=Paper=
{{Paper
|id=Vol-1674/iStar16_pp85-90
|storemode=property
|title=Applying Tropos Modeling for Smart Mobility Applications Based on the FIWARE Platform
|pdfUrl=https://ceur-ws.org/Vol-1674/iStar16_pp85-90.pdf
|volume=Vol-1674
|authors=Hugo Estrada,Karen Najera,Blanca Vazquez,Alicia Martinez,Juan Carlos Téllez,Juan Jose Hierro
|dblpUrl=https://dblp.org/rec/conf/istar/EstradaNVMTH16
}}
==Applying Tropos Modeling for Smart Mobility Applications Based on the FIWARE Platform==
https://ceur-ws.org/Vol-1674/iStar16_pp85-90.pdf
Applying Tropos modeling for Smart mobility
applications based on the FIWARE platform
Hugo Estrada1, Karen Najera1, Blanca Vázquez1, Alicia Martinez2, Juan Carlos Te-
llez1, Juan Jose Hierro3
1 Center of Research and Innovation in Information Technology and Communication
- INFOTEC, Mexico
{hugo.estrada, karen.najera,
blanca.vazquez,juancarlos.tellez}@infotec.com.mx
2 National Center of Research and Technological Development - CENIDET, Mexico
amartinez@cenidet.edu.mx
3 Telefonica Research and Development - Madrid, Spain
jhierro@tid.es
Abstract. One of the main issues in current cities is mobility. The unprecedent-
ed growth in urban mobility has led to multiple open challenges to be consid-
ered in the design of smart cities that improve the citizens' quality of life. There
are several factors involved in the development of urban mobility applications
as pollution monitoring, traffic congestion, and environmental conditions. One
of the most well-founded platforms to develop complex technological solutions
for smart cities is FIWARE, which proposes open standard specifications using
an open cloud infrastructure. In this context, modeling approaches are needed to
represent, in a high abstraction level, the to-be view of a mobility application,
with models that help the analysts to understand the semantic of the physical,
human, and software actors involved in the application development. In this pa-
per, we explore the use of Tropos framework for modeling in a high abstraction
level, the requirements of a smart mobility application that will be developed
with FIWARE. The paper discusses the issues found in using Tropos to repre-
sent the smart mobility solution.
Keywords: Tropos modeling, smart mobility, FIWARE platform
1 Introduction
Nowadays, the high density of vehicles in current cities has led to several complex
issues, such as traffic congestion, pollution, longer trip time, decreasing public safety
and increasing noise. For instance, in Mexico City, with 20 million people, there are 4
million vehicles on the road and a total of 22 million trips every day [1]. In this com-
plex scenario, mobility concerns represent key challenges in smart cities design,
where novel technologies such as Internet of Things (IoT) are required for developing
systems that produce intelligent decisions based on data provided by objects connect-
ed to Internet. One of the most recognized platforms for smart city development is
FIWARE [2], an open innovation ecosystem that allows creating new applications
Copyright © 2016 for this paper by its authors. Copying permitted for private and academic purposes.
Proceedings of the Ninth International i* Workshop (iStar 2016), CEUR Vol-1674
and Internet services. Mexico has an instance of the FIWARE platform called Mexi-
can FIWARE Lab Node hosted in the National Future Internet Laboratory
(http://lanif.infotec.mx/), where we will deploy a smart mobility application taking the
advantages provided by the platform. In this sense, more than 100 European cities are
already using FIWARE, most of them for mobility solutions, however, one of the
main issues of FIWARE, as well as of most of current smart city platforms, is that
they are programming-based approaches, where solutions are generated starting from
low design levels. In this paper, we present the use of the Tropos for modeling in a
high abstraction level, the requirements of a smart mobility application based on
FIWARE. A Release Planning (RP) approach will be used for developing application
[4]. As a result, some challenges have been identified in the use of Tropos concepts to
represent IoT components of technological solutions for smart mobility applications.
2 Objectives of the research
This paper has two main research objectives: a) to explore the use of Tropos to model
the complex scenario of a smart mobility application to be developed, called Green
Route, representing smart technology components as intentional actors since they
have the ability to make decisions by themselves, and b) to make explicit the chal-
lenges to be faced in Tropos to appropriately represent smart technological compo-
nents, specifically for mobility applications based on FIWARE.
3 Scientific contributions
We propose using the Tropos framework to model the scenario and the operational
environment of Green Route. The objective of Green Route is to help the final user to
determine the best route to follow to reach a destination, taking into account the user
profile (e.g. age, health conditions, disabilities, etc.), and the user preferences, such as
transport type. Green Route proposes the ideal route for the user, avoiding routes with
high levels of pollution, floods or pollen, etc., allowing for instance, to obtain the
preferred routes for people with respiratory diseases. Green Route obtains data from
mobile and fixed sensors (based on the IoT approach) to get real time information.
Additionally, Green Route obtains data from other sources, such as air conditions
provided as open data and data generated through the concept of human-as-a-sensor
to get the user feedback about specific conditions in its location (e.g., rain, pollution,
etc.). Green Route is based on the cloud capabilities of FIWARE, which enables to
connect sensors through standard specifications and provides software components to
store sensors data, enabling data analysis and querying. It is important to point out
that for developing the route system, we are considering to generate a series of releas-
es among the requirement process, in order to satisfy the user needs obtaining early
feedback and to satisfy important technical, resource, budget and risk constraints.
Hence, we don’t follow the conventional Tropos software development process.
86
Applying Tropos modeling for Smart mobility applications based on the FIWARE platform
Tropos modeling of a smart mobility application based on FIWARE
The goal of this phase is to present the generated Tropos models for Green Route1,
which mainly focuses on: a) the representation of all the elements that support Green
Route: data acquisition and infrastructure and platform as a service for data storage,
processing, analysis and consumption of data in the cloud, and b) the representation of
all the required functionalities of Green Route to answer user needs. The actor con-
cept has been used to represent the hardware and software involved in the solution,
since they communicate among them without human intervention. Additionally, these
components have skills to independently take decisions based on its context, making
use of mechanisms of the artificial intelligence. For this reason these components
were modeled as intentional elements.
The first model is the as-is setting model (Fig. 1) which describes, in a high level
view, the current scenario to determine the best route to follow to reach a destination
(without using a smart application), where the user needs to access to different data
sources to consolidate the information. The second model is the goal model (Fig. 2)
that represents the goal dependencies of the key actors of the Green Route in a high
level view. This model was generated as a first approximation to automate the as-is
setting model. In order to represent the sources to acquire environmental data, the
model contains two units: Mobile Data Acquisition Unit (Mobile Unit) and Fixed
Data Acquisition Unit (Fixed Unit). These units have been represented using the tech-
nology module concept [3], which allows to represent technology in a high abstraction
level including information about functionalities, specific requirements and quality
features. In our proposed scenario, the technology module concept allows to encapsu-
late all technology components related with data acquisition from several sensors,
data local storing, wireless communication and data transmission.
The Mobile Unit represents the actor and dependencies involved in the process of
acquiring environmental data from unit mounted on transport in movement, and the
Fixed Unit represents the actor and dependencies involved in the process of acquiring
environmental data from fixed monitoring stations. FIWARE Lab receives environ-
mental data captured by sensors from Mobile and Fixed Units, also it receives envi-
ronmental data obtained by a crawler from an open data website. The PaaS consumer
(Platform as a Service) actor depends on FIWARE Lab to provide cloud resources on
demand.
Obtain
Open data D Obtain
pollution level D
D forecast
D
website by zone Weather
D User forecasting
Obtain D Obtain current website
D
best route environmental
D
D D D conditions
Visualize
Traffic and D D
D
traffic
navigation
app Obtain Obtain
Obtain Obtain
average transportation
route distance
time options
D
D
D D
Maps
website
Fig. 1. The as-is setting model to reach a final destination manually
1 Models are available in: http://www.tagoon.semanticbuilder.com/Troposmodels/
87
Proceedings of the Ninth International i* Workshop (iStar 2016), CEUR Vol-1674
User
D D
Paas
Consumer
Determine route Provide real time
based-on user profile environmental data in current
and type of transport position (data source 4)
D
Technology Module:
Provide cloud resources
Manage D
D
Fixed Data environmental data
on demand
Acquisition Unit in the cloud
D
D
D
Acquire
D
environmental data D Provide data
(datasource 2) storage
D
Green
D
Technology Module: Route
Mobile Data D FIWARE D D
Acquisition Unit Acquire Publish Open Data
Lab
environmental data D
D
D
D
in movement
D
D
(data source 1) D Provide user
validation Spatial
D Data
Information
To be mounted on
the transport
D Crawling Management
visualization
Transport Environmental data
(data source 3)
Obtain context
data from cities
D D
D D D
Open data Provide Data Geographic
website D environmental Consumer Information
Crawler
data System
Fig. 2. Goal dependency model of the smart city mobility application
The Green Route depends on the FIWARE Lab for storing, management, analysis and
consumption of environmental data in the cloud. The user depends on Green Route to
determine the ideal route to follow based-on user profile and type of transport, and
provide real time environmental data in current position. The Green Route depends on
Geographic Information System to perform historical analysis and to get data visuali-
zation through maps. In Fig. 3, a fragment of the third model (strategic rationale: SR)
of the Green Route actor is presented. This model represents the social and intentional
relationships mainly of IoT technologies and the final user.
Fig. 3. Fragment of the SR model of the Green Route actor
In order to achieve the goal “Determine route based-on user profile and type of
transport”, the Green Route has to accomplish five main tasks: a) “Manage data in
the cloud”, which involves activities related with storage, security and data manage-
ment; b) “Get environmental information from different data sources”, which in-
volves activities related with getting environmental data from sensor networks, open
data websites and data provided by the user; c) “Get user information”, which in-
88
Applying Tropos modeling for Smart mobility applications based on the FIWARE platform
volves activities related with collecting all the information from the user required to
provide the best route; d) “Determine best user route”, which involves activities relat-
ed mainly with cross information and data analysis to generate the best route for the
user, taking into account all the available information from the user and from the en-
vironment; and e) “Provide data visualization”, which involves activities related with
the visualization of available data through maps. At the present time, models have
been used for analysis purposes. The next development phases will be incremental
using technologies such as DevOps and OpenShift. Therefore, we have not used a
specific Tropos tool for the modeling process.
3.1 Experiences
In this section, we detail some of our experiences using Tropos while representing the
IoT components of the smart mobility application. In general, Tropos enable analysts
representing most of elements of smart mobility scenario: actors of different nature,
technology, software and people, for instance, actors involved in cloud computing,
sensor networks, crawling and user interfaces. Some issues identified are:
- The crawler actor obtains environmental data directly from a website in an auto-
matic way. In this case, there is not goal, task or resource delegation between actors,
since the website is not aware of the crawler. This has been represented in the model
as a goal dependency where the crawler is the depender and the website is the depen-
dee, however, the crawler does not delegate the website the responsibility to provide
the data.
Open D
data D Crawler
Provide
website
environmental data
- Two examples of conflicts representing dependency between two actors have been
identified, since the delegation is not about a task execution or a resource provision-
ing: a) The Mobile Unit depends on the transport to be mounted on it, b) The
FIWARE Lab depends on OpenStack to provide Platform as a Service (PaaS). In these
cases, a goal dependency has been used to indicate a physical dependency between
two actors, however, these dependencies do not have the semantic of goal delegation.
<>
Mobile Data Acquisition Unit
To be mounted on
D Provide FIWARE
D the transport D Transport OpenStack
PaaS
D
Lab
4 Conclusions
The advantages of using Tropos for modeling smart mobility applications can be
summarized in its effectiveness to cover a deeper understanding of a complex envi-
89
Proceedings of the Ninth International i* Workshop (iStar 2016), CEUR Vol-1674
ronment, such as the smart mobility scenario, where both, software and hardware
must take decisions independently based on its context.
We consider that representing a smart mobility scenario, without a formal language to
present intentional social relations, will difficult to obtain a high level view of the
system to be developed and to clarify the role of each actor in the fulfillment of the
goals of the system. Moreover, it could not be determined the vulnerability level of
the system if some actors fail to accomplish its goals. In this sense, we consider the
effectiveness of using Tropos to enable the analysts to represent, in a high abstraction
level, the complex IoT technologies involved in a smart mobility application. Tropos
allows to obtain a deeper understanding of the environment where the application
must operate, and the kind of interactions between components and the final users,
considering that in the proposed scenario, there are complex relationships among
actors of different nature, such as: smart software applications, cloud platforms,
crawlers, sensor networks, human users, etc. It is important to point out that the de-
velopment of mobility applications is highly dependent on technology. This techno-
logical dependency has been appropriately represented using the concept of module.
Advantages and issues in using Tropos in modeling the to-be mobility application
were identified.
5 Ongoing and future work
One of the conclusions of the modeling activity is the need to extend Tropos concepts
for the representation and interaction of IoT technologies. We believe that the Tropos
semantic (delegation of responsibilities, the vulnerability associated to dependency,
the role of depender and dependee) does not comply the semantic of these new ways
of interacting. For instance, some tasks are executed over on the information of an
actor, without his permission. New modeling concepts are been developed to properly
represent the needed semantics. As future work, we will use of the generated model as
starting point of a semi-automatic process to develop the initial implementation of the
system to-be. This is possible because the interaction with FIWARE components are
standardized and they have well-defined functionalities.
References
1. Mexico City Mobility Challenge. http://mexicocity-mobility.devpost.com/ (June, 2016).
2. FIWARE ecosystem. https://www.fiware.org/ (June, 2016).
3. Alicia Martinez, Blanca Vazquez, Hugo Estrada, Luis Santillan, and Crispin Zavala: Incor-
porating Technology in Service-Oriented i* Business Models: A case study. In: Journal In-
formation Systems and e-Business Management, S.I.: Model-based engineering for next-
generation EIS, 2016, pp. 1-27, ISSN 1617-9854, DOI 10.1007/s10257-016-0316-9.
4. D Greer, G Ruhe, Software release planning: an evolutionary and iterative approach, Infor-
mation and Software Technology, Volume 46, Issue 4, 15 March 2004, Pages 243-253,
ISSN 0950-5849, http://dx.doi.org/10.1016/j.infsof.2003.07.002.
90