonstrations using NVESDs SIAMS and DSIF
  development environments
• Focus on standards to govern how sensor data is
  integrated into the common operating picture
• Compatible with the current Tactical Internet,
  but structured to take advantage of a future, and
  more robust FCS TI
• Define a common means to store, catalog, and
  maintain, and then facilitate the transfer of
  sensor data
• Evolve from focus on unattended systems to an
  architecture that addresses tactical sensor
  interfaces
• Vehicle or platform to-from a remote sensor
• Remote communications gateway to-from a remote
  sensor
• Vehicle or platform to-from an on-board sensor
• Soldier to-from soldier-carried or weapon mounted
  sensor
• Ground control or processing station to-from a
  remote sensor

37
SIAMS(Sensor Interface and Management System)

• The SIAMS effort is divided in two parts
• The development and implementation of a message
  protocol (Sensor Data Link future extensions
  called Portable Sensor Data (PSD) that
  communicates between sensor systems and control
  stations
• The development of a Sensor Information
  Management Layer (SIML) to facilitate the
  communication between distributed sensor systems
  and Command and Control Systems

MC2
S I M L
C 2 A P I
SUAV
Sensor Cloud SDL/PSD
UGS
MS
UGV CETS
SEAMS
SDL

• FBCB2

Legacy Sensors
PSD Translator
. . .
38
Bit Steam View of Transmitted Message
SDL Message
User Application Header
String of Data Key/Data Structures
Encryption/Special Functions
Data Link/Network
Portable Sensor Data
39
Identify the Information to be Sent
Depending on File Size, Bandwidth, etc. the
image may be included or sent separately.
40
Generate the PSD Message Portion
The SIAMS Message is formed by the concatenation
of the desired data structures.
41
Complete the Bit Stream and Transmit

• The PSD Portion is packed into the User Data
  Portion of the overall SDL Bit Steam
• Other necessary components are assembled (e.g.,
  User Application Header, Network Header/Footer)
• Bit Steam is then transmitted as a complete
  message

42
Summary

• CMS2 provides a proven and affordable platform to
  support such simulation and experiment events
• Government owned software
• 2-4 PC servers can support simulation with
  10,000s of entities
• CMS2 is now the primary simulation tool being
  used to represent and support development and
  evaluation of UGS and IMS
• Networked Sensors for the Objective Force ATD
• Spider APLA
• Intelligent Munitions Systems for FCS
• Tactical UGS
• UA MBL