Military reference books and manuals (2009-2023, Volume 6)

FOR OFFICIAL USE ONLY

Foreword

The United States Air Force has long envisioned a strategic role for remotely

piloted and autonomous aircraft. As early as May 1896, Samuel Pierpont Langley

developed an unpiloted heavier-than-air vehicle which flew over the Potomac River. On

V-J Day in August 1945, General Hap Arnold, US Army Air Forces, observed¹:

We have just won a war with a lot of heroes flying around in planes. The next war

may be fought by airplanes with no men in them at all … Take everything you’ve

learned about aviation in war, throw it out of the window, and let’s go to work on

tomorrow’s aviation. It will be different from anything the world has ever seen.

Since these early days, extended range, persistence, precision, and stealth have

characterized remotely piloted aircraft (RPA) advancements. RPAs have been employed

in multiple combat roles and increasingly contested environments. This year, for the first

time in history, the President’s budget proposed a larger investment in RPAs than

manned aircraft. A seemingly insatiable operational appetite for RPAs, however, has led

to an Air Force manning bottleneck. This is exacerbated by a lack of common ground

stations, unsatisfactory integration with civilian and international airspace, and

vulnerabilities in communications and command and control links. Further complicating

efforts, yet essential in irregular warfare, are directives to minimize civilian casualties.

General David Petraeus sees this need as a direct way to support a key center of gravity:

…We must fight the insurgents, and will use the tools at our disposal to both defeat the

enemy and protect our forces. But we will not win based on the number of Taliban we

kill, but instead on our ability to separate insurgents from the center of gravity - the

people …²

Our Panel conducted an extensive set of visits and received numerous briefings

from a wide range of key stakeholders in government, industry, and academia. Taking a

human-centered, evidence-based approach, our study seeks to address operational

challenges as well as point to new opportunities for future RPAs. That RPAs will be a

foundational element of the Air Force’s force structure is no longer debatable. The real

question is how to maximize their current and future potential. Our intention is that this

study will help provide both vector and thrust in how to do so in the irregular warfare

context, as well as other applications.

¹ Words on War: Military Quotes from Ancient Times to the Present, by Jay Shafritz, Prentice Hall, New

York, 1990, pg. 104

² Tactical Directive. Headquarters ISAF

6 July

2009. http://usacac.army.mil/cac2/coin/repository/

FOR OFFICIAL USE ONLY

iii

FOR OFFICIAL USE ONLY

Executive Summary

Background

RPAs are revolutionary surveillance and weapons delivery systems - changing

the way the Air Force builds situation awareness and engages enemy forces - but their

full potential has yet to be realized. To begin to address this issue, the Air Force initiated

this study to review the state-of-the-art in RPA operations, focusing on control and

connectivity in an irregular warfare (IW) environment. The Panel was specifically tasked

to identify RPA architectures and operational concepts centered on human-systems

integration, distributed systems operations, and effective command and control - a cluster

of concepts and technologies we subsequently labeled as

“mission management”

enablers. The Panel was also tasked to recommend mid- to far-term S&T development

roadmaps for advancing these technologies to improve the flexibility and capability of

RPA operations. The study terms of reference (TOR) identified a number of core issues

which were further articulated by the Study Panel to include:

Issue #1: Manning and personnel shortfalls are concerns in RPA deployment.

Exploiters represent the largest manning dependency (39 percent), exacerbated by

expected significant exploiter growth from new sensor suites (e.g., ARGUS-IS,

Gorgon Stare). Current sensors

(e.g., Constant Hawk and Angel Fire) and

expected sensors (e.g., ARGUS-IS) produce data at rates of 10 to over 1000 times

projected communications data transmission capacities, and will far exceed

human analytic capacity.

Issue #2: Manually intensive airspace management and integration requiring

exclusion zones and Certificates of Authorization (COAs) make inefficient use of

national and international airspace, will not scale to accommodate future RPA

growth, hampers manned/unmanned integration, and presents special challenges

for small RPAs.

Issue #3: Minimizing collateral damage (CD) and fratricide is not a requirement

unique to RPA strike operations. For manned and unmanned platforms, the lack

of positive ID (PID) and tactical patience are the most significant causes of

civilian casualties (CIVCAS) in current conflicts (8 percent CIVCAS compared

with

66 percent caused by insurgents). Persistence; up-close access; high-

resolution intelligence, surveillance, and reconnaissance

(ISR); improved

situation awareness; and improved mission management will permit RPAs to

minimize CD/fratricide. Small-focused lethality munitions and non-lethal options

for RPAs promise to further minimize CD and CIVCAS (e.g., as low as 5

percent).

Issue #4: In spite of current low RPA losses, inexpensive physical threats (e.g.,

MANPADS, low-end SAMs, air-to-air missiles) and electronic threats

(e.g.,

acoustic detectors, low cost acquisition radars, jammers) threaten future

operations.

FOR OFFICIAL USE ONLY

Findings

Following an extensive set of fact-finding meetings with operators, developers,

trainers, analysts, maintainers, and scientists and technologists, together with

data/evidence analysis where possible, the Panel distilled our findings into the following

top-level observations regarding RPA mission management in irregular warfare

environments:

Key Finding

1: Current RPA automation implementations either under-

automate, over-automate, or fail to provide a flexible human-centric solution.

Insufficient and inflexible automation/sensor processing increases pilot workload,

increases incident rates, degrades mission performance and agility, and inhibits

distributed cross-platform collaboration.

Key Finding 2: Poorly-designed Operator Control Stations

(OCSs) fail to

provide effective, robust, and safe RPA mission management because of a lack of

accepted systems engineering design practices, a lack of Human-System

Integration

(HSI) design and implementation, and closed and stovepiped

architectures that constrain “best of breed” component solutions.

Key Finding 3: Limited communications systems result in communications

latency, link vulnerabilities, and lost-link events, which limits mission roles

assigned to RPAs, operational flexibility, and resiliency in the face of

unanticipated events.

Key Finding 4: Successful transition of the Beta program to Undergraduate RPA

Training (URT) helps establish the RPA career field. This transition is a positive

step to address the manning shortfalls, but shortcomings in crew selection and

training continue to contribute to bottlenecks and compromise mission

effectiveness and safety.

Key Finding 5: Concepts of RPA operation (CONOPS) and tactics, techniques,

and procedures

(TTPs) appear to be developed

“after the fact,” following

development and deployment rather than as a concurrent effort.

Recommendations

As a result of our findings and extensive deliberations, the Study Panel makes the

following key recommendations with associated offices of primary responsibility

(OPRs):

1. Recommendation 1: Develop flexible levels of automation to enable situation-

adaptive human interaction for improved mission effectiveness in the near term

for single-platform task-level automation (ASC), in the mid term for mission-

level automation (AFRL), and in the far term for multi-RPA collaboration and

autonomy (AFRL).

2. Recommendation 2: Improve the operator control stations in the near term by

applying human systems engineering practices to correct existing operator control

stations (AFMC, ACC), in the mid term by developing a family of networked,

interoperable systems (ESC, ASC), and in the far term by advancing human-

FOR OFFICIAL USE ONLY

automation interaction science and technology to enable adaptive, mixed

initiative, human-centric control (AFRL).

Recommendation 3: Develop assured communications, appropriate security, and

a long term plan for a globally interoperable communications systems architecture

(AFMC/ESC). Reduce bandwidth requirements, latency impacts, and security

vulnerabilities by developing a unified communications architecture that in the

short term provides “good enough” sensor security with high C2 link reliability

and security (AFMC, AFRL/RB); in the mid term adopts increased platform

automation and latency compensation including platform forecasting and target

tracking (AFMC, AFRL/RB); and in the long term provides electronic warfare

countermeasures, addresses scale, and supports networked operations of

heterogeneous assets (AFMC/ASC, AFRL/RB, AF/A3/5).

Recommendation 4: Understand fundamental capabilities and skillsets required

for each RPA specialty, select appropriately, and exploit current training and

simulation technologies to reduce manning pipeline delays and deal with

contested environments (AF/A3/5, ACC). In the near term, activities should be

directed at developing scientifically based RPA career path enhancements that

generate targeted selection procedures and accelerated training programs

(AF/A3/5, ACC). In the far term, focus on developing higher level skillsets,

including team coordination, communication, and mission planning.

Recommendation 5: Develop RPA CONOPS and TTPs concurrently with

system development, considering all components that drive mission effectiveness

(ACC). In the near term define how the current SOC/WOC/AOC structures can

enable distributed global operations via

“brokering” among operators,

platforms/sensors, and “customers” (ACC). In the mid term, anticipate growth in

RPA operations, growth in sensor capabilities, transition to integrated

manned/unmanned operations, and more contested operating environments

(AF/A9). In the far term, co-evolve CONOPS/TTPs for new platforms and

missions as part of a focused acquisition process, leading systems development

and informing systems requirements choices (ACC).

Recommendation 6: Transition the successful RPA prototypes to the acquisition

process, incorporating rigorous systems engineering practices including Human

Systems Integration (SAF/AQ). Continue successful use of the ACTD process to

develop game-changing technology solutions, while upgrading, developing, and

acquiring new RPA systems using accepted systems engineering practices (e.g.

2010 NDAA Section 804) to address operator-related issues via incorporation of

HSI as part of the systems engineering process, and follow recommendations of

two previous SAB studies^3,4.

³ United States.

“Unmanned Aerial Vehicles in Perspective: Effects, Capabilities, and Technologies.

Volume 1: Summary (Public Released)” SAB-TR-03-01 September 2003.

United States. “Human-System Integration in Air Force Weapon Systems Development and Acquisition,”

Executive Summary and Annotated Brief, SAB-TR-04-04, July 2004.

FOR OFFICIAL USE ONLY

vii

FOR OFFICIAL USE ONLY

Report Structure

The report is structured as follows. Following a backg1round of and motivation for

the study, the Panel summarizes the key RPA mission management issues. The Panel

then details each of the key findings derived from a hypothesis-driven, evidence-based

analytic process. A subsequent recommendations section details specific actionable

recommendations designed to address the challenges uncovered in the findings section.

A section on the implications of under-investments follows. The report concludes by

summarizing the implications of our study. Report appendices detail assumptions,

hypotheses, and future data collection requirements, study members, briefings, visits,

acronyms, references, and report distribution.

FOR OFFICIAL USE ONLY

viii

FOR OFFICIAL USE ONLY

Table of Contents

Foreword……………………………………………………………………………………………………iii

Executive Summary

Background

Findings

Recommendations

Report Structure

viii

Table of Contents

Table of Figures

Chapter 1: Study Background

1.1 Background and Motivation

1.2 Scope of Study

1.3 Study Approach and Methodology

1.4 Challenges

Chapter 2: Mission Management Challenges

2.1 Issue 1: Manning Shortfalls and Personnel Selection/Training

2.2 Issue 2: Airspace Deconfliction and Management

2.3 Issue 3: Minimizing Collateral Damage/Fratricide

2.4 Issue 4: Operations in Contested Airspace

2.5 Summary of Issues

Chapter 3: Major Findings

3.1 Finding 1: Insufficient and Inflexible Platform Automation

3.2 Finding 2: Poorly Designed Operator Control Stations

3.3 Finding 3: Limited Communications Systems

3.4 Finding 4: Inadequate Selection Criteria and Training

3.5 Finding 5: Lagging CONOPS and TTPs

3.6 Summary of Findings

Chapter 4: Recommendations

4.1 Recommendation 1: Develop Flexible Levels of Automation

4.2 Recommendation 2: Enhance Operator Control Stations

4.3 Recommendation 3: Create Robust Communications

4.4 Recommendation 4: Develop Targeted Selection and Effective Training

4.5 Recommendation 5: Improve CONOPS and TTPs

4.6 Recommendation 6: Ensure Effective ACTD Transition

Chapter 5: Analysis and Summary of Recommendations

5.1 Impact Analysis of Recommendations

5.2 Summary of Recommendations

Chapter 6: Summary

Appendix A: Assumptions, Initial Hypotheses, and Data Requirements

A.1 Assumptions

A.2 Initial Hypotheses

A.3 Data Requirements

FOR OFFICIAL USE ONLY

Table of Figures

Figure

1-1: RPAs have Revolutionized Air Warfare

Figure

1-1: RPA Groups

Figure

1-2: Irregular Warfare Constraints

Figure

1-3: Study Process

Figure

1-3: Study Structure

Figure

1-4: Problem Space

Figure

1-4: Today’s Reality vs Tomorrow’s Potential

Figure

2-1: Manning Growth

Figure

2-2: Effect of Wide Area Surveillance on Personnel Requirements

Figure

2-3: Unmanned and Manned Aircraft Comparison

Figure

2-4: Beta Class Performance

Figure

2-5: Airspace Management and Integration

Figure

2-6: Improved Situation Awareness to Support Airspace Reallocation

Figure

2-7: German RPA in a Near Miss with an Airliner over Kabul (view from RPA)

Figure

2-8: Collateral Damage

Figure

2-9: Contested Environment Operations

Figure

2-10: Summary of Operational Issues

Figure

3-1: Automation Findings

Figure

3-2: HSI Issues with Operator Control Stations for Predator/Reaper

Figure

3-3: Operator Control Stations Findings

Figure

3-4: Limited Communications Systems

Figure

3-5: Inadequate Selection Criteria and Training

Figure

3-6: Lagging CONOPS and TTPs

Figure

3-7: Summary of Findings

Figure

4-1: Improved Platform Automation

Figure

4-2: Enhanced Operator Control Stations

Figure

4-3: Create Robust Communications

Figure

4-4: Training and Selection Recommendations

Figure

4-5: Improved Operating Concepts, CONOPS, and TTPS

Figure

4-6: Effective ACTD Transitions for RPAs

Figure

4-7: CONOPS Technology Double Helix

Figure

5-1: Effect of Addressing Mission Management Gaps on Operational Issues

Figure

5-2: Technology, People, and Process Investments fill Mission Gaps

Figure

5-3: Manning and Selection Recommendations

FOR OFFICIAL USE ONLY

Chapter 1: Study Background

1.1 Background and Motivation

Remotely piloted aircraft (RPAs) have been acquired to perform dull, dirty, and

dangerous missions for the Air Force. Dull missions include situation awareness

missions, especially those involving persistent surveillance of unmoving targets or

identifying the “normal” activities in an area so that unusual activities can be spotted.

Dirty missions include flights/missions into areas in which suspected chemical or

biological weapons or contamination are suspected. Dangerous missions include those in

which the potential for loss of the aircraft can be high, but for which there is a high

payoff for information that can bring weapons on targets.

As illustrated in Figure 1-1, RPAs have revolutionized operations in irregular

warfare. RPA operations have grown rapidly in both volume and application.

Approaching one half million annual flight hours across the services, RPAs have

provided persistent surveillance and engagement with adversaries. Through dedication,

ingenuity, and drive, the Air Force has overcome many obstacles to assure effective

support to our ground forces. Daily air operators are continuously innovating with new

missions and Tactics, Techniques, and Procedures (TTPs) to take advantage of RPAs.

Future visions include many new missions and roles, complementing and extending

manned operations.

The Air Force has introduced remote split operations to successfully reduce our

operational footprint in the battlespace, by locating the majority of personnel at CONUS

bases for mission command and control and the remainder at the operational airfields for

take-off and landing operations. Overall, remote split operations increases the number of

personnel involved in operating the aircraft while decreasing the risk of flight accidents

during take-offs and landing.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Persistent surveillance of areas, targets, and individuals in

OIF/OEF has paid off in finding and engaging the adversary

The growth in flight hours

450000

400000

speaks for itself

350000

Through dedication, ingenuity,

300000

250000

and drive, the USAF RPA team

200000

has overcome many obstacles

150000

to assure support to the

100000

50000

ground forces

New applications and TTPs

emerge each day

SOURCE: OSD

AIR FORCE ARMY NAVY & USMC

…and visions for the future

encompass a wide array of roles

complementing and extending

existing manned operations

Figure 1-1: USAF RPAs Have Revolutionized Air Warfare in the IW Arena.

The wide range of RPA Groups and platforms illustrated in Figure 1-2 provides

for a variety of capabilities to be used against a variety of targets, including surveillance,

reconnaissance, and engagement targets under conditions we control rather than those

controlled by adversaries. RPA payload and range is plotted on a log-log scale, showing

the wide variation in platforms from a few pounds to thousands of pounds, and from a

few miles range to thousands of miles. Although not plotted, altitude varies likewise,

from a few feet to many tens of thousands of feet. The Air Force operates RPAs in all

five Groups shown. Unfortunately this wide array of RPA vehicles, acquired under

individual agreements, and operated and maintained by separate units and personnel

further exacerbate the manpower demands to operate the aircraft in the field. Common

control stations, common training protocols and systems, and common maintenance

procedures could go a long way toward minimizing the number of personnel required to

maintain an effective battlespace presence of such aircraft.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

10,000

Group 5

Global

Hawk

1,000

Group 4

Reaper

Firescout

Sky Warrior

Warrior-A

Predator

A-160 Hummingbird

Hunter A/B

Group 3

100

Shadow

Tier II/ STUAS

M2UAS

EUAS

Tiger Shark

Sentry

Group 2

Neptune/VCUAS

Scan Eagle

Silver Fox

Group 1

AECV

BUSTER

Raven

Aqua Puma

Aqua Wasp

Wasp III/ BATMAV

100

1,000

10,000

Max Radius (nm)

Figure 1.2: RPA Groups by Payload and Range.

1.2 Scope of Study

The Study Panel was tasked to review the state-of-the-art in RPA

operations,

focusing on control and connectivity in an irregular warfare (IW) environment. The

Panel was also asked to identify RPA architectures and operational concepts enabling low

probability of collateral damage/fratricide. The Panel was also asked to identify enabling

[mission management] technologies to include:

 Human Systems Integration, and potential for reduced manning, enhanced

operator awareness, and multi-aircraft control (MAC)

 Distributed System Operations, and potential for multi-aircraft collaboration

(manned and unmanned), and better airspace management and integration

 Command and Control, and potential for more efficient bandwidth use, and

improved mission assurance in degraded or denied communications

environments

Finally, the Panel was tasked to recommend mid- to far-term S&T development

roadmaps for these technologies, with an emphasis on improving the flexibility and

capability of RPA operations.

As shown in Figure 1-3, Air Force Doctrine Document 2-3 (1 August 2007)

defines Irregular Warfare as “a violent struggle among state and nonstate actors for

legitimacy and influence over the relevant populations. IW favors indirect approaches,

though it may employ the full range of military and other capabilities to seek asymmetric

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

approaches to erode an adversary's power, influence, and will.”⁵ In addition, because

airpower roles in IW are almost as broad as they are in non-nuclear MCO⁶ and because

RPAs could conceivably be employed in most of those roles (given enough technological

and operational advancements), it follows that RPAs could conceivably be used to apply

a full range of (non-nuclear) airpower capabilities in an IW engagement. What this

means for this study, because of the broad potential for RPA employment and

engagement, is that collateral damage minimization is critical, because of the “struggle

for legitimacy” over relevant populations in an IW engagement. Finally, though present

IW operations enjoy uncontested airspace, we can expect contested airspace in the future,

although for the purposes of this study, denied airspace is considered out of scope.

IW’s objective and approach (AFDD 2-3):

“[A] struggle … for legitimacy and influence over the relevant

populations.”

“[IW] may employ the full range of military and other capabilities …to

erode an adversary's power, influence, and will.”

What does this mean for this study?

Airpower roles in IW are almost as broad

as they are in non-nuclear MCO (WP by

SECAF Donley and Gen Schwartz, 2009)

… and RPAs could conceivably be

employed in most of those roles

Collateral damage minimization is critical,

because of the “struggle for legitimacy”

Contested airspace can be expected in the

future. Denied airspace is out of scope.

Figure 1-3: Irregular Warfare Constraints.

1.3 Study Approach and Methodology

The study process is characterized in the model represented in Figure

1-4.

Focused by the bounds of the Terms of Reference (TOR), the Study Panel both reviewed

past studies and visited a number of industry, government, and academic facilities to

assess the current state of the art. The Panel brainstormed to generate hypotheses, but at

the same time attempted to explicitly capture underlying assumptions (Appendices A1

and A2). The Panel then collected data from a broad range of sources attempting to

confirm, deny, or refine its hypotheses, some of which became key operational issues to

be addressed (“Background” as shown in Figure 1-5), and some of which became

identified gaps in mission management associated with those issues (“Findings” in Figure

1-5). In a parallel and iterative fashion, we then turned our attention to generating

⁵ (AFDD 2-3)

⁶ Donley and Schwartz 2009

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

recommendations in terms of technology, people, and processes to address these mission

management gaps uncovered by our findings (“Recommendations” in Figure 1-5). A

final phase consisted of group assessment of findings and recommendations and

assignment of those to specific offices of primary responsibility

(OPRs). At the

beginning of the study, the study leadership met with key stakeholders to obtain clarity

on Air Force senior leadership priorities. The Panel then met again midcourse to perform

a vector check on the study’s progress. The study results were then reported to senior Air

Force leadership and key stakeholders.

Study Process

Review past studies

and state of art

Brainstorm to

Assess findings and

generate

recommendations

assumptions and

hypotheses

Collect data to

Generate

confirm/deny/refine

recommendations

hypotheses

Convert hypotheses

into findings

Figure 1-4: Study Process.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Study Structure

Background

Findings

Recommendations

Mission Management

Technology, People,

Operational Issues

Gaps

Processes

Figure 1-5: Study Structure.

1.4 Challenges

RPAs have provided the Air Force revolutionary capabilities such as persistence

for dull missions, and safe and stealthy access for dirty and dangerous missions.

However, there are challenges with current RPAs including:

 Significant personnel footprint per RPA combat air patrol (CAP)



“Soda straw” sensors (i.e., narrow field of view) that serve up full motion

video (FMV) that requires manual exploitation

 Human-intensive guidance and control (“in the loop” not “on the loop”),

based on low-level remote control technologies

 Low operator situation awareness (SA)

 Human-intensive airspace management and deconfliction procedures

 RPAs acquired and operated as individual platforms

 Manual management of collateral damage/fratricide risk

 Heavy reliance on continuous long-haul (e.g. space-based) communications

 Capabilities that generally rely on uncontested battle space

A more integrated view of these issues (and others not listed) and potential

capabilities, as they evolve over time, is given in Figure 1-6. Shown here are seven key

factors: increasing threats, increased aircraft density, airspace integration, operating in all

weather conditions, operating in multiple and increasingly challenging roles

(e.g.,

strategic strike, SEAD), multi-aircraft control (MAC), and increasing levels of autonomy.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Problem Space

contested

denied

Threat^uncontested

Low

medium

high

Density

Full Integration

Procedural

Partial

Airspace

(manned/unmanned,

Separation

Integration

military/civilian)

Integration

Some weather

All weather

Weather^{No weather}

Multirole

Long range strike,

Role

surveillance/strike

Unmanned fighter

Single

4 ship

N-ship

Semi Automation

Fully Autonomous

Autonom

Human in loop

Human in/on loop

Human on loop

TODAY

MID TERM

LONG TERM

Figure 1-6: Problem Space.

Figure 1-7 illustrates some current limitations of RPAs and the promise for

advanced solutions in the future. The Study Team believes the primary solution for these

shortcomings is a unified mission management capability that integrates unmanned and

manned assets into a more focused operational mission capability that relies on

unmanned assets for the portions of the mission in which they are most capable. RPAs

can perform many missions, but it is important for weapon delivery to have a person at

the stage of weapon delivery making the trigger decision based on timely intelligence and

accurate situation awareness to ensure low collateral damage.

Today

Tomorrow

Significant personnel footprint per RPA

Progressive personnel reductions in pilots, sensor operators, exploiters,

combat air patrol (CAP)

and maintainers through multiple advances in all areas.

Soda straw imagers serving up full motion

Wide Area Airborne Surveillance (WAAS) with some/all processing,

video (FMV), manually exploited

exploitation, dissemination (PED) done onboard.

Human-intensive guidance and control

Human supervisory control (“on the loop”) enabled by architectural and

(“in the loop”), based on “remote control”

algorithmic advances providing for multi-vehicle management, and better

technologies

engineered HCIs.

Low operator situation awareness (SA)

Mission-specific SA enabled by fused sensors, health monitoring,…

Airspace management and deconfliction

Airspace management and deconfliction for manned/unmanned

is human intensive

operations done more autonomously (eg, onboard sense and avoid)

Operated as individual platforms

Operated collaboratively via multi-platform comms and coordination,

across manned and unmanned platforms.

Collateral damage/fratricide risk managed

Collateral damage/fratricide risk management supported by decision-

entirely manually

aiding at mission control and/or onboard automation

Heavy reliance on continuous long-haul

More intermittent comms enabled by more autonomy when unlinked

comms

Uncontested battlespace (mostly)

Operating in contested battlespace (from kinetic air/ground, EW, and

cyber threats) via stealth, defensive systems, secure comms,…

Limited-mission capable (ISR, shooters)

Multi-mission capable (EW, counterair missile trucks, ...) via modular

platforms, payloads, comms, and operator interfaces.

Figure 1-7: Today’s Reality vs Tomorrow’s Potential.

FOR OFFICIAL USE ONLY

Chapter 2: Mission Management Challenges

A number of key challenges that hamper effective mission management of

remotely piloted aircraft (RPAs) were identified in the terms of reference and noted by

the Study Panel as key issues. In particular, the Study Panel focused on manning

shortfalls and personnel selection, airspace management and integration, minimizing

collateral damage and fratricide, and contested airspace. We consider each of these

issues in detail in the following sections.

2.1 Issue (1): Manning Shortfalls and Personnel

Selection/Training

RPA manning difficulties are a direct result of the increase in use of RPAs for

surveillance and interdiction missions. As Figure

2-1 illustrates, maintaining a

Predator/Reaper combat air patrol (CAP) (which represent the vast majority of Air Force

unmanned vehicle operations) for 24/7 coverage requires approximately 168 personnel⁷

and four aircraft. In terms of overall Predator/Reaper RPA operations, there are 4 basic

categories of personnel for a single CAP: exploiters

(31 percent), maintainers

(40

percent), pilots (6 percent), and sensor operators (6 percent), with the other 12 percent

representing other administrative and support personnel. This distribution is similar for

the Global Hawk, although the manning requirements are higher (approximately 300).

⁷ Deptula, David. Lt. Gen. “The Way Ahead: Remotely Piloted Aircraft in the United States Air Force,”

presented to the SAB, January 2010.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Planned growth in RPA deployment

Global Hawk

Predator

perceived to be limited by available

Pilots

12%

qualified personnel

17%

Sensor Operators

“Our #1 manning problem in the Air Force is

Maintainers

manning our unmanned platforms”

31%

36%

45%

34%

Exploiters

>70% exploiters and maintainers

Other

Exploiter growth accelerated by WAAS

3500

(Gorgon Stare, ARGUS)

Not an RPA-specific issue

3000

Maintainer problem improves with

MQ-9 only “spot”

2500

numbers with WAAS

experience & system maturity

2000

Pilots require longest & costliest

training investment; source of a

1500

MQ-1 and MQ-9

possible chokepoint

FMV “spot”

1000

numbers w/o WAAS

Recruitment low and retention poor

500

FY13 RPA total pilot requirement is 1250

FY13 Undergraduate RPA Training (URT)

production will be 168/yr

FY08

FY09

FY10

FY11

FY12

FY13

FY14

FY15

FY16

CAPs CAPs CAPs CAPs

Figure 2-1: Manning Growth.

Although the Air Force’s focus has been on dealing with RPA pilot shortfalls,

exploiters and maintenance personnel represent the largest manpower requirement.

Whereas recruitment and retention of these personnel may not be considered to be a

problem in the current climate, with systems like Gorgon Stare and ARGUS-IS coming

on line with 10 to 65 additional full motion video (FMV) feeds (“spots”) per platform, the

manpower to support a single CAP could grow significantly (if 65 people are needed for

a single video stream, 10 times 65 would be 650 people required for a Gorgon Stare feed

- See Figure 2-2). Without changes in how such images are processed (e.g., simply

delivering the imagery to ground forces versus exploiting them via reachback facilities),

the insertion of new wide area sensing technologies is likely to overwhelm this aspect of

operations, with the ultimate result that the majority of data will simply not be analyzed.

Unless this problem can be addressed, the effective use of these technologies appears to

be unrealistic.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

3500

3520

3000

2750

2500

2250

2000

MQ-9 only

1750

“spot” numbers

WAAS Increment Two

with WAAS

1500

65 spots per MQ-9

1250

MQ-1 and MQ-9

FMV “spot”

1000

WAAS Increment One

numbers w/o

13 spots per MQ-9

WAAS

750

500

250

FY08

FY09

FY10

FY11

FY12

FY13

FY14

FY15

FY16

CAPs CAPs CAPs CAPs

Figure 2-2: Effect of Wide Area Surveillance on Personnel Requirements.

The reality is that the RPA manning issue is more affected by FMV processing,

exploitation, and dissemination (PED) than operation of the aircraft. A comparison of

RPAs flying a 24-hour surveillance mission with a manned aircraft (MC-12 Liberty)

flying a similar mission (Figure 2-3) shows that, in fact, operating the RPA requires

fewer pilots and fewer aircraft because of the longer time on target offered by

Predator/Reaper. The number of maintainers appears abnormally large for the Predator

aircraft, but could be expected to decline as these aircraft mature and maintainers become

more familiar with the aircraft.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Unmanned (Predator) vs. Manned (Liberty)

120

100

Sensor

Other (CC,

Exploiters

Maintainers

Pilots

Total

Operators

MC, OH)

Predator CONUS

109

Predator AOR

Liberty CONUS

Liberty AOR

108

Figure 2-3: Unmanned and Manned Aircraft Comparison⁸.

Because exploiters are the real chokepoint for intelligence, surveillance, and

reconnaissance (ISR) missions, the focus needs to be on training and improving the

efficiency of the exploiters and the quality of the products they provide for the warfighter

(targeting, time, and location stamping, friendly forces identification). A more urgent

effort needs to be placed on providing better technology for these personnel to identify

and date/time stamp targets of interest. Also, there is a need to provide better integration

of target identification, sensor data, and pilot observations into the situation awareness

data fed back to the warfighter (in the air and on the ground).

Currently, the Air Force is using on-the-job training to train exploiters in applying

their capabilities to providing data to the warfighter. The problem with on-the-job

training is that there is no way to introduce improved data integration in that environment

or to improve the efficiency with which the product is developed. Additionally, there is

no control over training material (e.g., anomalies, critical threats) when training occurs on

the job.

If one’s assumption is that training pilots (and non-pilots) to become RPA pilots

is a major personnel issue (this may not be a valid assumption for long-term operations,

as just discussed), there are two accession solutions: 1) train more pilots and/or 2)

increase the number of vehicles pilots can manage. In addition, recruitment/retention is

critical for keeping these personnel. Note that there is a feedback loop between the multi-

vehicle solution (see below) and the retention issue, such that if pilots have higher job

satisfaction and are not so bored and/or marginalized, retention will improve.

⁸ AF/A2, “CAP Comparisons - MQ1~MQ9~RQ4~MC12 (20 Apr 10), Presented to SAB, June 2010; Brig.

Gen. Blair Hansen, “Project Liberty ‘MC-12’,” Presented January 2010.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

An analysis of costs per training hour⁹ reveals that transition (traditionally

trained) pilots cost approximately

$2100/hour at the end of RPA training, SUPT

(Specialized Undergraduate Pilot Training) pilots $500/hour, and Beta pilots¹⁰

$150/hour,

as shown in Figure 2-4. Thus, transition pilots cost more than an order of magnitude

more per training hour than do Betas. Of note is the fact that current grade sheet analysis

from the Predator schoolhouse (see Figure 2.4) shows that the transition pilots generally

lag the SUPT pilots in performance, and on some measures, perform similarly to Beta

pilots. So whereas transition pilots cost significantly more to train, they do not

outperform their counterparts as would be expected given their training investment.

MQ-1 Gradesheet Performance Across Common Training

Missions

Traditional

SUPT

Beta

Avg. $ cost per UAS

Common

2.75

Syllabus

training hour

Missions

2.5

1. Sim-1

2. Sim-2

$2,108.66

2.25

3. Sim-3

4. S-ISR-1

5. S-ISR-2

6. ISR-1

SUPT

$498.59

7. ISR-2

1.75

8. ISR-3

9. S-BSA-1

Beta

$150.18

1.5

10. S-BSA-3

11. S-CAS-1

1.25

12. S-CAS-3

13. CO-1

14. S-EP-1

15. S-EP-2

Common Syllabus Missions

16. CO-3

Beta classes generally seen as a failure across the stakeholders we interviewed:

Screening tools required external review

On several measures, TX performances were similar to Beta. SUPTs were clearly better than TXs and Betas

Beta class “failures” required on average 20 extra hours of training (1.25 extra

sorties)per pilot class of 8)

An addition of 2.2% off training time which = an additional $3003.60 to the $135,162 total

Air Force Audit Agency proposes monetary benefits of ~1.5B with the introduction of

Beta class UAS pilots

Figure 2-4: Beta Class Performance.

Given that the training approach to generate more pilots can only partially address

the RPA manning problem, one other solution is to have pilots control more than one

vehicle simultaneously. Research has shown that one person can control anywhere from

two to twelve independent RPAs, depending upon autonomy level and fidelity of

decision support.¹¹ Although this may appear to be a clear technical solution, more

⁹ Assuming that 18 weeks of UAS training = 900 hours for SUPT and Beta, and 693 for TX pilots, source:

Air Force Audit Agency Unmanned Aerial System Pilot Force Management Audit Report F2009-0005-

fd4000 17 December 2008.

¹⁰ The Beta program for training ab initio Remotely Piloted Aircraft (RPA) pilots was initiated to prove and

refine training concepts and to alleviate pressure on conventional pilot training assets by creating a new

pilot pipeline. A distinguishing characteristic of the Beta RPA pilot training track is reduced actual cockpit

hours (44 hours in flight, versus approximately 200 or more hours in flight for conventional students). This

results in reduced pipeline time and reduced expense, and helps evolve toward a professionally rewarding

stand-alone RPA pilot career track.

Cummings, M. L. et al. 2007. Automation Architecture for Single Operator-Multiple UAV Command

and Control. The International Command and Control Journal 1(2): 1-24.

FOR OFFICIAL USE ONLY

research is needed on how concepts of Operations (CONOPS) would and could change

under this paradigm-shift, (i.e., should a pilot be on call for emergency events, are there

certain mission types that require a pilot [like CAS or transit of the National Airspace

(NAS)] while en route; can loiter missions be left to sensor operators with a pilot on

call?).

2.2 Issue (2): Airspace Deconfliction and Management

A second issue identified in the TOR and illustrated in Figure 2-5 is airspace

management and integration. The Federal Aviation Administration

(FAA) is the

governing agency responsible for establishing flight rules for the operation of air vehicles

in the NAS, and their primary responsibility is to facilitate the flow of air traffic, while

keeping the NAS safe for all operational users. The NAS is sub-divided into categories

of airspace with various operational flight restrictions published to control traffic and

mitigate risk. Operational rules such as flight under Visual Flight Rules (VFR) and

Instrument Flight Rules

(IFR), in concert with airspace categories and associated

restrictions, help to manage the flow of a diverse variety of aircraft while deconflicting

traffic and keeping the airspace safe for all users. These “rules of the road” were written

with pilot capabilities as well as aircraft technical capabilities in mind, and the majority

of the rules since their inception have been intended for manned vehicles.

RPA operations in National Airspace (NAS) limited by FAA policy and flight rules

designed for manned systems (eg, “see and avoid”)

Conservative risk management concepts

Lack of national standards for RPAs, and heavy reliance on (local) COAs

Low confidence in RPA response predictability under lost-link

RPA operational issues under ICAO rules are more

Military Controlled Airspace

restrictive than NAS

RPA operations in Military Controlled Airspace

(MCA) present additional opportunities and

challenges in IW environment

Some host nation flight restrictions may exist

Joint doctrine creates management flexibility, but

integration of manned/unmanned operations limited b

Ground

“ROZ”

Forces

Manually intensive processes

Controlled

Airspace

”Reference Grid”

Inefficient geographic/temporal separation

Figure 2-5: Airspace Management and Integration.

Managing the risks associated with flight in the NAS by a rapidly growing

population of diverse RPAs, using VFR and IFR rulesets dependent upon pilot cockpit

actions, presents challenging airspace management and integration issues for both the

FAA and prospective NAS operators. While the FAA continues to study the full

implications and risks associated with RPA flight in the NAS, unmanned flight risks are

FOR OFFICIAL USE ONLY

currently mitigated and permitted on a case by case basis, based on the application and

approval of a formal FAA Certificate of Authorization (COA).

Approved COAs represent a “hand shake” between the Federal Government and

the unmanned aircraft operator, to operate a specific aircraft in a defined geographic area

under a set of flight rules for a specified period of time. COAs for governmental

operators are typically approved for a year, and can be updated annually. While the

COAs do permit flight in the NAS, and many are active today, most are very limited in

nature, restricted to sparsely populated areas, and tend to avoid all major flight airways.

COA approvals can take anywhere from months to years, which can seriously hamper

Small Unmanned Aerial Systems (SUAS) development and testing.

For RPAs to be able to effectively integrate with manned aircraft in the NAS, the

FAA continues to insist that the RPAs be capable of performing all of the flight rules

which have been codified for pilot occupied cockpits. Flight by RPAs under VFR flight

rules in Visual Meteorological Conditions (VMC) are fundamentally challenged by the

ruleset which requires the pilot/RPA to be able to “see and avoid” and comply with

published “right of way” maneuvers to avoid impending collisions. This requirement to

see and avoid has become a show stopper for most UAS operators attempting to procure

approved COA operations. As the RPA aircraft become larger, and the onboard systems

more robust, there are more potential solutions (Traffic Collision Avoidance System

[TCAS] with predictable avoidance maneuvers is one of these) which may provide the

risk mitigations and capabilities that the FAA mandates.

It should be noted that TCAS was considered for implementation on Global Hawk

and Predator Class vehicles; however, because of the bearing error data and low update

rate of TCAS, the FAA and the International Civil Aviation Organization (ICAO) have

stated that the TCAS display alone is not sufficient to provide the operator with enough

situation awareness to avoid the threat. The inability to perform visual acquisition means

that the traffic display information cannot be corroborated by the UAV operator, and

therefore the certification authorities are concerned about its use.

The results of a dynamic simulation study conducted for Global Hawk indicate

that a significant improvement in safety is provided by equipping the aircraft with TCAS

in addition to its currently installed Mode S transponder.¹² Risk ratios from the TCAS-

equipped cases in which the aircraft is permitted to fly a preplanned avoidance route are

in the range of 0.003-0.079, compared to 0.004-0.058 for conventional aircraft; however,

multi-aircraft avoidance, which might be required in military controlled airspace (MCA),

was not considered. Also, if RPA pilots are required to disconnect the autopilot and fly

the aircraft by hand, response times can vary significantly (because of communications

latencies) and the risk increases accordingly. In addition, because TCAS operates by

interrogating transponders on equipped aircraft, non-cooperative traffic, or aircraft

without transponders, are not tracked by TCAS.

As illustrated earlier in Figure

1-2, RPAs are currently grouped into five

categories, based upon size, and thus technical capability. In the top category, Group 5,

¹² Billingsley, 2^nd Lt. Thomas B., “Safety Analysis of TCAS on Global Hawk using Airspace Encounter

Models”, USAF Academy and MIT/Lincoln Labs, June 2006.

FOR OFFICIAL USE ONLY

the USAF Global Hawk is a full sized aircraft with typical systems such as would be

found in a manned aircraft. It includes all of the conventional instruments and navigation

capabilities, is pre-programmed to fly precise navigation routes, and is connected to

ground based operators thru the Ku Band Satellite infrastructure for real-time flight

maneuvering. Group 4 includes the Predator/Reaper family of RPAs, also manned

aircraft sized and very capable, comparable to the Global Hawk, and also flown thru the

Ku Band system. The Department of Homeland Security has been flying the Predator

Bin the NAS successfully on FAA COAs for more than five years. There are a very large

number and variety of SUAS in the three smaller, Groups 1-3, but these aircraft are all

subscale, with very limited technical capability in comparison those in Groups 4 and 5,

and these aircraft are going to have a very difficult time getting approval for unrestricted

flight in the NAS.

While the FAA works to adapt their legacy rulesets to accommodate RPA

operations in the NAS, they continue to hold users to a very high standard, one which is

unlikely to permit unfettered Group 1-3 RPA operation. Group 4-5 operation is a

different case. An illustrative example is the Department of Homeland Security’s (DHS)

Predator B operation. The Predator family of RPAs has more than a million operational

flight hours and has been in service for more than 15 years. It has all of the technical

equipment found on the most sophisticated manned platform, including electro-

optical/infrared

(EO/IR) and synthetic aperture radar (SAR) sensors, and all of the

necessary radios and navigation equipment to fly accurately and safely in the NAS. The

aircraft is virtually flown by FAA certified pilots thru a Ku Band Satellite Link, has IFF

(identification friend or foe) and TCAS (although the Air Force does not presently use

TCAS). The aircraft are only flown from Department of Defense (DOD) operating bases

on IFR Flight Plans, in the area of Positive Control

(PCA). These missions are

operational homeland security missions flown across the nation’s borders, and the flights

are under continual radar surveillance by local FAA and DOD radar sites that are actively

monitored by the Air and Marine Operation System in Riverside CA. This facility is

fully integrated with the FAA, with FAA controller presence and continual

communication with all FAA Air Traffic Control Centers. When the aircraft is linked to

the ground-based pilots, it can see and avoid on these missions better than a manned

aircraft (because of additional ground feeds not normally available in a manned cockpit),

but when link is lost, the aircraft navigates a pre-planned route, and it cannot see and

avoid when it descends out of the PCA and enters the airspace where VFR rules mandate

see and avoid capability.

In the last five years, DHS Predators have flown more than 6,000 operational

hours in the NAS, and during all of that time, the RPA lost link only .05 percent of the

time, with the longest lost-link period less than two minutes. The FAA considers lost-

link a serious issue, and lost-link combined with unexpected aircraft motions in altitude

control or navigation to be grounding errors.

Next Generation Air Traffic Control policies and technologies are being

developed, and it is anticipated that the NextGen system will depend on cooperative

onboard technologies to provide safe and effective airborne integration. The fact that the

FAA plans to move away from non-cooperative ground based radars in favor of new

onboard systems built on IFF, TCAS, and other emergent technologies does not bode

FOR OFFICIAL USE ONLY

well for small payload SUAS (Groups 1-3), but the larger RPAs in Groups 4 and 5 may

be able to carry the new technology and meet future FAA rules for NAS flight.

In combat AORs, illustrated in Figure 2-6, the rulesets are defined by Military

Controlled Airspace

(MCA), and there is more flexibility in options for

unmanned/manned integration. Combat theaters previously had more tolerance for

collateral damage, gaining military advantage by easing flight rulesets. Today, the

increased concern over collateral casualties among the civilian population and the ability

to target individuals instead of areas, has led to a more stringently applied control and

deconfliction between manned and unmanned systems, as well as joint operations by

manned systems from various partner nations. In the combat theater there is more

uniformity of operators and vehicle systems in various altitude regimes, and spatial-

temporal exclusion zones provide the requisite degree of safety and separation necessary.

The operation of RPAs in these environments is perhaps the best use for small subscale

aircraft with high combat utility, but limited capability to meet FAA or ICAO rules for

flight in the NAS. Group 4 and 5 RPAs are very capable in the MCA, and have the

necessary potential for safe and unconstrained long range navigation in the NAS and in

ICAO airspace on overseas missions. As onboard Next Gen technologies emerge,

aircraft in this class that have the capacity to carry this new equipment should, with little

problem, be able to safely fly the NAS along with manned aircraft .

In Military Controlled Airspace (MCA), the combat area is typically cordoned off

into regions for flight of RPA Groups 1-3 at altitudes below 3,000 ft., which is controlled

by ground forces (as shown in Figure 2-6). Closely associated with the requirement for

enhanced situation awareness is the need to provide planners with the ability to rapidly

conduct airspace reallocation to support mission re-planning. Currently, MCAs specified

in the ACO remain in effect until the next ACO is published (normally every 24 hrs).

This leads to an inefficient utilization of the air environment, as MCAs may only be

required for a short time period. Even when an MCA is in use, it may be possible, with

the ability to track assets within airspaces through enhanced SA, to temporarily reallocate

potions of active airspaces to higher priority tasks. For example, if situation awareness

allowed airspace managers to know that an airborne tanker was in the north of its

designated orbit, they could temporarily reassign part the southern portion of the orbit to

facilitate the transit of a Combat Search and Rescue (CSAR) package.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Figure 2-6: Improved Situation Awareness to Support Airspace Reallocation¹³.

A long-standing problem in air warfare is the need to coordinate air and ground

operations with the aim of preventing potential blue-on-blue engagements, both from the

ground and from the air. This is a particularly relevant issue for the coordination of

aircraft with ground-based air defense (GBAD). Airspace managers require improved

situation awareness of air and ground operations, along with the current weapons control

status, to ensure that appropriate measures are put in place to prevent fratricide incidents.

This is especially germane in high-tempo operations in which both air and ground units

are moving rapidly over wide geographical areas.

One of the greatest challenges facing airspace managers is their current inability

to coordinate and deconflict the operation of RPAs. The unprecedented proliferation of

RPAs in recent years, specifically within tactical level ground units and sub-units, has

dramatically increased the risk to air operations. In Afghanistan, an Airbus 300B4

airliner with 100 personnel on board came within 170 feet of a German EMT Luna

tactical RPA,¹⁴ (see Figure 2-7) and in Iraq reports have indicated that helicopters have

been struck by RPAs.¹⁵

¹³ “RPA Airspace Integration,” Presentation to SAB by A3O, January 2010.

¹⁴ “Near Misses Between UAVs And Airliners Prompt NATO Low-Level Rules Review,” Flight

International, March 2006.

¹⁵ Erwin, Sandra I., “Controlling Iraq's Crowded Airspace No Easy Task,” by Sandra I. Erwin, National

Defense Magazine, November 2005.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Figure 2-7: German RPA in a Near Miss with an Airliner over Kabul (view

from RPA)¹⁶.

This deconfliction issue becomes more complex when considering the integration

of manned and unmanned vehicles in the same airspace. In this case, see and avoid

becomes much more important, along with predictable avoidance maneuvers coupled

with predictable lost-link maneuvers. Today’s RPAs have not always exhibited these

capabilities and thus we have reverted to assigned domains for each in segregated

airspace.

However, the ability to exhibit these three attributes (sense and avoid, predictable

avoidance maneuvers, and predictable lost-link maneuvers) are critical to both integrated

military airspace and for integration into the national and international airspaces in the

future. Thus, the Panel feels that working together with the FAA to determine reliable

attributes for any RPA that would fly the NAS will not only reduce constraints in future

CONUS operations, but will also move the Air Force ahead toward integrated air

missions in the future in military controlled airspace.

Finally, operating RPAs in combat theaters around the world will necessitate

compliance with International Civil Aviation Organization

(ICAO) rules for flight

operation in International Airspace. These rules, though similar in nature in terms of

their flight procedures and risk mitigations, may be even more restrictive than flights in

the NAS. Large RPAs in Group 4 and Group 5 with long range transit capabilities will

be particularly susceptible to these issues during flights between operational areas in

which the military are in control of airspace management and deconfliction.

¹⁶ Bundeswehr photo, 30 August 2004.

FOR OFFICIAL USE ONLY

2.3 Issue (3): Minimizing Collateral Damage/Fratricide

A third issue identified by the Study Panel was collateral damage/fratricide (Figure

2-8). RPAs, originally developed for ISR operations, have become important weapons

platforms for tactical and special strike missions in IW. Their expanded use in CAS

missions in the future requires technology improvements for mission management to

minimize fratricide, collateral damage (CD), and civilian casualties (CIVCAS).

In IW, success requires winning the “hearts and minds” of the population in the

face of an adaptable and agile adversary hiding amongst them. A missile fired (e.g.

Hellfire missile) from a RPA is no different from a Hellfire missile fired from other

platforms like the AH-64 Apache. Causing collateral damage is not an issue unique to

RPAs. Data obtained from the Afghanistan AOR¹⁷ confirms that insurgents have caused

approximately two thirds of CIVCAS. The exact number of CIVCAS caused by US

forces was not reported, but an estimate from available data suggests the figure to be less

than 10 percent. Of these CIVCAS, approximately half were caused by air-to-ground

munitions, but the role of RPAs in these CIVCAS was also not reported. In the majority

of these CIVCAS, inadequate acquisition and maintenance of positive target

identification (PID) was the primary cause, and the ability to provide tactical patience

during operations would have improved mission success and minimized CIVCAS. In an

article by the Washington Post,¹⁸ it was reported that within a recent 15-month period, the

CIA conducted 70 RPA strikes using the low collateral damage focused lethality

Scorpion weapon, killing 400 terrorists and insurgents while causing 20 CIVCAS. This

CIVCAS figure was based on the use of RPAs to conduct pre-strike ISR and post-strike

battle damage assessments. Because of precision targeting and focused lethality,

CIVCAS is now primarily dependent on the human intelligence and situation awareness

upon which the targeting decision is based.

¹⁷ Joint Center for Operational Analysis, Civilian Casualties in Afghanistan, 2010.

¹⁸ Warrick, J., and Finn, P. “Amid outrage over civilian deaths in Pakistan, CIA turns to smaller missiles.”

Washington Post, April 26, 2010.

FOR OFFICIAL USE ONLY

“…we will not win based on the number of Taliban we kill, but instead on our

ability to separate insurgents from the center of gravity - the people….”

General David Petraeus

Primary factors behind US - Afghanistan civilian casualties from all causes*

Inadequate acquisition/maintenance of positive target ID

Limited tactical patience

Scorpion

IW Rules of Engagement demand:

Low Collateral Damage/

Directed Lethality

Persistent, highly precise ISR

Precision strike/low collateral damage (LCD) munitions

Collateral damage is not an RPA-unique issue, but RPA capabilities in IW

can help minimize the problem

Persistent wide-area surveillance ˘ Improved situation awareness

Increased connectivity ˘ More “eyes on target” and more opportunities for data

fusion from available feeds

LCD capable

*SOURCE: Joint Center for Operational Analysis

Figure 2-8: Collateral Damage.

An advantage of the RPA is greater pre-strike surveillance capability that can

better prevent collateral damage. RPAs provide the ability to conduct ISR to gather

“pattern-of-life” information and thus allow their human operators to distinguish between

non-combatants and legitimate targets. Importantly, the use of future RPAs in IW may

well prove to be a solution to the CD dilemma, rather than an additional source of the

problem.

In current IW operations, the Rules of Engagement (ROEs) emphasizing low

collateral damage generate challenging requirements. To minimize targeting of non-

combatants, persistent and highly precise ISR must be available. To minimize blast

effects beyond the desired target, focused lethality strike weapons must be employed. An

increased RPA ISR persistence capability needs to enable continuous tracking to improve

target confidence via multi-sensor pods

(e.g., Gorgon Stare), efficient cross-cueing,

multi-INT fusion, and automated processing of FMV for event detection. To increase the

likelihood of positive target identification future RPAs should provide high definition EO

and IR FMV, and have the capability for high-definition (HD) single-frame photography.

In future RPA operations, networked distributed operations promise to minimize

CD by providing more “eyes on target” and thus better situation awareness and rapid

transfer of flight control and munitions delivery to specialized operators best prepared for

the dynamic tactical situation.

Future RPA missions in IW can benefit from the development of low collateral

damage munitions that are optimized for focused lethality and directional kinetic effect.

In IW, it can be more appropriate to capture rather than kill the enemy, so RPAs can be

valuable in allowing for the use of less-than-lethal force to neutralize a target. Options

for new non-lethal approaches (e.g., acoustic, directed energy) and expanding RPA

capacity for non-kinetic operations

(e.g., show of force, PSYOP, strategic

FOR OFFICIAL USE ONLY

communication) hold promise to enable RPAs to be an effective weapons platform while

reducing the risk of collateral damage.

2.4 Issue (4): Operations in Contested Airspace

The nature of air warfare in irregular warfare is dramatically different from

operation in denied anti-access airspace protected by a sophisticated Integrated Air

Defense System (IADS). Today the Air Force employs theater-capable Group 4-5 RPA

platforms primarily for ISR, tactical ISR, counter insurgency, and time sensitive target

(TST) operations. The experiences in Iraq and Afghanistan in the absence of an IADS

are that counter-RPA action is more a

“pick up” collection of small arms, rocket

propelled grenades

(RPG), a few shoulder fired missiles, and some rudimentary

electronic warfare (EW) jamming devices.

As noted in Figure 2-9, the Study Panel is aware of only a few instances of enemy

fire destroying an RPA, all of that prior to 2006. In the midst of the challenges imposed

by providing additional support to the ground troops, previous and ongoing experience

makes preparations for operations in contested battlespace a low priority task. The

impact of this is apparent both in the limited use of technology to enhance survivability

(the notable exception is RQ-170), and in the limited development of TTPs and conduct

of warfighter training/exercises.

RPA losses due to enemy actions to date are minor

Planning for operations in contested battlespace is not a priority

RPA technology addressing survivability is limited (exception is RQ-170)

TTPs and warfighter training/exercises are limited at best

Mid- and far-term IW operations are likely to render assets and TTPs

less effective by increasingly capable adversaries

Target

Attack

Effect

Platforms

small IR MANPADS, directed

destroyed or disabled platforms

energy, small arms

Sensors

lasers, dazzlers, cyber attacks,

blinded/damaged sensors,

camouflage

deception, collateral damage

Communication Links

jamming, spoofing,

loss of ISR data, exploited data

eavesdropping, hijacking

links, loss of C2

Position, Navigation &

cyber attacks, GPS jamming &

manipulated/disrupted C2,

Timing

repeater attacks, spoofing

collateral damage

Figure 2-9: Contested Environment Operations.

Although the threat today is limited, one can expect that the defenses against

RPAs will improve in future IW scenarios, for two reasons: 1) as RPAs become more

effective against IW operations, adversaries will be driven to finding countermeasures

and 2) the simple diffusion of technology will provide the needed capability.

FOR OFFICIAL USE ONLY

2.4.1 Threat to Platforms

The relatively large size of Group 4-5 platforms offers several opportunities for

defense:



In the future, relatively simple acoustic receivers capable of detecting and

localizing RPAs are quite possible and could have a simple processing capability

to classify the aircraft.



Low-cost air acquisition radars could be in the hands of combatants, though

probably effective only against the larger, faster RPAs. A rudimentary network,

perhaps as simple as voice “forward-tell” will enhance the Find/Fix capability, but

it unlikely that an IADS as we know it will appear in the short term.



The most likely area of improvement in the air defenses of adversaries against the

larger, faster, and higher RPAs in the IW of the future will be in the use of small

shoulder-fired surface-to-air missiles (MANPADS), some developed in the US

(Stinger) and Russia (SA-16/18), stolen or copied by these adversaries, but more

likely manufactured by countries sympathetic to their cause. Though the range is

limited to a few miles, improvements are likely, and such weapons are easily

smuggled between countries. The use of mobile SAMs such as the SA-6, SA-11,

and SA-17 that would be effective at the operating altitudes of the

Predator/Reaper is less likely in IW, but with cases of strong backing by major

military powers, they could appear.



There have been a few RPA shoot-downs by aircraft. These have involved RPAs

comparable to the Group 3 RPAs and have generally been attributed to IR

missiles fired from fighter aircraft. The IW environment is unlikely to include

such threats.



The 2006 AF SAB Summer Study, “Air Defense Against Unmanned Aerial

Vehicles,” addressed the actions the US could take to defend against UAV/RPA

attacks at home and abroad, and suggested near-term actions as well as both near-

and far-term technology development programs. Though focusing on what

systems the US could employ against enemy RPAs, it suggested several other

possibilities worth considering as future threats. A potential enemy could obtain,

at modest cost, small vehicle-mounted radar-directed high-rate guns already

developed and tested in the US. Netted material could be launched into the flight

paths of smaller RPAs flying at low altitudes. The Study Report is recommended

for further review of potential threats.

For the larger Group 4-5 RPAs, there is a case for making the aircraft more

survivable for the sophisticated defenses expected during Major Combat Operations

(MCO). A modern IADS would quickly decimate the current Predator/Reaper fleet and

be a serious threat against the high-flying Global Hawk. In the IW scenario, however, the

case for highly survivable RPAs is less compelling. As discussed, MANPADS

(particularly IR) are likely the greatest threat for these RPAs, so highly survivable

designs may not be necessary.

FOR OFFICIAL USE ONLY

Self-protection equipment such as jammers, RF and IR flares, towed decoys, etc.

is not suggested for Group 1-3 RPAs, as the weight alone would require unacceptable

reductions in payload or fuel. Group 4-5 RPA development efforts that are more

advanced than Predator/Reaper (UCAV, UCAS, for example) have high price tags driven

largely by design, manufacturing, and maintenance costs associated with high

survivability requirements. Typically, active self protection techniques are additive and

are costly as well. A limited suite matched to the likely threats (IR flares against

MANPADS, etc.) that are fully automated should be considered for future IW

engagements. Given the smaller nature of the RPA signatures, a new class of expendable

decoys may be necessary.

The Panel feels that the costs of hardening a smaller RPA against even a modest

threat would quickly become prohibitive. A better approach for the smaller Group 1-3

RPAs should be to reduce unit costs to a level at which moderate losses would be

acceptable. Low cost, “expendable” RPAs will enable the extensive use of redundancy to

mitigate threats and improve resilience. The use of “swarms” to accomplish the tasks at

hand, coupled with multi-vehicle mission management concepts, including threat

avoidance (route planning, pseudo-random flight paths [jinking], low cost decoys, etc.),

and accepting the potential losses could be made cost effective.

Consideration should be given to manufacturing and deploying representative

RPAs without the expensive mission equipment to be used as expendable decoys,

drawing defensive fire and perhaps exhausting the adversary’s munitions. At a

minimum, the decoys, flown with well-thought-out tactics will expose the location of the

adversaries. These decoys (with visual, acoustic, and infrared signatures similar to those

of the mission-focused RPAs) could be operated within the swarm, perhaps at lower

altitudes to divert attention from the mission-focused RPAs. Critical to this concept is

lowering the cost of manufacture of the airframes, engines, and communications links.

2.4.2 Threat to Sensors

Effective ISR operations can be degraded by sensor attacks that may include a

number of methods. Examples include:

 Adversaries can use lasers and dazzlers to blind or damage EO/IR sensors.

 Widely available techniques for camouflaging can hide potential targets or can be

used for deception.

 Cyber attacks for spoofing or distorting sensor downloads may lead to

inaccuracies that can result in increased collateral damage.

The cost of RPA systems is dominated by the cost of the mission package

(sensors, sensor processing, wideband data links, weapons, etc.). The hunger for higher-

resolution and larger images, and bigger weapons effects is expanding in an uncontrolled

fashion. Improved electronic manufacturing techniques coupled with higher production

quantities can provide some cost reduction, but RPA mission packages will remain high

cost items for the foreseeable future.

FOR OFFICIAL USE ONLY

2.4.3 Threat to Communication Links

Current RPA operations depend on maintaining high reliability of communication

links; therefore they are a primary target of the adversary:

 Jamming of commercial satellite communications (SATCOM) links is a widely

available technology. It can provide an effective tool for adversaries against data

links or as a way for command and control (C2) denial.

 Operational needs may require the use of unencrypted data links to provide

broadcast services to ground troops without security clearances. Eavesdropping

on these links is a known exploit that is available to adversaries for extremely

low cost.

 Spoofing or hijacking links can lead to damaging missions, or even to platform

loss.

Recognizing that there is a possibility that a determined adversary could employ

relatively simple jammer techniques to disrupt navigation and communications, there are

some actions that can be taken to reduce the effect of the jamming. Given that the

character of a jammer used by an adversary in IW will probably have limited effective

range, global position satellite (GPS) jamming can be mitigated by including a simple

flight management auto-pilot that can just adopt/maintain a heading to clear the jammer

(effecting a simple threat avoidance maneuver). As for communications jamming, a

number of modest-cost electronic counter countermeasure (ECCM) radios exist and are

in use for Group 4-5 (and perhaps Group 3) which, along with a communications

architecture that includes alternate control communications paths, can similarly escape

the jammed region. For the smaller Group 1-2 platforms, simple spread spectrum radios

may escape detection and would survive some jamming. Such spread spectrum radios

are already in use by model aircraft enthusiasts who want to avoid unintentional

frequency conflicts and jamming.

Jamming or exploitation of wideband sensor communications is more complex as

is the protection against such actions. NSA Type 1 encryption has been tested on Rover

5 radios used by forward personnel to receive RPA imagery and could be easily included

in the future. Some current wideband communications systems have ECCM features that

would likely survive the nature of countermeasures characteristic of IW forces.

2.4.4 Threat to Position, Navigation, and Guidance

There is a wide range of methods that a determined adversary can use for

attacking RPA guidance and navigation systems. The report mentions here only three

categories of threats without going into the details:

 Small, simple GPS noise jammers can be easily constructed and employed by an

unsophisticated adversary and would be effective over a limited RPA operating

area.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

 GPS repeaters are also available for corrupting navigation capabilities of RPAs.

 Cyber threats represent a major challenge for future RPA operations. Cyber

attacks can affect both on-board and ground systems, and exploits may range

from asymmetric CNO attacks to highly sophisticated electronic systems and

software attacks.

2.5 Summary of Issues

RPA mission effectiveness will not reach full potential, unless the Air Force deals

directly with four key operational issues including manning shortfalls and personnel

selection, airspace management and integration, collateral damage and fratricide, and

operations in contested environments (Figure 2-10). These issues will grow in severity

with time for a variety of reasons including RPA growth in numbers and density,

expansion of assigned operational roles, increasing importance of information operations

in IW, and movement from advantaged (or benign) operating environments to more

disadvantaged environments, stressed by weather conditions and/or adversary action.

Having identified the major issues, in the following section the Study Panel identifies the

key Mission Management gaps at the root of these issues.

Background

Four key operational issues are likely to limit RPA

mission effectiveness…

Operational Issues

…and the situation will get more complex with RPA

growth, expansion of roles, and more contested

Manning & Selection

operations

Airspace Deconfliction &

Management

Collateral Damage

Contested Environment

Operations

Figure 2-10: Summary of Operational Issues.

FOR OFFICIAL USE ONLY

Chapter 3: Major Findings

Following a process of data collection, interaction with subject matter experts,

and Panel deliberation, the Study Panel identified several key findings that hamper the

effective mission management of RPAs. As detailed below, in terms of technical

findings, the Study Panel found insufficient and inflexible platform automation, poorly

designed operator control stations, and limited communications systems. In terms of

personnel and processes, the Study Panel found unsubstantiated selection criteria, costly,

sub-optimal training, and lagging Concepts of Operations

(CONOPS) and Tactics,

Techniques, and Procedures (TTPs). These findings are now considered in detail.

3.1 Finding (1): Insufficient and Inflexible Platform Automation

The Study Panel’s first finding was insufficient and inflexible automation.

Current RPA automation implementations contribute to poor operator situation awareness

(SA), increased workload, and reduced effectiveness (Figure 3-1).

Current RPA automation implementations either under-automate,

over-automate, or fail to provide a flexible human-centric solution

Under-automation in key capability areas

Auto-takeoff/landing (MQ-1/9), dynamic route replanning

Sense and avoid

Auto target cueing/tracking; sensor-slaved flight control

Transfer of RPA control between OCSs

Overly-automated tasks

Predator fully automated mission employment mode never used

Global Hawk automated mission execution resists response to ad hoc tasking

High levels of automation reduce human operator vigilance, skills

Inflexible human-automation interaction deficiencies

Difficult to adjust automation levels to best fit mission need

No fine-grained control of automation

Abrupt transitions between automation levels disrupt missions ops

Lack of automation transparency/feedback decreases operator SA, safety

Figure 3-1: Automation Findings.

There is a lack of RPA automation in several key capability areas including

airspace integration, contested airspace operations, and single/multi-vehicle control by a

single operator. Examples include flight management (e.g., dynamic mission replanning,

vehicle handoffs, sense and avoid, threat awareness/avoidance, terminal area operations)

and sensor management (e.g., change detection, object identification and tracking in

complex environments, event detection, cross sensor target cueing, sensor fusion)

FOR OFFICIAL USE ONLY

functions. As a result, these functions often require a significant amount of time (e.g., 20

minutes to transfer vehicular control of Predators between ground stations) and manual

effort to accomplish (e.g., many individuals must simultaneously inspect full motion

video for targets). In some cases, required automation is absent (e.g., there is currently

no threat detection capability on AF RPAs and multi-aircraft control capability is very

limited).

Secondly, automation that does exist is often poorly designed for the variety of

missions experienced in dynamic real-world operations. Automation seems to have been

applied where it was most feasible rather than where it was most needed. Current

automation capability is narrowly focused and is often characterized by poor and

inflexible human-automation interaction. As an example, the Predator/Reaper can

theoretically perform entire missions in a fully automated manner, but this capability is

never employed because the automation cannot accurately sense or respond to dynamic

mission requirements. The Global Hawk is highly automated but requires significant

time to fully plan the mission before a flight, and this automation is very cumbersome to

override in flight in response to ad hoc tasking. In general, currently implemented

automation is not adjustable, so that the RPA operator cannot quickly customize a

response to a new situation by applying automation at a level (from no automation, to

semi-automated, to fully-automated) that provides the most effective performance for the

specific context at hand. In some cases, a fully automated response may be appropriate,

while in others the human operator may need to more directly interact with the system at

some level (e.g., information acquisition, situation assessment, decision making, action

implementation) to achieve the desired outcome. As a simple example, the Predator pilot

cannot control bank angle while in autopilot, resulting in suboptimal sensor standoff

range during loiter unless the pilot completely disengages the autopilot. Additionally,

operators are often not made aware of exactly what the automation is doing or planning

to do, and why (a lack of mode awareness, lack of shared situation awareness, etc.). Two

such examples involved Global Hawk altitude deviation incidents due in part to the

operator being unaware of how the automation was responding to the current situation.

This lack of automation transparency decreases operator SA and trust in the automation

and it increases the potential for decision-making errors that reduce mission effectiveness

or lead to unrecoverable events.

With exponentially increasing amounts of ISR data generated from current as well

as soon-to-be-fielded wide-area persistent sensors (e.g., Gorgon Stare, ARGUS-IS), the

need for sensor/image processing automation is greater than ever. If all this sensor

information could be transmitted for viewing

(itself a huge issue due to data link

transmission rates), the existing cadre of sensor operators and image analysts would be

impossibly overloaded. As a result, increased sensor and image processing automation

will be required to assist sensor operators and image analysts in rapidly collecting and

exploiting imagery to support upcoming IW missions. In addition, recognizing that

current and envisioned sensor hardware can/will produce data at rates of 10 to over 1000

times available communications data transmission rates, significant sensor/image

processing automation will likely need to be handled onboard the RPA. For example,

experience with manned platforms (e.g., Constant Hawk and Angel Fire), as well as with

ARGUS-IS components, provides evidence that automated change detection is a tool

which can be used to identify within frame motion, support image-driven tracking of

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

multiple objects, determine the need for increased/decreased frame transmission rates,

and determine the need for resolution adjustments in selected portions of the imaged

scene.

3.2 Finding (2): Poorly Designed Operator Control Stations

Current RPA OCSs do not promote effective, safe mission mgmnt

Not designed using accepted systems

engineering practices

Requirements/standards for Human-System

Integration (HSI), and usability processes for

RPAs nonexistent

Poor baseline for robust multi-aircraft

control (MAC) ops

Closed architectures fail to take advantage

of “best of breed” OCS component

solutions; and constrain options for training

simulation

Current OCS re-architecture efforts inadequately address:

Compliance with established HSI process

Human-automation interaction for future capabilities

Distributed mission ops, leveraging cross-platform commonalities

Figure 3-2: Operator Control Stations Findings.

The term “unmanned aerial vehicle (UAV)” is a misnomer, as illustrated by the

Air Force desire to emphasize the term “remotely piloted aircraft.” Although the crew is

no longer on board the platform, there clearly remains a critical need for human

involvement for RPAs to successfully perform missions today and well into the future.

This is especially true for complex, rapidly changing, and time-critical mission areas such

as tactical reconnaissance and close air support. Algorithms to enable autonomous RPA

operations in these mission areas do not exist due to a variety of reasons: many relevant

mission inputs are not digitized (and thus not available for input into these algorithms in

real-time); target/friendly/non-combatant identification is complex, uncertain and

variable; mission objectives vary constantly as does the ground situation; and decisions

often must take into account abstract considerations such as ethics and secondary/tertiary

effects of actions. Thus, even as automation matures, the human RPA operator will

remain a critical element of the system, with the sole link connecting the human mind to

the RPA being the operator control station (OCS).

However, for the three largest Air Force RPA systems (Predator, Reaper, Global

Hawk), these key subsystems remain largely unimproved from their initial engineering

prototype configurations developed under an ACTD 10-15 years earlier. As illustrated in

Figure 3-3, current operator interfaces are replete with interface deficiencies, including

inefficient and ineffective acquisition and processing of information, cumbersome control

FOR OFFICIAL USE ONLY

implementations, poor situation awareness, high workload at inopportune times, lack of

automation mode awareness, and inadequate ergonomics.

As summarized in Figure 3-2, current OCSs are highly proprietary, platform-

specific designs that do not follow a standard-based open architecture. This limits the

potential for interoperability and plug-and-play modularity, while increasing personnel

training time and costs. These OCSs (and those for smaller RPAs as well) were not

designed using accepted systems engineering practices including Human System

Integration and usability processes. These deficiencies have contributed to several

mishaps as well as lower mission effectiveness.¹⁹ As RPAs become more capable,

mission sets expand, and network capabilities increase, the OCS will likely become a

primary limiting factor towards mission success and safety unless immediate steps are

taken to improve this critical component. As one example, a key vision of the AF RPA

Flight Plan (and one often advocated by CSAF at recent briefings) is the ability of single

crews to simultaneously control (i.e., supervise) multiple RPAs. Current OCS designs

are not readily extendable to these multi-aircraft control operations²⁰, requiring

significant advances in mission management of critical information and aircraft control

systems.

Lengthy process to

Highly loaded visual channel

handoff vehicle control

Visualization is poor and

perpetuates mistakes

Add-on systems (i.e.,

Multiple screens require

Falconview) provide needed

significant mental

functionality but cannot

Lack of system

integration to obtain

interface directly with core GCS

feedback regarding

mission/battlespace

task completion

awareness

Varying methods to input

Too many inputs required

data

to implement commands

Narrow visual field of

(e.g., 22 key strokes to

view

turn on the auto pilot )

Limited alert cues to warnings

where messages can be

hidden

Extra workspace

Poor ergonomics

required

Multiple keyboard/input

Multiple separate

devices required

comm devices

No decision

Non-intuitive

aiding /support

Numerous

multilayered menus

technology

alphanumeric status

displays

Figure

3-3: Human System Integration

(HIS) Issues with Operator Control

Stations for Predator/Reaper.

¹⁹ Cooke, N. J., Pringle, H., Pedersen, H., and Connor, O.

(Eds.) (2006). Human Factors of Remotely

Operated Vehicles. Volume in Advances in Human Performance and Cognitive Engineering Research

Series, Elsevier.

²⁰ ACC MAC Report.

FOR OFFICIAL USE ONLY

Current acquisition efforts are now underway to improve the OCS of the three

largest AF RPAs (e.g., Predator/Reaper Block 50 Advanced Cockpit Program, and the

Global Hawk Ground Segment Re-Architecture Project). These efforts appear to be

implementing some human-computer interface improvements as well as a more modular

software architecture for the OCS. These efforts represent steps in the right direction, but

they still do not fully address the need for:

1) improved system automation

implementations with associated improvements to human-automation interaction;

future capabilities such as multi-aircraft control and globally-distributed RPA control

(beyond designing for physical connectivity); 3) the potential for commonalities between

major AF RPA OCSs; and 4) the need to follow an established Human-System

Integration (HSI) process in acquiring systems that are fully matched to the knowledge,

skills, and abilities of the human operator.

3.3 Finding 3: Limited Communications Systems

The Study Panel also found that limitations in RPA communications and

networking impede distributed operations, hinder integration into the NAS, and

perpetuate vulnerabilities in contested environments (Figure 3-4). The Panel identified

issues with robustness, security, interoperability, and bandwidth.

Limitations in robustness, security, interoperability, and scalability of RPA

comms and networking impede distributed operations, hinder integration

into the NAS, and perpetuate vulnerabilities in contested environments

Robustness

Lost-link events are uncommon, but limit NAS integration

Latencies in satellite links inhibit real-time, emergency C2

Vulnerability and lack of alternative C2 uplinks (and PNT) risks loss of control

Security

Lack of encrypted downlinks has enabled simple and low-cost interception of

sensor data. But sensitivity varies by mission

Interoperability, Distributed Operations, and Scale

PED is distributed, but C2 is “stovepiped”˘ limited distributed ops capability

Critical dependence on limited (and vulnerable) SATCOM bandwidth

Lack of a systems-level communication architecture increases manpower

demands and fails to anticipate future platform and sensor bandwidth growth

Figure 3-4: Limited Communications Systems.

3.3.1 Security and Robustness

The Study Panel found limitations in security and robustness in terms of lost-link

events, communication latency, and limited encryption.

FOR OFFICIAL USE ONLY

3.3.2 Lost-Link Events

Lost-link events are operationally uncommon,²¹ yet of primary concern for

operations in the NAS. Risk mitigation techniques include ground-based and air-based

sense and avoid

(GBSAA and ABSAA, respectively). GBSAA may require

augmentation of existing air surveillance radar with 3D position information, with an

estimated installation time of 2-3 years (in excess of $3M per site). A cooperative

ABSAA approach is better for lost-link events, as it only requires air-to-air

communications over relatively short distances. Current Traffic Collision Avoidance

System (TCAS) implementations are thought to be insufficient for ABSAA because of

limited accuracy and deployment, but Automatic Dependent Surveillance Broadcast

(ADS-B), required in all non-exempt aircraft by 2020, should suffice. Unfortunately,

ABSAA with uncooperative aircraft is estimated to be more than 12 years out. In 2012,

the World Radio Conference (WRC-12) will consider allocation of a reserved spectrum

for implementing high reliability UAS command and control (C2) links (agenda item

1.3). This offers the potential of reducing interference from other spectrum users.²²

Lost-link, if persistent, can lead to loss of mission effectiveness while the aircraft diverts

to a holding pattern or returns home.

3.3.3 Latency

A geosynchronous satellite (GEO), used today in RPA Remote Split Operations,

imposes at least 250 milliseconds latency. Actual latencies may be twice these values or

more because of queuing, framing, error correction, and coding, assuming only a single

satellite hop is required; a multi-satellite hop could incur in excess of a second latency.

Because interactive human-in-the-loop control activities require no more than about 200

milliseconds latency, any fine-grain rapidly interacting control of the platform, such as

flight control during takeoff or landing must be accomplished in one of two ways:

automatically via an on-board flight controller, or manually, via an operator located in the

area of operations and linked to the platform by a line-of-sight (LOS) communications

link (in this case, C Band comms) Because the USAF has chosen not to automate these

flight critical functions (takeoff and landing), personnel need to be placed in or near the

Area of Responsibility

(AOR), operating OCSs near the airfield, connected to the

platform by LOS communications links, at least for the takeoff and landing portions of

the mission.

Even if some flight critical functions were to be automated, there could still be

advantages gained in reducing latencies associated with CONUS-based OCSs, for

example, in collision avoidance maneuvers in heavy traffic, or in jinking maneuvers to

evade threats. Alternative satellite technologies (Low Earth Orbit [LEO], Medium Earth

Orbit [MEO]) offer substantially decreased latencies (potentially under 50ms), but few

Sanitate,

Guy, AF/A30-AOZ,

(U)

“UAS Link Vulnerabilities and Mitigation,”

16 Nov 09.

Lacher, A., Zeitlin, A., Maroney, D., Markin, K., Ludwig, D., and Boyd, J. "Airspace Integration

Alternatives for Unmanned Aircraft", AUVSI Unmanned Systems Asia-Pacific Conference, Feb 2010.

FOR OFFICIAL USE ONLY

such systems are readily available and they do not in general guarantee assured access.

Nonetheless, these could be used as alternative C2 channels for the full range of RPAs

(Iridium has been demonstrated on Global Hawk). Sensor data, if delay-tolerant, could

remain on GEO or use store-forward satellites since storage density has increased

dramatically in recent years. Additional autonomy and onboard processing/exploitation

could also relax the need to send sensor data (like full motion video [FMV]) with low

latency to a remote operator.

3.3.4 Encryption and Potential C2 Link Vulnerabilities

Historically, sensor/data downlinks for some RPAs have not been encrypted or

obfuscated. Unencrypted sensor data (e.g., FMV) is beneficial because the downlink is

used to feed ROVER systems used by Joint Terminal Attack Controllers (JTAC) and

other ground personnel, including uncleared coalition members and contractors. This is a

life-saving capability. Nevertheless, not protecting against interception of sensor data has

been criticized.^

“Fixing” this security issue by mandating NSA Type 1 encryption is

likely to lead to an unacceptable key management burden because of the large number of

users of RPA data that have a wide variety of access rights. However, commercial-grade,

NSA-approved cryptography is available (“Suite B”). Commercial cryptography of this

kind does not require the same degree of rigor in handling key material and encryption

devices, and is not limited in operation to cleared personnel. There is relevant

Department of Defense (DOD) activity in this general area.²³

Encryption has generally been used on C2 messages because the risks associated

with compromise are higher (loss of the vehicle), and there is a greatly reduced need for

sharing of the C2 data as compared with sensor data. However, crypto issues will likely

be exacerbated when doing coalition/joint swarming across platforms that require shared

C2 across security domains - a capability that is desired to fully exploit the potential of

networked RPA operations.

3.3.5 Interoperability, Distributed Operations, and Scale

The Study Panel found limitations in interoperability, distributed operations, and

scale as evidenced by stovepiped operations, insufficient satellite communications

(SATCOM) bandwidth, and the need for a long-term communication architecture.

^ And not just because of the press coverage of Iraqi insurgent downlink capture (e.g., “Insurgents Hack

U.S. Drones,” Wall Street Journal, 17 December 2009). Although debated, access to RPA sensor data and

accompanying meta-data can provide at least the following potentially sensitive information: detailed

imagery of wherever the RPA travels (including takeoff and landing sites), effectiveness of adversary

camouflage, behavior of weapon systems, TTPs in use, and “movies” of high-regret operations with

potentially significant IO implications.

²³ see ASD/NII DTM-09-004 [April 2009]

FOR OFFICIAL USE ONLY

3.3.6 Stovepiped Operations

Current RPA C2 systems tend to be stovepiped in the sense that they provide

limited options for interfacing current mission state with other distributed system

elements and, conversely, limited options for interfacing different types of OCSs with an

aircraft after the commencement of a mission. AF operators reported a 20-minute delay

is required to (manually) effect a “handoff” from one OCS to another. In addition, both

the communications and software architectures of several fielded RPA systems (e.g.,

Predator) are proprietary, and do not allow processing of C2 or sensor data feeds from

multiple, heterogeneous systems. The upshot of these deficiencies is that situation

awareness is reduced, and more personnel are required (than should be) to produce a set

of mission effects. If designs are less stovepiped, air assets, pilots, and analysts could be

treated as common resources across multiple RPAs and missions, leading to a form of

statistical multiplexing gain in operational efficiency.

3.3.7 SATCOM Bandwidth

Current Remote Split Operations are dependent on availability of satellite

communications, much of which is purchased from the commercial sector. Assuming no

radical departure from current practice is undertaken, future RPA-based sensor

technologies (e.g., Gorgon Stare, ARGUS-IS) will require bandwidth on the order of 100

times more than that needed by current RPA sensors. This presents a significant

challenge, as the current demand for SATCOM is already outpacing its availability.

Depending on the operating environment, there may be alternatives to SATCOM (e.g.,

airborne relays, laser communications). One of these includes Battlefield Airborne

Communications Node (BACN) as an airborne RPA communications relay, although this

would introduce yet another ground station to integrate and operate and another airborne

system to deconflict. In addition, sufficiently robust on-board sensor processing may

reduce the need for off-platform communications, thereby reducing the impact on

SATCOM demand.

3.3.8 Long-Term Communication Architecture

The current approach to RPA management has evolved based on an ACTD

process leading to the stovepiped systems the Air Force has today. Consequently, RPAs

and their corresponding OCSs are not easily interfaced with a broader distributed system.

This is most acute in the command, control, and tasking aspects of RPA management. At

present, sensor/intelligence products are managed in a somewhat distributed fashion (i.e.,

DCGS), yet C2 for one aircraft tends to be performed from a single OCS and handoff

between one OCS and another is a laborious process. Greater efficiencies can be

achieved if pilots, sensor operators, and aircraft could be dynamically tasked from one

operation/AOR to another. However, achieving this capability requires common message

sets, network interconnection, and resolution of potential security issues across multiple

domains, potentially including coalition members and possibly non-government agencies

(NGAs). In addition, intelligence analysts, pilots, and support personnel must be trained

FOR OFFICIAL USE ONLY

in TTPs that provide for tasking and mission requirements that may arise from any point

on the network.

3.4 Finding (4): Inadequate Selection Criteria and Training

The identified gaps in crew selection and training relate directly to RPA manning

issues and ultimately to mission effectiveness and safety. Legacy selection and training

processes, established for manned flight positions have, until recently, created an RPA

manning bottleneck. For instance, with the exception of the recent Beta program, RPA

pilots have been experienced USAF fighter/bomber or airlift/tanker pilots on ALFA^

tours.²⁴ Training has been costly and inefficient given that the ALFA pilots have been

trained on other weapons systems first (Figure 3-5). In fact, the 2008 USAF Audit

Report estimates that these legacy procedures, if continued, would cost the USAF over

$1.3 billion dollars in the years from 2009-2014. This inefficiency is not only costly, but

contributes significantly to the manning bottleneck.

The recent USAF transitioning of the Beta program to formal Undergraduate RPA

Training (URT) establishes an RPA career field and is a positive step in addressing

training costs and the manning bottleneck. However, there are other shortcomings in

crew selection and training that continue to contribute to manning issues and compromise

mission effectiveness and safety. Current selection of qualified candidates for an RPA

path is not based on perceptual, cognitive, motor, and team capabilities required for each

RPA position. Likewise, training programs, simulation capabilities, and training

assessment methods are not guided by Mission Essential Competencies (MECs) and are

not exploiting currently available training technologies.²⁵ MECs are

“higher-order

individual, team, and inter-team competencies that a fully prepared pilot, crew, or

command and control team, requires for successful mission completion under adverse

conditions and in a non-permissive environment.”²⁶

^ ALFA is a Vietnam era acronym identifying the types of positions a pilot might fill on their tour, such as

Air Education and Training Command

(AETC) instructor, Lead-in Course instructor, Forward Air

Controller, or Air Liaison Officers.

Jensen, R. M. and Slater, J. W., Jr. 2008. Unmanned Aerial System Pilot Force Management, Air Force

Audit Agency Report F2009-0005-FD4000.

Gramm, Joshua and Pappan, Steven, “Insatiable Demand: ‘Manning’ the US Air Force’s Unmanned

Aircraft Systems with Capable Pilots,” Harvard Kennedy School of Government, March, 2009.

Alliger, G.M., Beard, R.M., Bennett, W., Jr., Colegrove, C.M. and Garrity, M. 2007. Understanding

mission essential competencies as a work analysis method. AFRL-HE-AZ-TR-2007- 0034). Mesa AZ: Air

Force Research Laboratory, Human Effectiveness Directorate, Warfighter Readiness Research Division.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Establishment of Undergraduate RPA Training (URT) - formerly Beta - and the

RPA career field is a positive step; but a limited scientific basis exists for crew

selection and training

MQ-1 Gradesheet Performance

Selection criteria not based on

Traditional

SUPT

URT (Beta)

capabilities associated with each RPA

position

Beta program evaluation ongoing

20% washout

Beta class FTU performance equivalent to

Traditional, SUPT trainees

But operational squadron instructors

critical of Beta’s combat readiness

Cause of discrepancy unclear

First Beta class operating unsupervised

Inadequate simulation facilities,

especially for sensor operator training

No structured continuation training for

mission management skills

Common Syllabus Missions

Figure 3-5: Inadequate Selection Criteria and Training.

Gaps in selection and training are evident in the ongoing assessment of the Beta

pilots. There were two classes of 10 and 11 with initial washout of 20 percent in each

class leaving 17 Beta pilots total. As seen in the graph in Figure 3-5, the Betas’ Formal

Training Unit (FTU) performance based on instructor grade sheet scores is statistically

equivalent to that of other trainees (Traditional and Specialized Undergraduate Pilot

Training [SUPT[). At the same time, operational squadron instructors are generally

critical of Beta pilots’ limited air sense and combat readiness, and as a result, the Beta

pilots went on for additional supervised training after Mission Qualification Training

(MQT).²⁷ Although the precise cause for these discrepancies in assessment of the Beta

program are not known at this time, they point to: 1) the potential benefits of selection

that is targeted for known capabilities required of each position; 2) the need for a richer

set of assessment methods that meaningfully assess skills required; and 3) the need for

training that is customized for the MECs required of the RPA career path. These issues

are all tied to gaps in understanding the fundamental capabilities and skillsets of RPA

positions.

RPA training is also limited at the more advanced end of the training spectrum.

There is minimal exploitation of simulation training, Distributed Mission Operations

(DMO) exercises, or the use of Live, Virtual, Constructive (LVC) entities. Sensor

simulation fidelity is also low. Simulation for all RPA positions has been hindered by

delays in simulator development and limited access to proprietary vehicle and system

models and data. Additionally, there is no structured continuation training for RPA

operators. As a result of this training gap, acquisition of team communication and

coordination, advanced mission management skills, and other skills that could be

exercised in continuation training have been largely relegated to on the job training.

²⁷ Although, at the time of this writing, currently eight Betas are flying unsupervised missions.

FOR OFFICIAL USE ONLY

In sum, though the establishment of a formal RPA career path is a positive step,

selection and training for RPA operators has not kept pace with the state-of-the-art in

selection and training including the exploitation of MECs, simulation technologies, and

distributed mission operations. The potential consequences of such deficiencies in

operational environments are compromised mission effectiveness and safety.

3.5 Finding (5): Lagging CONOPS and TTPs

The Study Panel also found that late development of CONOPS and TTPS has

reduced the usefulness of RPAs (Figure 3-6). In the case of the Predator, initial delivery

resulted from an Advanced Concept Technology Demonstration

(ACTD) and the

CONOPS significantly lagged the physical delivery. As a result, the system’s operation

was very much on an ad hoc basis. Lack of sufficient logistic support reduced the

effectiveness of the system in its initial deployment to the Former Republic of

Yugoslavia (FRY), for example. The same pattern seems to have been repeated in the

deployment of the Reaper and, currently, with the Gorgon Stare sensor. There is also a

need for a CONOPS for distributed management to enable the same flexibility of

distributed operations currently enjoyed by exploiters given the distributed command

ground station (DCGS). The Panel strongly supports the concept of developing an initial

CONOPS well before deployment and using this CONOPS as the basis for equipage and

operation.

CONOPS/TTPs created after development have yielded systems

difficult to operate in the NAS, a challenge for full manned/

unmanned integration, and vulnerable in contested environments

Predator/Reaper acquisition did not transition to the more traditional

front-end requirements analysis and CONOPS development for full

integration into the force structure, resulting in

Continuing NAS integration issues, and lack of clear plan for manned/

unmanned integration in MCA

No vision for dealing pro-actively with what may become an increasingly

contested environment

Gorgon Stare appears to be on the same path, with CONOPS TBD by

ACC, after delivery

A lack of CONOPS exists for distributed mission management, which

will be another “technology push” response

Figure 3-6: Lagging CONOPS and TTPs.

3.6 Summary of Findings

RPAs have provided significant value to the Air Force; however, their full

potential is limited in several key mission gaps (Figure 3-7) including human-automation

interaction, OCS design, communications, selection and training, and CONOPS/TTPs.

Having identified and analyzed these limitations, the Panel next considers how to address

these shortfalls.

FOR OFFICIAL USE ONLY

Chapter 4: Recommendations

Having considered the implications of the above findings, the Study Panel

identified and prioritized a number of possible recommendations. The Panel focused on

eliminating mission management gaps through technology, people, and process

enhancements. In terms of technology, the recommendations focus on platform

automation/autonomy, enhanced Operator Control Stations

(OCSs), and robust

communications systems. In terms of people, the Study Panel recommends targeted

selection and more effective training. In terms of process improvements, the Study Panel

recommends enhanced operations through innovative Concepts of Operations (CONOPS)

and Tactics, Techniques, and Procedures (TTPs), and improved acquisition through

improved systems engineering and human systems integration.

4.1 Recommendation (1): Develop Flexible Levels of Automation

To fully achieve the capability envisioned in the Air Force RPA (Remotely

Piloted Aircraft) Flight Plan, flexible levels of automation must be developed in several

key areas involving both flight management and sensor management functions.

However, increased automation alone is not sufficient. Just as critically, new automation

functionality must be implemented in a user-centered, flexible manner to enable rapid,

situation-adaptive human-automation responses for improved mission effectiveness

across a wide range of conditions. Accordingly, the first recommendation of the Study

Panel is to enable flexible levels of automation as depicted in Figure 4-1.

Develop flexible levels of automation enabling situationally-adaptive

human interaction to improve single-platform, multi-platform, and

mission-level effectiveness

Near: Single-platform task-level automation [ASC]

Flight control/guidance

Monitor environment, threat, health/status

Auto target cueing

Visualizations and alert aids: automation transparency,

mode awareness

Mid: Single-platform mission-level automation [AFRL]

Cross-cueing, information fusion

Contingency management (eg, dynamic mission planning)

Environment and threat prediction; sense and avoid capability

Far: Multi-platform mission-level automation [AFRL]

Collaborative goal-based tasking to RPA teams

Assured autonomous behaviors (V&V)

Figure 4-1: Improved Platform Automation.

FOR OFFICIAL USE ONLY

In the near term, focus should be centered on single-platform task-level

automation aiding (OPR: ASC). Rather than rely on the traditional “left-over” principle

(automate as much as possible and whatever is left over is handled by the human),

choices on what functions to automate should be based on that which most benefits the

combined human-automation RPA system in completing the mission. Examples of task-

level automation enhancements may include flight control/guidance

(e.g., auto

takeoff/landing, lost-link behavior), monitoring of health and status systems and the

environment including threats, sense-and-avoid capability (in-flight), and improved target

cueing/tracking technology. With regard to improved human-automation design,

automation status and mode alerts must be clearly conveyed to the human operator

through intuitive visualizations so that the operator is always aware of what the

automation is doing or planning to do, along with the associated rationale.

In the mid term, emphasis should be extended to single RPA mission-level

automation aids (OPR: AFRL). The human operator will generally assume a higher

supervisory role and the automation will support greater aspects of the RPA mission.

Examples of automation capability include enhanced contingency management (e.g.,

complex dynamic mission planning with alternative courses of action, vision-based

navigation), information fusion/management from disparate sources, multi-target

detection, tracking and identification, and transparent terminal area operations.

Environment and threat prediction capability can also be addressed in the mid-term.

Additionally, new RPA human-automation interaction methods need to be designed,

evaluated and matured (i.e., delegation interfaces, adaptive automation, mixed-initiative

interactions) to flexibly implement multi-level automation (optimized to function) across

each of the above functions. The human operator will need fine-grained control over the

initiation and application of supporting automation. Levels of automation and assigned

authority need to be thoroughly researched, along with its effects on operator

performance and trust, to determine the best methods of flexibly applying these systems

to a wide variety of situations under differing levels of automation fidelity.

Because current and envisioned sensor hardware produces data at rates of 10 to

over 1000 times the expected and available communications data transmission rates,

more research is needed underlying the development of algorithms to process sensor data

onboard to tag only a small portion of the recorded “scenes” for further immediate

transmission and human analysis. Examples of enabling technologies underlying these

future algorithms include pixel-velocity-based image stabilization, optimal

foreground/backg1round transmission rates for compression, moving-target indicators

incorporating stability/compression advances, geo-registration aided by enabling

techniques such as multi-scale geometric analysis, view interpolation and morphing

algorithms, and continuing advances in miniaturized on-board processors such as

graphics processing units. These investments will also lead to reductions in manpower

associated with sensor data interpretation. These investments could usefully cover a

range of efforts from signal processing techniques to implementation across a range of

current and developing systems. Depending on the specific deployment scenario and

environment

(including adequate communications links), the automated analysis of

sensor data from distributed sensors could in many cases provide improvement in signal-

to-noise

(e.g., through image comparison algorithms), reduction in convergence and

identification times, and better geo-registration accuracy.

FOR OFFICIAL USE ONLY

Finally, in the far term, efforts should be placed on developing and implementing

multi-RPA mission-level automation

(OPR: AFRL). Collaborative automation is a

critical enabler at this stage, especially with regard to distributed manned-unmanned

teams. Effects-based tasking (as opposed to platform-based tasking) will become the

preferred method of RPA control, as RPA machine cognition (perception, intelligence,

etc) and associated autonomy increases. Humans operators will need to collaborate with

(and possibly negotiate with) increasingly intelligent automation associates in the

planning and conducting of complex, multi-asset mission tasking. Collaboration across

(both vertically and horizontally) the various command and control (C2) echelons will

also be enhanced, allowing real-time visibility, oversight, and coordination of RPA

operations across the battlespace. Swarming technologies should be explored for cheaper

RPA systems. The key to an integrated battlespace is shared situation awareness that can

be seamlessly handed from one platform/team to another as a situation changes. It will

also become critical to develop verification and validation methods for increasingly

complex RPA systems, which, on the one hand, will be engaging in increasingly

“emergent” behaviors to satisfy the developer’s goal for dealing with new situations in an

autonomous fashion, and which, on the other hand, need to assure the user of

“predictable” behaviors within some acceptable envelope of autonomy.

It is important to note that the themes associated with each category (near, mid,

far) are intentionally simplistic to convey an easily understood framework for the general

progression of automation capability. Certainly not all single-RPA automation issues

will be solved in the near term or even the mid term, nor will collaborative automation

development be entirely absent in the near term and mid term (for example, the task-level

automation enabling a micro RPA to navigate an urban landscape and land on a window

ledge at night while experiencing strong wind gusts during a thunderstorm is likely to

remain a very difficult challenge well into the long-term).

4.2 Recommendation (2): Enhance Operator Control Stations

The Study Panel recommends that Operator Control Stations (OCSs) be improved

by applying existing design methods and processes to design/develop new operator aids

to reduce manpower requirements, enable airspace integration as well as force

multiplication, increase interoperability, reduce potential for collateral damage, and

increase resilience in contested environments

(as detailed in Figure

4-2). This is

facilitated in the near term by providing immediate fixes to existing OCSs that address

critical operational needs, focusing on increasing the integrated presentation and

management of information, and streamlining control inputs

(OPR: SAF/AQC,

ACC/A8). Additionally the Program Offices must be properly manned

(and

appropriately trained) for developing and applying Human-System Integration (HSI)

processes to their respective programs.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

Fix Human Computer Interfaces (HCIs) of existing systems, develop

new ones following established Human System Integration (HSI)

processes, and invest in advanced S&T for future systems

Near: Apply quick HCI fixes to

existing systems [SAF/AQC, ACC/A8]

Mid: Develop a family of networked,

interoperable, open architecture

OCSs incorporating HSI standards

[ASC, AFRL]

Far: Advance S&T in collaborative

aids, multi-platform heterogeneous

teaming, goal-based control, mixed-

initiative systems that adapt/learn,

natural language processing, 3D

visualizations [AFRL]

Figure 4-2: Enhanced Operator Control Stations.

In the mid term, employ a comprehensive HSI process and usability design/testing

to develop a family of networked, interoperable OCSs with new non-proprietary,

standards-based OCS architectures with plug-and-play components for a full range of

functions that facilitate interoperability and modularity (OPR: ASC, AFRL). This effort

should be aligned with direction provided by the Office of the Secretary of Defense

(OSD) Unmanned Control Segment Working Group and other similar Department of

Defense (DOD) activities. Human Computer Interfaces (HCIs) should be optimized to

the intended function of the OCS such as multi-aircraft control (MAC) for supervising

several RPAs during benign operations and transit, tactical operations

(for rapidly

evolving, uncertain, time critical, and collaborative combat operations), or sensor

operation (focused solely on sensor management tasks). Efforts should also concentrate

on improved information integration and presentation (including temporal displays) that

support rapid decision making, attention directing cues that highlight high priority tasks,

improved decision aids, and naturalistic multi-sensory interaction technologies, each

tailored to the intended operating environment of the OCS (man portable/wearable, fixed

facility, vehicle-based, etc). Future HCIs must also be designed to maximize human-

automation interaction through better operator awareness of automation state and

projected plans, and flexible intervention methods to intervene when automation does not

accurately support mission need.

Finally, in the far term, advanced science and technology (S&T) research is

needed in the areas of customized operator interfaces tailored to individual differences;

collaborative aids for multi-platform heterogeneous teaming; goal-based control methods;

advanced mixed-initiative systems that adapt/learn based on changes in the mission,

environment, or operator functional state; natural language processing; tools supporting

anticipatory decision making; and novel 3D visualizations of the battle space (OPR:

AFRL). Additionally, even in the far term, the key to success will be to identify and

FOR OFFICIAL USE ONLY

apply the appropriate levels of human skill/attention to each mission task and provide

operators powerful and intuitive automation tools so that they can focus their full

attention at the mission execution level.

4.3 Recommendation (3): Create Robust Communications

The Study Panel recommends creating more robust communications supporting

RPA operations, focusing on assured communications, appropriate security, and a

globally interoperable communications architecture (Figure 4-3).

Focus on assured communications, appropriate security, and

globally-interoperable communications architecture

Near: Implement/accelerate short-term

fixes for security and improved

interoperability [ESC]

Mid: Increase communications diversity,

efficiency, and robustness [ESC]

Far: Develop a globally-interoperable

communications system architecture

which accommodates manned and

unmanned operations, growth in sensor

system capabilities, and movement to

networked operation of multiple,

heterogeneous coalition assets in a

contested environment [ESC, AFRL, AF/A6]

Figure 4-3: Create Robust Communications.

4.3.1 Near-Term Fixes for Security and Improved Interoperability

In the near term, the Study Panel recommends implementing/accelerating near-

term fixes for security and improved interoperability (OPR: ESC). This would include

appropriate cryptography, CONOPS/TTPs for communications alternatives such as

relayed operations, and standardized, documented, and accessible networked interfaces

for ground elements. More specifically, we recommend the following.

 Cryptography: In April 2009, ASD/NII DTM-09-004 called for use of FIPS 140-

2 Security level 1 cryptography or better for unmanned aerial systems (UAS)

Groups 2-5. It also called for the establishment of a UAS Encryption Working

Group

(UASEWG) to coordinate implementation and make additional

recommendations. We recommend continuation of the UASEWG’s work, and

acceleration where possible. RPAs should in general be capable of operation

using stronger forms of cryptography if deemed necessary, based upon mission

objectives, though acknowledging that Type

1 cryptography may frustrate

coalition operations due to the additional protective measures required.

FOR OFFICIAL USE ONLY



CONOPS/TTPs for Relayed Operations: There are numerous platforms that

offer the potential to provide communications alternatives for RPAs. For

example, the Battlefield Airborne Communications Node

(BACN) package

currently being fitted to Global Hawk Block 20 provides numerous capabilities

that could be used to provide C2 for RPAs with less-advantaged communications

(such as Predator/Reaper). Although the technical means for doing so appears to

be available, CONOPS and TTPs are needed to exploit this capability further.

Relaying provides (at least) two potential benefits: as an alternative to denied

satellite communications

(SATCOM), and as a potentially lower-latency

communication path that enhances the effectiveness of remote operator controls.



Standardized, Documented, and Accessible Networked Interfaces for

Ground Elements: There are numerous ground elements involved in RPA

operations including the primary control stations (e.g. MCE for Global Hawk or

OCS for Predator), landing and recovery elements (LRE), and various ground

terminals used primarily for viewing sensor data within a line-of-sight (LOS) area

(e.g., ROVER and OSRVT). In addition, there is at present an effort to design a

new generation of interoperable OCSs being pursued by the UAS Control

Segment Working Group

(UCS-WG). In the short-term, we recommend

retrofitting existing OCSs with a physical network port (e.g., 100Mb/s Ethernet)

capable of providing (or possibly consuming) essentially all sensor and mission

meta-data using documented message sets over a TCP/IP based transport. Such a

port needs to be “accessible” in the sense that appropriate accreditation must be

provided to allow these messages to be processed by other equipment. Note that

as a short-term recommendation, we do not require the application-layer messages

to conform to a particular standard, but rather that they be accessible and

documented.

4.3.2

Mid Term: Increase Communications Diversity, Efficiency, and

Robustness

In the mid term, the Study Panel recommends increasing communications

diversity, efficiency, and robustness (OPR: ESC). This would include evaluating the

utility of potential solutions, including gateways, additional C2 diversity, and disruption

tolerant communications, as well as the utility of state-of-the-art sensors incorporating

on-board processing/exploitation, and non-geostationary satellites for C2 and sensors.

 Evaluate Gateways, Additional C2 Diversity, and Disruption Tolerant

Communications: At present, there are a variety of communication capabilities

available for each RPA type. For the less capable RPAs, SATCOM vulnerability

is a significant concern. To provide additional communication options, we

recommend evaluating a broad set of gateway packages, including BACN and the

Defense Advanced Research Projects Agency’s (DARPA) MAINGATE program,

in addition to equipping existing communications-limited RPAs with additional

link capabilities, including Common Data Link (CDL) and Link-16. Increasing

the set of link options improves resiliency in contested environments (e.g., via the

anti-jamming properties of Link-16) and reduces latency for other assets within

FOR OFFICIAL USE ONLY

the LOS area that wish to interact with the RPA. Size, Weight, and Power

(SWAP) may be a concern with some of the additional radio systems, but modern

radios have addressed this concern (e.g., L-3’s “mini-CDL” 200 radio weighs

1.4lbs and uses 25-45W). Disruption Tolerant Networking (e.g., the DARPA

DTN and WNaN programs) uses storage within the network to help overcome

modest link or long-term link outages, especially for environments that do not

require precise latency bounds for effective operation.



State-of-the-Art Sensors with On-Board Processing/Exploitation: The Study

Panel recommends improving algorithms for compression and sensor exploitation

to limit off-platform bandwidth demands. In addition, on-board cross-sensor

processing can be used to enhance resilience to sensor attacks

(e.g.,

communications jamming or GPS spoofing) by performing a secondary or tertiary

check on the validity of navigational and platform health information. In general,

algorithms produce better results when provided higher-quality input (e.g., sensor)

data. Thus, although the highest-resolution sensors may overwhelm certain (e.g.,

SATCOM) communication links, such sensors can lead to better autonomy with a

higher degree of trust, thereby reducing the need for human operators to be shown

all sensor data. Large onboard storage can be used to house raw sensor data for

subsequent, on-the-ground, exploitation. For example, if it is assumed that

uncompressed HDTV-quality video requires a storage rate of

1Gb/s

(conservative), a 20-hour mission would require approximately 9 Tb of storage.

With single drives now capable of storing 2.5 Tb, a modest amount (4 drives or

fewer) of onboard storage would be required.



Non-GEO Satellites: The Study Panel recommends evaluating options for non-

geostationary satellites for RPA C2, as well as for low-data-rate sensor backhaul.

There are two potential benefits, including lower latency (potentially allowing for

more effective tactical control of the aircraft) and greater diversity/assuredness.

In particular, if conventional GEO-based SATCOM were compromised or made

unavailable, LEOs may provide a viable option, at least for C2. Unfortunately,

there are a limited number of LEO constellations available

(e.g., Iridium,

Globalstar, Orbcomm). Nonetheless, encouraging results, including latencies

below 100 milliseconds, have been observed (e.g., the National Aeronautics and

Space Administration

[NASA] “Flight Modem” work²⁸). Note that Iridium

announced plans for its next-generation satellite constellation (Iridium NEXT) on

June 2, 2010.

4.3.3 Far Term: Develop RPA Communications Architecture

In the far term, the Study Panel recommends developing a globally interoperable

communications system architecture which accommodates manned and unmanned

²⁸ Morgan et al., “Telemetry Tracking & Control (TT&C) - First TDRSS, then Commercial GEO & Big

LEO and Now Through LEO,” 2001 (Document ID 20010020063)

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

operations and anticipates growth in sensor system capabilities for networked operation

of multiple, heterogeneous coalition assets, operating in a contested environment (OPR:

ESC, AFRL, AF/A6). The Study Panel recommends that a comprehensive

communications architecture study be undertaken to determine what interfaces, TTPs,

and capabilities are required to produce a fully-distributed network of RPAs, pilots,

analysts, and maintainers. At present, there are a collection of stovepiped systems with

limited interoperability and proprietary formats, leading to consequent difficulty in

performing distributed control.

4.4 Recommendation (4): Develop Targeted Selection and

Effective Training

The Study Panel recommends developing targeted selection and effective training

(Figure 4-4). The Air Force should understand fundamental capabilities and skillsets

required for each RPA specialty, use this knowledge to become more effective at

selection, and exploit current training and simulation technologies to reduce manning

pipeline delays, better enable NAS/MCA integration, and deal with future contested

environments.

Understand fundamental capabilities and skill-sets required for each

RPA specialty, select appropriately, and exploit current training and

simulation technologies to reduce manning pipeline delays and deal

with future contested environments

Performance Measurement and

Live

Tracking of Live and Virtual

Near: Reduce manning

bottleneck by targeted

selection and accelerated

training [AF/A1, AETC, AFRL]

Mid: Develop and use

training simulators more

effectively [ASC, AETC, ACC]

Virtual

Far: Develop skills in

team coordination,

communication and mission

Integrated live and

planning/After-Action Review

virtual

Constructive

performance-

[AFRL]

based debrief

Figure 4-4: Training and Selection Recommendations.

Improving available training and selection technologies can result in tremendous

cost savings and improved efficiencies in RPA training, while at the same time

addressing the manning bottleneck and generating a future RPA force capable of dealing

with contested environments. It is necessary to understand fundamental capabilities and

skillsets required for each RPA specialty, train and select accordingly, and exploit current

training and simulation technologies.

FOR OFFICIAL USE ONLY

Near-term activities should be directed at developing scientifically-based RPA

career path enhancements that generate targeted selection procedures and accelerated

training programs. The new Undergraduate RPA Training (URT) classes should be

appropriately selected, based on validated test methods and assessed in terms of training

and operational performance using objective state-of-the art performance metrics. In

addition, Mission Essential Competencies (MECs) should be used to optimize URT,

Mission Qualification Training (MQT), and continuation training and selection, with

independent paths developed for RPA pilots and sensor operators. (OPR: AF/A1, AETC)

Mid-term recommendations focus on the increased use of existing simulation

technologies, as well as the development and application of new technologies to increase

training effectiveness for non-LRE pilots and sensor operators. The use of flight

simulators, including part task trainers, should be optimized for all task levels (e.g.,

individual, team, multi-team). Sensor operator training should also be enhanced with

simulation, drawing from existing gaming technologies, Google Earth, and actual

operational video footage. It is important to recognize that the simulators must be

developed concurrently with an improved OCS. A simulator that replicates a poor

human-machine interface is a poor foundation for training. (OPR: ASC, AETC, ACC)

In the far term, RPA skills in team coordination, communication, and mission

planning/after action review (AAR) need to be developed. This can be accomplished via

the development and use of an advanced RPA training and simulation facility for

comprehensive continuation training and training research. In this context, simulation

training for RPAs should be advanced to the current state-of-the art in simulation for

fighter training, so that RPA operators can engage in multi-person, multi-vehicle

exercises on par with Distributed Mission Operations (DMO) exercises. Through the

advanced use of simulation, requisite skills can be trained for current and future RPA

operations to include: CONOPS in realistic manned/unmanned, joint, and contested

environments, team coordination, communication and mission planning, and after-action-

review skills.

(OPR: AFRL)

4.5 Recommendation (5): Improve CONOPS and TTPs

To more effectively drive mission effectiveness, the Study Panel recommends

development of RPA CONOPS and TTPs concurrently with system development,

considering all components including the payload, platform, communications, OCS,

Squadron Operation Center (SOC), Wing Operation Center (WOC), and Air and Space

Operations Center

(AOC) (Figure

4-5). To support this development, the Panel

recommends the use of modeling and simulation, operator-in-the-loop experimentation,

and performance evaluation. (OPR: ACC)

In the near-term, the Air Force should formalize the concept of distributed global

operations. The Air Force should define how current SOC/WOC/AOC structures can

enable distributed global operations via “brokering” among operators, platforms/sensors,

and

“customers.” This would enable increased efficiency and effectiveness if

“customers” could agree on a prioritization scheme which allowed for shifting assets to

the areas of greatest need.

FOR OFFICIAL USE ONLY

In the mid-term, the Air Force should develop CONOPS/TTPs that anticipate

growth in RPA operations, sensor capabilities, integrated manned/unmanned operations,

and contested operating environments. Development of models and simulations should

be pursued so that conjectural (“what-if”) analysis can be performed. This is particularly

important as the Air Force considers ways to operate new platforms and payloads in

mixed manned/umanned and contested environments.

In the far-term, the Air Force should enforce the necessary acquisition and

development discipline to ensure that CONOPS and TTPs are co-evolved with new

systems. This should be done as part of a well-structured systems engineering process

which leads systems development and informs systems requirements, taking into

consideration not just performance but also many of the standard “ilities”, including:

reliability, maintainability, and affordability.

Develop RPA CONOPS and TTPs concurrently with system

development, considering all components (payload, platform,

comms, OCS, SOC, WOC, AOC…) to drive mission effectiveness.

Use M&S, operator-in-the-loop experimentation, and performance

evaluation [ACC]

Near: Define how current SOC/WOC/AOC

structures can enable distributed global

operations via “brokering” among operators,

platforms/sensors, “customers”

Mid: Develop CONOPS/TTPs that anticipate

growth in RPA operations, sensor capabilities,

integrated manned/unmanned operations, and

contested operating environments

Far: Co-evolve CONOPS/TTPs for new platforms

and missions as part of a focused acquisition

process, leading systems development and

informing systems requirements choices

Figure 4-5: Improved Operating Concepts, CONOPS, and TTPS.

FOR OFFICIAL USE ONLY

4.6 Recommendation (6): Ensure Effective ACTD Transition

The Study Panel recommends the Air Force ensure effective transition of

successful RPA prototypes to the acquisition process (Figure 4-6), incorporating systems

engineering practices including HSI, in accordance with AFI 10-601 and 63-1201 (OPR:

SAF/AQ). The Air Force should continue successful use of the ACTD/JCTD (Advanced

Concept Technology Demonstration/Joint Capability Technology Demonstration) process

to develop game-changing solutions and transition them when appropriate. They should

upgrade, develop, and acquire RPA systems using accepted Systems Engineering

practices.

This includes employing Section

804 of the

2010 National Defense

Authorization Act which directs the Secretary of Defense to develop and implement a

new acquisition process that follows best practices to include “(a) early and continual

involvement of the user;

(b) multiple, rapidly executed increments or releases of

capability; (c) early, successive prototyping to support an evolutionary approach; and (d)

a modular, open-systems approach.”²⁹ Starting with a clear vision of the CONOPS and

accounting for operations in contested environments are essential. It is also important

that the Air Force establish RPA-specific HSI requirements (OPR: A5R) and guidelines

(OPR: AFRL). Operator-related issues should be addressed via incorporation of HSI as

part of the systems engineering process (OPR: AFMC), following the recommendations

of two previous SAB studies.^30,

³¹ The remainder of this section discusses this

recommendation in further detail.

²⁹ Further details of the new acquisition approach can be found in the March 2009 Defense Science Board

Task Force on Department of Defense Policies and Procedures for the Acquisition of Information

Technology.

³⁰ United States.

“Unmanned Aerial Vehicles in Perspective: Effects, Capabilities, and Technologies.

Volume 1: Summary (Public Released)” SAB-TR-03-01 September 2003.

United States.

“Human-System Integration in Air Force Weapon Systems Development and

Acquisition,” Executive Summary and Annotated Brief, SAB-TR-04-04, July 2004.

FOR OFFICIAL USE ONLY

Transition successful RPA prototypes to the acquisition process,

incorporating Systems Engineering practices including HSI, in

accordance with AFI 10-601 & 63-1201 [SAF/AQ]

Continue successful use of ACTD/JCTD process

to develop game-changing solutions and

Experimentation

transition when appropriate

Upgrade, develop, and acquire RPA systems

using accepted Systems Engineering practices

(2010 NDAA Section 804)

CONOPS

Establish RPA HSI requirements [A5R] and

guidelines [AFRL]

Address operator-related issues via

incorporation of HSI as part of the Systems

Technology

Engineering process [AFMC]

Figure 4-6: Effective ACTD Transitions for RPAs.

The ACTD process was initiated to permit the early and inexpensive evaluation of

mature advanced technology to meet the needs of the warfighter. The evaluation is

accomplished by the warfighter to determine military utility before a commitment is

made to proceed with formal acquisition. ACTDs also allow the warfighter to develop

and refine operational concepts to take full advantage of the new capability. Upon

conclusion, a successful ACTD may leave behind a residual operational capability. The

capability can be replicated, if only a few are required, or can be transitioned into the

appropriate phase of formal acquisition.

The ACTD program is a pre-acquisition activity that allows warfighters to use and

assess leading-edge capabilities. ACTDs are an integral part of reforming and

revolutionizing the acquisition process. Success should be measured by the ability to

quickly meet the warfighter's critical needs and by the extent to which ACTDs provide a

vehicle for new concepts and technologies to be evaluated before acquisition decisions

are made. Figure 4-6 highlights ways to ensure effective transition for RPAs.

ACTDs continue to provide a valuable streamlined approach to demonstrate

feasibility and potential of emerging capabilities. They strike the right balance between

fidelity and affordability for the intended purpose. The ACTD process and approach

does not, however, apply well to the system engineering process required in complex

acquisition programs. These acquisition processes are iterative; they take process inputs

such as customer needs through exhaustive system design, requirements, and analysis

loops while balancing all these needs through an overarching system analysis and control.

The result is a decision database highlighting system configuration, and project baselines.

Nearly all current RPA procurements are the product of ACTD or ACTD-like

approaches that have never benefited from the rigor of the systems engineering process.

By simply continuing or extending the ACTD approach, the RPAs in use today

completely missed or received limited engagement in the areas of CONOPS development

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

and application, HSI, systems engineering, and sustainment. Understanding the user

environment and mission objectives has been limited by lack of upfront user CONOPS

input as exemplified in the DARPA Technology CONOPS Double Helix approach

depicted in Figure 4-7.

Understanding of 2^nd and

Techniques, Tactics and

New CONOPS

3^rd order effects

Procedures; Training

Methodology

Figure 4-7: CONOPS Technology Double Helix.

With the RPA fleet, CONOPS development significantly lagged behind

development, and resulted in mission performance deficiencies and inefficiencies.

Systems engineering across these multi-element systems did not conform to DOD 5000

standards and led, in the case of Predator/Reaper, to fielding a closed architecture that is

sole source

(due to lack of data rights), unresponsive to meeting new operational

requirements, and expensive to change. As a related issue, and as noted earlier,

significant HSI deficiencies continue to exacerbate the development of CONOPS.

ACTDs also typically do not focus on sustainment engineering and associated life cycle

cost trades. This situation delayed deployment and sub-optimized ongoing logistics and

support functions that continue to detract from mission effectiveness.

A more effective path is to use ACTDs for their intended purpose and cleanly

transition to DOD 5000.02 after the decision to transition the demonstrated capability

into a program of record is accomplished. ACTDs are not intended or structured to

deliver production capabilities and should not be expected to do so. Two excellent

examples of successful ACTD transitions are the Have Blue to F-117, and the

Lightweight Fighter (YF-16) to the F-16 program.

Importantly, ACTC/JCTDs are fertile ground for groundbreaking, game-changing

technologies. Without doubt, they are key sources of innovation; however, successful

RPA prototypes must be transitioned into the acquisition process, incorporating Systems

Engineering practices, including HSI in accordance with AFI 10-601/63-1201. To

achieve their full potential, RPA programs must start with a clear vision of use and

CONOPS, and anticipate operations in contested environments.

FOR OFFICIAL USE ONLY

Chapter 5: Analysis and Summary of Recommendations

This report began with a number of identified RPA issues (Chapter 2) associated

with manning and selection, airspace deconfliction and management, collateral damage,

and contested environments. We then identified mission management gaps contributing

to these issues (Chapter 3), and put forth recommendations as to how these gaps might be

eliminated or reduced with appropriate investments in technology, people, and processes

(Chapter 4). We now present two sets of matrices, one showing the impact of mission

management gaps on the operational issues, and the second showing the impact of

technology investments on closing the mission management gaps. Although the impact

assessments are qualitative, the matrices clearly illustrate that dealing with the issues

requires a multi-pronged approach, across multiple dimensions.

This chapter concludes with a brief summary of how each of the issues can be

dealt with directly with a set of recommended actions, to provide a more compact

summary of the study’s overall findings and recommendations.

5.1 Impact Analysis of Recommendations

This section presents a qualitative assessment of the impact of the Study Panel’s

recommendations on identified gaps. Importantly, this analysis did not consider cost or

quantitative impacts of various recommendations, but rather reflects expert judgment by

the Panel based on individual and Panel expertise and experience. The assessments

reflect the collective best judgment given the short study time. Accordingly, the Panel

caveats its analysis as not being a formal life cycle cost/benefit analysis, which would

require further effort.

The Study Panel’s analytic method included considering the effects of closing the

identified mission management gaps discussed in Chapter 3, in terms of ameliorating the

operational issues articulated in Chapter

(i.e., manning and selection, airspace

deconfliction and management, collateral damage, and contested environments). Figure

5-1 depicts a matrix illustrating the benefits of addressing known mission management

gaps (rows) on the operational issues (columns). For example, investments in solving the

mission management gaps associated with automation, Operator Control Station (OCS)

design, communications, selection/training, and CONOPS/TTP

(Concepts of

Operations/Tactics, Techniques, and Procedures) will improve the identified operational

issues. Each cell in the matrix is numbered and colored as green (high impact), light

green

(medium impact), and white (little or no impact) to indicate the degree of

contribution an investment has in terms of correcting the operational issue. A

justification for each assessment is captured in Appendix B, indexed by the cell number

for ease of access. Rows indicate how improvements affect issues: thus cells 1 through 4

indicate how flexible levels of automation will contribute to the four operational issues.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

For example, cell number 2 indicates how investments in automation will improve

airspace deconfliction and management.

█ High

Airspace

Contested

█ Med

Manning &

Collateral

Deconfliction &

Environment

Selection

Damage

█ None

Management

Operations

Flexible Levels of

1-H

2-H

3-H

4-H

Automation

Enhanced OCS

5-H

6-M

7-M

8-H

Robust

Communications

9-M

10-H

11-M

12-H

Systems

Targeted

Selection &

13-H

14-M

15-M

16-H

Training

Improved

17-M

18-M

19-M

20-H

CONOPS & TTPs

Figure 5-1: Effect of Addressing Mission Management Gaps on Operational

Issues.

Figure 5-2 depicts a matrix illustrating the benefits of investing in technology,

people, and processes (shown down columns) to fill known mission gaps (shown across

in rows). That is, investments in solving automation/processing, OCS, communications,

selection/training, CONOPS/TTPs, and Advanced Concept Technology Demonstration

(ACTD) transitions will address the identified mission gaps such as the need for flexible

automation. Each cell in the matrix is numbered and colored as green (high impact), light

green

(medium impact), and white (little or no impact) to indicate the degree of

contribution an investment has to correcting a mission gap. A justification for each

assessment is captured in Appendix B, indexed by the cell number for ease of access. For

example, cells 21 through 25 indicate that implementing the Study Panel’s mid-term

enhanced OCS recommendation will significantly improve all gap areas, from enabling

flexible automation to improving future CONOPS/TTPs. In particular, employing a

cohesive Human-System Integration

(HSI) process and usability design/testing to

develop a family of networked, interoperable OCSs with new non-proprietary, standards-

based OCS architectures with plug-and-play components for a full range of functions will

enable automation, facilitate interoperability and modularity, improve networked

communications, ease training, and support and foster novel CONOPS/TTPs. As is

evident, all the main diagonals between major columns and their respective mission

management gaps are equivalent (e.g., automation for automation), and are therefore left

out of the detailed analysis.

FOR OFFICIAL USE ONLY

FOR OFFICIAL USE ONLY

█ High

Technology, People and Processes

█ Med

Robust

Automation

Enhanced

Selection

CONOPS and

ACTD

█ None

Communicati

& Processing

OCSs

and Training

TTPs

Transition

ons

Flexible

Automation/

1-H

2-H

3-H

4-N

5-H

6-M

7-N

8-M

9-M

10-N

11-M

2-M

13-H

14-M

15-H

16-M

17-H

18-M

Processing

System Based

19-H

20-H

21-H

22-H

23-H

24-H

25-M

26-N

27-H

28-N

29-M

30-N

31-H

32-H

33-H

34-H

35-H

36-H

Design of OCS

Networked

37-M

38-H

39-H

40-N

41-H

42-N

43-M

44-M

45-H

46-N

47-N

48-N

49-H

50-H

51-M

52-M

53-H

54-M

Communicatio

nsSystems

Targeted

55-M

56-M

57-M

58-M

59-H

60-M

61-N

62-N

63-M

64-H

65-H

66-H

67=M

68-N

69-N

70-M

71-M

72-M

Selection &

Training

Future CONOPS

73-M

74-H

75-H

76-M

77-H

78-M

79-N

80-M

81-H

82-N

83-M

84-H

85-H

86-H

87-H

88-H

89-H

90-M

& TTPs

Figure 5-2: Technology, People, and Process Investments Fill Mission Gaps.

5.2 Summary of Recommendations

This chapter concludes with a brief summary of how each of the key issues can be

dealt with directly with a set of recommended actions.

5.2.1 Manning and Selection Recommendations

The Study Panel recommendations directly addressing manning and selection,

shown in Figure 5-3, are grouped in four main categories. First, the Air Force should

develop more automation and smoother transitions between hands-on and hands-off

operation. Ways to implement this would include, for example, enabling multi-aircraft

control

(MAC), especially during low activity and boring mission segments (a side

benefit would be the relief of operators from loss of vigilance caused by boredom).

Automation could also provide cross-cueing or tagging and tracking of targets, reducing

or eliminating manually intensive processing; or to make mission planning easier and

more rapid. Finally, automation can help implement automated landings, a cause for as

much as 20 percent of RPA loss. Second, the Air Force should develop better OCSs

which would simplify mission handoffs, enable multi-aircraft operations, improve overall

SA with better integrated displays and via automatic recognition of key targets, and

enable target hand-off between assets (manned or unmanned). Third, the Air Force

should have training address sensor operators, exploiters, and maintenance personnel, as

well as pilots. Training and training simulation need to incorporate both manned as well

as unmanned missions, as well as training for contested environments. Finally, the Study

Panel recommends continuing RPA career path development.

FOR OFFICIAL USE ONLY

Manning: Recommendations

Develop more automation and smoother transitions between hands

on and hands off operation

Allow for multiple aircraft operation / Relieve the boring times

Make tag / track targets and mission planning easy/quick

Implement automated landings / remote landings

Develop OCS that simplify mission handoffs

Enable multi-aircraft operations

Include overall SA and automatic recognition of key targets

Enable target hand-off between assets (manned or unmanned)

Training must address sensor operators, exploiters and

maintenance personnel as well as pilots

Affects manned as well as unmanned missions

Train for contested environments

Continue RPA career path development

Figure 5-3: Manning and Selection Recommendations.

5.2.2 Airspace Deconfliction and Management Recommendations

The Study Panel recommendations directly addressing airspace deconfliction and

management, shown in Figure 5-4, are grouped in four main categories. First, in terms of

autonomy, the Air Force should focus on technology investments (“see and avoid”

technologies and others) which meet safety requirements for Group 4-5 RPA flight in the

National Airspace (NAS) (those vehicles that can carry the necessary payloads while

maintaining their mission utility) Second, in the area of robust sensing and

communications, the Air Force should focus future technology investments for Group 1-2

small RPAs towards enhancing communications and sensor technologies to enhance

combat mission capability in the Military Controlled Airspace

(MCA), where risk

management margins have greater flexibility. Third, in terms of policy and CONOPS,

the Air Force should accelerate DOD/DHS/FAA (Department of Defense/Department of

Homeland Security/Federal Aviation Administration) joint efforts in Executive

Committee to formalize national standards which mitigate risks of lost-link and see and

avoid in the NAS for Group 4-5 RPAs. Finally, in the area of acquisition, the Air Force

should pursue formal DOD acquisition programs, in place of the ACTD approach, for

next generation RPAs, to deliver an integrated system of systems type approach to ensure

unrestricted and integrated flight in the NAS, MCA, and in the International Civil

Aviation Organization

(ICAO) with manned aircraft

(although the Panel also

recommends that ACTD efforts still be pursued to explore and evaluate promising and

potentially game-changing technologies in the RPA “space”).

FOR OFFICIAL USE ONLY

Airspace Management & Integration:

Recommendations

Enhance airspace management via:

Autonomy - Focus USAF technology investments (AFRL “see and avoid”

technologies and others) which meet safety requirements for RPA flight in

the NAS to those large aircraft in Group 4/5 that can carry these payloads

while maintaining their mission utility (AFRL)

Robust Sensing and Communications - Focus future USAF technology

investments for Group I/2 small RPA towards enhancing comm/sensor

technologies which enhance combat mission capability in MCA where risk

management margins have greater flexibility (AFRL)

Policy/CONOPS - Accelerate DOD/DHS/FAA joint efforts in Executive

Committee to formalize national standards which mitigate risks of “Lost

Link” and “See and Avoid” in NAS for Group 4/5 RPA (SAF/AQ)

Acquisition- Pursue formal DOD acquisition program vice ACTD

approach for Next Generation RPA which deliver integrated system of

systems type approach to ensure unrestricted and integrated flight in the

NAS, MCA, and in the ICAO with manned aircraft (SAF/AQ)

Figure 5-4: Airspace Management and Integration Recommendations.

5.2.3 Collateral Damage Recommendations

While the Study Panel found that collateral damage is not an RPA-unique issue,

RPA capabilities in IW can help minimize the problem. First, persistent wide-area

surveillance helps improve warfighter situation awareness. It also can enable tactical

patience, a key enabler in limiting civilian casualties. Increased connectivity of split

remote operations results in more “eyes on target” in the CONUS ground stations and

more opportunities for data fusion from available feeds. Increased sensor fidelity can

enhance positive identification, the lack of which is a key cause of civilian casualties.

Finally, the employment of low collateral damage/directed lethality munitions on RPAs

can help limit civilian casualties.

Contested Environments

The Study Panel recommendations directly addressing contested operating

environments, shown in Figure 5-5, are grouped in two main categories. The Air Force

can improve mission robustness in contested environments via increased platform

redundancy.

Enablers include low-cost design and airframe/engine/electronics

manufacturing (as envisioned by the DARPA META and META-2 programs), multi-

vehicle mission planning and management concepts, including pseudo-random auto-

jinking, swarming, short-range jam-resistant inter-platform communications, and cyber

intrusion protection exploiting redundancy-based methods. Second, for highly contested

environments, the Air Force should conduct comprehensive systems-level engineering

analyses to establish effective design tradeoffs that balance the cost of passive/active

protection measures verses the cost of platform or sensor loss. This could include low

FOR OFFICIAL USE ONLY

observable (LO) measures (including acoustic, visual, and radar), on-board dynamic

mission planning/re-planning for threat avoidance, redundant links with electronic

protection (EP) waveforms to mitigate jamming of communications, automated self-

protect electronic warfare (EW) tactics, and alternative navigation techniques to reduce

effects of GPS loss.

Contested Environment Operations:

Recommendations

Improve mission robustness in contested environments via

increased platform redundancy. Enablers are:

Low-cost design and airframe/engine/electronics manufacturing (DARPA META

and META-2 programs)

Multi-vehicle mission planning and management concepts, including pseudo-

random auto-jinking, swarming

Short-range jam-resistant inter-platform communications

Cyber intrusion protection exploiting redundancy-based methods

For denied environments, conduct comprehensive systems-level

engineering analyses to establish proper design point between

passive/active protection measures, and against cost of loss

LO measures including acoustic, visual, and radar

Dynamic mission planning/re-planning for threat avoidance

Redundant links with ECCM waveforms to mitigate jamming of comms

Automated self-protect EW tactics

Alternative navigation techniques to reduce effects of GPS loss

Figure 5-5: Contested Environments Recommendations.

FOR OFFICIAL USE ONLY

Military reference books and manuals (2009-2023, Volume 6) - page 17