Interpersonal distance in VR: reactions of older adults to the presence of a virtual agent
Grzegorz Pochwatko, Barbara Karpowicz, Anna Chrzanowska, Wies?aw Kope?
IInterpersonal distance in VR: reactions of olderadults to the presence of a virtual agent
Grzegorz Pochwatko − − − ,Barbara Karpowicz − − − ,Anna Chrzanowska − − − , andWies(cid:32)law Kope´c − − − Institute of Psychology, Polish Academy of Sciences [email protected]://psych.pan.pl Polish-Japanese Academy of Information Technology [email protected]
Abstract.
The rapid development of virtual reality technology has in-creased its availability and, consequently, increased the number of itspossible applications. The interest in the new medium has grown dueto the entertainment industry (games, VR experiences and movies). Thenumber of freely available training and therapeutic applications is also in-creasing. Contrary to popular opinion, new technologies are also adoptedby older adults. Creating virtual environments tailored to the needs andcapabilities of older adults requires intense research on the behaviour ofthese participants in the most common situations, towards commonlyused elements of the virtual environment, in typical sceneries. Comfort-able immersion in a virtual environment is key to achieving the impres-sion of presence. Presence is, in turn, necessary to obtain appropriatetraining, persuasive and therapeutic effects. A virtual agent (a humanoidrepresentation of an algorithm or artificial intelligence) is often an ele-ment of the virtual environment interface. Maintaining an appropriatedistance to the agent is, therefore, a key parameter for the creator ofthe VR experience. Older (65+) participants maintain greater distancetowards an agent (a young white male) than younger ones (25-35). Itmay be caused by differences in the level of arousal, but also culturalnorms. As a consequence, VR developers are advised to use algorithmsthat maintain the agent at the appropriate distance, depending on theuser’s age.
Keywords: virtual reality · embodied virtual agents · older adults. Virtual environments are a new communication medium. Interactions in VRhave become particularly common in recent years when the quality of sound andgraphics offered by modern devices has come close to natural (see [2], [3]). Nowa-days, they are not just an amusing novelty. There are more and more serious a r X i v : . [ c s . H C ] J a n G. Pochwatko et al. applications and virtual experiences emerging. Some of them rely on interper-sonal mediated communication, but, thanks to advanced algorithms, naturallanguage processing and machine learning, an increasing number of apps in-corporate embodied virtual agents. Users got used to intelligent virtual agents.Google Assistant or Siri is becoming a part of our everyday life, and the popu-larity of Alexa and Google Home devices is increasing (to name just a few mostpopular ones). But embodied agents are something more than that. People havea tendency to treat computers as if they were other people. Reeves and Nass [11]show that users treat computers in a social and natural way. Giving them a bodythat responds naturally to the environment by mimicking human-like movement,reacting to the user’s presence, following with the gaze and maintaining socialdistance, makes this tendency even stronger. One doesn’t have to imagine theghost in the machine any more and suspend his/her disbelief. It’s (almost) notan illusion of copresence any more as well. Instead, we have (almost) genuinesocial interaction. One of the recent applications is shown by Kim et al. [6] whoproposed giving intelligent assistants a human body with the use of augmentedreality. According to the expectations, contact with embodied agents resulted inhigher confidence, trust and feeling of social presence.Serious games, training and therapy applications are increasingly used inplace of traditional methods. They are cheaper, sometimes more effective, theyoffer users almost unlimited access, especially where it is difficult for any reason.It is particularly important in times of pandemic when traditional services arelimited and safer, remote methods are preferred. This refers primarily to olderadults, who are highly vulnerable. However, serious games must constantly evolveto adapt to the changing conditions and needs of new user groups [4].
Older adults are the group in which the increase in interest in VR and experiencewith the use of various applications is the largest (according to the Global WebIndex from 0.5% in 2016, an increase to 6% in 2018). The global virtual real-ity software market is predicted to double in size till 2023. Simultaneously, thepercentage of older adults in the European population (according to the UnitedNations Population Division) will exceed 25%. Similarly, the World Health Or-ganization is predicting the dawn of the super-aged society. We will face thepossible decline in both physical and mental health, which may lead to the de-terioration of quality of life and well being of older adults. Can VR reduce thosenegative effects and assure active ageing?At the same time, older adults can gain a lot from using applications de-veloped in new technologies. This is not only about entertainment, which hasdominated recently, but also about improving mental well-being and generalfunctioning (e.g. [5], [10], [1]). It seems we have a pretty good foundation forthis. According to Syed-Abdul et al. [13], older adults have positive attitudestowards the use of VR as a supportive technology. Perceived ease of use is high,which makes the experience enjoyable and regraded as useful. These are the main lder adults’ responses to the virtual agent 3 factors influencing the intention to use. Complying to social norms, on the otherhand, is less important.
Objective
One of the critical challenges is to develop guidelines for creatingeffective interfaces in virtual environments targeted at older adults. An elementof the VR interface can be a virtual agent, i.e. a humanoid representation ofan application or artificial intelligence algorithms. An important task is to findout if there are any differences between younger and older adults in interactingwith the virtual agency. An important parameter is a distance to the virtualagent. Maintaining a comfortable distance to the user may translate into theeffectiveness of a training or therapeutic application. In the world of physicalreality, interpersonal distance is an essential feature of individuals’ social be-haviour in relation to their physical environment and social interactions (for areview see [12]). The same is true of virtual reality [2].
Participants were invited to take part in the study conducted in the Virtual Re-ality and Psychophysiology Lab of the Institute of Psychology, Polish Academyof Sciences. Older adults were recruited from among the participants of VRworkshops organised by the KOBO Association with XR Lab, Polish-JapaneseAcademy of Information Technology (PJAIT) and Emotion Cognition Lab ofSWPS University. The VR workshops were organised as a part of follow-upstudies of previous research on VR experience with older adults by Living LabKobo and HASE research initiative (Human Aspects in Science and Engineer-ing) [9], based on previous intergenerational participatory studies by Living LabPJAIT [7, 8]. Six older adults, aged 65+, were in the experimental group. Therewere three women and three men among them. The control group consisted of6 people aged 18-35 (4 women and 2 men), volunteers from VRLab IP PANparticipants’ panel. The results of 12 people were analysed. All participants hada normal or corrected vision. All participants in the experimental group expe-rienced immersive VR for the first time. The participants in the control groupwere selected to meet the same condition. People in the experimental and con-trol groups were recruited in such a way that they differed only in belonging toa specific age group. The reason for this was that according to Sorokowska etal. [12], gender and age are the only personal characteristics associated with in-terpersonal distance. The third feature, closing the list of significant predictors,is the air temperature (which in our case, in laboratory conditions, was kept ata constant level of 22 degrees Celsius).
Lenovo ThinkSTation P510 workstation with Intel (R) Xeon (R) E5-1630 3.70GHz processor with 32 GB RAM, with Nvidia GeForce GTX 1080 Ti graph-
G. Pochwatko et al. ics card, and 64-bit Windows 10 Pro operating system was used. The virtualenvironment was prepared in the Worldviz Vizard Virtual Reality software, dis-played using HMD HTC Vive with AudioStrap (AMOLED screen, 2160x1200resolution, i.e. 1080x1200 per eye, 90Hz refresh rate, 110 degrees field of viewwidth). It is worth noting that using a worse or much better display could affectthe results.
The participants performed initial activities aimed at getting used to the virtualenvironment and the way of moving around. Display parameters, IPD, sharpness,HMD mount and controllers were also adjusted.
Fig. 1.
The view of the immersive virtual environment that participants saw (topleft: initial view, introduction and practice; top right: the beginning of the procedure;bottom left and right: close-ups of an agent.
The participants were immersed in a virtual environment which was a map-ping of a room corresponding to the dimensions of the laboratory’s physical lder adults’ responses to the virtual agent 5 space. We used the standard HTC VIVE play area: 3.5 by 3.5 meters. The phys-ical space was slightly larger to ensure the safety of the participants. On each sideof the play area, we have left at least 1.5 meters of space to the nearest barrier.One virtual agent, a white young man, was placed in the room. The agent didnot interact with the participants. We used the so-called idle pose animation,in which the character stands still and performs only small natural movements.More specifically, it is a cyclical animation in which the character stands relaxed,you can see breathing movements, slight movements of the head, arms, handsalong the body, the center of gravity is shifted to the left or right side. The targettask of the participants was to approach the agent and read the inscriptions onbadges placed on his clothes, on the front and the back (cf. Bailenson et al. [2]).The lettering on the agent’s chest uses a font 3 cm high so that it cannot be readfrom the starting position. The inscription on the back uses an 8 cm high font.In both cases, white letters with black borders were used. The position of theparticipants’ heads and hands was measured during the task (with a frequencyof 90 Hz). It should be noted that traditional research on social distancing differsin approaches and methods from research using immersive VR. In the traditionalstudy cited above, conducted in 42 countries, participants marked the declareddistance towards a stranger, acquaintance and a close person on a paper diagramshowing schematic drawings of two human figures, representing the participantand another person. This approach is convenient for mass intercultural research.In a virtual environment, the participant (in the case of this study) or the agentphysically moves through space and adopts convenient positions. In our opin-ion, this type of measurement is more reliable and accurate. It’s worth notingthat there are studies on social distance, also carried out in immersive VR, thatcontain elements such as the movement of an agent in the space of a motion-less participant or instructions for the participant to take specific positions inspace of an agent. Undoubtedly they may cause discomfort to the participants,which can be registered objectively, in the form of physiological response (HRand skin conductance changes resulting from activation of the sympathetic ner-vous system) or declaratively. For ethical reasons, we consciously resigned fromprocedures that could cause discomfort to participants.
The research was preliminary, and therefore a small group of volunteers tookpart. In this situation, it can be assumed that the distribution of results willdiffer from normal. The Shapiro-Wilk distribution tests confirmed it for thegroup of seniors (for body distance: W older adults = .76, p < .05; W control = .91, p= .41; for hand distance: W older adults = .90, p = .36; W control = .86, p = .18).Failure to meet the assumption of normal distribution forces the application ofthe non-parametric Wilcoxon test for two independent samples.At first glance, the manner in which the older participants and the con-trol group move in the space differs. The older adults move more slowly and G. Pochwatko et al. cautiously than younger people from the control group. Their relative positiontowards the embodied virtual agent also differs (see Fig. 2).
Fig. 2.
Position towards embodied virtual agent (m).
The Wilcoxon test for independent samples revealed that older adults keep agreater distance from the virtual agent than younger people in the control group(0.54 m vs 0.38 m). The Wilcoxon test for two independent samples was foundto be significant (W = 2, p < .05; see Fig. 3).The same relationship occurred in the case of the distance between the handsand the embodied virtual agent. The Wilcoxon test for independent samplesrevealed that older adults keep the hands further from the virtual agent thanyounger people in the control group. The Wilcoxon test for two samples wasfound to be significant (W = 3, p < .05; see Fig. 4). An analysis of the position lder adults’ responses to the virtual agent 7 Fig. 3.
Mean minimal body distance (m) towards embodied virtual agent in olderadults and control group.
Fig. 4.
Mean minimal hand distance (m) towards embodied virtual agent in olderadults and control group. G. Pochwatko et al. of the hands in relation to the body showed that the participants did not try tostretch them towards the virtual agent. There were also no attempts to touchthe agent’s virtual body in any of the groups.
Differences in the pace and manner of movement of the older and younger partic-ipants can be explained in two ways: first, in the case of older adults, there maybe deficits in mobility and the sense of balance. Secondly, the lower speed and no-ticeable caution of movements could also result from the general less familiarityof the older participants with virtual environments. However, it should be notedthat both the older adults and those in the control group experienced virtualreality for the first time in their lives. It was possible to repeat and extend theresults of Bailenson et al. [2]. As for the interpersonal distance maintained, theparticipants treated the virtual agent in the same way that strangers are usuallytreated in a public place. Older participants avoided entering the agent’s inti-mate space ( <
45 cm) and kept the average distance above 50 cm. The youngerparticipants maintained a significantly smaller distance; they slightly entered theagent’s intimate space (38 cm on average). Hand position analysis showed thatparticipants from both groups did not try to touch the agent. This may indicatethat the virtual agent is treated like another living person in the same room.However, an alternative explanation cannot be ruled out. In case of negativeemotions towards the virtual agent, it could be an aversive reaction. Differencesbetween younger and older participants in the distance kept from the virtualagent may have a cultural background [12]. Another possible explanation is thatyounger people have greater experience with virtual environments (such as videogames) and virtual agents (such as non-player characters - NPCs), which maylead to them to treat the agent in the simulation in a more objective manner.Maybe the older adults are more under the illusion of social presence (or cop-resence). However, this requires more systematic research and larger trials.
There are many VR applications on the market that are designed for older adults(although most often not only about them). Among them, there are relaxationapplications that allow one to regain peace and move into a pleasant, relaxingspace, even when it is impossible in the real world. There are also virtual socialenvironments that allow you to stay in touch with family and friends, and spendtime together. A special place among them is occupied by applications aimedat residents of retirement homes or other such communities, thanks to whichmembers can jointly participate in specially prepared simulations. Thanks tosome, you can explore the world and watch various exciting places both in theform of spherical photos and 360 films, and even travel back in time and explorethe same places at different times in history. The increasing percentage of olderadults in the population, as well as their noticeably greater activity, will raise lder adults’ responses to the virtual agent 9 the interest in this type of solutions. The creators of VR applications shouldtake into account the preferences of users if they want to provide them with acomfortable experience of their simulations. In applications whose interface is anembodied virtual agent, care should be taken to increase the distance towards theolder user, as well as to avoid gestures that violate the intimate zone and virtualtouch, if it is not accepted in a given culture. For younger users, it is possible toshorten the distance. The use of VR applications by older adults requires furtherin-depth research on much larger trials. It is worth taking a closer look at theissues of reaction to facial expressions, and pantomimics of a virtual agent, gazetracking, voice communication, issuing and listening to commands, compliancewithout pressure. An interesting development of the described research is alsothe manipulation of gender and age of a virtual agent.
We would like to thank our participants, older adults from our Living Lab, thoseaffiliated with Kobo Association and volunteers from VRLab IP PAN panel whoparticipated in this study. We would also like to thank all transdisciplinaryexperts involved with the HASE research initiative (Human Aspects in Scienceand Engineering).
References
1. Amorim, J.S.C.d., Leite, R.C., Brizola, R., Yonamine, C.Y.: Virtual reality therapyfor rehabilitation of balance in the elderly: a systematic review and meta-analysis.Advances in Rheumatology (2018)2. Bailenson, J.N., Blascovich, J., Beall, A.C., Loomis, J.M.: Interpersonal distance inimmersive virtual environments. Personality and social psychology bulletin (7),819–833 (2003)3. DeFanti, T.: Better than being there: Next millennium networks. IEEE ComputerGraphics and Applications (1), 60–61 (2000)4. Fleming, T.M., Bavin, L., Stasiak, K., Hermansson-Webb, E., Merry, S.N., Cheek,C., Lucassen, M., Lau, H.M., Pollmuller, B., Hetrick, S.: Serious games and gam-ification for mental health: current status and promising directions. Frontiers inpsychiatry , 215 (2017)5. Garcia-Betances, R.I., Jim´enez-Mixco, V., Arredondo, M.T., Cabrera-Umpi´errez,M.F.: Using virtual reality for cognitive training of the elderly. American Journalof Alzheimer’s Disease & Other Dementias ® (1), 49–54 (2015)6. Kim, K., Boelling, L., Haesler, S., Bailenson, J., Bruder, G., Welch, G.F.: Doesa digital assistant need a body? the influence of visual embodiment and socialbehavior on the perception of intelligent virtual agents in ar. In: 2018 IEEE In-ternational Symposium on Mixed and Augmented Reality (ISMAR). pp. 105–114.IEEE (2018)7. Kope´c, W., Balcerzak, B., Nielek, R., Kowalik, G., Wierzbicki, A., Casati, F.: Olderadults and hackathons: A qualitative study. Empirical Software Engineering (4),1895–1930 (2018). https://doi.org/10.1007/s10664-017-9565-60 G. Pochwatko et al.8. Kope´c, W., Skorupska, K., Jaskulska, A., Abramczuk, K., Nielek, R., Wierzbicki,A.: Livinglab pjait: Towards better urban participation of seniors. In: Proceedingsof the International Conference on Web Intelligence. pp. 1085–1092. WI ’17, ACM,New York, NY, USA (2017). https://doi.org/10.1145/3106426.31090409. Kope´c, W., Wichrowski, M., Kalinowski, K., Jaskulska, A., Skorupska, K.,Cnotkowski, D., Tyszka, J., Popieluch, A., Voitenkova, A., Mas(cid:32)lyk, R., Gago, P.,Krzywicki, M., Kornacka, M., Biele, C., Kobyli´nski, P., Kowalski, J., Abramczuk,K., Zdrodowska, A., Pochwatko, G., Mo˙zaryn, J., Marasek, K.: Vr with older adults:Participatory design of a virtual atm training simulation. IFAC-PapersOnLine (19), 277–281 (2019). https://doi.org/10.1016/j.ifacol.2019.12.11010. Korn, O., Buchweitz, L., Rees, A., Bieber, G., Werner, C., Hauer, K.: Using aug-mented reality and gamification to empower rehabilitation activities and elderlypersons. a study applying design thinking. In: International Conference on AppliedHuman Factors and Ergonomics. pp. 219–229. Springer (2018)11. Reeves, B.N., Nass, C.: C.(1996) the media equation: How people treat computers.Television, And New Media Like Real People And Places. Cambridge UniversityPress, New York, Ny, Usa12. Sorokowska, A., Sorokowski, P., Hilpert, P., Cantarero, K., Frackowiak, T., Ah-madi, K., Alghraibeh, A.M., Aryeetey, R., Bertoni, A., Bettache, K., et al.: Pre-ferred interpersonal distances: a global comparison. Journal of Cross-Cultural Psy-chology (4), 577–592 (2017)13. Syed-Abdul, S., Malwade, S., Nursetyo, A.A., Sood, M., Bhatia, M., Barsasella,D., Liu, M.F., Chang, C.C., Srinivasan, K., Raja, M., et al.: Virtual reality amongthe elderly: a usefulness and acceptance study from taiwan. BMC geriatrics19