DEPARTMENT OF THE AIR FORCE

deliriousattackInternet και Εφαρμογές Web

4 Δεκ 2013 (πριν από 3 χρόνια και 11 μήνες)

97 εμφανίσεις


DEPARTMENT OF THE AIR FORCE

COMMANDANT OF CADETS

US AIR

FORCE

ACADEMY






DEVELOPING AN UNMANNED
AIRCRAFT
SYSTEM FOR INTELLIGENCE,
SURVEILLANCE
,

AND RECO
N
NA
ISS
ACE (ISR) MISSIONS


United States Air Force Academy UAS Research

Team
:
C1C
Keil Bartholomew,
C1C
Matthew Bender,

C1C
Neil Delaney
, C1C
Emily Fishe
r,
C1C
Kyle Moses,
C1C
Andrew Sainsbury
, C1C
Bradley Sapper,
C1C

Vai
Schierholtz, C1C
Frank Schmidt
, C1C
Taylor Soster
,
C1C
Brain Wilson
,

C2C
Tristan Latchu,

C2C
Clifford
Peterson
, C2C John Welch
, and C1C
Russell Wilson


5 May 2011



Abstract


This report describes

the

control, sensing, and communication
capabilities of an unmanned
aircraft

system

developed

through a systems engineering process

at the U.S.
Air Force Academy
.

A p
articular emphasis is placed

on finding targets
autonomously and

relaying the target
information to
a ground station

using a
user
-
friendly

graphical interface
.
The goal of
the
unmanned

aircraft

system
(UAS)
is to fly

an unmanned aerial vehicle (UAV)
over a mission
area
,
search for ground targets,
locate
, and
report
detected targets

to a human
operator
.

The
UAS

consists of

two

modules
: the ground station
and the aircraft. The target recognition
algorithm
resides on the
ground
station, which combines
telemetry data and imag
es produced by an
onboard
electro
-
optical (
EO
)

camera

to compute locations of detected targets
.

The
target

recognition software scans each image and informs the
operator

if there is a possible target in
an

image.
The UAV
was

designed
and developed
to carry

the

system

payload
w
hile maintaining
stabile
aerodynamic
flight

for
a
t least

30 minute
s
.
Propulsion is provided by a Hacker Brushless
Electric Motor and two Thunder Power 37V 5 a
mpere
h
our

(Ah)
lithium polymer batteries
.


The
payload is powered by a single Thunder Power 11.1V 5
Ah
lithium polymer battery.

The aircraft
weighs 20 pounds and has a wingspan of
80

inches.
The UAV is controlled autonomously
via a
commercial

autopilot
,

which utilizes the 900 Mhz frequency, and
backup
manual radio control
within
the 2.4
GHz
frequency

band
.

United States Air Force Academy UAS Research


Page
2

of
19


Table of Contents

List of Figures


.

.

.

.

.

.

.

.

.

3

1. Introduction


.

.

.

.

.

.

.

.

.

4


1.1 Mission Requirements

.

.

.

.

.

.

.

4


1.2
System

Analysis


.

.

.

.

.

.

.

4


1.3
System Overview


.

.

.

.

.

.

.

4


2.
System Selection

.

.

.

.

.

.

.

.

.

5


2.1 Payload Selection

.

.

.

.

.

.

.

.

5


2.2 Airframe Selection

.

.

.

.

.

.

.

6


3.
Airborne System

.

.

.

.

.

.

.

.

.

7


3.1 Autopilot


.

.

.

.

.

.

.

.

7


3.2

Camera Payload


.

.

.

.

.

.

.

8


4.
Ground S
tation

.

.

.

.

.

.

.

.

.

8


4.1
UAS Control Unit

.

.

.

.

.

.

.

8


4.2
Target Recognition/Localization Unit

.

.

.

.

.

9


4.3
Communications

Unit

.

.

.

.

.

.

.

11

5.
Safety

.

.

.

.

.

.

.

.

.

.

1
2


5.1
Autopilot Safety Measure
.

.

.

.

.

.

.

1
2


5.2 Safety
Boards

.

.

.

.

.

.

.

.

1
2


5.3 FAA Coordination/Spotters

.

.

.

.

.

.

1
2

5.4 Checklist Discipline

.

.

.

.

.

.

.

1
3

6.
Tests/Results

.

.

.

.

.

.

.

.

1
3

6.1 Hardware in the Loop

Simulation Tests
.

.

.

.

.

.

14

6.2 Flight
Tests

.

.

.

.

.

.

.

.

1
4

7.
Conclusion


.

.

.

.

.

.

.

.

.

1
4


Appendix I

.

.

.

.

.

.

.

.

.

.

1
6


Appendix II

.

.

.

.

.

.

.

.

.

.

1
7


Appendix III

.

.

.

.

.

.

.

.

.

.

1
8


List

of Terms

.

.

.

.

.

.

.

.

.

.

19



United States Air Force Academy UAS Research


Page
3

of
19




List of Figures

Figure












Page



1.

Cover Image




.

.

.

.

.

1


2.

Onboard Autopilot System

.

.

.

.

.

.

9


3.

Autopilot GUI

.

.

.

.

.

.

.

10


4.

Target Recognition/Localization Unit

.

.

.

.

.

11


5.

Target Recognition Flowchart
.

.

.

. . . .
12



6.

Results
of Target Recognition Program
.

.

.

.

.

.

1
2


7.

Safety Board .

.

.

.

.

1
4




List of Tables

Table












Page



1.

Evaluating Two Possible Payloads

.

.

.

.

.

6


2.

Aircraft Performance


.

.

.

.

.

.

7


3.

Mission Averages


.

.

.

.

.

.

.

8










United States Air Force Academy UAS Research


Page
4

of
19


1. Introduction

The
Association for Unmanned Vehicle Systems International (
AUVSI
)

Colligate

Unmanned Aircraft Systems (UAS) competition
organ
izers have

defined
system
requirements
that
are used as

the basis of developing our unmanned system. We have
designed and developed
our system
using a systems engineering approach
that

prioritized
the
given
requirements

and
guided our development
efforts
. We have devel
oped and tested
our unmanned

system to sat
isfy
these requirements.


1.1
Mission
Requirements

The AUVSI
competition rules
were

used to derive the system requirements
. The
competition requires that
an

aircraft takeoff and follow

a path

to a
designated
search area. The
search area
contains a number
of

different targets
on the ground
. The team must be able to report
information on these targets, such as the latitude, longitude, color, shape, orientation, and
alphanumeric

designator. Our system
should also

be able to cover a
pop
-
up search area

in order

to find additional targets
within

an efficient timeframe.
An additional constraint to be considered
is designing

an

aircraft
capable of
autonomous flight, inc
luding takeoff and landing.
Furthermore
, having an autonomous targeting system would be optimal f
or quick, real
-
time
reporting

of target information
, which is
an
essential
military capability.

In order to track our
requirements, our team created a Requirements
Traceability Matrix

(RTM) in which our team
grouped requirements into sections based on the
different parts of the mission
. Each requirement
was assigned a reference number along with
threshold and

objective requirement values.
Verification methods
were listed
for each numbered requirement. The RTM was used to
manage
requirements for our project
.
Our RTM

can be found in Appendix I of this document.


1.2
System

Analysis


Our team was

given two candidate payload modules and three available aircraft from
which we could select our system. We prioritized the functionalities
based

on requirements and
preferred systems engineering trade off studies as shown in Appendix II
. Our team rated safety,
communication, and autonomous flight as the top three priorities for the mission requirements.
The other
categories

that
were
prioritized

c
losely behind the
first three are
accurate
target

recognition, cost, and
weight.


1.3 System Overview


The

complete system has two
modules,
which contain subsystems

which will be
discussed in further detail in
S
ection
s

3 and 4. The first main
module

is the
UAV
. We
chose

to
use two
different
airframes:

the Kadet
-
Senior

and the
Academy

Hauler

due to aircraft availibility
and the desire for multiple backup aircraft
. The Kadet Senior is a
commercial
RC aircraft that we
have
modified for

our
mission
. The
Academy

Hauler is a customized airframe

to meet Academy
research requirements
.
Each

a
irframe ha
s

a payload that consists of a camera,
an
autopilot,
a
laser
-
altimeter, and
a
power system. The camera and the autopilot, with the laser altimeter being
connect
ed to the autopilot, transmit real
-
time information back down to the ground

station
. The
ground station
, the second module,

is made up of two main components, the autopilot
control
unit

and the image capture
/target recognition

system. The autopilot
control

unit

provides
an
operator

with a capability to control
and monitor the UAV including its trajectories and airspeed
.
The image capture
/target recognition

system is designed to provide a user with target
United States Air Force Academy UAS Research


Page
5

of
19


information. The autonomous target system captures im
ages, performs
image
-
processing

algorithms, and
deploys

resulting target information
for

an operator
.


2. System Selection


In order to determine

the best payload/ airframe selection, members of the
UAS system
engineering team

reviewed
the

AUVSI competition

rules and requirements
.

This
group, which is
made up of c
adets with various backgrounds, used a combination of
research, systems
engineering

procedures
, and 2010
competition results to determine the best airframe and payload
for the give
n mission. The main device, however, for selecting our airframe and payload was the
use of
the
systems engineering

process

based
on
feasibility analyses.


2.1 Payload Selection


The

team
considered

two different
sensor
payloads
: Payload A and Payload B. Payload A
has an onboard computer, a digital camera, and a digital communication system while Payload B
has no onboard computer, an analog camera, and an analog communication system. Based on
the
AUVSI competition guidelines and
rules,
the

design team determined that the following
items
that support system requirements

should be analyzed for

each option:

s
ensor

capabilitie
s,

c
ommunications

range
,
p
ower

requirements
,
s
afety,
c
ontrol,
w
eight, and
s
implicity. Although
each of these r
equirement families were determined to be very important, our team felt that the
payload sensing, communications

range
, and
weight

were the most important factors and these
were weighted at 0.
3
, 0.2
5
, and 0.
4
, respectively.

Refer to
Table 1.

In the sensin
g category,
Payload A was a clear winner due to its ability to support fully autonomous target recognition,
output images in
Joint Photographic Experts Group

(
JPEG
)

format, and take digital pictures
which
provided

clearer images compared to the
ones genera
ted by the
analog camera in Payload
B. Next, in the communications category, Payload A was once again the winner. Despite the fact
that Payload A uses an Ethernet signal which must be converted to a wireless signal increasing
time delay, Payload B

based on

test flights

has an issue with consistently transmitting
uninterrupted signals. Last but not least is the simplicity category
, or the selection of a
component producing the least technical risk due to complex componenets
. With the belief that








Property:

Field

of
View

Image
S
ize

Weight

Range

Num
.

of
Batt
eries



Weight:

0.15

0.15

0.40

0.25

0.05


Normalizing Rule:

Amount/
Greatest

Size
/

Greatest

Lowest
/
Amount

Highest/
Amount

Lowest/

Amount


Candidate

Ratings

Totals

Payload A

Raw Score

"+/
-
70
pan
+/
-
52 tilt"

640x480

2064.00

1 mile

3.00

0.88

(Normalized)

(1.00)

(1.00)

(0.70)

(1.00)

(1.00)

Payload B

Raw Score

"+/
-

45 pan
+/
-

45 tilt"

640x480

1444.00

1 mile

3.00

0.95

(Normalized)

(0.64)

(1.00)

(1.00)

(1.00)

(1.00)

Table
1

-

Evaluating Two Possible Payloads

United States Air Force Academy UAS Research


Page
6

of
19



fewer interfac
es between subsystems on our UAS

will create a more reliable UAS,

our team
decided that Payload B was the simpler system. This was determined because Payload B has no
onboard compu
ter which will eliminate
additional, unneeded
data connections. Payload A, on the
other hand, has an

onboard computer, which makes it necessary to have these extra
connections.
Additionally, Payload B will make the system
analysis and debugging
easier beca
use there are
fewer components to interface ultimately making the system easier to use, maintain, and train
operators in the long run.
As a result, we chose a combination of payloads A and B, selecting the
onboard system option with a superior digital came
ra, no onboard computer, and a digital
communication system.



2.2
Airframe
Selection


The primary means of selecting our aircraft was based on the correct
mass

to

wing siz
e
ratios
.
These ratios are

determined by looking at the mass to wing area ratios of the different
components of the aircraft and then
deriving

the appropriate wing area

for the desired ratio
. The
derivation begins with realizing the total mass is the sum of
masses of
the many different aircraft
components
.
For simplicity and ease of measurement, the airframe and motor masses were
combined.























Since we are not dealing with known values of

wing area, it
is
more appropriate to look at
the ratio of mass to areas. This is done by dividing all the masses by the wing area.



























With s
ome simple rearrangement

of the
equation,

we can isolate the payload ratio, as
shown below.

























Now we can solve for the wing area from the payload mass
-
to
-
area ratio.

























The mass of the payload can be easily measured and the mass
-
to
-
area ratios of the total
and airframe plus motor can
be

found with a simple calculation of the weight divided by a
referen
ce wing area. The battery ratio is slightly more complicated and involves such factors as
mission duration, charge density, and efficiencies.



















In the above equation,










and




stand for the lift
-
to
-
drag ratio, mission
duration, the efficiency coefficients of the motor and propeller used, respecitvely.
By measuring
the masses and areas of our components, we were able to find the values and ratios needed to
solve the equation. The value used for the wing area was taken from an average of the three
aircraft considered
, seen below in Table 2
.




United States Air Force Academy UAS Research


Page
7

of
19



Sig Rascal

Academy

Hauler

Kadet Senior

Average

m
Aircraft/Motor

(kg)

2.880

4.706

2.795

-

m
Payload

(kg)

1.586

1.586

1.586

1.586

m
Total

(kg)

6.852

8.678

6.023

-

S (m
2
)

0.604

0.928

0.440

0.657

m
Aircraft/Motor
/S
(kg/m
2
)

4.768

5.071

6.352

5.397

m
Total
/S (kg/m
2
)

11.344

9.351

13.689

11.461

Table
2

-

Aircraft Performance


Based on our mission and available components, we were able to choose value
s to
estimate our battery ratio, as shown in Table 3.


Mission Average Velocity

11.1 m/s

Mission

Duration

40 min

Aircraft Lift
-
to
-
Drag Ratio

8

Battery Charge Density

163 W
-
hr/kg

Motor Efficiency

0.9

Propeller Efficiency

0.85

Table
3

-

Mission Averages





(







)

(







)

















(







)




















With all the values of the sizing equation known, we were able to determine the proper
wing area for our aircraft.







(







)

(







)

(







)







All of the models investigated for this selection have a wing area greater than that which would
be ideal.
The drawback with
the

Rascal aircraft was its tail dragger design, which introduced an
additional challenge
during

aircraft take
-
off and landing.
We, therefore, chose the Kadet Senior
and Academy Hauler with their tripod gear configurations.


3.
Airborne System


The airborne payload is
made of

two
subsystems
, each with their own power system. The
first
sub
system

is
a commercial

autopilot,
and t
he se
cond is the
sensor/communication system.
We describe both
sub
systems in this section
.


3.1 Autopilot

The team chose to use
a commercial

autopilot system
to provide control for
the UAV's
autonomous flight.
This
Cloud Cap
Technology Piccolo
SL

was an easy choice because it has
been used in many of the US military

s operational UAVs and it has the capability to
easily
integrate with other
custom
-
made software
.

United States Air Force Academy UAS Research


Page
8

of
19


The onboard
autopilot

shown

in

Figure
2

has inputs for a pitot tube and static tube.

These
inputs are used by the autopilot to calculate the speed of the aircraft using differential pressures.
The autopilot has a connection for a GPS antenna
, which is used

to determine the location of the
aircraft at all times. The
imbedded
acceleromete
rs

are used to determine the pitch, roll, and yaw
angles of the aircraft. Finally, the onboard autopilot has
outputs

that provide servo
control
signals
.
This allows the
autopilot

to control all aspects of flight with little or no human
interaction.


F
igure
2
-

Onboard

Autopilot System

3.2 Camera Payload

Our system utilizes an Axis 212 PTZ Network Camera with a resolution range of 160x90
to 640x480 pixels that is capable of streaming up to 30 frames per second in either
Moving
Picture Experts Group

(
MPEG
)
-
4 format or Motion JPEG format. The camera is capable of

performing a digital

zoom
.
O
ur team
,

however
,

desired a constant field of view and therefore the
function is not
u
s
ed
. The camera mount was constructed by USAFA Training Devices in
conjunction
with the Engineering Mechanics Depar
tment and is an in
-
house design
. The camera,
coupled with the Ubiquiti Ethernet
-
to
-
Wireless communication system transmits one 480x360
image per second t
o the image processing system in

the ground station.

4. Ground Stat
ion


The ground station consists of three main subsystems: the
UAS control unit
, the
target
recognition/localization unit
, and the communication
unit
.
The three units are described in this
section


4.1
UAS Control Unit



T
he
Cloud Cap UAV
control unit has

a Graphical U
ser
I
nterface

(GUI)
, shown in Figure
3,

that allows
an operator

to control the aircraft. The program is capable of
controlling
multiple
aircraft

simultaneously. The control unit

track
s

each aircraft's location based on the
downlinked
onboard

telemetry

data
.
The operator can

switch
control of
the
aircraft

between autonomous and
manual control
and
fly the aircraft via an RC controller
when in the manual mode
.


The
control unit

allows
one

to create, delete, or modify a set of waypoints or ind
ividual
waypoints

within the GUI which the UAV will fly to
. Modifications include changing location
s

of waypoint
s

by clicking and dragging
them on the displayed map. The
operator
is also capable
of

adjusting altitude, speed,
and

direction

of the UAS at ea
ch waypoint

by opening the waypoint

United States Air Force Academy UAS Research


Page
9

of
19



Figure
3

-

Autopilot GUI


settings page. Creating a set of waypoints is as easy as clicking on the displayed map
, entering

the des
ired location of each waypoint and the altitude, and
naming the set of waypoints.


The
control unit

also allows
the operator

to monit
or the aircraft

status
as well as set
specific
operational
limits. It displays a type of Heads Up Display which shows the aircraft's
attitude, altitude, airspeed, and headi
ng. It also has
win
dows that mon
itor

bank angle, GPS
position, servo inputs, altitude

(AGL and MSL

using a laser altimeter

and GPS
, respectively)
.


Finally, the
control unit can support

auto takeoff and auto landing with the use of
the

laser altimeter. It is capable of creating a standa
rd landing pattern, calculating altitudes based on
mean sea level and the desired glide slope, as well as executing an abort if the aircraft falls
outside of set conditions, such as too far from the runway, or too high when it reaches the
landing point.

4.2
Target Recognition/Localization Unit

To accurately detect and locate
ground
targets, the team uses an image processing system
consisting of three parts

(Figure 4)
. The first component is a telemetry console written in C++,
the second component is a dat
abase written in SQL, and the third is a target recog
nition program
written in C++
. Each of the subsystems contributes to the operation of the autonomous target
recognition and location system. All three subsystems are located on
a
separate

telemetry
lapto
p
,
or console,
the team uses as part of the Ground Station. The telemetry console queries the
Cloud
Cap autopilot console

once per second for the most recent telemetry data for the aircraft and also
requests an image from the Axis digital camera once per s
econd. The telemetry console also
displays current telemetry and imagery to the team
'
s sensor operator. Once both the telemetry
from the autopilot and an image from the camera are received, the telemetry console pairs the
telemetry data and the image
using

their

respective
time stamp
s
. The telemetry console writes the
telemetry data, including the aircraft’s latitude, longitude, altitude, heading, and speed, to the
database while also writing the name of the associated image to the same row in the database.

At
the same time, the telemetry console writes the image in
JPEG

format to a file on the image
United States Air Force Academy UAS Research


Page
10

of
19


processing laptop’s hard drive. After the telemetry and image name have been written to the
database and the
JPEG

image has been written to the hard drive, the
program queries the
database for the mo
st recent database entry. Once
the program

recognizes a new entry to the
database, it reads the jpeg image from the hard drive associated with the most recent database
entry.
The target recognition program

then search
es the image for possible targets and displays
the results of the search to the user.

Figure
4

depicts the
target recognition/localization unit
.



Figure
4
-

Target Recognition
/Localization

Unit


The team uses a
seperate target recognition
program

written in C++

to autonomously
search for targets in the images received from the aircraft camera system. The search process
consists
of
multiple steps.
The program

first reads the most recent jpeg image from the hard
drive.
Next,

the program converts the image from the Red, Green,
and Blue

(RGB) spectrum into
the Hue, Saturation,
and Value

(HSV) spectrum. This facilitates the next step in the process,
which is to isolate the hue values for each pixel in the image. Once only hue va
lues exist in the
image, the program converts the hue image into a black and white image based on a threshold
pre
-
established for each color. Once only certain colors remain in the image, the program
identifies objects larger than a pre
-
determined pixel si
ze and evaluates those objects based on
their character
istics. If an object meets
the physical characteristics of a target,
the program

marks
the object as a possible target and displays th
e result to the user. Figure 5
provides a flowchart of
the target r
ecognition process.

The program

also calculates the GPS coordinates of the target based on the known
coordinates of the center of the image. The team developed an algorithm to calculate the actual
GPS coordinates of the target assuming that the telemetry i
nformation
associated with

the image
provided the correct latitude and longitude for the center of the image. This algorithm introduces
some error to the estimation of the target’s actual location, but tests
verified
, under

the conditions
expected
for the
competition,
the algorithm to be accurate within
15
feet
,

given the correct
latitude and longitude coordinates for the center of the image. If no targets are found in the
image, the program displays “No Targets Found” to
the sensor operator. Figure
6

illu
strates the
processes that take place when the
target recognition program

is executed
.

Target
Recognition

United States Air Force Academy UAS Research


Page
11

of
19




Figure
5
-

Target Recognition Flowchart




Figure
6
-

Results of T
arget Recognition Program


4.3
Communications

Unit

The communication
s

unit receives
flight
telemetry and image data from the UAV
.

It also
sends command signal
s

to the UAV.

The communication
s

unit receives images from the aircraft
which

uses the
U
biquiti

B
ullet

M2
-
HP
radio
equipped with an omni
-
directional an
tenna to send
images captured by

the

A
xis

212 PTZ Network Camera.

The images are transmitted at 2.4GHz
from the
UAV
transmitter to
the
receiver

at the ground station
.

The telemetry data is received
and command signals are transmitted on
the 900 MHz Unlicensed
industrial, scientific, and
medica
l (
ISM
)

radio frequency

band
.

United States Air Force Academy UAS Research


Page
12

of
19


5. Safety

The USAFA AUVSI Team implemented four different safety

mechanisms

into their
UAV system.
First, w
e programmed the
autopilot

with

a lost communications
orbit location
.

T
his
prevents the aircraft from flying away from

the

mission area when a communication link is lost.
Second,

we have a safety board, which allows the pilot in command to take complete
independent
radio
control of the aircraft

in case the autopilot does not respond appropriately
.
Third, w
e clear each flight through our local
airfield
aircraft controller and
verified FAA
approval

to fly our UAS in the local Class D airspace.
Finally, t
his airspace is also watched with
spotters from our team to make sure there are eyes on the aircraft at all

times. Along with the
safety measures that we take for each flight, we are serious about
following
checklists in order to
ensure each flight is carried out with

the appropriate
procedures
. As an additional safety
measure, we
switched
the radio communicati
on frequency for
independent
manual backup
control from

72
MHz

to 2.4
GHz

to prevent interference. We describe the four safety
measures

in detail next
.

5.1
Autopilot Safety Measure

The first safety
mechanism

uses

the options available to us in the
autopilot

program, the
most prominent being
a lost communication waypoint.
The lost communication
orbit

is a preset
circular

path
around the lost communication waypoint, which is chosen to be
near the ground
station

but not directly over objects or areas o
f personnel
. If the UA
V

loses communication with
the ground station for more the five seconds
,

the UA
V

will automatically start
fly
ing
to
the lost
communication
orbit

until communication is reestablished with the ground station
. If

communication is not regained within a predetermined time
,

the UA
V

will go into a controlled
crash by putting the plane into a spin.

We also program

mission limits into the
autopilot
. If the
aircraft goes
outside of these mission limits, the operator is
g
iven a visual warning
through the
ground station interface
.


5.2 Safety Boards

The second safety
measure adopted

is
the
safety board
, seen in Figure
7
. The safety board
gives a
human
radio
-
control (
RC
)

pilot on the ground the ability to take control of th
e aircraft at
anytime using a
n off
-
the
-
shelf

commercial
2.4 GHz RC controller

and receiver
. I
f the UA
V

is
not performing as it should, the
RC

pilot can take control of the UA
V

and fly it like a regular
remote control plane.
In addition
, the safety board will automatically give control to the safety
pilot if the battery for the
autopilot

fails. The safety board is a better solution than
a separate RC
controller we used in the past since
the
switch

is instantaneous and
there is no need t
o switch a
controller
. The safety board is designed,
built
, and thoroughly tested

by the Electrical

and
Computer

Engineering Department at the United States Air Force Academy.


5.3 FAA Coordination/Spotters


In order to ensure that there will be no mish
aps with local air traffic, we coordinate all of
our flights with the local
airfield
air traffic control. We notify the
air traffic controller (
ATC
)

before we take off and after we land. This provides the ATC with situational awareness of the
traffic in
th
e

airspace so
he/she can
divert any traffic away from
our operations as necessary
. We
have approval from the FAA to fly our system in the local airspace for testing, provided we have
spotters. For each
flight,

we keep

multiple

eyes on the aircraft with our

spotters. The spotters

United States Air Force Academy UAS Research


Page
13

of
19



Figure
7

-

Safety Board


keep both the
ground station control unit operator

and the
human

pilots aware of local air traffic,
birds in the area, and the whereabouts of the aircraft.


5.4 Checklist Discipline

The checklists
include

necessary steps discovered from lab and flight experiments
.

A
copy of the checklist that we u
se for each flight
can be found

in Appendix II
I
.

6
.
Tests/Results

Mission tests were conducted to
prove our UAS worked and to
prepare for the
competition. The
tests include
d

hardware
-
in
-
the
-
loop simulations
in
the lab, which
were
always
conducted before an aircraft
was
flown in the field
.

The hardware
-
in
-
the
-
loop and flight tests are
described in this section.


By combining a full year of data with many hours of

simulations and real world testing
,

our team has brought together a completed platform capable of achieving desired results.
Since
January of this year, each of our main subsystems (aircraft, payload, sensors, and ground control)
ha
ve

undergone simulation

tests in our lab at least three days a week, all exceeding thirty hours
of software
-
in
-
the
-
loop since completion. The primary flight control and aircraft have been
tested in real world scenarios and at various altitudes since mid
-
April. We utilized an agi
le
systems engineering process in conjunction with a previously developed system to allow us to
demonstrate results and perform testing at intervals
during

the design process. This agile process
allowed us to make changes continuously as needed during the
integration process.

Each of our subsystem specialists was able to
complete
several hours of testing with
multiple versions of their
sub
systems. The sensor team tested both analog and digital packages
including several different cameras, camera arrays, transmitters, receivers, and processing
software. Our flight team evaluated several aircraft ranging from the Kadet Senior, Rascal, and
the

Academy

Hauler models to the
P
redator
, and
R
eaper
used in active
operations

by the United
States Air Force.

Our extensive feasibility analysis of the Rascal, Kadet Senior, and
Academy

Hauler
airframes allowed us to clearly

see the benefit in transitioning to the more stable, larger, and
more spacious
Academy

Hauler. Our multivariable decision process
used

weighted criteria and
requirements in the decision process.

(reference Appendix II
) We broke into
five

smaller teams:
Se
nsors, Operations, Controls, Ground Station, and Management.
Each did a separate evaluation
United States Air Force Academy UAS Research


Page
14

of
19


in order to catch any differing opinions; however, each team ultimately decided on the same
system.
Similarly, when we evaluated the sensor
/communication

packages,

each team returned a
unanimous decision that the digital package utilizing the 2.4 GHz Wireless
-
to
-
Ethernet
transmitter was a better option than the analog system.

Once fully integrated, our system was able to consistently achieve successful image
collect
ion
and 50% reliability in identifying

targets. The live video feed provides an
uninterrupted view of the target area below the aircraft and the still images are undistorted and of
high quality. The
target recognition/localization

software can detect a ta
rget automatically and
display its location to the operator within an average
error
of 50ft based on GPS coordinates. All
image and location data from our flight are processed into Google
E
arth
immediately following
the aircraft’s departure of the mission
space. This gives us an overview of the targets on a map
with approximate

target

locations. Backup measures including spare aircraft are in place in case
any subsystem fails, and safety is always our top priority.


6.1 Hardware in the Loop
Simulation
Tests

H
ardware
-
in
-
the
-
loop is the name for the simulation that we run in our laboratory at the
Air Force Academy. It is a key component to our overall success
preparing for

the competition
because it allows us to simulate flying without actually putting t
he aircraft in
-
flight. The basic
components of the hardware
-
in
-
the
-
loop system are the airframe with its
entire payload,

the
ground control station, and a flight simulator.

While the hardware
-
in
-
the
-
loop simulation is
running,
all
our
UAS subsystems
opera
te

as if they were in
-
flight.
The

flight
simulator program
simulates

air speed, GPS location, orientation, direction, and altitude

of the UAV
. The ground
control station receives data from the simulator on the aircraft and transmits that to the
autopilot
.
We can also run the target recognition/localization program with t
he imagery database where the
recognition
program retrieves images and then searches through them for targets based on
differences in color and shape.

The hardware
-
in
-
the
-
loop simulation
s ar
e useful since they
show us
results
very similar
to what would
be seen

in
-
flight
.
These simulations played a critical part of our development and
the
necessary
modifications
.

A
ny
modifications to our sub
systems
had to

pass this simulation
before being test
ed on the flight line.

6.2 Flight
Tests


The team accomplished several flight tests
, which
initially

uncovered

several

problems
.

For example, when
the

laser altimeter was flight tested,
the

data was random. To validate
the

proper function of the altimeter, we flew the UAV with a known

elevation and estimat
ed

the
correct altitude based on observations during the test flights. The digital camera
also

introduced
many new issues. Among those issues
we
re
the clarity of images an
d range of transmission to the
ground station. By
repeatedly
testing the camera and communications in the air, the team
was
able to solve the problems
, mostly configuration
-
related issues
.
Overall,
these flight
tests helped
the team to ensure that the airc
raft is properly prepared for

the competition
.






United States Air Force Academy UAS Research


Page
15

of
19


7. Conclusion

Using
the

systems engineeri
ng approach, the requirements to participate successfully at
the 2010 AUVSI competition were analyzed. We prioritized the
competition
requirements and
performed
a feasibility analysis. Our team
evaluated

two separate payloads
and
three aircraft to
meet
mission requirements.


Safety is a key concern in our team’s endeavors,
for

which we have taken
multiple

measures to keep the risks low.


The

system has gone thro
ugh a substantial amount of testing
during
which we were able
to
fine tune our UAS into

the most proficient system possible.
We conducted several missions
inside and outside the lab, either using real world, or simulated tests.
There were several minor
cha
nges made to our system throughout the semester and checklists were developed during our
testing stages in order to have consistency in our flights. We were sure to take key notes of the
errors that we ran across in our tests so they
could
be analyzed.
Our

team feels
completely
confident that we are prepared to

compete in this year’s competition
with a well
-
tested

system.
We are excited to see the competition that we have to face from the other schools

and look
forward to competing
.



United States Air Force Academy UAS Research


Page
16

of
19


Appendix I

-

RTM (page 1 of 4)



Rqmt Nbr

Source of
Rqmt

Para. # (Prb
St or
Derived)

Weight/

Importance

Technical Requirement

Threshold

Objective

Verification
Method

(Demonstration,

Analysis,

Test, or

Inspection)

Achieved
Value

Sub
-
System

Team

1


4

Overall






1.1

C.3.5.2


Autonomy

During waypoint navigation and area
search

All phases of flight

Test/ Dem


CCT

1.2


Table 1 Row
2


Imagery

ID 2 targets

ID all 5

Test/ Dem


SST

1.3

Table 1 Row
4


Mission Time

Less than 40 mins

20 minutes

Test/Dem


Ops

1.4

Table 1 Row
5



Operational Availability


50% of msn

100 % of msn



Analysis



Ops

1.5

C.4.3.2,

C.4.3,3


Weather

Surface temp. up to 110 def F, fog,
visibility of 2 mi or greater with no
precipitation


Test/ Dem


PM

2

C.3.5.1, Table
1 Row 1,
M.3.3.2.3.1

3

Take Off
Requirements

From ground or static launcher


Inspection


PM

2.1

C.3.5.1


Take off area

Paved asphalt, 100 ft wide, no height
obstacles


Test/ Dem


PM

2.3

C.3.5.1,

Table 1 Row
1


Autonomous take
-
off

no

yes

Test/ Dem


CCT

2.4

C.4.3.1


Winds

Crosswinds
of 8kts with gusts
to 11kts


Test/ Dem


PM

3


4

In Flight







3.1

C.3.5.1,
C.3.5.2.2


Cruise altitude

100
-
750 ft MSL

(approx. 90
-
740 ft AGL)


Test/ Dem


GC

3.2

C.3.5.2,
C.3.5.3


Autonomously overfly
selected waypoints

yes




Test/ Dem


CCT

3.3

C.3.5.2


Flight area

Avoid no
-
fly zones



Test/ Dem


GC/ CCT

3.4

C.3.5.2, Table
1 Row 1,
M.3.3.2.2



Waypoints

Be achieved in order


Test/ Dem


GC/ CCT

4

C.3.5.3,C.4.4.
10, M.3.3.2.2,
M.3.3.2.3.5,
M.3.3.2.3.6

4

Search Area

Autonomously search for targets


Test/ Dem



CCT

4.2

C.3.5.2.2


Target Location (1)

Directly on route, while at alt of
500 ft (
±
50) MSL


Test/ Dem


SST

United States Air Force Academy UAS Research


Page
17

of
19


Appendix I
I

-

Payload
Decision Matrix






Candidate






Payload A

Payload B

Requirement Family

Property

Weight

Property
Weight

Normalizing
Rule:

Raw Score

(Normalized)

Raw Score

(Normalize)

Weight/size



0.1













Battery Weight



0.2

Lowest/#

528 g

1

528 g

1

Autopilot Weight



0.2

Lowest/#

212 g

1

212 g

1

Battery Size



0.15

Lowest/#

647680
mm^3

1

647680
mm^3

1

Autopilot Size



0.15

Lowest/#

404984
mm^3

1

404984
mm^3

1

Additional Weight



0.15

Lowest/#

820 g

0.304878049

250 g

1

Camera Weight



0.15

Lowest/#

504 g

0.900793651

454 g

1

Total



1





0.880850755



1

Target Recognition



0.15













Field Of View



0.3

#/Highest

140 degrees

1

25 degrees

0.178571429

Gimbal Pan



0.2

#/Highest

70 degrees

1

45 degrees

0.642857143

Gimbal Tilt



0.2

#/Highest

52 degrees

1

45 degrees

0.865384615

Camera Zoom



0.1

#/Highest

3 x

0.1

30 x

1

Image Size



0.15

#/Highest

640 x 480

1

640 x 480

1

Image Cap
ture rate



0.05

#/Highest

30 fps

1

30 fps

1

Total



1





0.91



0.65521978

Autonomous
flight



0.2













Way point
Navigation



0.5

Yes=1,No =0

Yes

1

yes

1

Flight Duration



0.5

Lowest/#

30

min

1

30 min

1

Total



1





1



1

Communication



0.2













Range



0.7

#/Highest

1 mile

1

1 mile

1

Bandwidth



0.3

#/Highest

20 MHz

1

20 MHz

1

Total



1





1



1

Safety



0.2













Servo Switch Board



0.25

Yes=1,No =0

yes

1

yes

1

RC control



0.25

Yes=1,No =0

yes

1

yes

1

Lost Comm
.

Waypoint



0.25

Yes=1,No =0

yes

1

yes

1

Transponder



0.25

Yes=1,No =0

yes

1

no

0

Total



1





1



0.75

Cost



0.1













Camera Cost



0.25

Lowest/#

$618.95

0.644559334

$398.95

1

Autopilot



0.25

Lowest/#

$7,500.00

1

$7,500.00

1

Battery Cost



0.25

Lowest/#

$1,245.88

1

$1,245.88

1

On board Computer
Cost



0.25

Lowest/#

$323.00

0

$0.00

1

Total



1





0.661139834



1

Simplicity



0.05













set up time



1

Lowest/#

30 minutes

1

30 minutes

1

Total



1





1



1










United States Air Force Academy UAS Research


Page
18

of
19


Appendix III

-

Hauler Start up Checklist


1.

Piccolo Ground Station

................................
...................

set up

a.

Ensure serial cable in Link 1

.............................

check

b.

Ensure UHF antenna in Link 1
..........................

check

2.

Piccolo Ground Station

................................
...................

power on

3.

PCC laptop

................................
................................
......

power on

4.

PCC software
................................
................................
...

start

5.

Video Capture System

................................
.....................

set up

6.

Hauler propeller

................................
...............................

install

7.

Piccolo battery

................................
................................
.

connect

8.

Check all Piccolo servo connections……………….connected

9.

Install wing

................................
................................
......

check

10.

C
enter of
G
ravity

................................
.............................

check

11.

R/C transmitter

................................
................................

power on

12.

Hauler servos

................................
................................
...

power on

13.

R/C only control
surfaces

................................
................

check

14.

Hauler Piccolo Autopilot

................................
.................

power on

a.

PCC autopilot reset report

................................
.

check

15.

Kill engine using PCC

................................
.....................

engine off

16.

Piccolo switch to manual

................................
.................

check

17.

Piccolo manual control surfaces

................................
......

check

18.

Piccolo switch to autopilot

................................
..............

check

19.

Preflight Window: Piccolo autopilot co
ntrol surfaces

.....

check

20.

PFD: Pitch, Bank, Yaw angles

................................
........

check

21.

PFD: Check airspeed

................................
.......................

blow in pitot

22.

Piccolo System Window

................................
.................

open

a.

Voltage > 11 Volts

................................
............

check

b.

Channel = 1 and power = 1

...............................

request and check

c.

Fast Telemetry

................................
..................

send

d.

RSSI is near
-
71

................................
................

check

e.

Link is 100

................................
........................

check

23.

Piccolo
Mission Limits Window

................................
.....

open

a.

Lost comm waypoint set to 99

..........................

set

b.

Autoland waypoint set to 90

.............................

check

c.

Comm. timeout set to 2.0 seconds

....................

check

d.

Flight termination If Lost GPS and Comm.

......

check

e.

Close throttle on flight termination

...................

check

24.

Telemetry Window: GPS 3D and PDOP

< 3.0

...............

check

25.

Preflight Window: GPS field elevation

...........................

enter

a.

Cover pitot, Zero air data

................................
..

check

26.

Hauler Camera
................................
................................
.

power on

27.

Receiving Video and Telemetry in Image Capture System …….check

28.

Focus Camera at 250 ft.................................................... chec
k

29.

Enable engine using PCC

................................
................

engine on

30.

R/C only throttle operation

................................
..............

check

31.

Piccolo manual throttle operation

................................
....

check

32.

Airspace operations and permission

................................

clear

33.

Call airspeed and altitude

................................
...............

during climb out

34.

Call GPS/INS and autopilot enable

................................
.

when GPS/INS green



United States Air Force Academy UAS Research


Page
19

of
19


List
of Terms


ISR



I
nformation,
S
urveillance,
R
econnaissance

GP
S
-

G
lobal
P
ositioning
S
ystem

EO

Earth Observation

AUVSI

Association for Unmanned Vehicle Systems International

RC



R
adio
C
ontrol

UAS



U
nmanned
A
erial
S
ystem

SE


Systems Engineering

JPEG
-

Joint Photographic Experts Group
,

method of
lossy

compression

for photographic images

MPEG
-
4

-

Moving
Picture Experts Group
,

compression

of audio and visual (AV) digital data

USAFA



United States Air Force Academy

AGL



A
bove
G
round
L
evel


MSL



M
ean
S
ea
L
evel

SQL

-

Str
uctured Query Language

RGB

Red, Green, Blue

HSV

Hue, Saturation, Value

ISM

-

Industrial, Scientific and M
edical

FAA



Federal Aviation Administration

ATC



A
ir
T
raffic
C
ontrol

GCS



G
round

C
ontrol
S
tation

PCC


Piccolo
C
ommand
C
enter

LiPo


Lithium
P
olymer


CG


C
enter of
G
ravity

PFD



Primary Flight Display

INS


I
nertial
N
avigation
S
ystem