A Model-driven Framework for the Analysis and Verification of Smart Parking Systems

Abstract

A few decades ago, the swift increase in population and developments in technology caused an extraordinary increase in vehicle population. This leads to several problems including environmental pollution, traffic congestion, unnecessary time consumption, and fuel wastage. To tackle these issues, smart parking is an emerging area of research. The smart parking solutions are of real-time nature where a prompt response is essential and therefore, the architecture of Fog computing is frequently utilized. Moreover, it is necessary to analyze the diverse nature of resources and their characteristics before system development, so that, effective and economical Fog architecture can be recommended for a particular smart parking system as per requirements. Hence, Resource assessment is one of the major concerns to ensure the QoS measurements for Smart Parking Systems. Although different solutions have been proposed by practitioners to assist the developments in Smart Parking Systems, the existing solutions usually target typical parking system features while overlooking the early resource assessment. In addition, existing tools for the analysis of hardware resources operate on lower abstraction level with higher complexities. Therefore, there is a strong need to develop a framework at a high abstraction level to allow modeling of resources to facilitate early analysis. In this thesis, a Model-Driven Smart Parking Transformation Engine (MSPTE),) is proposed by utilizing the concepts of Model-Driven Engineering (MDE). In particular, a Unified Modeling Language (UML) Profile (i.e., UML Profile for Smart Parking - UPSP) is developed by extending UML-meta-model to adapt the concepts of smart parking system framework. This allows to model parking system functions along with resources. As a part of framework, a transformation engine named Model-Driven Smart Parking Transformation Engine (MSPTE) is developed to transform high level USPS models to low level java code. A model-to-text approach is implemented to develop the transformation engine using Acceleo tool. The generated code from transformation engine not only supports smart parking system development but also provides early resource assessment capabilities through iFogSim tool. Lastly, the framework is validated with the help of two benchmark case studies. The experimental results indicate that the proposed framework offers an easy and reusable solution for smart parking system development.