Interactive Vehicle Level Human Perfomance Modeling

1
Interactive Vehicle Level Human Perfomance
Modeling

• Presented by
• Mr. Tim Lee
• DCS Corporation
• 1330 Braddock Place
• Alexandria, VA 22302
• Phone [phone] X 203
• Fax [phone]
• tlee_at_dcscorp.com

2
Introduction

• TARDEC Embedded Simulation Team and the Joint
  Virtual Battlespace (JVB) Program conducted a
  collaborative effort to develop a technique to
  implement interactive human performance models
  for the crew of simulated vehicles on a virtual
  battlefield, aka Vehicle Level Human Performance
  Model (VLHPM)
• Initial development and integration based on
  TARDEC Vetronics Technology Testbed (VTT) Vehicle
  model

Vehicle Model Sensors Actuators Automation System
Architecture
Joint Virtual Battlespace Battlefield Friendlies
Enemies
Human Performance Model Operators MM
Interfaces Decision Aides
HPM ICD
JVB FOM
Vehicle Level Human Performance Model
3
Introduction (continued)

• VLHPM utilized ARL Human Research Engineering
  Directorates Improved Performance Research
  Integration Tool (IMPRINT) by Micro Analysis
  Design
• Design event driven models (not scripted static
  models)
• Utilize prioritized goals concept
• TARDEC expanded the effort to model operators of
  Command Vehicle (CV) and Armed Robotic Vehicle
  (ARV) in FCS LSIs Unmanned Combat Demo
• Unmanned Combat Demo
• Utilizing Crew integration Automation Testbed
  (CAT) vehicle assets
• Evaluate maturity of key robotics technologies,
  CV / ARV concept
• Two CV operators Two (or more) ARV operators
• Experiment with different operator to robotic
  asset ratios to determine optimum

4
Introduction (cont.)

• CAT vehicle limitations for UCD
• Two man CAT crew
• Experiments limited to two ARV operators no
  simultaneous CV operations
• Augmentation with constructive operator model
• Four constructive operators, Two CV operators,
  two ARV operators, feasible
• Low cost additions and modifications of crew,
  crewstation model
• Low cost experimentation (vs with real HW,
  operator)
• Mutual refinement and validation with real system

5
CV/ARV VLHPM Overview
CV/ARV Vehicle Model
CV/ARV Operator HPM
Video (for monitoring)
Embedded Simulation System
IMPRINT 6.0 Runtime
JVB C3 Grid Surrogate
CV Dpty
CV Cdr
Simulation Control ESS state/mode control, CGF
control AAR, data log control
Visualizer
Operational Commands
ARV Operator 1
ARV Operator 2
CAT Vehicle Sensor simulation Sensor motion,
visualization, automation (ATR simulation, auto
tracking)
Control Status
Crewstation Interface
ARV Operator 3
ARV Operator n
CAT vehicle Weapon simulation Weapon motion,
control, round Flyout
COM
Execution Time Control
CAT Vehicle Mobility Simulation Automotive
simulation, auto-drive Navigation sim
DIS / HLA Interface
Vehicle I/F
Route/ Tactics Planner
Communications Simulation JVMF message
generation, reception
WSTAWG OE
UGV Simulation UGV/ARV mobility, tele-op, RSTA,
OCSW, Javelin simulations
Internal OTB
Modified from VTT Model
New Stuff
External SAF Environment DIS / JVB / RPR
Disabled when interfaced to external SAF
Optional Stuff
6
CV/ARV Vehicle Model

• Extensive modifications and additions to baseline
  model of VTT Embedded Simulation System (ESS)
• Automation intensive CAT architecture
• Multiple UGV simulation
• Stryker variant for CV
• Wheeled vehicle mobility
• Concept 5.5 ton wheeled ARVs with configurable
  payload
• Weapon simulation based on Cougar turret (OCSW,
  Javeline)
• Simulation of RSTA sensor, Target tracking sensor
  
• Semi-autonomous mobility

Example screen shot of vehicles. Not from actual
experiments
7
CV/ARVHuman Performance Model

• IMPRINT Goal oriented task network paradigm
• Goals can be prioritized
• Execution of tasks (task networks) to achieve a
  goal can be interrupted by execution of tasks of
  a higher priority goal
• Behavior models designed for real-time
  interaction, not static analysis
• Operators are never loaded more than 100
• Reactions to external stimuli represented as
  mutually exclusive prioritized goals
• Continuous main mission tasks that monitor for
  external stimuli
• Awareness and Interaction via MMI
• Assume plausible level of automation
• Processing of audio visual input aggregated as a
  task load factor for each task
• Operations that require sight are approximated
  with, assisted by, or replaced by automation that
  provide digital data (route / tactics planner,
  Automated Target Recognition, Automated Target
  Tracking).

8
CV/ARVHuman Performance Model ARV Operator
7 Report In Defilade Goal -Tell CDR In Defilade
6 View RSTASCAN Goal -View Image - Send Resume
Msg
ARV In Defilade Report
ARV RSTA SCAN

• 2
• View RSTA Target Goal
• View Target
• Add Target to Queue
• Tell Cdr Targets

Monitor ARV
1 Teleop ARV Goal - Teleop ARV back 20m -Update
ARV Route -Send Resume Msg
ARV ATR Alarm
ARV Stuck Alarm

• 4
• Engage Threat Goal
• -Confirm Targets
• -Engage Targets
• Clear Targets
• -Tell CDR Engagement
• Complete

5 Plan to Objective Goal - Develop ARV Plan -Send
ARV Plan
Engage ARV Targetsgt0
Plan to Objective
Monitor Internal Communications
3 Route to Defilade Goal -Enter ARV plan -Send
ARV plan
Defilade
ARV Operator Mission and Goal Breakdown
9
CV/ARVHuman Performance Model CV Commander
1 Process C2 Message Goal -Enter CV Route -Tell
ARV Plan to Obj -Set Fire Perm. Free -Tell
DPTY Drive Route
C2 Alarm (Unit March)
Monitor External Communications
5 View Threat Goal -Confirm Targets -Report
TypeSPOT
CV ATR Alarm !FREE
Monitor Surveillance Systems

• 2
• Seek Defilade
• Goal
• -Update CV Route
• Say Defilade

Targets ARV Targetsgt1
CV ATR Alarm FREE
4 Engage Threat Goal -Set Fire Perm.
Free -Confirm Targets -Engage Targets -Report
TypeSIT
ARV Targetsgt1Free In Defilade1 In
Defilade2 At Destination OR ARV Targets1
FREE
3 Unit Engage Goal Say Engage
Monitor Internal Communications
At Destination
7 Send C2 Report Goal -Enter C2 Report -Send C2
Report
Detonation CV Targetsgt1
6 Resume Mission Goal -Update CV Route -Tell ARV
Plan to Objective -Report TypeSIT
Engagement Complete
Monitor Self Protection Systems
CV Commander Mission and Goal Breakdown
10
CV/ARVHuman Performance Model CV Deputy/Driver
1 Drive Route Goal - Steer to Waypoint -Accelerate
/Brake to Waypoint -Say At Destination
Drive Route
Monitor Internal Communications
CV Driver Mission and Goal Breakdown
11
CV/ARVHuman Performance Model
Example Task Network Driving Task network
12
CV/ARVHuman Performance Model Automation

• Both manual and autonomous control of mobility
  designed to follow a route generated by a route /
  tactics planner
• Considers terrain elevations
• Considers terrain type
• Manual control model adjusts actuator values in
  real time to keep vehicle on route
• Autonomous control model issues a plan that
  contains the route and RSTA scan commands to the
  vehicle model
• Target detection by RSTA Automatic Target
  Recognition
• Target engagement with Automatic Target Tracking

13
CV/ARVHuman Performance Model Automation
Open Terrain
Non-traversable area
Roads
Routes for two ARVs And CV
Non-traversable area
Pictorial representation of example terrain type
data and generated routes Elevation data is not
shown
14
CV/ARVHuman Performance Model User Interface

• Real time display of Visual Auditory, Cognitive,
  Psychomotor (VACP) loading on each operator
• Real time indication of vehicle position
  navigation data
• All interfaces and data logging of IMPRINT 6.0

15
Experiments

• Unmanned Combat Demo
• One CV, two ARVs
• Initial integration demo of maneuvers
• Capstone Demo
• One CV, one ARV
• Limited participation due to risk mitigation
• Virtual Distributed Lab for Modeling Simulation
  (VDLMS) First Application
• One CV, one ARV
• Modified vehicle model, and HPM for participation
  as a Forward Observer / Laser Designator of
  Netfires concept evaluation
• Maneuver in recon mission
• Spot report or call for fire (for Precision
  Attack Missile)
• Perform laser designation
• Participated in majority of recorded runs
  simultaneously with CAT crewstations

16
Lessons Learned

• Real-time interactive simulation of complex
  behaviors of combat vehicle operator(s) is
  feasible
• With goal oriented task model of IMPRINT
• With a high fidelity vehicle model that has a
  rich, portable data interface
• With sensor automation such as ATR, ATT that
  minimizes the need to convert and process visual
  information
• Enhancements to IMPRINT can simplify dynamic HPM
  development, such as
• Nested IF or While constructs
• Temporary variables in macros
• Improved External Model Call facility
• Inter-operability with cognitive models (SOAR,
  ACT-R)
• Improvement of command data protocol / interface
  is needed
• JVB C3 grid interactions afforded limited
  flexibility
• Direct manual command injection via a C3 Grid
  surrogate was implemented for non-JVB
  experiments
• Additional manual input of decision thresholds
  for target engagement were added for First App

17
Future Plans

• Enhancement of C3 interactions, processing,
  decision making
• Higher echelon platform(s) or organization
• Improvement of C3 interaction protocols, data
  sets
• Run-time manual command input
• Simulation C2 interface JVB, Netfires, MATREX
• Modeling of or integration with other automation
  or simulation tools
• On board decision aids
• Higher fidelity component or functionality models
  of VDLMS (other HLA federates)
• Integrated Unit Simulation System (IUSS) of
  SBCCOM
• Cognitive simulation tools (ACT-R, SOAR)

18
Future Plans

• Higher fidelity route / tactics planner
• Higher fidelity environment data
• More rigorous algorithms
• Controlled data collection and analysis
• Manned crewstations HPM executing same
  operation (achieved at First App, but not enough
  correlated data)
• Compare HPM predicted performance with actual
• Adjust HPM
• Continued support of simulation analysis
  objectives of FCS SDD
• Enhancements based on all the above
• CAT HPM developments in parallel with crewstation
  enhancements / modifications
• Unmanned Combat Demo II