Intelligent Social Networking in CPS

Organization of the Chapter

the conclusion of this chapter.

Section 1 provides the terms and terminologies, to help the reader to understand this chapter. Section 2 presents an introduction to cyber physical social systems, highlighting the applications, challenges and opportunities that exist in this realm. Section 3 outlines IoT in the context of social networks. Section 4 discusses community detection, with the emphasis on smart communities. Section 5 presents the link prediction models utilized in CPS and provides a brief overview of the different metrics. Finally, Section 6 presents

Terms and Terminologies

Cyber Physical Systems (CPS): A confluence of cyber systems encompassing computers and communication devices together with physical systems consisting of sensors, actuators and users.

Social networking: Using Internet-based social media sites to stay connected with friends, family, colleagues, customers, or clients.

Smart community: A community consisting of government, business, and, residents who decide to use information technology to transform life /work in their region.

Sensors: Devices that measures a physical property and responds to it.

Social computing: Aims to recreate social conventions and social contexts using software and technology.

Community Detection: Identifies highly connected groups of individuals or objects inside a network.

Introduction

Cyber physical systems (CPS) are a confluence of cyber systems, encompassing computers and communication devices, together with physical systems consisting of sensors, actuators and users. The interactions between these systems are tightly integrated at multiple scales and levels by a computing and communicating core. The operations of CPS are monitored, controlled and co-ordinated in realtime, using computer-based algorithms securely linked through the Internet (Ashibani et al. 2017; Lee et al. 2015; Rajkumar et al. 2010), and consist of a feedback loop, as shown in Figure 12.1. In CPS, the sustainability, safety and efficiency of interactions between the physical and cyber world is achieved by actuation of unconventional objects, such as machines and switches, using the Internet (Negenborn et al. 2014; Sisinni et al. 2018). The term CPS initially originated in the USA in the year 2006.

As shown in Figure 12.2, CPS connects the cyber and physical world through three key enablers, namely, computation, communication and control (Gao et al. 2013). This is widely referred to as the 3C architecture of CPS (Liu et al. 2017; Xiong et al. 2015). The development of the CPS domain requires an integration of expertise across engineering disciplines, such as human- computer interaction, learning theories, material science and biomedical engineering. Traditionally, physical processes and control flows are modelled using differential equations and automata, respectively. Whereas such modelling approaches may suffice for component level associations, these approaches exhibit shortcomings during system-level physical and behavioral interactions between distinct components in CPS (Baheti et al. 2011).

FIGURE 12.1

Components of CPS

In spite of being an emerging topic, the system components of CPS are well defined, as depicted in Figure 12.3. The physical world, interfaces and cyber systems constitute the components of CPS. The devices and entities that will be monitored or controlled are designated as the physical world. Embedded devices, incorporating cutting-edge technology which can exchange data with their distributed environment, are termed cyber systems. The intermediate components, such as the sensors and actuators, analogue-to-digital (A/D) and digital-to-analogue converters (D/A), which aid the cyber systems to communicate with the physical world, are referred to as interfaces. Sensors and actuators are widely utilized in CPS to reciprocally convert from a multitude of energy forms to analogue signals in the form of electricity (Gunes et al. 2014).

The forthcoming sections present the need for CPS and its various characteristics.

Need for CPS

The advent of the World Wide Web in the 1980s-1990s led to an exponential increase in the usage of computers for day-to-day applications. Simultaneously, advances in hardware, such as shrinking memory spaces, better graphical interfaces, affordable sensors and higher computational speeds, resulted in extensive, innovative applications, such as digital libraries, online communities, e-commerce and social networks (Rajkumar et al. 2010). These advances were aided by the rapid growth in technology, such as sources of renewable energy, augmented internet bandwidth, portable laptops and smartphones. The proliferation of the Internet, coupled with technological advances, have resulted in the demand for linking the cyber and

FIGURE 12.2

3CArchitecture of CPS

physical world in domains such as aerospace, transportation vehicles, industrial automation, defence systems, safety-critical processes, medical devices and healthcare. CPS can therefore be envisioned as embedded systems with real-time high-power computing capabilities, blended with distributed sensors and controls. CPS have wide-ranging applications, from compact heart pacemakers to expansive power grids.

Characteristics of CPS

In order to perform functions, such as collecting data, monitoring and controlling systems, a CPS should pass with respect to specific issues, namely flexibility, security, usability and privacy concerns (Xiong et al. 2015). Consequently, a system is required to possess the following characteristics (Liu et al. 2017) to be categorized as a CPS, as shown in Figure 12.4.

Physical System as the Principal Component

Physical systems consist of perceptive devices with sensing, computing and wireless communication capabilities, such as sensors, Global Positioning System (GPS), Radio Frequency Identification (RFID) tags and actuators.

FIGURE 12.3

Holistic view of CPS

The physical system is the principal component of any CPS, as it is responsible for recognizing and interpreting real-time data collected from the physical world (Ashibani et al. 2017). The data, gathered by different sensors in the form of sound, vibration, temperature, location, humidity etc., is then transmitted via communication networks such as Infrared, 4G internet, Bluetooth and WiFi, in order to be assimilated and analyzed by the information system. Due to the wireless nature of these networks, reliability of connections and, therefore, predictability of interactions is an issue in CPS. Particularly, the use of mobile devices leads to great uncertainties in the network. For instance, connected automobiles in a vehicular network may interact between one another only when they are at close range.

Integration with Information System

The information system can be envisioned as the central hub of any CPS, as this system should be competent to handle the large volume of data

FIGURE 12.4

Characteristics of CPS

generated by a multitude of physical systems, and simultaneously provide real-time feedback. Additionally, information systems should possess the ability to handle the various tradeoffs, such as network latency, memory management, data validation and reconciliation (DVR) and hierarchical storage management (HSM). A close integration between the physical and information modules is achieved, using embedded systems.

Homogenization of Heterogeneous Systems

A CPS is not a stand-alone system and is an intelligent network of heterogeneous distributed systems, such as sensors, high-power servers and handheld mobile devices. In a large-scale CPS, with components distributed world-wide, spatiality and time synchronization are of paramount interest during integration and interaction among heterogeneous systems.

Robustness and Security

In contrast to online systems integrated over the internet, CPS encompasses a host of hardware, users and software systems, communicating with each other. This necessitates infallible integration at a deeper level. Therefore, robustness and security are key requisites of CPS. Robustness is the ability of a processing system to deal with errors during input and execution. In a network, data authentication, validation and verification of the sending device's identity are the principal factors to be examined during data input. The receiving device must decipher the true identity of the sender to avert bogus data being transferred and executed. Hence, security of the information transmitted over the network should be ensured, using appropriate encryption and decryption algorithms. Another aspect of a robust system is the ability to predict requirements and to scale resource allocations to fulfil real-time tasks. In essence, a large-scale CPS should possess the ability to dynamically adapt, reconfigure and reorganize, in addition to manifest attributes, such as robustness, security, reliability, interoperability and credibility.

CPS have been extensively deployed in cross-domain applications involving autonomy, interoperability and adaptability, such as entertainment services, industrial mass production and intelligent transportation systems. Furthermore, CPS can be envisioned as the elementary form of the Internet of Things (IoT) since, in CPS, real-world objects are mapped to the cyber world as cyber entities (Wang et al. 2015; Ning et al. 2016).

Introduction to Cyber Physical Social Systems (CPSS)

As mentioned earlier, CPS is an indivisible integration of cyber and physical systems. These systems are designed and built to interact with humans. The contemporary surge in the utilization of social networks generates an abundance of user-specific data, which can be integrated with CPS to provide large-scale, real-time services, such as air pollution monitoring, venue recommendations, smart parking systems, smart cities and smart homes. This paradigm of taking into account the social dynamics as a part of CPS is referred to as Cyber Physical Social Systems (CPSS)(Amin et al. 2019). CPSS can be perceived as a three-dimensional space, incorporating close interaction and involvement of humans with the physical and cyber space. CPSS act as the fundamental enabler of social computing. In CPSS, the physical data shared by users in a network is collected using numerous sensors. In conditions, such as commenting and recording preferences, the users themselves act as sensors (Zheng et al. 2017). Social computing typically consists of two aspects, namely dynamic social interaction and sensing social phenomena. Whereas social interaction pertains to interpreting data gathered from immense online interactions among users, sensing social phenomena refers to analyzing and comprehending the pattern of interaction among users (Zeng et al. 2016). A CPSS, while amalgamating the four realms of information, physical, social and cognitive domains, also performs auto-synchronization and parallel processing. Therefore, CPSS are utilized in applications involving real-time command and control, such as military operations (Liu et al. 2011). In addition to the physical entities in CPS, social attributes, such as affiliations and ownerships, are appended, leading to an interfusion of CPS and social realms to form CPSS, an enhanced version of IoT (Ning et al. 2016).

CPSS consists of a deeper integration and semantic interdependence between physical layers, composed of sensors, and collective intelligence, signified by futuristic cognition to blend machine and human perceptions (Sheth et al. 2013). A major concern when supplementing the existing CPS with social data, is the breach of privacy of users. When blending physical and social systems, confidential data such as individual habits, user location, stay and travel details, are evidently shared through the cyber world. Unlike privacy in social networks, long-term physical profiles and behavioral patterns of users are expected to be privacy preserved in CPSS. Handling of the enormous data thus generated, is one major challenge in CPSS (Zheng et al. 2017).

Applications of CPSS

CPSS have been beneficially applied to numerous fields. Table 12.1 summarizes the field and scale of application in various sectors.

Challenges and Opportunities in CPSS

Several challenges are yet to be addressed for large-scale implementation of CPSS to solve day-to-day problems. Table 12.2 summarizes a few of the problems outlined in the literature.