HumanAutomation Interface Model to Guide Automation Design of System Functions

1
Human-Automation Interface Model to Guide
Automation Design of System Functions
Josh Kennedy Michael McCauley, Ph.D. Army
Research Lab Naval Postgraduate School
2
Background

• Armys Future Combat Systems (FCS)
• Manned Ground Vehicle (MGV) fleet
• Sub-contractor UDLPBAE Systems
• CRADA NPS and UDLP/BAE for HSI
• Many HSI risk areas
• ORD requirement 2-Soldier Crew!

3
Too Many Engineers

• Promise advanced automation to take on tasks
  formerly performed by Soldiers
• Tendency automate the easiest tasks (technically
  feasible), leave the rest to the operator(s)
• This dignifies the human operator(s)
• This approach to automation design does not
  relieve a human operator of tasks
• Merely shifts manual tasks to more supervisory
  tasks
• Soldier(s) become super-supervisor

4
Will Lead to

• But, leads to a hodgepodge of partial automation
• Remaining control tasks to be performed by humans
  are less coherent, less meaningful, and more
  complex than need be
• Wholly inappropriate.many examples in
  transportation, industrial accidents
• No consistent plan in place for an overall
  human-automation interface scheme

5
Problem Statement

• The manned ground vehicle (MGV) fleet of the
  Armys Future Combat Systems (FCS) lacks an
  overarching, top-down approach to its
  human-automation interface scheme.
• Given the restriction of a 2-soldier crew, we
  must design a human-automation interface model
  that
• Can be applied to the MGV common crew station
  (CCS)
• Develop a functional architecture between human
  and automation for the total system

6
Proposal

• Develop a qualitative model to drive the
  functional architecture and the human-automation
  interface scheme on the MGV fleet
• Apply new interface scheme to the parts of the
  MGV common functional/task analysis (from
  BAE-Santa Clara)
• Quantitatively apply new interface scheme via
  Advanced IMPRINT models

7
Qualitative Model

• Function Allocation
• LOA can vary across a continuum
• Simple 4-stage info processing model

8
Kennedy 5-Stage ModelInfo Processing for
Human-Automation Interaction Scheme
Expertise
9
FCS Manned Ground Vehicle Fleet (Common Crew
Station)
Analysis
Acquisition (See First)
COA Development
Action(s) (Act First)
Decision(s)
(Understand First)
High
High
High
High
High
10
10
9
9
Function A
8
8
(Finish Decisively)
7
7
Function A
6
6
LOA
LOA
5
5
4
4
System B
3
3
2
2
System B
1
1
Low
Low
Low
Low
Low
10
FCS Manned Ground Vehicle Fleet (Common Crew
Station) Local Defense (CFM 5)
Acquire/Track/Engage Threat (CFM 51-52)
Detection
ID / Track
Shoot / Report
Engagement
COA Development
High
High
High
High
High
10
10
9
9
8
8
7
7
6
6
LOA
LOA
5
5
Reject
Reject
Reject
Reject
4
4
Proposal
3
3
2
2
Current
1
1
1
2
3
1
2
3
Low
Low
Low
Low
Low

• Characteristics of target (e.g. bearing, speed,
  altitude

• Arm Weapon
• Fire Weapon
• Prepare/Xmit Digital SITREP

• Possible COAs listed

• Symbology of Target
• Status of ownship sensors
• Status of CTP/LINK

• Engage the Target?
• Which weapon?
• Which ammo?

11
Quantitative Implementation

• CFM5 (Local Defense) Current vs. Proposal
• Using IMPRINT
• Predict human task-loading
• Equate to mental workload
• One (of many) human performance indicators
• Five metrics from previous IMPRINT usage

12
The Stats
13
Discussion

• Results of paired comparison are what they
  arenot to be overemphasized
• Possible reductions in task-loading
• Thus a lowering of the mental workload (only a
  construct)
• Theoretically, human performance benefits as a
  result
• If, and only if, automation design is possible to
  proposed levels
• Supports efforts to meet 2-Soldier crew
  requirement
• Method allows clear Soldier involvement
• Gain some understanding of human performance
  ramifications in automated systems

14
Conclusions

• Provides HFE/SE a top-down, overarching approach
• Explicitly define and design the interaction
  between proposed automation schemes and the human
  crew.
• Constitutes the design methodology and automation
  philosophy as espoused by Rouse et al (1987).
• Coherent plan for automation will help ensure
  soldier performance and system effectiveness
• Model and analytical processes proposed are
  certainly useful in a wide array of complex
  systems in multiple domains (aviation, space,
  maritime, manufacturing, etc.)

15
HSI Implications

• Manpower Personnel
• Trade-off in crew-size reqt
• Higher caliber enlistee
• Training
• Clear mental model of automation scheme
• HFE
• Define human role in overall system
• Keep MWL at consistently acceptable level
• System Safety/Soldier Survivability
• Expand FMEA/FMECA efforts to Soldier-automation
  interactions

16
Relevant Quote

• In response to article Blame the Umpires in
  Wall Street Journal, 14 Oct 05
• Letter to the Editor, 25 Oct 05
• Ive always felt that the strike zone should be
  automated. It seems like we should be able to use
  advanced technology to remove the need for a
  human to make this incredibly difficult call.
• - Steve Clark (Cary, NC)

17
BACKUPSLIDES
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
CFM Functional Flow Level 1
22
CFM Functional Flow Level 2
23
CFM Functional Flow Level 3 Function 5.0
24
Local Defense (CFM5) with Automation Model Applied
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
Baseline MGV CFMWorkload Predictions
29
Proposed Interface Scheme Applied to MGV CFM