Вы находитесь на странице: 1из 6

Future Generation Computer Systems xxx (xxxx) xxx

Contents lists available at ScienceDirect

Future Generation Computer Systems


journal homepage: www.elsevier.com/locate/fgcs

Editorial

Editorial: Smart Cyber–Physical Systems: Toward Pervasive


Intelligence systems
Flavia C. Delicato a ,1 , Adnan Al-Anbuky b ,2 , Kevin I-Kai Wang c ,2
a
Department of Computer Science, Federal University of Rio de Janeiro, Brazil
b
Auckland University of Technology, New Zealand
c
The University of Auckland, New Zealand

article info a b s t r a c t

Article history: Cyber–physical systems (CPSs) are the next generation of engineering systems, in which computing,
Available online xxxx communication and process control technologies are transparently integrated, constituting novel type
Keywords:
of autonomous system. In these systems, embedded computers and networks monitor (through
Cyber–Physical Systems sensors) and control (through actuators) the physical processes, usually with feedback loops where
Pervasive systems physical processes and computations affect each other. By equipping physical objects with interfaces
Computational intelligence to the virtual world, and by incorporating intelligent mechanisms to leverage collaboration between
Smart systems these objects, the boundaries between the physical and virtual realms become blurred. Interactions
Security occurring in the physical world can change the processing behavior in the virtual world, in a causal
relationship that can be exploited for the constant improvement of processes. Intelligent, self-aware,
self-managing and self-configuring pervasive systems can be built to improve quality of process in
many important application domains such as transportation, energy, industrial and medical systems.
The goal of this special issue is to report high-quality research on recent advances toward the
realization of the Smart Cyber–Physical Systems paradigm. By presenting a selection of papers on
various topics related to Smart Cyber–Physical Systems, we hope to shed light on the multiple aspects
of this emerging paradigm. The papers included in this issue propose solutions for data processing and
analysis, architecture, platforms and technology enablers, energy and security management, as well as
its application to building smart and sustainable spaces.
© 2019 Published by Elsevier B.V.

1. Introduction virtual components), and applying intelligence techniques, are


able to efficiently manage real-world processes and support a
Cyber–Physical Systems (CPS) [1–4] refer to the seamless in- broad range of novel applications and services.
tegration of computation with physical processes, possibly with By equipping physical objects with interfaces to the virtual
humans in the loop. In these systems, embedded computers and world, and by incorporating intelligent mechanisms to leverage
networks monitor (through sensors) and control (through actua- collaboration between these objects, the boundaries between the
tors) the physical processes, usually with feedback loops where physical and virtual realms become blurred. Interactions occur-
physical processes and computations affect each other. A key ring in the physical world can change the processing behavior in
point in these systems is the control of physical processes from the virtual world, in a causal relationship that can be exploited for
the monitoring of variables and the use of computational intelli- the constant improvement of processes. Intelligent, self-aware,
gence to obtain a deep knowledge of the monitored environment, self-managing and self-configuring pervasive systems can be built
thus providing timely and more accurate decisions and actions. to improve quality of process across a variety of application
The growing interconnection of physical and virtual worlds, and domains, helping to address several contemporary social and
the development of increasingly sophisticated intelligence tech- environmental issues.
niques, has opened the door to the next generation of CPS, that Components of a sCPS must have a high degree of auton-
is referred to as smart cyber–physical systems (sCPS). sCPS are omy while cooperating with each other in a robust, scalable
large-scale software intensive and pervasive systems, which by and decentralized way. However, several challenges need to be
combining various data sources (both from physical objects and overcome in order to realize such a paradigm, which is highly
multidisciplinary. These challenges range from the design of in-
1 Lead Guest Editor. telligent and energy efficient physical infrastructures for sens-
2 Guest Editor. ing and communication, data stream processing, data analytics

https://doi.org/10.1016/j.future.2019.06.031
0167-739X/© 2019 Published by Elsevier B.V.

Please cite this article as: F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang, Editorial: Smart Cyber–Physical Systems: Toward Pervasive Intelligence systems, Future Generation
Computer Systems (2019), https://doi.org/10.1016/j.future.2019.06.031.
2 F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang / Future Generation Computer Systems xxx (xxxx) xxx

and machine learning techniques to build the intelligence core issue to consider in event detection processes is the presence of
of these systems through the development of self-adaptive and outliers. Outliers can be defined as readings that are outside what
context-aware software. Moreover safety, social and behavioral is considered as a ‘‘normal state’’ of the data being collected (rep-
issues also need to be considered, when including human beings resented by a model, for example). These ‘‘abnormal’’ readings
as an integral part of these highly complex systems. are considered ‘‘events with extremely small probability of occur-
This special issue aims to report high-quality research on re- rence’’. That is, outliers are elements that differ significantly from
cent advances toward the realization of the Smart Cyber–Physical others in a dataset, which does not necessarily mean that they
Systems paradigm. Our goal is to shed light in the multifaceted
represent errors. Sometimes, outliers can represent important
aspects of this emerging paradigm, including solutions for data
information about the CPS application in question. For example,
processing and analysis, architecture, platforms and technology
a health risk situation in a patient being monitored, or some
enablers, energy and security management, as well as its appli-
cation to building smart and sustainable spaces. After a careful environmental risk [9]. Therefore, the detection of outliers in CPS
review process performed by reviewers selected by the Guest is very important for the correct analysis of the data and for
Editors, 17 papers were accepted addressing subjects that were producing reliable information. In [10], the authors propose a
classified in the following categories: (i) data processing and method to explore the multiway nature of urban spaces data in
analysis; (ii) energy management; (iii) architecture, platforms outlier’s detection. The proposed method includes three stages:
and technology enablers for Smart CPS; (iv) security manage- (i) dimensionality reduction, where data is modeled as a 3rd-
ment; and (v) Cyber-enabled Smart Applications. The detailed order tensor; from this reduction, they extract a set of latent
contributions of each paper are summarized in the next section. factors to obtain the best fit for the next classification step;
(ii) classification of latent factors, where the latent factors from
2. Content of this issue the stage (i) are used to generate instances of similar events
in monitoring smart urban spaces which result in high-quality
2.1. Data processing and analysis in Smart CPS clusters from the factorization; and (iii) combining steps (i) and
(ii) to generate a refined urban space pattern identification model.
Environmental sensors of all kinds, social media, and large-
The authors analyzed a real large-scale dataset with valuable data
scale computing infrastructures have produced a variety of cyber–
captured and streamed by urban sensors from 4 cities: Elda and
physical–social data, ranging from values of the current temper-
ature in a room, Twitter posts, geographical location of people, Rois (Spain), Nuremberg (Germany), and Tallinn (Estonia). The
up to the trajectories of cars on highways. Analyzing these data obtained results provided important insights as the observation
for extracting knowledge and producing useful information for that there is a kind of cyclic time patterns of urban sensing, which
humans and computer systems is the key to obtaining the real can be exploited for different purposes in an sCPS.
benefits of Smart CPS [5–7]. After being analyzed, the acquired Another interesting avenue of research explored in this special
cyber–physical–social data is capable of providing inputs for in- issue is data fusion techniques, particularly in the context of
telligent process control and aiding decision making. Data analy- wireless sensor networks (WSN) [11], one of the possible compo-
sis in CPS encompasses from simple tasks such as data cleansing nents of a Smart CPS. Wireless sensor networks have evolved in
and filtering, to complex data fusion processing and event de- recent years, moving from single-application, monolithically and
tection. The inherent volume, heterogeneity and distribution of inflexibly designed systems toward highly virtualized systems
cyber–physical–social data pose challenges for the techniques to aimed at supporting multiple heterogeneous applications [12].
be adopted in order to perform efficient and timely analyses. With this new design approach, such networks can be viewed as
Event detection has been a significant topic in the context
an integrated CPS infrastructure for a multitude of applications
of data analyses techniques for a long time, since the initial
which can be owned and developed by different users. As the
development of pervasive systems. Recently, with the explosion
number of applications in a WSN increases, a growing amount
of social networks, event-detection technologies have played an
increasingly important role for various purposes, such as recom- of sensor-generated data will be produced, from which useful
mending future events, and instant sharing of events of interest. information can be extracted. In spite of the advances in energy
Users of social media act as human sensors, providing data and harvesting and in designing increasingly efficient and powerful
information in real time about entities and events. In the pa- devices, wireless sensors and actuators still commonly rely on
per [8], the authors claim that most of the research on event batteries as their energy sources, whose replacement is undesir-
detection – using human or non-human sensors – focuses only on able or infeasible. Therefore, a major requirement in WSN is to
identifying events. These models assume an event to be a single save energy in order to extend their operational lifetime. Among
entity and ignore that it can be composed of other new events the methods employed to extend network lifetime, Multisensor
over time. The detection of subevents enriches the understanding data fusion (MDF) is one of the most widely used. According
of the main event, contextualizing it and creating a powerful to the authors in [13], traditional MDFs are not able to identify
knowledge about the scenario. To capture the parts of an event different contexts, since they are designed using an application-
and the information changing over time, they proposed a scalable specific approach for the network. As the number of applications
and modeling based algorithm. The proposed algorithm identifies increases, the application data ranges overlap and it becomes
subevents and creates labels to properly represent them. The
more complex to identify the origin of each data sample to deliver
proposed sub-event detection approach was evaluated using two
data to the correct application, with the consequence of reducing
large-scale Twitter corpuses. The first one is related to Brazil’s
data accuracy. In order to overcome these limitations, the authors
political protests scenario. The second analyzes the Zika epidemic
in the world. Their approach detected several subevents, most propose a MDF technique that divides the monitored interval into
of them related to real subevents. Due to the nature of social a set of intervals (non-overlapped intervals and overlapped inter-
networks, with a minimum delay between an event occurrence vals) and attributes each interval to an abstract sensor. Then they
and its dissemination, these results can open an opportunity for use MDFs to identify (hidden) correlations in abstract sensors and
temporal tracking of emergence and outbreak scenarios. to exploit such knowledge to monitor the behavior of sensors
In addition to the issue of detecting subevents cyber–physical– during their working life. Their proposal was validated through
social data, explored in the previous paper, another relevant simulations and tests on real nodes.

Please cite this article as: F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang, Editorial: Smart Cyber–Physical Systems: Toward Pervasive Intelligence systems, Future Generation
Computer Systems (2019), https://doi.org/10.1016/j.future.2019.06.031.
F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang / Future Generation Computer Systems xxx (xxxx) xxx 3

2.2. Energy management in Smart CPS 17]. Challenges in this context include (iii) the development and
adaptation of technologies for the interconnection between phys-
The importance of saving energy has already been mentioned ical and virtual elements that support their communication and
and explored in the previous discussed paper [13], since Cyber– integration in a ubiquitous, seamless and reliable manner, (i) the
Physical Systems are usually composed of a myriad of battery- development and implementation of comprehensive frameworks
powered devices. For better results, holistic solutions are needed and architecture that include the physical and cybernetic ele-
that increase the energy efficiency from the hardware devices ments of sCPS, and (ii) the adoption and adaptation of computing
to the software that executes on them, through the use of data platforms capable of effectively extracting useful information
aggregation and data fusion techniques such as the one men- from large amounts of raw data, in an effective and efficient way.
tioned above. Therefore, developers should pay attention to the Regarding the interconnection of physical devices with the
energy consumption of the global system so as not to compromise cyber world, in recent years we have witnessed the continuous
the system lifetime. In [14] the authors contribute to increase evolution of the Internet of Things (IoT) [18], which is one of the
the energy-awareness of cyber–physical systems developers, in pillars on which contemporary physical-cybernetic systems have
order to help making informed decisions that promote the system flourished. The IoT can be defined as an ensemble of different sys-
energy efficiency. According to the authors, there are plenty of tems (e.g., Smart Roads, Smart Buildings, Smart Grids) composed
experimental studies that give hints about how to reduce the of heterogeneous but interacting components (e.g., humans, cars,
energy consumption. However, this knowledge is not readily smartphones, gateways, smart meters) both individually and col-
available for the software developers of cyber–physical systems. lectively providing innovative cyber–physical services. Due to the
They normally use software development environments that do varying data volumes produced by smart devices, enterprises and
not provide useful advice about the energy consumption of the researchers seek methods that enable them to manage those
software solutions being implemented. In the paper, they pro- devices and process the real-time streams of data to generate
pose a Developer Eco-Assistant to integrate the experimental useful information in a timely manner [19]. Containers [20] have
results obtained by researchers into the software development been a recent and promising approach to help rapidly develop,
environments. In the proposed solution, the energy information is run, test and update IoT applications. The autoscaling capacity of
obtained in real-time from a repository of energy consuming con- containers can adaptively allocate computing resources for vari-
cerns, where researchers store their experimental measurements. ous data volumes over time. Therefore, elasticity, a critical feature
Developers use the repository to perform sustainability analyses, of a cloud platform, is significant to measure the performance of
which, in turn, will lead to greener design/implementation de- lightweight containers. In the paper ‘‘Quantifying Cloud Elasticity
cisions. In the paper, they illustrate the use of the approach in with Container-based Autoscaling’’, authored by Xuxin Tang, Fan
the context of cyber–physical systems development using both Zhang, Xiu Li, Samee Khan, Zhijiang Li, the authors propose a
open source environments (e.g. JetBrains IDEs) and proprietary framework with container auto-scaler. The proposed framework
environments (e.g. Waspmote development environment). They monitors containers resource usage and accordingly scales in
experimentally demonstrated that cyber–physical systems can or scales out containers in need. Further, they define elasticity
reduce more than 40% of its energy consumption depending on mathematically in order to quantify the cloud elasticity using
the scenario, reaching approximately 90% in some certain cases. the proposed framework. Extensive experiments were carried out
When adopting strategies for saving energy in CPS, one should with different workload modes, workload durations, and scaling
not lose sight of the overall quality of data and of the service cool-down period of times. Results shown that the framework
provided to applications and end users. A highly energy efficient captures the workload variation firmly with a very short delay.
system that does not deliver correct data, makes accurate and They also found out that the container-enabled platform shows
timely decisions does not meet the objectives for which it was de- the best elasticity in repeat workload mode due to its recurring
signed. An over-usage of periodic sensors (i.e. too frequent mea- and predictable feature, and that the length of the cool-down
surements) may easily lead to quick battery depletion, besides period should be properly set up in order to balance system
producing network congestion, and conversely, a lower usage stability and good elasticity.
is likely to cause poor measurement quality. Therefore, the ele- In [21], the authors claim that despite a decade of research, the
ments of a CPS, including the smart sensors, need to be carefully IoT is still into an emergent phase: instead of achieving the en-
configured in order to properly manage the energy-quality trade- visaged state of global interconnectivity of systems, devices and
off. In [15], the authors propose a middleware that continuously people, it is currently shaped by few isolated IoT systems/devices
generates and exposes to the sensor network an energy-efficient that provide conventional computing services mainly designed
sensors configuration based on data live observations. Using a for static environments with a-priori interactions. Nevertheless,
live learning process, their contributions dynamically act on two there is a clear and prominent trend toward a fully realized
configuration points: (i) sensors sampling frequency, which are IoT, in which arguably the key drivers will be cyber–physical,
optimized based on machine-learning predictability from pre- highly dynamic and contextualized services, called by the authors
vious measurements, (ii) network usage optimization according ‘‘Opportunistic IoT Services’’. For such services, provisioning may
to frequency of requests from deployed software applications. be meaningful only in certain space/time configurations and may
As a major outcome of their work, a self-adaptive platform was be subject to multiple heterogeneous constraints and conditions,
designed with an extended sensors battery life, while ensuring e.g., current user status, location, policies, etc. In this direction,
a proper level of data quality and freshness. Through theoretical their work presents a full-fledged approach to the development
and experimental assessments, they shown the capacity of the of Opportunistic IoT Services from high-level metamodeling up to
proposed approach to constantly find a near-optimal tradeoff their concrete implementation and simulation.
between sensors and network usage, and measurement quality. In addition to the interconnection of heterogeneous devices
among themselves and with the Internet, largely enabled by
2.3. Architecture, platforms and technology enablers for Smart CPS IoT, Smart Cyber–physical systems can be highly connected and
integrated in many other ways, often crossing the boundaries of
Innovative architectures, software platforms and technolo- multiple application domains. Achieving the design of effectively
gies are needed to support the construction and execution of interconnected, cooperative, autonomous systems that maintain
highly complex and interconnected cyber–physical systems [16, the man in the loop is a major challenge for the widespread

Please cite this article as: F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang, Editorial: Smart Cyber–Physical Systems: Toward Pervasive Intelligence systems, Future Generation
Computer Systems (2019), https://doi.org/10.1016/j.future.2019.06.031.
4 F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang / Future Generation Computer Systems xxx (xxxx) xxx

adoption of such systems in the future. One scenario where such speculation mechanism to detect abnormal tasks during runtime.
a challenge becomes more evident is in building CPSs for Smart However, such mechanism cannot detect all abnormal tasks and
Cities [22]. Smart Cities combine advances in IoT, Big Data, Social does not provide the root causes of degraded performance. In [30]
Networks, and Cloud Computing technologies with the demand an offline tool is proposed for Log-based Abnormal tasks Detec-
for cyber–physical applications in areas of public interest, such tion and Root-cause Analysis (LADRA) for big data processing with
as Health, Public Safety, and Mobility. The end goal is to leverage Spark. LADRA can detect abnormal tasks by examining features
the use of city resources to improve the quality of life of its extracted from logs through statistical spatial–temporal analysis,
citizens. Achieving this goal, however, requires advanced support which consist of Spark execution logs and JVM garbage collec-
for the development and operation of applications in a com- tion (GC) logs related to resource usage. Moreover, a General
plex and dynamic environment. Middleware platforms can pro- Regression Neural Network (GRNN) is used to identify root causes
vide an integrated infrastructure that enables solutions for smart for abnormal tasks. The experimental results demonstrated that
cities by combining heterogeneous city devices and providing LADRA outperformances the Spark speculation mechanism.
unified, high-level facilities for the development of applications Coordination in such complex environments as Smart CPS
and services [23]. Although several smart city platforms have is undoubtedly one of the most challenging tasks and various
been proposed in the literature, there are still open research algorithmic and technologic solutions have been investigated in
and development challenges related to their scalability, main- this context. A very interesting possibility is investigated by the
tainability, interoperability, and reuse in the context of different author in [31]. He proposes extending the applicability of de-
cities, to name a few. Moreover, available platforms lack scientific centralized blockchain consensus protocols to the broad field of
validation, which hinders a comparative analysis of their applica- sCPS. Blockchain technology consists in its core of the use of
bility. Aiming to close this gap, in [24], the authors propose the consensus algorithms for coordination between parties involved
InterSCity platform for Smart Cities. InterSCity is a microservices- in a transaction. It was initially adopted in the context of digital
based and open-source platform that enables the collaborative currencies and financial transactions but has recently attracted
development of large-scale systems, applications, and services huge attention from industry and academia and its use was
for the cities of the future, contributing to turn them into truly expanded for a myriad of applications. The author contribution
sCPS environments. In the paper, they present the architecture in the paper includes a proposal of a new family of cyber–
of the InterSCity platform, followed by a comprehensive set of physical systems, an in-depth discussion of new possibilities and
experiments that evaluate its scalability properties. These exper- inherent shortcomings. The proposed coupling of the data stor-
iments were conducted using a smart city simulator to assess the age mechanisms, communication and consensus protocols, allows
platform. The experimental results shown that the platform can for deployment of self-sustaining cyber–physical environments
scale horizontally to handle the highly dynamic demands of a in which all the mission-critical aspects in both the cyber and
large smart city while maintaining low response times as well physical layer are effectively incentivized, coordinated and main-
as the usefulness of the synthetic workload generation technique tained. This includes the essential data storage, data exchange,
used in the experiments. but also the production, transmission and distribution of energy.
As the volume and diversity of data collected through Smart The paper presents a comprehensive undertaking toward sym-
CPS increases, how to efficiently process and store data becomes biotic cyber–physical systems and environments. The proposed
a challenging topic. In order to offer large enough data storage family of systems is decentralized and does not imply the ne-
with fast data access, a hybrid storage system that consists of cessity of reliance on trusted third parties. An architecture is
both traditional mechanical hard disk drive (HDD) and higher devised and discussed in the paper which incorporates physical,
performance solid state drive (SSD) is typically used. In hybrid blockchain-enabled sensors and actuators capable of monitoring
storage systems, different data is distributed on difference devices and altering parameters of the underlying physical layer. A par-
according to their performance benefits, where performance- ticular instance of a blockchain-aided CPS’ is analyzed, in the
critical data are ones that can lead to significant performance gain context of the power grid at both the micro and macro-grid
if located on SSD. The key issue in hybrid storage systems is how levels. At the macro-grid level, they explore the possibility of
to identify performance-critical data properly. Existing identifi- utilizing blockchain-enabled physical actuators. These theoretical
cation schemes mainly focus on two categories: hotness-based actuators are able to introduce savings by a near real-time routing
identification (i.e. how frequent is the data accessed) [25,26] and of energy and enable for tracking of power losses thanks to de-
cost-based identification (e.g. latency cost of I/O access) [27,28]. centralized accountability. At the micro-grid level, they propose
In [29], a novel identification approach called FvRS was proposed a network of blockchain enabled smart meters to coordinate
to properly identify high-cost hot data and to maximize the real-time auction-based energy trade. To glue together the el-
utilization of SSD to improve system performance. The I/O cost ements of such an open, trustless, decentralized system, they
was more accurately modeled by evaluating both the request size discuss an intrinsic communication protocol capable of rewarding
and data access sequentially. The FvRS system was implemented intermediaries of a data exchange for each transmitted byte.
on a Linux-based hybrid storage system and evaluated using
three real-workload traces and a famous benchmark Postmark. 2.4. Security management in Smart CPS
The results showed that FvRS reduces I/O response time by up
to 45.6% and 25.1%, respectively, with existing hotness-based As the number of smart and connected things grows in a sCPS,
identification and cost-based identification approaches. responsibility on the transmitted data also needs to increase,
In addition to storage, processing large amount of data col- especially in terms of security aspects, as authentication, access
lected from sCPS and providing timely information and decision control, data protection, and trust, just to name a few. There is no
are also critical to the success of real-world applications. MapRe- question that reliability, safety and security issues play a crucial
duce paradigm is commonly used to handle batch and streaming role in the acceptance and use of today’s and future sCPS. For a
data and among all the existing frameworks, Spark, has gained sCPS to be reliable, secure, and safe, the system must be able to
a wide adoption. However, Spark is vulnerable to the existence adapt to the physical environment and withstand cybernetic and
of abnormal tasks (i.e. showing significant delay in comparison physical attacks while maintaining data integrity and robustness.
with other tasks), which can lead to significant performance In addition to primitives and architectures that manage typical
degradation. To mitigate such performance issues, Spark uses a security aspects such as authorization, access control and data

Please cite this article as: F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang, Editorial: Smart Cyber–Physical Systems: Toward Pervasive Intelligence systems, Future Generation
Computer Systems (2019), https://doi.org/10.1016/j.future.2019.06.031.
F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang / Future Generation Computer Systems xxx (xxxx) xxx 5

integrity, trust between the parties (physical and virtual) of a sCPS the role of smart homes in monitoring the activities of the el-
must be established and maintained. derly, identifying safety hazards in the home, and understanding
In [32] the authors investigate security aspects in the context environmental changes that may correlate with the deterioration
of vehicular Cyber–Physical Systems. In such complex network of cognitive and physical health. They present LaPlace, a system
environments, there are trust-based recommendation schemes used to manage context enriched behavior patterns learned in a
that could effectively filter most of the false data. However, ac- smart home with sensing devices. LaPlace is based on a formal
cording to the authors, these schemes may exhaust the resources model to represent intelligible, context enriched behavior pat-
of the vehicular network, including energy, computation abil- terns, an online adaptive learning algorithm called TIMe which
ity, and storage, causing a network outage. To ensure real-time was created for learning such patterns. Context-awareness pro-
data transmission and security in a vehicular CPS network, they vides insights about behavior that may otherwise go unnoticed.
propose a novel trust-based recommendation scheme (TBRS). TIMe is a one-pass algorithm that uses a stream processing model.
The paper has three main contributions. First, the isomerism of The TIMe algorithm is presented in this paper along with an
vehicular sensor nodes in CPS networks, where differences of extensive evaluation of it using real life datasets.
mobility between normal nodes and selfish/malicious nodes is Fundamental to the Smart CPS is the understanding of and
analyzed. Second, a trust model is designed based on delivery interaction with human through various types of sensor inputs
credibility and position intimacy of nodes; such a model can and actuator outputs. Human Activity Recognition (HAR) is one
adjust the weight coefficient of direct trust parameters, which can of the most critical elements in many sCPS such as smart homes,
be utilized to analyze the secure and trustable tasks in data trans- elderly healthcare, and child monitoring. The advancement of
mission. Third, to address attacks caused by selfish/malicious embedded and communication technologies has led to the pro-
nodes and sparsity issues of nodes in vehicular CPS, a secure liferation of a wide variety of sensors embedded in the ambient
filtering algorithm based on K-Nearest Neighbor (KNN) cooper- environment and attached on humans as wearable devices. The
ative computing is proposed. The trust value is computed by the diversity and interleaved nature of heterogeneous sensor data
proposed trust model. The cooperative computing-based filtering contributes to the major challenge of identifying and segmenting
algorithm is utilized to filter false recommendation from self- sensor events and associating them with ongoing Activities of
ish/malicious nodes, thus greatly reducing interference of these Daily Living (ADL). Existing HAR proposals rarely explore se-
nodes on the performance of vehicular CPS network. Experi- mantically based approach for distinguishing sensor events and
mental results shown that the proposed scheme is superior to linking with the relevant human activities. The authors in [35]
the comparative schemes in terms of delivery rate, transmission proposed a new approach using Semiotic theory inspired on-
delay and reliability. Besides, the resistance against illegal eaves- tological model to capture generic knowledge and inhabitant-
dropping attacks has increased by an average of 32.53% when specific preferences for sensor event segmentation and HAR. The
compared to other algorithms. proposed method was implemented and tested against 30 test
In [33] the authors present a solution for security in a different scenarios. The results demonstrated that all sensor events were
domain of SCPS. The massive proliferation of sophisticated tech- adequately segmented with 100% accuracy for single ADL scenar-
nologies into the heart of traditional Industrial Control Systems ios and minor improvement of 97.8% accuracy for composite ADL
has given birth to ‘‘smart Industrial Cyber–Physical Systems’’ scenarios.
(ICPS). While this industrial revolution has brought upon a wide In addition to identifying sensor events and human physical
range of advantages, it also raised new design challenges and activities, one of the main features of Smart CPS is to be able
exposed ICPS to a new breed of cyber–physical attacks. This pa- to evolve and adapt toward user’s needs. In order to achieve
per aims to integrate security primitives (e.g., enforcing/verifying this goal, an accurate and efficient cognitive model that can
data authenticity) in control applications by formulating an inno- learn, memorize and recall past knowledge, in a way, mimicking
vative architectural paradigm shift. More specifically, their pro- human brains, is required. However, the processing efficacy and
posal is twofold. The authors elaborated a novel security-aware efficiency of such cognitive model pose critical challenge to small,
control application architecture, which: (i) defines two security battery powered IoT devices. Therefore, using minimal resources
primitives that are called at the beginning and at the end of to facilitate learning and recall is of critical importance to the
each program to verify and to enforce the required security success of IoT applications. In [36], the authors proposed a 3-
properties; and (ii) runs the key management code as a separate layer associative memory and recall (AMR) model, that makes use
program in order to isolate its implementation and to ensure its of memory chunking mechanisms and knowledge–information–
minimal interference with the rest of the programs. They also data (KID) cognitive model to efficiently organize and form asso-
designed a lightweight key distribution protocol exploiting the ciation of human activity knowledge. The proposed approach is
characteristics and computational advantages of symmetric key evaluated using the well know ‘‘Activities of Daily Living (ADLs)
cryptography and hash functions. Extensive experimental results dataset’’ and demonstrated better performance comparing with
on a testbed replicating the precise hardware and software of a the state-of-the-art approaches in finding associations between
node from a Romanian gas transportation network shown the human activities efficiently.
effectiveness of the proposed scheme and its applicability to Another important application area of sCPS is the digitization
resource constrained ICPS. of medical applications, which is beneficial to the entire human
race in both developed and developing countries which are facing
2.5. Cyber-enabled smart applications scarcity of medical resources and personnel. A typical application
that is becoming increasingly common is the collection of gene
In addition to the domains of smart applications already ex- expression data and cyber-enabled tumor risk analysis. In this
plored in the previous papers, such as Industrial, Vehicular and application, a key challenge is to identify the distinct genes that
Smart Cities, another very relevant domain is that of Smart are the most likely to cause tumors. In [37], an improved SSO
Homes. In a society with a growing population of elders, provid- (Simplified Swarm Optimization) algorithm is proposed with a
ing efficient and cost-effective long-term care has become central hybrid filter and local search strategy to achieve a higher clas-
to reducing the economic and societal impact of this demographic sification accuracy with fewer selected genes. 10 tumor gene
shift. Part of the solution to improving the well-being of the expression datasets are used to evaluate the proposed algorithm
elderly at home are smart homes. In [34] the authors examine against nine well-known benchmark classification methods and

Please cite this article as: F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang, Editorial: Smart Cyber–Physical Systems: Toward Pervasive Intelligence systems, Future Generation
Computer Systems (2019), https://doi.org/10.1016/j.future.2019.06.031.
6 F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang / Future Generation Computer Systems xxx (xxxx) xxx

other approaches from 6 referenced studies. The experimental [16] E.A. Lee, Cyber physical systems: Design challenges, in: 2008 11th IEEE
results showed that the proposed algorithm achieves better clas- International Symposium on Object and Component-Oriented Real-Time
Distributed Computing (ISORC), Orlando, FL, 2008, pp. 363–369, http:
sification accuracy in comparison to the existing approaches. In
//dx.doi.org/10.1109/ISORC.2008.25.
addition to tumor risk analysis, this improved algorithm also has [17] J. Zeng, et al., A survey: Cyber-physical-social systems and their system-
the potential to be used in pervasive patient monitoring and level design methodology, Future Gener. Comput. Syst. (2016) [online]
health management to analyze physiological parameters, such as Available: http://dx.doi.org/10.1016/j.future.2016.06.034.
[18] L. Atzori, A. Iera, G. Morabito, The internet of things: A survey, Comput.
ECG, blood pressure, and SPO2 .
Netw. (ISSN: 1389-1286) 54 (15) (2010) 2787–2805, http://dx.doi.org/10.
1016/j.comnet.2010.05.010.
3. Conclusions [19] J. Gubbi, R. Buyya, S. Marusic, et al., Internet of Things (IoT): A vision,
architectural elements, and future directions, Future Gener. Comput. Syst.
In conclusion, the papers included in this Special Issue show 29 (7) (2013) 1645–1660.
[20] D. Merkel, Docker: lightweight linux containers for consistente develop-
the diversity of research being conducted in the field of Smart
ment and deployment, Linux J. 239 (2014).
Cyber–Physical Systems. While this issue offers limited contribu- [21] R. Casadei, G. Fortino, D. Pianini, W. Russo, C. Savaglio, M. Viroli, Modelling
tion to this field, we strongly believe it has opened the door for and simulation of opportunistic IoT services with aggregate computing,
more involved research on this important area of research. Future Gener. Comput. Syst. (ISSN: 0167-739X) 91 (2019) 252–262, http:
//dx.doi.org/10.1016/j.future.2018.09.005.
[22] R. Petrolo, V. Loscri, N. Mitton, Cyber-physical objects as key elements for
Acknowledgments a smart cyber-city, in: A. Guerrieri, V. Loscri, A. Rovella, G. Fortino (Eds.),
Management of Cyber Physical Objects in the Future Internet of Things.
The guest editors would like to thank all the authors and the Internet of Things (Technology, Communications and Computing), Springer,
reviewers for their hard work in helping us organize this Special Cham, 2016.
[23] E.F.Z. Santana, A.P. Chaves, M.A. Gerosa, F. Kon, D.S. Milojicic, Software
Issue. They would also like to express their heartfelt gratitude
platforms for smart cities: Concepts, requirements, Challenges, and a uni-
to the Editor-in-Chief, Prof. Peter Sloot, for giving us this great fied reference architecture, ACM Comput. Surv. 50 (6) (2017) 78:1–78:37,
opportunity, and the members of the Editorial Staff for their URL http://doi.acm.org/10.1145/3124391.
support during the process. [24] A. de M. Del Esposte, E.F.Z. Santana, L. Kanashiro, F.M. Costa, K.R. Braghetto,
N. Lago, Fabio Kon, Design and evaluation of a scalable smart city software
platform with large-scale simulations, Future Gener. Comput. Syst. (ISSN:
References 0167-739X) 93 (2019) 427–441, http://dx.doi.org/10.1016/j.future.2018.10.
026.
[1] W. Wolf, Cyber-Phys. Syst. Comput. 42 (03) (2009) 88–89, http://dx.doi. [25] I. Koltsidas, S.D. Viglas, Flashing up the storage layer, Proc. VLDB
org/10.1109/MC.2009.81. Endowment 1 (1) (2008) 514–525.
[2] K. Kim, P.R. Kumar, Cyber–physical systems: A perspective at the cen- [26] Y. Lv, B. Cui, X. Chen, J. Li, Hotness-aware buffer management for
tennial, Proc. IEEE 100 (Special Centennial Issue) (2012) 1287–1308, http: flash-based hybrid storage systems, in: Proceedings of the 22nd ACM
//dx.doi.org/10.1109/JPROC.2012.2189792. International Conference on Conference on Information & Knowledge
[3] F. Wu, Y. Kao, Y. Tseng, From wireless sensor networks towards cyber Management, ACM, 2013, pp. 1631–1636.
physical systems, Pervasive Mob. Comput. (ISSN: 1574-1192) 7 (4) (2011) [27] F. Chen, D.A. Koufaty, X. Zhang, Hystor: making the best use of solid
397–413, http://dx.doi.org/10.1016/j.pmcj.2011.03.003. state drives in high performance storage systems, in: Proceedings of the
[4] J. Wan, M. Chen, F. Xia, D. Li, K. Zhou, From machine-to-machine com- International Conference on Supercomputing, ACM, 2011, pp. 22–32.
munications towards cyber-physical systems, Comput. Sci. Inf. Syst. 10 (3) [28] C.K. Kang, Y.J. Cai, C.H. Wu, P.C. Hsiu, A hybrid storage access framework
(2013) 1105–1128. for high-performance virtual machines, ACM Trans. Embedded Comput.
[5] A.J. Jara, D. Genoud, Y. Bocchi, Big data for cyber physical systems: Syst. 13 (5s) (2014) 157.
An analysis of challenges, solutions and opportunities, in: 2014 Eighth [29] Gaoxiang Xu, Zhipeng Tan, Dan Feng, Laurence T. Yang, Wei Zhou, Xinyan
International Conference on Innovative Mobile and Internet Services in Zhang, Yang Zhang, Jie Xu, FvRS: Efficiently identifying performance-critical
Ubiquitous Computing, Birmingham, 2014, pp. 376–380, http://dx.doi.org/ data for improving performance of big data processing, Future Gener.
10.1109/IMIS.2014.139. Comput. Syst. (ISSN: 0167-739X) 91 (2019) 157–166, http://dx.doi.org/10.
[6] Y. Sun, H. Song, A.J. Jara, R. Bie, Internet of things and big data analytics 1016/j.future.2018.09.003.
for smart and connected communities, IEEE Access 4 (2016) 766–773, [30] Siyang Lu, Xiang Wei, Bingbing Rao, Byungchul Tak, Long Wang, Liqiang
http://dx.doi.org/10.1109/ACCESS.2016.2529723. Wang, LADRA: Log-based abnormal task detection and root-cause analysis
[7] X. Xu, Q. Hua, Industrial big data analysis in smart factory: Current
in big data processing with Spark, Future Gener. Comput. Syst. (ISSN:
status and research strategies, IEEE Access 5 (2017) 17543–17551, http:
0167-739X) 95 (2019) 392–403, https://doi.proxy.ufrj.br/10.1016/j.future.
//dx.doi.org/10.1109/ACCESS.2017.2741105.
2018.12.002.
[8] D. Nolasco, J. Oliveira, Subevents detection through topic modeling in social
[31] Rafał Skowroński, The open blockchain-aided multi-agent symbiotic cyber–
media posts, Future Gener. Comput. Syst. (ISSN: 0167-739X) 93 (2019)
physical systems, Future Gener. Comput. Syst. (ISSN: 0167-739X) 94 (2019)
290–303, http://dx.doi.org/10.1016/j.future.2018.09.008.
430–443, http://dx.doi.org/10.1016/j.future.2018.11.044.
[9] Y. Zhang, M. Qiu, C. Tsai, M.M. Hassan, A. Alamri, Health-CPS: Healthcare
[32] Wei Liang, Jing Long, Tien-Hsiung Weng, Xuhui Chen, Kuan-Ching Li,
cyber-physical system assisted by cloud and big data, IEEE Syst. J. 11 (1)
Albert Y. Zomaya, TBRS: A trust based recommendation scheme for
(2017) 88–95, http://dx.doi.org/10.1109/JSYST.2015.2460747.
vehicular CPS network, Future Gener. Comput. Syst. (ISSN: 0167-739X) 92
[10] T. Souza, A. Aquino, D.G. Gomes, A method to detect data outliers from
(2019) 383–398, http://dx.doi.org/10.1016/j.future.2018.09.002.
smart urban spaces via tensor analysis, Future Gener. Comput. Syst. (ISSN:
[33] Béla Genge, Piroska Haller, Adrian-Vasile Duka, Engineering security-aware
0167-739X) 92 (2019) 290–301, http://dx.doi.org/10.1016/j.future.2018.09.
control applications for data authentication in smart industrial cyber–
062.
physical systems, Future Gener. Comput. Syst. (ISSN: 0167-739X) 91 (2019)
[11] E.F. Nakamura, A.A.F. Loureiro, A.C. Frery, Information fusion for wireless
206–222, http://dx.doi.org/10.1016/j.future.2018.09.001.
sensor networks: Methods, models, and classifications, ACM Comput. Surv.
[34] Paula Lago, Claudia Roncancio, Learning and managing context enriched
39 (2007) 3, http://dx.doi.org/10.1145/1267070.1267073, Article 9.
behavior patterns in smart homes, Future Gener. Comput. Syst. (ISSN:
[12] C.M. de Farias, W. Li, F.C. Delicato, L. Pirmez, A.Y. Zomaya, P.F. Pires, J.N.
0167-739X) 91 (2019) 191–205, http://dx.doi.org/10.1016/j.future.2018.09.
de Souza, A systematic review of shared sensor networks, ACM Comput.
004.
Surv. 48 (4) (2016) 51:1–51:50.
[35] Darpan Triboan, Liming Chen, Feng Chen, Zumin Wang, A semantics-based
[13] C.M. de Farias, L. Pirmez, G. Fortino, A. Guerrieri, A multi-sensor data
approach to sensor data segmentation in real-time Activity Recognition,
fusion technique using data correlations among multiple applications,
Future Gener. Comput. Syst. (ISSN: 0167-739X) 93 (2019) 224–236.
Future Gener. Comput. Syst. (ISSN: 0167-739X) 92 (2019) 109–118, http:
[36] Runhe Huang, Peter Kimani Mungai, Jianhua Ma, Kevin I-Kai Wang,
//dx.doi.org/10.1016/j.future.2018.09.034.
[14] D. Munoz, J.A. Montenegro, M. Pinto, L. Fuentes, Energy-aware envi- Associative memory and recall model with KID model for human activity
ronments for the development of green applications for cyber–physical recognition, Future Gener. Comput. Syst. (ISSN: 0167-739X) 92 (2019)
systems, Future Gener. Comput. Syst. (ISSN: 0167-739X) 91 (2019) 312–323, http://dx.doi.org/10.1016/j.future.2018.09.007.
536–554, http://dx.doi.org/10.1016/j.future.2018.09.006. [37] Chaochao Ye, Julong Pan, Qun Jin, An improved SSO algorithm for cyber-
[15] C. Cecchinel, F. Fouquet, S. Mosser, P. Collet, Leveraging live machine enabled tumor risk analysis based on gene selection, Future Gener.
learning and deep sleep to support a self-adaptive efficient configuration Comput. Syst. 92 (2019) 407–418.
of battery powered sensors, Future Gener. Comput. Syst. (ISSN: 0167-739X)
92 (2019) 225–240, http://dx.doi.org/10.1016/j.future.2018.09.053.

Please cite this article as: F.C. Delicato, A. Al-Anbuky and K.I.-K. Wang, Editorial: Smart Cyber–Physical Systems: Toward Pervasive Intelligence systems, Future Generation
Computer Systems (2019), https://doi.org/10.1016/j.future.2019.06.031.

Вам также может понравиться