aerospace

Article
Analysis of Radar and ADS-B Influences on Aircraft
Detect and Avoid (DAA) Systems
William Semke 1, *,† , Nicholas Allen 1,† , Asma Tabassum 1,† , Matthew McCrink 2,† ID ,
Mohammad Moallemi 3,† , Kyle Snyder 4,† ID , Evan Arnold 4,† , Dawson Stott 4,† and
Michael G. Wing 5,†
1 Department of Mechanical Engineering, University of North Dakota, Grand Forks, ND 58202, USA;
nicholas.allen@und.edu (N.A.); asma.tabassum@und.edu (A.T.)
2 Department of Mechanical and Aerospace Engineering, Ohio State University, Columbus, OH 43210, USA;
mccrink.2@osu.edu
3 Next-Generation Applied Research Lab, Embry-Riddle Aeronautical University, Daytona Beach, FL 32114,
USA; moallemm@erau.edu
4 Institute for Transportation Research and Education, North Carolina State University, Raleigh, NC 27695,
USA; kyle_snyder@ncsu.edu (K.S.); eparnold@ncsu.edu (E.A.); ddstott@ncsu.edu (D.S.)
5 Department of Forest Engineering, Resources, and Management, Oregon State University, Corvallis,
OR 97331, USA; Michael.Wing@OregonState.edu
* Correspondence: william.semke@engr.und.edu; Tel.: +1-701-777-4571
† These authors contributed equally to this work.

Received: 19 July 2017; Accepted: 8 September 2017; Published: 18 September 2017

Abstract: Detect and Avoid (DAA) systems are complex communication and locational technologies
comprising multiple independent components. DAA technologies support communications between
ground-based and space-based operations with aircraft. Both manned and unmanned aircraft systems
(UAS) rely on DAA communication and location technologies for safe flight operations. We examined
the occurrence and duration of communication losses between radar and automatic dependent
surveillance–broadcast (ADS-B) systems with aircraft operating in proximate airspace using data
collected during actual flight operations. Our objectives were to identify the number and duration
of communication losses for both radar and ADS-B systems that occurred within a discrete time
period. We also investigated whether other unique communication behavior and anomalies were
occurring, such as reported elevation deviations. We found that loss of communication with both
radar and ADS-B systems does occur, with variation in the length of communication losses. We also
discovered that other unexpected behaviors were occurring with communications. Although our data
were gathered from manned aircraft, there are also implications for UAS that are operating within
active airspaces. We are unaware of any previously published work on occurrence and duration of
communication losses between radar and ADS-B systems.

Keywords: radar; ADS-B; drop out; data anomaly; air traffic control; NextGen; Detect and Avoid

1. Introduction
Detect and Avoid (DAA) systems are complex communication and locational assemblies that
consist of several independent components that are used throughout the aviation industry. Each of
these components is assumed to be an integral portion of an overarching DAA system. A complete
DAA system would include an automatic dependent surveillance broadcast (ADS-B) module,
traffic collision avoidance system (TCAS), radar, secondary surveillance radar (Mode S) transponder,
global positioning system (GPS), and an altimeter. DAA components have extensive history and are
used in manned aircraft systems. These technologies allow communications between ground-based

Aerospace 2017, 4, 49; doi:10.3390/aerospace4030049 www.mdpi.com/journal/aerospace

Aerospace 2017, 4, 49 2 of 14

and space-based operations with aircraft. Manned and unmanned aircraft systems (UAS) rely on
DAA communication and location technologies to maintain safe flight operations. The positions of
aircraft relative to one another can be managed by air traffic control (ATC) and pilots when all systems
are operating correctly. Some of these components, including GPS and altimeter, are common UAS
components in civilian applications. It is expected that ADS-B will become a more common component
of future civilian UAS applications as manufacturers continue to improve upon the miniaturization
and affordability of this technology.
Altitude data is the most crucial data for aircraft vertical separation. Currently, altitude data
can be found from different DAA components. The barometric altimeter is the oldest and most used
device, other sources like radar altimeter and ADS-B are considered to be potential sources for reliable
altitude data. By 2020, having ADS-B onboard is mandatory for manned aircraft operating in controlled
airspace, and altitude data found within ADS-B messages can be used for vertical separation. However,
previous work [1] showed differences detected between the altitudes coming from different sources
(barometric vs. geometric altitude) along with variation in elevation detected in aircraft reports by
ADS-B (In) devices [2].
We investigated the occurrence and duration of communication losses between radar and ADS-B
systems with aircraft that were operating in proximate airspace. Our analyses considered data that
were collected during actual flight operations. Our objectives were to identify the number and duration
of communication losses for both radar and ADS-B systems that occurred within a discrete time period.
In addition, we also considered whether other unique communication behavior and anomalies were
occurring, such as reported elevation deviations. Our results revealed that the loss of communication
with both radar and ADS-B systems does occur, with variation in the length of communication losses.
We also found that unexpected behaviors, such as aircraft sharing the same identification number and
altitude discrepancies, were occurring with communications. Although our data were gathered from
manned aircraft, there are also implications for unmanned aircraft systems that are operating within
active airspaces due to similar proposed equipage requirements for group 4–5 (large platform) UAS.
From the literature review performed, this is the first study of its kind available to the public.
Over the past 100 years radar systems have evolved substantially. Despite this fact, radar, just like
any other complex engineering design is prone to design flaws and problems. Previous studies [1–10]
have been performed addressing and identifying the problems with radar. Some work has been done
on the reliability/performance of radar, while others have been on future applications to address the
current problems. However, from the literature review performed, no work to date has classified radar
data anomalies like the ones found in this paper.

1.1. Background
Several publications have documented surveillance resolution, measurement errors, and the
coverage of air traffic control radars. Thomas [3] created maps of the radar coverage throughout the
continental United States. An operational test and evaluation plan for the ARSR-4 demonstrated
that radar successfully performed its basic functions [4]. A comparable study was also done on the
ASR-11 radar by Weber and Schanne [5]. Azimuth errors on the ARSR-4 [6], ASR-9 and ASR-11 [7]
were evaluated as well. Additional studies identified vulnerabilities in secondary radar [8,9]. Even the
effects of wind turbines on air traffic control radars were evaluated [10].
While some studies addressed the current problems of radar, others looked to future applications
to address the existing problems. Immoreev and Taylor discuss how increasing the bandwidth of radars
can significantly increase their capabilities [11]. The use of phased array radars can add numerous
benefits, as described by Buckler [12]. However, with phased array radars, the multiple secondary
radar signals will need to be dealt with and separated by computer algorithms [13]. The Federal
Aviation Administration (FAA) recently published a document on the national requirements for phased
array radars [14].
Aerospace 2017, 4, 49 3 of 14

FAA conducted flight inspection to analyze the data and coverage capability for ADS-B in I90
TRACON [15]. The performance criteria for the test were measured by integrity and accuracy indicator
values present in the ADS-B message, such as Navigation Integrity Category, Navigation Accuracy for
Position, and System Integrity level. A value was set for each indicator to define the performance of
the system. The coverage analysis of this report reveals that in some area geometric altitude wasn’t
reported by 1090ES link, where the universal access transceiver (UAT) link is maintained to report
geometric altitude. FAA also published a performance report [16] to provide public information
to aid in the interpretation of ADS-B data. Studies have been conducted in Europe analyzing the
performance of ADS-B ground station data [17]. Integrity and accuracy analyses were also conducted
in China [18]. Other studies have been done on the fusion and comparison of ADS-B and radar
data [19,20]. Castle, Trinh, Mayer and Parry [20] demonstrated that the comparison between ADS-B
and Radar was occasionally difficult due to the some of the same radar failures identified in this paper.
One group of researchers is currently studying how an autonomous ATC will work [21].
There are multiple components within DAA systems. We provide a brief description and summary
of altimeters, GPS, radar and ADS-B below.

1.1.1. Altimeters
In aircraft, altimeters are used to measure the atmospheric pressure from a static port outside the
aircraft [22]. In principle as altitude increases, pressure decreases. An aneroid barometer is calibrated
on aircraft to show pressure directly as an altitude above mean seal level (MSL). The altimeter system
takes a barometric pressure reading from the Pitot-Static tube on the outside of the aircraft and uses a
vacuum system to display the altitude above MSL. Modern aircraft use a sensitive type altimeter in
which the actual MSL value can be adjusted based on current weather conditions.
Altimeters are used by both manned and unmanned aircraft as a primary means of determining
aircraft altitude. The barometric altitude is used in all systems of the aircraft from navigation to DAA.
ADS-B, TCAS, and transponders all rely on altimeters to provide aircraft altitude. Additionally, GPS
altitude as well as ground radar altitude readings can be used as an accuracy check of barometric
altitude or in the case of a barometric altimeter failure as the primary means of altitude information.

1.1.2. GPS
Manned and unmanned aircraft use GPS to determine altitude, location, and navigation. The GPS
system feeds into ADS-B and TCAS, and is used by those DAA systems for location data. This reliance
on GPS for DAA as well as basic location and navigation services makes GPS a critical system for
aviation [23]. The basic GPS service provides users with approximately 8-meter accuracy, 95% of the
time anywhere on or near the surface of the earth. To accomplish this, each of the GPS satellites emits
signals to receivers that determine the user location by computing the difference between the time that
the signal is sent and the time it is received. The time information is placed in the codes transmitted by
the satellites so that the receivers can constantly determine the time a signal was broadcast. The signal
also contains data that a receiver can use to compute the locations of the satellites as well as make
adjustments to maintain accuracy. With a minimum of four separate satellite transmissions, a GPS
receiver can triangulate its own three-dimensional position by determining the distance (referred to as
a range) from each of the satellites.

1.1.3. Radar
Detect and Avoid radar is an air-to-air radar that is being developed in order to provide an
additional layer of collision avoidance and separation for manned and unmanned aircraft in the
national airspace. The system will have one or more antenna elements in order to cover the radar field.
The electronics of the radar provide transmit, receive, control, status, and tracking functions [24,25].
Until an air-to-air system is developed and even after a system is developed, ground based radar
will be used. Ground based radar is one of the primary means for ATC to provide aircraft separation
Aerospace 2017, 4, 49 4 of 14

Aerospace
services for 2017,
manned4, 49 flight, and likely unmanned flights in the future as they are integrated 4 ofinto
14 the

national airspace
Aerospace (NAS). The FAA and other government agencies use primary and secondary
2017, 4, 49 4 of 14radar
national airspace (NAS). The FAA and other government agencies use primary and secondary radar
systems for aircraft
systems detection.
for aircraft Primary
detection. Primaryradarradar sends highfrequency
sends high frequency pulses
pulses that
that are are reflected
reflected of targets
of targets
national
(aircraft) and airspace
the (NAS).
signal is The FAA back
received and other
as government
an echo [26]. agencies
This is use
a primary
passive and secondary
system, which radar the
means
(aircraft) and the signal is received back as an echo [26]. This is a passive system, which means the
systems for aircraft detection. Primary radar sends high frequency pulses that are reflected of targets
aircraft doesdoes
aircraft not not
need anyany
need additional
additionalequipment
equipment for thesystem
for the systemtotowork.
work. The
The primary
primary datadata collected
collected
(aircraft) and the signal is received back as an echo [26]. This is a passive system, which means the
include the range
include the rangeandandazimuth
azimuthofofthetheaircraft
aircraft relative tothe
relative to theradar
radarsite.
site. The
The major
major
aircraft does not need any additional equipment for the system to work. The primary data collected
components
components to a to a
primary
primary radar radar
systemsystem include:
include: a atransmitter,
transmitter, antenna,
antenna, receiver,
receiver,and
and signal processor.
signal A schematic
processor. A schematicof of
include the range and azimuth of the aircraft relative to the radar site. The major components to a
the primary
the primary radar system can be seen in Figure 1 [27].
primary radar systeminclude:
radar system can be aseen in Figure
transmitter, 1 [27]. receiver, and signal processor. A schematic of
antenna,
the primary radar system can be seen in Figure 1 [27].

Figure1.1.Primary
Figure radarschematic.
Primary radar schematic.
Figure 1. Primary radar schematic.
Secondary radar systems differ from the primary radar system slightly, where the aircraft needs
Secondary radar systems differ from the primary radar system slightly, where the aircraft needs
to have a transponder to reply to interrogations from the secondary radar transmitter [27]. Meaning
to have Secondary radar tosystems todiffer from the primary radar system slightly, where the aircraft needs
thisaistransponder
an active system, reply
because interrogations
the aircraft needs from the secondary
a transponder for radar transmitter
the system to work.[27].
The Meaning
two
to have a transponder to reply to interrogations from the secondary radar transmitter [27]. Meaning
this isprimary
an active system, because the aircraft needs a transponder for the system
classes of transponders include Mode A and Mode C. Mode A reports the aircraft ID number, to work. The two
this is an active system, because the aircraft needs a transponder for the system to work. The two
primary
and classes
Mode Cofreports
transponders include
the altitude. Modethe
Typically, A and Mode radar
secondary C. Mode A reports
is mounted on the
top aircraft ID number,
of the primary
primary classes of transponders include Mode A and Mode C. Mode A reports the aircraft ID number,
and Mode C
radar, and reports
with the the altitude.
combination Typically,
of the data the secondary
received betweenradar
primaryis mounted on
and secondary
and Mode C reports the altitude. Typically, the secondary radar is mounted on top of the primary top of
radar, the
ATC primary
is
radar,able
andtowith
get a the
complete picture ofofan
combination theaircraft’s
data location.between
received A schematic of the and
primary secondary radar system
secondary radar, is
ATC is
radar, and with the combination of the data received between primary and secondary radar, ATC is
provided
able able
to get in Figure 2 [28]. The ground unit is typically called the interrogator, which sends a coded
to aget
complete
a completepicture
pictureofofananaircraft’s
aircraft’s location.
location. AAschematic
schematic of of
thethe secondary
secondary radarradar
systemsystem
is is
pulsein
provided to Figure
the transponder.
2 [28]. TheOnce the pulse is is
ground received, it iscalled
decoded, and a reply is send back downato
provided in Figure 2 [28]. The groundunitunit is typically
typically called the the interrogator,
interrogator, which
which sendssends
a codedcoded
the ground station.
pulsepulse
to theto transponder.
the transponder. Once
Once thepulse
the pulseisis received,
received, ititisisdecoded,
decoded, andand a reply is send
a reply backback
is send downdown
to to
the ground
the ground station. station.

Figure 2. Secondary Surveillance Radar Schematic.

Figure2.2.Secondary
Figure Secondary Surveillance Radar
Surveillance Radar Schematic.
Schematic.
Aerospace 2017, 4, 49 5 of 14
Aerospace 2017, 4, 49 5 of 14

1.1.4. Automatic Dependent Surveillance-Broadcast (ADS-B)

1.1.4. Automatic Dependent Surveillance-Broadcast (ADS-B)
ADS-B is a system by which aircraft and fixed ground locations can share position, velocity, and
ADS-B is a system by which aircraft and fixed ground locations can share position, velocity,
other information with one another [29–31]. ADS-B periodically transmits its state vector, which
and other information with one another [29–31]. ADS-B periodically transmits its state vector, which
includes horizontal and vertical position, and velocity. The system is broken down into two separate
includes horizontal and vertical position, and velocity. The system is broken down into two separate
components, ADS-B Out and ADS-B In. The transponder mode is the ADS-B Out portion which
components, ADS-B Out and ADS-B In. The transponder mode is the ADS-B Out portion which
broadcasts all state vector information. The receiving part of the system is ADS-B In which receives
broadcasts all state vector information. The receiving part of the system is ADS-B In which receives
communication from other aircraft as well as ADS-B messages from ground locations. With state
communication from other aircraft as well as ADS-B messages from ground locations. With state
vector information available from other proximate aircraft as well as information re-broadcasted
vector information available from other proximate aircraft as well as information re-broadcasted from
from ground locations, it is possible to establish the relative position and movement of those
ground locations, it is possible to establish the relative position and movement of those proximate
proximate aircraft with reference to the ownship aircraft.
aircraft with reference to the ownship aircraft.
ADS-B is automatic in the sense that no pilot or controller action is required for the information
ADS-B is automatic in the sense that no pilot or controller action is required for the information
to be broadcast. It is a dependent surveillance because it requires that the aircraft state vector and
to be broadcast. It is a dependent surveillance because it requires that the aircraft state vector and
additional information be derived from the on-board navigation equipment. The aircraft originating
additional information be derived from the on-board navigation equipment. The aircraft originating
the broadcast may or may not have knowledge of which users are receiving its broadcast.
the broadcast may or may not have knowledge of which users are receiving its broadcast.
The overall system could be used to replace secondary radar as the primary surveillance method
The overall system could be used to replace secondary radar as the primary surveillance method
for air traffic control. The ADS-B system is currently used in the United States as a component of the
for air traffic control. The ADS-B system is currently used in the United States as a component of the
NextGen national airspace strategy for upgrading and enhancing the overall aviation infrastructure.
NextGen national airspace strategy for upgrading and enhancing the overall aviation infrastructure.
ADS-B increases air traffic safety by making aircraft visible in real time to Air Traffic Control (ATC)
ADS-B increases air traffic safety by making aircraft visible in real time to Air Traffic Control (ATC)
and to other appropriately equipped aircraft. The system allows for the possibility of increased
and to other appropriately equipped aircraft. The system allows for the possibility of increased
situational awareness, improved visibility, weather reporting, flight information broadcasts, traffic
situational awareness, improved visibility, weather reporting, flight information broadcasts, traffic
capacity improvement, cockpit display of traffic information (CDTI) and airborne collision
capacity improvement, cockpit display of traffic information (CDTI) and airborne collision avoidance.
avoidance. A high-level block diagram of ADS-B can be seen in Figure 3 [32].
A high-level block diagram of ADS-B can be seen in Figure 3 [32].

Figure 3. High level universal access transceiver (UAT) automatic dependent surveillance–broadcast
Figure 3. High level universal access transceiver (UAT) automatic dependent surveillance–broadcast
(ADS-B) block diagram.
(ADS-B) block diagram.

2.2.Materials
Materialsand
andMethods
Methods
2.1. Radar
2.1. Radar
Harris (an aeronautical research organization) provided access to radar data from two Red River
Harris (an aeronautical research organization) provided access to radar data from two Red River
Valley (ND, USA) radar sites located near the cities of Fargo and Finley. The amount of data accounted
Valley (ND, USA) radar sites located near the cities of Fargo and Finley. The amount of data
for hundreds of thousands of lines of logged information each day for each site. To better understand
accounted for hundreds of thousands of lines of logged information each day for each site. To better
the data and what could be done with it, altitude information was analyzed for one day (21 June 2015).
understand the data and what could be done with it, altitude information was analyzed for one day
This date was chosen because of the Summer solstice, which provided the longest duration of daylight.
(21 June 2015). This date was chosen because of the Summer solstice, which provided the longest
This was particularly of interest, because many of the aircraft detected on the radar are student operated
duration of daylight. This was particularly of interest, because many of the aircraft detected on the
from the University of North Dakota (UND) fleet during daylight hours. Dropout rates and unique
radar are student operated from the University of North Dakota (UND) fleet during daylight hours.
behavior were looked at for both sites and categorized based on time. A dropout is identified if the
Dropout rates and unique behavior were looked at for both sites and categorized based on time. A
dropout is identified if the time between two successive logged data points is greater than the radar
Aerospace 2017, 4, 49 6 of 14

scan rate. The radar scan rate from Fargo is 4.8 s whereas Finley has a radar scan rate of 12.0 s. Unique
time between
behavior was two successive
identified logged
by any data points
unusual is greater
behavior otherthan
thanthea radar
dropout.scanOnce
rate. Thethe radar
dropoutsscan and
rate
from Fargo is 4.8 s whereas Finley has a radar scan rate of 12.0 s. Unique behavior
unique behavior were identified, the information was analyzed based on time duration, location, and was identified by
any unusual
source of the behavior other than a dropout. Once the dropouts and unique behavior were identified,
data anomaly.
the information was analyzed based on time duration, location, and source of the data anomaly.
2.2. ADS-B
2.2. ADS-B
Also, ADS-B message dropouts are a concern, which this study investigated. ADS-B should
updateAlso,
andADS-B
send its message dropouts
position are a concern,
status every second but which this study
we found this investigated.
is not always ADS-B the case. should
This
update and send its position status every second but we found this is not always
analysis is split into two phases. The first phase discusses the dropout rates. This issue is explored the case. This analysis
is split into
through two phases.
a closure The first
examination of phase
seven discusses the mostly
days of data dropout rates. This
composed issue
from theisUND
explored
fleetthrough
collecteda
closure examination of seven days of data mostly composed from the
from the Grand Forks Airport. ADS-B dropouts were analyzed utilizing the data from 15 June UND fleet collected from tothe
21
Grand Forks Airport. ADS-B dropouts were analyzed utilizing the data from
June 2015 from Grand Forks Airport. The data were archived in a GDL-90 raw pass-through format 15 June to 21 June 2015
fromparsed
and Grand forForks Airport. TheDropout
interpretation. data were archived
rates in a GDL-90
were counted as theraw pass-through
number of times format
the dataanddidparsed
not
for interpretation. Dropout rates were counted as the number of times the
indicate a one second update period. A total of 642 aircraft data were analyzed including 534,736 datadata did not indicate
a one second update Each
points/transmissions. period.
dataApoint
totalcontained
of 642 aircraft data were
the aircraft analyzed
position (latitude,including 534,736
longitude), data
altitude,
points/transmissions.
aircraft status (Ground/Air) Each along
data point
with contained
additional the aircraft position (latitude, longitude), altitude,
information.
aircraft status (Ground/Air) along with additional information.
The ADS-B unit at Grand Forks Airport is a GDL-90 datalink transceiver. Aircraft altitude
The ADS-B
discrepancy was unit at GrandinForks
investigated Airportphase
the second is a GDL-90
with thedatalink transceiver.
data collected from Aircraft
an openaltitude
source
discrepancy was investigated in the second
website (ADS-B Exchange 2016) with world-wide coverage.phase with the data collected from an open source website
(ADS-B Exchange 2016) with world-wide coverage.
3. Results
3. Results
To adhere to the standards set forth by the Federal Aviation Administration and the standard
To adhere to the standards set forth by the Federal Aviation Administration and the standard
units for the United States. National Airspace System, the altitude readings and subsequent analysis
units for the United States. National Airspace System, the altitude readings and subsequent analysis
are provided and performed in United States Customary System units, rather than SI units.
are provided and performed in United States Customary System units, rather than SI units.
3.1.
3.1. Radar
Radar
Cyclic
Cyclic behavior,
behavior, multiple
multiple aircraft
aircraft with
with the
the same
same IDID number,
number, and
and outliers
outliers (a
(a deviation
deviation of of more
more
than
than 10% of the expected altitude based on the past and projected behavior of the aircraft) were the
10% of the expected altitude based on the past and projected behavior of the aircraft) were the
primary
primary unique
unique radar
radarbehaviors
behaviorsspotted
spottedininthis
thisanalysis.
analysis.Figures
Figures4–7 show
4–7 showthethe
four
fourdifferent types
different of
types
anomalous
of anomalous behavior (drop outs and the three unique behaviors). Figure 4 shows dropouts as aa
behavior (drop outs and the three unique behaviors). Figure 4 shows dropouts as
function
function ofof altitude and dropout
altitude and dropout duration
duration length.
length. Note
Note the
the durations
durationspast
past6.33
6.33×× 10
104 ss where
4
where there
there are
are
missed data points, these drop outs ranged from 9 to 24 s. Figure 5 shows unique
missed data points, these drop outs ranged from 9 to 24 s. Figure 5 shows unique cyclical behavior cyclical behavior
with
with altitude
altitude readings
readings fluctuating
fluctuating between
between 35,000
35,000 and
and 36,000
36,000 feet
feet over
over aa 200
200 ss time
time span.
span. Figure
Figure 66
shows a case where there were many aircraft with the same aircraft Mode A
shows a case where there were many aircraft with the same aircraft Mode A ID number being reported ID number being
reported during the same time period. Lastly, Figure 7 shows numerous outliers,
during the same time period. Lastly, Figure 7 shows numerous outliers, where the reported altitudewhere the reported
altitude
repeatedlyrepeatedly
jumps fromjumps from nominal
nominal (expected)
(expected) to 0 feet to 0 feet elevation.
elevation.
Altitude [ft]

Figure
Figure 4.
4. Altitude
Altitude [ft]
[ft] vs.
vs. time
time[sec
[secfrom
from 21
21 June
June 2015
2015 12:00
12:00 a.m.]
a.m.] plot
plot showing
showing dropouts.
dropouts. Note
Note the
the time
time
gaps/dropouts on the right side of the plot within the red
gaps/dropouts on the right side of the plot within the red box. box.
Aerospace 2017,
Aerospace 4, 49
2017, 4, 49 7 of 14
7 of 14
Aerospace 2017, 4, 49 7 of 14

4
10 4
10
3.65
3.65 10 4
3.65

3.6
3.6
3.6

3.55
3.55
3.55

3.5
3.5
3.5

Altitude[ft][ft]
3.45
3.45
3.45
Altitude
3.4
3.4
3.4

3.35
3.35
3.35

3.3
3.3
3.3

3.25
3.25
3.25

3.2
3.2
3.2 00 50
50 100
100 150
150 200
200 250
250 300
300 350
350 400
400
0 50 100 150 200 250 300 350 400
Time[sec
Time[sec from
from June
June 21,
21, 2015
2015 12:00
12:00 am]
am]
Time[sec from June 21, 2015 12:00 am]

Figure 5. Altitude
Figure [ft] [ft]
5. Altitude vs. time [sec [sec
vs. time fromfrom
21 June
21 2015
June 12:00
2015 a.m.]
12:00 plot
a.m.]showing cyclicalcyclical
plot showing unique unique
behavior.
Figure 5. Altitude [ft] vs. time [sec from 21 June 2015 12:00 a.m.] plot showing cyclical unique
Note the altitude
behavior. jumping
Note the between
altitude jumping 35,000 (red
between circles)
35,000 and
(red 36,000
circles) feet
and (blue
36,000 circles).
feet (blue circles).
behavior. Note the altitude jumping between 35,000 (red circles) and 36,000 feet (blue circles).
[ft]
Altitude[ft]
Altitude

Figure 6. Altitude
Altitude [ft] vs.time
[ft] time [secfrom
from 21June June 2015 12:00 a.m.]
12:00 a.m.] plot showing multiple aircraft with
Figure 6. 6.
Figure Altitude [ft]vs.
vs. time[sec
[sec from21 21 June 2015
2015 12:00 a.m.] plot
plotshowing
showingmultiple
multiple aircraft
aircraft with
with
the
thethe same
same aircraft
aircraft ID number. This plot illustrates an example of when there were many readings that
same aircraftID
IDnumber.
number. This Thisplotplotillustrates
illustratesanan example
example of when
of when therethere
were were
many many readings
readings that
occurred
that occurred at at
thethe
same same time period
time without
period without the ability to resolve
the ability them independently,
to resolve indicating that
occurred at the same time period without the ability to resolve them them independently,
independently, indicating
indicating that
multiple
that multiple aircraft
aircrafthadhad the same ID number. As the ID numbers are identical it is not possible to
multiple aircraft had thethe same
same IDID number.
number. AsAsthethe
IDID numbers
numbers areare identical
identical it is
it is notnot possible
possible to to
distinguish
distinguish individual
individual aircraft, resulting in significantly degraded situational awareness.
distinguish individualaircraft,
aircraft,resulting
resultingin insignificantly
significantly degraded situationalawareness.
degraded situational awareness.

Figure 7. Altitude [ft] vs. time [sec from 21 June 2015 12:00 a.m.] plot showing outliers. Note how the
Figure 7. Altitude [ft] vs. time [sec from 21 June 2015 12:00 a.m.] plot showing outliers. Note how the
Figure 7. Altitude
altitude readings[ft]
jumpvs.from
timetheir
[sec nominal
from 21 reading
June 2015 12:00
(blue a.m.]toplot
circles) showing
0 ft (red outliers.
circles) Note
back to the how the
nominal
altitude readings jump from their nominal reading (blue circles) to 0 ft (red circles) back to the nominal
altitude readings jump
value many times. from their nominal reading (blue circles) to 0 ft (red circles) back to the nominal
value many times.
value many times.
Aerospace 2017, 4, 49 8 of 14

Dropouts and unique behavior were classified into time intervals. Table 1 summarizes the dropout
data, while Table 2 summarizes the unique behavior data (note: the multiple aircraft unique behavior
was omitted from the data in the table). The average, minimum, and maximum dropout/unique
behavior duration is listed, along with the categorized time intervals. As shown from the tables,
the average dropout was approximately five radar scans (23.69 s) for Fargo, and just under three radar
scans (32.41 s) for Finley. The unique behavior statistics were much lower, with an average of about one
radar scan (5.30 s) for Fargo and approximately two radar scans (23.52 s) for Finley. Fargo had 67% of
its dropouts and 98% of its unique behavior occur for less than three radar scans. While Finley had 82%
of its dropouts and 84% of its unique behavior for less than three radar scans. However, there are some
dropouts that occur for a significant period, with some dropouts persisting for hundreds of seconds.

Table 1. Summary of dropout statistics for both the Fargo and Finley sites. The average, minimum,
and maximum dropout time is noted. The dropouts are also ranked by time intervals, showing the
number of dropouts for that time period and the percentage.

Fargo Finley
Dropout Duration - Dropout Duration -
Average (s) 23.69 - Average (s) 32.41 -
Minimum (s) 9.00 - Minimum (s) 23.53 -
Maximum (s) 265.46 - Maximum (s) 336.27 -
Categorized Data - Categorized Data -
Number of Dropouts 132 - Number of Dropouts 352 -
Less than 10 s 65 49% Less than 24 s 146 41%
10–15 s 24 18% 24–36 s 143 41%
15–20 s 1 1% 36–48 s 24 7%
20–25 s 5 4% 48–60 s 9 3%
25–30 s 5 4% 60–90 s 19 5%
30–60 s 26 20% 90–120 s 9 3%
Greater than 60 s 9 7% Greater than 120 s 2 1%

Table 2. Summary of unique behavior for the Fargo and Finley sites. The average, minimum,
and maximum unique behavior time is noted. The unique behavior occurrences are also ranked by time
intervals, showing the number of unique behavior instances for that time period and the percentage.

Fargo Finley
Unique Behavior - Unique Behavior -
Average (s) 5.31 - Average (s) 23.52 -
Minimum (s) 4.66 - Minimum (s) 11.84 -
Maximum (s) 19.24 - Maximum (s) 108.16 -
Categorized Data - Categorized Data -
Unique Behavior 52 - Unique Behavior 25 -
Less than 10 s 51 98% Less than 24 s 19 76%
10–15 s 0 0% 24–36 s 2 8%
15–20 s 1 2% 36–48 s 1 4%
20–25 s 0 0% 48–60 s 1 4%
25–30 s 0 0% 60–90 s 0 0%
30–60 s 0 0% 90–120 s 2 8%
Greater than 60 s 0 0% Greater than 120 s 0 0%

Time and location were investigated to see if either demonstrated an influence on dropouts or
unique behavior in system performance. Time did not appear to play a large role, however location
had a large effect. This can be seen by Figures 8 and 9, which shows the dropouts and unique behavior,
with the effective radar radius overlaid on top of the points. It can be seen that a significant number of
Aerospace 2017, 4, 49 9 of 14

the dropouts and unique behavior occurred either in the approach path to the airport (Fargo or Grand
Aerospace
Forks) or2017,
near4, the
49 effective radar radius. 9 of 14
Aerospace 2017, 4, 49 9 of 14

(a) (b)
Figure 8. The Finley Radar (a)
Map showing the effective radar radius (blue circle) and the location of each
(b)
Figure
dropout or 8. The Finley
unique Radar
behavior (redMap
dot): showing
(a) the effective
Dropouts; (b) radar
Unique radius (blue circle) and the location of
Behavior.
Figure 8. The Finley Radar Map showing the effective radar radius (blue circle) and the location of each
each dropout or unique behavior (red dot): (a) Dropouts; (b) Unique Behavior.
dropout or unique behavior (red dot): (a) Dropouts; (b) Unique Behavior.

(a) (b)
(a) (b)
Figure 9. The Fargo Radar Map showing the effective radar radius (blue circle) and the location of
each dropout
Figure or unique behavior (red dot):the
(a)effective
Dropouts; (b) Unique Behavior.
Figure 9.9.The
TheFargo
Fargo Radar
Radar Map
Map showing
showing the effective radar
radar radius
radius (blue(blue circle)
circle) andlocation
and the the location of
of each
each dropout
dropout or unique
or unique behavior
behavior (red(a)
(red dot): dot): (a) Dropouts;
Dropouts; (b) Unique
(b) Unique Behavior.
Behavior.
An overlapping study was also done to determine if the dropouts and unique behavior were
radarAnor overlapping
transponder induced.
study was Fargo’s
also doneradartoradius fits entirely
determine inside Finley’s,
if the dropouts so theoretically,
and unique all of
behavior were
An overlapping study was also done to determine if the dropouts and unique behavior were radar
Fargo’s
radar or planes should
transponder be detected
induced. Fargo’sbyradar
Finley (unless
radius fits the planes
entirely are Finley’s,
inside too low so to theoretically,
be intercepted all by
of
or transponder induced. Fargo’s radar radius fits entirely inside Finley’s, so theoretically, all of Fargo’s
Finley’s radar beam). It was found that only 2 of the 13 (15%) planes
Fargo’s planes should be detected by Finley (unless the planes are too low to be intercepted by from Fargo showed unique
planes
behaviorshould be detected
on both radars, by Finley
and 1 out of(unless 64the
the only planes are too
(1.5%) low to beexperienced
intercepted dropouts
by Finley’s radar
Finley’s radar beam). It was found that 2 of planes from
the 13 (15%) Fargo
planes from Fargo showed on both
unique
beam). It
radars. This was found
indicates that only
thatand 2 of
in most the 13 (15%) planes from Fargo showed unique behavior
issueon both
behavior on both radars, 1 out cases
of theof64dropouts and unique
(1.5%) planes behavior,
from Fargo it is a radar
experienced dropouts and
on not
both
radars, and
a transponder 1 out of the 64
issue. that(1.5%)
However, planes from Fargo experienced dropouts on both radars. This indicates
radars. This indicates in most transponder
cases of dropouts issues
anddouniquestill occur,
behavior,which
it is aisradar
shownissuefrom the
and not
that in most cases of dropouts and unique
aircraftbehavior, it is a radar issue andanot a transponder issue.
aoverlapping
transponder analysis.
issue. The multiple
However, transponder unique behavior
issues do stillis also likely
occur, whichtransponder
is shownerror,fromwiththe
However,
either thetransponder
wrong planeissues do still occur,
ID multiple
number being which
entered,is shown
ATCfrom the overlapping analysis. Theplane
multiple
overlapping analysis. The aircraft uniqueorbehavior assigning multiple
is also likely of the same
a transponder IDs.
error, with
In conclusion,
either the wrong many
planedropouts
ID number andbeing
unique behavior
entered, or ATC phenomenon occur onofathe
assigning multiple daily
samebasis,
planewhich
IDs.
warrants
In futuremany
conclusion, investigation.
dropouts and unique behavior phenomenon occur on a daily basis, which
warrants future investigation.
Aerospace 2017, 4, 49 10 of 14

aircraft unique behavior is also likely a transponder error, with either the wrong plane ID number
being entered, or ATC assigning multiple of the same plane IDs. In conclusion, many dropouts and
unique behavior phenomenon occur on a daily basis, which warrants future investigation.

3.2. ADS-B
A total of 642 aircraft data were analyzed including 534,736 data points/transmissions. Each data
point contained the aircraft position (latitude, longitude), altitude, aircraft status (Ground/Air) and a
few other messages. Dropouts occurred 35,063 times, but it should be noted that this number (35,063)
does not represent time. In other words, the ADS-B message was lost for at least 1 s over 35,063 times
transponding. The dropouts ranged from 2 s to over 5 min in some cases. Table 3 summarizes
these results.

Table 3. Dropout statistics over 7 days (15 June to 21 June 2015). Dropouts are also categorized by
duration interval including occurrences and interval percentages.

Total Data Set Dropout % Dropout

534,736 35,063 6.6%
Categorized Data
Total 35,063 -
Less Than 5 s 25,066 71.5%
5–10 s 512 1.5%
10–20 s 313 0.9%
20–40 s 92 0.3%
40–60 s 163 0.3%
60–300 s 3401 9.7%
Greater 300 s 5469 15.6%

We found that 71.5% of dropouts were less than 5 s in duration, and approximately 13% of total
dropouts fell between 5 and 300 s. One unique finding was that 15.6% of the total dropouts lasted
longer than 300 s (5 min), which is quite significant. However, it might not be solely a transponder
issue. This data includes UND’s training fleet, so there is a possibility that the aircraft piloted by
students took a flight path that went out of range of the ADS-B receiver. To check this hypothesis,
the dataset from 21 June 2015 was considered because it had the least number of aircraft. The aircraft
showing dropouts more than 300 s were identified, and the position information was scrutinized.
We did find that some (3 out of 5) of the aircraft were taking a route that went outside the range of the
ground ADS-B and came back after several minutes and repeated that maneuver again. This is one of
the potential reasons for the higher dropout time in this initial study.
In the second phase of the ADS-B study, a representative sample of data starting from 12:00 a.m.
to 12:14 a.m. on 20 October 2016, was utilized. A total of 1282 aircraft were found, of which only
744 aircraft reported position, and both geometric altitude and barometric altitude. This means
approximately 42% (538) missed at least some portion of the position or altitude data. In addition,
only 1155 aircraft reported both barometric and geometric altitude, which means that 10% of aircraft
missed at least one of the altitude measurements.
Deviation was found between barometric and geometric altitudes ranging from 9 to 500 feet in
the ADS-B data. Messages of 121 (10%) aircraft showed deviation in altitude data with approximately
48% of the aircraft having barometric altitude higher than geometric altitude. This trend was reversed
in the remaining aircraft, comprising approximately 52% of the aircraft data. Table 4 summarizes
the results.
Primary analysis revealed that the altitude discrepancy could be as high as 500 feet. About 28%
of discrepancies fell above 200 feet which might increase the likelihood of violation of the vertical
separation threshold (Table 5). A significant percentage of these cases could produce the possibility
Aerospace 2017, 4, 49 11 of 14

of a near mid-air collision condition. The discrepancies were nearly normally distributed above and
below the actual altitude, so combinations where one is high and the other low can lead to dramatic
vertical separation errors.

AerospaceTable
2017, 4. Aircraft demonstrating different characteristics of altitude data from ADS-B messaging. 11 of 14
4, 49

Primary analysis revealed that the altitude discrepancy

Behaviors couldofbeAircraft
Number as high asPercentage
500 feet. About 28%
of discrepancies fell above
Report200 feet
Both which might increase the likelihood
Altitude 1155 of violation
- of the vertical
No deviation
separation threshold (Table found
5). A significant percentage of these1034
cases could produce 90%the possibility
Discrepancy
of a near mid-air collision condition. found
The discrepancies were nearly121 normally distributed
10% above and
Out of 121 cases of Consistent Discrepancy 99 82%
below the actual altitude, so
altitude discrepancy
combinations where one
Fluctuating Discrepancy
is high and the
22
other low can lead
18%
to dramatic
vertical separation errors.
Table 5. Aircraft altitude discrepancy ranges.
Table 5. Aircraft altitude discrepancy ranges.

Percentage
Percentage(Out
(Outofof121
121Aircraft
AircraftThat
That
Altitude
Altitude Discrepancy
Discrepancy Range
Range Numberof
Number ofAircraft
Aircraft
ShowedAltitude
Showed AltitudeDiscrepancy)
Discrepancy)
11 feet–50
feet–50 feet
feet 32
32 26%
26%
51 feet–100
51 feet–100 feet
feet 23
23 19%
19%
101
101 feet–150
feet–150 feet
feet 16
16 13%
13%
151
151 feet–200
feet–200 feet
feet 16
16 13%
13%
201
201 feet–250
feet–250 feet
feet 17
17 14%
14%
251
251 feet–350
feet–350 feet
feet 11
11 9%
9%
351 feet–450 feet 04 3%
351 feet–450 feet 04 3%
451 feet–500 feet 02 2%
451 feet–500 feet 02 2%

It
It isis important
importantto to recognize
recognize the characteristics
the characteristics of theof the discrepancy
discrepancy in orderin order to
to resolve resolve
them. The
them. The initial study revealed that approximately 82% aircraft display consistent
initial study revealed that approximately 82% aircraft display consistent discrepancy, however there discrepancy,
however there are
are cases (~18% cases
of the (~18%where
aircrafts) of the
theaircrafts)
deviation where
is not the deviation
a constant value is and
not fluctuates.
a constantAlso,
valuephase
and
fluctuates.
of flight appearsAlso, to
phase
playofanflight appears
important roletoinplay
the an important Level
discrepancy. role inflight
the discrepancy.
showed higher Level flight
deviation
showed higher deviation than
than ascending/descending ascending/descending
phases. Figure 10 represents phases. Figure of
an example 10therepresents anin
deviation example of the
two different
deviation
flight phases. in two different flight phases.

(a) (b)
Figure 10.
Figure 10. Altitude
Altitude [ft]
[ft] vs.
vs. time
time [sec
[sec from
from 20
20 October
October 2016
2016 12:00
12:00 a.m.]
a.m.] plots
plots showing
showing discrepancies
discrepancies in
in
ascending and level flight: (a) Altitude discrepancy is small in ascending phase; (b) discrepancy
ascending and level flight: (a) Altitude discrepancy is small in ascending phase; (b) discrepancy range range
is higher
is higher in
in level
level flight.
flight.

A unique anomaly was discovered by the analysis that can be called “altitude jump” where the
A unique anomaly was discovered by the analysis that can be called “altitude jump” where the
altitude data set jumps every other sample. This was similar to the cyclical behavior identified in the
altitude data set jumps every other sample. This was similar to the cyclical behavior identified in the
radar analysis. The analysis was done on a small sample of data; a more diverse and comprehensive
study may provide more insight on the anomalies found in this preliminary study and lead to
solutions to overcome these behaviors. Figure 11 represents the “altitude jump” anomaly.
Aerospace 2017, 4, 49 12 of 14

radar analysis. The analysis was done on a small sample of data; a more diverse and comprehensive
study may provide more insight on the anomalies found in this preliminary study and lead to solutions
Aerospace 2017, 4, 49 12 of 14
to overcome these behaviors. Figure 11 represents the “altitude jump” anomaly.

Figure Altitude
11.11.
Figure [ft][ft]
Altitude vs.vs.
Time [sec
Time from
[sec 2020
from October 2016
October 12:00
2016 a.m.]
12:00 plot
a.m.] showing
plot cyclic
showing change
cyclic inin
change
both geometric
both geometric and barometric
and barometric altitude data.
altitude data.

4. 4.Discussion
Discussion
Thisstudy
This studysought
soughtto to further explore
exploreDAA DAAcriticality
criticalitythrough
through examination
examination of live
of liveradar and and
radar ADS-
B thatthat
ADS-B were made
were madeavailable
availablefrom twotwo
from radar sitessites
radar in proximity
in proximity to theto University
the University of North
of North Dakota. Over
Dakota.
hundreds
Over hundreds of thousands
of thousands of oflines
lines ofofdata
datawerewereexamined
examined for for these sitessites representing
representingone onedaydayofof
observations
observations atat each
each site.The
site. Thedropout
dropout ratesand
rates andunique
uniquebehavior
behavior were
were tallied
tallied from
from thetherecorded
recorded radar
radar
data.The
data. Theaverage
averagedrop dropout outlength
lengthoccurred
occurredfor for23.69
23.69s sand and32.41
32.41s,s,atatFargo
Fargoand andFinley,
Finley,respectively.
respectively.
While
While thethe average
average unique
unique behavior
behavior length
length waswas 5.31
5.31 s at
s at Fargo
Fargo and and 23.52
23.52 s atFinley.
s at Finley. InIn fact,
fact, some
some drop
drop
outs lasted for several minutes. Considering the average or even
outs lasted for several minutes. Considering the average or even the maximum time length of drop outs the maximum time length of drop
outs
and and unique
unique behavior, behavior,
they are they are a significant
a significant period of period
time forofATC timetofor ATC erroneous
receive to receivedata. erroneous data.
The large
The large
majority majority
of drop outs ofand drop
uniqueoutsbehavior
and unique behavior
occurred on the occurred
edge of on the the edge radar
effective of theradius
effective radar
or near
radius or
airports. near
This airports.
is logical This isthe
because logical
radar because
is being thepushed
radar is tobeing
the edgepushedof itstooperational
the edge oflimitations
its operationalat
limitations
the edge of itsatscanthe edge
area. of its scanalso
Airports area. Airports
have some alsoof the have sometraffic
highest of thedensity,
highestwhich trafficlikely
density, which
leads to
likely
these leads to these
phenomena. We phenomena.
determined that We determined
when dropouts thatandwhen dropouts
unique and unique
behavior occur, itbehavior occur, it
is often caused
byisthe
often
radarcaused
groundby the radar
station andground
not the station and notHowever,
transponder. the transponder.it was alsoHowever,
found that it was thealso found that
transponder
the transponder
functionality appearedfunctionality appeared
to be responsible fortosome
be responsible
of the unique forbehavior
some of cases.
the unique behavior
The results cases.
suggest The
that
results suggest that dropouts and unique behavior occur every
dropouts and unique behavior occur every day. This is concerning as the unique behavior included day. This is concerning as the unique
behavior
altitude included
readings altitude
that readings
alternated that alternated
by hundreds of feetbyathundreds
successive of feet
scans at successive scans and large
and large discrepancies
discrepancies
between altitude between
readings. altitude
Thisreadings.
could prove This problematic
could prove problematic for vertical
for vertical aircraft aircraft separation.
separation. Further
Furthershould
research research shouldhow
explore explore how pervasive
pervasive and common and common these failures
these failures are over aretime
overperiods
time periodsthat gothat
go beyond
beyond a single
a single day,atand
day, and otherat other
radarradar locations.
locations.
WeWe examined
examined ADS-Bdropout
ADS-B dropoutrates ratesover
overone oneweek
weekofofdata datafrom fromananairport
airportlocated
locatednear nearUND,
UND,
involving642
involving 642aircraft
aircraftand and 534,736
534,736 data
data point
point transmissions.Data
transmissions. Datawas waslostlost35,063
35,063timestimes(6.6%(6.6%ofofallall
transmissions)for
transmissions) fora least
a least oneone second
second during
during transponding.InInsome
transponding. somecases,
cases,dropouts
dropoutslasted lasted over
over five
five
minutes but the vast majority of the dropouts (71.5%) were less
minutes but the vast majority of the dropouts (71.5%) were less than five seconds. These dropouts than five seconds. These dropouts
may
may notnotallall
bebe attributable
attributable to to transponder
transponder failure
failure as some
as some of these
of these dropouts
dropouts werewere likely likely attributable
attributable to
to aircraft
aircraft leavingleaving the range
the range of the ofground
the ground ADS-B ADS-B and later
and later returning.
returning.
ToTo further
further examine
examine ADS-B
ADS-B performance,we
performance, weconsidered
considereddata datafromfroma world-wide
a world-widedatabase databaseofof
1282aircraft
1282 aircraftoperating
operatingover overa abriefbrieftimetime span.Only
span. Only744 744(58%)
(58%)ofofthesetheseaircraft
aircraftreported
reportedposition,
position,
geometric altitude, and barometric altitude. Only 1155 (90%)
geometric altitude, and barometric altitude. Only 1155 (90%) of these aircraft had both barometricof these aircraft had both barometric
and
and geometricaltitude.
geometric altitude.AmongAmongthe theaircraft
aircraftthat
thatreported
reportedboth both altitudes,
altitudes, elevationdifferences
elevation differenceswere were
present in 121 (10%). We found elevation variation as high as 500 feet when we compared barometric
and geometric altitudes. Approximately 28% (34) of the aircraft had elevation differences in excess of
200 feet, a level which could lead to a mid-air collision.
These two ADS-B studies raise concerns about ADS-B reliability and functionality. Further study
should examine whether the trends we observed are consistent over longer time spans and larger
Aerospace 2017, 4, 49 13 of 14

present in 121 (10%). We found elevation variation as high as 500 feet when we compared barometric
and geometric altitudes. Approximately 28% (34) of the aircraft had elevation differences in excess of
200 feet, a level which could lead to a mid-air collision.
These two ADS-B studies raise concerns about ADS-B reliability and functionality. Further study
should examine whether the trends we observed are consistent over longer time spans and larger
aircraft fleets. Further study also might investigate whether the anomalous patterns and behavior we
witnessed with radar and ADS-B also occur with other facilities and locations.
Our results drew from data that involved communications with manned aircraft. Our findings
are also relevant, however, for UAS. UAS are increasingly becoming part of airspace throughout the
world, and it is expected that unmanned aircraft will number more greatly than manned aircraft at
a not-too-distant future, particularly in the transport and delivery role. Radar and ASD-B will also
be technologies that UAS rely upon for airspace deconfliction and will therefore be affected by the
identified anomalies and failures. In addition to understanding these anomalies to correctly build
robust DAA systems for UAS, findings from this study could also be applied to autonomous ATC
systems. In the future, autonomous ATC systems will likely be used in airspace that is not heavily
congested. The radar and ADS-B anomalies identified in this work will need to be understood to help
with the logic used in the future ATC systems.

Acknowledgments: The Federal Aviation Association has sponsored this project through the Center of Excellence
for Unmanned Aircraft Systems. However, the agency neither endorses nor rejects the findings of this research.
The presentation of this information is in the interest of invoking technical community comment on the results
and conclusions of the research.
Author Contributions: William Semke, Nicholas Allen, and Asma Tabassum contributed to all phases of this
research. Matthew McCrink and Mohammad Moallemi participated in the analysis, review, and writing of results.
Kyle Snyder was the team lead in organizing this research effort, reporting with our sponsor, and reviewing all
research products. Evan Arnold and Dawson Stott contributed to the experiment design and assisting with the
synthesis of the results in the scope of the unified research project. Michael G. Wing coordinated the writing team
effort and contributed to the writing and review of research results.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Taib, N.A.; Ali, B.S. An Analysis of Geometric Altitude Data in ADS-B Messages. In Proceedings of the ION ITM,
Monterey, CA, USA, 25–28 January 2016.
2. Federal Aviation Administration. ADS-B IN Device Performance Analysis Report; Simulyze Inc.: Reston, VA,
USA, 2016.
3. Thomas, M.L. Atlas of Radar Coverage of the Lower 48 Border States—Final. Available online:
http://psugeo.org/Publications/Atlas%20of%20Radar%20Coverage%20of%20the%20Lower%2048%
20Border%20States%20-%20Final.pdf (accessed on 8 March 2017).
4. Healy, T.A.; McDonald, R.K.; Pomrink, R.F.; Conklin, W.P. Air Route Surveillance Radar Model 4 (ARSR-4)
Operational Test and Evaluation (OT&E) Final Report; Federal Aviation Administration Technical Center:
Atlantic City, NJ, USA, 1997.
5. Weber, R.; Schanne, J. Airport Surveillance Radar Model 11 (ASR-11) FAA Test and Evaluation Master Plan
(TEMP); Federal Aviation Administration Technical Center: Atlantic City, NJ, USA, 1998.
6. Busch, A.C.; Bradbury, P. Measurement and Analyses of ASR-4 System Error. Part I. Overview; Federal Aviation
Administration: Atlantic City, NJ, USA, 1974.
7. Mayer, C.; Tzanos, P. Comparison of ASR-11 and ASR-9 Surveillance Radar Azimuth Error. In Proceedings
of the 2011 IEEE/AIAA 30th Digital Avionics Systems Conference, Seattle, WA, USA, 16–20 October 2011.
8. Air Traffic Control Radar Beacon System (ATCRBS). Available online: http://web.mit.edu/6.933/www/
Fall2000/mode-s/atcrbs.html (accessed on 30 January 2017).
9. Phillips, D. Mode A and Mode C: The Straight Scoop on How It Works. Available online: http://www.
aeroelectric.com/articles/Altitude_Encoding/modec.htm (accessed on 30 January 2017).
10. Lemmon, J.L.; Carrol, J.E.; Sanders, F.H.; Turner, D. Assessment of the Effects of Wind Turbines on Air Traffic
Control Radars; United States Department of Commerce: Washington, DC, USA, 2008.
Aerospace 2017, 4, 49 14 of 14

11. Immoreev, I.J.; Taylor, J.D. Future of Radars. In Proceedings of the IEEE Conference on Ultra Wideband
Systems and Technologies, Baltimore, MD, USA, 21–23 May 2002.
12. Buckler, L.M. The Use of Phased Array Radars at Civilian Airports. In Proceedings of the IEEE Phased Array
Systems and Technology Conference, Boston, MA, USA, 15–18 October 1996.
13. Petrochilos, N.; Galati, G.; Mene, L.; Piracci, E. Separation of Multiple Secondary Surveillance Radar Sources
in a Real Environment by a Novel Projection Algorithm. In Proceedings of the IEEE Conference on Signal
Processing and Information Technology, Athens, Greece, 21 December 2005.
14. Federal Aviation Administration. Multifunction Phased Array Radar (MPAR) Notional Functional Requirements
Document; Federal Aviation Administration: Washington, DC, USA, 2012.
15. Federal Aviation Administration. Houston/I90 TRACON Automatic Dependent Surveillance—Broadcast (ADS-B)
Flight Inspection Analysis and Coverage Report; Federal Aviation Administration: Washington, DC, USA, 2011.
16. Federal Aviation Administration. Public ADS-B Performance Report (PAPR) User’s Guide; Federal Aviation
Administration: Washington, DC, USA, 2016.
17. Ali, B.S.; Schuster, W.; Ochien, W.; Majumdar, A.; Kian, C.T. A Study of ADS-B Data Evaluation and Related
Problems. Proceeding of the 2013 International Technical Meeting of the Institute of Navigation, San Diego,
CA, USA, 27–29 January 2013.
18. Zhang, J.; Lui, W.; Zhu, Y. Study of ADS-B data evaluation. Chin. J. Aeronaut. 2011, 24, 461–466. [CrossRef]
19. Baud, O.; Honore, N.; Taupin, O. Radar/ADS-B Data Fusion Architecture for Experimentation Purpose.
In Proceedings of the IEEE Conference on Information Fusion, Florence, Italy, 10–13 July 2006.
20. Castle, M.W.; Trinh, T.; Mayer, C.; Parry, C. Evaluation of Separation Performance with ADS-B at the
Philadelphia Key Site. In Proceedings of the IEEE/AIAA 29th Digital Avionics Systems Conference,
Salt Lake City, UT, USA, 3–7 October 2010.
21. Wolff, C. Radar Tutorial. Available online: http://www.radartutorial.eu/druck/Book2.pdf (accessed on
15 February 2017).
22. Tang, W.; Howell, G.; Tsai, Y.H. Barometric altimeter short-term accuracy analysis. IEEE Aerosp. Electron.
Syst. Mag. 2005, 20, 24–26. [CrossRef]
23. Kaplan, E.D.; Hegarty, C. Understanding GPS: Principles and Applications; Artech House: Norwood, MA,
USA, 2005.
24. Wolff, C. Radar Sets. In Radar Basics; RadarTutorial.eu: Berlin, Germany, 1998; pp. 1–11.
25. Jain, P. Classification of RADARs. Available online: https://www.engineersgarage.com/articles/types-of-
radars (accessed on 21 March 2017).
26. Richards, M.A.; Scheer, J.A.; Holm, W.A. Principles of Modern Radar; SciTech Publishing: Raleigh, NC, USA,
2010; ISBN 978-1-891121-52-4.
27. Jain, P. Classification of RADARs. Available online: https://www.engineersgarage.com/articles/types-of-
radars (accessed on 29 April 2017).
28. International Civil Aviation Organization. Manual for the Universal Access Transceiver (UAT); ICAO: Montreal,
QC, Canada, 2003.
29. Wichgers, J.M. Automatic Dependent Surveillance—Broadcast. In Digital Avionics Handbook, 3rd ed.;
CRC Press: Boca Raton, FL, USA, 2015.
30. Federal Aviation Administration. Airworthiness Approval of Automatic Dependent Surveillance—Broadcast
(ADS-B) Out Systems; Federal Aviation Administration: Washington, DC, USA, 2015.
31. Federal Aviation Administration. NextGen: Automatic Dependent Surveillance—Broadcast. Available online:
https://www.faa.gov/nextgen/update/progress_and_plans/adsb/ (accessed on 4 May 2017).
32. Nikoleris, A.; Erzberger, H.; Paielli, R.; Chu, Y. Performance of an Automated System for Control of Traffic
in Terminal Airspace. In Proceedings of the 16th AIAA Aviation Technology, Integration, and Operations
Conference, Washington, DC, USA, 13–17 June 2016.

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).