use style: paper title

0 downloads 0 Views 1MB Size Report
Point of Interest (POI). ... there is no information about what really impacts on the POI, ..... GPS in network-based barrier-free street map creation system,” in.
Personalizing Pedestrian Accessible Way-finding with mPASS Silvia Mirri, Catia Prandi, Paola Salomoni Department of Computer Science and Engineering Università di Bologna Bologna, Italy silvia.mirri, catia.prandi2, paola.salomoni@unibo.it

Abstract—This work presents users evaluation of mPASS (mobile Pervasive Accessibility Social Sensing), a system to provide citizens with personalized accessible way finding. mPASS collects data both from crowdsourcing and from crowdsensing, in order to obtain a detailed georeferenced description of the urban environment accessibility. It combines these data with a user profile, with the aim of tailoring paths and maps to users’ preferences and needs. To drive the design of our application, we assessed our first proposal through a preliminary questionnaire, showing mPASS mockups to 60 uses with disability and elderly users. On the basis of the results of such questionnaire, we developed a prototype and we tested it with a small group of pilot users. This paper presents results of both the assessments. Keywords— Mobile accessibility; accessible way-finding; personalization; users with special needs; crowdsourcing

I. INTRODUCTION Urban mobility can be defined as the ability of people to move around the city, living and interacting with the space. This ability can be strongly affected by those architectural obstacles that represent a physical impediment to the exercise of citizenship for people with special needs, as people with disabilities and elderly people. Another barrier, that prevent and discourage some citizens in moving independently in the urban space, is represented by the lack of information about the urban environment and its accessibility [1]. In fact, the availability of information about urban accessibility is very limited and, consequently, personalized accessible way-finding services based on these data cannot be effective in becoming a means for social inclusion [2]. With the aim of collecting data about urban accessibility and providing citizens with effective accessible paths and maps, we have designed and developed mPASS (mobile Pervasive Accessibility Social Sensing) [3, 4]. This system uses information provided by experts combined with data produced by sensors and data gathered via crowdsourcing by users. All these data are exploited so as to provide users with paths and maps that are tailored to their special needs and/or requirements in terms of urban mobility. Examples of users who can benefit from our system are: -

wheelchair users and users with mobility impairments: they would avoid steps and stairs or uneven surfaces on

their paths and they would appreciate ramps, curb cuts and similar facilities; -

blind users and users with low vision: they would prefer paths equipped with tactile paving and acoustic cues, avoiding obstacles and unsafe crossings;

-

elderly people: generally, they would prefer safe and comfortable paths; more specifically, they would choose paths that are personalized according to their actual abilities.

Moreover, mPASS can meet the needs of a wider variety of users, with specific requirements, such as kids coming back from school, women who prefer to avoid unsafe areas at night, parents pushing a pram or tourists carrying a heavy luggage. The mPASS system is described in details in some previous papers [3, 4, 5, 6, 7]. In particular, the system architecture and data model are described in [3], while [4] focuses on the user profiling system and on the maps personalization mechanisms. Some mPASS cases studies are presented in [7]. Issues related to differences among data coming from different sources and their trustworthiness are discussed in [5] and [6], together with mPASS trustworthiness model and its assessment. With the aim of evaluating our system, this paper illustrates a two-phase assessment we have performed and key results of tests with users are presented here. Such tests confirm the effectiveness of mPASS: users appreciated the interface and the interaction with the app and provided useful suggestions and comments that are currently at the basis of our research, as described in the following of the paper. The remainder of this paper is organized as follows. Section 2 presents main related works and compares them with our approach. Section 3 briefly presents the mPASS system, while Section 4 describes tests with users, detailing a preliminary questionnaire and a prototype evaluation we have conducted. Section 5 illustrates open issues we are working on and, finally, Section 6 concludes the paper and presents some future works. II. RELATED WORK In analyzing related work, we considered several groups of researches and applications, ranging from maps

personalization to accessible maps and navigation. In particular, we have focused on platforms and systems devoted to urban accessibility for citizen with special needs. Several projects and publications are devoted to support users in collaborating each other with the aim of improving the quality of life in their urban environments [8], gathering, storing and providing data about urban accessibility. The application Wheel Map [9] offers different tools, letting users (i) find wheelchair accessible toilets and parking spaces, and (ii) rate the accessibility of a service (for instance, services related to tourism, sport, education, etc.). Such services are geo-referenced and identified as specific type of Point of Interest (POI). The main lack of this system is that there is no information about what really impacts on the POI, which effectively affects its accessibility level, in terms of barriers and/or facilities. This means that it is not possible to compute personalized paths according to real needs of users. The authors of [10] has designed and developed a mobile app that permits users to add photos and comments related to barriers and obstacles on sidewalks. This kind of data is integrated with other different data sources, in particular data gathered by sensing with data from crowdsourcing. The proposed system provides multiple geo-referenced services and computes accessible routes. These systems and projects do not take into account official reviews done by experts. On the contrary, [11] involves users and accessibility experts working for an organization which supports people with disabilities. They can provide official reviews of indoor accessibility of POIs, which are located in Bologna (Italy). It is worth mentioning that such official reviews about accessibility are not geo referenced and not structured. These data are delivered by means of a Web site. Another project, which involves users as well as experts, is AccessibleMaps [12]. Such a project provides blind people and wheelchair users with web-based urban maps that show icons representing accessibility barriers and facilities. The authors of [13], [14] explored the potential benefits of exploiting open accessibility data. The main aim of these works is addressing the information barriers between users’ needs and accessible facilities or places. In order to obtain such result the authors integrate heterogeneous accessibility related resources and then apply the Linked Data principles to establish a linked open accessibility data repository. Many sensing apps have been developed to monitor human activities and a part of them could be effectively used to detect accessibility/pedestrian barriers (such as stairs) and facilities (such as zebra crossing). These researches present sensing architectures and algorithms studied to be used in different contexts, so they need to be adapted in order to be exploited in detecting barriers and facilities (see for example [15] and [16]). In [17], the authors (by using data obtained by a smartphone accelerometer) aim to recognize the position where a pedestrian stops and crosses a street ruled by a traffic light. Some barriers and facilities could be recognize more easily by using cooperative sensing, working on detecting movement of groups of people [18].

Some mobile applications and systems, which support urban accessibility, are devoted to compute personalized paths for citizens with special needs. They perform routing algorithms based on geo-referenced data about accessibility barriers and facilities that are usually collected by crowdsourcing. Some interesting examples are OpenTripPlanner [19] and Path 2.0 [20, 21]. These systems provide personalized routing to mobility impaired pedestrians and wheelchair users, but they fail in achieving the critical mass needed to provide effective services, due to the difficulty in collecting a large amount of trusted and updated data. In [22], the authors describe a system that use GIS and GPS to support the creation and the use of network based barrier-free street maps, using specific hardware. RouteCheckr [23] is a client/server system for collaborative multimodal annotation of geo-referenced data. It provides personalized routing to mobility impaired pedestrians thought the configuration of a user profile. Some other works are devoted to find route for elderly people [24, 25]. In particular, in [25] the authors present a barrier notification service running on cellular phones equipped with GPS sensor. III. MPASS AT A GLANCE mPASS collects georeferenced data about urban accessibility and more generally about urban characteristics, which are relevant to pedestrians. This information is used in order to provide the community with personalized pedestrian paths, on the basis of preferences and needs expressed by each user. mPASS architecture and data flow are outlined in Figure 1 above. Data are georeferenced trough OpenStreetMap and are collected from three different sources: 1.

authoritative data, by local authorities and disability organizations, including open data about the urban environment provided by municipalities;

2.

data crowdsourced by registered users, who contribute in reporting barriers and facilities both while moving and from home;

3.

data sensed by smartphones or tablets owned by the same community of users, while they are moving.

Urban GIS Open data

mPASS mPASS DB

Crowdsourcing S S

S

Profile Module Routing Module

Crowdsensing

Open Street Map Figure 1. mPASS architecture

The use of multiple data sources arises form need to compute path and maps on a complete and effective database of accessibility barriers and facilities. A partial mapping of the urban environment can induce the user to go through a (wrong) route without considering the presence of an undetected barrier. This can prevent the user from reaching his/her destination/goal compromising the effectiveness of mPASS services. For example if there is a stair on the path, a wheelchair user could be forced to stop and find an alternative route.

paths, on the basis of their specific needs and preferences. 88% claim they would afford a path longer of 30% to reach their destination if the path is tailored on to their preferences and needs, while 12% of the users would afford a personalized path longer more than 30%. In particular, 6 users declared they would afford a personalized path longer of 5%, 27 users would afford a path longer of 10%, and 20 users would afford a path longer of 30%; while 3 and 4 users claimed they would afford a path longer of 50% and more than 50% respectively. Such details as shown in the chart depicted in Figure 2.

Collected data are provided to users on the basis of a user profile, which describes her/him in terms of barriers and facilities s/he likes, dislikes or wants to avoid. The routing algorithm removes all paths that include barriers to avoid and provides 1 to 3 possible solutions, where liked and disliked barriers and facility are weighted together with the total lengths of the path. All the solutions are summarized to the user who can select the more appropriate on the basis of her/his actual preferences. A complete description of mPASS architecture can be found in [3, 4].

Being aware of the possible presence of uncertain data (that can be produced by information coming from sensing and crowdsourcing activities), 73% users would prefer affording a longer path (in order to avoid a detected barrier which is not actually present on the path), instead of meeting an undetected barrier in their path or instead of having a path with a detected facility which is not actually present.

IV. EVALUATING MPASS WITH USERS In order to evaluate our prototype, we have planned a two-phase assessment: a preliminary questionnaire with mockups showing the mPASS interface (subsection A) and a prototype evaluation conducted by some users with disabilities, testing some specific mPASS characteristics, through paths with barriers and facilities which have been mapped in advanced (subsection B). The following subsections present the obtained results. A. Preliminary questionnaire In this first phase of evaluation we have involved 60 European users (including blind and people with low vision, wheelchair users and users with physical impairments, deaf and hard of hearing users and elderly people), thanks to the engagement in our project of organizations supporting people with disabilities. We have invited these users in answering an online questionnaire, asking them some general questions about their potential interest in using our application, according to its aims and some specific characteristics; moreover, we have shown them some mockups and asked them to provide their feedbacks. The group of users involved in such a first phase were composed by 60 people (26 female, 34 male), with ages ranged from 19 to 68 (with an average value of 44). Some general questions reveal that 66% of the users exploits some assistive technologies on their smartphone (including font magnifications, speech-to-text applications, and so on). 63% of them usually exploits GPS navigation systems to get information about urban pedestrian paths in their activities (i.e. Google Maps, Ariadne GPS, etc.) and 70% of the users declare they trust systems which provide geo-referenced information on the basis of crowdsourced data (including foursquare, TripAdvisor, and so on). All the users declare their willing in exploiting a system on their mobile phone that provides personalized pedestrian

Most of the users (81%) expressed their willing in sharing their personal data and information about their preferences, including details about the routes they usually perform in their daily life (77%); only 4% of the involved users declare they would not share any data about the paths they use to go along. Almost all the users (98%) declare they would be very interested in exploiting information about public means of transport routes which can be involved in their habitual paths. Then, we have asked the users to express their interests in having some specific personalized pedestrian paths. The users chose among different choices, expressing multiple preferences. Details about the resulting preferences are reported in Table 1 and show a strong interest in having above all personalization on the basis of pedestrian safety issues (i.e. zebra crossing, traffic lights, etc.) and then in having personalization on the basis of element of urban accessibility (avoiding barriers and meeting facilities). The users declared their preferences in terms of categories of urban elements they are interested in having details along their personalized paths. The users can set multiple choice, and they showed interested above all in having details about public means of transports stations (including bus stops) and vehicles. They expressed preferences also for detailed data about sidewalks, stairs (including steps, curbs and ramps), obstructions (including trees, garbage bins and traffic signs) and crossing elements (including zebra crossing, traffic lights and audible traffic lights). While surfaces (including road pavement) and parking spaces obtained less interest. Details about these results are reported in Table 2. We have concluded this first phase, asking the users to evaluate some HTM mockups, showing information about a complete urban pedestrian path and details about a specific part of a path, warning about the presence of a barrier. Obviously, we have prepared two different sets of mockups: one set provides graphical representation of the maps, while the other one provides textual representations. The users generally appreciated the proposed interface and interaction (72% of positive feedbacks), considering them clear and easy to understand and to interact with.

Figure 2. Users' willing about affording a personalized, but longer path TABLE 1. PREFERENCES ABOUT PERSONALIZED PATHS Path type

Preferences (%)

Accessible paths

44%

Safe pedestrian paths

73%

Safe paths (avoiding unsafe areas)

37%

Lit paths

15%

Most crowded paths

27%

Less crowded paths

14%

Less polluted paths

12%

Less noisy paths

29%

TABLE 2. PREFERENCES ABOUT HAVING DETAILS ON URBAN ELEMENTS Details on urban elements

Preferences (%)

Stairs, steps, curbs and ramps

31%

Obstructions

31%

Zebra crossing and traffic lights

29%

Parks

14%

Surfaces and road pavement

19%

Sidewalks

32%

Buses and trains

34%

Public means of transport stations

37%

Some users provided suggestions to improve mPASS, commenting the presence of data considered needless or considered not so clearly understandable (10% of feedbacks in this sense) or the lack of information they would like to exploit at a certain step of the interaction with the app (16% of comments in this sense). We have exploited such detailed comments in order to build the second phase of the assessment, as described in the following subsection. B. Prototype Evaluation In this second phase of tests with users, we have emulated the mPASS operation in the city of Cesena (Italy), for a set of users with different disabilities. In particular, we have involved 3 blind people, 3 wheelchair users and 4 elderly

people, equipped with their mobile devices, mounting different versions of Android, the assistive technologies they usually exploit (whenever necessary) and our mPASS prototype. We have mapped in advance three different urban paths, with the same starting point (the building hosting the degree course in Computer Science and Engineering, University of Bologna) and the same destination (the railway station). We have collected detailed information about barriers and facilities along such paths, to adequately populate the mPASS DB, and we have prepared suitable profiles together with the users, according to their preferences and needs. Then the users have exploited the mPASS prototype to reach the railway station from the University, going along the proposed personalized path, applying the Thinking Aloud method. We have followed them during the trials and we have collected their comments and suggestions. Moreover, they have filled a post-test questionnaire. Figure 3 and 4 show two screenshots of the graphical representation of a whole path and of a specific detail of another path. The users involved in this second phase appreciated the mPASS prototype; all of them declare that the interface and the interaction mechanisms are clear and easy to be used. 90% of the users found that all the provided information (shown in Figure 3 and 4) are clear to be understand and all the functions are easy to interact with. 100% of the users declare that all the information (as shown in Figure 3 and 4) are useful, while 50% of the users declare that they would appreciate additional data about public means of transport and bus stops and stations eventually involved in the proposed path. Moreover, 60% of the users are interested in some more details about the path, in particular more significant comments and suggestions are related to add: -

More information about pavements and surfaces. In particular, such details should include the position and the dimension of the uneven road pavements aPOI (shown in Figure 6), information about the feasibility of that aPOI, clarifying if it is necessary to bypass it or quantifying the effort in crossing it. Features that allow users to upload and share pictures of aPOIs would be really appreciated and would support the users in being better aware of the aPOI.

-

Information about the presence of one-ways streets involved in the path and the directions of the vehicles in such streets.

-

Information about the presence in the paths of landmarks that can help users in orienting (i.e. schools, stores, café and so on) and of squares, open spaces and roundabouts.

-

Features that let users provide comments and that let users exploit comments coming from users with a similar profile who have already gone along the same path, meeting the same aPOIs. Such comments could be related to the whole path and/or to the single aPOIs.

-

Information about the presence of works in progress and road works.

-

Details about the availability of rest areas (such as benches) along the paths and in their nearby.

-

Information about Wi-Fi areas available along the paths.

All these comments are at the basis of the current phase of our research related to mPASS. They convinced us about the need of enlarging our data model and confirmed the need of a strong personalization mechanism based on a dense and granular database of information related to the urban environments. V. DISCUSSION As witnessed by the questionnaire and by the prototype evaluation results, quality of data (in terms of trustworthiness and credibility) and quantity of data (in terms of data density on the maps), are key features in providing effective customized paths. In order to improve the quantity of data, there is the need of extending the mapping and the possible personalization, with the aim of involving more groups of users, so as to increase not only the users who can benefit by personalized paths and maps, but even the set of collected data thanks to crowdsourcing and sensing activities [3].

Path 1 Distance: 650 meters Time: 15 minutes Barriers/Facilities: 9

7

3

Start

Path 2

(600 meters, 14 minutes)

Path 3

(700 meters, 16 minutes)

Figure 3 - Screenshot showing the graphical representation of a path on the mPASS prototype

Figure 4 - Screenshot showing the graphical representation of an aPOI on a path on the mPASS prototype.

Enlarging the community of mPASS mainly relies on two possible strategies: 

Extend the group of people who can benefit from personalized paths, e.g. offering information about safe pedestrian routes, path avoiding unsafe areas, pollution or noise or other information that is useful to a wider audience.



Use gamification approaches to push people who is not actually interested in services offered by mPASS to contribute in mapping the urban environment.

With the aim of involving more user, supporting them in providing more data, we are exploiting gamification techniques. Together with the Madeira Interactive Technologies Institute (M-ITI), at the University of Madeira, we are studying gamification strategies: we have defined some examples of games and we have conducted some usability evaluations of mockups [26, 27]. An interesting strategy could be involving local administrations, such municipality and providers of public means of transports with the aim of offering free bus tickets or free parking tickets to mPASS users who are very active and very trustworthy. Another issue emerged from the preliminary questionnaire, related to how to manage unreliable reports. Such unreliable reports can generate both false positives (due to non-existing barriers and facility detected) and false negatives (due to undetected existing barriers and facilities). Both of them can cause difficulties to users, in particular the presence in the path of undetected barriers and/or the absence of a facility incorrectly detected have been recognized as highly critical by the users involved in our evaluations. In order to manage uncertain data, we have studied and defined a trustworthiness model [6, 7] and we are testing it by means of simulations. Some preliminary results in this sense can be found in [6, 7]. Another interesting issue arose from the evaluation: the inclusion (in the route module and in the mPASS DB) of open data about public means of transport. We are working at the integration of open data, which comes from T-Per, the Regional Bus operating company. In particular, we are integrating mPASS with another app: WhereIsMyBus. This latter one has been designed and developed with the aim of supporting citizens who travel by bus in the city, equipping them with a dedicated service. It directly interacts with official open data, providing real time information about public means of transport availability and equipment (in terms of accessibility facilities for citizens with disabilities). Hence, the final resulting whole service can equip users with personalized multimodal urban paths and can provide information related to travelling time, tailored to users’ abilities to move and to the bus real arrival time, as described in [5]. VI. CONCLUSION In this paper, we introduced mPASS, a system with the goal of equipping citizen with personalized pedestrian paths and a mapping of urban accessibility. mPASS collects data from many sources to improve the accessibility of urban environments, meeting special needs of citizen with disabilities and elderly people. Tests with users confirmed the

effectiveness of our system, as reported in the paper. A third phase evaluation has been planned, which will be based on a dense and detailed set of data we are in the process of collecting. We are now conducting further studies with the aim of: (i) defining a suitable trustworthiness model, (ii) properly applying gamification strategies (with the aim of engaging a wider audience of users) and (iii) including information about public means of transport (with the aim of equipping users with personalized multimodal urban paths). ACKNOWLEDGMENT We want to thank all the colleagues who supported us during this work: Marina Vriz (ASPHI), Franco Callegati and Aldo Campi (CIRI, University of Bologna), Stefano Ferretti (DISI, University of Bologna), Valentina Nisi (M-ITI, University of Madeira). REFERENCES [1]

C. Prandi, “Accessibility and smart data: the case study of mPASS,” in Proceedings of the 13th International Web for All Conference (W4A’14), ACM Press, 2014. [2] H. Filiz Alkan Meshur, “Accessibility for people with disabilities in urban spaces: A Case Study of Ankara, Turkey,” in International Journal of Architectural Research. Archnet-IJAR, Vol. 7, Issue 2, pp. 43-60, July 2013. [3] C. Prandi, P. Salomoni, and S. Mirri, “mPASS: Integrating People Sensing and Crowdsourcing to Map Urban Accessibility,” in Proceedings of the IEEE International Conference on Consumer Communications and Networking Conference (CCNC ’14), IEEE, 2014. [4] S. Mirri, C. Prandi, and P. Salomoni, “A Context Aware System for Personalized and Accessible Pedestrian Paths,” in Proceedings of the 2nd International Workshop on Location-based Services and Applications in Ubiquitous Computing (LSAUC 2014) - International Conference on High Performance Computing & Simulation (HPCS 2014), IEEE, 2014. [5] S. Mirri, C. Prandi, P. Salomoni, F. Callegati, and A. Campi, “On combining crowdsourcing, sensing and open data for an accessible smart city,” in Proceedings of the 3rd International Conference on Technologies and Applications for Smart Cities (I-TASC ’14), IEEE, 2014. [6] S. Mirri, C. Prandi, and P. Salomoni, “Trustworthiness assessment in mapping urban accessibility via sensing and crowdsourcing,” in Proceedings of the First International Conference on IoT in Urban Space (Urb-IOT 2014), 2014. [7] S. Ferretti, S. Mirri, C. Prandi, and P. Salomoni, “Trustworthiness in Crowd- Sensed and Sourced Georeferenced Data,” in Proceedings of the 2nd International Workshop on Crowd Assisted Sensing, Pervasive Systems and Communications (CASPer 2015) - in conjunction with IEEE PerCom 2015. [8] F. Zambonelli, “Pervasive Urban Crowdsourcing: Visions and challenges,” in Proceedings of the International Conference on Pervasive Computing and Communications Workshops, PERCOM ’11. IEEE, pp. 578-583, 2011. [9] Wheel Map. Available from: http://wheelmap.org/en/ [Retrieved: October 2015]. [10] C. Cardonha, D. Gallo, P. Avegliano, R. Herrmann, F. Koch, and S. Borger, “A crowdsourcing platform for the construction of accessibility

[11] [12] [13]

[14]

[15]

[16]

[17]

[18]

[19] [20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

maps,” in Proceedings of the 10th International Cross-Disciplinary Conference on Web Accessibility (W4A’13), ACM Press, 2013. Ingresso Libero. Available from: http://www.ingressolibero.info/ [Retrieved: October 2015]. AccessibleMaps. Available from: http://www.accessiblemaps.org/ [Retrieved: October 2015]. C. Ding, M. Wald, and G.A. Wills, “Survey of Open Accessibility Data,” in Proceedings of 13th International Web for All Conference (W4A’14), ACM Press, 2014. C. Ding, M. Wald, and G.A. Wills, “Open Accessibility Data Interlinking,” in Proceedings of 14th International Conference on Computers Helping People with Special Needs (ICCHP), pp. 73–80, 2014. S. Choi, R. LeMay, and J. Youn, “On-board processing of acceleration data for real-time activity classification,” in Proceedings of the IEEE International Conference on Consumer Communications and Networking Conference (CCNC 2013), IEEE, pp. 68-73, 2013. A. Anjum and M. U. Ilyas, “Activity recognition using smartphone sensors,” in Proceedings of the IEEE International Conference on Consumer Communications and Networking Conference (CCNC 2013), IEEE, pp. 914-919, 2013. A. Bujari, B. Licar, and C.E. Palazzi, “Movement pattern recognition through smartphone’s accelerometer,” in Proceedings of the IEEE International Conference on Consumer Communications and Networking Conference (CCNC 2012), IEEE, pp. 502-506, 2012. M.B. Kjærgaard, M. Wirz, D. Roggen, and G. Tröster, “Detecting pedestrian flocks by fusion of multi-modal sensors in mobile phones,” in Proceedings of the 2012 ACM Conference on Ubiquitous Computing (UbiComp '12), ACM Press, pp. 240-249, 2012. OpenTripPlanner, http://opentripplanner.usf.edu/ [retrieved: October, 2015] C.E. Palazzi, M. Roccetti, and G. Marfia, “Realizing the unexploited potential of games on serious challenges,” in Computers in Entertainment 8 (4), 23, 2010. C.E. Palazzi, L. Teodori, and M. Roccetti, “Path 2.0: A participatory system for the generation of accessible,” in Proceedings of the International Conference in Multimedia and Expo (ICME), IEEE, pp.1707-1711, July 2010. M. Kurihara, H. Nonaka and T. Yoshikawa, “Use of highly accurate GPS in network-based barrier-free street map creation system,” in Proceedings of IEEE International Conference on Systems, Man and Cybernetics, IEEE, pp. 1169-1173, 2004. T. Völkel and G. Weber, “RouteCheckr: personalized multicriteria routing for mobility impaired pedestrians,” in Proceedings of the 10th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS ‘08), ACM Press, pp. 185-192, 2008. T. Kawamura, K. Umezu, and A. Ohsuga, “Mobile navigation system for the elderly - preliminary experiment and evaluation,” in Ubiquitous Intelligence and Computing, Lecture Notes in Computer Science, 5061 (2008), pp. 578-590, 2008. K. Umezu, T. Kawamura, and A. Ohsuga, “Context-based Barrier Notification Service Toward Outdoor Support for the Elderly,” in International Journal of Computer Science & Information Technology (IJCSIT), 5:3, 2013. P. Salomoni, C. Prandi, M. Roccetti, V. Nisi, and N.J. Nunes, “Crowdsourcing Urban Accessibility: Some Preliminary Experiences with Results,” in Proceedings of CHIItaly, ACM Press, 2015. C. Prandi, V. Nisi, P. Salomoni, and N.J. Nunes, “From gamification to pervasive game in mapping urban accessibility,” in Proceedings of CHIItaly, ACM Press, 2015.