Though smart cities are facing many problems during their development, including socioeconomic and political issues, but the most important hurdle here is the technical issues. In technical problems, along with the other issues like system interoperability and cost-efficient technology, the concern of security and privacy is very important. All these factors motivate us to take up this survey work. Initially, we define the need for smart cities along with the inherent challenges faced by them. We also review the different architectural approaches for smart cities. Further, we provide detailed description of the real-life implementation of such existing schemes. Finally, we give an insight into the possible directions for future work in smart cities.
This chapter has been organized as follows. In Sect.5.2, we briefly discuss the various architectures that were proposed for undertaking the works on smart cities.
In Sect.5.3, we study existing works on real-life applications ranging from food management to energy management andfinally transport management. Section 5.4 presents a summary of the key issues and challenges in building intelligent systems for smart city. Section5.5draws possible open research issues. Finally, Sect. 5.6 concludes this chapter.
5.2 Architectural Approaches for Smart City
responsible for creating a virtual environment of the cities through 3D spaces and 2D maps. Thefinal layer is the interaction layer where agent systems are used for communicating with each other.
In another interesting work, the smart city of Trikala in Greece was developed by Anthopoulos et al. [15], where the architecture is made up offive layers. The first as well as thefifth layers are user layers that consist of the stakeholders of a smart city, i.e., the designers of the services and the end users. The infrastructure layer is the second layer consisting of technologies, platforms and networks that are needed to create and offer the services. The third layer is named as the information layer that contains the necessary data such as geospatial data needed for operations in smart cities. The fourth layer is named as the service layer that consists of the application provided for the city and allows for interaction between citizens and organizations.
In contrast, Su et al. [16] proposed three stages for the building of smart city.
Thefirst stage is the manufacture of public infrastructure while the second stage is the manufacture of public platform. The second stage consists of network infras-tructure, cloud computing platform and Wireless Sensor Networks (WSNs).
Finally, the third stage is the manufacture of application systems such as, some basic applications like construction of wireless city, smart home, etc.
Recently, Carretero [17] developed a self adaptive system for smart city using an architecture named ADAPCITY. It provides heterogeneous devices with the capability of reacting effectively in different environments. Also, this system has the ability of immediate recovery as well as updates and creates new operations. Here, the architecture is divided into four layers. First layer, the physical layer, consists of the state and behaviour of the devices and objects. Second layer, the grid layer, is responsible for the process, storage and communication among the data coming from the physical layer. Third layer, the management layer makes use of statistics, data mining and prediction techniques for managing the processed data from the grid layer. Thefinal layer is the control layer that includes the provided services, taking into consideration the desires of the account users and optimization measurements.
Different from earlier works, Vilajosana et al. [18] proposed a generic archi-tecture by combining the common features of several existing platforms. The bottom layer of the platform named as capillary network layer consists of sensors and actuators needed for data collection, data warehouses for storage of historical, real time and metadata. The service layer receives the incoming data from the capillary network layer. It then processes, combines and secures the received data.
It manages different types of data, such as big, open and streaming data and also provides analytics services. The last layer is the application layer where the data are analyzed and converted into useful information, which is eventually provided to people through predefined interfaces.
Apart from the aforesaid works based on architecture, IBM has also defined the structure of smart city based on three layers, viz. perception, network and appli-cation. The perception layer is responsible for recognizing the device and gathering of data using sensors, GPS, RFID, etc. The second layer is the network layer that
processes the data obtained by the perception layer using components related to the intelligence and communication capabilities of the network. While the third layer which is the application layer examines and evaluates the total amount of data through advanced technologies, such as cloud computing and fuzzy techniques.
5.2.2 Service Oriented Architecture (SOA)
In SOA, the primary objective is collection, communication and interaction between services. The communication between different services in a computer system is implemented by data exchange among them. Every interaction is con-sidered to be unconstrained as services are unrelated, loosely coupled and self-sufficient.
In one state-of-the-art work, Anthopoulous et al. [19] developed a common architecture for smart cities based on SOA, named as enterprise architecture. It contained information regarding urban development and service delivery in such environments. The enterprise architecture combines the logical and physical architectures. The authors developed this architecture keeping in mind the draw-backs of the architecture that was used for developing Trikala as a smart city. This architecture was used to overcome the problems faced by Trikala.
5.2.3 Event Driven Architecture (EDA)
In EDA, creation, identification utilization and response to events are handled. The events that are dealt here with are generally uncommon and are related to uncertain modifications and asynchronous conditions. This architecture produces results that invoke production of notification of events. For example, a change is detected by the sensors and the events resulting from this change are processed by the system.
This architecture can also be combined with SOA. For example, in one work [20], Filipponi et al. designed the SOFIA project by combining the architectures EDA and SOA. This project was developed for monitoring cities for security threats and also for detecting emergency and abnormal situations.
5.2.4 Internet of Things (IoT)
In IoT, a number of heterogeneous devices are connected to the Internet and they identify themselves using IP addresses and protocols. Here, the devices are embedded with sensors and actuators and are wirelessly connected to the internet.
IoT provides for connectivity and communication between the sensors to facilitate
various applications for users. IoT-based architecture has played a prime role in the coming up of smart cities. Some of the works that have dealt with IoT in smart cities are described as follows.
An infrastructure for smart city termed as smart city critical infrastructure is developed by Attwood et al. [21]. The purpose of this infrastructure is protecting critical infrastructures from failure or assists the system to recover and continue functioning if a failure was inevitable. For operating all the functions mentioned previously, sensor actuator networks are very much essential. These actuator net-works connect themselves to IoT for collecting data needed by the smart city. The basic elements of smart city critical infrastructure developed by researchers such as smart cities systems annotation and aggregation service, critical response reasoning instance, sensor actuator network overlay state management, etc.
In [22], Asimakopoulou and Bessis concentrated on disaster management using crowdsourcing techniques to create smart buildings. Crowdsourcing technology is used for detection of emergency events and hazards by citizens using APIs in their mobile phones. The authors also proposed other technologies such as grid com-puting to integrate heterogeneous resources, cloud comcom-puting to enable access to these resources and pervasive computing to collect and handle data from devices.
In another work, Wang et al. [23] demonstrated the use of world wind geo-graphic software developed by NASA for reconstructing a city. The software is an open-source platform that allows visualization, simulation and interaction in all living sectors of a smart city. The two main components of this technology are data collection and visual display. The data are collected through IoT, network analysis and web map services.
Samaras et al. [24] developed a smart city platform SEN2SOC for implemen-tation in the SmartSantander City of Spain. The objective was to increase the interaction between sensor and social networks using the system of natural lan-guage generation (NLG) for improving the standard of living of the citizens and visitors in a smart city. The architecture of SEN2SOC platform is component based and includes mobile and web applications, sensor and social data monitoring, statistical analysis and interface. The NLG system is embedded in the platform and has the capability of receiving information from sensors and converts it into mes-sages that can be easily understood by humans.
In [25], a smart parking system is proposed by Horng. The proposed system enables the people in locating parking spaces easily and quickly, which eventually helps in reduction of fuel congestion and air pollution. The proposed smart parking system uses WSNs for searching the presence of vehicles near a parking space. In smart parking system, an internal recommendation mechanism of the specific place informs the Parking Congestion Cloud Centre (PCCC). The PCCC transmits these data to the cloud server. The user ultimately receives the required information through his/her mobile device, which also acts as a sensor for the cloud server at the same time.
Besides the above-mentioned architectures, some works are undertaken by combining the features and technologies of the different architectures mentioned
above. Examples of such architectures include combinations of IoT-AL [26], IoT-SOA [27,28], IoT-SOA-AL [29], IoT-EDA [30], etc.