Α capability-oriented modelling and simulation approach for autonomous vehicle management
Introduction
Much has been written in recent years in specialist publications and indeed in the popular press about the idea of Smart Cities, that is the offering of novel services, supported by Information and Communication Technologies (ICT) . There have been several attempts to provide descriptive definitions of the term “smart city”. According to Giffinger [18], smart city is a city well performing in a forward-looking way in economy, people, governance, mobility, environment, and living, built on the smart combination of endowments and activities of self-decisive, independent and aware citizens. In the authors focus on a city's infrastructure in terms of the city's ability to connect the physical infrastructure, the IT infrastructure, the social infrastructure, and the business infrastructure to leverage the city's collective intelligence.
One important service provision in a smart city is that of smart mobility. Mobility in large cities presents many challenges, which often result in losses of time, decrease in the level of safety, pollution, degradation of life quality, and huge waste of non-renewable fossil energy. Mobility affects citizens, public authorities and businesses.
This paper presents an approach to developing smart mobility applications and demonstrates this approach on a specific aspect of smart mobility, namely that of Autonomous Vehicular Safety (AVS). An AVS is a software-enabled system that is embedded in a vehicle and facilitates the provision of better information to the driver, as well as the achievement of faster response than existing systems, for enabling safety management in road transport. Most modern cars include several warning systems, whereas Global Positioning System (GPS) units provide information about traffic. An AVS system operates on the basis of collecting contextual, profiles and policies information from various sources, processing it intelligently, integrating knowledge and experience, and visually informing the driver proactively for forthcoming emergencies, based on a set of warning functions. Moreover, an AVS system can be seen supporting Smart Mobility Operations, which demand close attention to the effectiveness of the software engineering process since the focus is shifted from engineering of individual systems and components towards the generation, adaptation and maintenance of software-intensive ecosystems consisting of software, hardware, and human agents.
The contribution of the work presented in this paper lies in proposing an approach that follows a systematic process whose central focus is to answer the key question of “what kind of capabilities are required for a Smart Mobility system and how do these capabilities collaborate in a systemic manner and visualised in simulation settings?”. By focusing on capabilities, it is possible to begin the process by considering the essential elements on which the Smart Mobility applications will function, without having to consider how this functioning comes about. Thus, the problem is scoped by considering the city's current set of capabilities, whether any of these capabilities need to be modified or if new ones are to be acquired or developed. When attempting to answer this key question, inevitably one has to consider some related issues. Specifically, “what are the goals of the city's stakeholders?” i.e., “what is the raison d'être for the identified capabilities?”. Also, “who are the agents (human, software, devices) and how would they need to collaborate in order for these goals to be realized?” and in doing so “what kind of data sources do they need?”. Given that a central requirement of such systems is their interaction with human agents a related question is “how will humans interact with the system?” the answer of which will lead to the two subsequent questions of “what is the functionality of the system for facilitating this human-system interaction?” and “what kind of algorithms are required for dealing with the ‘smartness’ of the system itself and how are they externalised in such a way as to assist decision makers in the choices that can be made?” Given that decision makers would have a clear view of the choices than can be fed back to the capabilities and perhaps the evolution of these capabilities result in feasible solutions. All in all, the paper presents a systematic software engineering process, dedicated to Smart Mobility operations, satisfying the challenges identified and discussed for such systems. In order to ground the approach on a realistic application, the approach is exemplified using the AVS as a target application.
The structure of the paper is as follows. First, in section II, an overview of related work is presented together with the description of the example AVS application which will be referenced throughout the paper. Section III introduces the capability-oriented approach in terms of its foundational concepts and the way of working from a methodology perspective. Section IV demonstrates how the concepts and the way of working are applied to the application previously introduced in section II. Section V focuses on the algorithm implemented as a result of the analysis and design that was carried out in preliminary steps of the development process. Section VI provides some simulation results from the execution of the algorithm, on a custom developed simulator, whereas section VII compares the results to similar methodological and algorithmic works. Finally, section VIII concludes this paper with a set of observations, reflections and aspirations for extending this work.
Section snippets
Related work on intelligent transport systems and autonomous vehicular safety
Current mobility strategies are inefficient, leading to enormous losses of time, safety compromises, pollution and degradation in the quality of life, as identified by research community of both, public agencies and private industry [16], [35]. Given the current energy source mix, the above inefficiencies lead to a huge waste of non-renewable fossil energy, indicating the necessity for more efficient and safer mobility.
The response to these challenges has been an emergence of innovative,
3.1. Background
The approach being presented in this paper, known as Capability Oriented Requirements Engineering (CORE), falls in the scientific field of ‘Requirements engineering’ (RE). RE activities take place on the boundary between the stakeholder's business view and the developer's technical view aiming to discover and document the system stakeholders and their needs in a manner that is amenable to analysis, communication and implementation [39].
Traditionally, RE has focused on software systems
4.1. Information elicitation and textual analysis
The process of developing a system whose requirements were defined at a high level in section II, begins with a textual analysis of these high-level expressions. This textual analysis is shown in Fig. 3.
The task of developing the AVS system and indeed any other technological system, could not be achieved with any degree of confidence about its functionality, by considering only the narrative expressed by stakeholders. It is of profound importance that the text is analysed in order to define key
Algorithm specification and implementation
The fourth major process step in the CORE methodology (see section III.C) is concerned with the development of the software component of the AVS management system, which comprises the algorithm specification, the algorithm implementation and the simulation. In this respect, this section (V) is briefly concerned with the algorithm specification. The next section (VI) presents some information on its implementation, along with extensive simulation results that prove its efficiency.
The system
Simulation results
This section presents the 2 last steps of the way of working presented in section III.C. The implementation of the AVS algorithm is presented below (setup of the simulator), whereas section VI.B presents extensive simulation results that demonstrate AVS's efficiency and response to the expectations set out in section III.
Discussion
Due to the dramatic increase of safety-critical, software-intensive embedded systems in modern vehicles, requirements engineering specifically and methodologies for developing such systems in general, have become a crucial consideration in the automotive domain [4], [6], [43], [52]. Specific challenges to this domain stem from the fact that more and more safety functions in modern vehicles need to be realized by integrating on-board as well as environmental interaction systems into higher
Conclusion and future work
This paper presented a complete methodologically based approach on developing a system for smart mobility applications. In order to situate the approach on a pragmatic case, the application chosen for demonstrating the approach and its results is that of a system supporting autonomous vehicular safety.
The paper follows a theoretically strong and practically valid process that involves four major phases, each phase dealing with a clear development goal, whilst ensuring that each one of these
Acknowledgment
The authors wish to acknowledge the Qatar National Research Fund project i-Doha (Proj. No. NPRP 7-662-2-247) project, under the auspices of which the work presented in this paper has been carried out.
References (53)
Evaluating the service quality in multi-modal transport networks
Transport. Res. Procedia [online]
(2016)- et al.
"The brave new world of design requirements."
Inform. Syst.
(2011) - et al.
"Requirements engineering in automotive development: experiences and challenges."
IEEE Softw.
(2003) LTE for vehicular networking: a survey
IEEE Commun. Mag.
(2013)- et al.
"Physical-cyber-social computing: looking back, looking forward,"
in IEEE Internet Comput.
(2015) CrowdSC: building smart cities with large-scale citizen participation
Internet Comput. IEEE
(2013)- et al.
A tour of secure software engineering solutions for connected vehicles
Softw. Qual. J.
(2017) The Smart City Market: Opportunities for the UK, 2013
(2013)- et al.
Guiding requirements engineering for software-intensive embedded systems in the automotive industry
Comp. Sci. - Res. Develop.
(2014) On the role of understanding models in conceptual schema design
Autonomous driving in urban environments: ap-proaches, lessons and challenges
Philosoph. Trans. Royal Soc. A
Capability as requirement metaphor
Requirement analysis for building practical accident warning systems based on vehicular ad-hoc networks
Modeling enterprise capabilities with i*: reasoning on alternatives
A proactive, knowledge-based intelligent transportation system based on vehicular sensor networks
IET Intell. Transp. Syst. J.
A holistic framework for embedded safe and connected automation in vehicles
Decentralized environmental notification message
Integration of domain-specific simulation models into descriptive system models by using SysML
Chauffeur assistant: a driver assistance system for commercial vehicles based on fusion of advanced acc and lane keeping
Smart cities: ranking of european medium-sized cities
E-government success factors: Mapping practical tools to theoretical foundations
Govern. Inform. Quart.
Integrated description of functional and non-functional requirements for automotive systems design using SysML
Mobility as a service – the new transport model?
Eurotransport
Strategic business modeling: representation and reasoning
Softw. Syst. Model.
Feasibility study for “mobility as a service” concept for London
Modelling of organisational change using the EKD framework.
Commun. Assoc. Inform. Syst.
Cited by (17)
Identifying concerns when specifying machine learning-enabled systems: A perspective-based approach
2024, Journal of Systems and SoftwareRequirements engineering framework for human-centered artificial intelligence software systems
2023, Applied Soft ComputingRequirements practices and gaps when engineering human-centered Artificial Intelligence systems
2023, Applied Soft ComputingRequirements engineering for artificial intelligence systems: A systematic mapping study
2023, Information and Software TechnologyRequirements Engineering for Cyber Physical Production Systems: The e-CORE approach and its application
2022, Information SystemsCitation Excerpt :Reflecting upon the definition of capabilities in the fields of those of strategic management and frameworks it behoves that the concept of capability is a most suitable metaphor when considering requirements for major transformations from PPS to CPPS. The conceptual modelling framework underpinning e-CORE has been developed and applied in recent work [39–41]. It employs a set of complimentary and intertwined modelling paradigms based on enterprise capabilities, goals, actors, and information objects.