This article was downloaded by: [DUT Library]

On: 04 October 2014, At: 07:36

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK

The International Journal of

Aviation Psychology
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/hiap20

Supporting 4D Trajectory
Revisions on the Flight Deck:
Design of a Human–Machine
Interface
a a
Bart J. A. van Marwijk , Clark Borst , Mark Mulder
a a a
, Max Mulder & Marinus M. van Paassen
a
Department of Aerospace Engineering, Control &
Simulation Division , Delft University of Technology ,
Delft, The Netherlands
Published online: 13 Jan 2011.

To cite this article: Bart J. A. van Marwijk , Clark Borst , Mark Mulder , Max Mulder
& Marinus M. van Paassen (2011) Supporting 4D Trajectory Revisions on the Flight
Deck: Design of a Human–Machine Interface, The International Journal of Aviation
Psychology, 21:1, 35-61, DOI: 10.1080/10508414.2011.537559

To link to this article: http://dx.doi.org/10.1080/10508414.2011.537559

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any
losses, actions, claims, proceedings, demands, costs, expenses, damages,
 and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.

This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions
Downloaded by [DUT Library] at 07:36 04 October 2014
 THE INTERNATIONAL JOURNAL OF AVIATION PSYCHOLOGY, 21(1), 35–61
Copyright © 2011 Taylor & Francis Group, LLC
ISSN: 1050-8414 print / 1532-7108 online
DOI: 10.1080/10508414.2011.537559

Supporting 4D Trajectory Revisions

on the Flight Deck: Design
of a Human–Machine Interface
Downloaded by [DUT Library] at 07:36 04 October 2014

Bart J. A. van Marwijk, Clark Borst, Mark Mulder,

Max Mulder, and Marinus M. van Paassen
Department of Aerospace Engineering, Control & Simulation Division,
Delft University of Technology, Delft, The Netherlands

To accomplish air traffic growth in a safe and efficient way, future air traffic manage-
ment concepts require aircraft to accurately execute 4-dimensional (4D) trajectories.
A trajectory planned prior to takeoff, might, however, require in-flight revision. To
support the flight crew in their task of accurately replanning a flight plan up to a me-
tering fix, in 4 dimensions, a dedicated planning interface has been designed, adopt-
ing a cognitive systems engineering approach. The interface allows direct manipula-
tion of the ground track and the descent profile. Constraints on trajectory planning
are mapped onto alternative waypoint locations, highlighting the possibilities for ac-
ceptable ground track geometry in the horizontal situation display. In the vertical sit-
uation display, these constraints are mapped onto candidate top- and bottom-of-de-
scent locations. It is hypothesized that the designed interface enables pilots to
efficiently plan suitable 4D trajectories, while allowing for adaptive behavior and
supporting situation awareness, even under high workload conditions.

To increase airspace capacity, future air traffic management (ATM) environments

will not only require greater diversity and flexibility in the routes that can be
flown, but also greater accuracy and timeliness with which aircraft adhere to these
routes. This has major consequences for both ground-based ATM and airborne
navigation planning. Indeed, in most of the proposed new ATM systems, the capa-
bility to accurately plan, implement, and execute a flight plan in four dimensions,

Correspondence should be sent to Mark Mulder, Department of Aerospace Engineering, Control &
Simulation Division, Delft University of Technology, Kluyverweg 1, 2629 HS Delft, The Netherlands.
E-mail: mark.mulder@tudelft.nl
 36 VAN MARWIJK ET AL.

that is, in space and time, is a central ingredient (Joint Planning and Development
Office, 2007; SESAR Consortium, 2007; Swenson, Barhydt, & Landis, 2006) The
planning, guidance, and navigation tasks of the flight crew will change when ad-
hering to strict time constraints becomes of key importance.
Currently, airborne planning, implementation, and execution of a flight plan is
automated with the help of the flight management system (FMS). Although the
FMS has evolved at an exceptional rate in available features and functionality
(Lidén, 1994), reprogramming a flight plan during flight can be a cumbersome
task. The specification of the sequence of waypoints, flight levels, and speed and
time constraints typically needs to be entered alphanumerically through the key-
pad of the command and control display unit. Not surprisingly, in the case of ob-
Downloaded by [DUT Library] at 07:36 04 October 2014

structive weather cells, pilots often do not bother with altering the flight plan in the
FMS but simply request permission from air traffic control (ATC) to deviate from
it using the autopilot mode control panel. This provides a workaround in tradi-
tional ATM operations, be it at the cost of reliable arrival time and fuel burn pre-
dictions. Needless to say, deliberate deviation from contracted trajectories is in
contrast to the plan-based approach envisioned in the future.
Increasing demands on punctuality in future ATM environments, in particular
the American NextGen (Joint Planning and Development Office, 2007) and Euro-
pean Single European Sky ATM Research (SESAR) concepts (SESAR Consor-
tium, 2007), will make airborne flight plan amendment increasingly complex. The
need for pilots to manipulate the powerful functionality of the FMS quickly and
accurately to perform this task proficiently calls for a redesign of the navigation
planning interface. This article proposes such a redesign to allow direct manipula-
tion of the flight plan during the task of airborne trajectory revision while adhering
to strict time constraints at some waypoint ahead, for example, a traffic merging
point, or other metering fix. The proposed interface allows the flight crew to di-
rectly manipulate their flight plan in space and time, using the support of automa-
tion. In part, this article is a continuation of previous studies exploring this concept
(Huisman, Verhoeven, Houten, & Flohr, 1997; Kaber, Riley, & Tan, 2002; van
Marwijk, Mulder, Mulder, van Paassen, & Borst, 2009; Mulder, Winterberg, van
Paassen, & Mulder, 2010).
A limitation of the proposed interface is that it assumes all possible flight plan
changes can be made without interfering in the flight plans of surrounding traffic.
Other traffic, however, does impose constraints on the availability of airspace for
flight plan changes and vice versa. Ultimately, the interface design should be ex-
tended with visualizations of traffic constraints, for example with traffic conflict
support tools such as those developed by van Dam, Abeloos, Mulder, and van
Paassen (2004) or Battiste, Johnson, and Bochow (2000). However, the aim of the
design presented here is to demonstrate the potential of a redesigned planning in-
terface in simplifying in-flight 4D trajectory modification. Integration with other
information and functionality is beyond the scope of this article.
 SUPPORTING 4D TRAJECTORY REVISIONS 37

The structure of the article is as follows: First, we describe the background of

designing for adaptive behavior and situation awareness and then a cognitive work
analysis of the 4D planning task. Next, we discuss the mapping of constraints on
the display, the underlying computations, and the use of the new cues, leading up
to the discussion and conclusion.

DESIGNING FOR ADAPTIVE BEHAVIOR

AND SITUATION AWARENESS

Although the ongoing development of flight deck automation has resulted in many
Downloaded by [DUT Library] at 07:36 04 October 2014

improvements to safety and crew workload, new problems have also emerged.
Some exemplary categories of suboptimal human–machine interaction reported in
the literature are as follows (Harris, 2004; Sherry, Feary, Polson, Mumaw, &
Palmer, 2000):

• Automation failing to reveal its intent, for example during autonomous mode
changes, undermining the pilot’s ability to foresee the aircraft state and antic-
ipate imminent danger.
• Automation selecting and executing control strategies that the pilot regards
as suboptimal, prompting him or her to bypass or work against the automa-
tion.
• Automation that provides support in anticipated situations, but becomes less
supportive, or even counterproductive, when unanticipated events occur.

In general, any form of automation support that (unintentionally) hides infor-

mation that is important to the task at hand seems to be in conflict with the respon-
sibilities and competencies of the pilot even if it might result in low workload and
good performance during designer-anticipated procedures. As a consequence, we
believe the particular task of trajectory replanning to be ill-suited for allocation to
autonomous or directive automation. Instead, a cooperative process, in which the
automation enables the pilot to function to his or her full potential, is considered
the optimal outcome of the design. Essentially, the design problem can be reduced
to the following, closely related questions:

• How can the subtasks in trajectory replanning be allocated between pilot and
automation?
• What information could help the pilot perform essential tasks?
• How can this information best be presented to the pilot?
• How can the system facilitate pilot decisions?
 38 VAN MARWIJK ET AL.

The Ecological Interface Design (EID) framework can help provide answers to
these questions. EID was developed by Vicente and Rasmussen (1990) as a frame-
work for designing human–machine interfaces for complex socio-technical sys-
tems. Their approach focuses on the constraints that the work domain imposes on
the worker.
When operating a complex socio-technical system, an operator typically expe-
riences problems in predicting, or even perceiving, the system state and the effect
of his or her control actions on it. For example, BASI (1999) reported that out of
1,268 questioned pilots, 11% reported that “they did not always know what mode
the autopilot, autothrottle and flight director was in.” More specifically, when the
operator is unable to form an accurate functional mental model of the system, the
Downloaded by [DUT Library] at 07:36 04 October 2014

control action that will bring the system toward its goal state cannot be selected ef-
fectively. Depending on the system under consideration, there might be various
causes hindering the operator in forming a functional model. These could include,
for example, high-order system dynamics, delayed response to control input, cou-
pled subprocesses, or insufficient knowledge of co-actor intent (human or auto-
mated). EID aims to support the operator in forming a functional model of the sys-
tem. As a consequence, it does not advocate interfaces that provide optimal control
directions based on predefined procedures, but promotes designs in which the op-
portunities for control are revealed. This allows operators to understand, or at least
perceive, the system state and the effect of their control actions on it (van Paassen,
Mulder, & van Dam, 2005). They can then use their expertise in selecting the opti-
mal control strategy under varying circumstances.
Recent studies in interface design have suggested that EID has the potential
to support situation awareness and decision making (Borst, Sjer, Mulder, van
Paassen, & Mulder, 2008; Borst, Suijkerbuijk, Mulder, & van Paassen, 2006; van
Dam, Mulder, & van Paassen, 2008).

SUBJECTIVITY OF OPTIMALITY IN TRAJECTORY

REPLANNING

Trajectory replanning is very much a trade-off between performance criteria,

which will rank differently, in terms of importance, under varying circumstances.
For example, on a delayed flight, the crew might wish to trade in some fuel effi-
ciency to make up lost time.
More generally, performance criteria can be ranked differently by different op-
erators, or might depend on the flight type. An airline might, for instance, instruct
its passenger transport crews to avoid as much adverse weather as possible, to
maximize customer comfort but instruct its cargo crews to focus on fuel efficiency.
Additionally, between-pilot variance in replanning strategy preference is to be ex-
pected, even under seemingly similar circumstances.
 SUPPORTING 4D TRAJECTORY REVISIONS 39

Although nonuniqueness of the trajectory optimization cost function, in itself,

does not imply that an automated optimization approach is inappropriate, design-
ing an exhaustive set of such functions would be difficult. Whether a designer
would be able to anticipate all relevant situations (or at least their archetypes) and
predefine a corresponding cost function is uncertain. Even with robust autono-
mous planning software, a mismatch between the system’s optimal solution and
the subjective pilot perception of what is optimal can degrade pilot acceptance of
the system.
A pragmatic alternative to automated trajectory selection might thus lie in a
form of human–machine cooperation in which pilot expertise and decision-mak-
ing ability are supported by a relatively unassuming automated co-actor. Active in-
Downloaded by [DUT Library] at 07:36 04 October 2014

volvement of the pilot in the replanning task might additionally help reduce confu-
sion of intent between pilot and automation during the trajectory execution phase,
an issue that has, for example, been reported in relation to vertical navigation
(VNAV) functionality of the FMS in particular (Sherry et al., 2000).

COGNITIVE WORK ANALYSIS OF THE AIRBORNE

TRAJECTORY REVISION TASK

The design of an ecological interface starts with a cognitive work analysis. The
process of a cognitive work analysis consists of five steps (Vicente, 1999):

1. Work domain analysis: What are we working with? With what purpose?
2. Control task analysis: What must be done?
3. Social organization and cooperation analysis: Who can best perform each
(sub)task?
4. Strategies analysis: How can it be done?
5. Worker competencies analysis: How can human actors be supported in
their task?

Work Domain Analysis

System boundary. A detailed scenario of an aircraft requesting a trajectory
modification would involve numerous interacting systems. However, this research
focuses on supporting the crew during the replanning task, not on subsequent in-
teractions with, for instance, the air navigation service provider (ANSP). The sys-
tem considered for work domain analysis will therefore be confined to an aircraft
flying in airspace with obstructive elements.

Abstraction hierarchy. The abstraction hierarchy uses different levels, from

abstract to concrete, to describe the same systems in terms of means and ends. The
 40 VAN MARWIJK ET AL.

highest level of abstraction provides insight into the system’s overall goals. The
lower levels provide a more detailed representation of means and subgoals. Four
levels of abstraction were considered:

1. Functional purpose: What is the purpose of the work domain?

2. Abstract function: What are the underlying laws and principles?
3. Generalized function: What specific processes are involved?
4. Physical function: What tools are available to influence these processes?

The elements in the work domain most important to trajectory planning have been
highlighted in the abstraction hierarchy (Figure 1), along with the mental abstraction
Downloaded by [DUT Library] at 07:36 04 October 2014

jump,’ between goals and controls, which a pilot needs to make to modify a flight plan.

FIGURE 1 Abstraction hierarchy of the flight replanning problem. FMS = flight management
system; VNAV = vertical navigation; LNAV = longitudinal navigation.
 SUPPORTING 4D TRAJECTORY REVISIONS 41

Existing navigation and planning interfaces enable pilots to perceive the flight plan as
a sequence of altitudes, speeds, waypoints, and other navigation-related information
(generalized function level, Figure 1), but do not explicitly represent what effects a
modification of the flight plan would have on the higher level goals, such as efficiency
and punctuality (functional purpose level, Figure 1). By clarifying cross-abstrac-
tion-level interactions through automation support (indicated by the boxed elements
in Figure 1), the demands on pilot cognitive behavior can be lowered, as a “transla-
tion” of the high-level goals in terms of the low-level replanning controls, and vice
versa, would reveal the functional meaning of control possibilities.

Social Organization and Cooperation Analysis

Downloaded by [DUT Library] at 07:36 04 October 2014

As a starting point for interface design, the subtasks that make up the replanning
task must be allocated to the available actors. For the system considered, this is
equivalent to dividing tasks between the pilot and the automation. The relation-
ships between actions and subtasks of each actor in a system can be represented in
a so-called decision ladder (see Figure 2). Because the pilot is responsible for, and
capable of, deciding which route would satisfy the appropriate levels of safety,
productivity, and efficiency, the selection of a particular route from a broad range
of solutions is ideally left to him or her. However, defining an alternate flight plan,
in both space and time, within various constraints, is a complex task. The crew
cannot be expected to perform this task proficiently unless the interaction between
spatial and temporal constraints is easily observable. It is not inherently so.
As illustrated in Figure 2, the spatial trajectory planning task is allocated to the
pilot and temporal planning of the trajectory is left to the automation. Once the pi-
lot has defined the new spatial trajectory, the automation completes it by suggest-
ing fuel-efficient speed and altitude profiles that satisfy the 4D constraints at the
metering fix—a fix along a planned trajectory from over which an aircraft will be
metered prior to entering terminal airspace. The interactions between the decision
ladders in Figure 2 illustrate how the outcome of spatial planning affects temporal
planning, and vice versa.
To allow the pilot to make a well-informed decision on the spatial resolution
quickly, the effect of his or her actions on adherence to constraints can be pre-
viewed in a graphical representation of the spatial maneuvering space the current
situation affords. This so-called affordance zone is realized through automated
preevaluation of numerous spatial trajectories (Mulder et al., 2010). After defini-
tion of a new spatial trajectory by the pilot, the automation adds a corresponding
temporal plan, so as to ultimately obtain a complete 4D trajectory.

Control Task Analysis

Decision ladder. The control task is mapped as a sequence of subtasks
(Vicente, 1999). The four dimensions of the trajectory to be defined are interdepen-
 42 VAN MARWIJK ET AL.

Spatial Planning: Pilot Temporal Planning: Automation

Knowledge-based
domain Evaluate performance Evaluate
criteria

Ambi- Ultim. Ambi- Ultim.

guity goal guity goal

Interpret: choose Interpret: time required

resolution strategy vs RTA and tolerance

Rule-based
domain System System Goal
Goal
state state state state
Downloaded by [DUT Library] at 07:36 04 October 2014

Identify trajectory Determine satisfactory Identify attainable speed Update speed profile
acceptability spatial trajectory over spatial trajectory

Set of Set of
Task Task
obs. obs.

Observe obstruction Establish waypoint Observe spatial trajec- Determine thrust

relative to trajectory geometry tory and time available settings

Prode- Prode-
Alert Alert
dure dure

Activation Update 3D trajectory Activation Complete 4D trajectory

Skill-based
domain

FIGURE 2 Decision ladder of the 4D planning task, separated into spatial planning, which is allo-
cated to the pilot, and temporal planning, which is allocated to the automation.

dent. For example, once a spatial trajectory has been defined, the time constraint at
the metering fix, and the aircraft performance capabilities, limit the possibilities for
temporal planning considerably. A distinction of the task in terms of temporal and
spatial planning is therefore considered useful (see Figure 2). Decisions made by ei-
ther pilot or automation are shaped by internal and external constraints of the work
domain. Internal constraints are determined by the aircraft and all its subsystems.
External constraints are imposed by the environment of which the aircraft is part.

Internal and external constraints. Assuming that the aircraft has the re-
quired guidance and navigation capability to execute user-preferred routes in fu-
ture ATM operations, the remaining internal constraints are as follows:

• Flight envelope. For replanning, the most important constraints stemming

from the flight envelope are the limits on air speed and Mach number. Mod-
 SUPPORTING 4D TRAJECTORY REVISIONS 43

ified trajectories will have to be checked for adherence to these constraints

before they are negotiated with the ANSP. Because 4D trajectories with con-
trolled time of arrival are necessarily defined in terms of ground speeds
rather than air speeds, the combination of high ground speed and strong wind
might push the required airspeed toward the limits of the flight envelope.
• Aircraft dynamics. The aircraft dynamics, in particular accelerations, decel-
erations, and turn performance, should be simulated to ensure that the
planned trajectories can be physically executed by the aircraft. The dynamic
equations used in this study are discussed in the section on computation of
affordances later.
• Fuel available. The amount of fuel on board can limit the alternatives during
Downloaded by [DUT Library] at 07:36 04 October 2014

trajectory modification. For any particular candidate trajectory, the required

fuel consumption can be predicted and compared to the fuel available. There
must be an acceptable margin between fuel required and fuel available if the
candidate trajectory is to be considered feasible.

The external work domain constraints are as follows:

• Obstruction of the flight path. Obstruction can come in the form of traffic,
weather cells, airspace restrictions, and terrain. The latter two, due to their
predictability, are highly unlikely to cause obstruction of the planned trajec-
tory, but they can limit the set of alternatives during trajectory replanning.
Adverse weather is a likely cause of trajectory revision, due to its locality and
unpredictability. Interaction with other air traffic is beyond the scope of this
study.
• Operational regulations. Aviation regulations could constrain the possibili-
ties for replanning. Some examples are:
• Maximum speed depending on flight level (e.g., maximum 250 kt below
flight level 100).
• Maximum bank angle or rate of turn (e.g., maximum 30° bank).
• Maximum rate of descent (e.g., maximum 4,000 ft/min).
• Arrival requirements at a metering fix. Metering typically occurs near the
destination; the fix can designate the transition from user-preferred routing
airspace into airspace managed by air traffic control. Concerning the exact
location of the fix, the SESAR concept definition states, “Ideally, the con-
trolled times of arrival would be set at the runway threshold (to focus on the
optimization of runway throughput) but in reality a merging point further out
is more likely to be practicable” (SESAR Consortium, 2008). Consequently,
merging points further out, such as the initial approach fix, are regarded as
likely candidates in this study. The exact location of the merging point has,
however, only minor consequences (if any) on the replanning task in user-
preferred airspace, for which an interface is proposed in this article.
 44 VAN MARWIJK ET AL.

Strategies Analysis
To complement the engineering perspective on system development, three experi-
enced professional pilots (see Table 1), with backgrounds in civil aviation and re-
search, were consulted for input on the design. After an introductory discussion of
future ATM concepts (NextGen and SESAR) and their implications on in-flight
trajectory modification, the participants were questioned on their preferences with
regard to in-flight trajectory replanning and interface content of a new FMS plan-
ning interface design. Three preferred trajectory alternatives in case of obstruc-
tions were expressed:

1. To plan descents at constant throttle setting.

Downloaded by [DUT Library] at 07:36 04 October 2014

2. To separate speed changes from flight level changes.

3. To minimize fuel consumption resulting from the revision.

Concerning the interaction with the automation, a preference for decoupled

planning of the ground track and vertical profile was expressed. All pilots were in
favor of direct manipulation of the flight plan geometry through a cursor control
device. When asked to name display content that would be useful in performing a
trajectory revision, the following answers were given:

• A visualization of time constraints at the metering fix.

• A preview of throttle settings and speeds per planned trajectory segment.
• A preview of the maximum rate of descent.
• An estimate of the fuel consumption corresponding to the modified trajec-
tory.
• An outline of the original trajectory during the editing process.
The interface design is intended to reflect these pilot preferences as much as
possible.

Worker Competencies Analysis

The purpose of the worker competencies analysis is to identify the level of cogni-
tive behavior required to perform the tasks allocated to the human. Using the skills,

TABLE 1
Age, Gender, and Experience of Interview Participants

Pilot Gender Age Flight Hours Aircraft Type

A Male 32 2,000 Cessna Citation II, Piper PA-31

B Male 64 9,000 Boeing 737–200/300, 757–200, 767–300 ER
C Male 69 14,000 Boeing 747–300/400
 SUPPORTING 4D TRAJECTORY REVISIONS 45

rules, knowledge taxonomy (Rasmussen, 1999) as a qualitative framework for

assessment, the hypothetical benefits of the proposed planning interface can be
explained.

Skill-based behavior. Establishment of the waypoint geometry is achieved

through direct manipulation, which is best categorized as skill-based behavior,
combined with a basic understanding of the possible manipulations and symbols
(some rule-based behavior). Additionally, the navigation display allows immedi-
ate perception of the planned spatial trajectory relative to constraints such as
weather, terrain, and so on.
Downloaded by [DUT Library] at 07:36 04 October 2014

Rule-based behavior. Choosing and implementing a satisfactory resolu-

tion strategy is supported by signs on the display. The primary function of the
affordance overlays in both the horizontal and vertical situation displays is to pro-
vide additional support for rule-based behavior during replanning and, as a result,
prevent unnecessary knowledge-based demands.

Knowledge-based behavior. If the complexity of a situation calls for knowl-

edge- based behavior, the automated representation of constraints supports the pi-
lot in his or her tasks of interpretation and decision making. Of course, the purpose
of the display design is not to prevent the pilot from applying knowledge-based be-
havior when it is useful, such as in a trade-off between several performance crite-
ria, but to lower the demands where possible.

INTERFACE DESIGN

The display design is a combination of conventional and novel display elements.

There are several reasons for building on conventional displays, rather than de-
signing from scratch. First of all, the existing representations used for aircraft navi-
gation, the horizontal (HSD) and vertical situation displays (VSD), have already
proven their value both in experiments and practice (Prevôt & Palmer, 2000). Sec-
ond, by designing a replanning display that looks similar to the navigation display,
pilots are not required to adjust to a different representation to make modifications.
This is considered useful, because, compared to navigation, airborne trajectory
replanning is an infrequent, intermittent, and brief task. Additionally, pilots are
likely to engage in edit mode only after they have identified a problem with, and
possibly even an intended modification of, the active trajectory in navigation
mode. A practical advantage of extending conventional display functionality is
that it might facilitate implementation of the proposed replanning interface in fu-
ture systems.
 46 VAN MARWIJK ET AL.

Relation to Previous Work

As stated earlier, this study can be considered a continuation of previous design
work (Kaber et al., 2002; Mulder et al., 2010). As such, it makes sense to briefly
and explicitly discuss the added value of the new design.

Visualization of affordances. Figure 3 illustrates how waypoint locations,

which result in equal ground track length, form an ellipse when they are connected.
A necessary assumption underlying this mapping is that the ground track consists
of straight lines only.
Downloaded by [DUT Library] at 07:36 04 October 2014

Correcting the affordance for aircraft urn radius. To improve the valid-
ity of the mapping of time constraints onto alternate waypoint locations, this study
incorporates the effect of aircraft turn radius and waypoint fly-by in the computa-
tion of the affordance overlay. By simplifying actual turn performance to turns
with constant radius, an approximation of waypoint fly-by effect on the shape of
the waypoint affordance is obtained. The resulting shape, shown in Figure 4a, ap-
proximates an ellipse only for waypoint geometries in which the heading change
per waypoint is relatively small at all three waypoints. As the heading change at

(a) Moderate heading change (b) Large heading change

FIGURE 3 When neglecting turn radius, the locations of the intermediate waypoint that would result
in equal ground track length make up an ellipse. The simplified ground track in the left and right figures
has the same length.
 SUPPORTING 4D TRAJECTORY REVISIONS 47
Downloaded by [DUT Library] at 07:36 04 October 2014

FIGURE 4 Visualization of fly-by effect, exaggerated by large turn radius and small distance
between waypoints.

the intermediate waypoint is increased (see Figure 4b), the shape of the affordance
differs substantially from an ellipse. If the waypoint geometry is made such that a
large heading change at a particular waypoint is required, the location of this
waypoint might hardly be indicative of the trajectory that the aircraft will follow.
Because, during trajectory replanning, it is the shift in trajectory, rather than the
shift in waypoint location, that is of interest to the pilot, the relatively useless part
of the waypoint affordance, where the heading change at this waypoint is larger
than, for example, 70°, is omitted from the affordance. Although this limits the set
of possibilities for manipulation presented to the pilot, the practicality of the dis-
play overlay is expected to increase from such a limit on heading change per
waypoint. Figure 4c shows the result of discarding solutions that require a very
large heading change at a single waypoint.

Supporting direct manipulation in the VSD. To complement the improve-

ments to overlays in the HSD, this article introduces direct manipulation of the
descent profile in the VSD in combination with a representation of vertical
affordances. Because the time constraints of most importance for replanning are
likely to be those at the metering fix, the part of the trajectory to be replanned will
typically involve a descent. Understandably, exploratory studies on direct manipu-
 48 VAN MARWIJK ET AL.

lation interfaces for trajectory replanning used paths of constant glideslope to con-
nect waypoints in the VSD. Although such a constant glideslope has its practical
advantages for runway approach, there is little reason to perform an initial descent
(e.g., from cruise altitude toward a merging point) in this way. To minimize fuel
consumption, for example, a postponed descent at idle thrust and favorable Mach
number would be more appropriate. Given the importance of the vertical profile in
trajectory replanning, the added support for direct manipulation in the VSD is hy-
pothesized to be a valuable addition.

Increased complexity of the temporal plan. In previous work by Mulder

et al. (2010), adaptation of the speed profile to compensate for ground track length
changes was limited to an along-track shift of a single deceleration phase. In the
Downloaded by [DUT Library] at 07:36 04 October 2014

current design, temporal flexibility is considerably increased by using an interme-

diate segment flown at suitable target ground speed, complemented with speed
changes before and after it. This is discussed in more detail in the section
“Working with the Cues.”

Novel Representation
The proposed interface, shown in Figure 5, inherits all display elements of a conven-
tional navigation display. The most obvious additions, both in HSD and VSD, are
the affordance zones that appear when a waypoint or a top or bottom of descent is se-
lected. As the name suggests, these zones visualize the control actions that the pilot
can afford to take in terms of waypoint relocation without violating replanning con-
straints. It is important to note that the affordance zone does not provide any infor-
mation as to whether obstructions would be avoided by a particular modification of
the trajectory. Air traffic avoidance is beyond the scope of the research reported
here, and en-route obstruction by adverse weather, restricted airspace, or terrain is
assumed to be visible on the display. The decision of which weather conditions are
acceptable to fly through is deliberately left to the pilot. All novel display elements
can be hidden by pilot request (pressing a toggle button) to reduce display clutter.

Horizontal situation display. In the HSD, control action is equivalent to

modification of the location of the selected waypoint. The constraints that bound
the affordance zone are the time available in which to reach waypoint FIX and the
achievable ground speed range. Modification of the trajectory by means of way-
point relocation will generally result in a different distance-to-fly to waypoint FIX.
The estimated time of arrival (ETA) of the new trajectory can be made close or
equal to the required time of arrival (RTA) of the contracted trajectory by (auto-
matic or manual) adjustment of the speed profile to the (new) trajectory length. As-
suming a tolerance window around the RTA, candidate trajectories can be classi-
fied in three types, according to the smallest possible difference between ETA and
RTA at waypoint FIX:
 SUPPORTING 4D TRAJECTORY REVISIONS 49
Downloaded by [DUT Library] at 07:36 04 October 2014

FIGURE 5 Display design of the replanning interface, showing the waypoint affordance
zone for estimated time of arrival (ETA) equal to required time of arrival (RTA) (1) and ETA
within tolerances (2); labels regarding the speed (3) and altitude profile (4); the affordance zone
of the top of descent (5) and an outline of the descent envelope (6); the maximum rate of descent
(7); and a ground speed and throttle preview (8).

1. The ETA equals the RTA. Candidate locations that would result in this
type of a trajectory are represented by the light shade of the affordance
zone (1).
2. The ETA lies within tolerances, but the RTA itself cannot be achieved.
Waypoint locations corresponding to such trajectories are represented by
the dark shade of the affordance zone (2).
3. The ETA is outside RTA tolerances. Corresponding waypoint locations
are not part of the affordance zone, as the resulting trajectory would require
renegotiation.
 50 VAN MARWIJK ET AL.

The RTA tolerance is assumed to be defined in trajectory management require-

ments, and will most likely be selected by the ANSP to fit the traffic density at the
destination airport. In addition to the affordance zone, the HSD includes a repre-
sentation of the speed profile selected by the automation (3), and the location of the
top and bottom of descent (4).

Vertical situation display. Because manipulation of waypoints in the HSD

only results in a definition of the ground track, the VSD is used to facilitate modifi-
cation of the vertical profile. After manipulations to the ground track, the automa-
tion will present the corresponding energetically optimal vertical profile on the
VSD by default. Analogous to replacement of waypoints in the HSD, the vertical
Downloaded by [DUT Library] at 07:36 04 October 2014

profile can be modified by manipulation of its nodes, which are the top and bottom
of descent. The affordance zone in the VSD consists of a horizontal band (5),
which appears when either the top or bottom of descent is selected, and highlights
the alternative locations for it that would result in a feasible vertical profile.
Placing the top or bottom of descent within this band ensures that the resulting tra-
jectory is not too steep to allow for descent with constant ground speed. The sec-
ond element of the vertical affordance zone, is an outline of the descent envelope
(6), which is bounded by the steepest descent from the forward most and rearmost
top of descent location, and the initial and final altitudes. To assist the pilot in eval-
uating the vertical profile, a numeric representation of the maximum vertical speed
is included (7). Finally, because it is important for pilots to form an accurate men-
tal picture of the 4D flight plan during evaluation and replanning, the trajectory
edit display reveals the time dimension of the trajectory through ground speed tar-
gets, along with predicted throttle settings (8). Showing this information is hypoth-
esized to increase situation awareness and reduce the risk of the crew being sur-
prised by, for example, automatically executed speed changes.

COMPUTATION OF AFFORDANCES

Simplified Aircraft Dynamics

For the purpose of trajectory replanning, rotational aircraft dynamics are of little
relevance. Therefore, a point mass model and the assumption of steady non-
side-slipping banked turns are used to generate trajectories.

Boundary Conditions
The shift toward 4D navigation in future ATM implies that every trajectory must
contain at least one metering fix with a corresponding RTA. The affordance zones
provide support for modification of the part of the trajectory that lies in user-pre-
 SUPPORTING 4D TRAJECTORY REVISIONS 51

ferred airspace between the current aircraft position of the aircraft, and the fix.
However, to allow the crew some time in which to make a modification and have it
negotiated, the predicted aircraft position after a margin of 3 min, rather than the
current position, is taken as the starting point for trajectory modification. This
value of 3 min is based both on findings in literature (Battiste et al., 2000) and on
an estimate of the time required by a pilot for flight plan revision.
Once the replanning mode is started by the flight crew, the initial and final con-
ditions of the trajectory to be revised are known. The initial condition corresponds
to the aircraft state projected 3 min forward along the original trajectory and the
final condition corresponds to the required aircraft state at the metering fix.
Downloaded by [DUT Library] at 07:36 04 October 2014

Ground Speed and Altitude Profiles

Because manipulation of the spatial trajectory in the HSD under invariant meter-
ing constraints calls for a corresponding redefinition of the ground speed profile,
alternative trajectories will generally introduce additional speed changes. To re-
duce trajectory complexity, as was suggested during pilot input on design, these
speed changes are not performed simultaneously with altitude changes. Instead a
typical trajectory consists of three consecutive segments in terms of groundspeed:

1. Initial speed change to target ground speed.

2. Flight at constant target groundspeed.
3. Final speed change to meet constraints at the metering fix.

Lookup Table for Waypoint Affordances

On initialization of the replanning mode, the boundary conditions of the flight plan
are used to generate a set of feasible 3D trajectories that will allow rapid evaluation
of affordances at waypoint selection. The three dimensions in which lookup table
trajectories for waypoint affordances are defined are: along-ground-track distance,
altitude, and time. The reason for generating such a lookup table is that it allows a
nearly infinite number of waypoint geometries to be linked to a finite set of
preevaluated trajectories. As illustrated in Figure 6, the lookup table is filled by it-
eration through ground track lengths and intermediate target speeds by checking
ETA at the fix against the time constraint (i.e., the RTA and tolerances). For the
purpose of generating the lookup table, the descent is assumed to be performed at
idle thrust. Additional possibilities for vertical planning will be evaluated during
generation of a lookup table for descent affordances. Each available descent in the
lookup table is called a descent node.
The second segment, flown at target ground speed, corresponds to the maxi-
mum length available for the descent phase, s2max. The minimum length of a de-
 52 VAN MARWIJK ET AL.
Downloaded by [DUT Library] at 07:36 04 October 2014

FIGURE 6 Schematic overview of affordance zone computation for waypoints.

scent, s2min, corresponds to the steepest descent. The possibilities for manipulating
the top and bottom of descent increase as the difference between s2max and s2min
increases.
The altitude change (if any) is performed somewhere within the segment flown
at constant target ground speed. Apart from the fixed maximum length and ground
speed, the descent profile can be freely manipulated by the pilot within the descent
envelope. The latter is bounded by the steepest (i.e., idle thrust) descent from the
forward most and rearmost top of descent location. For the computation of descent
paths, constant throttle setting and ground speed are used to iteratively define the
required pitch angle. This results in a slight curve. As a result of the constant
ground speed, air temperature gradient, and decreasing pitch angle, airspeed in-
creases slightly and Mach number decreases slightly during a descent performed
in this way.

Waypoint Affordance Zone

Because lookup table generation is performed instantaneously when the replan-
ning mode is initialized, evaluation of the affordance zone, which is performed on
waypoint selection through a cursor click, is reduced to comparing the ground
 SUPPORTING 4D TRAJECTORY REVISIONS 53

track length corresponding to alternate waypoint locations to the ground track

lengths in the lookup table. In contrast to the aforementioned lookup table genera-
tion, this does not require numeric integration of differential equations and is thus
not as computationally intensive. During affordance computation, only alternative
locations for the selected waypoint are evaluated; the current locations of all other
waypoints are kept constant. If the waypoint geometry results in a feasible ground
track length, the corresponding trajectory can be obtained from the lookup table.
The result is a 4D trajectory, as the 3D distance–altitude–time trajectory from the
lookup table is now instantiated on a path in the horizontal plane. Assuming the
availability of a predictive weather model, the effects of wind can be taken into ac-
count in the prediction of required airspeed and Mach number, as well as fuel con-
Downloaded by [DUT Library] at 07:36 04 October 2014

sumption. If these are all within acceptable ranges and, additionally, the trajectory
does not violate operational regulations, the waypoint location under consider-
ation can be considered part of the waypoint affordance, as illustrated in Figure 6.

Lookup Table for Descent Node Affordances

As stated earlier, definition of a ground track does not necessarily imply a unique
altitude profile. On initialization of the replanning mode, as well as after each re-
definition of the ground track, a lookup table of descents is generated. This table
contains the constant-throttle descents, at required ground speed, which can be
performed within the length available for descent. For increasing throttle settings,
the resulting descent at constant ground speed will be longer and more gradual.
The descents in the lookup table are obtained by iteration through throttle settings,
while checking adherence to the distance constraint. Figure 7 shows the generation
of the lookup table and its relation to the descent node affordance zone.

Descent Node Affordance Zone

On selection of a descent node, alternative along-track locations for it are evalu-
ated in terms of descent length resulting from these locations. As was the case for
waypoint affordances, a predictive weather model can be used to account for wind
effects in checking for adherence to constraints stemming from the flight enve-
lope, fuel available, and operational regulations.

WORKING WITH THE CUES

Defining the Ground Track

To manipulate the ground track, the pilot can drag a waypoint to a different loca-
tion, as illustrated by Figure 8. The affordance zone of the selected waypoint will
 54 VAN MARWIJK ET AL.
Downloaded by [DUT Library] at 07:36 04 October 2014

FIGURE 7 Schematic overview of affordance zone computation for descent nodes.

highlight alternative locations that allow adherence to constraints. A flight plan re-
vision can involve replacement of several waypoints, in which case it might be
logical for the pilot to place one or more waypoints outside their affordance zone
during intermediate steps. Placing a waypoint outside its affordance zone is by no
means a wrong action. During such a replacement, the affordance zone communi-
cates that the trajectory would violate constraints if left as is. Subsequent selection
of a different waypoint will nevertheless highlight an affordance for it, provided
that there are alternate locations that would restore the trajectory’s adherence to
replanning constraints. After one waypoint is placed outside its affordance all
other waypoints will, necessarily, be outside their respective affordances as well.
As long as the last manipulation in the sequence constitutes the placement of a
waypoint (which one is irrelevant) into its affordance, metering constraints will be
met. The affordance overlay supports the pilot only in evaluating the currently de-
fined trajectory without hindering revision in multiple steps.
During manipulation (i.e., dragging), the length of the ground track is translated
into corresponding ground speed and required throttle settings, which the pilot can
read from the display. Because the affordance zone of a waypoint is computed us-
ing the current location of all other waypoints, sequentially replacing two way-
points will result in differently shaped affordance zones.
 SUPPORTING 4D TRAJECTORY REVISIONS 55
Downloaded by [DUT Library] at 07:36 04 October 2014

Original trajectory, obstructed Upon selection of UD1, its UD1 is replaced within its
by weather cell. affordances are highlighted. affordance zone.

Upon selection of UD2, its UD2 is replaced within its Revised trajectory to be
affordances are highlighted. affordance zone. negotiated with the ANSP.

FIGURE 8 Manipulations to the trajectory in the horizontal situation display, using the
waypoint affordance overlays to ascertain that metering constraints can be satisfied through au-
tomated adjustment of the speed profile. ANSP = air navigation service provider.

Defining the Vertical Profile

Manipulation of the vertical profile is achieved by along-track repositioning of de-
scent nodes. Figure 9 illustrates how the top of descent can be dragged to a new lo-
cation within its affordance zone to obtain a more gradual descent than the energet-
ically optimal idle descent, which is selected by default after manipulations to the
ground track. Because the affordances of the selected node are computed based on
the current position of other nodes, the affordance zone of the bottom of descent is
constrained by the new position of the top of descent.
 56 VAN MARWIJK ET AL.
Downloaded by [DUT Library] at 07:36 04 October 2014

FIGURE 9 Manipulations to the vertical profile by direct manipulation of descent nodes,

using the affordance overlays to preview suitable locations of the selected node.

Similar to the HSD, the changing vertical speed and throttle settings re-
sulting from the modified trajectory are visible while the node is being
dragged.
 SUPPORTING 4D TRAJECTORY REVISIONS 57

Negotiation
As long as the FMS replanning mode is active, the pilot can freely define and re-
vise an alternate trajectory. The assumption made in this study is that once the pilot
is satisfied with the revision, he or she can submit the trajectory to the ANSP
through digital data link. The ANSP is assumed, after (automated) evaluation, to
either accept it or propose an alternative. On acceptance, the modified trajectory
becomes the new reference for both the ANSP and the airspace user(s).
In the event that modifications made by the pilot take such a substantial amount
of time that the aircraft closely approaches, or passes, a waypoint that marks the
start of the modified trajectory (labeled EDIT in Figures 5 and 8), negotiation of
the new trajectory will no longer be possible. The pilot can then use the automation
Downloaded by [DUT Library] at 07:36 04 October 2014

to define a new waypoint EDIT, projected 3 min from the present aircraft position,
before restarting the revision process. Tasks such as redefinition of the transition
waypoint EDIT, submission of the revised flight plan, and cancellation of the
replanning process are facilitated through corresponding text buttons. Depending
on screen dimensions, these are displayed either directly below or adjacent to the
navigation display.

COGNITIVE DEMANDS

The proposed interface can be related to the goals of interface design, which were
stated with regard to the skills, rules, knowledge taxonomy by Vicente and Ras-
mussen (1990) as “[d]esign interfaces that do not force cognitive control to a
higher level than the demands of the task require, but that also provide the appro-
priate support for all three levels.”

Lowering the Required Level of Cognitive Behavior

As discussed earlier, the interface is designed to allow pilots to access information
that is relevant at all levels of cognitive behavior—either skills-, rule-, or knowl-
edge-based behavior. Given the allocation of the temporal planning task to auto-
mation, cognitive support has mainly been provided for the spatial planning task.
Applying the skills, rules, knowledge taxonomy to this task, each subtask allocated
to the pilot in Figure 2 is categorized in Table 2 according to the level of cognitive
behavior it primarily requires (admittedly, a subtask often involves more than one
level).
The arrows in Table 2 represent changes in weight of the required levels of cog-
nition per subtask, which the interface is hypothesized to induce. That is, the tasks
of evaluation and interpretation, which required mostly knowledge-based behav-
ior without automation support, are partly reduced to rule-based behavior by the
 58 VAN MARWIJK ET AL.

TABLE 2
Effect of Supporting Cognitive Behavior Through Interface Design

Subtask Skills Rules Knowledge

Observe obstructions relative to trajectory ×

Identify trajectory acceptability ×
Evaluate performance criteria × <<
Interpret: Choose resolution strategy × <<
Determine satisfactory spatial trajectory ×
Establish waypoint geometry × <<

Note. × = level of cognitive behavior primarily induced by subtask; << = hypothesized reduction
in cognitive demands.
Downloaded by [DUT Library] at 07:36 04 October 2014

proposed interface, as constraints are expressed in terms of waypoint location and

visualized in the affordance zone as a preview of resulting performance. The task
demands are reduced through support of pilot perception of cross-abstraction-
level interactions (recall the abstraction hierarchy space in Figure 1). This allows
the pilot to determine a satisfactory spatial trajectory by working according to a
rule, such as the following: Replace a node within its affordance zone to obtain an
alternative trajectory that satisfies replanning constraints. The pilot can envision
the effect his or her control action will have on high-order goals.
Reduction of the rule-based behavioral component involved in establishing a
new node geometry is a hypothesized effect of using direct manipulation, rather
than keypad input, for editing.

DISCUSSION

Multiple-to-One Constraint-to-Sign Mapping

The fact that the affordance zones represent information from a high abstraction
level (timeliness, productivity, safety) on a lower abstraction level (node location)
leads to an arguable design weakness. Several lower level constraints have been
lumped together in the affordance overlay. The mapping of lower level work do-
main constraints on the cues, or signs, provided by the interface, is thus not consis-
tently one-to-one. This is contrary to design principles laid out by Vicente (1999).
The result is that although the pilot can observe the space of effective control, he or
she might not be able to tell which constraints disqualify a particular waypoint
location outside the affordance zone. The most obvious resolution of this issue
would be to represent each relevant constraint by its own display element (i.e.,
sign). However, anticipating unacceptable display clutter as a result, such informa-
tion content was not added. By focusing on the high-level purpose, overview is
gained at the expense of completeness.
 SUPPORTING 4D TRAJECTORY REVISIONS 59

Limited Flexibility in Supported Speed and Altitude Profiles

The proposed interface simplifies the well-informed manipulation of the vertical
profile by means of along-track replacements of top and bottom of descent, allow-
ing the pilot to select the spatial path of the descent and thus control the vertical
speed resulting from it. The possibilities for manipulation of the vertical profile
are, however, limited, as vertical replacement of nodes is not supported. Addi-
tionally, the only type of descent that is supported is descent at constant ground-
speed and constant throttle setting. In the generated trajectories, altitude changes
and speed changes are performed sequentially, rather than simultaneously. The
latter requirement reduces the complexity of the trajectories, but admittedly limits
the possibilities for replanning.
Downloaded by [DUT Library] at 07:36 04 October 2014

Representation of Fuel Consumption

The affordance zone helps the pilot rule out trajectories that would require too
much fuel, but does not differentiate between feasible trajectories with different
fuel consumption characteristics. Ideally, the pilot would be able to foresee the ef-
fect of his or her actions on fuel efficiency before making manipulations, so as to
“get it right” during his or her first attempt.

Experimental Evaluation
In preparation of an experiment to evaluate the proposed planning interface, a soft-
ware prototype has been developed. The interface can be evaluated in the follow-
ing areas of interest:

1. Objective replanning performance, for example, avoidance of obstruc-

tions, deviation from the RTA, and time required for editing.
2. Situation awareness (or trajectory awareness), measured explicitly with a
questionnaire such as Endsley’s Situation Global Awareness Global As-
sessment Technique (SAGAT), to assess awareness on the levels of per-
ception, comprehension, projection, and metacognition (Endsley & Gar-
land, 2000).
3. Subjective measures, such as interface usability, satisfaction with re-
planned path and display clutter, measured by means of a questionnaire.
4. Workload, measured with NASA’s Task Load Index method (Hart &
Staveland, 1988).

The recommended first step would be an experiment that focuses on cognitive util-
ity, with the aim of revealing possibilities for further improvement of the interface
 60 VAN MARWIJK ET AL.

design. This initial evaluation might, at a later time, be complemented by an exper-

iment that compares the interface to a design alternative.

CONCLUSIONS

We have described development of a new style of ecological interface for in-flight

trajectory revision to accommodate the anticipated changes in ATM operation in
the not-too-distant future. In accordance with the EID framework, the possibilities
for effective control, rather than directions, are presented to the pilot. Direct ma-
nipulation of the flight plan, combined with a preview of the effects of control ac-
Downloaded by [DUT Library] at 07:36 04 October 2014

tion on performance, reduces the required level of cognitive demand. The system
performs temporal planning automatically in response to spatial manipulation by
the pilot. The design additionally includes a speed and throttle setting representa-
tion, which is aimed at increasing situation awareness regarding the execution of
the planned trajectory.

REFERENCES

BASI. (1999). Advanced-technology aircraft safety survey report. In Flight safety digest: Special issue
(pp. 137–216). Alexandria, VA: Flight Safety Foundation.
Battiste, V., Johnson, W. W., & Bochow, S. H. (2000, October). Enabling strategic flight deck route
re-planning within a modified ATC environment: The display of 4-D intent information on a CSD.
Paper presented at the 2000 World Aviation Conference, San Diego, CA.
Borst, C., Sjer, F. A., Mulder, M., van Paassen, M. M., & Mulder, J. A. (2008). Ecological approach to
support pilot terrain awareness after total engine failure. Journal of Aircraft, 45(1), 159–171.
Borst, C., Suijkerbuijk, H. C. H., Mulder, M., & van Paassen, M. M. (2006). Ecological interface design
for terrain awareness. The International Journal of Aviation Psychology, 16(4), 375–400.
Endsley, M. R., & Garland, D. J. (2000). Situation awareness analysis and measurement. Mahwah, NJ:
Lawrence Erlbaum Associates, Inc.
Harris, D. (2004). Human factors for civil flight deck design. Burlington, VT: Ashgate.
Hart, S. G., & Staveland, L. E. (1988). Development of NASA–TLX (Task Load Index): Results of em-
pirical and theoretical research. In P. A. Hancock & N. Meshkati (Eds.), Human mental workload
(pp. 139–183). Amsterdam, The Netherlands: Elsevier Science.
Huisman, H., Verhoeven, R. P. M., Houten, Y. A. V., & Flohr, E. L. (1997). Crew interfaces for future
ATM. In Proceedings of the AIAA IEEE Digital Avionics Systems Conference (Vol. 2, pp. 6.3-33–
6.3-40). Irvine, CAAIAA/IEEE.
Joint Planning and Development Office. (2007). Concept of operations for the next generation air
transportation system (NextGen) (Tech. Rep.). Washington, DC: Author.
Kaber, D. B., Riley, J. M., & Tan, K. W. (2002). Improved usability of aviation automation through di-
rect manipulation and graphical user interface design. The International Journal of Aviation Psy-
chology, 12(2), 153–178.
Lidén, S. (1994). The evolution of flight management systems. In Proceedings of the AIAA IEEE Digi-
tal Avionics Systems Conference (pp. 157–169). Phoenix, AZ.
 SUPPORTING 4D TRAJECTORY REVISIONS 61

Mulder, M., Winterberg, R., van Paassen, M. M., & Mulder, M. (2010). Direct manipulation interfaces
for in-flight four-dimensional navigation planning. The International Journal of Aviation Psychol-
ogy, 20(3), 249–268.
Prevôt, T., & Palmer, E. (2000, October). Staying ahead of the automation: A vertical situation display
can help. Paper presented at the 2000 World Aviation Conference, San Diego, CA.
Rasmussen, J. (1999). Ecological interface design for reliable human–machine Systems. The Interna-
tional Journal of Aviation Psychology, 9(3), 203–223.
SESAR Consortium. (2007). D3: Definition of the future ATM target concept (Tech. Rep.). Brussels,
Belgium: Author.
SESAR Consortium. (2008). D4: The ATM deployment sequence (Tech. Rep.). Brussels, Belgium: Au-
thor.
Sherry, L., Feary, M., Polson, P., Mumaw, R., & Palmer, E. (2000). A cognitive engineering analysis of
the vertical navigation (VNAV) function (Tech. Rep.). RAND/Honeywell Intl Inc., SJSU/NASA-
ARC, Department of Psychology University of Colorado, Boeing Commercial Airplane Group.
Downloaded by [DUT Library] at 07:36 04 October 2014

Swenson, H., Barhydt, R., & Landis, M. (2006). Next Generation Air Transportation System (NGATS)
air traffic management (ATM) airspace project (Tech. Rep.). Washington, DC: National Aeronau-
tics and Space Administration.
van Dam, S. B. J., Abeloos, A. L. M., Mulder, M., & van Paassen, M. M. (2004). Functional presenta-
tion of travel opportunities in flexible use airspace: An EID of an airborne conflict support tool. In
Proceedings of the IEEE International Conference on Systems, Man and Cybernetics (pp. 802–808).
Den Haag, The Netherlands: IEEE.
van Dam, S. B. J., Mulder, M., & van Paassen, M. M. (2008). Ecological interface design of a tactical
airborne separation assistance tool. IEEE Transactions on Systems Man and Cybernetics, Part A:
Systems and Humans, 38(6), 1221–1233.
van Marwijk, B. J. A., Mulder, M., Mulder, M., van Paassen, M. M., & Borst, C. (2009). A human-
machine interface for replanning of 4D trajectories. In Proceedings of the 15th International Sympo-
sium of Aviation Psychology (pp. 1–6). Dayton, OH: Wright State University.
van Paassen, M. M., Mulder, M., & van Dam, S. B. J. (2005). “Meaningful physics” or finding a system
description suitable for ecological interface design. In Proceedings of the 13th International Sympo-
sium of Aviation Psychology (pp. 772–776). Oklahoma City, OK.
Vicente, K. J. (1999). Cognitive works analysis: Towards safe, productive, and healthy computer-
based work. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Vicente, K. J., & Rasmussen, J. (1990). The ecology of human–machine systems II: Mediating “direct
perception” in complex work domains. Ecological Psychology, 2(10), 207–249.

Manuscript first received: September 2010