At present time, Ambient Intelligence (AmI) is a revolutionary computing paradigm that promises to have a deep effect on the way we interact with the computers, devices, physical spaces, and other people. The development of applications for this paradigm represents a new challenge for analysts and software engineers due to the complexity to consider physical, human and system actors interacting among them to give support to existing human activities. In this context, requirements engineering plays a very relevant role in AmI applications development because it allows the analysts to contextualize the expected functionalities of the system to-be before its implementation. In this paper, a technology modeling method, based on i*, has been proposed as a tool to model the software requirements for Ambient Intelligence applications.
The vision of Ambient Intelligence (AmI) is to create a “physical world that is richly
and invisibly interwoven with sensors, actuators, displays, and computational
elements, embedded seamlessly in the everyday objects of our lives and connected
through a continuous network” [
Copyright © 2016 for this paper by its authors. Copying permitted for private and academic purposes. However, the complexity of AmI systems, its complexity in the definition of the requirements for new types of interaction, intelligent systems, sensitivity to context and pervasiveness are fundamental concepts that must be captured. Therefore new tools are needed to facilitate the definition of requirements for AmI systems. This paper presents a methodology to face the modeling of AmI applications that considers the basic elements of an ambient intelligence applications using as basis the i* framework notation.
The paper is structured as follows: Section 2 describes the objectives of the research work. Section 3 presents the proposed approach for modeling software requirements for ambient intelligence applications. Section 4 provides the conclusions and future work. 2
The main objective of this research work is to propose a software requirements methodology for AmI applications that helps analysts to understand the role of the key actors of a system of this kind. To do this, two main objectives were identified: i) the development of a modeling methodology that extends i* models, and b) the development of a case study to provide an empirical validation of the proposed approach. 3
The core of our contribution is the development of a modeling methodology that extends i* models with elements specific for AmI systems. Three steps were followed to generate the proposed methodology: a) The analysis of existing methodologies for requirements engineering, b) the discovery of new modeling elements and c) the proposal of a requirements engineering methodology based on i* framework. 3.1
The methodologies selected are methodologies of requirements engineering. These
are: (i) Documentation of requirements approached to users (DoRCU) [
The main objective of this analysis was to identify advantages and disadvantages of methodologies when they are used in a software requirements project of an AmI application. We also wanted to know if these methodologies could be adapted for representing the key elements of AmI. In this analysis, some techniques and tools of requirements engineering like requirements classification, questionnaires, business and technology modeling, and basic notation of i* were analyzed. As result, we identified that they do not provide basic modeling concepts for AmI systems due to non-graphical way to represent the requirements, gaps in the representation of interaction forms and clarity in the specification. Therefore, the framework i* allow us to generate modules, model graphically and incorporate technology, which is useful for representing the elements of an ambient intelligence application. However, i* does not modeling a natural interaction among human and technology. For this reason we propose to extend this framework with a natural interaction, i.e. how humans interact naturally through his senses with the environment. 3.2
Ambient Intelligence is the vision that technology becomes invisible, embedded in
our natural environment, present whenever necessary, accessible through systems of
simple interaction, addressed to all our senses and adaptive to users and context [
In this research work we unify the basic elements of an ambient intelligence application in: ubiquity, intelligence, context sensitivity and natural interaction. These concepts are modeled with proposed methodology. 3.3
The proposed methodology models basic elements of an AmI application: context, intelligence, ubiquity and natural interaction. The AmI applications need an implicit interaction with the users, unlike to the traditional systems that need continuous interventions of users for indicating to the system each action to perform. Our methodology is composed by four phases (Fig. 1): business modeling phase, technology modeling phase, interaction modeling phase, and finally, general modeling phase.
Business modeling phase. In this phase, the organization has been modeled
through of an i* service model. The key idea of the service-oriented approach is to
use business services as building blocks that encapsulate internal and social
behaviors. In this model, we try to capture the context sensitivity through the business
context [
Technology modeling phase. This phase represents all technological components
of the software system to-be. It is compose of two models: 1) Model of technology for
development; 2) Model of Technology to operate.
a) Model of technology for development: This model shows a representation of the
technological elements required for application development. We use i* Framework
notation to represent the technological elements using Agent concept and the system
that will be generated by using Actor concept. The objective of this model is to
provide the analyst with the information about technologies involved in the
application.
b) Model of technology to operate: This model shows a representation of the
technological elements required for the operation of the application. We use a
protocol model [
Interaction modeling phase. This phase shows the interactions in the software system between users and technology. We define three types of interactions: 1) human-technology interaction; 2) technology-technology interaction; 3) humanhuman interaction.
In this phase we propose the user-technology interaction model. This model shows the interactions that exist in an AmI application. The model incorporates the technological elements defined in the previous phases. The goal of this model is to provide the analyst with the information about the types of interactions that exist within the application. This model describes the following elements: Ubiquity, the technological elements that allows the system runs anywhere. Natural interaction, the specification of new ways to interact with the humans and technology.
General modeling phase. This phase provides information about organizational
goals and the goals of the system to-be using a goal-refinement tree model. This
model allows us to elicit the business goals and to represent these in a goal structure.
To do this, we use a goal classification, which permits us to construct a
GoalRefinement Tree (GRT) using refinement and abstraction strategies [
The validation of our approach was carried out in a real case study. This case study
implements a software system for the detection of social isolation through AmI [
The second model is the model of technology, which shows all technologic elements involved in the software system. Third model is the technology to operate, which capture the intelligence, and the context of each environment. Fourth model is the model of users and technology interactions, which capture the ubiquity and the natural interaction. Fifth model is goal refinement tree, which contains general information about the goals and objectives of the system software-to-be.
Due to space limitations, in this paper we only show the model of user and technology interaction (see Fig. 2) because we considered that this model visualizes more adequately the elements of an AmI system. In this model the actors are represent as users system, technological elements and the environment. The associations among users and technologic elements can be: tactile, gesture and voice. The associations, named communicates, show the interactions among the technological elements. The visibility between users and technologic elements are represented as white or black rectangle.
In this paper, a methodology for modeling software requirements of AmI applications is presented. This methodology consists of five models. In phase 1 we captured part of the context sensitivity with the business modeling. In phase 2 we captured part of the context sensitivity defining the environments in addition to capture the intelligence with the tasks and goals. In phase 3 we capture the ubiquity defining the technological elements of communication and also we capture the natural interaction with the actor associations indicating the type of interaction, the phase 5 show us overview of goals of the application. The results on the validation of our approach with the case study showed better clarity in the specification of the interaction between the user and the system. In addition to a concrete definition of the objectives and goals of the analysts. Currently, we are working on refining our proposed methodology for validating the identification of all elements of an AmI application. For to do this, we are working in several case studies reals of AmI applications.