lable at ScienceDirect

Applied Ergonomics 58 (2017) 198e207

Contents lists avai

Applied Ergonomics

journal homepage: www.elsevier .com/locate/apergo

Vibrotactile perception assessment for a haptic interface on anantigravity suit

Sang Min Ko a, Kwangil Lee b, Daeho Kim c, Yong Gu Ji a, *

a Department of Information and Industrial Engineering, Yonsei University, 50 Yonsei-Ro, Seodaemun-gu, Seoul, 03722, Republic of Koreab Jin Air Co., Ltd., 453 Gonghang-dearo, Gangseo-gu, Seoul, 07570, Republic of Koreac Republic of Korea Air Force Safety Center, P.O. Box 8, Yeouidaebang-ro 22-gil 77, Dongjak-gu, Seoul, 07056, Republic of Korea

a r t i c l e i n f o

Article history:Received 27 April 2015Received in revised form16 June 2016Accepted 24 June 2016Available online 5 July 2016

Keywords:HapticsVibrotactileAnti-G suit

* Corresponding author.E-mail address: yongguji@yonsei.ac.kr (Y.G. Ji).

http://dx.doi.org/10.1016/j.apergo.2016.06.0130003-6870/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

Haptic technology is used in various fields to transmit information to the user with or without visual andauditory cues. This study aimed to provide preliminary data for use in developing a haptic interface for anantigravity (anti-G) suit. With the structural characteristics of the anti-G suit in mind, we determined fiveareas on the body (lower back, outer thighs, inner thighs, outer calves, and inner calves) on which toinstall ten bar-type eccentric rotating mass (ERM) motors as vibration actuators. To determine the designfactors of the haptic anti-G suit, we conducted three experiments to find the absolute threshold, mod-erate intensity, and subjective assessments of vibrotactile stimuli. Twenty-six fighter pilots participatedin the experiments, which were conducted in a fixed-based flight simulator. From the results of ourstudy, we recommend 1) absolute thresholds of ~11.98e15.84 Hz and 102.01e104.06 dB, 2) moderateintensities of 74.36 Hz and 126.98 dB for the lower back and 58.65 Hz and 122.37 dB for either side of thethighs and calves, and 3) subjective assessments of vibrotactile stimuli (displeasure, easy to perceive, andlevel of comfort). The results of this study will be useful for the design of a haptic anti-G suit.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

The haptic sensation has different characteristics from those ofvisual and auditory sensations. The latter are perceived byspecialized organs, i.e., the eyes and ears, whereas the hapticsensation can occur at any part of the human body via physicalcontact (Iwata, 2008). Thus, the body surface, the largest organ ofthe body (~1.8e2.0 m2 for the average male) (Montagu, 1986), canbe used as a channel for communication. Haptic sensations canprovide an additional method of communication between systemsand operators in environments in which there is visual or auditoryoverload (Spence et al., 2010). In 1960, Geldard proposed the senseof touch as a means of communication (Geldard, 1960) and therehave been considerable advances and developments in thisresearch area over the past 50 years. Recently, haptic (or tactile)interfaces have been studied and applied in diverse fields, includingthe medical sciences (Mayer et al., 2007), mobile devices (Qianet al., 2011), vehicles (Baldwin and Lewis, 2014; Gu Ji and Jin,2010), virtual environments (Hale et al., 2009; Lahav and

Mioduser, 2008), simulator training (Gerling and Thomas, 2005),and assistive technologies (Kim et al., 2013; Nam et al., 2012). Theuse of haptic technology in military aviation also has been studied(Albery, 2007; Salzer et al., 2011; van Erp et al., 2006).

The design of modern fighter aircraft continues to trend towardsignificantly enhanced technical capabilities, complexity, and so-phistication to ensure the safety and reliability of the aircraft. As aresult, pilots often face increasing demands on their perceptual,cognitive, and physiological abilities (Hettinger and Haas, 2000).Fighter pilots have to maintain the proper altitude, course, andspeed of the aircraft in addition to performing tactical actions suchas managing weapons, communicating, and monitoring the envi-ronment of the aircraft. To perform these tasks successfully, pilotsneed to acquire and combine information from different sources,including the flight system, local airspace, and the terrain (Tannenet al., 2004). Under such circumstances, the pilot could experiencephysical stress and mental overload. These difficulties can be madeworse by high levels of fatigue, sustained acceleration, hypoxia, andvery poor visual conditions (Van Erp and Self, 2008). The visualchannel is the dominant method of communication between theflight system and the pilot. In addition, the auditory channel isconsidered an alternative or supplemental to the visual channel.

mailto:yongguji@yonsei.ac.krhttp://crossmark.crossref.org/dialog/?doi=10.1016/j.apergo.2016.06.013&domain=pdfwww.sciencedirect.com/science/journal/00036870http://www.elsevier.com/locate/apergohttp://dx.doi.org/10.1016/j.apergo.2016.06.013http://dx.doi.org/10.1016/j.apergo.2016.06.013http://dx.doi.org/10.1016/j.apergo.2016.06.013

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207 199

The flight system provides a great deal of information to the pilot invisual and auditory formats (Veen and Erp, 2001). Both channels ofa pilot, working under extreme pressure, are considerably loadedwith data so the ability of the pilot to process the data is degraded(van Erp and Verschoor, 2004).

The military aviation community and researchers from diversefields have begun to investigate the use of a haptic (or tactile)interface to counteract the challenges of sensory and cognitiveoverload. The United States Air Force (USAF) developed the spatialorientation retention device (SORD), a multisensory system meantto reduce spatial disorientation. Navathe and Singh (1994) reportedthat “spatial disorientation is the failure of a pilot to correctly sensethe attitude or motion of the aircraft, or of him or herself, resultingfrom inadequate or erroneous sensory information (from the re-ceptors).” Spatial disorientation is a primary concern to both civiland military aviation because it is a major cause of fatal accidents.The SORD consists of head-mounted display (HMD) symbology,three-dimensional audio, and tactile vest, all of which provide vi-sual, auditory, and tactile cues to the pilot. As a result, the SORD canhelp reduce pilot workload by providing complementary cues andthus improve aviation performance (Albery, 2007). The NavalAerospace Medical Research Laboratory (NAMRL) developed thetactile situation awareness system (TSAS) to help the pilot maintainspatial orientation and effective control of the aircraft with no vi-sual cues. The TSAS was flight-tested on a U.S. Navy T-34 aircraft.The test demonstrated that it is possible to improve pilot perfor-mance by providing intuitive three-dimensional spatial orientationand awareness information via a tactile display. Following thesuccess of the T-34 TSAS project, the U.S. Army integrated the tactiledisplay into a UH-60 helicopter (Chouvardas et al., 2005; McGrath,2000). Van Erp et al. (2003) investigated whether a tactile torsodisplay could compensate for degraded visual information whenflying while wearing night vision goggles and found that itimproved the performance of subjects under reduced (night) visionand full-vision conditions. The study of Van Erp et al. providedevidence that information can be delivered effectively to subjectsvia somatic tactile sense.

Previous studies demonstrated the possibility of the use of ahaptic (tactile) interface in military aviation. Haptic technology inmilitary aviation can improve performance and reduce workload,thereby reducing the risk of aviation accidents. However, previousstudies on the subject had some limitations. Most had presentedthe effectiveness of the haptic (tactile) interface in military aviationonly through a proof-of-concept, without any practical data for usein the development of actual equipment. Therefore, the aim of thepresent study was to obtain and provide preliminary data for use indeveloping a haptic (i.e., vibrotactile) anti-G suit to be worn in thefighter aircraft environment. We made a prototype of the suit andrecruited currently serving Air Force fighter pilots to participate inthe study. The experiments were conducted in the fixed-basedflight simulator to ensure ecological validity.

2. Methods

2.1. Participants

Twenty-six fighter pilots (24 men) from the Republic of KoreaAir Force (ROKAF) were recruited for the study after participating inregular safety education. The mean age of the participants was 31.9years (range ¼ 29e35 years), mean height was 172.4 cm(range ¼ 162e186 cm), and mean weight was 69.5 kg(range ¼ 57e82 kg). The flying experience of the participants var-ied; the aviation careers ranged from 5 to 12 years(mean ¼ 8.5 ± 1.9 years) and total flight time ranged from 514 to1592 h (mean ¼ 965.8 ± 271.8 h). Previous studies generally

classified study participants on the basis of their flying experience.Wiggins and O’Hare (2003) classified participants in their study onthe basis of their cross-country flying experience. Those who hadmore than 1000 h of cross-country flying experiencewere classifiedas experts and the remainder were classified as novices. In addition,Schriver et al. (2008) classified their participants (28 pilots) on thebasis of their flight-related expertise and flying experience. Par-ticipants were divided into two groups: “less expert” (14 pilots) and“more expert” (14 pilots). The more-expert group had more totalflight hours (481.9 vs. 110.5 h) and more total instrument flighthours (80.5 vs. 10.8 h) and outperformed the less-expert group onthe pilot skills test (14.6 vs. 13). The differences between the totalflight hours and total instrument flight hours for the less-expertand more-expert groups were statistically significant. However,there was no statistically significant difference in the pilot skill testscores.

Participants in the present study were formally trained tobecome fighter pilots and are currently serving as fighter pilots inthe ROKAF and participating in the regular classes. Based on theirflying experience (aviation career and total flight time), 13 of ourparticipants were less expert and 13 weremore expert pilots. Therewas onewoman in each group. Themore-expert group had a longeraviation career (10.2 vs. 6.8 years) andmore total flight time (1194.3vs. 737.1 h) than the less-expert group (Table 1).

2.2. Apparatus and materials

2.2.1. Vibrotactile actuatorThe vibrotactile stimuli were generated by a bar-type (cylinder-

type, 21.9 mm long) ERM vibration motor comprising a directcurrent (DC) motor and a rotor with an eccentric mass of 4 g. TheERM vibration motor has been used extensively to generatevibrotactile stimuli in various studies and commercial productsbecause of its low cost and structural simplicity. The generation ofvibrotactile stimuli depends on the revolutions per minute (rpm) ofthe DC motor and the weight of the eccentric mass. Small DC mo-tors are simple to control and can produce vibrations, but they havelimited power-to-mass ratios and the frequency and amplitude ofthe vibrations are difficult to control independently. Therefore,small DC motors are generally activated at a fixed frequency andamplitude (Jones and Sarter, 2008).

We used the following instruments to identify the relationshipbetween the voltage input and the physical characteristics of thevibrotactile stimuli, i.e., frequency and amplitude. The accelerationsensor (8692C50 triaxial accelerometer, Kistler Group, Winterthur,Switzerland) measured vibration along the x, y, and z axes simul-taneously. Data acquisition (NI USB-6251, National Instruments,Austin, TX, USA) translated the sampling data into digital signals tobe manipulated by a computer. The power supply and signal pro-cessor was type 5134B1 from the Kistler Group, and the data wereacquired and analyzed using LabVIEW 2012 software (NationalInstruments). These instruments acquired, measured, amplified,and converted the vibrotactile stimuli generated by the vibrationmotor (Ji et al., 2011). During the experiment, the vibration motor,which was attached to the acceleration sensor, was suspended inair to measure the data without any interference. After analyzingthe measured data, we estimated the relationship between theinput voltage (V), the output frequency (Hz), and the output vi-bration level (dB) of the vibrotactile stimuli (Fig. 1). Consequently,when the input voltage throughout the main experiment wasrecorded, it was possible to identify the frequency and the vibrationlevel.

2.2.2. Flight simulatorTo ensure the ecological validity of the experiments, we used the

Table 1Characteristics of the study participants (n ¼ 26).

Less-expert group (n ¼ 13) More-expert group (n ¼ 13) t (df) pAge (years) 30.5 ± 0.7 (29e31) 33.4 ± 1.0 (32e35) 8.533 0.000*Height (cm) 172.2 ± 6.0 (162e186) 172.6 ± 5.7 (166e183) 0.167 0.869Weight (kg) 68.8 ± 4.7 (57e78) 68.8 ± 4.7 (59e82) 0.706 0.487Aviation career (years) 6.8 ± 0.9 (5e8) 10.2 ± 0.9 (9e12) 9.453 0.000*Total flight time (h) 737.2 ± 139.4 (514e909) 1194.3 ± 145.8 (1040e1592) 8.177 0.000*

Values are mean ± standard deviation (SD) and ranges (in parentheses).*p < 0.001.

Fig. 1. Characteristics of the vibration actuator. (a) Relationship between the voltage (V) and the frequency (Hz). (b) Relationship between the voltage (V) and the vibration level(dB).

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207200

T-50 HQS flight simulator (Dodaam Systems Ltd., Daejeon-city,South Korea), which is used for pilot training, and the anti-G suitto simulate the environment of the cockpit in a fighter aircraft. TheT-50 HQS is a fixed-based flight simulator with a cockpit identical tothat of the T-50 and a visualization system that provides a virtualflight environment (Fig. 2). It is based on a geographic informationdatabase that uses real satellite photos and displays images in thefront of the cockpit through a three-channel rear-projectiondisplay. Moreover, the avionic software built into the simulatorprovides the same functions as that of the T-50, such as checkingthe operations and flight characteristics of the aviation electronicsequipment.

2.2.3. Anti-G suitThe CSU-13B/P anti-G suit, which is used by the ROKAF and the

USAF (Clere et al., 1995), is a pair of trousers with two layers of

Fig. 2. (a) T-50 HQS simulator and (b) a

nonstretch fabric and rubber bladders inserted at the abdomen, thefront of the thighs, and the front of the calves [Fig. 2(b)] (Balldin,2002). Pilots of high-performance fighter aircraft frequently expe-rience gravity-induced loss of consciousness (G-LOC) while per-forming aerobatic maneuvers such as inside loops and fallingleaves. G-LOC is a major threat to fighter pilots and presumably iscaused by reduced cerebral blood flow resulting in a decrease inoxygen supply to the brain (Kurihara et al., 2007). The anti-G suit, apiece of military equipment, uses air pressure to adjust blood flowand thus prevent G-LOC. For our experiments, the participantswore the anti-G suit over their flight suit [Fig. 2(b)].

The appropriate site for the vibrotactile actuator was deter-mined by first considering whether the participant could perceivethe vibrotactile stimuli while wearing the anti-G suit. Theabdomen, thighs, and calves of the fighter pilot were selected aspreferred sites for the actuators because the anti-G suit covered

nti-G suit worn over the flight suit.

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207 201

those areas. In this setup, the abdominal bladder is connected to theG-valve and all the bladders are interconnected. The G-valve sensesacceleration and provides gas pressure to the anti-G suit auto-matically (Balldin, 2002). Placement of the actuator over or under abladder could interfere with the primary function of the anti-G suitso these sites were excluded. In addition, the area of the anti-G suitthat directly touched the pilot seat was excluded.

We selected the five areas of body at which to place the bar-typeERM vibration motors inside the anti-G suit to generate thevibrotactile stimuli. Each actuator was mounted in the middle ofthe installation site. Two actuators were placed on the lower backand one each was placed on the outer thighs, the inner thighs, theouter calves, and the inner calves (Fig. 3). Three sizes of the anti-Gsuit, medium regular, medium long, and large regular, were pre-pared to accommodate the participants.

2.3. Experimental design

The goal of this study was to obtain preliminary data for use inthe development of a haptic (or vibrotactile) anti-G suit. To this end,we needed to determine the design factors involved in imple-menting the haptic interface of the anti-G suit in the fighter aircraftenvironment. Van Erp (2002) discussed four factors that affectvibrotactile parameters: stimulus detection, tactile informationcoding, comfort, and pitfalls. Tactile stimuli can be perceived whenthe energy of the stimulus exceeds the absolute or detectionthreshold. The absolute threshold dependsmainly on the frequencyof the stimulus and the location on the body where the stimulus isapplied. Furthermore, for the tactile stimuli to be perceived, theremust be actual contact between the actuator and the body. Thus,the vibrotactile stimuli must feel comfortable and unobtrusive tothe user during their use over long periods. The other guidelinesdiscussed the ability to discriminate tactile information and thepossible pitfalls in the application of tactile stimuli.

Self et al. (2008) discussed the human factors relevant to thedesign and implementation of a tactile display for in the militaryenvironment, including perceptual issues, coding principles,cognitive issues, multisensory integration, and user acceptance.Perceptual issues relevant to the perception of tactile stimuliinclude spatial acuity and absolute localization of the tactile stimuli.User acceptance includes acceptance and subjective assessment ofthe tactile stimuli, such as familiarity, intuitiveness, accuracy, andease of use. The remaining issues included the development of theintuitive tactile patterns, cognitive overload, and an appropriateway to integrate the tactile sense with the visual and auditorysenses.

Previous studies have involved tasks to determine perception oftactile stimulation to help design a tactile interface and gatherrelevant information. In addition, these studies suggested a need toinvestigate the subjective assessments of tactile stimulation.Accordingly, we aimed to examine the following three factors thatare fundamental to the design of the haptic interface of an anti-Gsuit for fighter pilots. The first factor is the absolute threshold,

Fig. 3. Installation sites for the vibrotactile actuators: (a) lower back, (b) both oute

which is the minimum magnitude of stimulus required for a testsubject to perceive vibration in an environment identical to that ofthe cockpit of an actual fighter plane. The second factor is the levelof vibrotactile intensity felt by the test subject before feelingoverburdened by the vibration. The third factor is the subjectiveassessment of what is felt by the user according to the site andintensity of the vibration. Consequently, we conducted three ex-periments to (1) identify the absolute threshold at each vibrotactileactuator installation site, (2) identify the moderate intensity of thevibrotactile stimuli, and (3) investigate the subjective assessmentswith the vibrotactile stimuli.

2.4. Procedures

Participants were provided with a verbal and written descrip-tion of the objective of the study and completed a questionnaire ondemographic information and flying experience. They were askedto wear the prototype haptic anti-G suit over their flight suit andthen take their position in the flight simulator, which providedinformation to the participants about the standby status of theaircraft prior to takeoff. During the experiment, the only task to beperformed was vibration recognition to prevent the effects ofexternal factors from other tasks. We provided the discrete vibro-tactile stimuli to the participants. Kirman (1974) suggested that theinterstimulus interval had to be less than the duration of thestimulus. Therefore, we used a 0.5-s interval between 1-s vibro-tactile stimuli. Noise-cancelling headphones blocked the sound ofthe actuators.

The first experiment determined the absolute threshold, i.e., theminimal magnitude of stimulus required for the test subject toperceive the vibration. The absolute threshold was determined bythe method of limits (Ehrenstein and Ehrenstein, 1999), a well-known method for detecting sensory threshold that has beenused in many studies on the sense of touch (Stuart et al., 2003;Zadeh et al., 2007). The method of limits consists of an equalnumber of descending and ascending series. The descending seriesof stimuli begins with a stimulus of clearly detectable intensity. Inthe present study, the descending series started at 121.495 Hz(equivalent to 134.065 dBe4.0 V). If the participant felt the stimulusand answered “Yes,” the stimulus intensity was decreased in smallsteps until the response was “No.” The ascending series of stimulibegins with a stimulus well below the expected threshold. In thepresent study, the ascending series started at 0 Hz (equivalent to0 dBe0 V). If the participant did not feel the stimulus and answered“No,” the stimulus intensity was increased in small steps until theresponse was “Yes.” This test was conducted three times, and theaverage of the transition points from the six descending andascending series was the absolute threshold.

The second experiment determined the moderate intensity ofthe vibrotactile stimuli presented to the five areas of the bodythrough the vibrotactile actuators. The intensities of the vibrotactilestimuli were classified into nine levels ranging from 27.23 Hz(equivalent to 109.77 dBe1.0 V) to 152.92 Hz (equivalent to

r thighs, (c) both inner thighs, (d) both outer calves, and (e) both inner calves.

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207202

133.16 dBe5.0 V) in 11.52-Hz (equivalent to 101.78 dBe0.5 V) in-tervals. The stimuli were presented in random order to the par-ticipants who were asked to rate each stimulus using the five-pointLikert scale (not detected, weak, moderate, strong, and too strong)(Ji et al., 2011; Kaaresoja and Linjama, 2005). This process wasperformed two times.

The third experiment evaluated the subjective assessments ofthe vibrotactile stimuli given by the test participants who perceivedthe vibrotactile stimuli depending on their application site andintensity. We used three modified evaluation questions that hadbeen used in other related studies (Ji et al., 2011; Koskinen, 2008)and that had the anticipated responses of displeasure, easy toperceive, and level of comfort. The participants provided subjectiveassessments of a vibrotactile pattern presented at the five actuatorsites. The intensity of the vibrotactile pattern was classified intothree levels: (1) 58.65 Hz (equivalent to 122.37 dBe2 V), (2)105.78 Hz (equivalent to 132.83 dBe3.5 V), and (3) 152.92 Hz(equivalent to 133.16 dBe5 V), and each stimulus was presented for5.5 s in random order. After each trial, participants were asked torate the evaluation questions using the five-point Likert scale,where a score of 1 was total disagreement and 5 was totalagreement.

2.5. Statistical analysis

Statistical analyses of the data were conducted using SPSS 21(SPSS Statistics, IBM, Chicago, IL, USA). An analysis of variance(ANOVA) and a t-test were used to assess the statistical significanceof the experimental factors. Significance was set at a ¼ 0.05. Anystatistical significance was followed up using Tukey’s honest sig-nificant difference (HSD) post hoc test to determine whether therewere significant differences.

3. Results

3.1. The absolute threshold at each installation site

We measured the absolute threshold of vibrotactile stimulipresented at five areas of the body via vibrotactile actuatorsinstalled in the anti-G suit (Fig. 3) and perceived by the test par-ticipants. The 26 test participants were divided into two groups,less expert and more expert, on the basis of their flying experience.The absolute thresholds for both groups at each site are shown inFig. 4 and the results of the analysis of these values are presented in

Fig. 4. Absolute thresholds at each installation site for the less-e

Table 2.Fig. 4 and Table 2 show that the minimal magnitude of stimulus

needed by the test participants to detect the vibrotactile stimuliwas highest at the lower back, where a stimulus of 15.15 Hz(equivalent to 103.70 dB) was needed by the less-expert group and16.54 Hz (equivalent to 104.41 dB) was needed by the more-expertgroup, and the lowest magnitude stimulus was at the inner thighs,i.e., 11.52 Hz (equivalent to 101.75 dB) and 12.43 Hz (equivalent to102.26 dB) for the novice and the expert groups, respectively. Thissignifies that stronger vibrotactile stimuli must be used and pre-sented to the lower back compared to the stimuli at the inner thighswhen information is presented to the fighter pilot via vibrotactilestimuli through the haptic interface to the anti-G suit. Furthermore,the magnitude of the absolute threshold was the highest at thelower back, followed by outer thighs, outer calves, inner calves, andinner thighs, in descending order, for both groups. However, therewere no statistically significant differences in the absolutethreshold at all sites between the less-expert and more-expertgroups. Therefore, we analyzed the data for all 26 test partici-pants together and the absolute thresholds at each site are asshown in Fig. 5 and the average and standard deviation are given inTable 3.

The data for all 26 test participants show that the absolutethreshold was highest at the lower back [15.84 Hz (equivalent to104.06 dB)] and lowest at the inner thighs [11.98 Hz (equivalent to102.01 dB)]. Furthermore, the magnitude of the absolute thresholdhad the same descending order as for the two separate groups. AnANOVA test for the absolute thresholds at the five vibrotactilestimuli sites showed statistically significant differences, in partic-ular, (1) voltage (V): F4,125 ¼ 7.722, p < 0.001; (2) frequency (Hz):F4,125 ¼ 7.723, p < 0.001; and (3) vibration level (dB): F4,125 ¼ 7.635,p < 0.001. A post hoc analysis using the Tukey’s HSD test showedthere were statistically significant differences between the absolutethreshold of the lower back and the other four sites but no statis-tically significant difference between the remaining four sites.

3.2. The moderate intensity of vibrotactile stimuli

The nine levels of vibrotactile stimuli were presented at the fivesites of the anti-G suit where the vibrotactile actuators wereinstalled as they were in the first test, and the moderate intensitiesperceived subjectively by the test participants were investigated.There were no statistically significant differences in the moderateintensity between the two groups at all the sites. These results are

xpert group (n ¼ 13) and the more-expert group (n ¼ 13).

Table 2Absolute thresholds at the each installation site for the less-expert group (n ¼ 13) and the more-expert group (n ¼ 13).

Installation sites

LB OT IT OC IC

Voltage(V)

LEG 0.615 (0.100) 0.539 (0.073) 0.500 (0.127) 0.531 (0.066) 0.521 (0.087)MEG 0.660 (0.139) 0.567 (0.075) 0.529 (0.081) 0.546 (0.081) 0.544 (0.093)p 0.271 0.722 0.191 0.298 0.558

Frequency(Hz)

LEG 15.15 (3.14) 12.73 (2.28) 11.52 (4.00) 12.49 (2.09) 12.19 (2.75)MEG 16.54 (4.38) 13.63 (2.35) 12.43 (2.56) 12.97 (2.55) 12.91 (2.92)p 0.270 0.724 0.191 0.298 0.559

Vibration level(dB)

LEG 103.70 (1.66) 102.42 (1.23) 101.75 (2.14) 102.30 (1.13) 102.13 (1.47)MEG 104.41 (2.27) 102.91 (1.25) 102.26 (1.39) 102.55 (1.37) 102.52 (1.56)p 0.304 0.675 0.185 0.311 0.555

LEG ¼ less-expert group; MEG ¼ more-expert group; LB ¼ lower back; OT ¼ outer thighs; IT ¼ inner thighs; OC ¼ outer calves; IC ¼ inner calves.The values are the mean and standard deviation (in parentheses), and the p-value is for the absolute threshold at each site tested.

Fig. 5. The absolute thresholds at each installation site for all participants (n ¼ 26).

Table 3Absolute thresholds at each installation site for all participants (n ¼ 26).

Installation sites

LB OT IT OC IC

Voltage (V) 0.638 (0.116)a 0.553 (0.068)b 0.514 (0.100)b 0.539 (0.064)b 0.533 (0.081)b

Frequency (Hz) 15.84 (3.66)a 13.18 (2.14)b 11.98 (3.13)b 12.73 (2.01)b 12.55 (2.56)b

Vibration level (dB) 104.06 (1.92) 102.67 (1.15)b 102.01 (1.68)b 102.43 (1.08)b 102.32 (1.38)b

LB ¼ lower back; OT ¼ outer thighs; IT ¼ inner thighs; OC ¼ outer calves; IC ¼ inner calves.Values are mean and standard deviation (in parentheses) for the absolute threshold. The values with different superscripts across the rows indicate significant differences.

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207 203

identical to the absolute threshold results. Similar to the first test, itwas confirmed that the difference in the moderate intensities be-tween the two groups created on the basis of the total flight timewas insignificant. Accordingly, we analyzed the data of all 26 testparticipants and present the results in Fig. 6.

Analysis of the moderate intensity at each site showed that65.4% (i.e., the highest) of the test participants determined that74.36 Hz (equivalent to 126.98 dBe2.5 V) was appropriate at thelower back. The appropriate level of moderate intensity wasdetermined to be 58.65 Hz (equivalent to 122.37 dBe2.0 V) at theouter thighs and the inner thighs by 63.5% and 59.6% of the testparticipants, respectively. Similarly, 58.65 Hz (equivalent to122.37 dBe2.0 V) was deemed an appropriate level of moderateintensity at the outer calves and the inner calves by 65.4% and 61.5%of the test participants, respectively. There were slight differencesin the range of 58.65e74.36 Hz for the moderate intensity deter-mined to be appropriate by the test participants at each location.

Furthermore, the proportion of test participants that determinedthat vibrotactile stimuli of 105.78 Hz (equivalent to132.83 dBe3.5 V) or higher was significantly lower, i.e.,

Fig. 6. Results from subjective scoring by all participants (n ¼ 26) of perception of moderate intensity at each installation site.

Table 4Subjective assessments at each installation site of all participants (n ¼ 26).

Displeasure Easy to perceive Level of comfort

2 V 3.5 V 5 V p 2 V 3.5 V 5 V p 2 V 3.5 V 5 V p

LB 2.52a

(0.806)3.88b

(0.816)4.38b

(0.697)0.000 4.25a

(0.764)4.69b

(0.471)4.77b

(0.514)0.004 2.68

(1.085)2.50(1.105)

2.19(0.981)

0.244

OT 3.09a

(1.017)3.81b

(0.849)4.23a

(0.815)0.000 4.08a

(0.796)4.77b

(0.430)4.77b

(0.430)0.000 2.65

(0.843)2.35(0.892)

2.15(0.834)

0.120

IT 3.52a

(0.943)4.00ab

(0.800)4.35b

(0.797)0.003 4.48

(0.500)4.73(0.452)

4.73(0.533)

0.118 2.29(0.874)

2.08(0.744)

1.88(0.766)

0.195

OC 2.85a

(0.925)3.12ab

(0.952)3.69b

(0.838)0.004 4.20

(0.693)4.27(0.604)

4.54(0.582)

0.129 2.60(0.849)

2.54(0.647)

2.38(0.941)

0.616

IC 3.20a

(1.058)3.58ab

(0.809)3.96b

(0.916)0.017 4.17

(0.834)4.50(0.583)

4.58(0.578)

0.072 2.49(0.944)

2.38(0.637)

2.23(0.765)

0.501

LB ¼ lower back; OT ¼ outer thighs; IT ¼ inner thighs; OC ¼ outer calves; IC ¼ inner calves.Values are mean and standard deviation (in parentheses) and indicate the ratings of displeasure, easy to perceive, and level of comfort. The values with different superscriptsacross the rows indicate significant differences.

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207204

participants together with the results summarized in Table 4,where the three levels (58.65, 105.78, and 152.92 Hz) of vibrotactileintensity at the five installation sites can be compared with respectto the three subjective assessments.

An ANOVA test of the “displeasure” responses at the threedifferent vibrotactile intensities showed statistically significantdifferences at all sites: lower back: F2,75 ¼ 40.358, p < 0.001; outerthighs: F2,75 ¼ 10.743, p < 0.001; inner thighs: F2,75 ¼ 6.213,p ¼ 0.003; outer calves: F2,75 ¼ 5.918, p ¼ 0.004; and inner calvesF2,75 ¼ 4.285, p ¼ 0.017. A post hoc analysis using Tukey’s HSD testshowed there were significant differences in the averages of“displeasure” at every intensity level for the lower back and theouter thighs and at 58.65 Hz (equivalent to 122.37 dBe2.0 V) and152.92 Hz (equivalent to 133.16 dBe5.0 V) for the three remainingsites. The results of the analysis of the “displeasure” data confirmedthat as the vibrotactile intensity increased, the number of testsubjects who felt “displeasure” increased consistently at every site.

The ANOVA test of the “easy to perceive” responses showedstatistically significant differences at the lower back (F2,75 ¼ 5.817,p ¼ 0.004) and outer thighs (F2,75 ¼ 12.145, p < 0.001). However,there was no statistically significant differences at the inner thighs(F2,75 ¼ 2.199, p ¼ 0.118), outer calves (F2,75 ¼ 2.108, p ¼ 0.129), andinner calves (F2,75 ¼ 2.728, p ¼ 0.072). The post hoc analysis usingTukey’s HSD test showed there were significant differences for theaverages of the “easy to perceive” responses at every vibrotactileintensity level for the lower back and outer thighs. The results ofthe analysis confirmed that as the vibrotactile intensity increased,

more test subjects easily perceived the vibrotactile stimuli at everysite.

Finally, an ANOVA test of the “level of comfort” responsesshowed no statistically significant differences at all sites: lowerback: F2,75 ¼ 1.436, p ¼ 0.244; outer thighs: F2,75 ¼ 2.180, p ¼ 0.120;inner thighs: F2,75 ¼ 1.672, p ¼ 0.195; outer calves: F2,75 ¼ 0.488,p ¼ 0.616; and inner calves F2,75 ¼ 0.697, p ¼ 0.501. The results ofthe analysis confirmed that as the vibrotactile intensity increased,the number of the test subjects who felt comfortable decreasedconsistently at every site.

4. Discussion

This study was meant to be a first step in the introduction of ahaptic interface in the cockpit of fighter aircraft. The anti-G suit,which must be worn by fighter pilots for flight safety, was thesubject of the study. The basic design factors of the study werebased on the general properties of a haptic interface and thestructural characteristics of the anti-G suit, and their practicalguidelines were assessed. We reviewed previous studies on haptictechnology and its use in the military aviation field. On the basis ofprevious studies, we identified three design factors that should beconsidered with respect to the haptic interface of the anti-G suit,i.e., absolute threshold, moderate intensity, and subjectiveassessments.

Preliminary data for use in developing a haptic interface for theanti-G suit can be derived from the results of this study. First, the

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207 205

results recommend the use of a minimal magnitude vibrotactilestimulus as follows: (1) lower back, 15.84 Hz and 104.06 dB; (2)outer thighs, 13.18 Hz and 102.67 dB; (3) inner thighs, 11.98 Hz and102.01 dB; (4) outer calves, 12.73 Hz and 102.43 dB; and (5) innercalves, 12.55eHz and 102.32 dB. The participants instantlyperceived these magnitudes of vibration at each installation site.Second, the results recommend vibrotactile stimuli of 58.65 Hz and122.37 dB for both sides of the thighs and calves and 74.36 Hz and126.98 dB for the lower back because participants perceived thosevalues as appropriate or moderate. Finally, we obtained subjectiveassessments (displeasure, easy to perceive, and level of comfort) forvibrotactile stimuli at each installation site.

Twenty-six fighter pilots (24 male) participated in the experi-ments. We could not recruit more female fighter pilots as testsubjects because their number in the ROKAF is very low. Weignored gender differences in this study so the results could beproblematic. Procacci et al. have suggested that there are differ-ences in the sensory functions between males and females, e.g.,females are more sensitive to pain than males at all ages (Procacciet al., 1974). However, according to Verrillo (1979), males and fe-males show no significant difference with respect to vibrotactilethresholds. Therefore, the results of our study may be applicable tofuture research, regardless of gender.

There is a fundamental difference between direct and mediatedtouch. When using a tool, the perceiver faces problems because ofthe variations introduced by the tool, such as the object at the placeof contact and the materials in contact (Hayward, 2008). In addi-tion, Sofia and Jones (2013) reported that the frequency of vibrationdecreased significantly when the motor generating the vibrationwas attached to the skin. In the present study, the experimentswere designed so that the vibrotactile stimuli generated by avibrotactile actuator were delivered to the subjects after theypropagated through the flight suit and the anti-G suit. Thus, thevibrotactile stimuli were delivered to the participants via mediatedtouch, i.e., there was a difference between the vibrotactile stimuligenerated by the vibrotactile actuator and the actual vibrotactilestimuli delivered to the participant. However, the flight suit and theanti-G suit used in the experiments were manufactured accordingto specific standards and with the same material (aramid fiber).Thus, similar results would be obtained if vibrotactile actuators areutilized with the same specifications as the ERM vibration motorused in the present study.

The participants were divided into two groups on the basis oftheir flying experience, i.e., less expert and more expert. Analysis ofthe experimental data showed that there were no statistically sig-nificant differences between the two groups in terms of all thestudied factors. This is attributed to the age distribution of thesubjects in the two groups. In general, touch perception is nega-tively affected by increasing age in all areas (Reuter et al., 2012). Inaddition, Stuart et al. (2003) confirmed that older subjects (55e90years) have significantly higher detection thresholds than youngsubjects (17e27 years) at all areas of the body, except the fingertips.The age range of the 26 participants in the present study was29e35 years, which is a narrow range. This explains the lack ofsignificant differences between the two groups.

Separating the fighter pilots into less-expert and more-expertgroups on the basis of flying experience should be considered infuture studies. The environment inside a fighter aircraft is highlyvolatile and dynamic and the pilots make very important instantjudgments and decisions based on situation awareness of thatenvironment. In addition, situation awareness differs according topilot expertise (Doane et al., 2004). A number of studies have usedthe total flight time to differentiate fighter pilots (Schriver et al.,2008; Wiggins and O’Hare, 2003), as was done in this study. Inthe present study, the perceptions of vibrotactile stimuli by the

fighter pilots was not a function of their expertise, so future studiesshould determine any differences in situation awareness and taskperformance between the two groups, in situations where hapticanti-G suits might be utilized.

We investigated the applicability of haptic technology to anti-Gsuits and presented several basic results that are required for actualimplementation and information coding. As is typical of mostmilitary aviation-related studies, this study has limitations becausethe experiments could not be conducted in a real fighter aircraftenvironment. Our experiments were conducted in a fixed-basedflight simulator that provided an environment similar to that ofan actual fighter cockpit. The fixed-based flight simulator, anti-Gsuit, and flight suit used in this study are equipment used inactual military training exercises by fighter pilots. A flight simulatorrealistically imitates actual tasks and pilot performancewhile flying(Yl€onen et al., 1997) and is a valuable tool for pilot training andlaboratory experiments. In addition, several studies have suggestedthat there are similarities in the reactions of a pilot in simulated andactual flights. Magnusson (2009) reported that there is very littledifference in the psychophysiological variables of reaction patternsbetween simulated and real flight. In addition, Veltman (2002)reported that physiological measures (e.g., heart rate, heart ratevariability, and respiratory frequency) were similar in the flightsimulator and in real flight. Previous studies have shown that thereare no significant differences between simulated and actual flight.Accordingly, the results of this study can be used as preliminarydata for further studies on a haptic anti-G suit in both simulatedand actual flight.

The results of this study can be utilized as design factors infuture studies of haptic interaction between various fighter aircraftsafety systems and anti-G suits. For example, the ground proximitywarning system (GPWS) is an electronic aviation tool that ensuresthe safety of the aircraft in flight. The GPWS device warns ofdangerous situations via an alarm light and a voice message whenthe aircraft approaches terrain that presents abnormal obstaclessuch as ground andmountains that cannot be perceived by the pilot(Breen, 1999). Thus, in addition to the current visual and auditorywarnings, further studies could investigate the presentation ofwarnings via the sense of touch by interlinking the GPWS with ahaptic anti-G suit to raise the awareness of the pilot about anabnormal situation. Situational awareness and reduction ofresponse time could be investigated by calculating the collisiontime using the signals received from the GPWS and the presenta-tion of a vibrotactile warning, in addition to the usual visual andauditory warnings that are appropriate in such situations. A hapticanti-G suit could also be used in situations where an instrumentapproach is needed because of limited visibility under severeweather conditions. The applicability and usability of a haptic anti-G suit to provide minimum altitude warnings, which could bepresented via vibrotactile stimuli during instrument approachwhen an aircraft reaches the decision height, could also beinvestigated.

Verrillo and Bolanowski (1986) reported that the maximumsensitivity of human skin is at approximately 250 Hz. However, theresults of such studies may depend on the equipment used and theconditions under which the studies were conducted. In addition,there could be differences in the subjective assessment of thevibrotactile system (Lieberman and Breazeal, 2007). Bikah et al.(2008) reported the mean frequency threshold at the head, neck,upper arm, wrist, waist, and ankle; interestingly, every locationtested had an absolute threshold of

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207206

et al. (2011) designed a haptic interface in vehicle seats based on theguidelines described in Ji et al. (2011). They showed that a hapticseat interface performed better than visual and auditory interfaces.In addition, it provides participants with information via multi-modal interfaces rather than with the haptic interface alone.

Future studies will focus on the implementation of a haptic anti-G suit to help perform various tasks in combat flying. We willconduct a study on the effect of a haptic anti-G suit on the cognitiveload and situation awareness of the fighter pilot. Moreover, we aimto study the overall impact of a haptic anti-G suit on the perfor-mance and safety of the flight system and on the evaluation of thesystem by fighter pilots.

5. Conclusion

This study investigated the perceptual characteristics and sub-jective assessments of vibrotactile stimuli when fighter pilots werewearing the anti-G suit. Despite some weak points in the flightsimulator and prototype of the haptic anti-G suit, we could set upthe fairly appropriate experiment environment. The results of thisstudy recommend absolute thresholds, moderate intensities, andsubjective assessments at the five areas of body. In addition, it isconcluded that flying experience did not affect the perceptualcharacteristics and subjective assessments for tests conducted inthis study.

The results of this study can be used to help develop a new wayfor a fighter pilot to interact with the fighter aircraft, such as ourproposed modification to the anti-G suit. Different types of flightinformation delivered to pilots from the fighter aircraft can beperceived readily and intuitively via applying haptic technology inthe anti-G suit. However, to ensure the reliability and validity of ourresults, it is necessary to perform additional validation procedures,which we will do by testing scenarios that take into account actualcombat flying situations and military operation environments,including the above-described flight safety systems. We will vali-date the study results and expand the scope of our study in addi-tional experiments.

References

Albery, W.B., 2007. Multisensory cueing for enhancing orientation informationduring flight. Aviat. Space. Environ. Med. 78, B186eB190.

Baek, Y., Myung, R., Yim, J., 2006. Have you ever missed a call while moving?: theoptimal vibration frequency for perception in mobile environments. In: Pro-ceedings of the 6th WSEAS International Conference on Applied Informaticsand Communications. World Scientific and Engineering Academy and Society(WSEAS), Elounda, Greece, pp. 241e245. Retrieved from. http://www.wseas.us/e-library/conferences/2006elounda1/papers/537-356.pdf.

Baldwin, C.L., Lewis, B.A., 2014. Perceived urgency mapping across modalitieswithin a driving context. Appl. Ergon. 45, 1270e1277.

Balldin, U., 2002. Acceleration effects on fighter pilots. Med. Asp. Harsh Environ. 2,1014e1027. Retrieved from. http://citeseerx.ist.psu.edu/viewdoc/download?doi¼10.1.1.175.3309&rep¼rep1&type¼pdf.

Bikah, M., Hallbeck, M.S., Flowers, J.H., 2008. Supracutaneous vibrotactile percep-tion threshold at various non-glabrous body loci. Ergonomics 51, 920e934.

Breen, B.C., 1999. Controlled flight into terrain and the enhanced ground proximitywarning system. IEEE Aerosp. Elctron. Syst. Magazine 14, 19e24.

Chang, W., Hwang, W., Ji, Y.G., 2011. Haptic seat interfaces for driver informationand warning systems. Int. J. Hum-Comput. Int. 27, 1119e1132.

Chouvardas, V.G., Miliou, A.N., Hatalis, M.K., 2005. Tactile display applications: astate of the art survey. In: Proceedings of the 2nd Balkan Conference in Infor-matics, pp. 290e303. Retrieved from. http://citeseerx.ist.psu.edu/viewdoc/download?doi¼10.1.1.60.2942&rep¼rep1&type¼pdf.

Clere, J., Ossard, G., Kergu�elen, M., Center, F.T., 1995. Anti-G Suits. AGARD, CurrentConcepts on G-Protection Research and Development 8 p(SEE N 95-34050 12-54). Retrieved from. http://www.dtic.mil/dtic/tr/fulltext/u2/a296761.pdf#page¼40.

Doane, S.M., Woo Sohn, Y., Jodlowski, M.T., 2004. Pilot ability to anticipate theconsequences of flight actions as a function of expertise. Hum. Factors 46,92e103.

Ehrenstein, W., Ehrenstein, A., 1999. Psychophysical methods. In: Windhorst, U.,Johansson, H. (Eds.), Modern Techniques in Neuroscience Research. Springer,

Berlin Heidelberg, pp. 1211e1241.Geldard, F.A., 1960. Some neglected possibilities of communication. Science 131,

1583e1588.Gerling, G.J., Thomas, G.W., 2005. Augmented, pulsating tactile feedback facilitates

simulator training of clinical breast examinations. Hum. Factors 47, 670e681.Gu Ji, Y., Jin, B.S., 2010. Development of the conceptual prototype for haptic interface

on the telematics system. Int. J. Hum-Comput. Int. 26, 22e52.Hale, K.S., Stanney, K.M., Malone, L., 2009. Enhancing virtual environment spatial

awareness training and transfer through tactile and vestibular cues. Ergonomics52, 187e203.

Hayward, V., 2008. Haptic shape cues, invariants, priors and interface design. In:Grunwald, M. (Ed.), Human Haptic Perception: Basics and Applications.Birkh€auser, Basel, pp. 381e392.

Hettinger, L.J., Haas, M.W., 2000. Current research in advanced cockpit displayconcepts. Int. J. Aviat. Psychol. 10, 227e229.

Iwata, H., 2008. History of haptic interface. In: Grunwald, M. (Ed.), Human HapticPerception: Basics and Applications. Birkh€auser, Basel, pp. 355e361.

Ji, Y.G., Lee, K., Hwang, W., 2011. Haptic perceptions in the vehicle seat. Hum. Factor.Ergon. Man. 21, 305e325.

Jones, L.A., Sarter, N.B., 2008. Tactile displays: guidance for their design andapplication. Hum. Factors 50, 90e111.

Kaaresoja, T., Linjama, J., 2005. Perception of short tactile pulses generated by avibration motor in a mobile phone. In: Eurohaptics Conference, 2005 andSymposium on Haptic Interfaces for Virtual Environment and TeleoperatorSystems, 2005. World Haptics 2005. First Joint, pp. 471e472.

Kim, H.N., Smith-Jackson, T.L., Nam, C.S., 2013. Elicitation of haptic user interfaceneeds of people with low vision. Int. J. Hum-Comput. Int. 29, 488e500.

Kirman, J.H., 1974. Tactile apparent movement: the effects of interstimulus onsetinterval and stimulus duration. Percept. Psychophys. 15, 1e6.

Koskinen, E., 2008. Optimizing Tactile Feedback for Virtual Buttons in Mobile De-vices. Helsinki University of Technology. Retrieved from. http://www.soberit.hut.fi/T-121/shared/thesis/Master's-Thesis-Emilia-Koskinen.pdf.

Kurihara, K., Kikukawa, A., Kobayashi, A., Nakadate, T., 2007. Frontal corticaloxygenation changes during gravity-induced loss of consciousness in humans:a near-infrared spatially resolved spectroscopic study. J. Appl. Physiol. 103,1326e1331.

Lahav, O., Mioduser, D., 2008. Haptic-feedback support for cognitive mapping ofunknown spaces by people who are blind. Int. J. Hum-Comput. Stud. 66, 23e35.

Lieberman, J., Breazeal, C., 2007. TIKL: development of a wearable vibrotactilefeedback suit for improved human motor learning. IEEE Trans. Robot. 23,919e926.

Magnusson, S., 2009. Similarities and differences in psychophysiological reactionsbetween simulated and real air-to-ground missions. Int. J. Aviat. Psychol. 12,49e61.

Mayer, H., Nagy, I., Knoll, A., Braun, E.U., Bauernschmitt, R., Lange, R., 2007. Hapticfeedback in a telepresence system for endoscopic heart surgery. Presence Tel-eop. Virt. Environ. 16, 459e470.

McGrath, B.J., 2000. Tactile Instrument for Aviation. DTIC Document. Retrievedfrom. http://www.dtic.mil/dtic/tr/fulltext/u2/a531665.pdf.

Montagu, A., 1986. Touching: the Human Significance of the Skin, third ed. Harperand Row, New York.

Nam, C.S., Li, Y., Yamaguchi, T., Smith-Jackson, T.L., 2012. Haptic user interfaces forthe visually impaired: implications for haptically enhanced science learningsystems. Int. J. Hum-Comput. Int. 28, 784e798.

Navathe, P.D., Singh, B., 1994. An operational definition for spatial disorientation.Aviat. Space. Environ. Md 65, 1153e1155.

Procacci, P., Zoppi, M., Maresca, M., Romano, S., 1974. Studies on the pain thresholdin man. Adv. Neurol. 4, 107e113.

Qian, H., Kuber, R., Sears, A., 2011. Towards developing perceivable tactile feedbackfor mobile devices. Int. J. Hum. Comput. Stud. 69, 705e719.

Reuter, E.M., Voelcker-Rehage, C., Vieluf, S., Godde, B., 2012. Touch perceptionthroughout working life: effects of age and expertise. Exp. Brain Res. 216,287e297.

Salzer, Y., Oron-Gilad, T., Ronen, A., Parmet, Y., 2011. Vibrotactile “on-thigh” alertingsystem in the cockpit. Hum. Factors 53, 118e131.

Schriver, A.T., Morrow, D.G., Wickens, C.D., Talleur, D.A., 2008. Expertise differencesin attentional strategies related to pilot decision making. Hum. Factors 50,864e878.

Self, B., Van Erp, J., Eriksson, L., Elliott, L., 2008. Human Factors Issues of TactileDisplays for Military Environments, Tactile Displays for Orientation, Navigationand Communication in Air, Sea and Land Environments (Report No. TRHFM-122, p. 3e1-3-18). Neuilly-sur-Seine Cedex, France: North Atlantic Treaty Or-ganization, Research and Technology Organization. Retrieved from. http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page¼53.

Sofia, K.O., Jones, L.A., 2013. Mechanical and psychophysical studies of surface wavepropagation during vibrotactile stimulation. IEEE Trans. Haptics 6, 320e329.

Spence, C., Ngo, M.K., Lee, J.-H., Tan, H., 2010. Solving the Correspondence Problemin Haptic/Multisensory Interface Design, pp. 47e74. Retrieved from. https://engineering.purdue.edu/~hongtan/pubs/PDFfiles/B3_Spence_etal_AdvancesInHaptics_Chap3_2010.pdf.

Stuart, M., Turman, A.B., Shaw, J., Walsh, N., Nguyen, V., 2003. Effects of aging onvibration detection thresholds at various body regions. BMC Geriatr. 3, 1.

Tannen, R.S., Nelson, W.T., Bolia, R.S., Warm, J.S., Dember, W.N., 2004. Evaluatingadaptive multisensory displays for target localization in a flight task. Int. J.Aviat. Psychol. 14, 297e312.

http://refhub.elsevier.com/S0003-6870(16)30124-7/sref1http://refhub.elsevier.com/S0003-6870(16)30124-7/sref1http://refhub.elsevier.com/S0003-6870(16)30124-7/sref1http://www.wseas.us/e-library/conferences/2006elounda1/papers/537-356.pdfhttp://www.wseas.us/e-library/conferences/2006elounda1/papers/537-356.pdfhttp://refhub.elsevier.com/S0003-6870(16)30124-7/sref3http://refhub.elsevier.com/S0003-6870(16)30124-7/sref3http://refhub.elsevier.com/S0003-6870(16)30124-7/sref3http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.175.3309&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.175.3309&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.175.3309&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.175.3309&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.175.3309&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.175.3309&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.175.3309&rep=rep1&type=pdfhttp://refhub.elsevier.com/S0003-6870(16)30124-7/sref5http://refhub.elsevier.com/S0003-6870(16)30124-7/sref5http://refhub.elsevier.com/S0003-6870(16)30124-7/sref5http://refhub.elsevier.com/S0003-6870(16)30124-7/sref6http://refhub.elsevier.com/S0003-6870(16)30124-7/sref6http://refhub.elsevier.com/S0003-6870(16)30124-7/sref6http://refhub.elsevier.com/S0003-6870(16)30124-7/sref7http://refhub.elsevier.com/S0003-6870(16)30124-7/sref7http://refhub.elsevier.com/S0003-6870(16)30124-7/sref7http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.60.2942&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.60.2942&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.60.2942&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.60.2942&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.60.2942&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.60.2942&rep=rep1&type=pdfhttp://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.60.2942&rep=rep1&type=pdfhttp://www.dtic.mil/dtic/tr/fulltext/u2/a296761.pdf#page=40http://www.dtic.mil/dtic/tr/fulltext/u2/a296761.pdf#page=40http://www.dtic.mil/dtic/tr/fulltext/u2/a296761.pdf#page=40http://refhub.elsevier.com/S0003-6870(16)30124-7/sref10http://refhub.elsevier.com/S0003-6870(16)30124-7/sref10http://refhub.elsevier.com/S0003-6870(16)30124-7/sref10http://refhub.elsevier.com/S0003-6870(16)30124-7/sref10http://refhub.elsevier.com/S0003-6870(16)30124-7/sref11http://refhub.elsevier.com/S0003-6870(16)30124-7/sref11http://refhub.elsevier.com/S0003-6870(16)30124-7/sref11http://refhub.elsevier.com/S0003-6870(16)30124-7/sref11http://refhub.elsevier.com/S0003-6870(16)30124-7/sref12http://refhub.elsevier.com/S0003-6870(16)30124-7/sref12http://refhub.elsevier.com/S0003-6870(16)30124-7/sref12http://refhub.elsevier.com/S0003-6870(16)30124-7/sref13http://refhub.elsevier.com/S0003-6870(16)30124-7/sref13http://refhub.elsevier.com/S0003-6870(16)30124-7/sref13http://refhub.elsevier.com/S0003-6870(16)30124-7/sref14http://refhub.elsevier.com/S0003-6870(16)30124-7/sref14http://refhub.elsevier.com/S0003-6870(16)30124-7/sref14http://refhub.elsevier.com/S0003-6870(16)30124-7/sref15http://refhub.elsevier.com/S0003-6870(16)30124-7/sref15http://refhub.elsevier.com/S0003-6870(16)30124-7/sref15http://refhub.elsevier.com/S0003-6870(16)30124-7/sref15http://refhub.elsevier.com/S0003-6870(16)30124-7/sref16http://refhub.elsevier.com/S0003-6870(16)30124-7/sref16http://refhub.elsevier.com/S0003-6870(16)30124-7/sref16http://refhub.elsevier.com/S0003-6870(16)30124-7/sref16http://refhub.elsevier.com/S0003-6870(16)30124-7/sref16http://refhub.elsevier.com/S0003-6870(16)30124-7/sref17http://refhub.elsevier.com/S0003-6870(16)30124-7/sref17http://refhub.elsevier.com/S0003-6870(16)30124-7/sref17http://refhub.elsevier.com/S0003-6870(16)30124-7/sref18http://refhub.elsevier.com/S0003-6870(16)30124-7/sref18http://refhub.elsevier.com/S0003-6870(16)30124-7/sref18http://refhub.elsevier.com/S0003-6870(16)30124-7/sref18http://refhub.elsevier.com/S0003-6870(16)30124-7/sref19http://refhub.elsevier.com/S0003-6870(16)30124-7/sref19http://refhub.elsevier.com/S0003-6870(16)30124-7/sref19http://refhub.elsevier.com/S0003-6870(16)30124-7/sref20http://refhub.elsevier.com/S0003-6870(16)30124-7/sref20http://refhub.elsevier.com/S0003-6870(16)30124-7/sref20http://refhub.elsevier.com/S0003-6870(16)30124-7/sref21http://refhub.elsevier.com/S0003-6870(16)30124-7/sref21http://refhub.elsevier.com/S0003-6870(16)30124-7/sref21http://refhub.elsevier.com/S0003-6870(16)30124-7/sref21http://refhub.elsevier.com/S0003-6870(16)30124-7/sref21http://refhub.elsevier.com/S0003-6870(16)30124-7/sref22http://refhub.elsevier.com/S0003-6870(16)30124-7/sref22http://refhub.elsevier.com/S0003-6870(16)30124-7/sref22http://refhub.elsevier.com/S0003-6870(16)30124-7/sref23http://refhub.elsevier.com/S0003-6870(16)30124-7/sref23http://refhub.elsevier.com/S0003-6870(16)30124-7/sref23http://www.soberit.hut.fi/T-121/shared/thesis/Master%27s-Thesis-Emilia-Koskinen.pdfhttp://www.soberit.hut.fi/T-121/shared/thesis/Master%27s-Thesis-Emilia-Koskinen.pdfhttp://refhub.elsevier.com/S0003-6870(16)30124-7/sref25http://refhub.elsevier.com/S0003-6870(16)30124-7/sref25http://refhub.elsevier.com/S0003-6870(16)30124-7/sref25http://refhub.elsevier.com/S0003-6870(16)30124-7/sref25http://refhub.elsevier.com/S0003-6870(16)30124-7/sref25http://refhub.elsevier.com/S0003-6870(16)30124-7/sref26http://refhub.elsevier.com/S0003-6870(16)30124-7/sref26http://refhub.elsevier.com/S0003-6870(16)30124-7/sref26http://refhub.elsevier.com/S0003-6870(16)30124-7/sref27http://refhub.elsevier.com/S0003-6870(16)30124-7/sref27http://refhub.elsevier.com/S0003-6870(16)30124-7/sref27http://refhub.elsevier.com/S0003-6870(16)30124-7/sref27http://refhub.elsevier.com/S0003-6870(16)30124-7/sref28http://refhub.elsevier.com/S0003-6870(16)30124-7/sref28http://refhub.elsevier.com/S0003-6870(16)30124-7/sref28http://refhub.elsevier.com/S0003-6870(16)30124-7/sref28http://refhub.elsevier.com/S0003-6870(16)30124-7/sref29http://refhub.elsevier.com/S0003-6870(16)30124-7/sref29http://refhub.elsevier.com/S0003-6870(16)30124-7/sref29http://refhub.elsevier.com/S0003-6870(16)30124-7/sref29http://www.dtic.mil/dtic/tr/fulltext/u2/a531665.pdfhttp://refhub.elsevier.com/S0003-6870(16)30124-7/sref31http://refhub.elsevier.com/S0003-6870(16)30124-7/sref31http://refhub.elsevier.com/S0003-6870(16)30124-7/sref32http://refhub.elsevier.com/S0003-6870(16)30124-7/sref32http://refhub.elsevier.com/S0003-6870(16)30124-7/sref32http://refhub.elsevier.com/S0003-6870(16)30124-7/sref32http://refhub.elsevier.com/S0003-6870(16)30124-7/sref33http://refhub.elsevier.com/S0003-6870(16)30124-7/sref33http://refhub.elsevier.com/S0003-6870(16)30124-7/sref33http://refhub.elsevier.com/S0003-6870(16)30124-7/sref34http://refhub.elsevier.com/S0003-6870(16)30124-7/sref34http://refhub.elsevier.com/S0003-6870(16)30124-7/sref34http://refhub.elsevier.com/S0003-6870(16)30124-7/sref35http://refhub.elsevier.com/S0003-6870(16)30124-7/sref35http://refhub.elsevier.com/S0003-6870(16)30124-7/sref35http://refhub.elsevier.com/S0003-6870(16)30124-7/sref36http://refhub.elsevier.com/S0003-6870(16)30124-7/sref36http://refhub.elsevier.com/S0003-6870(16)30124-7/sref36http://refhub.elsevier.com/S0003-6870(16)30124-7/sref36http://refhub.elsevier.com/S0003-6870(16)30124-7/sref37http://refhub.elsevier.com/S0003-6870(16)30124-7/sref37http://refhub.elsevier.com/S0003-6870(16)30124-7/sref37http://refhub.elsevier.com/S0003-6870(16)30124-7/sref38http://refhub.elsevier.com/S0003-6870(16)30124-7/sref38http://refhub.elsevier.com/S0003-6870(16)30124-7/sref38http://refhub.elsevier.com/S0003-6870(16)30124-7/sref38http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page=53http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page=53http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page=53http://refhub.elsevier.com/S0003-6870(16)30124-7/sref40http://refhub.elsevier.com/S0003-6870(16)30124-7/sref40http://refhub.elsevier.com/S0003-6870(16)30124-7/sref40https://engineering.purdue.edu/~hongtan/pubs/PDFfiles/B3_Spence_etal_AdvancesInHaptics_Chap3_2010.pdfhttps://engineering.purdue.edu/~hongtan/pubs/PDFfiles/B3_Spence_etal_AdvancesInHaptics_Chap3_2010.pdfhttps://engineering.purdue.edu/~hongtan/pubs/PDFfiles/B3_Spence_etal_AdvancesInHaptics_Chap3_2010.pdfhttp://refhub.elsevier.com/S0003-6870(16)30124-7/sref42http://refhub.elsevier.com/S0003-6870(16)30124-7/sref42http://refhub.elsevier.com/S0003-6870(16)30124-7/sref43http://refhub.elsevier.com/S0003-6870(16)30124-7/sref43http://refhub.elsevier.com/S0003-6870(16)30124-7/sref43http://refhub.elsevier.com/S0003-6870(16)30124-7/sref43

S.M. Ko et al. / Applied Ergonomics 58 (2017) 198e207 207

Van Erp, J., Self, B., 2008. Introduction to Tactile Displays in Military Environments,Tactile Displays for Orientation, Navigation and Communication in Air, Sea andLand Environments (Report No. TRHFM-122, p. 1e1-1-18). Neuilly-sur-SeineCedex, France: North Atlantic Treaty Organization, Research and TechnologyOrganization. Retrieved from. http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page¼17.

Van Erp, J., Veltman, J., Van Veen, H., Oving, A., 2003. Tactile Torso Display asCountermeasure to Reduce Night Vision Goggles Induced Drift. DTIC Document.Retrieved from. http://www.dtic.mil/dtic/tr/fulltext/u2/a413343.pdf#page¼477.

Van Erp, J.B., 2002. Guidelines for the use of vibro-tactile displays in human com-puter interaction. In: Proceedings of Eurohaptics, pp. 18e22. Retrieved from.http://www.eurohaptics.vision.ee.ethz.ch/2002/vanerp.pdf.

van Erp, J.B., Groen, E.L., Bos, J.E., van Veen, H.A., 2006. A tactile cockpit instrumentsupports the control of self-motion during spatial disorientation. Hum. Factors48, 219e228.

van Erp, J.B., Verschoor, M.H., 2004. Cross-modal visual and vibrotactile tracking.Appl. Ergon. 35, 105e112.

Veen, H.H.C., Erp, J.F., 2001. Tactile information presentation in the cockpit. In:Brewster, S., Murray-Smith, R. (Eds.), Haptic Human-computer Interaction.Springer, Berlin Heidelberg, pp. 174e181.

Veltman, J.A., 2002. A comparative study of psychophysiological reactions duringsimulator and real flight. Int. J. Aviat. Psychol. 12, 33e48.

Verrillo, R.T., 1979. Comparison of vibrotactile threshold and suprathreshold re-sponses in men and women. Percept. Psychophys. 26, 20e24.

Verrillo, R.T., Bolanowski Jr., S.J., 1986. The effects of skin temperature on the psy-chophysical responses to vibration on glabrous and hairy skin. J. Acoust. Soc.Am. 80, 528e532.

Wiggins, M.W., O’Hare, D., 2003. Expert and novice pilot perceptions of static in-flight images of weather. Int. J. Aviat. Psychol. 13, 173e187.

Yl€onen, H., Lyytinen, H., Leino, T., Lepp€aluoto, J., Kuronen, P., 1997. Heart rate re-sponses to real and simulated BA Hawk MK 51 flight. Aviat. Space. Environ.Med. 68, 601e605.

Zadeh, M.H., Wang, D., Kubica, E., 2007. Perception-based lossy haptic compressionconsiderations for velocity-based interactions. Multimed. Syst. 13, 275e282.

http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page=17http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page=17http://www.dtic.mil/dtic/tr/fulltext/u2/a492500.pdf#page=17http://www.dtic.mil/dtic/tr/fulltext/u2/a413343.pdf#page=477http://www.dtic.mil/dtic/tr/fulltext/u2/a413343.pdf#page=477http://www.eurohaptics.vision.ee.ethz.ch/2002/vanerp.pdfhttp://refhub.elsevier.com/S0003-6870(16)30124-7/sref47http://refhub.elsevier.com/S0003-6870(16)30124-7/sref47http://refhub.elsevier.com/S0003-6870(16)30124-7/sref47http://refhub.elsevier.com/S0003-6870(16)30124-7/sref47http://refhub.elsevier.com/S0003-6870(16)30124-7/sref48http://refhub.elsevier.com/S0003-6870(16)30124-7/sref48http://refhub.elsevier.com/S0003-6870(16)30124-7/sref48http://refhub.elsevier.com/S0003-6870(16)30124-7/sref49http://refhub.elsevier.com/S0003-6870(16)30124-7/sref49http://refhub.elsevier.com/S0003-6870(16)30124-7/sref49http://refhub.elsevier.com/S0003-6870(16)30124-7/sref49http://refhub.elsevier.com/S0003-6870(16)30124-7/sref50http://refhub.elsevier.com/S0003-6870(16)30124-7/sref50http://refhub.elsevier.com/S0003-6870(16)30124-7/sref50http://refhub.elsevier.com/S0003-6870(16)30124-7/sref51http://refhub.elsevier.com/S0003-6870(16)30124-7/sref51http://refhub.elsevier.com/S0003-6870(16)30124-7/sref51http://refhub.elsevier.com/S0003-6870(16)30124-7/sref52http://refhub.elsevier.com/S0003-6870(16)30124-7/sref52http://refhub.elsevier.com/S0003-6870(16)30124-7/sref52http://refhub.elsevier.com/S0003-6870(16)30124-7/sref52http://refhub.elsevier.com/S0003-6870(16)30124-7/sref53http://refhub.elsevier.com/S0003-6870(16)30124-7/sref53http://refhub.elsevier.com/S0003-6870(16)30124-7/sref53http://refhub.elsevier.com/S0003-6870(16)30124-7/sref54http://refhub.elsevier.com/S0003-6870(16)30124-7/sref54http://refhub.elsevier.com/S0003-6870(16)30124-7/sref54http://refhub.elsevier.com/S0003-6870(16)30124-7/sref54http://refhub.elsevier.com/S0003-6870(16)30124-7/sref54http://refhub.elsevier.com/S0003-6870(16)30124-7/sref54http://refhub.elsevier.com/S0003-6870(16)30124-7/sref55http://refhub.elsevier.com/S0003-6870(16)30124-7/sref55http://refhub.elsevier.com/S0003-6870(16)30124-7/sref55

Vibrotactile perception assessment for a haptic interface on an antigravity suit1. Introduction2. Methods2.1. Participants2.2. Apparatus and materials2.2.1. Vibrotactile actuator2.2.2. Flight simulator2.2.3. Anti-G suit

2.3. Experimental design2.4. Procedures2.5. Statistical analysis

3. Results3.1. The absolute threshold at each installation site3.2. The moderate intensity of vibrotactile stimuli3.3. The subjective assessments of vibrotactile stimuli

4. Discussion5. ConclusionReferences