A Techies Guide for Monitoring Urban Accessibility in Smart Cities

Abstract

Keywords: Smart City, RFID, smart sensor network, urban planning, SOA, Internet of Things, urban accessibility

1. Introduction

The continuous development of Information and Communication Technology (ICT) has a high impact in different fields of human society, contributing to development through the proposal of new approaches that solve existing problems, such as human-computer interaction, social development, sustainability, and city planning [1]. The growing body of knowledge regarding ICT allows us to propose solutions that exploit the full potential of technological advances. This fact enables organizations to manage social, industrial and business processes in a different manner to increase efficiency and user satisfaction.

Currently, some of these proposals can be seen through emerging concepts such as Smart Cities, Sustainable Cities and Smart Grids, where it is possible to deploy ubiquitous sensor networks for acquiring distributed data and generating valuable information used to improve citizens’ quality of life. In this way, the new ICT innovations drive us towards a Smart World concept as a global framework for the collaboration between these smart systems, where different stakeholders will exploit the information provided to achieve their specific purposes according to Smart World’s goals [2].

This idea has been increasing in significance in political agendas and public services. The most representative example is the intensive promotion of Smart City initiatives [3,4]. Cities are complex systems typified by massive numbers of interconnected citizens and companies and numerous modes of transport and communication networks, including services and utilities. The Smart City concept helps municipalities face new challenges due to population growth, such as resource consumption, urban design and design of communication routes [5,6]. They can be addressed with solutions that leverage the full potential that technological development offers. Several major transitions in technology—each one of them important in its own scope—are combining to make the Internet of Everything possible. These proposals use new computing, sensing and telecommunication technologies to provide knowledge and intelligence to the city. In general, the goal is to make better use of resources. The convergence and maturity experienced by previous technologies allow for deployment of applications in the Smart City context to generate knowledge for the management of urban processes. Additionally, they improve efficiency in the design of policies for sustainable resource management and public accessibility [7,8].

Despite the growth of this field, there are still many issues to resolve. One of these issues is focused on achieving Accessible Cities. Accessibility is an element of life quality that is of universal interest, a right of all citizens, a determining factor of the liveability of cities, and an essential element in modern society. It provides security, comfort and autonomy to pedestrians, cyclists, public transport and private motor vehicle users, in its rational use. It is respectful of the urban fabric and should be approached from the many aspects that affect accessibility. In this case, our research focuses on the field of pedestrian mobility, particularly in the most vulnerable group of people with disability. It deals with the study of urban accessibility—urban streets and urban public open spaces—as a means to enhance their participation and contribution in all aspects of social life.

Unfortunately, some spaces of the public built environment are not accessible enough. Pedestrian crossings with curb ramps on only one end, excessively narrow sidewalks and those occupied by street furniture, unsuitable slopes, lack of rest areas over excessively long distances or streets to ensure a leisurely pace, and a lack of markers or clear organization of different traffic flows are just some examples. Its design does not take into account the requirements of people with mobility difficulties and other physical or sensory limitations (of understanding, communication or perception). Much remains to be done to achieve barrier-free areas [9].

In addition, one of the main current problems is the difficulty of knowing what the accessibility deficiencies of cities are. In many cases, regulations are only applied, without checking if they are effective or if there is a degradation that makes the results unusable after a short period of time. For example, consider the construction of a ramp on a street sidewalk that is followed by the installation of an element of street furniture that prevents its use. There is no ongoing analysis over time that confirms that actions are effective or that identifies if they have degraded. The current techniques are focused on conducting biased manual surveys, in very specific environments and time periods, with high cost and without guarantees that these actions are an improvement for the citizens.

For this reason, the main objective of this research is to design a model for evaluating the effectiveness of urban accessibility, improving the habitability of cities and citizens’ quality of life, and achieving equal opportunities for people with disabilities. This model provides automatic monitoring tools supported by technology to dynamically discover, assess and classify urban accessibility issues, generating value-added information regarding accessibility.

To achieve these aims, the starting hypothesis of this research is based on new technologies to yield opportunities for this dynamic and automatic analysis of urban accessibility and overcome difficulties in the proceedings of urban effectiveness evaluation.

To carry out the proposal, three key aspects must be jointly considered to monitor events that occur in the city related to accessibility issues: the advantages of sensor-based ubiquitous computing, the characteristics of wireless communication technologies, and the possibilities of cloud-based application systems. In this regard, the main novelty of this work is the integration of several strategies, paradigms, techniques, new technologies and Internet of Things solutions to build a comprehensive system that is effective in a dynamic manner.

The information generated will allow us to establish criteria for actions and urban planning concerning disability through a set of Key Accessibility Indicators (KAIs). They will be obtained by two methods: (a) by means of the inference of the patterns of movement for individuals according to their degree of disability and by comparing the routes followed in each case to conclude: for example, if people with disabilities move along the same routes as people without disabilities, or if instead they find some impediment (stairs, ramps, obstacles, etc.) along the way; (b) through the self-reporting of citizens of all accessibility deficiencies that are found anywhere when they move around the city. The remainder of the paper is organized as follows: Section 2 describes the related work on urban activities and technology used to address the analysis of the accessibility in the city; Section 3 introduces the overall system architecture for acquiring knowledge on the urban accessibility, which is explained in depth in the next sections; Section 4 explains the infrastructure needed for dynamically gathering the citizens’ locations; Section 5 describes the cloud infrastructure that supports the system; Section 6 deals with the urban accessibility information services to provide the information to stakeholders; Section 7 conducts the evaluation and discusses the results and their possibilities; and finally, Section 8 presents that study’s conclusions.

2. Related Work

Social awareness towards maintaining urban accessibility is growing in modern societies. The authorities have established abundant legislation at all administrative levels to create accessible and socially inclusive urban spaces. However, the methodologies used up to the present have served to eliminate some barriers in controlled urban environments but not to preserve the urban spaces over time. There are still many movement barriers, and few efforts are dedicated to analysing possible causes or elements that contribute to generating them.

The possibilities of new technologies provide innovative methods for monitoring and preserving the accessibility of urban areas. In this section, the existing methods for monitoring urban accessibility and the active technologies for performing it are reviewed. Recent and representative works are discussed to show the intensive research developed in this area. A final subsection is added to summarize the contributions to previous work.

3. System Architecture

Once the main challenges have been pointed out and the findings of previous studies have been obtained, the following aspects must be covered to adequately address the accessibility of cities:

• Consolidating the cross-cutting nature of the accessibility from the perspective of the activities of citizen participation, that is, establishing relationships between different areas, equipment and services, different barriers to mobility and their causes.
• Providing mechanisms for developing the assessment of the current conditions of accessibility and monitoring their effectiveness over time.
• Enabling smart scheduling of urban planning actions that aim not only for the timely removal of barriers but also attempt to prevent the occurrence of barriers in a dynamic manner.
• Including citizen participation in the city development towards accessibility models.

The achievement of these aspects represents the desirable features of the required system. The proposed computational model consists of a distributed architecture that takes advantage of the new ICT technologies in a context characterized by deployment of a wireless communication infrastructure and the cloud computing paradigm in Smart City environments.

The proposed architecture is based on the experience and recent previous research works regarding distributed IoT architectures [55] and citizens’ traceability systems performed by this research group [26,27,56,57], which improve the current track-and-trace system proposals for the acquisition of locations of citizens in heterogeneous and uncontrolled environments.

This proposal combines the advantages of the previous works and integrates some of the modules tested to extend the functionalities of the entire system in several significant ways: (1) using a combination of acquisition technologies and establishing a homogeneous layer for other communication technologies; (2) designing new movement route inference methods from discrete locations; (3) improving some components of the architecture for handling large amounts of data; (4) expanding the scope for both indoor and outdoor environments; (6) especially focusing on urban accessibility issues for citizens with disabilities; and (7) integrating citizens in the urban accessibility process through user-centric mobile applications and IoT solutions.

The architecture is made up of three main layers to contain the infrastructure components and provide the services: the Citizen Location Acquisition infrastructure, Cloud Support Infrastructure, and Urban Accessibility Information Services. The basic scheme of the new architecture is shown in . The following sections describe the design and the implementation details of each of the layers of the system.

 

Overall architecture of the System for Accessibility Monitoring in Smart Cities.

4. Citizen Location Acquisition Infrastructure

The main objective of the acquisition infrastructure is to obtain the positions of citizens when they walk through the city, both in indoor and outdoor environments. This aim has the following design constraints: (1) automation: to be capable of acquiring and processing the individual location information in an unassisted way and transparently for citizens; (2) scalable: to be able to handle a variable amount of data; (3) reliable: to ensure the delivery and reception of each location to generate the different citizen flows; (4) completeness: to offer complete information about citizen movements despite the lack of certain individual locations, and their experiences in such displacements; (5): interoperable: to be able to obtain information from different location formats or methods, as the system is intended to take advantage of the already deployed elements incorporating them for users who might work with different protocols or formats.

The selection of the localization technology takes into account the experience acquired in the review of other works on the traceability of users and the performance and the benefits they currently offer. Two complementary technologies are used in this work to illustrate the capabilities of the system: RFID and GPS. Other wireless connections can be included in this system under the same operating principles, such as Bluetooth Low Energy (Bluetooth LE—BLE) or Near Field Communication (NFC), which are also built into most mobile phones and many consumer electronics devices. In addition, the proposed system could be valid for discrete location acquisition from users by means of widespread communication methods, such as Global System for Mobile communications (GSM) and wireless local area networking (Wi-Fi).

The next figure () shows the detailed acquisition infrastructure for various communication technologies. The figure illustrates the following parts of the acquisition infrastructure: citizen components and acquisition components. These parts depend on the type of technology used in each case. In this work, RFID and GPS work together to obtain the user position. They are complementary and have some advantages over other alternatives, as described in related works, to make them suitable for the aims of this system. Obtaining positioning through GPS systems takes advantage of the wide range of phones and wireless devices that incorporate this technology already in use among the population. In this case, collaboration and user permission regarding sharing of their location are required. To facilitate this collaboration, it is necessary to guarantee the total disassociation of personal data.

 

Acquisition infrastructure deployment.

This technology is scalable to the deployment of infrastructure, with the main advantage being the non-requirement of additional elements besides the satellites. Alternately, the use of RFID technologies is continuously increasing. In addition, many devices, including smartphones, currently incorporate such technology.

4.2. Acquisition Components

The acquisition components are those responsible for reading the user positions and sending them to the cloud infrastructure. This part of the architecture has been designed to cover the set of previously defined goals that are necessary to achieve the aims of the proposal. This approach covers issues that have not been taken into account in other related and previous works. Specifically, some aspects, such as environmental heterogeneity and most of the design constraints of this work, are not considered.

There are two ways to acquire the user location: by means of: (a) some communication technology for data transmission and/or (b) GPS positioning.

The communication technology requires the installation of a network of Smart Readers to capture the users’ locations, including those areas where GPS is inaccessible. These devices are mainly responsible for reading the locations automatically, as well as ensuring the delivery of such information to a central cloud server. In this part, the IoT paradigm plays an important role in providing the capabilities to match the constraint set. The design of these devices includes communication and computing features provided through an embedded system architecture, which was connected to the reader by means of some communication interface, for example, through of a Serial or Ethernet connection. This component gives it the required local intelligence and the desired functionality.

The GPS user devices (such as smartphones) already include necessary communication and computational features. In this manner, the system architecture will be deployed, directly, in these devices by means of a Mobile Application or ‘App’. The main modules of the computational architecture of the acquisition components are depicted in .

 

Computational architecture of the acquisition components.

A full description of the Smart Sensor architecture for RFID was provided in our previous work [56]. In this work, the architecture has been generalized to support several methods of acquiring the user locations. Thus, different types of components can be deployed to cover a broader range of acquisition possibilities (RFID, BLE, NFC and others). The Location Acquisition Module handles the data according to the technology used and transforms it into an interoperable scheme described in the next sections.

In the previous proposal based on RFID, when a reading was received through the proximity of an RFID tag, an optimization process was called to avoid redundant readings of the same individual (Optimizer module). This module has been modified for the case of GPS and similar technologies. In this case, a period of periodic reading is established. In each reading of the GPS position it is compared with the previous reading. In case both readings are brought closer, based on a radius previously configured, the new reading is discarded. As a result, sending unnecessary information is avoided. Once the locations are obtained, optimized and stored, the generation of the message is performed through the Generating locations message module. This generation is dependent on the technology used to locate the position of citizens. In the case of technologies such as RFID, a message is generated with the position of the RFID reader but with multiple time readings of different citizens. In the case of technologies such as GPS, where readings only belong to an individual, the generated message contains multiple temporary positions of that same citizen. By default the message is generated by the JSON standard as shown in .

 

JSON locations Message formats. (a) RFID Message Format; (b) GPS Message Format.

In both cases, the communication channel is optimized according to the protocol used. When there are enough readings, the generated message is transmitted through the Locations message sender using the corresponding protocol (i.e., HTTP). In addition, a software version of this component has been developed to benefit from the presence of GPS components in many devices worn by citizens. The collected data will, in this manner, be formatted and integrated with the system.

5. Cloud Support Infrastructure

The Cloud Support Infrastructure is the centralized part of the system for the collection of all the individual locations and issues and computation for the generation of the KAIs of the city area. This infrastructure is installed on the cloud to serve for many city areas and be accessed by stakeholders. It consists of four main components: the accessibility middleware, the structuring of urban information, the accessibility issues detection, and the KAI generator for each city. depicts the scheme of this part of the system. The components of the system are described in the next subsections.

 

Components of the Cloud Support Infrastructure.

5.2. Structuring Information Component

This component performs the calculation of the possible routes followed by the citizens across the area under study. This process is performed according to three criteria: (a) completeness, (b) diversity, and (c) segmentation of routes:

• (a)The completeness is related to complete information regarding the movements of citizens.
• (b)The routes generated may be diverse depending on the time of the day, the day of the week and the month of the year.
• (c)Another important aspect is providing separated information for citizens with disabilities and citizens without disabilities, and even the disability type.

Each route is composed by a set of single individual points read by the deployed acquisition components along the area. The single location (LOC) will be the most basic and essential entity to determine the citizen flow. This entity will be defined by the following structure:

LOC ≡ ⟨ts, id, ac, pos⟩

where

• ts: timestamp in which the reading takes place.
• id: tag or mobile device identifier that is carried by citizens.
• ac: acquisition component identifier. In the case of the application for mobile devices, this identifier is the same as id.
• pos: acquisition component position. This represents the single location of the acquisition component device determined by its longitude (lo), latitude (la) and height (h). The mobile devices are carried by the users, and the readers can be installed in a fixed manner (for example, in the street furniture) or in a mobile manner (for example, in a bus door). This scheme supports all of these approaches.

Each single route (R) is composed of a list of single locations read by the deployed acquisition components along the area under study when the citizen moves in it. That is,

R ≡ ⟨LOC₁, LOC₂, …, LOC_n⟩

where for each LOC_i ∊ R, LOC_i(ts) < LOC_i+1(ts), that is, the time is increasing while the citizen goes along the route.

Each single route (R_AB) draws a path from a start point A to a destination B of the city. The points A and B are defined by the position field (pos) of LOC₁ and LOC_n, respectively. In addition, the routes can be decomposed on a set of subroutes where their starting and finishing points are in the initial route. illustrates this idea.

 

Route decomposition example.

The following list of subroutes can be deducted from the initial route RAB depicted in :

sub(R_AB) = {R_Ax1, R_Ax3, R_x1x5, R_x2x4, R_x4x6, R_x4B, …}

where for each R_ij, the points i and j are defined by single locations of the route R.

The inference process consists of defining the path followed by a citizen from the set of her/his discrete locations read. The first step is to complete the different routes. As defined above, the routes are a list of discrete individual points. In the case of the data read with GPS technology, the single locations are close enough together to easily infer it. However, for the other communication technologies, such as RFID or BLE, the consecutive locations can be distant in time and/or space, and an inference process is needed to build the probable path. In this case, it is useful to have a neighbouring map of the deployed acquisition component network in the city. This information will help to complete the routes from the individual locations because two consecutive reads must provide neighbour acquisition components.

Let SN ≡ {S₁, S₂, …, S_n} be the set of acquisition components of the network. For each S_k ∊ SN, let n(S_k) ⊂ SN be the neighbouring sensors of S_k. In this manner, shows an outdoor example of the neighbouring sensors and how the route can be completed.

 

Neighbouring sensors in blue of the sensor located in X₂.

In the case shown in , if a citizen is read by the sensor located in X₂, and the next sensor that reads the same citizen is in X₄, then a lack of reading takes place because X₄ is not a neighbour of X₂. The system can insert an automatic read by the most likely sensor and complete the route. It seems to be the sensor n₅. Other sensors could be considered according to the time difference between the readings and the options of frequent routes. For mobile acquisition components, this method does not apply.

Once the routes have been completed, the information is ready for issue detection and querying. The queries for the diversity goal could be sent to the set of the single routes calculated, as well as all their subroutes. However, the calculation of aggregate routes provides more meaningful information about the real movement habits of the citizens. The route aggregation is performed by grouping the set of single routes that have the same starting and finishing points in a time interval, for example, all routes and subroutes that go through A and B.

The segmentation goal needs to register the accessibility issues of the area under study. To achieve this knowledge, the mentioned collaboration of the citizens with and without disabilities comes into play. The cooperating users must enter their disability type or none in the system.

5.3. Issue Detection Component

This component is focused on detecting the urban accessibility issues for citizens with disabilities. The proposed detection process is achieved through two methods: by means of an automatic method using the acquisition system described previously and by means of the citizen collaboration using an App for Smartphone and IoT solutions.

5.3.1. Automatic Detection Method

In the first method, the basic idea consists of comparing the routes followed by citizens with and without disabilities to identify significant differences among them caused by the existence of a lack of accessibility paths. This idea was proposed in our previous works [54,56]. In this research, the methods have been improved using the advanced structured information proposed in the previous subsection.

The main aspect examined to automatically detect an accessibility problem is the path length. Usually, when there are differences in length, it is because the people with disabilities have to take a longer path to get to their destination. For example, describes this situation (coordinates: 38.387298, −0.512315). The shortest route has stairs, which prevent the passage of wheelchair users. It can be noted that the length difference between the two routes is more than twice over (57.1 m vs. 27.8 m). Then, the system raises an accessibility issue at this point.

 

Example 1. Different routes detected due to movement disability.

Secondly, if there are no significant differences in length, the frequency of the routes followed by people with disabilities may show the problems, i.e., when such people always take the same path. For example, describes this situation (coordinates: 38.384863, −0.510106). In this case, both routes have the same length (252 m), but in all cases, the people with disabilities go along the green path because this way is inaccessible for wheelchairs due to the grass and the stones of the road. The system can detect an accessibility issue in this route.

 

Example 2. Different routes detected due to movement disability.

Algorithm 1 describes the pseudocode of the previously proposed methods. There are some high-level functions in the method whose code is not provided, but it is easy to deduce how they work.

The selected area can be defined by rectangle coordinates, circular coordinates or a predefined area, such as a street, a building or a city district. The system must be calibrated by adjusting the threshold values and examining the false positive and negative issues detected.

Finally, a simple method consists of analysing a specific place where an accessibility action has been made to verify if this is used properly, for example, adapted toilets, ramps, lifts, etc. After a time, it can be deduced if it works correctly. Of course, once the system raises the list of issues, an in-place human monitoring is needed to check the validity of the data obtained and to identify the actions needed to overcome each issue.

Algorithm 1. Detection Accessibility Issues.
1.	Select set of aggregation routes (GR) from Selected Area
2.	For each route r of GR
3.	let r.A be the start point of r
4.	let r.B be the destination point of r
5.	let dr = dist(r.A, r.B) be the distance from r.A to r.B
6.	let sr = dssb(r) be the disability type of r
7.	For each route s of GR
8.	if r.A and r.B belong to s then
9.	let ds = dist(s.A, s.B) be the distance from A to B following the route s
10.	let ss = dssb(s) be the disability type of s
11.	if |dr-ds| < threshold(sr, ss) then raise_issue(difference_in_length)
12.	else if |difference(r-s)| > threshold_d then add(s) to ListOut(r)
13.	endif
14.	endif
15.	endfor
16.	For each type of disability d of D
17.	let cr = card(d, r.A, r.B) be the number of routes from A to B followed by people with the disability d
18.	let cs = card(d, ListOut(r)) be the number of routes added in ListOut(r) followed by people with the disability d
19.	if (cr>0 and cs/cs) > threshold_q then raise_issue (difference_in_path)
20.	endif
21.	endfor
22.	endfor

5.3.2. Accessibility-Issues Self-Reporting Service

This App is oriented towards obtaining the end-user experience. Its aim is to provide a user interface for mobile devices to allow citizens to add their accessibility experiences while moving around the urban environment anywhere, anytime.

The analysed works focused on the detection of urban accessibility issues are mainly based on evidence or manual surveys. With these techniques, the results obtained are from limited groups of individuals regarding a specific environment, and with more economical cost and time. In contrast, the aim of the proposed application is to change end-users into active participants in the effort to detect accessibility problems. Through this application, comments, photos or reports can be sent by citizens anonymously, and it allows information collecting on urban accessibility provided directly by the citizens. The idea behind this functionality is that the accessibility issues are notified at the moment when the user finds them. The application can report the exact location of this claim and take a photograph of it.

The application can be installed on user mobile devices by downloading the App. No personal data is needed, just the disability type of the user. The next picture shows an example of the interface of this application with real data ().

 

App for Accessibility Self-Reporting Issues.

The information gathered explicitly complements that obtained through the automatic monitoring system and provides valuable information to the administrators on the status of accessibility and urban sustainability.

6. Urban Accessibility Information Services

This layer is in charge of providing the accessibility services to the key stakeholders: citizens, city administrators, and public service provider companies, such as transportation, emergencies, etc. The provision of the services is designed according to the Service Oriented Architecture (SOA) principles paradigm [60]. The resulting information is provided in a structured manner using eXtensible Markup Language (XML) or JavaScript Object Notation (JSON) languages to enhance the interoperability. The core of the service provision architecture is an Enterprise Service Bus (ESB) platform. This is a powerful infrastructure for software integration and communication that enables interoperability, enforce security and compliance, logging and monitoring and provides agility and flexibility through of the micro-services orchestration (service logic) [61]. shows the overview of the information service architecture.

 

Urban Accessibility Information Service Architecture.

The results of the system should be provided following two models, with a first approach based on the paradigm of Business to Customer (B2C), oriented to offer a human-machine interface that directly allows a user to interact with the system. In this case, a presentation layer builds web or mobile interfaces for thin clients. The second approach is based on the paradigm of Business to Business (B2B), which allows consumers to connect with the service system to incorporate it as process of their organizations. This will be achieved throughout Web Services to improve the information decision task. Both protocols can define search filters to monitor a specific urban area according to the structure of the information provided.

This interface allows surfing of the worldwide map and shows the area of interest of the user. Within this area, the application visualizes the result of the Urban Accessibility Information service and shows the KAIs corresponding to the places with accessibility issues. The user has the option of selecting one of these marks to get details about it. depicts an example of the web interface developed.

 

Urban Accessibility Monitoring Service Web Application Screenshot.

7. Evaluation and Discussion

The aim of this section is to evaluate and validate the described system and the operation of the modules and components described in the previous sections. It is expected to verify the reading technology and data processing performance in acquiring the citizens’ locations and traces along the evaluation context. In this manner, a simulated use case for people with and without disability in a real-world environment has been designed.

7.1. Evaluation Setup & Implementation

The technologies used to read the citizens’ locations are RFID and GPS. The development has been improved to include the new functionalities described, and both the communication and positioning technologies have been combined to acquire the locations from different sources and to support outdoor and indoor deployments.

We have implemented two acquisition prototypes to validate the proposal. First is a RFID prototype, based on our previous research described in [56]. The second is a GPS prototype through a native App for the Android system deployed over different mobile devices such as the Aquaris BQ M5.5 and E5 (BQ, Madrid, Spain).

Now, outdoor areas can be analysed by both types of technology. However, the use of RFID requires a specific installation of the acquisition components in the places of special interest, such as building entrances, bus stops, schools, etc. In contrast, the GPS technology requires the citizen’s smartphone and for the App to be installed.

The ‘Cloud Infrastructure’ and ‘Information Services’ have been implemented using Mule ESB together with the NodeJS framework and Express Library to develop RestFul Services as a backbone element of the proposed architecture. To provide data persistence, the MySQL community version database manager has been used. In addition, ActiveMQ has been used as message service to implement the guaranteed delivery pattern. The presentation layer has been developed through Web technologies such as Cascading Style Sheets (CSS), HyperText Markup Language version 5 (HTML5) and JavaScript using the AngularJS framework. It provides more accessibility and usability to the final users. In addition, a web-based Geographical Information System from a third-party application (Google Maps JavaScript API v3) has been used to represent the results of the examples utilized. In addition, the ‘Accessibility-issues self-reporting service’ has been developed by means of Android Development Tools for Mobile Apps. shows the general prototype architecture.

 

Urban accessibility monitoring service web application screenshot.

7.2. Case Studies

In this section, two case studies are developed to explain the methods and how the system works in different scenarios. The simulation performed does not aim to be exhaustive or to collect all existing situations but rather to show the possibilities of the proposed system for detecting existing accessibility issues. In this manner, the case studies will provide a better understanding of how the accessibility of infrastructures can be assessed.

The area under study is in the campus of the University of Alicante in Spain. This site is complex enough and similar to a city but on a smaller scale that facilitates the deployment of the system. Because it has different areas (rooms, labs, small shops, restaurants, open areas, etc.) and diverse activities (academic, administrative, social, small health centres, sports, etc.), the university campus can be seen as a small city.

The evaluation was focused on analysing the daily paths of the students and lecturers across the campus (outdoors) and their movements inside of the buildings (indoors), such as the access to the classrooms, laboratories, libraries, toilets, etc.

shows the scenario used to carry out the evaluation and the infrastructure deployed in the indoor scenario. There are two RFID-based acquisition components installed in the first floor with the objective of checking the appropriate use of the WC rooms.

 

Indoor scenario and RFID-based acquisition components deployed.

After one week of operation, several KAIs were registered in the system. They come both from self-reporting application and items inferred by the system. and show different KAIs identified during the evaluation period carried out in the context described above.

 

Different KAIs obtained in the validation process in Politécnica IV building. (a) KAIs 234 & 237; (b) KAI 235; (c) KAI 239.

Table 3

Different KAIs obtained in the validation process in Politécnica IV building.

Test 1. Student Movements in EPS IV Building
KAI #	Description	KAI	Source
KAI 234	Inaccessible WC in module 3 of Politécnica IV building, first floor.		Self-reported & Inefficient inferred
KAI 235	Impossible access to lecturers’ offices due to the height and the type of the closed door.		Self-reported
KAI 237	WC adapted for disabled is closed. Keys are required for entry.		Self-reported
KAI 239	Ramp of access to different classrooms of theory and laboratories too long and too slope.		Inefficient inferred

As shown in the previous table, the acquisition components detect that people with movement disabilities cannot use the adapted toilet. In addition, there are self-reported KAIs of the same issue. This is an example of an inefficient accessibility action since it does not meet its function. The other inefficient issue is the outdoor ramp between two floors. The GPS paths detect that it is not used by people with disabilities. A simple observation after this finding that anyone can take note of is that this ramp is too long and too steeply inclined.

The next use case is focused on some common paths of the students and lecturers on a normal working day: from classroom building to administrative services building and lecturers’ offices. In this case, the routes are outdoors, but some acquisition components can be installed in the entrances of the buildings to trace the users’ access to them. and show the area of this evaluation case and a sample of the issues registered.

 

Outdoor scenario and RFID-based acquisition components deployed.

Table 4

Different KAIs obtained in the validation process along some common paths.

Through the proposed system, different paths taken by people with and without disabilities to reach the same place or access the same building have been detected. In addition, the self-reporting contributions by users complement the information generated to build the full list of accessibility issues. The knowledge of this experience of citizens where they move in the area is very much of interest to verify the validity of the accessibility actions made. All interested users were especially helpful with this tool and, as noted in the results obtained, will serve as a source of information to prioritize some actions over others. An example of this is the multitude of inaccessible and inefficient points reported. In these cases, existing alternatives are commonly used. All these results have demonstrated the complementarity of the approaches of this proposal.

Based on the results obtained in the case studies, we have observed how the proposed system is able to automatically identify problems and deficiencies related to the accessibility of a specific area. The KAIs registered will help those responsible for managing accessibility to identify such deficiencies, prioritize them, and make decisions. We have verified that the combination of different technologies (GPS + RFID) allows for the detection of situations in heterogeneous areas in which not all the acquisition alternatives are available.

Finally, regarding the performance of the system, the computing cost of the system is variable and highly dependent on the daily information gathered. A wide deployment of the system should retrieve many locations points and routes to process. However, this is not a problem since the depth processing can be performed offline by the cloud infrastructure. In addition, the involved tasks are very parallelizable. That is, several routes can be processed in parallel by different computing elements of the cloud. In this manner, updated information can be provided to the stakeholders without delay.

8. Conclusions

The research presented in this paper proposes a system for evaluating the effectiveness of urban accessibility in a dynamic manner that is sustainable over time. This system provides a user-centric tool supported by technology to discover, assess and classify urban accessibility issues and generate value-added information regarding accessibility through the setting of Key Accessibility Indicators (KAI).

The proposal infers the real routes of citizens across the area under study by means of acquisition methods based on communication and positioning technologies, smart sensor networks and a distributed cloud-based architecture. In addition, the system allows acquisition of citizens’ accessibility experiences while moving around the urban environment, any place and at any time, through an App for mobile devices. From this information, the proposal allows an automatic smart scheduling of urban planning actions focused not only on the removal of barriers but also on the prevention of their occurrence.

The knowledge of movement flows of citizens with and without disability, the frequency of routes and the distribution of the main origin and destination locations will allow the efficient management of the resources of cities and ensure long-term sustainability in terms of enhanced service provision. The analysis of the identified routes together with the employed time to route them provides an accurate picture of the difficulty faced by people with disabilities, for example: urban spaces without access for people with disabilities even though these areas have high activity; urban spaces or areas highly frequented by collective with a specific type of disability; and accessible urban spaces that, temporally or with a certain frequency, become inaccessible.

Author Contributions

Higinio Mora, Virgilio Gilart-Iglesias, Raquel Pérez-delHoyo and María Dolores Andújar-Montoya wrote the paper and read and approved the final manuscript.

References