Skip to main content

Full text of "DTIC AD1002837: The Impact of Conflicting Spatial Representations in Airborne Unmanned Aerial System Sensor Control"

See other formats


Spatial Discordance 1 


Running head: SPATIAL DISCORDANCE IN AIRBORNE UAS OPERATIONS 


The impact of conflicting spatial representations in airborne unmanned aerial system sensor 

control 


Joseph W Geeseman, James E Patrey, Caroline Davy, Katherine Peditto, & Christine Zernickow 


Naval Air Systems Command 
AIR 4.6, Human Systems Department 
NAS Patuxent River, MD 20670 


DISTRIBUTION A. Approved for public release: distribution unlimited. 



Spatial Discordance 2 


Abstract 

The purpose of this study is to extend the basic research findings on spatial representations 
to a unique application area and extend the small number of applied research studies in this 
general area to a more robust, Navy-specific application set (i.e.. Unmanned Aerial Systems). In 
this study, participants assumed the role of a sensor operator for an unmanned aerial system 
(UAS) simulation while riding in the fuselage of an airborne Lockheed P-3 Orion. The P-3 flew 
a flight profile of intermittent ascending, descending, and turning profiles (in strict accordance 
with an emphasis on safety of flight) to induce a maximum level of spatial discordance to the 
sensor operator screen where the participant tracked simulated targets. Participants also 
performed trials on the ground with the laptop-based UAS sensor operator simulation to establish 
baseline performance. In a counterbalanced design, the participants performed trials while 
airborne and on the ground. The multiple frames of reference for the participant induced spatial 
discordance and an overall decrease in tracking performance compared to trials during straight 
and level flight and ground baseline trials. 



Spatial Discordance 3 


The impact of conflicting spatial representations in airborne unmanned aerial system sensor 

control 

Emerging concept of operations (CONOPs) for Unmanned Aerial System (UAS) 
employment for the US Navy calls for the potential control of UAS while airborne, such as 
controlling a broad area maritime surveillance (BAMS) UAS from an airborne P-8A or an 
unmanned carrier-launched surveillance and strike (UCLASS) UAS from a fighter-class aircraft. 
The utility of pairing these two types of aircraft is primarily to reduce the distance of the UAS 
and control station, thus reducing latency of control input and increasing bandwidth for 
information relay. The focus of this project is to assess the influence of disparate spatial 
representations on the human operator while executing sensor (i.e., camera) control. 

In general, performance on spatial tasks worsens as a result of conflicting and/or inconsistent 
spatial representations. Shepard & Metzler (1971) used 3D objects to test peoples’ ability to 
recognize new objects across different viewpoints and noted that participants performed faster 
when tested with the original learned view than with a new viewpoint. Moreover, they found 
that participants’ reaction time was a linear function of the angle between learned and presented 
view - the greater the change in position, the slower the response. This oft-cited “mental 
rotation” task implicates some cognitive “cost” to mentally rotate an object. Could the mental 
processes underpinning the rotation task be similar to those utilized when a UAV sensor operator 
tracks a target? We predict that the mental processes are indeed similar and, thus, target tracking 
performance will decline while in an airborne environment. 

The general finding of performance decrements due to mental manipulation of the spatial 
orientation of objects has been well-established over the past four decades. More recently, 



Spatial Discordance 4 


however, it has been determined to be subject to further distortions. For example, Simons & 
Wang (1998) discovered that the ability to assess the spatial configuration of objects is not 
merely dependent on the visual stimuli. They found that a viewpoint change does not impair the 
understanding of where objects are in relation to each other unless the viewpoint change is 
decoupled from a participant’s control. In other words, when a participant changed their 
viewpoint by physically changing their position, their ability to detect subtle changes in the 
spatial configuration of objects around a table went unchanged. When the scene was moved 
independent of the participant’s control, however, the ability to detect changes in the spatial 
layout of the objects was impaired. Similarly, Wraga, Creem-Regehr, and Proffitt (2004) found 
that the relative position of objects in a simulation was most accurately and quickly identified 
when the participant’s motion was embedded within the simulation. (For further examples of the 
value of coupling movement and viewpoint, see Jurgens & Becker, 2011; Klatzky et al., 1998; 
Rokos-Ewoldsen et al., 1998; Simons, Wang, & Roddenbery, 2002; Wraga, Creem-Regehr, & 
Proffitt, 2000; and/or see Previc & Ercoline, 2004, for an overview of spatial disorientation in 
aviation.) 

These effects have been shown to have real-world applicability as well. Muth, Walker, and 
Fiorello (2006) demonstrated that controlling one vehicle while riding in another reduced 
performance in their measure of accuracy; increased latency to complete their task; and resulted 
in a four-fold increase in motion sickness. This study used ground vehicles rather than aircraft, 
but clearly indicates that discordant spatial cues are a cause for concern. Similar findings have 
been discovered for tethered Remotely Piloted Vehicles (RPVs; see Hollands & Lamb, 2011). 

Performance decline has also been noted for aviation tasks in spatially discordant 
environments. Reed (1977) demonstrated that simulated turbulence impaired control of a RPV. 



Spatial Discordance 5 


Likewise, Olson and colleagues (2007) demonstrated that discordance between the platform 
motion and a controlled UAS significantly impairs performance. They conducted these 
experiments controlling a UAS simulator on a motion-platform (2006) and in an aircraft (2007), 
with vertical axis control and runway alignment impairments evident in both experiments. 
Additionally, their 2007 results suggest that the presence of horizon information creates greater 
impairment. Their methodology, however, was limited - participants were only seated in a 
forward configured seat in a civilian aircraft and only rudimentary maneuvers were used (e.g., 
turn, climb, and landing). 

Due to these limitations, the current experiment included multiple seat configurations and a 
highly dynamic flight profile. In this experiment, participants used a joystick to track a vehicle 
traversing a specified path around the airfield at NAS Patuxent River, MD. The tracking 
crosshairs were depicted via a “camera” on a simulated UAV flying an elliptical flight pattern 
around the base (see Figure 1). The system recorded target tracking error while the participants 
attempted to maintain the cursor on the center of the simulated vehicle as it “drove” around the 
base (see Figure 2). In a counterbalanced design, trials took place at a desk (i.e., ground trials) 
and in the air flying a figure-eight flight profile while quickly ascending and descending in a P-3 
(i.e., airborne trials). 

The current experimental environment posed a unique problem to the visual and vestibular 
systems of the participants. Participants operated an unmanned aerial system sensor simulator 
while riding in the aft fuselage of a large military aircraft (i.e., P-3 Orion). During airborne 
trials, the aircraft flew alternating straight-and-level and highly dynamic flight profiles to induce 
variable magnitudes of spatial disparity between observed visual stimuli and perceived vestibular 
information. Consider the conflicting vestibular and visuospatial representations among the 



Spatial Discordance 6 


movement of the aircraft and ones’ own position within the aircraft; the simulated UAS position 
and movement; and the ventral sensor (i.e., camera) position and orientation on the UAS. The 
translation and rotation difficulties within the aircraft (e.g., heave, sway, surge, pitch, yaw, roll) 
are easily and quickly resolved by most people, but an interesting problem arises when the 
participant is tasked to control a simulated camera on a simulated UAS that is in a different flight 
profile than they are experiencing. This task triples the degrees of freedom that must be resolved 
by the participant - the aircraft, UAS, and sensor all have their own six degrees of freedom. 

To provide an example of multiple degrees of freedom for this experiment, envision the 
following: An aircraft is banking left and descending; the simulated UAS is banking right and 
ascending; and the camera is slewing aft - for a person in a starboard seat, facing towards the 
center of the aircraft, great discord between these spatial representations and their relevant 
sensory inputs would result. This creates the likelihood not only of noteworthy momentary 
performance lapses until these spatial discrepancies are reconciled, but also pervasive and 
sustained fatigue and motion sickness - not to mention the consequent creation of significant 
safety risks. In addition to the behavioral measures previously discussed, we administered the 
motion sickness assessment questionnaire (MSAQ) after ground trials and airborne trials to 
evaluate the effect of the multiple degrees of freedom on the well-being of the participants (see 
Appendix A). 

Although this operational environment is quite unique, especially for experimentation, it 
should not be treated any different from other experimental environments (Gibson, 1966; 
Stroffregen & Riccio, 1988). In other words, the exceptionality of this operational environment 
elicits behavior specific to this environment as a result of the perceptual experience and, 
although unique, still provides meaningful information about conflicting spatial information 



Spatial Discordance 7 


resolution and motor behavior in a tracking task. This could be considered a limitation of the 
generalizability of the current study, but given the recent rapid advancement of technology in 
aviation, space, and underwater environments, value can be found through this examination of an 
infrequent environmental condition. 

Through the use of this distinct experimental environment, we predict that when participants 
are airborne, their ability to track the target with the simulated UAV camera will not be as 
accurate or consistent via several measures. First, the error distance between the cursor and the 
target is predicted to be larger when participants are airborne than during ground trials. Second, 
the variability of the cursor movement about the center of the target is predicted to be higher 
during trials in the plane. Third, the amount of time on target (TOT), or when the cursor is 
within five meters of the center of the target, is predicted to be longer during ground trials than 
during airborne trials. Finally, we predict that self-reports of motion sickness will indicate 
higher magnitude of motion sickness during air trials than ground trials. 

Method 

Participants 

Eight Naval Officers participated in this experiment. Participants had current flight 
physicals and were qualified to ride in military aircraft. These requirements were necessary 
given that data collection occurred on an aircraft during flight. All efforts to recruit an equal 
number of male and female participants were executed, but only males participated in this 
experiment - ages ranged from 33-49 (M = 40.75, SD = 6.88). Participants read and signed 
consent forms approved by the Naval Air Warfare Center - Aircraft Division Institutional 
Review Board (NAWC-AD IRB). Potential risks and their mitigations can be found in 


Appendix B. 



Spatial Discordance 8 


Fortunately, none of these risks or discomforts occurred to the point that experimentation 
had to be paused or cancelled. Headaches and nausea were reported, but will be discussed later 
in the results section. 

Materials 

Four networked Dell Precision M6800 laptop computers (17.3” screen) presented the 
UAS sensor simulators to the participants. Three of the computers ran a modified version of 
MetaVR™ v5.10 Scenario Creation Tool (SCT), which is a 3D real-time PC-based virtual 
environment generator. This software was modified to generate the sensor information (i.e., 
camera) from the ventral side of a GlobalHawk UAS. The flight profile approximated that of the 
GlobalHawk, and maintained an unclassified level of fidelity. 

The fourth computer ran the MetaVR™ v5.10 Virtual Reality Scene Generator (VRSG). 
This software generates surface information based on satellite information. Surface information 
of NAS Patuxent River, MD, was chosen because all participants were familiar with the 
geography and surface structures in and around this location. The fidelity of all visual stimuli 
approximated that of current (2014) video game technology. 

Thrustmaster™ T-Flight Hotas X Flight Stick joysticks connected to the three participant 
computers. The default transfer function of the joystick to the cursor on the screen was non¬ 
linear, in that, small movements in the joystick moved the cursor to a lesser extent than larger 
joystick movements in a non-linear manner (i.e., sigmoidal transfer function). In other words, 
small joystick movements would not noticeably move the cursor, whereas long-distance cursor 
movement could be achieved easily without moving the joystick to its maximal range. 

For this experiment, a Lockheed P-3 Orion provided the flight environment. The P-3 
Orion is a four-engine turboprop aircraft developed for the United States Navy. Although 



Spatial Discordance 9 


normally used for anti-submarine or maritime surveillance missions, the aircraft used for this 
experiment had a nearly empty fuselage rather than the usual configuration for military 
operations. The empty fuselage allowed participants to walk freely in between trial runs and 
socialize away from the experimental set-up in the aft area (i.e., galley) of the aircraft. For 
inexperienced readers, the length of the aircraft is 116’ 10”, leaving ample room for movement 
of the participants. 

Procedure 

In a counterbalanced design, participants completed ground trials either before or after 
the airborne trials. Due to the small sample size, initial conditions were pseudo-randomly 
assigned between one condition and the other in the order that participants signed up for the 
study rather than through truly random assignment. 

Regardless of initial condition assignment, each participant was trained to criterion on 
one of the laptop-based UAS sensor operator simulators. Criterion is a performance measure that 
is based on an individual’s performance, not a predefined level of proficiency. This simulation 
training occurred on the ground and included a basic introduction to the system displays and 
controls and an explanation of the mission. As previously mentioned, the mission was to keep 
crosshairs of a UAS sensor on a simulated ground vehicle driving a pre-programmed path around 
a local military base. 

For ground trials, participants chose a one-hour time period that suited their schedule 
within one week prior or after airborne testing depending on their initial condition assignment. 
Participants were seated in front of one of the UAS sensor operator laptops and they completed 
approximately 30-45 minutes of trials that lasted approximately 1-2 minutes each. 



Spatial Discordance 10 


At the beginning of each trial, the participant began with their “camera” fully zoomed-out 
in an attempt to prevent participants from easily replicating motor behavior to move the cursor 
over the target vehicle. Once the participant indicated that they were ready for the trial to begin, 
they were instructed to “zoom in” their camera to find the target vehicle and to keep the cursor 
on the center of the target as best they could. When the participant moved the cursor over the 
target vehicle, the researcher began a one-minute timer to designate the duration of a trial. After 
a minute of target tracking elapsed, the trial was over and the participant was instructed to zoom 
out their camera and wait for the next trial to begin. Upon completion of the ground trials, the 
participants were administered the MSAQ for comparison with the airborne trials. We 
implemented the same procedure for airborne trials, but with a few key differences. 

After take-off, the Principle Investigator set-up three test stations in the aft area of the P-3 
aircraft. This area is where the galley (i.e., kitchen) is located and contains bench seating at a 
large table and mounted stool seating at a small table. This seating configuration provided space 
for three participants to be run at a time in three different seating orientations: forward, 
backward, and center of the aircraft facing outboard. In addition to seating configuration, pilots 
flew two different flight profiles. 

The first flight profile, Profile Alpha, was a standard racetrack (i.e., oval) profile flown at 
one altitude. The second flight profile, Profile Beta, was a figure eight path flown in a 2000 feet 
per minute ascending and descending pattern. Two groups of three and one group of two 
participants completed 30 trials each. The Principal Investigator communicated with the pilots 
via electronic means to indicate when to change from Profile Alpha to Profile Beta - this 
manipulation was counterbalanced among the groups of participants as well. 



Spatial Discordance 11 


Data Analyses 

Analyses included data from all eight participants - no adverse events or instructions 
misunderstandings required any data exclusions. 

Residual distribution and link-function assignment to transform the data to a normal 
distribution were identified with Box-Cox analysis, and planned analyses were conducted with 
linear-mixed effects modeling (LME) in R (available at www.r-project.org) using the lme4 
package (Nelder & Wedderbum, 1972). Graphical representations of the data were produced in 
R, Microsoft Excel, and JMP v9, a graphical analysis product developed by SAS. 

Two within-subject variables (e.g., baseline/airborne trials, elapsed time) and one 
between-subject variable (e.g., seat configuration) were used as model predictors for three 
dependent variables (e.g., tracking error, tracking variability, proportion of time on target). 
Student’s t-test analyses of the MSAQ data were conducted as well. 

Residual distributions for tracking data were positively skewed, indicating that 
participants were more often near the target than far away from the target in both airborne and 
ground trials (see Figures 3 & 4). A Box-Cox analysis indicated that a modified log-transform 
normalized the distributions (see Figures 5 & 6). After data normalization, LME analyses 
revealed that the experimental manipulations influenced the participants’ ability to track targets. 
Results are written in a manner analogous to simple-effects effects tests to ease interpretation for 
the reader. 

Results 

Flight profile influenced tracking performance [t=l8.83, pc.OOl]. Post-hoc analyses 
revealed that all three trial types resulted in significantly different tracking performance of the 
participants. Ground trials resulted in the lowest tracking error, followed by trials conducted 



Spatial Discordance 12 


during Profile Alpha, and trials conducted during Profile Beta resulted in the worst tracking 
performance of the participants (see Figure 7). 

Further analyses of tracking performance revealed that seat position also influenced 
tracking performance [t=-6.48, pc.Ol]. Post-hoc analyses indicate that the best tracking 
performance was found during ground trials, forward and aft seating positions were significantly 
worse, and, surprisingly, an outboard-facing orientation resulted in the worst tracking 
performance (see Figure 8). Although we did not assign specific hypotheses to seat 
configuration as a predictor of performance, these results suggest that further investigation is 
needed and will be discussed further in the next section. 

Our second hypothesis predicted that the variability of cursor movement about the center 
of the target would be higher during airborne trials than ground trials. This relationship was 
found to be true, but there was a significant interaction between trial type and trial number for 
the variability of tracking performance [t=-18.01, pc.OOl]. This interaction reveals that although 
performance during airborne trials began more variable at the beginning of a block of data 
collection, performance became more stable as the experiment progressed. 

Next, we predicted that participants would spend less time on target (TOT) during 
airborne trials than during ground trials. An interaction of trial type and trial number predicted 
TOT performance [F(2,2) = 47.15. pc.OOl]. As predicted, TOT was worse during airborne trials, 
but only initially. Similar to tracking variability, as trials progressed TOT increased for airborne 
trials. Notably, however, TOT performance during ground trials decreased as trials progressed 
(see Figure 10). 

Finally, participants indicated that symptoms of motion sickness were more prevalent 
during airborne trials than during ground trials [F(l,14) = 4.70, pc.05] (see Figure 11). 



Spatial Discordance 13 


Headaches, fatigue, and nausea were the most common symptoms to become more pronounced 
during airborne trials. 

Discussion 

This experiment investigated the effects of multiple disparate sources of spatial 
information on motor tracking behavior. The environment in which this experiment occurred 
was novel and highly dynamic causing mismatched information from the visual and vestibular 
systems to negatively influence tracking performance and overall comfort of the participants. 

Due to the uniqueness of this testing environment, some may ask about the utility of the project - 
how are these findings extensible or applicable in other, not so novel, environments? 

At first glance, one may be inclined to side with the previously mentioned skeptical 
reader, but after reviewing the results of the current study, this section will outline other military 
applications and provide further questions that are better suited for the laboratory. The results 
generally supported the hypotheses posited for this study and some unexpected findings came to 
light as well. 

We asserted that flight condition (i.e., ground/airborne) would influence tracking 
performance and found that not only were differences in performance found between the 
baseline (i.e., ground) and testing (i.e., airborne) trials, but the phase of flight (i.e., Profile Alpha, 
Profile Beta) also resulted in performance differences. These results suggest that not only does 
highly dynamic motion negatively influence tracking performance, but fairly inert, stable motion 
can similarly lead to poorer tracking performance. 

During analyses of tracking performance, we noticed fairly obvious differences based on 
seat orientation. Although we did not address this topic as an original hypothesis due to a lack of 
existing evidence in distinguishing between forward, backward, or side-facing orientation in 



Spatial Discordance 14 


performance metrics, we decided to further scrutinize these differences. Similar to the first 
hypothesis, performance during ground trials was better than during airborne trials, but facing 
outboard during testing resulted in poorer tracking performance than facing either forward or aft. 
No interactions with seat orientation were found, but due to the small number of participants, it 
would not be surprising to find more interesting results that could elucidate the relationship of 
motion and spatial orientation with a larger number of participants. 

While not a significant difference, outboard-facing participants tended to report higher 
levels of headaches and nausea than forward or aft facing participants. Perhaps this unnatural 
traversing of space induces more spatial discordance than more natural forward or backward 
motion. To the uninformed reader, the solution to this issue seems simple - don’t orient seats 
facing inboard or outboard. It should be noted, however, that most military aircraft have side- 
oriented payload operator seats to save space within the aircraft. Therefore, it is cost-prohibitive 
to suggest reconfiguration of seating across hundreds of aircraft as a solution. In the next few 
paragraphs, we will discuss results that suggest perhaps training or repeated exposure to 
conflicting spatial information may result in performance analogous to baseline conditions, 
which is a much less costly solution. 

We predicted that tracking performance would be more variable while airborne than 
during ground trials. For the beginning of each data collection session while airborne, we found 
this relationship to be true - performance was more variable while airborne than during ground 
trials. As time progressed, however, performance variability returned to a profile similar to that 
of baseline. Would this initial increase in variability persist over time, or would repeated 
exposure eliminate variable tracking performance? 



Spatial Discordance 15 


Another interaction resulted between time on target (TOT) and trial number. As 
previously mentioned, TOT is when the crosshair was within 5 meters of the center of the 
tracking target. This interaction revealed that as time passed, TOT was initially highest during 
ground trials; initially lowest during Profile Beta (i.e., highly dynamic); and TOT was stable 
during Profile Alpha (i.e., not dynamic). Interestingly, during Profile Beta, performance became 
less variable and TOT increased as the experiment progressed, and the opposite relationship was 
found to be true for ground trials. The former result suggests that exposure to high motion 
environments while performing motor skill tasks warrants further investigation. Whereas the 
latter is indicative of task boredom; a more thorough investigation of this performance decrement 
should be considered as well. 

Finally, as expected, participants reported higher levels of motion sickness during 
airborne flights than during ground trials. Although these results could be confounded by 
demand characteristics of the experiment, the increased trend was only found in a subset of 
questions. In particular, questions involving headache, nausea, and fatigue indicated an increase 
in magnitude while other questions of perspiration, anxiety, and mood remained stable. Since an 
increase across all measures was not found, we can conclude that participants answered the 
MSAQ in a manner consistent with their self-assessment and not due to demand characteristics. 

Although this project took place in a difficult to access and highly specific environment 
outside of the laboratory, the lessons learned are extensible to other more highly controlled 
environments. For example, pilot studies for this project took place in a flight simulator with a 
6-DOF (i.e., degrees of freedom) motion base. This environment provided discrepant vestibular 
and visual information to the participants without requiring access to a large experimental 
aircraft - the results of the pilot studies were similar, but performance degraded to a smaller 



Spatial Discordance 16 


magnitude. We expect to see further research in similar environments to investigate the effect of 
repeated exposure in dynamic motion environments to evaluate if training is sufficient to 
overcome performance changes and motion sickness. 



Spatial Discordance 17 


References 

Antonov, G.S., Domogala, M.C. & Olson, W.A. (2007). Operating an unmanned aerial system 
from a moving platform. Institute For Information Technology Applications, United 
States Air Force Academy, November 2007. 

Gianaros, P. J., Muth, E. R., Mordkoff, J. T., Levine, M. E., & Stem, R. M. (2001). A 

questionnaire for the assessment of the multiple dimensions of motion sickness. Aviation, 
Space, and Environmental Medicine, 72, 115-119. 

Hollands J. G., & Lamb, M. (2011). Viewpoint tethering for remotely operated vehicles: effects 
on complex terrain navigation and spatial awareness. Human Factors, 53(2), 154-167. 

Jurgens, R & Becker, W. (2011). Human spatial orientation in non-stationary environments: 

relation between self-turning perception and detection of surround motion. Experimental 
Brain Research, 215(3/4), 327-344. 

Klatzky, R. L., Loomis, J. M., Beall, A. C., Chance, S. S., & Golledge, R. G. (1998). Spatial 
updating of self-position and orientation during real, imagined, and virtual locomotion. 
Psychological Science, 9, 293-298. 

Muth, E.R. (2009). The challenge of uncoupled motion: duration of cognitive and physiological 
aftereffects. Human Factors, 51(5), 752-761. 

Muth, E.R., Walker, A.D., & Liorello, M. (2006). Effects of uncoupled motion on performance. 
Human Factors, 48(3), 600-607. 


Nelder, J., & Wedderburn, R. (1972). Generalized linear models. Journcd of the Roycd Statistical 
Society, 135(3), 370-386. 

Olson, W. Al, DeLauer, E.H., & Fale, C. (2006). UAS control from a moving platform - a 
preliminary study. 3 ld Human Factors of UAVs Workshop, Mesa, AZ. 

Previc, F. & Ercoline, W.R. (2004). Spatial Disorientation in Aviation. Progress in Astronautics 
and Aeronautics, 203, AIAA. 

Reed, L.E. (1977). Visual-proprioceptive cue conflicts in the control of remotely piloted 

vehicles. Brooks AFB, TX: Air Force Resources Laboratory; Report No.: AFHRL-TR- 
77-57. 

Roskos-Ewoldsen, B., McNamara, T. P., Shelton, A. L., & Carr, W. (1998). Mental 

representations of large and small spatial layouts are orientation dependent. Journcd of 
Experimented Psychology: Learning, Memory and Cognition, 24, 215-226. 

Shepard, R and Metzler. J. (1971). "Mental rotation of three dimensional objects." Science 
171(972):701-3. 



Spatial Discordance 18 


Simons, D. J., & Wang, R. F. (1998). Perceiving real-world viewpoint changes. Psychological 
Science, 9, 315-320. 

Simons, D. J., Wang, R. X. F., & Roddenberry, D. (2002). Object recognition is mediated by 
extraretinal information. Perception & Psychophysics, 64, 521-530. 

Wraga, M., Creem, S. H., & Proffitt, D. R. (2000). Updating displays after imagined object and 
viewer rotations. Journal of Experimental Psychology: Learning, Memory, and 
Cognition, 26, 151-168. 


Wraga, M., Creem-Regehr, S. H., & Proffitt, D. R. (2004). Spatial updating of virtual displays 
during self- and display rotation. Memory and Cognition, 32, 399-415. 



Spatial Discordance 19 


Appendix A 

MOTION SICKNESS ASSESSMENT QUESTIONNAIRE (MSAQ) 

Instructions. Using the scale provided above each statement, please rate how accurately the 
following statements describe your experience. 

Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt sick to my stomach. 

Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt faint-like. 


Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt annoyed/irritated. 

Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt sweaty. 


Not at all 

1-2-3-4-5-6-7- 

I felt queasy. 


Severely 
8-9 


Not at all 

1-2-3-4-5-6-7- 

I felt lightheaded. 


Severely 
8-9 


Not at all 

1-2-3-4-5-6-7- 

I felt drowsy. 


Severely 
8-9 


Not at all 

1-2-3-4-5-6-7- 

I felt clammy/cold sweat. 


Severely 
8-9 


Not at all 
1 - 2 - 


■3-4-5-6-7- 

I felt disoriented. 


Severely 
8-9 



Spatial Discordance 


20 


Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt tired/fatigued. 

Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt nauseated. 


Not at all 

1-2-3-4-5-6-7- 

I felt hot/warm. 


Severely 
8-9 


Not at all 

1-2-3-4-5-6-7- 

I felt dizzy. 


Severely 
8-9 


Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt like I was spinning. 

Not at all Severely 

1-2-3-4-5-6-7-8-9 

I felt as if I may vomit. 


Not at all 
1 - 2 - 


-3-4-5-6-7- 

I felt uneasy. 


Severely 
8-9 



Spatial Discordance 21 


Appendix B 

Several potential risks and discomforts existed for participants. These risks and 

discomforts and their potential mitigations are listed below: 

Spatial Disorientation - the inability to correctly interpret aircraft attitude, altitude or 
airspeed, in relation to the Earth or point of reference, especially after a reference point 
(e.g., the horizon) has been lost. Spatial disorientation can escalate from a novel and 
interesting experience (e.g., amusement park rides) to a quite uncomfortable experience. 

If a participant experienced extreme disorientation to the point of discomfort, the 
experiment would have been temporarily stopped while the participant regained a point 
of reference to regain spatial orientation. If they chose to end their participation in the 
experiment, the aircraft would have returned to base and testing for that participant would 
have finished. 

Nausea - a very common symptom that is often described as a feeling of queasiness or 
wooziness, or a need to vomit. If a participant needed to vomit, a proper receptacle 
would have been provided. The experiment would have been temporarily stopped while 
the participant took measures to relieve their nausea. If they chose to end their 
participation in the experiment, the aircraft would have returned to base and testing for 
that participant would have finished. 

Hyperhidrosis - condition characterized by abnormally increased sweating/ perspiration 
in excess of that required for regulation of body temperature. This secondary effect of 
spatial disorientation and/or nausea is benign and no extraneous prevention or treatment 
was necessary. 

Headache - pain anywhere in the region of the head or neck. If a participant experienced 
headache, the experiment would have been temporarily stopped while the participant took 
measures to relieve their headache. If they chose to end their participation in the 
experiment, the aircraft would have returned to base and testing for that participant would 
have finished. 

Flight Mishap (including injury or death) - an occurrence associated with the operation 
of an aircraft, which takes place between the time any person boards the aircraft with the 
intention of flight until such time as all such persons have disembarked, where a person is 
fatally or seriously injured, the aircraft sustains damage or structural failure or the aircraft 
is missing or is completely inaccessible. All safety measures and training were 
completed by all aircrew aboard the aircraft prior to flight. 



Spatial Discordance 22 


Figure Captions 

Figure 1. View from simulated UAV at 10,000’. Trials begin in “zoomed-out” camera position 
to vary cursor starting position which eliminates rote motor movements. 

Figure 2. “Zoomed-in” view of trial where cursor is approximately 45° off target in the center of 
the screen. Note that Figure 2 is a zoomed-in image of Figure 1. 

Figure 3. Distribution of residuals for distance from target for airborne trials indicate a strong 
positive skew in the distribution. This distribution suggests participants kept the cursor closer to 
the target for more time than farther distances. 

Figure 4. Log transform of the distribution of residuals for distance from target for airborne trials 
normalizes the data for analyses. 

Figure 5. Distribution of residuals for distance from target for ground trials indicate a strong 
positive skew in the distribution. This distribution suggests participants kept the cursor closer to 
the target for more time than farther distances. 

Figure 6. Log transform of the distribution of residuals for distance from target for ground trials 
normalizes the data for analyses. 

Figure 7. Ground trials produced the least error distance from the target center. Profile Alpha 
(e.g., racetrack) trials led to significantly worse tracking than ground trials, and Profile Beta 
(e.g., dynamic) trials led to the worst average tracking performance of participants. 

Figure 8. Inboard-facing seat orientation resulted in the worst tracking performance. Forward 
and aft seating resulted in significantly better performance, and ground trials resulted in the best 
tracking performance. 

Figure 9. An interaction of trial type and trial number for tracking variability indicates that 
although tracking performance is initially more variable during airborne trials, it stabilizes and 
returns to baseline variability levels as trials progress. 

Figure 10. The proportion of time on target (i.e., within 5 meters) was determined by trial type 
and trial number. Performance improved for participants while airborne, but degraded during 
ground trials as trials progressed. 

Figure 11. Participants indicated a higher magnitude of motion sickness symptoms during 
airborne trials than during ground trials. 




Figure 1. 











Distribution of Residuals for Distance from 
Target for Airborne Trials 



Figure 3. 







Distribution of Residuals for Distance from 
Target for Airborne Trials - Log Transform 



Distance from Target (log) 


50 


Figure 4. 




















Distribution of Residuals for Distance from 
Target for Ground Trials 



Distance from Target (m) 


Figure 5. 








Distribution of Residuals for Distance from 
Target for Ground Trials - Log Transform 



Distance from Target (log) 


Figure 6. 























log(Distance) 


12.0 


Distance from Target by Flight Profile 


148.0 


11.5 


11.0 


10.5 


10.0 


9.5 


9.0 



Ground 



Profile Beta 


90.0 


55.0 


33.0 


20.0 


13.0 


Figure 7 . 


Profile Alpha 
Flight Profile 


7.5 


meters 





log(Distance) 


12.0 


Distance from Target by Seat Position 


148.0 


11.5 


11.0 


10.5 


10.0 


9.5 


9.0 

Ground 




Forward Aft Inboard 

Seat Position 


90.0 


55.0 


33.0 


20.0 


13.0 


7.5 


Figure 8. 


meters 











Tracking Variability across Trial Type and Trial Number 



Trial Type 


Ground 

Profile Alpha 

Profile Beta 


10.0 

9.0 


8.0 



0 2 4 6 8 10 14 18 22 0 2 4 6 8 10 14 18 22 0 2 4 6 8 10 14 18 22 


Trial Number 


Figure 9. 









Proportion of Time of Target (TOT) 


0.50 


Time on Target across Trial Type and Trial Number 


0.40 


0.30 


0.20 



Trial Type 

- Ground 

- Profile Alpha 

- Profile Beta 


0.10 


0.00 

-2 

Figure 10. 


0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 

Trial Number 





Motion Sickness Assessment Questionnaire and Trial Type 


35 -i 

30 

25 


0 ) 

k_ 

o 

u 



Ground 


Trial Type 



Airborne 


Figure 11.