A systematic review to aligning research paths: Energy cyber-physical systems

Abstract Cyber-Physical Systems (CPS) is a new generation of digital technology that is concerned with the integration and inter-dependencies of the cyber and physical world alongside computational elements. As a new and leading technology, its applications are seen in different projects involving energy and this has however generated lots of interest from the industry, researchers in academia and the government. This paper presents an extensive overview and modernistic research on the applications relating to energy CPS and the security issues and challenges revolving around this research path. In order to achieve this, a systematic literature review was carried out which enabled the analysing and classifying of different applications and security issues described in selected publications. Furthermore, the systematic review permitted the discourse stringing these research areas as well as providing future lines of research. Also, results from the review show the paths where there are increasing research focus and expected research trend in the future years to come. These results will not only be useful in serving as a guide to researchers but will also create new research paths to more experienced researchers to actively follow.


PUBLIC INTEREST STATEMENT
Cyber-physical systems is a type of heterogeneous system that integrates computational elements with physical properties to ensure a common goal.The interaction between the physical components and the cyber or computational components is done by embedding computers and networks to control and monitor the physical processes.In this paper, an extensive and systematic literature review is carried out on the applications of energy cyber-physical systems and the security issues in the system.The results from this research will assist and enable scholars and researchers to learn new approaches and pave research paths in this research field.

Introduction
From years past, researchers in the field of controls and systems have been working on developing efficient and resourceful tools and techniques for evaluating, examining and controlling physical processes, while researchers in the field of computing and electronics have also been trying to achieve a similar feet in cyber systems (Baheti & Gill, 2011).Nevertheless, there still lie a huge disconnect between the physical world where energy and mass transformation takes places and the cyber world where there is transformation and exchange of data and information (Dillon & Chang, 2010).Recently, there have been a massive revolution in the Information Technology (IT), computing and communications sectors where computing gadgets and devices are now more efficient, smaller and faster, communications and communication networks are now less expensive, faster and better.This no doubt have also had a significant impact on energy systems.To further buttress this transformation, objects, structures and entities in the physical environment are now equipped with communication and computing capabilities as well as the components of energy-related systems (Gilroy, 2010).The merger of cyber and physical systems with energyefficient capabilities that possesses the features of flexibility, security, reliability, sustainability and efficiency poses new challenges to researchers and engineers in this field (Farhangi, 2010;Ilic, Xie, Khan, & Moura, 2010;Lemmon & Venkataramanan, 2009).Hence, a system that possesses cyber abilities integrated into each physical components and all unified to form a large Networked Control Systems (NCS) with efficient control and modelling methodologies can be seen as an alternative tool to resolve these challenges.That means an ideal methodological combination of both systems (physical and cyber) is the main ingredient in forming a Cyber-Physical System (CPS).
CPS can be termed as a new generation of digital systems that combines functional components of the cyber and physical world.It consists of advanced connectivity that guarantees the process of acquiring data from the physical world as well as feedback information from the cyber space with intelligent data computational, analytics and management abilities that constructs the cyber space to ensure a common goal (Baheti & Gill, 2011;Lee, 2008Lee, , 2015)).

Concepts, features and characteristics
CPS has categorically been in existence since the early '70s when the first microprocessors were invented and put into use (Wolf, 2009), but it was not until the year 2006 at the General National Science Foundation meeting, in the USA that Helen Gill coined the name Cyber-Physical System as a term used to describe a system connecting both the physical and cyber world (Gill, 2006).After then, CPS has continued to be an emerging multi-disciplinary area which includes the juxtaposing, overlapping and involvement of various fields in engineering and science.The concept of CPS is so broad and encompassing that it allows an automation and control engineer, a mechanical engineer, chemical engineer, civil engineer and a biologist to collaborate closely with computer scientist and network professional experts to model a workable system.Due to this, the definition of CPS is given by how these different scholars view it in their own disciplines, perspectives and scientific communities.
In the paper by (Kim & Kumar, 2013), CPS is defined as the next generational engineering systems that involves the combination of communication, hi-tech computing and control technologies to accomplish high performance and efficiency, reliability, robustness and stability in physical systems applications involving transportation, medical, defence and energy.Lee (2007) also described CPS as a scalable, controllable and credible network of physical equipment systems that is deeply integrated with control, communications and computational abilities on the foundation of environmental perception.In Ding, Han, Xiang, Ge, and Zhang (2018), CPS is expressed as a heterogeneous, geographically dispersed, life-critical and large-scaled systems incorporated with devices like actuators and sensors to enable the system monitor, control and sense the physical world.Also in Gunes, Givargis, and Vahid (2014), CPS is defined as "a complex, physically-aware next generation engineered systems that integrates embedded computer technology (cyber part) into the physical phenomena by using transformative research approaches.This integration mainly includes observation, communication and control aspects of the physical systems for multi-disciplinary perspective".It however can be very safe to say that several research branches and areas are closely linked or interconnected to CPS. Lee (2015) outlined the differences and similarities between CPS and several research terminologies and fields including Cybernetics, Machine-to-Machine (M2M), Big Data and Internet of Things (IoT).
CPS normally consists of three major components which are the physical system, networking and communication elements and the distributed cyber system.In a typical CPS, the distributed software component plays a key role as it is responsible for processing, filtering and storing information.CPS operates in a three-layer mode which are perception, transmission and application (Ashibani & Mahmoud, 2017).The perception layer which is sometimes regarded as the sensory layer are embedded in devices such as cameras, GPS, readers, actuators, sensors and RFID tags to enable the collection of real-time data from the application layer.The transmission layer which is also referred to as network layer assists in the processing and interchanging of data between the processing layer and application layer.This is achieved with the aid of several communication technologies like Zigbee, Infra-red, Bluetooth, Wi-Fi, LANs or even the Internet.
In the application layer, information obtained from the transmission layer is processed and commands are issued to the actuators and sensors for execution which in turn enables a smart environment.
There exist several characteristics of CPS such as mass computing, dynamic reconfiguration, large-scaled diversified networks and highly automated control circuits, but the major characteristics of CPS comprise of reliability, scalability and distributed real-time applications involving realtime control, real-time forecasting and real-time monitoring.
Energy in Cyber-Physical Systems is basically looking at CPS in the energy point of view or the integration of computing, communication and control in the energy domain.This is seen as an important class of CPS as it contributes significantly to the economic, environmental and the social outlook and perspective of CPS.It encompasses the generation (renewables and conventional), distribution and transportations as well as energy storage of a system (Stamatescu, Stamatescu, Arghira, F˘ag˘ar˘asan, & Iliescu, 2014).There exist plenty of research interest on how energy in CPS can be integrated and deployed in various systems.Hence, the main aim of this study is to review cyber-physical systems relating to energy with the objective of providing valuable insight to its applications and security challenges.Our key motivation for this research article is to assist researchers with well-detailed and upto-date information that would enable them access this field easily and also to enlighten experts in the field on what has been developed and accomplished so far and with a possible avenue to creating future research paths.The remainder of this paper is organised thus-section 3 presents the methodology used for this study while section 4 details the findings from our analysis of different publications.In section 5, discussions arising from the findings in section 4 are presented and section 6 concludes the paper.

Methodology
For this review, the Kitchenham's systematic literature review methodology or procedure (Kitchenham et al., 2010) is used.This procedure is often used in software engineering researches and has been very successful in this regard.Due to its success and easy implementation, other fields like education, economics, management and nursing have adopted its usage.In conducting our research, we address our study based on the following actions in this order- When conducting research, the very first process is to formulate research questions that entail the basis and focus of the study.These questions are rooted from the analysis of the different issues expressed in different works in the research field.The academic research papers and paper reports examined in this article have been read in order to answer the two formulated research questions deduced for this study.
RQ1.What are the applications of cyber-physical systems as it relates to energy?RQ2.What are the security issues and concerns in energy cyber-physical systems?
The next process is the search process.The strategy adopted for searching the literature for this research is to systematically search citation databases and relevant digital libraries using keywords and search strings.The keywords combined to create a search string for these research are cyber physical system, cyber physical energy systems, energy cyber physical system, application of cyber physical systems, security cyber physical systems, issues in cyber physical system, energy CPS, security CPS and cyber physical systems issues.
The inclusion and exclusion criteria for this review article are the criteria for including and excluding research papers from the study in order to refine the search.For our inclusion criteria, papers that are related and relevant to any of the keywords used in the search are included.Papers from January 2010 to July 2019 were considered and included in this study.While for the exclusion criteria, papers that were not written in English were excluded.Also, papers that do not have concrete evidence to back up the research or are not clear enough or have focused entirely on another topic area were excluded.Papers that do not deal explicitly on CPS were also excluded.Table 1 shows the trend and statistics of publications on CPS applications and security.
In the next process, data is collected and analysed from the selected papers.The data collected from these papers are analysed applying a standard template across the board.Afterwards, the work is then classified and categorized using the research questions previously formulated.

Results
In this section, we evaluate the findings of the literature review carried out.We start holistically and then begin to narrow down to the concentration of our study by categorizing them into different parts depending on their area of application or security.
4.1.Energy applications in cyber-physical systems RQ.1 means that this study is to concentrate on the applications of cyber-physical systems in an energy perspective.For this purpose, the literature previously gathered was broken down into different parts to enable us to identify the applications of CPS.It is identified that there are various applications of CPS as it applies to different areas of research.Most of the areas found in the literature that studied the applications of CPS are in agriculture, medicine and health care, process control, intelligent transportation and smart cars, smart grid and smart cities, smart manufacturing, environmental monitoring, energy management and robotics.We will, however, focus on the applications related to energy and group our research findings into different categories.

Smart grid and smart cities
Smart grids are intelligent types of electric network that uses modern technologies to deliver a more reliable and efficient energy supply.Smart grids combine power network infrastructure which we can regard as the physical system with sensors, actuators, and the ICT and communications aspect which we can regard as the cyber system to deliver an intelligent and smarter type of electric network.The publications in this area of research are seen in Table 2.
The paper by Yu and Xue (2016) evaluates the challenges associated with smart grid in the cyber-physical system context.They also identified the possible and likely contributions that CPS can offer to smart grids and showed an architectural design system that incorporates communication, computation and control for the deployment of a CPS.Also, open questions were also raised and these questions will affect the future development of CPS and smart grids.In the paper by Cintuglu, Mohammed, Akkaya, and Uluagac (2017), the authors present a range of testbeds relating cyber-physical systems to smart grid with an aim to providing important features, a wider categorization and design decisions for its development.The paper went on to categorize and classify the cyber-physical smart grid domain into four steps and investigated them individually according to their various criteria.It was also concluded by the authors that improved communication infrastructure in the system is very vital as its data communication, privacy and security will give a better opportunity to combine testbeds with different capabilities.Smart Grid Testbeds (Cintuglu et al., 2017(Cintuglu et al., ) 2017 Social smart grids (Cheng et al., 2016) 2016 (Zhang et al., 2018(Zhang et al., ) 2018 Software associated smart grids (Horcas et al., 2019) 2019 (Xin et al., 2017(Xin et al., ) 2017 Smart cities (Cassandras, 2016(Cassandras, ) 2016 Smart home (Seiger et al., 2016a) 2016 (Mazumdar et al., 2016) 2016 (Seiger et al., 2016b) 2016 The only two models that studied energy in cyber-physical systems in a social view point was found in Cheng, Zhang, and Yang (2016) and Zhang, Xu, and Yu (2018) as also documented in Table 1.A consumer-centred energy system is proposed in Cheng et al. (2016) where a cyber-physical social system combines a physical component which is the power grid and the cyber component which involves computing, communications and consumer interactions (social).The importance of consumer-centred energy system with respect to effective energy utilization was also discussed and an architecture capable of achieving flexible and intelligent energy management for smart grids was also presented.The work in (Zhang et al., 2018) also presents a novel cyber-physical social system with parallel learning for microgrid systems.The work was carried out to solve the human participation and interaction in microgrids.A case study of a microgrid with 11 energy suppliers and 7 energy consumers was used and the technique showed that computer-human collaboration can achieve a higher quality of distributed energy management when compared to other centralized heuristic algorithms.Farraj, Hammad, and Kundur (2018) presents a cyber-physical control framework for transient stability in smart grids.In this work, a parametric feedback linearization (PFL) controller is utilized to solve the issues associated with the stability of power systems.The PFL controller uses an external vitality capacity framework to absorb and infuse power from the system in order to ensure the stability of the rotor speed.In the paper by Allen, Liu, Lozano, and Yuan (2012), it was confirmed that smart grid is critical in terms of infrastructure and a decentralized, hierarchical framework was proposed to detect and mitigate potential cascade failures.The authors experimented the technique in real life and it was deduced that cascading failures are initiated by interactions developed between contingencies, protective relays and power flow switching among and around the studied zones.
The paper by Horcas, Pinto, and Fuentes (2019) addresses the problem of developing context-aware energy-efficient application for CPS with the aid of a green eco-assistant that utilizes innovative software engineering techniques.This assist software inventors and developers to be increasingly aware of the energy consumption in the associated hardware systems.Also in Xin et al. (2017), an information-energy flow model is developed using a matrix-based computational approach.The energy information flow was calculated in an IEEE 14-node system to help estimate the impacts of various cyber contingencies on a power system.
Smart cities and smart homes are seen in literature to use the CPS applications for innovative facilities like energy distribution, emergency response services, businesses, commerce and so on.Four papers were found that linked energy cyber-physical systems with smart cities and smart homes.The paper (Cassandras, 2016), identifies the main defining characteristics of a smart city as it relates to CPS and considered the smart city as a CPS with a new software platform.In Seiger, Huber, and Schlegel (2016a), an integrated system called PROtEUS is designed for process implementation in CPS.This integrated system unifies components for event processing, human interaction, dynamic service selection and data routing.Two case studies of a smart home domain are used to test its feasibility.Mazumdar and other authors in Mazumdar et al. (2016) introduce a project named AXIOM with an aim to developing a hardware-software platform for CPS to enable easy parallel programming and to easily scale up the performance of the system by adding multiple boards.Smart home living applications and smart video surveillance are used to achieve this design.Lastly, the authors in Seiger, Struwe, Lemme, and Schlegel (2016b), developed a mobile control centre that is user-friendly for cyber-physical systems in the smart home domain with the aim of lessening the hitches in utilising CPS with numerous control options and enabling the system to be more user-friendly to non-expert users.

Intelligent transportation and smart cars
Energy can be conserved and utilized efficiently through the use of smart cars while safety, traffic management, coordination and throughput can be improved through the sharing of real-time information using intelligent transportation systems.In these systems, advanced technologies involving computational software, communication, sensing and control mechanisms are all imbedded in transportation systems like cars, trains and aeroplanes to reduce traffic fatalities.The involvement of CPS in these areas has been found in 14 papers in literature as documented in Table 3.
The papers (Bradley & Atkins, 2015;Kumar, Bali, Iqbal, Chilamkurti, & Rho, 2015;Ma, An, Huang, & Cao, 2015;Wan et al., 2014), all examined the concept of vehicular cyber-physical systems.In Bradley and Atkins (2015), cyber-physical vehicle systems is presented where the authors examined various factors like sensor scheduling, task and motion planning, resource sharing, time-varying sampling patterns and feedback scheduling to co-optimize and co-regulate applications in CPS vehicular systems.In Kumar et al. (2015), a stochastic coalition game in a typical vehicular CPS setting is proposed where the vehicles acting as players in the game receive a limited number of resources from the cloud.Also in Ma et al. (2015), Dempster's rule is utilized to improve and resolve the problem of sensor data fusion in vehicular CPS.The results obtained showed that eliminating the issue of conflicting evidence combination yields an improved convergence performance.
The author (Kantarci, 2015), initiates a minimum Steiner tree approach for path selection for paramedics with the aid of cyber-physical systems consisting of human interaction components and vehicular communications.This approach yields an alternative routing selection for paramedics with minimal cost, minimal delay and increased resilience.Traffic measurements in CPS-based transportation are documented in Zhou, Mo, Xiao, Chen, and Yin (2016) and Zhou, Xiao, Mo, Chen, and Yin (2013) where measurements are obtained from various vehicles commuting from one geographical location to the other with the aid of an intelligent cyber-physical road system that is capable of collecting traffic data.In Wan, Chen, et al.(2013), it was proved that CPS is an evolution of Machine2Machine with additional interactive operations and intelligent software.In doing this, two CPS platforms are reviewed including a vehicle with WSNs navigation and cyber transportation system.In Rawat, Bajracharya, and Yan (2015), a study of how different parameters affect communication in smart transportation CPS was carried out.
Not much work have been carried out in terms of intelligent transportation relating to the aviation industry.Nonetheless, a novel CPS scheme that focuses on the cyber component and the cyberphysical interaction in aviation is proposed in Sampigethaya and Poovendran (2013) and Sampigethaya and Poovendran (2012).The work in Sampigethaya and Poovendran (2013) evaluates the integration of CPS in aircraft management including aircraft performance, air traffic, aviation users, flight deck and so on while the work in Sampigethaya and Poovendran (2012) proposes a new CPS framework for air traffic management and airspace system design to improve interactions in aircraft and airspace.
A Vehicular CPS architecture with the integrated mobile cloud computing system is analysed in Wan et al. (2014) where sensors deployed in the system capture data from a vehicle on a daily basis and store it in a vehicle storage space.A large transportation system based on CPS is studied in Besselink et al. (2016) where an approach sub-divided into layers is used to control platooning vehicles and coordinate transport planning.A CPS for connected car-based e-tourism is presented in Smirnov, Shilov, and Gusikhin (2017) using the cloud to deliver tourist context services, car context services or recommendation services.The system can be integrated into the car information system for planning and scheduling customized visual information and guides based on location.Transportation in CPS including travel time prediction, applications in smart traffic light system and small transportation are all documented in Möller and Vakilzadian (2016).

Industries and smart manufacturing
In the research area of industries and smart manufacturing, researchers evaluate and design architectural platforms to apply CPS in production and manufacturing in order to alleviate problems associated with flexibility, large work force and high energy demand.There are a total of 13 papers found in this area that concentrates on the use of hardware and embedded software technologies to optimize productivity in the manufacturing sector as tabulated in Table 4.The work by Basile, F, Chiacchio, Coppola, and Gerbasio (2015) presents elementary results from designing a suitable distributed control architecture for automated warehouse systems with the aid of functional blocks and cyber-physical systems.The work in Wiesner, Marilungo, and Thoben (2017) introduces a product service-oriented CPS tagged as "cyber-physical product-service systems" and its application in the industry is evaluated.In Monostori et al. (2016), a production design architecture consisting of different levels is analysed.The authors used a cyber-physical module as a case study to research on various production and work applications such as adaptive scheduling, maintenance strategy and automated generation.A self-controlling manufacturing system that sees humans as a component is discussed in Graessler and Poehler (2018) where intelligent controls for manufacturing systems and digital twins approach are proposed for industrial level usage.
Lu and Ju (2017) introduces a service-oriented architecture for industry and smart manufacturing using the CPS concept.The proposed architecture is used to enhance communication between service request and service consumption and provides appropriate conceptual models for describing and developing manufacturing in CPS and enabling its composition.In Agbor, Cao and Ehmann (2018), a conceptual framework that assists in the selection of correct manufacturing process is conceptualized with the aid of a constraint satisfaction problem.It also provide guidelines for implementers in the manufacturing sector most especially for the industry 4.0 revolution.A human-based and industrial based cyber-physical system is introduced in Lemoine, Berdal, Enjalbert, and Trentesaux (2018) where the authors analysed how possible it will be to involve a human being in the cyber physicalbased industry controls.This approach combined industrial CPS and Human-Machine Systems to achieve the goal.In order to deliver an adaptive and autonomous system in machining line in industries, authors in Herwan, Kano, Oleg, Sawada and Kasashima (2018) introduce a CPS production system that will aid tool wear and breakages and also surface roughness monitoring.However, in Novak, Kadera, and Wimmer (2017), the same goal was achieved with the aid of a P'X5 configurator for the Industry 4.0 applications.In O'Donovan, Gallagher, Bruton, and O'Sullivan ( 2018), the research introduces an industrial CPS based on fog computing paradigm using Predictive Modelling Markup Language (PMML) encoded machine learning models for industry 4.0 applications.Also in Hiang (2018), a smart factory based on CPS that includes a smart factory data center based on the cloud environment and an industrial IoT gateway is looked at for the concept of industry 4.0 application.The paper by Kim and Park (2017) proposes a new optimization strategy for manufacturing systems through the combination of CPS and IoT.Lastly, a paper that proposes a scheme for modelling and simulation of CPS-based factories is seen in Weyer, Weyer, Ohmer, Gorecky, and Zuhlke (2016).This framework was applied in the automotive industry for validation.

Environmental monitoring and energy management
In terms of environmental monitoring and energy management, energy can be conserved and managed since CPS is regarded as a distributed system.In environmental monitoring systems, CPS is distributed relatively over a large geographical location without human involvement for a lengthy period of time with minimal energy consumption.In this sense, Man-made and natural disasters can be averted through CPS by instantly responding to various sensors spread over a geographical area.In this research path, seven papers were found to have targeted this area of research as seen in Table 5.In the paper by Sanislav et al. (2014), the authors introduce an open CPS architecture with a high degree of scalability and reliability that can be applied in environmental monitoring.The architecture is broken into three different layers, the bottom layer permits the acquiring of data and their transfer to other layers while the middle and top layer is responsible for the appropriate and correct function of the CPS based on decision rules and complex data analysis.In the work by Sierla, O'Halloran, Karhela, Papakonstantinou, and Tumer (2013), an architecture for modelling communications between CPS and the environment is presented.The approach has to identify component failures emerging from abnormal environmental conditions and also determine the effects with respect to the degradation or loss of the functionality of the system.A CPS for environmental monitoring that encompasses the ecological states such as Table 5. Publications in the environmental monitoring and energy management research area
Managing energy in systems was also captured in some academic papers in literature since a lot of devices and components in CPS require energy for operation.In the paper by Li, Wu, and Li (2016), a dynamic-distributed technique designed for different classes of power systems associated with data attack are presented.In this technique, a dynamic state estimator using a nine-bus power system is used to attain an ideal control of large scale distributed systems.Authors in Wan, Yan et al. (2013) propose a novel CPS energy management application framework that has the ability to collect realtime data of power consumed as well as demand from autonomous electric vehicle and charging centres in a smart grid.The energy-efficient-designed application also incorporates path planning of autonomous electric vehicle in real-time and event-based control using wireless sensor networks.Another conceptual framework of CPS for managing energy in building structures is explored in Zhao, Simoes, and Suryanarayanan (2010).The framework analyses a cyber-enabled efficient building energy management system with a goal of attaining a high level of energy self-sufficiency.The paper by Aksanli and Rosing (2017) conceptualizes the human behavioural awareness for energy management in residential CPS.They showed how human behaviour-based decisions and actions affect residential energy consumption and their results showed high energy savings using the approach.

Security in energy cyber-physical systems
RQ.2 raises the issue of security in CPS as it relates to energy.In as much as CPS is able to utilize the information obtained in the physical domain to generate low cost and energy-efficient functionalities, it can also be prone and vulnerable to threats and attacks.These attacks which can also be in the form of cyber are constantly on the increase since it is a system that promotes connectivity and smart technologies.Security no doubt has earned a lot of attention in the field of CPS due to its important nature and this is evident from the number of papers found in this category as detailed in Table 6.
Issues relating to privacy and information security and trade-off between security and performance in CPS are highlighted in AlDosari (2017).The author presents a common attack known as Side Channel Attacks (SCA) on cryptographic algorithms and the countermeasures against these attacks.In Alguliyev, Imamverdiyev, and Sukhostat (2018), the main types of attacks on CPS are examined where issues relating to CPS components authentication methods, data security and trust levels are analysed and countermeasures were developed.To enable a feasible and extended life time system architecture of a software-defined CPS, the authors in Wu, Luo, Wang, and Wang (2018) proposes a systematic virtual networking architecture and a novel life time extension scheme that performs a global virtualization control and monitoring of a CPS.A theoretic topology decision system was used and the results show that the proposed architecture improves the energy efficiency and life time of the system.In Wu, Kang, and Li (2015) a quantitative hierarchal assessment model that includes attack success probability, attack severity and attack consequence is proposed.The model has the ability to access the risk caused by an attacker on the host and system stratum of CPS.Also in Vegh and Miclea (2016), another hierarchical system of accessing data which is sent through a digital signature algorithm is proposed.This system is designed to equip CPS with an efficient attack detection signature so as to prevent attacks in CPS.
Historical vulnerability data from databases is used to create a framework that characterizes a set of attributes for each given subsystems in CPS in Bakirtzis, Carter, Elks, and Fleming (2018).This framework is used to capture necessary specificity and match possible attack vectors with the model's attributes.The authors in Khalid et al. (2018) propose a security framework that allows the application of humans in collaboration with robots in the industry 4.0 context of CPS.The paper provided mitigation strategies analysed the cyber vulnerabilities in the system and examined the cyber-attack impacts.In Hong, Wang, Ji, and Beyah (2018), the authors propose an attacker location assessment scheme which is based on fake source scheduling (FSSE) for source location

2018
A method for modeling and quantitative examination of attacks in CPS is presented (Orojloo & Azgomi, 2014) 2014 A new method of quantitative evaluation of adversarial attacks on CPS is proposed (Orojloo & Azgomi, 2015) 2015 Mitigation of cyber-attacks on CPS processors with TAIGA (Lyn et al., 2015) 2015 Cyber-attack description language is developed for structured description of attacks on CPS (Yampolskiy et al., 2015) 2015 Game theoretic method application for attacker and defender interactions.Exploring cyber physical testbed developed using engineering models in CPS (Nguyen et al., 2017) 2017 privacy in CPS.This scheme assists in maintaining the system performance and enhancing the privacy level, energy consumption and transmission delay.The author in Axelrod (2013) proposes a technique that assists in determining the factors responsible for the high levels of combined risk in a CPS and recommended appropriate methods to conveniently mitigate such risk.Risk across a broad range of CPS both in the private and public sectors is accessed.The work in Zhang, Wang, and Tian (2013) investigates security threats common to CPS while security measures and recommendations to these threats are identified.Authors in Gao et al. (2014), Rekik, Gransart, and Berbineau (2018) also examine the security threats associated with CPS.In Gao et al. (2014), the threats were examined in the application, physical and network layers with their vulnerabilities listed in a management and policy, platform and network point of view while in (Rekik et al., 2018), the threats were examined in new generation of CPS trains and vulnerabilities listed in terms of train control and monitoring systems.
Trustworthiness in CPS is examined in Boyes (2013) and techniques to evaluate the risks and resilience in complex CPS are identified.Attacks that disrupt information flow such as Stuxnet is investigated in Howser and McMillin (2014).In this work, the authors explain the attack's implication on trust and how the attack can be dealt with since it hides behind the multiple security domain non-deducibility.In Orojloo andAzgomi (2014, 2015), the authors propose new approaches for qualitative evaluation and modeling security attacks in CPS and also model the classes of adversarial attacks on systems.The authors in Lyn, Lerner, McCarty, and Patterson (2015) identify the potential outcomes of cyber-attacks launched on CPS embedded processor and creates a trustworthy autonomic interface guardian architecture that ensures communication between the physical processes and the embedded controller of a CPS.The paper by Yampolskiy, Horvath, Koutsoukos, Xue, and Sztipanovits (2015) introduces a type of description language fuelled cyber-physical attacks which serves as a basis for a structured characterization of attacks on cyber-physical systems while the paper by Puttonen, Afolaranmi, Moctezuma, Lobov, and Lastra (2015) surveys security-related issues in cloud-based CPS and also evaluates the problem of security and threats in smart mobility services.Security approaches in CPS and various threat detectors in a CPS environment are evaluated in Bou-Harb (2016) while authors in Fernandez (2016) describe how threats in CPS can be modeled and how it can be unified to build a secure system.(Sanjab & Saad, 2016), proposes a game-theoretic method to study the interactions of an attacker and a defender in a typical CPS and develop a novel optimal strategy to solve the game putting energy markets and wide-area protection into consideration.Vegh and Miclea (2014), provide solutions for ensuring data security and confidentiality in CPS by combining methods involving steganography and cryptography while Wang et al. (2016) studies the fragility of an electrical CPS under numerous forms of attacks like the false data injection attacks and the denial of service attacks.In the paper Yan, Mestha, and Abbaszadeh (2019), an attack detection scheme was proposed at the physical layer of a CPS by using advanced machine learning techniques or obtaining salient signatures and features from a physical component.To deal with the problem of attackers simultaneously launching multiple cyber-attacks on a CPS, an efficient attack detection scheme base on a random set theory is proposed in Yang, Shi, Zhang, Wu, and Shi (2019).In Lanotte, Merro, Muteanou, and Vigano (2019), a comprehensive model for physics-based attacks, a means of accessing attack tolerance, estimating the impact of a successful attack and its chances and a tool for carrying out static security analysis on a CPS when exposed to attacks is documented.Learning-based attacks in CPS are scrutinized in Khojasteh, Khina, Franceschetti, and Javidi (2019) where an attacker uses an arbitrary learning algorithm to estimate the dynamics of the system.
In the paper by Huda et al. (2017), a novel semi-supervised malware detection system for CPS is proposed.The method is seen to fight against new malware variants without any manual effort or signature generation strategies and also offer advanced protection to CPS.Novel security strategies in CPS are proposed in Rebaï, Voos, and Darouach (2018) and Rebai, Voos, and Alamdari (2018) with the aid of an event-based implementation.Simulation results show that there is an attack-tolerant control in the system despite the infusion of data deception attacks.Authors in Konstantinou et al. (2015) focus on a research on privacy and security challenges at different layers of the composition of a CPS and also provide solutions to guaranteeing trust in a modern CPS.A systematic mapping study is conducted in Nguyen, Ali, and Yue (2017) to identify 48 primary security engineering models in CPS for data extraction and analysis.The results show that more investigation is needed into security solutions for CPS.It is quite interesting to note that cyber security is different from cyber-physical systems (CPS) security.Cyber-physical systems security is more important than cyber security and the latter is often considered as a sub-unit of security in CPS since cyber security is involved in the cyber side of the system and not the physical component part of the system (Northcutt, 2013).

Discussion and future line of research
CPS is constantly gaining more research grounds amongst researchers and this innovative line of research has the potential and ability to revolutionize our relation, communication and interactions with various complex systems which the physical world is largely dependent upon (Kim & Kumar, 2013).Virtually all applications of CPS provide pioneering technologies that are tangible enough to make a great impact in the world.The application categories chosen in this study are highly dependent on energy efficiency, resource reduction or consumption and increasing reliability or performance.
From this study, it is seen that although research on the applications of CPS has made tremendous progress over the years, it is still very new and in a developmental stage.From Figure 1, it can be noticed that lots of researchers have preferably been focusing on the application of CPS in the area of intelligent transportation and smart cars as to when compared to other applications.This is evident from the improved sensing and intelligent transport system and self-driving cars that easily interact with the physical environment.Environmental monitoring and energy management attract less research interest and only a few publications are seen in this field.Studies in various applications of CPS started gaining substantial in 2016 when more researchers from different multidisciplinary fields became involved in this cutting-edge line of research.It is however expected that this trend will continue and more ground-breaking research will emerge from these applications.CPS applications no doubt face numerous challenges especially in the area of efficiency, security and privacy.Due to the characteristic and intrinsic properties, CPS possesses amongst which are dynamicity, complexity and scalability, it is subjected to various attacks both on the cyber and the physical part of the system.Factors that normally contribute to the vulnerability of CPS making them susceptible to threats and attacks are the use of digital connectivity like the internet, unsecured communication protocols, the use of commercial off-the-shelf technologies and dependency on legacy systems (Gunes et al., 2014).Figure 2 shows the cumulative number of publications on the issue of security as it relates to energy CPS research.It is seen that there is a continual increase in the amount of research in combatting security in this field.From the year 2017, there have been a sharp increase in research activity as issues relating to security in large-scaled systems are being treated.
It is worth mentioning that from Table 1, it is observed that there are a lot of conference papers published in this field of research and this implies that most of the ideas are still in the conceptual stage and a lot of them are novel.Based on the publication search process undertaken in this study, it was observed that CPS is currently gaining more grounds with other interrelated technologies like Machine-2-Machine, Internet of Things, big data, storage and cloud and so on.The combination and merger of these research areas could lead to more technological advancement and breakthrough in the future which is also worth looking at by researchers venturing into this research field.But it will be important if researchers also focus on how these technologies will affect CPS applications and the underlining issues of interoperability, reliability and safety.Also, for future applications, CPS will play a very vital and prominent role in Germany's "Industry 4.0" and China's "Made in China 2025" projects and other CPS related projects in other countries.

Conclusion
Cyber-physical system (CPS) is a new technology that integrates cyber and physical components alongside computational elements to achieve high efficiency and performance.The main objective of this paper is to present an up-to-date information and overview on applications relating to energy CPS and also security concerns in the system in order to furnish researchers and scholars with a comprehensive overview on the topic.This study was performed using a Systematic Literature Review methodology that enabled us to search, locate, assess, analyse, and categorize existing selected papers in literature and also identify the connection and relation between papers so as to draw a valid conclusion about its contents.An extensive review of papers was carried out from 2010 to mid-2019 from various databases to assist scholars to learn new approaches in this research area.From our results, it was seen that this research area is exponentially growing in terms of publications especially in recent years and this confirms the fact that researchers are now more interested in exploring recent theories and technologies from this subject area.
The main limitation of this work is that only publications relating to energy in CPS was selected for this study so as to keep the volume and number of pages of this paper to a required size.Quite a few other CPS research fields were not treated in this paper such as agriculture, robotics, medicine, process control, aerospace systems and education as this will be looked at in future works.You are free to: Sharecopy and redistribute the material in any medium or format.Adaptremix, transform, and build upon the material for any purpose, even commercially.The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms: Attribution -You must give appropriate credit, provide a link to the license, and indicate if changes were made.You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.

No additional restrictions
You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.

(
Figure 1.Cumulative number of articles published in the category of applications in energy CPS.

Figure 2 .
Figure 2. Cumulative number of articles published in the category of security issues in Energy CPS.

©
2019 The Author(s).This open access article is distributed under a Creative Commons Attribution (CC-BY) 4.0 license.
Cogent Engineering (ISSN: 2331-1916) is published by Cogent OA, part of Taylor & Francis Group.Publishing with Cogent OA ensures: • Immediate, universal access to your article on publication • High visibility and discoverability via the Cogent OA website as well as Taylor & Francis Online • Download and citation statistics for your article • Rapid online publication • Input from, and dialog with, expert editors and editorial boards • Retention of full copyright of your article • Guaranteed legacy preservation of your article • Discounts and waivers for authors in developing regions Submit your manuscript to a Cogent OA journal at www.CogentOA.com

Table 1 .
Article sources and the number of articles found

Table 2 .
Publications in the smart grid and smart cities research area

Table 3 .
Publications in the intelligent transportation and smart cars research area

Table 4 .
Publications in the industries and smart manufacturing area of research

Table 6 .
Key characteristics and research areas identified in security in CPS