Authors: Ryan Younger, University of West London. John Moore, The Open University. Jose Abdelnour-Nocera, University of West London. Antonio Kheirkhahzadeh, University of West London.
Abstract: In this paper we describe CASIK, a Culturally Adaptive Sustainable Information Kiosk that has been designed for deployment within institutions that have a requirement for kiosks that can cater to users from a mixture of cultural backgrounds. We show how a ubiquitous technology such as sound, as well as innate interactions such as gestures, can be used as a method of interaction between high and low- end phones and an Internet-connected kiosk based around an Intel ‘Compute Stick’, monitor, and microphone.
Author Keywords: Screen, interaction, sound-based interface, ICT4D, smart phone, feature phone.
ACM Classification Keywords: H.5.m. Information interfaces and presentation (e.g., HCI): Miscellaneous.
Introduction: Information kiosks have become a prevalent aspect of our society, providing an on-demand method of convenient information retrieval to members of a community with minimal attendance or management required by any one individual. Most kiosks in operation today are based on either bespoke tangible buttons (such as an ATM keypad) or, more recently, touchscreen virtual controls.
Despite the prevalence and popularity of touchscreen controls a body of research has been growing within the area of alternate input methods, such as multimodal interactions, that provide a richer and more natural user experience [1],[2]. Kiosks are able to be categorised depending upon their intended purpose Borchers et al. [3] defines four different categories of kiosk; information kiosks, advertising kiosks, service kiosks and entertainment kiosks. We are primarily concerned with information kiosks and entertainment kiosks for the purposes of our research.
One interesting approach to interacting with information kiosks, that we intend to explore further with our research, is the use of a mobile cellular telephone as an input mechanism. Bergweiler et al. [4] describe Calisto, a system that enables users to connect their mobile telephones to a large public terminal and share interesting media and facts via an intuitive multimodal interaction paradigm [2].
Instead of connecting to the kiosk using wireless radio communications such as Bluetooth or 802.11x we intend to utilise sound as a transmittal mechanism, through DTMF tones, enabling a user to recall information previously submitted via a web app. Facilitating the social sharing of said information on a communal screen within a physical location. Using a pervasive technology such as sound opens our device up to a far more diverse range of users in both developed and developing countries. It also minimises the requirement for complicated setup procedures such as pairing and entering passphrases/words.
Sound as a transmittal method mandates as a minimum requirement only that a device that can store and playback a WAV file accurately with minimal sound degradation. This puts the list of potential devices that can be used to interact with the kiosk in the range of billions, from low-end legacy devices through to the latest smartphones or tablets.
The utilisation of sound as a ubiquitous transmittal medium removes the barrier of cost as a prohibitive factor in making our information kiosk accessible, however, in order to create a truly accessible global solution one must consider the usability of the interface within an international context. That is to say one must consider how usable the system would be to users from manifold cultural backgrounds, some even having cultural influences from two or more countries. Designing a bespoke interface for each Country would one approach, however, this would prove prohibitively costly and has shown to be ineffective as a canonical approach to this problem. [5]. A more holistic and sustainable approach would be a culturally adaptive user interface, which allows an end user to specify their cultural background at sign up and then provides a culturally bespoke interface based on an algorithm that uses a weighted ranking to customise the interface based on predefined criteria.
The research questions we will be aiming to answer within our research are: (1) Is the CASIK engaging for users, and what impact do the different interaction mechanisms have on the kiosks usability and performance? (2) Does adapting the kiosk to the users cultural background have an impact on the perceived usability of the kiosk?
The only work to our knowledge that has tackled this specific area is by Heimgartner (2005) [6] in the area of cultural adaptivity in navigation systems, Kamentz (2006) [7] in the area of e-learning, and Reinecke (2011) [5] in the area of adaptive user interfaces within task management software. Culturally adaptive driver information and assistance systems have been studied extensively by Heimgartner, with the results of two online studies revealing interaction behaviour differences which were dependant upon the cultural background of the user. However, in order to create these user models from which the customisation would be derived, a rules engine of some sort would need to be used. Kamentz (2006) opted to use machine learning for this task, while Reinecke (2011) developed a user-modelling engine. As scholars get a handle on the process of designing culturally adaptive user interfaces, profiling of different cultures is beginning to reveal that the mental models for user interface design of different cultural groups may have been shaped by the influence of multinational websites (Marcus, A., & Gould, E. W., 2000) [9].
There have not been any studies focussing specifically on culturally adaptive user interfaces within kiosk- based systems. Kiosks are of particular note as they are one of the most popular channels for information delivery in developing countries (Kuriyan, Ray and Kammen, 2008) [10].
System Description: Figure 1 provides a high-level overview of our Culturally Adaptive Sustainable Information Kiosk. The system is based on an Intel Compute Stick, which provides a quad-core ATOM CPU and 2GB of RAM – more than adequate for our purposes. An HDMI LED screen is connected to the Compute Stick via its HDMI port and a USB microphone is connected to receive the DTMF sound tones. An HP Leap Motion keyboard provides traditional keyboard input and the ability to input hand gestures. A persistent Internet connection is provided to the Compute Stick via it’s Ethernet port, which will be connected to the academic JANET LAN for the purposes of our testing.
The basis of our system is that of an information retrieval and consumption kiosk, the screens being used are situated in a communal area and are highly visible to students walking through the University or stopping for refreshments or a meal. We have developed a web application that allows a user to add images via a ‘creation interface’. An end user will need to register with the website initially, providing details of their cultural background when registering, and can then add a new post which has it’s own unique URL. When adding content to the system from their own computer, the end users interface will be tailored to their specific cultural background through the use of a user-modelling engine. The user-modelling engine is, in essence, an algorithm that assigns values based on the users cultural background selections provided at registration (as shown in figure 2) and then uses this weighted score to derive the required interface modifications based on adaptation rules derived from work on the effect of Hofstede’s cultural dimensions on UI design. A western version of the posting dialog is displayed in figure 3.
End users access this creation interface via a public website (http://www.uwlkiosk.com). Having a centralised website allows us to monitor registrations remotely as well as various website metrics we have configured. Were there any abuse of this system we could block offending usernames and IP addresses and we can also set manual approval of images to prevent abuse. Users will initially sign up by scanning a QR code
on the kiosk booth with their smartphone or by manually visiting the short URL listed next to the QR code on a device of their choosing. The web application is responsive so should display correctly across a variety of different screen sizes and configurations.
The sign up page will ask them for basic user information as well as which countries they have lived in before and how long they lived in each of those places. When the user visits the homepage this information is used to customise the UI based on the users cultural dimensions.
There will be two versions of our interface: a non- adaptive version and a culturally adaptive version. The version a user will see will be counter-balanced using an algorithm that shows one of the two interfaces to a user on an alternating basis and tracks what percentage of each version of the interface has been viewed by the current total users. The user will however be able to switch back to the interface they prefer should they wish by using a menu on the left and this will be recorded as a preference metric.
We will gather data in two ways: Firstly we will have useful quantitative data in the form of web application metrics such as the clickstream (where the user clicked on screen), page bounces (where the user exited the app), time spent on each page, number of clicks between the two versions, any errors that occurred etc. We will also provide a prominent feedback button that the user can use to submit any feedback. Secondly, we will gather subjective data from users by conducting a questionnaire with a representative sample of our subjects using the 7-point Likert scales, with 7 being “I don’t agree at all” and 1 being “I am in complete agreement”. There will also be observation of the users at regular intervals, as the author of this paper will be visiting the booths regularly to witness their usage.
This data will be useful in allowing us to assess whether subjects preferred the aesthetics and usability of our culturally adaptive user interface and, through our quantitative metrics, whether their task performance was improved.
Introducing an open-ended creation/consumption interaction paradigm is intentional as we hope to see interesting uses of the system evolve without constraining the users workflow or application domain. Reinecke (2011) identified that user generated content was most useful when assessing the effectiveness of a culturally adaptive user interface as it is free from cultural biases.
Conclusion: In this paper we have described CASIK, a Culturally Adaptive Sustainable Information Kiosk that adapts it’s post creation interface to a users cultural background and then allows the consumption and recall of content in a novel manner through the use of sound file transmittal or gestural interactions. The purpose of the project is to build a low-cost, sustainable, adaptable, and culturally inclusive kiosk solution that can be employed within a range of different use cases globally with the aim of harnessing ubiquitous technologies to facilitate rich interaction across cultures within communal physical social spaces. Our sustainable solution aims to provide an inclusive kiosk design that is accessible by both the information rich and information poor and prevents the exclusion of a particular group due to either economic or social factors.
References:
[1] Bergweiler, S., Deru, M. & Porta, D. 2010, “Integrating a multitouch kiosk system with mobile devices and multimodal interaction”, ACM, , pp. 245. [2] Karpov, A.A. & Ronzhin, A.L. 2009, “Information enquiry kiosk with multimodal user interface”, Pattern Recognition and Image Analysis, vol. 19, no. 3, pp. 546.
[3] Borchers, J., Deussen, O., & Knörzer, C. (1995). Getting it across: layout issues for kiosk systems. ACM SIGCHI Bulletin, 27(4), 68-74.
[4] Bergweiler, S., Deru, M., and Porta, D. Integrating a Multitouch Kiosk System with Mobile Devices and Multimodal Interaction. In ACM International Conference on Interactive Tabletops and Surfaces, ITS ’10, ACM (New York, NY, USA, 2010), 245–246.
[5] Reinecke, K., & Bernstein, A. (2011). Improving performance, perceived usability, and aesthetics with culturally adaptive user interfaces. ACM Transactions on Computer-Human Interaction (TOCHI), 18(2), 8.
[6] Heimgärtner, R. (2007). Towards cultural adaptability in driver information and-assistance systems. In Usability and Internationalization. Global and Local User Interfaces (pp. 372-381). Springer Berlin Heidelberg.
[7] Kamentz, E. (2006). Adaptivität von hypermedialen Lernsystemen (Doctoral dissertation, Phd thesis, University of Hildesheim).
[8] Kolb, A., & Kolb, D. A. (2013). Kolb Learning Style Inventory: LSI Workbook. HayGroup.
[9] Marcus, A., & Gould, E. W. (2000). Crosscurrents: cultural dimensions and global Web user-interface design. interactions, 7(4), 32-46.
[10] Kuriyan, R., Ray, I., & Kammen, D. (2008). How to use technology to spur development. Issues in Science & Technology, 24(2), 9.
[11] Moore, J., Younger, R., and Abdelnour-Nocera, J. WP-117: A tone driven offline information kiosk. In ACM International Conference on Interactive Experiences for Television and Online Video, TVX 2014, ACM (Apr. 2014).