Proceedings of the

44th IEEE Conference on Decision and Control, and

TuC15.1
the European Control Conference 2005
Seville, Spain, December 12-15, 2005

Airborne Technology for Distributed Air Traffic Management

Dr. Michael R. C. Jackson, Dr. Vivek Sharma, Dr. Christine M. Haissig, and Dr. Michael Elgersma

Abstract — Worldwide air traffic levels are growing at a rate support tools, are being developed and fielded to assist
expected to double the current traffic level by 2020. The controllers in handling more aircraft and larger airspace per
current technology Air Traffic Control systems are stretched to controller station with some success [38]. These systems are
their limit and are prone to large delays during the peak typically built upon predicting the trajectories of aircraft and
summer travel season. There is doubt that the current systems
scanning for future conflicts. Early conflict detection is
can be scaled up to meet the expected demand levels.
Many Air Traffic Management automation systems have useful, but it is still up to the controller to determine
been proposed to increase controller capability, and some are appropriate action and communicate instructions to the
in operation. While ATM automation systems will help handle aircraft verbally. It is doubtful that these systems can grow
more traffic, it is still doubtful that they can grow to meet the to meet the doubling in traffic levels foreseen.
doubling in traffic levels foreseen. A variety of airborne technologies have the potential to
This paper presents an introduction to Distributed ATM – further reduce controller workload by distributing the air
using the capability of airborne electronic systems to further traffic management responsibility between the aircraft and
relieve the controller workload. An overview of avionics
ground-based systems, and many concepts have been
capabilities is presented, followed by a detailed description of
five specific examples of airborne capability that can be used to developed to take advantage of these capabilities to improve
increase airspace capacity, as listed below. how aircraft are controlled. In the simplest concepts the
1. An onboard method to control an aircraft to cross a aircraft send accurate current state and future intent data to
terminal area waypoint at a Required Time of Arrival. ground automation systems to improve their tracking and
2. A trajectory negotiation process whereby the ground- prediction accuracy [5], [9], [18], [29], [62], [65], [66], [67].
based ATM system uses the 4D predicted trajectory In the most complex concepts the aircraft automation
computed by the aircraft, gives the aircraft RTA systems monitor their own traffic situation and take full
constraints to solve traffic conflicts, and contracts the responsibility for their own separation assurance [2]. There
aircraft to stay within a specified tolerance of the predicted
is also a broad spectrum of integrated air/ground ATM
4D trajectory.
3. A formation flight system whereby multiple aircraft can be concepts between these two extremes [1], [6], [12], [15],
flown close together and controlled as a single aircraft. [16], [39], [42], [45], [74].
4. An onboard alerting system for closely spaced parallel
approaches that can increase the capacity of an airport in II. BACKGROUND OVERVIEW
instrument meteorological conditions.
5. An onboard collision avoidance system that generates a A. Flight Management System (FMS)
conflict-free trajectory through complex airspace. Most modern transport aircraft are equipped with a Flight
Management System (FMS) that assists the pilot with flight
I. INTRODUCTION planning and automatically controlling the aircraft. The FMS

T here is doubt that the current ATM systems can be

scaled up to meet expected demand levels, since further
subdividing airspace sectors into smaller pieces to reduce the
blends all navigation data on the aircraft from various
sources to maintain an accurate status of aircraft position,
velocity, and atmospheric conditions. The pilot enters the
number of aircraft that a controller is working increases the flight plan into the FMS, including lateral route and vertical
number of airspace handoffs for each flight, and controllers requirements, and the FMS predicts the full aircraft
need a minimum size of airspace to allow room for trajectory including altitude, speed, fuel, and time. Some
maneuvering aircraft to avoid conflicts. The procedures for FMS systems are capable of optimizing the vertical profile
handling flights manually introduce delay by putting aircraft to minimize the cost of fuel consumption and flight time by
on standard tracks at similar airspeeds so that the controllers selecting cruise altitude(s) and climb/cruise/descent
can deal with them [47]. airspeeds. The FMS can automatically guide the aircraft to
Air Traffic Management automation systems, or decision follow the flight plan, sending appropriate commands to the
autopilot and autothrottle systems.
Manuscript received May 1, 2005. B. Automatic Dependent Surveillance-Broadcast (ADS-B)
M. R. C. Jackson, V. Sharma, C. M. Haissig, and M. Elgersma are with An ADS-B equipped aircraft broadcasts a message via a
Honeywell Aerospace Electronic Systems, Guidance, Navigation and digital datalink, which includes position, velocity, and
Control Center of Excellence, 3660 Technology Drive, Minneapolis, MN
55418 (e-mail: Mike.Jackson @ honeywell.com, Vivek.Sharma @
possibly other information. Other ADS-B equipped aircraft
honeywell.com, Christine.Haissig @ honeywell.com, Mike.Elgersma @ or ground systems can receive this information and use it for
honeywell.com).

0-7803-9568-9/05/$20.00 ©2005 IEEE 3947

distributed air traffic management. References [3], [4], [24]- or closure rate may also be shown. Traffic information may
[29], [52], [53], and [60] provide more information on be obtained from multiple sources (ADS-B, TIS-B, TCAS).
current and potential use of the ADS-B technology. The operational goals of CDTI applications are to
A typical ADS-B message may contain one or more improve the safety and efficiency of flight operations
reports: either periodic reports (State Vector Report or Mode through enhanced surveillance capability [54][57]. The
Status Report), or on-condition reports (Air Referenced CDTI may enhance current air traffic operations by
Velocity Report, Target State Report, or Trajectory Change providing enhanced visual acquisition and visual
Report) [53]. approaches, and in-trail climb and descent in non-radar
ADS-B technology aims to improve aviation safety by airspace.
giving pilots and controllers reliable, accurate, real-time
F. Conflict Detection, Prevention, and, Resolution
information about aviation traffic, even where radar is
ineffective or unavailable. ADS-B technology can also be A conflict is defined as the condition under which an
scaled and adapted for use in general aviation and in ground aircraft’s protected airspace zone (PAZ) is violated. The
vehicles. PAZ is a variable sized zone based on legal separation
requirements where they exist. For example, a typical PAZ
C. Traffic Information Service (TIS-B) may be a cylinder of 5 nm radius and 2000 ft altitude. A
TIS-B is a surveillance service that derives traffic smaller zone, called the collision avoidance zone (CAZ),
information from ground surveillance sources and defines a separation standard between aircraft that represents
broadcasts to equipped aircraft or surface vehicles with the a risk for a collision. A typical CAZ may have a 0.15 nm
intention of supporting aircraft surveillance applications radius and 600 ft altitude.
[28], [56]. TIS-B service is provided where there is both A Conflict Detection (CD) system alerts the pilot that a
adequate surveillance coverage from ground sensors loss of separation would occur if the aircraft involved
(Surveillance Coverage Volume) and adequate RF coverage maintain their current flight state and intent. One definition
from TIS-B Ground Stations (RF Coverage Volume). An of loss of separation is if at some point of time, the minimal
aircraft in the RF Coverage Volume can receive traffic distance between these two aircraft, called the Closest Point
information on targets located in the Surveillance Coverage of Approach (CPA), is less than the PAZ/CAZ limits.
Volume for which the Ground Station is providing TIS-B A Conflict Prevention (CP) system provides the pilot with
messages. knowledge of potential conflicts. A "potential conflict"
References [10] and [11] propose an extension to TIS-B exists if a conflict may be created through a change of the
called TIS-Contract, or TIS-C, for providing traffic ownship state and/or intent.
information in a distributed fashion. After a conflict has been detected and an alert issued, the
conflict must be resolved. A Conflict Resolution (CR)
D. Flight Information Services (FIS-B)
system provides the pilot with advisories which must be
Flight Information Services-Broadcast is an automated followed by the pilot unless the pilot deems them unsafe or
datalink system that will provide timely information to has better information.
pilots, such as weather graphics and text, Special Use A variety of research exists on algorithm development and
Airspace (SUA) information, and Notice to Airmen analysis for conflict detection, prevention, and resolution,
(NOTAMs). The FIS-B datalink systems use a one-way both ground and air-based. [13], [21], [34], [37], [40], [41],
broadcast protocol. Information flows only from the server [46], [49], [51], [75].
to the receiving aircraft without the need for the aircraft to
request information from the server, nor to acknowledge G. Autonomous Operations Planner (AOP)
receipt. NASA is developing a flight deck decision support tool to
The goal of FIS-B datalink systems is to provide weather support research into autonomous operations in a future
and other non-control flight advisory information to pilots in distributed air/ground traffic management environment [2],
a way that enhances their awareness of flight conditions and [1], [6], [74]. This interactive real-time decision aid, referred
enables better strategic route planning consistent with to as the Autonomous Operations Planner (AOP), will
guidance provided by Federal Aviation Regulations and enable the flight crew to plan autonomously in the presence
corporate policy [55]. of dense traffic and complex flight management constraints.
The crew interacts with the AOP to perform trajectory
E. Cockpit Display of Traffic Information (CDTI)
planning that accounts for conflicts with traffic hazards,
The CDTI is the pilot interface portion of the surveillance ownship aircraft performance limitations, system flow
system [27]. CDTI functions include a traffic display constraints that are imposed by the ATSP, airspace
function, associated control functions, annunciation constraints, such as severe weather and dynamic SUA, and
functions, and alert functions. The traffic display function operator flight goals, such as efficiency and schedule. The
may present information on a dedicated display device or a AOP assists crew decision-making through presentation of
shared/multi-function display (MFD) device. Additional the most relevant information, the accommodation of crew
information about the traffic such as ground speed, distance, planning preferences, alerting advisories, and automated

3948
negotiations with crews of other aircraft (if needed) and Many FMS systems today can already automatically
ground-based air traffic controllers. modify the aircraft speed profile during cruise to meet an
RTA constraint. However, the FMS RTA systems fielded
H. Merging and In-Trail Separation today can only adjust the aircraft speed profile prior to the
A merging and self-spacing capability provides a method top-of-descent point, so they are not useful for controlling
for aircraft that are maneuvering in a free-flight environment the time of arrival in the terminal phase of flight.
to enter a terminal area in an organized, safe, and efficient Honeywell has developed an advanced RTA system
fashion. The merging capability provides the ATSP with a capable of automatically adjusting the speed profile through
means of specifying to an aircraft how it is to merge with the descent phase to meet an RTA low in the descent phase
other aircraft. Two methods are envisioned: 1) Specification [58]. There are two main challenges in accomplishing RTA-
of a ghost aircraft trajectory, with which the aircraft then in-Descent. First is ensuring that the aircraft has the
performs self-spacing; 2) Specification of an RTA at a capability of adjusting the aircraft speed in the descent
merge point that is based upon the lead aircraft’s crossing phase. Since the descent trajectories are typically generated
time plus an appropriate time separation. at (or close to) idle throttle setting there is no automatic way
The self-spacing capability provides a means for the flight the FMS can decelerate. Since it is unlikely that the FMS
crew to maintain a desired spacing (in time) behind a lead will be given automatic control of the speed-brakes, the only
way to ensure the FMS can decelerate is to plan the
aircraft in instrument meteorological conditions. The time
trajectory at an off-idle throttle setting, resulting in a more
separation between the two aircraft is calculated by the ATM
shallow descent path. Since this would result in increased
automation system. The lead aircraft provides state and fuel consumption, this would only be used when accurate
trajectory information using ADS-B to assist the trailing time control is important.
aircraft in maintaining the desired separation. The self- The second challenge is efficiently computing the speed
spacing system provides the trailing aircraft with guidance adjustment required to meet the RTA constraint, brought
cues to the autopilot/autothrottle and the pilot displays. This about by the desire to have the speed adjustment appear to
capability is described in further detail in [22], [23], [33], the pilot as a simple change to the descent Mach and CAS
[35], [44], [48], and [50]. profile. The sensitivity of time of arrival to changes in the
descent speed profile is not known in current FMSs, leaving
The following five sections describe specific airborne only an iterative search to find a feasible speed adjustment.
capabilities that the authors have contributed to developing. Since each iteration requires a full trajectory prediction to
compute the arrival time, this solution is impractical to
III. TIME-BASED AIRCRAFT CONTROL implement onboard present-day FMSs. Going forward,
Honeywell has developed a method to analytically compute
While traditional ATC methods separate aircraft by space,
the sensitivity of the arrival time to a generic speed
several concepts are founded on time-based separation. The
adjustment parameter (SAP) that can include any
simplest time-based systems ensure that aircraft cross choke-
combination of Mach, CAS, or other means of adjusting the
points at specified times. These systems can require no
speed profile by a single parameter, such as cost index.
special avionic equipment if the controller “closes the loop”
The following derivation describes the basic approach
by issuing speed instructions to pilots to keep them on time.
used to compute the speed adjustment. First, consider an
But, the main promise of the time-based systems to relieve
equation that approximates how an FMS computes the
controller workload is by issuing crossing time clearances
estimate time of arrival (ETA) using numerical integration
directly to the aircraft and allowing the aircraft to adjust
over a series of steps that include waypoints, vertical events,
speed autonomously to meet the crossing restriction.
and intermediate integration steps for large legs.
Reference [67] presents the results of a study investigating
a ground-based ATM system that uses accurate 4D LegDist
ETA = ∑ + Current _ Time
trajectories computed by the FMS to improve the accuracy WPTS GndSpd
of the conflict detection, and is representative of several
other time-based ATM systems. This method involves nego- The sensitivity of the ETA to a speed adjustment
tiating 4D trajectories between the ground controllers and parameter can be found by simple differentiation to be:
the aircraft FMS prior to departure, and as needed during δETA − LegDist δGndSpd
flight, to ensure that the strategic flight plan remains conflict = ∑
δSAP wpts GndSpd 2 δSAP
free. To deal with changing conditions that cause unplanned
conflicts, RTA constraints are uplinked to the aircraft, and This sensitivity can be computed by the FMS during the
the FMS replans the trajectory to meet the constraint. While normal trajectory prediction calculations with only a small
this strategy reduces controller workload by increasing increase in computations. Once the sensitivity is known, the
trajectory prediction accuracy, the study found that it raised speed adjustment can be computed by representing the ETA
controller workload by requiring a trajectory negotiation as a Taylor’s series expansion
process to deal with effects of remaining uncertainties in the
predictions. In essence, this approach only uses the FMS to δETA
ETA( SAP + ∆SAP ) ≅ ETA( SAP ) + ∆SAP = RTA
increase the accuracy of trajectory prediction but still closes δSAP
the feedback loop through the controller.

3949
 and solving for the speed adjustment trajectory objectives while monitoring the compliance with
− TimeError the 4D trajectory contract and taking further action to stay
∆SAP = within the bubble only when necessary.
δETA It is preferable to fly the aircraft with smooth airspeed
δSAP commands chosen to achieve the trajectory constraints
where TimeError = ETA − RTA imposed by ATM. This strategy responds more smoothly to
While the above derivation works well for small time wind modeling errors, essentially attempting to remove time
errors, nonlinearities in the equations of motion and errors before crossing the RTA waypoint, instead of
performance model merit a higher order solution for large instantaneously. Since this opens up the possibility of
errors using a 2nd order partial derivative. The details of this leaving the bubble-in-tube contract, the predicted aircraft
approach are described in [58]. trajectory must be monitored to ensure that it does not leave
To ensure that the ground-based conflict probe retains the tube contract. If it is projected to leave the contract, the
accuracy with the aircraft autonomously changing the speed flight crew / FMS has several options available: a)
profile, the 4D trajectory can be periodically down-linked to Modify/add trajectory constraints in the vicinity of the
the ATM system. By placing the RTA constraint(s) at predicted violation. E.g., if the aircraft is predicted to leave
appropriate choke points or aircraft trajectory crossing the front or back of the bubble, an additional RTA could be
points, the autonomous speed modifications by the aircraft added to the flight plan. b) Switch aircraft guidance modes
ensure that the trajectory remains conflict-free with a to explicitly track the bubble-in-tube contract with a
minimum of controller workload. The resulting aircraft feedback guidance law. c) Reopen trajectory negotiation by
trajectory retains the desired FMS speed profile shape down linking the current predicted trajectory that meets the
familiar to pilots, while pushing the errors due to the set of trajectory constraints previously uplinked by ATM
uncertainty into portions of the trajectory that are not critical automation.
to the traffic situation.
V. CIVIL FORMATION FLIGHT
IV. AIR/GROUND TRAJECTORY NEGOTIATION Military transport aircraft such as the C-130 and C-17
In addition to [67] discussed above, several other regularly fly in formation as an effective way to move large
trajectory negation concepts have been presented in the numbers of aircraft safely. Procedures and on-board avionics
literature [1] [71] [72]. In these concepts, the ATM system have been developed to make formation flight possible in all
uplinks time constraints to the aircraft, which computes a weather conditions. Military aircraft typically fly with less
compliant 4D trajectory and downlinks the trajectory to the than 1 nm in-trail separation, which is one-half to one-fourth
ATM system. As discussed in the previous section, using the of the spacing standards used today in terminal airspace and
aircraft FMS to generate an accurate and flyable 4D substantially less than the spacing in en route and oceanic
trajectory and downlinking it to the ATM system for conflict environments [14], [32], [47].
probe has some benefit to reducing controller workload, but It is interesting to consider whether the procedures and
further benefit could be realized by transferring additional avionics developed for military formation flight could be
responsibility to the aircraft to automatically adjust the speed adapted for civil formation flight. Civil formation flight,
profile to meet the specified time constraints. where aircraft take on some self-separation responsibility,
As part of the NASA VAMS program, Honeywell and could complement other initiatives to reduce or better
Seagull Technologies developed a detailed trajectory manage aircraft spacing and alleviate congestion in oceanic,
negotiation process that takes advantage of additional en route or terminal air space. Air traffic controller workload
airborne capability to further reduce controller workload might be reduced, for example, if a controller could provide
[61], [63], [64]. This operational concept considers the
route deviations to the lead aircraft of a civil formation and
negotiated 4D trajectory to be a contract between the ATM
all other aircraft would follow the lead.
provider and the aircraft. The aircraft is expected to stay
within a specified tolerance of the negotiated trajectory or Military formation flight typically uses on-board avionics
inform the controller if that cannot be achieved. to assist the flight crew and provide the capability to fly in
Several other references have proposed 4D trajectory low visibility (Figure 1). There is a similarity between the
contracts, such as [72], however, these studies provide no avionics required for military formation flight and those that
method for the aircraft to track the contract other than are available for commercial aircraft. While the military may
explicitly and continuously controlling the aircraft to track choose to use specialized communication links and sensing,
the center of the 4D trajectory bubble. This has several and typically have a mission computer acting as the flight
negative aspects. Under realistic wind modeling errors, this management system, the avionics functions are essentially
tracking would require airspeed deviations on the order of the same as those available commercially.
the wind modeling errors, along with associated throttle
activity, reduced passenger comfort, increased fuel
consumption, and increased engine wear.
Honeywell and Seagull proposed a blended approach that
uses the FMS RTA capability to maintain the gross

3950
 Flight Crew Interface

ETCAS
Displays
Surveillance data Formation Flight System
Formation Display
Formation commands Formation Flight Algorithms Display
Position, velocity from other formation aircraft • Surveillance measurements data Guidance Display
Direct range/bearing/altitude measurements • Surveillance data extrapolation
• Surveillance accuracy monitoring
• Proximity detection
Interaircraft •Deviation/guidance commands Flight Crew Input
communication
Operator MCK
Interaircraft datalink data
• Surveillance data exchange
Formation commands • Flight commands
Position, velocity to other formation aircraft • Text messages Specialized Proximity Alerting

Nav and other state data Surveillance data

On-Board Sensors
Mission Computer/
Mode S/IFF GPS Air Data
Flight Management
(military only) System
IRU Radar Altimeter

Air Propulsion and Data Mgmt Computer

Figure 1. Example Formation Flight System Architecture

• Surveillance sources could include datalinked that runway, not on the aircraft on the parallel runway.
navigation data from other formation aircraft and/or a When Instrument Meteorological Conditions are in effect,
direct range and bearing measurement system such as aircraft can only operate independently on parallel runways
TCAS/Mode S [13]. separated by at least 4300 ft. With dependent operations,
• A low-bandwidth inter-aircraft datalink to relay airport landing capacity decreases dramatically. This affects
navigation data and formation information, such as many major airports across the world.
Automatic Dependent Surveillance-Broadcast [52]. Alerting systems that assist the controllers and/or the
• A flight crew interface that includes a cockpit display of pilots have the potential to increase airport capacity
traffic information and deviations from assigned substantially by providing arrival rates during Instrument
position and the ability to select which formation Meteorological Conditions that approach those during
aircraft to follow and the required relative spacing. Visual Meteorological Conditions. The Precision Runway
• An optional link to a flight management system or flight Monitor (PRM) is a ground-based alerting system based on a
control system for autocoupled flight. high update rate, high accuracy surveillance system
There are challenges to implementing civil formation combined with special display software and procedures.
flight. Procedures need to be established to form and PRM allows aircraft to fly independent approaches for
dissolve formations, including transferring separation runways spaced as close as 3400 ft. Minneapolis-St. Paul
responsibility from the controllers to the flight crew. and St. Louis airports have installed the Precision Runway
Military procedures are not directly applicable, since the Monitor and are currently evaluating it.
formation lead controls the formation. For a civil formation, With Differential Global Positioning System (DGPS)
the responsibility of controlling the formation spacing and landing systems for precision navigation and aircraft-to-
prescribing formation maneuvers largely will be ceded to aircraft communications, combined with Automatic
ATC. ATC is responsible for aircraft separation and must Dependent Surveillance-Broadcast (ADS-B), it is possible to
play an integral role in the creation and dissolution of develop on-board alerting systems for closely spaced parallel
formations. ATC must permit an aircraft to join a formation approaches. Rockwell-Collins and United Airlines are
and specify the lead aircraft, in-trail spacing, and the lateral collaborating on an airborne alerting system for closely
spacing. Once an aircraft leaves a formation, ATC must take spaced dependent approaches that they plan to demonstrate
back controlling authority. at the San Francisco airport, where the runway spacing is
750 ft.
VI. CLOSELY SPACED PARALLEL APPROACHES Honeywell and NASA Langley Research Center
demonstrated in 1999 an airborne alerting system called
Aircraft landing on parallel runways during Visual
AILS (Airborne Information for Lateral Spacing) that was
Meteorological Conditions can operate independently as
designed for independent closely spaced parallel approaches
long as the runways are at least 700 ft apart. Pilots are
to runways separated by 2500 ft to 4300 ft. [70]. A
responsible for separation, and the longitudinal spacing of
prototype system was tested at NASA Wallops in
each aircraft along a runway depends only on the aircraft on

3951
September, 1999, and demonstrated in Minneapolis two current states forward a finite amount of time to see if it will
months later using the NASA 757 ARIES aircraft and a penetrate the other aircraft’s protected zone (a linear
Honeywell Gulfstream-IV. In the both flights the distance above and below the aircraft), and an elliptical
Honeywell aircraft flew intrusion maneuvers at the NASA protection area in the horizontal plane. The forward
757 and the system alerted the pilot to the threat and projection assumes that the aircraft will continue turning at
instructed the pilot when to perform an emergency escape the current turn rate, but also that the aircraft may roll level
maneuver. at any time. This concept is known as the AILS “fan” and
An airborne alerting system for closely spaced parallel represents an added level of safety check and conservatism.
approaches has three key elements—precision navigation,
accurate state and approach (intent) information for the other VII. AIRBORNE COLLISION AVOIDANCE
aircraft in the area, and algorithms to alert the pilot to This section examines the use of Laplace’s equation for
dangerous situations. Precision navigation is provided by the airborne collision avoidance for UAVs. While not immed-
Differential Global Positioning System (DGPS). Differential iately applicable to commercial airspace conflict prevention,
corrections to the GPS signal are uplinked from a GPS this method illustrates advanced avionics capabilities in
Landing System ground station, such as a Local Area development that may be useful in the future. This technique
Augmentation System (LAAS). constructs paths, r(t), through a 3D domain by assigning a
Each aircraft broadcasts its position, velocity, and potential value of v(r)=0 for r on any boundaries or obstacle,
approach using ADS-B. The airborne alerting system uses and a potential of v(r)=1 for r on the goal region. Then
the received ADS-B information from the other aircraft in Laplace’s equation is solved in the interior of the 3D region,
combination with its own aircraft data to determine if either guaranteeing no local minima in the interior of the domain,
aircraft is a threat to the other with a specialized alerting leaving a global maximum of v(r)=1 for r on the goal region,
algorithm such as AILS. The ADS-B information is also and global minima of v(r)=0 for r on any boundaries or
broadcast to the ground so that controllers can stay informed obstacle. A path from any initial point, r(0), to the goal, is
of the status of aircraft on parallel approaches. made by following the gradient of the potential, v.
The Emergency Escape Maneuver is a fixed procedure Previous applications of Laplacian path-planning include
[7], [8], [43], and [69]. An analogy for paths obtained by
consisting of a climbing turn. The procedural EEM may
Laplace’s equation is to apply a voltage of 0 to all boundary
change for different airports or different runways, but it is
and obstacle locations, a voltage of 1 to goal region, fill
independent of the blunder characteristics. A procedural interior region with a conductor, then electrons will follow
EEM simplifies system design since no aircraft-to-aircraft paths from anywhere in the interior to the goal region.
maneuver coordination is required. It also simplifies pilot Laplace’s equation sets the divergence of a potential to
training, since the EEM is fixed and can be presented on an zero in the interior of a domain. Solutions of Laplace’s
approach plate in the same way as a missed approach equation are harmonic functions, which have no local
procedure. The major drawback of the procedural EEM is minima in the interior of their domain.
that it has to work for all possible blunder scenarios. Once Numerical solutions of Laplace’s equation are obtained by
the pilot performs the EEM, there is nothing more that the gridding the domain, then iteratively setting the potential at
alerting system can do to protect the aircraft, since active each interior point, equal to the average of its nearest
guidance is not provided. To address this drawback, the neighbors. By varying the grid size (halving or doubling cell
system was designed to hand-off to TCAS as soon as length at each step) from the crudest that stills leaves paths
possible, but after no less than 5 seconds. TCAS can then between obstacles, to the finest that is required for smooth
provide vertical maneuver coordination, if necessary, as the paths, the iteration can be made to converge in a time
aircraft performs the EEM. proportional to the number, N, of cells in the finest grid. The
The AILS alerting algorithms provide multiple levels of solution on crude grids is cheap, and is used to initialize the
alerting for pairs of aircraft that are in parallel approach solution on finer grids. This multigrid technique is described
situations. AILS uses current aircraft states as well as known in [68] and is applied to robotic path planning in [43].
“intent” information to project ahead for threat Textbook convergence proofs for empty domains, give the
total number of computations to be c*N, where N is the
determination. This forward projection is based on current
number of cells in the finest grid, and c ≈ 5 is some small
positions, velocities, altitudes, turn rates, climb rates, and
number of iterations at each grid size. That convergence
intended approach path. An alert is raised if one aircraft is speed relies on being able to set the crudest grid cell size
likely to penetrate another’s protected region within a equal to the entire domain for the crudest solution.
specified amount of time. However, we have found that in a domain with obstacles, the
AILS performs two types of alerting checks: Adjacent number of needed iterations is c*(path length)/(path width),
ship threat to own ship and Ownship threat to adjacent ship. since path-width between obstacles, limits the largest cell
Each check includes a caution level indicating a potential for size of the crudest grid that still preserves the topology of the
alert, and a warning level which represents potential computed paths. With largest grid cell size equal to path
imminent collision. width, the number of grid cells along the path is (path
The AILS’s threat evaluation projects each aircraft’s length)/(path width). The iterative process of setting a cell’s

3952
potential equal to the average of its neighbor’s potentials [6] B. Barmore, E. Johnson, D. J. Wing, and R. Barhydt, Airborne
Conflict Management within Confined Airspace in a Piloted Sim. of
propagates a nonzero solution value a distance of one more
DAG-TM Autonomous Aircraft Operations, 5th USA/ATM R&D
grid cell along the path each iteration. So it takes (path Seminar, 2003
length)/(path width) iterations for a nonzero solution to [7] C. I. Connolly, and J. B. Burns, “Path planning using Laplace’s
propagate along the entire path length, when crudest grid equation,” IEEE Int. Conference on Robotics and Automation, 1990.
[8] C. I. Connolly, and R. A. Grupen, “The Application of Harmonic
cell size is equal to path width. After c*(path length)/(path Functions to Robotics,” J. of Robotic Systems, 10(7), pp931-946,
width) iterations, with c ≈ 5, the iteration converges on the 1993.
crudest grid. Using iterations on each finer grid size results [9] R. A. Coppenbarger, G. Kanning, and R. Salcido, Real-Time Data
in the bulk of the work being done by c*(path length)/(path Link of Aircraft Parameters to the Center-TRACON Automation
width) iterations on the finest grid size, for a total number of System (CTAS), 4th USA/Europe ATM R&D Seminar, 2001.
[10] R. Ehrmanntraut, “Enabling Air-Ground Integration: Concept
operations of approximately c*(path length)/(path width)*N, Definition for TIS in the Contract Mode (TIS-C),” IEEE 2003.
where N is the total number of cells in the finest grid. [11] R. Ehrmanntraut, “Enabling Air-Ground Integration: Validation of the
An example region of size 128x128x16 is shown in Figure Traffic Information Service in the Contract Mode (TIS-C) Concept
over VDL Mode 2 with the ACTS Simulator,” Eurocontrol, Jan 2004.
2. The low ceiling in the front half of the region forces the
[12] H. Erzberger, Transforming the NAS: The Next Generation Air Traffic

Figure 2. UAV obstacle avoidance path with ceiling constraint

UAV to fly between buildings, while in the rear of the Control System, 24th International Congress of the Aeronautical
region, the high ceiling allows UAV to fly over obstacle Sciences, ICAS 2004, Yokohama, Japan, Aug 29, 2004.
[13] Federal Aviation Administration. Introduction to TCAS II, Version 7
buildings. This example took 6 seconds in Matlab, on a 1.8 (Nov), http://www.arinc.com/tcas. 2000.
GHz Pentium. [14] FAA. Order 7610.4, Special Military Operations, Section 12,
Formation Flight,
http://www.faa.gov/ATpubs/MIL/Ch12/mil1212.html
REFERENCES [15] D. Gianazza, N. Durand, “Separating Air Traffic Flows by Allocating
[1] M. G. Ballin, J. M. Hoekstra, D. J. Wing, and G. W. Lohr, “NASA 3D-Trajectories,” 23rd Digital Avionics Systems Conference, 2004.
Langley and NLR Research of Distributed Air/Ground Traffic [16] S. M. Green, K. Bilimoria, and M. G. Ballin, Distributed Air-Ground
Management,” AIAA-2002-5826. Traffic Management for En route Flight, ATCQ, Vol 9(4), 2001.
[2] M. G. Ballin, V. Sharma, R. A. Vivona, E. J. Johnson, and E. [17] R. Graham, E. Hoffman, C. Pusch, K. Zeghal, “Absolute versus
Ramiscal, “A Flight Deck Decision Support Tool For Autonomous Relative Navigation: Theoretical Considerations from and ATM
Airborne Operations”, AIAA Guidance, Navigation, and Control Perspective,” Eurocontrol ATM Workshop in Capri, Sept. 2002.
Conference and Exhibit, 5-8 August 2002, Monterey, California, [18] S. M. Green, T. Goka, and D. H. Williams, Enabling User
AIAA 2002-4554. Preferences Through Data Exchange, AIAA GN&C Conference,
[3] R. Barhydt, M. T. Palmer, W. W. Chung, and, G. W. Loveness, ADS- 1997.
B Within A Multi-Aircraft Simulation For Distributed Air-Ground [19] C. M. Haissig, “Military Formation Flight as a Model for Increased
Traffic Management, 23rd Digital Avionics Systems Conference, Capacity in Civilian Airspace,” 23rd DASC, 2004.
2004. [20] C. Haissig, B. Corwin, and M. R. Jackson, (1999), “Designing an
[4] Richard Barhydt and AnthonyW. Warren, Newly Enacted Intent Airborne Alerting System for Closely-Spaced Parallel Approaches,”
Changes to ADS-B MASPS: Emphasis on Operations, Compatibility, Paper AIAA-99-3986, AIAA GN&C Conference, 2000.
and Integrity, AIAA. [21] K. A. Harper, S. L. Guarino, A. White, M. L. Hanson, K. D.
[5] D. R. Barker, B. M. Haltli, C. Laqui, P. MacWilliams, K. L. McKee, Bilimoria, and D. Mulfinger, An Agent-Based Approach to Aircraft
“Assessment of Terminal RNAV Mixed Equippage,” 23rd Digital Conflict Resolution with Constraints, AIAA GN&C Conference 2002.
Avionics Systems Conference, Salt Lake City, UT, Oct 24-28, 2004.

3953
[22] C. Hébraud, E. Hoffman, N. Pène, L. Rognin, and K. Zeghal, [53] RTCA DO-242A, Minimum Aviation System Performance Standards
“Assessing the impact of a new air traffic control instruction on flight for Automatic Dependent Surveillance Broadcast (ADS-B).
crew activity,” AIAA GN&C 2004, AIAA-2004-5104. [54] RTCA DO-259, Applications Descriptions for Initial Cockpit Display
[23] A. A. Herndon, J. S. DeArmon, J. Spelman, “Use of Lateral/Parallel of Traffic Information (CDTI) Applications.
FMS Procedures and Implementation Issues,” 23rd DASC, 2004. [55] RTCA DO-267A, Minimum Aviation System Performance Standards
[24] E. Hoffman, D. Ivanescu, C. Shaw, and K. Zeghal, “Effect of ADS-B (MASPS) for Flight Information Services – Broadcast (FIS-B).
Transmission Quality on the Ability of Aircraft to Maintain Spacing in [56] RTCA DO-286A, Minimum Aviation System Performance Standards
a Sequence,” ATCQ Vol 11(3), Sept 2003. (MASPS) for Traffic Information Service – Broadcast (TIS-B),
[25] http://ads-b.com/ [57] RTCA SC-186, Minimum Operational Performance Standards for
[26] http://adsb.tc.faa.gov/ADS-B.htm Cockpit Display of Traffic Information, June 2001.
[27] http://adsb.tc.faa.gov/WG1.htm [58] J. R. Rumbo, M. R. Jackson, B. E. O’Laughlin, Aircraft Control
[28] http://adsb.tc.faa.gov/WG2.htm System for Reaching a Waypoint at a Required Time of Arrival, U.S.
[29] http://www.alaska.faa.gov/capstone/ Patent 6,507,782, Jan 2003.
[30] http://www.alaska.faa.gov/capstone/docs/uat.htm [59] P. Samanant, M. R. Jackson, C. M. Haissig, and B. Corwin,
[31] http://www.eurocontrol.fr/Newsletter/2003/December/AVT/AVT_Ov “CASPER/AILS: An Integrated DGPS/ADS-B Airborne Alerting
erview.htm System for Closely Spaced Parallel Approaches”, AIAA/IEEE PLANS
[32] http://www.faa.gov/programs/oep/Archive/V3/oep%20version%203.0 Conference, March 2000.
/enrte_congestion/er6.html [60] J. Scardina, Overview of the FAA ADS-B Link Decision, FAA, 2002.
[33] J. Hull, B. Barmore, T. Abbott, “Technology-Enabled Airborne [61] D. R. Schleicher, A. S. Huang, B. V. Kiger, and K. A. Ramanoorthy,
Spacing and Merging,” 23rd DASC, 2004. “Benefit-Cost Analysis of a 2022 Point-to-Point ATM Concept,”
[34] I. Hwang, J. Hwang, and C. Tomlin, Flight-Mode-Based Aircraft AIAA GN&C ,2004, AIAA-2004-5410.
Conflict Detection using a Residual-Mean Interacting Multiple Model [62] D. R. Schleicher, E. Jones, D. Dow, and R. A. Coppenbarger,
Algorithm, AIAA GN&C, 2003. AIAA 2003-5340 Improved Lateral Trajectory Prediction through En Route Air-Ground
[35] D. Ivanescu, D. Powell, C. Shaw, E. Hoffman, and K. Zeghal, “Effect Data Exchange, AIAA ATIO Conference, 2002.
of aircraft self-merging in sequence on an airborne collision avoidance [63] D. R. Schleicher, J. Sorensen, B. V. Kiger, and A. Huang, “A Point-
system,” AIAA GN&C Conference 2004, AIAA 2004-4994. to-Point Concept for Improved National Airspace System Capacity,”
[36] M. R. Jackson, P. Samanant, and C. Haissig, “Design and Analysis of AIAA GN&C Conference 2003, AIAA 2003-5768.
Airborne Alerting Algorithms for Closely-Spaced Parallel [64] J. Sorensen, et al., “Massive Point-to-Point (PTP) and On-Demand Air
Approaches,” AIAA GN&C Conference, August 2000. Transportation System Concept Description,” TR04231-01, NASA
[37] W. Johnson, K. D. Bilimoria, L. C. Thomas, H. Q. Lee, and V. Contract No. NAS2-02076, Seagull Technology Inc., 2004.
Battiste, Comparison of Pilot and Automation Generated Conflict [65] D. V. Stapleton, J. J. Cieplak, “Alaska’s Capstone Program – Systems
Resolutions, AIAA GN&C 2003. AIAA 2003-5400 Engineering for Comm. Nav. Surv.,” 23rd DASC, 2004.
[38] D. B. Kirk, Initial Functional Performance Assessment of A Terminal [66] C. Tamvaclis, N. McFarlane, B. Josefsson, “Use of Aircraft Derived
Airspace Conflict Probe Application, AIAA GN&C, 2003. Data for More Efficient ATM Operations,” 23rd DASC, 2004.
[39] J. Krozel, and M. Peters, “Decentralized Control Techniques for [67] D. J. Teutsch, E. Hoffman, “Aircraft in the Future ATM System –
Distributed Air/Ground Traffic Separation,” NASA TR 99RTO36-03. Exploiting the 4D Aircraft Trajectory,” IEEE 2004, 0-7803-8539-X.
[40] J. Kuchar, (1996), “Methodology for Alerting-System Performance [68] U. Trottenberg, C. Oosterlee, and A. Schuller, “Multigrid,” Academic
Evaluation,” J. of Guidance, Control, and Dynamics, 19(2). Press, 2001.
[41] J. K. Kuchar, and L. C. Yang, “A Review of Conflict Detection and [69] Valavinis, Herbert, Kollura and Tsourveloudis, “Mobile Robot
Resolution Modeling Methods,” IEEE Transactions on Intelligent Navigation in 2-D Dynamic Environments Using an Electrostatic
Transportation Systems, Vol. 1, No. 4, Dec. 2000, pp. 179-189. Potential Field,” IEEE Trans Sys and Cybernetics, v 30, 2002.
[42] P. U. Lee, J. S. Mercer, L. Martin, T. Prevot, S. Sheldon, S. Verma, N. [70] M. Waller, and C. Scanlon (editors), NASA Workshop on Flight Deck
Smith, V. Battiste, W. Johnson, R. Mogford, and E. Palmer, “Free Centered Parallel Runway Approaches in Instrument Meteorological
Maneuvering, Trajectory Negotiation, and Self-Spacing Concepts in Conditions, NASA CP 10191, 1996.
Distributed Air-Ground Traffic Management,” 5th USA/Europe ATM [71] D. H. Williams, P. D. Arbuckle, S. M. Green, and W. den Braven,
R&D Seminar, Budapest, Hungary, 2003. Profile Negotiation: An Air/Ground Automation Integration Concept
[43] Z. X. Li, and T. D. Bui, “Robot Path Planning Using Fluid Model,” for Managing Arrival Traffic, AGARD CP, 1993, pp. 22-1 / 22-11.
Journal of Intelligent and Robotic Systems, v. 21, pp29-50, 1998. [72] I. A. B. Wilson, “Trajectory Negotiation in a Multi-Sector
[44] G. W. Lohr, R. M. Oseguera-Lohr, T. S. Abbott, and B. Capron, Environment,” PHARE DOC 97-70-14, June 1998.
“Flight Evaluation of a Time-Based Airborne Inter-Arrival Spacing [73] L. Winder, and J. Kuchar, “Evaluation of Collision Avoidance
Tool,” 5th USA/Europe ATM R&D Seminar, Budapest, Hungary, Maneuvers for Parallel Approach”, Journal of GCD, Vol. 22, No. 6.
2003. [74] D. J. Wing, R. J. Adams, J. A. Duley, B. M. Legan, B. E. Barmore,
[45] S. C. Mohleji, D. Bhadra, P. A. Ostwald, F. P. Wieland, “A Process and D. Moses, “Airborne Use of Traffic Intent Information in a
for Estimating Cost/Benefits of Future Air Transportation System Distributed Air-Ground Traffic Management Concept”.
Operational Concepts Based on 4D Navigation,” 23rd DASC, 2004. [75] S. Wollkind, J. Valasek, T. Ioerger, “Automated Conflict Resolution
[46] Huy-Hoang Nguyen, Survey of Coordination of En Route Air Traffic for Air Traffic Management Using Cooperative Multiagent
Conflicts Resolution Modelling Methods, Proceedings of the Research Negotiation,” AIAA GN&C Conference 2004, AIAA 2004-4992.
Informatic Vietnam & Francophone Conference, Hanoi, 2003.
[47] M. Nolan, Fundamentals of ATC, Belmont, CA, Brooks/Cole, 2004
[48] R. M. Osequera-Lohr, G. W. Lohr, T. Abbott, T. M. Eischeid,
“Evaluation of Operational Procedures for Using a Time-Based
Airborne Interarrival Spacing Tool,” AIAA-2002-5824.
[49] A. U. Raghunathan, V. Gopal, D. Subramanian, L. T. Biegler, and T.
Samad, 3D Conflict Resolution of Multiple Aircraft via Dynamic
Optimization, AIAA GN&C Conference, 2003. AIAA 2003-5675
[50] E. Ramiscal, J. Murdoch, J. McNabb, “Feasibility of Self-Separation
and Sequencing During GPS Instrument Approaches,” 23rd DASC,
2004.
[51] Jie Rong, Sangeeta Bokadia, Surya Shandy, and John Valasek,
Hierarchichal Agent Based System For General Aviation CD&R
Under Free Flight, AIAA GN&C, 2002. AIAA 2002-4553
[52] RTCA/DO-242, Minimum Aviation System Performance Standards
for Automatic Dependent Surveillance-Broadcast (ADS-B), 1998.

3954