Figure 1: Theprototype of the multimodal assistive device designed by Sarfraz et al. (2017)and worn by blind individuals.
Thisprototype is an example of the functional design goals this report proposed asit satisfies the criteria of providing the essential visual information thatblind individuals require during social interaction. This is evident as thetechnology offers features for three of the unmet needs of blind individualsidentified by Krishna et al. (2008). Specifically, knowing the identity (‘ID’function) and location (both functions) of individuals in the room and knowingwhere a person is directing their attention (‘GAZE’ function). However, thisdesign could have incorporated more features to serve more needs (Krishna etal.
, 2008). For example, as there are already algorithms developed whichidentify facial expressions (Panchanathan, Chakraborty & McDaniel, 2016),such algorithms could also be integrated into the computer element of thedesign and then transferred to the user via audio feedback. Such visualinformation will enable users to infer emotional states of individuals, whichis vital for improved social interactions (Horstmann, 2003). However, although is it possible with currentcomputer vision research to incorporate such features to assist additionalneeds (Krishna et al., 2008), designers must be aware of the user’scapabilities, not just their disabilities.
For example, blind individuals havebeen found to be as accurate as sighted individuals in detecting specificemotions (i.e., happiness or sadness) from vocalisations (Gamond, Vecchi, Ferrari, Merabet& Cattaneo, 2017).
Therefore, as emotional states can be inferred from other sensoryinformation, visual information is not necessarily required to develop thevital emotional recognition skills in improved social interaction. Thus, byadding a facial expression feature to Sarfraz et al.’s (2017) design, withadditional audio feedback, it could interfere with the user normal abilities toprocess audio information (i.e.
, the valance of vocalisations). Therefore,designers of technology to support social interactions must incorporatefeatures that provide theessential visual information users require, but must be cautious as too muchinformation could have a detrimental effect on their currentcapabilities. Furthermore, testing the prototype on blindindividuals in a normal interaction scenario determines whether the devicestands up to the cognitive demand and usability criteria proposed. Sarfraz etal. (2017) tested their device on 12 blind individuals while they took part ingroup discussions with sighted individuals. Blind individuals conversed withthe group not using the device, using the device with only audio feedback andthen using the device audio-haptic feedback. Cognitive load and usability wereassessed via specific questionnaires after each condition. Results showed thatcognitive load was significantly greater when users received audio-hapticfeedback than when then received just audio feedback.
Haptic feedback possiblyincreased the cognitive load due to users’ having to remember, almost like anew language, what the vibrations meant (Adebiyi et al., 2017). Specifically,the magnitude of vibration and location of vibration on the belt infers thedistance and location of an individual respectfully in relation to the user.Therefore, a recommendation for designers is to just include one feedbackmodality (i.e., audio) as an additional feedback modality (i.
e., haptic) isintroducing an extra level of mental processing, while only providingcomplementary information. However,it might not be fitting for future designs of such assistive technology to onlyhave the option of audio feedback. Audio feedback can potentially affect ausers’ ability to process other audio information in the environment (Velázquez, 2010). This isobserved in the Sarfraz et al.’s (2017) evaluation of the multimodal prototypewhereby two users mentioned that the number of people looking at them wasannounced at the same time a question was directed at them and in turn, theymissed part of the question. In this situation perhaps haptic feedback wouldhave been more beneficial for the users. Thus, as receiving both audio andhaptic feedback significantly increases cognitive load but there are situationswhere specific feedback is more suitable, a recommendation for future designsis that they incorporate a feature whereby users can choose which feedback theyreceive, based on their current situation.
Blind individuals found both device variants(just audio feedback and audio-haptic feedback) easy to use, meeting theusability design criterion. In addition, all blind individuals said that usingthe device enhanced their participation in the discussion in comparison to notusing the device, with eight favouring audio-haptic feedback and four favouringjust audio feedback. Therefore, although audio-haptic feedback had asignificantly greater cognitive load than just audio feedback, preference andusability scores suggest that participants are able to manage the extra mentalprocessing. Furthermore, these scores have strong external validity, as thestudy involved a normal social interaction. Therefore, future designsincorporating the recommended feature of feedback choice also means users havethe option of choosing their preferred feedback modality, whether one or both.In turn, technology would be able to support individual differences inenvironment and feedback preference. Furthermore, it appears as if the prototypemeets the aesthetic design criterion. It is portable and comprises of wirelesscomponents whereby the belt can be subtly worn under clothes.
However, most ofthe blind individuals in Sarfraz et al.’s (2017) evaluation study felt that thedevice was too big and that the straps for the camera were too noticeable.Therefore, future designers should focus on reducing the amount of hardware,for example, incorporating vibrators into the camera straps would mean the beltis not required. Alternatively, designers could focus on trying toaesthetically improve the hardware itself. This is a current trend inprosthetic design, whereby highly stylised prosthetic limbs are created as arepresentation of the wearer’s personality (Profita, 2016).
However, such designs, althoughaesthetically pleasing, will not subtle and will draw the undesired attentionthat has been found to create social barriers and result in abandonment ofblind individuals using assistive devices (Pape et al., 2002). Consequently, a recommendation for technology aiming to support socialinteraction would be to focus on restructuring the hardware design to make itless noticeable and bulky. Furthermore, although research into the users’perception of the device is important for the functional and usability designcriteria, crucially it is the perception of the sighted individuals that shouldbe investigated to determine whether such technology is socially acceptable.
For example, the sighted individuals in Sarfraz et al.’s (2017) study couldhave felt uncomfortable having a camera pointing at them and thus felt uneasycommunicating with the user. Therefore, future research should evaluate the sightedindividuals’ perceptions of the prototype to be able to determine aspects ofdesign that may or may not be considered socially unacceptable, fuellingfurther design recommendations. Finally, for any of the design recommendationsfrom this report to be able to overcome the challenges blind individuals facein social interaction, misperceptions and social stigma associated withassistive devices must be dealt with. Undesired attention towards assistivedevices occurs as a result of sighted individuals being unaccustomed with them(Shinohara & Wobbock, 2011). As a result, assumptions are inaccuratelydrawn, namely that technology removes one’s disability and without it, disabledindividuals are completely unable (Shinohara, 2012).
It is not the users’ role to deal with thesemisperceptions by explaining the device, it is society’s role to place moreimportance on trying to understand it. Furthermore, as mentioned above, assistive device users are aware thatthe technology is a symbol of difference and highlights disability (Shinohara & Wobbrock, 2011).Such social stigmas are extremely discrediting for the blind population(Goffman, 1963).
Therefore, although aesthetic design considerations, forinstance trying to make the device more subtle, is a step in the rightdirection to remove social barriers, ultimately it is the responsibility ofsociety to become more familiar with assistive technology and change people’sattitudes towards them. To conclude, this report has proposed importantdesign criteria must be met for assistive technology to overcome the challengesblind individuals face in social interaction. The report firstly identified thevital design criteria; incorporating functional features with the users’ needsat the forefront of design, being easy to use, not too cognitively demandingthat it effects blind individuals’ normal capabilities and that it is subtleand socially acceptable.
The report then evaluated how a current prototypedevice that has been designed to support social interaction stands up to theproposed criteria. Although it is aligned with some of the criteria, theevaluation has enabled further design recommendations for other such socialinteraction devices. Specifically, being cautious of not providing too muchvisual information to a user, incorporating a feedback choice feature, for bothuser’s preference, as well as the situation they are in and making the designmore subtle by restructuring the hardware. Furthermore, understanding sightedindividuals’ perceptions of the prototype are vital to creating further designrecommendations. Ultimately, with the appropriate design, as well as societybeing more familiar with assistive devices, assistive technology to supportsocial interaction should be brought more closely to the market to help theblind population in what is such an essential part of humanity.