- Email: [email protected]
Integrated Design of Spatial Navigation Displays
Copyright © IFAC Man-Machine Systems, Kyoto, Japan, 1998
INTEGRA TED DESIGN OF SPATIAL NAVIGA TION DISPLAYS
Eric Theunissen*, Henk Stassen* and Durk van Willigen*
*Faculty of Information Technology and Systems, Delft University of Technology, Mekelweg 4, 2628 CD Delft, The Netherlands
tFaculty of Design, Engineering and Production, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands
Abstract: Compared to current command displays, a spatially integrated presentation of the future trajectory has the potential to reduce the task demanding load for guidance and control while at the same time increasing spatial and navigational awareness. To allow a structured design of spatial navigation displays, this paper illustrates how the information contents of the visual cues can be described as properties of the optic flow field. The type of visual cues are related to the potential control strategies, and the results of experiments in which these control strategies are evaluated are discussed. Copyright © 19981FAC
Keywords: displays, navigation, human-machine interface, aircraft control, design.
introduction of more complex approach trajectories is likely to reduce safety. By providing pilots with the information they need to accomplish their task more easily, while maintaining spatial and navigational awareness without the need to scan additional displays, it becomes possible to fly more complex approaches without a reduction in safety. It is highly unlikely that with all future developments, safety can be increased by extrapolating current concepts. New functionality and new technology cannot simply be layered onto existing concepts, because the current system complexities are already too high. Better humanmachine interfaces may require a fundamentally new approach. Navigation, guidance, and control are not three independent tasks, but with the current flight director display the control task is isolated. Displays providing a spatially integrated presentation of the future trajectory, and thus presenting guidance requirements instead of control commands have advantages relative to current non-spatial displays. These advantages result from the fact that the pilot has to perform less integration of information and the fact that the more natural presentation requires less effort for interpretation and evaluation. Fig.l shows a spatial navigation display during a flight test.
1. INTRODUCTION Due to the increasing number of aircraft passengers and the resulting increase in aircraft, bottlenecks in airspace capacity are beginning to emerge. Once the number of aircraft has reached a certain threshold, any further increase will cause unacceptable delays. Since the bottlenecks mainly occur in the vicinity of airports, this is also the place where a solution should start. The basic idea behind improvements is to abandon today's straight-in approaches and allow aircraft to intercept the final straight glidepath at predetermined locations. This provides air traffic control with more possibilities to manage the traffic flow, creating the opportunity to increase capacity. An additional advantage is that it becomes possible to avoid noise sensitive areas by using noise abatement procedures. This concept, however, increases the task demandinK load of the pilot. In contrast to the current straight-in approach, pilots will have to fly curved approaches. Due to the more frequent changes in direction, it becomes harder to maintain an adequate level of spatial and navigational awareness. As a result, pilots will have to scan the navigation display more frequently. Because of the already high workload during the approach, an
299
F,.",.oIR.'.,..,-,c. \1.Mng
"."ne.
F,./r:I-cI·vi.1III "~oIconstta.inU
20 qmboIogy ruI.s
Fig. 1. Flight test of an experimental spatial navigation display. Fig. 2. Overview of the elements involved in the spatially integrated presentation of navigation data.
2. AN INTEGRATED DESIGN APPROACH A factor which has prevented the introduction of spatial displays for navigation and guidance is the complexity of their design. Fadden et al. (1987) indicated this problem by stating that 'while the promise of spatial displays is great, the cost of their development will be correspondingly large. The knowledge and skills which must be coordinated to ensure successful results is unprecedented. From the viewpoint of the designer, basic knowledge of how human beings perceive and process complex displays appears Jragmented and largely unquantified'. Hardly any detailed guidelines to the design of these types of displays exist which take specific human capabilities in the areas of perception, cognition, and control into account.
A proper selection of the design parameters requires an understanding of their relation with the magnitude of the visual cues conveyed by the display and the potential task strategies. Furthermore, tasks and measures are needed to evaluate the potential of a certain perspective flightpath display format for a range of control strategies. An approach is needed which supports a structured design process for perspective flightpath displays in which technical possibilities and human factors are truly integrated. To develop such an approach, the question on how to utilize the existing knowledge in the domains of perception, cognitive science, and control theory, has been addressed. To accomplish this, it was decided to translate specific design questions to these domains. In the remaining part of this paper, the approach used to describe the information contents of the visual cues and the resulting control strategies will be discussed.
2. J Design questions The design of a spatial navigation display requires the specification of a frame of reference, a field of view, and a number of properties of the object representing the flightpath . Fig. 2 provides an overview of the different elements involved in the spatially integrated presentation of navigation data. In Fig. 2 Static Synthetic Data refers to data which describes abstractions of real-world objects. Symbology specification refers to data describing symbology which due to their specific representation have a particular meaning. Properties of the symbology such as position, orientation, col or, and size can be used to convey information. Dynamic Synthetic Data refers to data which describes the geometry of objects according to a set of representation rules and a forcing function . An example is the representation of the required trajectory. Based on the selection rules, the selection logic controls which data is to be presented. The transform rules determine the dynamic properties of the objects to be presented such as position, orientation, size, col or, and style. The data transformation applies the transform rules.
2.2 Analysis of the visual cues To benefit from research performed into perception and control of self motion, a model is needed which describes the relation between the data which must be observable and the magnitude of the visual cues which convey this data as a function of the display design parameters. For perspective flightpath displays the basis for this approach was developed by Wilckens (1973). Later, Warren (1982) derived similar equations to describe changes in the visual cues as properties of the optic flow field and introduced the concept of the splay angle. For a perspective flightpath display, the information content of the presentation is described by deriving a relation between the position and orientation errors of the aircraft and the resulting changes in the position and orientation of the perspectively presented trajectory. Fig. 3 shows the influence of a position error on the spatially integrated presentation of the flightpath .
300
~~~"""""~
::::::--....."'~~":,,,""
'\. \ \ \ \ I f / / / / ,,\\
~~~~,~"
=----~---------"
=-____________
'-. . .
.r"" /
~~/////
/
./ " "'., . ,
I
/
/
/ / / I
~/////////I
Fig. 3. Influence of a cross-track error on splay angle So. The dashed tunnel indicates the symmetrical reference condition and the solid tunnel is seen when the viewpoint is displaced to the left yielding a cross track error. Tunnel lines 1 to 4 rotate around the vanishing point as a result of the cross-track error. The change is splay angle is referred to as 6.S;. in which the index i is used to identify the tunnel line.
/1 ,/
/'./
,/
________ ....--".---........, / . / /'
~
-:;::::---,....---/"'/////
1///
\
\
\
\ \
\ \ \
"",
"
\ \\\'\.
Fig. 4. Velocity streamer pattern for the situation in which the track of the aircraft is aligned with the tunnel axis and the aircraft flies with a crab angle of 10 degrees due to crosswind. The direction of flight can be perceived from the location of the center of optic outflow.
2.3 Control strategies The presentation of the future forcing function and its constraints allows continuous compensatory, anticipatory (McRuer et aI., 1977), and error neglecting (Godthelp, 1984) control strategies to be applied. Continuous compensatory control is the strategy which is typically used with today's glides lope and localizer displays and flight directors . As indicated earlier in this paper, when splay rate is the functional variable for position control, an equal ratio increment in the variable should yield an equal interval improvement in tracking performance.
The rate of change of the angles between the tunnel lines and the line perpendicular to the horizon provides visual cues which are also known as splay rate cues. Owen (1990) describes several studies which all indicate that splay rate is the functional variable for altitude control. A synunetrical object as the one depicted in Fig. 3 provides splay rate cues both for vertical and horizontal position control. It can be assumed that when presenting both the horizontal and the vertical constraints. splay rate is the functional variable for horizontal and vertical position control.
The presence of preview on changes in the required trajectory combined with information about the temporal range towards these changes provides pilots with the possibility to apply anticipatory control. The accuracy of the anticipatory control action is detennined both by the accuracy with which the pilot can extract temporal range towards a change in the trajectory and the accuracy with which the magnitude of this change can be estimated.
Besides cues obtained through changes in splay angle, the optic flow field provides additional information about the direction of motion relative to the trajectory. These directional cues allow the observer to distinguish between the orientation of the vehicle body axis and the direction of vehicle motion. Fig. 4 shows the optic flow field combined with a snapshot of the tunnel for a situation in which a directional error is present.
When the task of the pilot is to remain within certain position constraints, instruments which force the pilot to continuously minimize position errors will unnecessarily increase task demanding load. In contrast, instruments which provide information about the current and future position margins that are available, allow the pilot to detennine whether a certain position or orientation error requires a control action or not. The resulting tracking accuracy is detennined by the threshold(s) the pilot uses when making a decision whether to intervene. Such an error-neglecting control strategy provides the pilot with the opportunity to make a trade-off between workload and tracking accuracy. With error-neglecting control there is a similarity with car driving in terms of control task (boundary control)
The dynamic presentation also conveys so-called temporal range information. This temporal range information can be used for the timing of anticipatory control actions, e.g. the moment to initiate a control action to enter a curve. In the situation of recti-linear motion, the dynamic perspective presentation of the trajectory allows the pilot to estimate the time until the center of projection reaches a certain reference point without knowing tunnel dimension or vehicle velocity. Lee (1976) refers to this phenomenon as time-tocontact (TIC). Kaiser and Mowafy (1993) showed that for rectilinear motion temporal range can already be extracted from the motion of a single point. They refer to this temporal range as time-to-passage (TTP).
301
and the type of visual cues (spatially integrated presention of constraints). When looking at the results from the time-to-line crossing (lLC) experiments performed by Godthelp (1984), this suggests that with this particular control-strategy, the pilot bases the moment of an error-correcting action on an estimate of the remaining time before the aircraft crosses one of the imaginary tunnel walls, the so-called time-to-wall crossing (1WC).
3. J Continuous compensatory control
Grunwald (1984) showed that with the addition of a posItIon predictor, tracking performance with perspective flightpath displays can be improved. To investigate the combined effects of error gain and position prediction on tracking performance and control activity for a closed loop compensatory control strategy on both the straight and the curved segments, two experiments have been conducted. In these experiments, six pilots had to fly a curved approach trajectory, which presented them with quite a challenging tracking task. The experiments were performed in the moving base flight simulator of Delft University of Technology. Both experiments were a two factor (error gain and position prediction) repeated measures design.
In the context of the Delft Program for Hybridized Instrumentation and Navigation Systems (DELPHINS), it has been investigated how the various design aspects influence the type and magnitude of the visual cues, what the consequences are for the translation of the data into useful information, and how useful the information is with respect to the ability to apply a particular control strategy. This analysis served to derive guidelines for the design aspects. To be able to investigate certain design questions in more detail, the concept has been implemented in a way which allows the manipulation of the design aspects for evaluation purposes. The following section discussed the results from experiments focusing on compensatory, anticipatory, and error-neglecting control strategies.
Fig. 5 shows the lateral tracking performance as a function of splay gain and Fig. 6 shows the aileron control activity as a function of splay gain. The label FPV indicates a condition in which a flightpath vector was presented, and the label FPP indicates the condition in which a position predictor was presented. The results have been split into those for tracking of straight segments (indicated by the -S) and those for tracking of curved segments (indicated by the -C). IS
3. EVALUATION
'\
For pilot-in-the-loop evaluation, tasks and measures are needed to quantitatively rate a certain design in terms of performance for the range of potential control strategies. Displays for guidance and control are typically compared with each other in terms of maximum tracking performance and control activity. With a well designed command display, maximum tracking perfonnance is achieved when the pilot applies a continuous compensatory control strategy. Since perspective flightpath displays allow a mix of compensatory, anticipatory, and error-neglecting control strategies to be applied, tasks and measures should be used to assess and to rate the possibility of applying each of these different types of control strategies.
....
I"
'" ~
1
~~
r-
Ih
-'~
[b.
o o
-- --
2
•
•
10
12
14
.,.lay •• /11 (tI."",,}
Fig. 5. Lateral tracking performance on straight and curved segments as a function of splay gain.
1.4 ", - - - - - - - - - - - - - - - - - ,
For evaluation purposes, the ability to apply a range of control strategy is a mixed blessing. A general problem when analyzing performance data obtained from pilotin-the-loop studies with perspective flightpath displays is that as a result of the range of potential control strategies, more trade-offs between effort and tracking performance are possible. This can be compensated for by defining tasks in which the pilot is forced to apply a particular control strategy. Although an isolated control strategy might not always result in typical control behavior, it allows the potential of different display formats to convey task relevant information to be compared. Some examples of tasks and measures to rate compensatory-, anticipatory-, and error-neglecting control strategies will be discussed next.
0.2
o : --------------~ o 2 4 6 8 10 12 14 spl.y ga/II [deglm]
Fig. 6. Aileron control activity on straight and curved segments as a function of splay gain. Fig. 5 illustrates that lateral tracking performance does not improve with a splay gain above approximately 2.5 deglm. Fig. 6 shows that both on straight and curved
302
segments the additional of a position predictor results in reduced control activity. Analysis of the data (Theunissen, 1997) led to the following conclusions: 1.
2.
i. -1 0
The variation in the size of the tunnel shows the effect on tracking perfonnance which is to be expected when changing the gain of the functional variable for position control, strengthening the hypothesis that with a perspective flightpath display the splay rate is a functional variable. With a position predictor, the splay rate is no longer the functional variable for position control.
+---:--~-"~\.~~'\."'--,,-+---.,----+-----j ~",
c:
== -15 +---;----..,.----""'\.~.,~.,:"""':"I~.--+--~-_-+~=....-1
e -20
-25
t-~--------~_T"~~~'~£.!:.~--;;....'-'---I
+
;............................; ......_. . ;....... -, . . . . . . . .·.·i·······_·+ ········--·I
-30 ! - - - - ' - - - - ' - - - - - ' - - - - ; - - ' - - - '
o
5
10
15
20
time {s} Fig. 8. Time histories of roll angle in the presence of a position predictor.
3.2 Anticipatory control
3.3 Error neglecting control
Each trajectory contained two curved segments, forcing the pilots to transition from a reference condition in which the wings are level to a reference condition in which the aircraft is banked to approximately 17 degrees for the commanded velocity and the specified turn radius. Theunissen (1997) illustrates that in an egocentric frame of reference, the magnitude of the visual cues conveying information about the rate of change of the direction of the future trajectory for a given turn radius makes it hard to accurately estimate the turn radius. Two parameters which can be used to rate the potential of a display format for anticipatory control are the deviation from a reference time which indicates the moment the open-loop control action should be initiated, and the deviation from a reference bank angle which is achieved as a result of the anticipatory control action. Fig. 7 shows a number of time histories of the bank angle during a change in the direction of the trajectory for a perspective flightpath display described in Theunissen (1997). Fig. 8 shows a number of time histories for the same part of the trajectory, but with the addition of a flightpath predictor in the display. As can be seen from Fig. 7, the basic tunnel does not provide adequate cues to support accurate timing and magnitude of the required anticipatory control action. Fig. 8 shows that the addition of the position predictor compensates for this deficiency. The results of the experiment clearly demonstrate that the integration of a position predictor allows the pilot to increase the accuracy of the anticipatory control actions.
An experiment was performed to verify whether the moment the pilot initiates an error-corrective control actions is related to a prediction of the time-to-wall crossing (TWC), and, if so, whether the TWC can be approximated with a first or second-order model. In this experiment, four pilots had to fly a number of trajectories consisting of two straight and one curved segments. To reduce the need to apply many compensatory control actions, a relatively low position error gain of 0.42 deglm (corresponding to a tunnel width of 135 m) was selected.
5
The task of the subjects was to try to minimize the number of control inputs while staying inside the tunnel, and only apply an error corrective action when the aircraft would otherwise violate the position constraints. In this way, the TWC measured at the moment a control action is initiated can be used as a measure to determine the ability to extract information about the constraints. Fig. 9 presents an estimate of the probability density function of TWC C2 , the 2nd order prediction of the time-to-wall crossing at the moment the pilot performs an error-corrective control action. 0.11
D.f4
O,f2
•
D.f
/1, ""lit
0.01
D
0,01
....---.. : -........ .---f.--.---...------D.H 0.02
o ·20
·f5
.fD
.,
0
5
10
,.
20
TWC .. re]
Fig. 9. Estimate of the probability density function of TWCC2 ' the 2nd order prediction of the timeto-wall crossing at the moment the pilot performs an error-corrective control action.
-25
-30
~----------------~--------~
o
5
10
15
20
time {s} Fig. 7. Time histories of roll angle in the absence of a position predictor.
Fig. 9 shows that with the type of perspective display
303
used in the experiment, a minimum temporal spacing of approximately 5 seconds was used. When forcing pilots to apply an error-neglecting control strategy, a distribution as depicted in Fig. 9 can be used to compare different display fonnat for their potential to convey infonnation about the margins towards the constraints. The experiment resulted in the following conclusions: I.
2.
3.
4.
ACKNOWLEDGMENTS The research into perspective flightpath displays is performed in the context of the Delft Program for Hybridized Instrumentation and Navigation Systems phase 11 (DELPHINS 11), which is being sponsored by the Dutch Technology Foundation STW.
As a result of the multitude of visual cues conveying position and orientation infonnation relative to the constraints, the perceptual response is not based on a single position or orientation error exceeding a certain threshold, but on a combination of these cues. The spatially integrated presentation of guidance data allows pilots to extract information which allows them to make better than first-order estimates of the time when the aircraft would cross a tunnel wall, enabling them to apply an efficient error-neglecting control strategy. Since pilots do not have to mentally integrate the values of position and angular errors and error rates and to verify whether the outcome exceeds a certain threshold which would be required for error-neglecting control with non-integrated displays, a perspective flightpath display reduces the task demanding load required for error neglecting control. When comparing different display concepts for the guidance task, an analysis of the TWC variable provides more insight into the pilot'S ability to utilize information about constraints.
REFERENCES Fadden, D.M., Braune, R & Wiedemann, J. (1987). Spatial Displays as a Means to Increase Pilot Situational Awareness. In: Spatial Displays and Spatial Instruments, NASA CPlOO32, pp. 35-1 to 35-12. Godthelp, H. (1984). Studies on Human Vehicle Control, PhD Thesis, Institute for Perception TNO, Soesterberg, The Netherlands. Grunwald, A.J. (1984). Tunnel Display for FourDimensional Fixed-Wing Aircraft Approaches, Journal of Guidance, Vol. 7, No. 3, pp. 369-377. Kaiser, M.K., and Mowafy, L. (1993). Visual Infonnation for Judging Temporal Range, Proc. of Piloting Vertical Flight Aircraft, pp. 4.23-4.27, San Francisco, CA. Lee, D.N. (1976). A Theory of Visual Control of Braking based on Infonnation about Time-toCollision, Perception, Vol. 5, pp. 437-459. McRuer, D.T., AlIen, RW., Weir, D.H., Klein, R.H. (1977). New Results in Driver Steering Control Models, Human Factors, Vol. 19, No. 4, pp. 381397. Owen, D.H. (1990). Perception & control of changes in self-motion: A functional approach to the study of information and skill. In: The Perception and Control of Self Motion (R Warren and A.H. Wertheim (Eds.)), Hillsdale, N.J. Theunissen, E. (1997). Integrated Design of a ManMachine Interface for 4-D Navigation. ISBN 90407-1406-1, Delft University Press, Delft, The Netherlands. Warren, R (1982). Optical transfonnation during movement: Review of the optical concomitants of egomotion, Technical Report AFOSR-TR-821028, Bolling AFB, DC: Air Force Office of Scientific Research (NTIS no. AD-A122 275). Wilckens, V (1973). Zur J..jjsung der FlugfUhrungsprobleme vomehmlich bei der Nullsicht-Landung mit der echt-perspektivischen, bildhaftquantitativen Kanal-Information. Technical University of Berlin, Germany.
4. CONCLUSIONS Spatial navigation displays are superior to current 2-D command displays because the former provide information about the current and future guidance requirements rather than just plain control commands. The design of display formats providing a spatially integrated presentation of data is more complex as compared to conventional 2-D command displays. One of the requirements for a more structured design process is a model describing how the various design parameters influence the magnitude of the task related visual cues. Based on findings from research into perception and control of self motion and findings from research into control strategies, an analysis of the information contents of the visual cues conveyed by a perspective flightpath display has been used to provide more insight into the relation between the design parameters and the potential control strategies. Pilot-inthe-loop experiments have confirmed that it is possible to use the information to apply compensatory, anticipatory, and error-neglecting control strategies, and that the tracking performance can be influenced by the splay gain.
304
INTEGRA TED DESIGN OF SPATIAL NAVIGA TION DISPLAYS
Eric Theunissen*, Henk Stassen* and Durk van Willigen*
*Faculty of Information Technology and Systems, Delft University of Technology, Mekelweg 4, 2628 CD Delft, The Netherlands
tFaculty of Design, Engineering and Production, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands
Abstract: Compared to current command displays, a spatially integrated presentation of the future trajectory has the potential to reduce the task demanding load for guidance and control while at the same time increasing spatial and navigational awareness. To allow a structured design of spatial navigation displays, this paper illustrates how the information contents of the visual cues can be described as properties of the optic flow field. The type of visual cues are related to the potential control strategies, and the results of experiments in which these control strategies are evaluated are discussed. Copyright © 19981FAC
Keywords: displays, navigation, human-machine interface, aircraft control, design.
introduction of more complex approach trajectories is likely to reduce safety. By providing pilots with the information they need to accomplish their task more easily, while maintaining spatial and navigational awareness without the need to scan additional displays, it becomes possible to fly more complex approaches without a reduction in safety. It is highly unlikely that with all future developments, safety can be increased by extrapolating current concepts. New functionality and new technology cannot simply be layered onto existing concepts, because the current system complexities are already too high. Better humanmachine interfaces may require a fundamentally new approach. Navigation, guidance, and control are not three independent tasks, but with the current flight director display the control task is isolated. Displays providing a spatially integrated presentation of the future trajectory, and thus presenting guidance requirements instead of control commands have advantages relative to current non-spatial displays. These advantages result from the fact that the pilot has to perform less integration of information and the fact that the more natural presentation requires less effort for interpretation and evaluation. Fig.l shows a spatial navigation display during a flight test.
1. INTRODUCTION Due to the increasing number of aircraft passengers and the resulting increase in aircraft, bottlenecks in airspace capacity are beginning to emerge. Once the number of aircraft has reached a certain threshold, any further increase will cause unacceptable delays. Since the bottlenecks mainly occur in the vicinity of airports, this is also the place where a solution should start. The basic idea behind improvements is to abandon today's straight-in approaches and allow aircraft to intercept the final straight glidepath at predetermined locations. This provides air traffic control with more possibilities to manage the traffic flow, creating the opportunity to increase capacity. An additional advantage is that it becomes possible to avoid noise sensitive areas by using noise abatement procedures. This concept, however, increases the task demandinK load of the pilot. In contrast to the current straight-in approach, pilots will have to fly curved approaches. Due to the more frequent changes in direction, it becomes harder to maintain an adequate level of spatial and navigational awareness. As a result, pilots will have to scan the navigation display more frequently. Because of the already high workload during the approach, an
299
F,.",.oIR.'.,..,-,c. \1.Mng
"."ne.
F,./r:I-cI·vi.1III "~oIconstta.inU
20 qmboIogy ruI.s
Fig. 1. Flight test of an experimental spatial navigation display. Fig. 2. Overview of the elements involved in the spatially integrated presentation of navigation data.
2. AN INTEGRATED DESIGN APPROACH A factor which has prevented the introduction of spatial displays for navigation and guidance is the complexity of their design. Fadden et al. (1987) indicated this problem by stating that 'while the promise of spatial displays is great, the cost of their development will be correspondingly large. The knowledge and skills which must be coordinated to ensure successful results is unprecedented. From the viewpoint of the designer, basic knowledge of how human beings perceive and process complex displays appears Jragmented and largely unquantified'. Hardly any detailed guidelines to the design of these types of displays exist which take specific human capabilities in the areas of perception, cognition, and control into account.
A proper selection of the design parameters requires an understanding of their relation with the magnitude of the visual cues conveyed by the display and the potential task strategies. Furthermore, tasks and measures are needed to evaluate the potential of a certain perspective flightpath display format for a range of control strategies. An approach is needed which supports a structured design process for perspective flightpath displays in which technical possibilities and human factors are truly integrated. To develop such an approach, the question on how to utilize the existing knowledge in the domains of perception, cognitive science, and control theory, has been addressed. To accomplish this, it was decided to translate specific design questions to these domains. In the remaining part of this paper, the approach used to describe the information contents of the visual cues and the resulting control strategies will be discussed.
2. J Design questions The design of a spatial navigation display requires the specification of a frame of reference, a field of view, and a number of properties of the object representing the flightpath . Fig. 2 provides an overview of the different elements involved in the spatially integrated presentation of navigation data. In Fig. 2 Static Synthetic Data refers to data which describes abstractions of real-world objects. Symbology specification refers to data describing symbology which due to their specific representation have a particular meaning. Properties of the symbology such as position, orientation, col or, and size can be used to convey information. Dynamic Synthetic Data refers to data which describes the geometry of objects according to a set of representation rules and a forcing function . An example is the representation of the required trajectory. Based on the selection rules, the selection logic controls which data is to be presented. The transform rules determine the dynamic properties of the objects to be presented such as position, orientation, size, col or, and style. The data transformation applies the transform rules.
2.2 Analysis of the visual cues To benefit from research performed into perception and control of self motion, a model is needed which describes the relation between the data which must be observable and the magnitude of the visual cues which convey this data as a function of the display design parameters. For perspective flightpath displays the basis for this approach was developed by Wilckens (1973). Later, Warren (1982) derived similar equations to describe changes in the visual cues as properties of the optic flow field and introduced the concept of the splay angle. For a perspective flightpath display, the information content of the presentation is described by deriving a relation between the position and orientation errors of the aircraft and the resulting changes in the position and orientation of the perspectively presented trajectory. Fig. 3 shows the influence of a position error on the spatially integrated presentation of the flightpath .
300
~~~"""""~
::::::--....."'~~":,,,""
'\. \ \ \ \ I f / / / / ,,\\
~~~~,~"
=----~---------"
=-____________
'-. . .
.r"" /
~~/////
/
./ " "'., . ,
I
/
/
/ / / I
~/////////I
Fig. 3. Influence of a cross-track error on splay angle So. The dashed tunnel indicates the symmetrical reference condition and the solid tunnel is seen when the viewpoint is displaced to the left yielding a cross track error. Tunnel lines 1 to 4 rotate around the vanishing point as a result of the cross-track error. The change is splay angle is referred to as 6.S;. in which the index i is used to identify the tunnel line.
/1 ,/
/'./
,/
________ ....--".---........, / . / /'
~
-:;::::---,....---/"'/////
1///
\
\
\
\ \
\ \ \
"",
"
\ \\\'\.
Fig. 4. Velocity streamer pattern for the situation in which the track of the aircraft is aligned with the tunnel axis and the aircraft flies with a crab angle of 10 degrees due to crosswind. The direction of flight can be perceived from the location of the center of optic outflow.
2.3 Control strategies The presentation of the future forcing function and its constraints allows continuous compensatory, anticipatory (McRuer et aI., 1977), and error neglecting (Godthelp, 1984) control strategies to be applied. Continuous compensatory control is the strategy which is typically used with today's glides lope and localizer displays and flight directors . As indicated earlier in this paper, when splay rate is the functional variable for position control, an equal ratio increment in the variable should yield an equal interval improvement in tracking performance.
The rate of change of the angles between the tunnel lines and the line perpendicular to the horizon provides visual cues which are also known as splay rate cues. Owen (1990) describes several studies which all indicate that splay rate is the functional variable for altitude control. A synunetrical object as the one depicted in Fig. 3 provides splay rate cues both for vertical and horizontal position control. It can be assumed that when presenting both the horizontal and the vertical constraints. splay rate is the functional variable for horizontal and vertical position control.
The presence of preview on changes in the required trajectory combined with information about the temporal range towards these changes provides pilots with the possibility to apply anticipatory control. The accuracy of the anticipatory control action is detennined both by the accuracy with which the pilot can extract temporal range towards a change in the trajectory and the accuracy with which the magnitude of this change can be estimated.
Besides cues obtained through changes in splay angle, the optic flow field provides additional information about the direction of motion relative to the trajectory. These directional cues allow the observer to distinguish between the orientation of the vehicle body axis and the direction of vehicle motion. Fig. 4 shows the optic flow field combined with a snapshot of the tunnel for a situation in which a directional error is present.
When the task of the pilot is to remain within certain position constraints, instruments which force the pilot to continuously minimize position errors will unnecessarily increase task demanding load. In contrast, instruments which provide information about the current and future position margins that are available, allow the pilot to detennine whether a certain position or orientation error requires a control action or not. The resulting tracking accuracy is detennined by the threshold(s) the pilot uses when making a decision whether to intervene. Such an error-neglecting control strategy provides the pilot with the opportunity to make a trade-off between workload and tracking accuracy. With error-neglecting control there is a similarity with car driving in terms of control task (boundary control)
The dynamic presentation also conveys so-called temporal range information. This temporal range information can be used for the timing of anticipatory control actions, e.g. the moment to initiate a control action to enter a curve. In the situation of recti-linear motion, the dynamic perspective presentation of the trajectory allows the pilot to estimate the time until the center of projection reaches a certain reference point without knowing tunnel dimension or vehicle velocity. Lee (1976) refers to this phenomenon as time-tocontact (TIC). Kaiser and Mowafy (1993) showed that for rectilinear motion temporal range can already be extracted from the motion of a single point. They refer to this temporal range as time-to-passage (TTP).
301
and the type of visual cues (spatially integrated presention of constraints). When looking at the results from the time-to-line crossing (lLC) experiments performed by Godthelp (1984), this suggests that with this particular control-strategy, the pilot bases the moment of an error-correcting action on an estimate of the remaining time before the aircraft crosses one of the imaginary tunnel walls, the so-called time-to-wall crossing (1WC).
3. J Continuous compensatory control
Grunwald (1984) showed that with the addition of a posItIon predictor, tracking performance with perspective flightpath displays can be improved. To investigate the combined effects of error gain and position prediction on tracking performance and control activity for a closed loop compensatory control strategy on both the straight and the curved segments, two experiments have been conducted. In these experiments, six pilots had to fly a curved approach trajectory, which presented them with quite a challenging tracking task. The experiments were performed in the moving base flight simulator of Delft University of Technology. Both experiments were a two factor (error gain and position prediction) repeated measures design.
In the context of the Delft Program for Hybridized Instrumentation and Navigation Systems (DELPHINS), it has been investigated how the various design aspects influence the type and magnitude of the visual cues, what the consequences are for the translation of the data into useful information, and how useful the information is with respect to the ability to apply a particular control strategy. This analysis served to derive guidelines for the design aspects. To be able to investigate certain design questions in more detail, the concept has been implemented in a way which allows the manipulation of the design aspects for evaluation purposes. The following section discussed the results from experiments focusing on compensatory, anticipatory, and error-neglecting control strategies.
Fig. 5 shows the lateral tracking performance as a function of splay gain and Fig. 6 shows the aileron control activity as a function of splay gain. The label FPV indicates a condition in which a flightpath vector was presented, and the label FPP indicates the condition in which a position predictor was presented. The results have been split into those for tracking of straight segments (indicated by the -S) and those for tracking of curved segments (indicated by the -C). IS
3. EVALUATION
'\
For pilot-in-the-loop evaluation, tasks and measures are needed to quantitatively rate a certain design in terms of performance for the range of potential control strategies. Displays for guidance and control are typically compared with each other in terms of maximum tracking performance and control activity. With a well designed command display, maximum tracking perfonnance is achieved when the pilot applies a continuous compensatory control strategy. Since perspective flightpath displays allow a mix of compensatory, anticipatory, and error-neglecting control strategies to be applied, tasks and measures should be used to assess and to rate the possibility of applying each of these different types of control strategies.
....
I"
'" ~
1
~~
r-
Ih
-'~
[b.
o o
-- --
2
•
•
10
12
14
.,.lay •• /11 (tI."",,}
Fig. 5. Lateral tracking performance on straight and curved segments as a function of splay gain.
1.4 ", - - - - - - - - - - - - - - - - - ,
For evaluation purposes, the ability to apply a range of control strategy is a mixed blessing. A general problem when analyzing performance data obtained from pilotin-the-loop studies with perspective flightpath displays is that as a result of the range of potential control strategies, more trade-offs between effort and tracking performance are possible. This can be compensated for by defining tasks in which the pilot is forced to apply a particular control strategy. Although an isolated control strategy might not always result in typical control behavior, it allows the potential of different display formats to convey task relevant information to be compared. Some examples of tasks and measures to rate compensatory-, anticipatory-, and error-neglecting control strategies will be discussed next.
0.2
o : --------------~ o 2 4 6 8 10 12 14 spl.y ga/II [deglm]
Fig. 6. Aileron control activity on straight and curved segments as a function of splay gain. Fig. 5 illustrates that lateral tracking performance does not improve with a splay gain above approximately 2.5 deglm. Fig. 6 shows that both on straight and curved
302
segments the additional of a position predictor results in reduced control activity. Analysis of the data (Theunissen, 1997) led to the following conclusions: 1.
2.
i. -1 0
The variation in the size of the tunnel shows the effect on tracking perfonnance which is to be expected when changing the gain of the functional variable for position control, strengthening the hypothesis that with a perspective flightpath display the splay rate is a functional variable. With a position predictor, the splay rate is no longer the functional variable for position control.
+---:--~-"~\.~~'\."'--,,-+---.,----+-----j ~",
c:
== -15 +---;----..,.----""'\.~.,~.,:"""':"I~.--+--~-_-+~=....-1
e -20
-25
t-~--------~_T"~~~'~£.!:.~--;;....'-'---I
+
;............................; ......_. . ;....... -, . . . . . . . .·.·i·······_·+ ········--·I
-30 ! - - - - ' - - - - ' - - - - - ' - - - - ; - - ' - - - '
o
5
10
15
20
time {s} Fig. 8. Time histories of roll angle in the presence of a position predictor.
3.2 Anticipatory control
3.3 Error neglecting control
Each trajectory contained two curved segments, forcing the pilots to transition from a reference condition in which the wings are level to a reference condition in which the aircraft is banked to approximately 17 degrees for the commanded velocity and the specified turn radius. Theunissen (1997) illustrates that in an egocentric frame of reference, the magnitude of the visual cues conveying information about the rate of change of the direction of the future trajectory for a given turn radius makes it hard to accurately estimate the turn radius. Two parameters which can be used to rate the potential of a display format for anticipatory control are the deviation from a reference time which indicates the moment the open-loop control action should be initiated, and the deviation from a reference bank angle which is achieved as a result of the anticipatory control action. Fig. 7 shows a number of time histories of the bank angle during a change in the direction of the trajectory for a perspective flightpath display described in Theunissen (1997). Fig. 8 shows a number of time histories for the same part of the trajectory, but with the addition of a flightpath predictor in the display. As can be seen from Fig. 7, the basic tunnel does not provide adequate cues to support accurate timing and magnitude of the required anticipatory control action. Fig. 8 shows that the addition of the position predictor compensates for this deficiency. The results of the experiment clearly demonstrate that the integration of a position predictor allows the pilot to increase the accuracy of the anticipatory control actions.
An experiment was performed to verify whether the moment the pilot initiates an error-corrective control actions is related to a prediction of the time-to-wall crossing (TWC), and, if so, whether the TWC can be approximated with a first or second-order model. In this experiment, four pilots had to fly a number of trajectories consisting of two straight and one curved segments. To reduce the need to apply many compensatory control actions, a relatively low position error gain of 0.42 deglm (corresponding to a tunnel width of 135 m) was selected.
5
The task of the subjects was to try to minimize the number of control inputs while staying inside the tunnel, and only apply an error corrective action when the aircraft would otherwise violate the position constraints. In this way, the TWC measured at the moment a control action is initiated can be used as a measure to determine the ability to extract information about the constraints. Fig. 9 presents an estimate of the probability density function of TWC C2 , the 2nd order prediction of the time-to-wall crossing at the moment the pilot performs an error-corrective control action. 0.11
D.f4
O,f2
•
D.f
/1, ""lit
0.01
D
0,01
....---.. : -........ .---f.--.---...------D.H 0.02
o ·20
·f5
.fD
.,
0
5
10
,.
20
TWC .. re]
Fig. 9. Estimate of the probability density function of TWCC2 ' the 2nd order prediction of the timeto-wall crossing at the moment the pilot performs an error-corrective control action.
-25
-30
~----------------~--------~
o
5
10
15
20
time {s} Fig. 7. Time histories of roll angle in the absence of a position predictor.
Fig. 9 shows that with the type of perspective display
303
used in the experiment, a minimum temporal spacing of approximately 5 seconds was used. When forcing pilots to apply an error-neglecting control strategy, a distribution as depicted in Fig. 9 can be used to compare different display fonnat for their potential to convey infonnation about the margins towards the constraints. The experiment resulted in the following conclusions: I.
2.
3.
4.
ACKNOWLEDGMENTS The research into perspective flightpath displays is performed in the context of the Delft Program for Hybridized Instrumentation and Navigation Systems phase 11 (DELPHINS 11), which is being sponsored by the Dutch Technology Foundation STW.
As a result of the multitude of visual cues conveying position and orientation infonnation relative to the constraints, the perceptual response is not based on a single position or orientation error exceeding a certain threshold, but on a combination of these cues. The spatially integrated presentation of guidance data allows pilots to extract information which allows them to make better than first-order estimates of the time when the aircraft would cross a tunnel wall, enabling them to apply an efficient error-neglecting control strategy. Since pilots do not have to mentally integrate the values of position and angular errors and error rates and to verify whether the outcome exceeds a certain threshold which would be required for error-neglecting control with non-integrated displays, a perspective flightpath display reduces the task demanding load required for error neglecting control. When comparing different display concepts for the guidance task, an analysis of the TWC variable provides more insight into the pilot'S ability to utilize information about constraints.
REFERENCES Fadden, D.M., Braune, R & Wiedemann, J. (1987). Spatial Displays as a Means to Increase Pilot Situational Awareness. In: Spatial Displays and Spatial Instruments, NASA CPlOO32, pp. 35-1 to 35-12. Godthelp, H. (1984). Studies on Human Vehicle Control, PhD Thesis, Institute for Perception TNO, Soesterberg, The Netherlands. Grunwald, A.J. (1984). Tunnel Display for FourDimensional Fixed-Wing Aircraft Approaches, Journal of Guidance, Vol. 7, No. 3, pp. 369-377. Kaiser, M.K., and Mowafy, L. (1993). Visual Infonnation for Judging Temporal Range, Proc. of Piloting Vertical Flight Aircraft, pp. 4.23-4.27, San Francisco, CA. Lee, D.N. (1976). A Theory of Visual Control of Braking based on Infonnation about Time-toCollision, Perception, Vol. 5, pp. 437-459. McRuer, D.T., AlIen, RW., Weir, D.H., Klein, R.H. (1977). New Results in Driver Steering Control Models, Human Factors, Vol. 19, No. 4, pp. 381397. Owen, D.H. (1990). Perception & control of changes in self-motion: A functional approach to the study of information and skill. In: The Perception and Control of Self Motion (R Warren and A.H. Wertheim (Eds.)), Hillsdale, N.J. Theunissen, E. (1997). Integrated Design of a ManMachine Interface for 4-D Navigation. ISBN 90407-1406-1, Delft University Press, Delft, The Netherlands. Warren, R (1982). Optical transfonnation during movement: Review of the optical concomitants of egomotion, Technical Report AFOSR-TR-821028, Bolling AFB, DC: Air Force Office of Scientific Research (NTIS no. AD-A122 275). Wilckens, V (1973). Zur J..jjsung der FlugfUhrungsprobleme vomehmlich bei der Nullsicht-Landung mit der echt-perspektivischen, bildhaftquantitativen Kanal-Information. Technical University of Berlin, Germany.
4. CONCLUSIONS Spatial navigation displays are superior to current 2-D command displays because the former provide information about the current and future guidance requirements rather than just plain control commands. The design of display formats providing a spatially integrated presentation of data is more complex as compared to conventional 2-D command displays. One of the requirements for a more structured design process is a model describing how the various design parameters influence the magnitude of the task related visual cues. Based on findings from research into perception and control of self motion and findings from research into control strategies, an analysis of the information contents of the visual cues conveyed by a perspective flightpath display has been used to provide more insight into the relation between the design parameters and the potential control strategies. Pilot-inthe-loop experiments have confirmed that it is possible to use the information to apply compensatory, anticipatory, and error-neglecting control strategies, and that the tracking performance can be influenced by the splay gain.
304