Event-based sensor data exchange and fusion in the Internet of Things environments

https://doi.org/10.1016/j.jpdc.2017.12.010Get rights and content

Highlights

  • A event-based data fusion, where heterogeneous devices exchange notifications of events among each other.

  • A solution for stream processing.

  • A new approach for sensor communications integrated it in the publish/subscribe service that we have designed and simulated within the context of a sensing infrastructure.

Abstract

Internet of Things (IoT) is a promising technology for improving our lives and society by integrating smart devices in our environment and paving the way for novel ICT application, spanning from smart cities to energy efficiency and home automation. However, such a vision encompasses the availability of thousands of smart devices, or even more, that continuously exchange a huge volume of data among each other and with cloud-based services, raising a big data problem. Such a problem can be approached by properly applying data fusion practices within an IoT infrastructure. Due to the characteristics and peculiarities of the communications among smart devices within the IoT, an event-based data fusion is needed, where devices exchange notifications of events among each others. Such data fusion should be focused on special devices where notification heterogeneity, and data source trust issues have to be faced with. Accordingly, the contribution of this work is proposing (i) a novel broker-less event-based communication protocol specifically tailored to sensors with constrained resources, (ii) a solution for flexible event-based communications among heterogeneous data sources, and (iii) an approach based on the theory of evidence for data fusion processes that depend on the matching and trust degree of the data to be fused.

Introduction

Internet of Things (IoT) [[4], [51]] is a novel promising vision for ubiquitous and pervasive computing that is facing an increasing success and widespread adoption thanks to the availability of Internet connections almost everywhere and to the miniaturization of computing devices. Basically, it consists of a set of smart devices, embedded in objects of daily use such as cups, household electrical appliances, cars and/or street furniture, just to cite some of them, or hold by human users, such as smart phones, tablets and/or sensors for health and wellness monitoring. Such devices are able to perform simple monitoring or remote control functions, by collecting data about the environment within which they are running and/or influencing it through simple actuators, as well as to execute some processing tasks, whose complexity depends on their computing and storage capabilities. In addition, any device is equipped with proper hardware for short-range radio frequency-based networking, and communicates by using one, or even more, widely-known wireless technologies and protocols, such as Wi-Fi, Bluetooth and/or ZigBee. Some devices may also be equipped with long-range communications interfaces for connecting to cellular networks, such as GPRS, HSDPA/HSUPA and/or LTE. All these interfaces and protocols may be used both for exchanging data among the individual devices as in Wireless Sensor Networks (WSNs), and to establish a connection to the Internet. Such a last peculiarity is the main difference between IoT and WSNs [1] allowing devices to outsource data storage and runtime tasks on cloud-based service infrastructures so as to take advantage of their elastic behavior and scalability for massive data collection and analytics purposes. In addition, these cloud services are also used to exchange the data collected/processed within the context of the IoT with external services and/or to visualize them through traditional browser by end users.

IoT represents the foundation for the latest ICT innovations [[22], [51]], such as smart cities & environment monitoring, critical infrastructure protection, domotic & home automation or e-health, just to refer the most cited ones [10]. Such applications are characterized by an extremely large horizontal and vertical scale, where the first characteristics refers to the geographical area covered by the application, while the second one considers the overall number of users and devices and the amount of generated/collected/stored/exchanged data. Just to mention a practical example, one of the above-mentioned application may reasonably include 50 billion active “users”, which may constantly talk among each other through the cloud-based service. If every device that is active sends a message per second, the IoT application might deal with multiple exabytes of data, which are likely to represent crucial messages (e.g., “This equipment is overheating”, “The driver’s heartbeat is over a certain threshold” or “An accident occurred at a certain road”) that could be lost in the shuffle. Within this data deluge, however, it is possible that not all the exchanged messages are significant, since the devices may repeat the same thing (e.g., a e-health sensor may emit a message with the monitored heartbeat every second, which may be the same with respect to the previous message, or a thermostat could report the current temperature in a car, which does not change respect to the previous monitoring period). Such useless notifications represents a kind of noise that let more important messages to be un-noticed and overwhelm the overall IoT infrastructure by wasting networking resources, cloud capabilities and device power, and hence should be carefully considered in order to reduce the data pressure affecting the IoT [18].

The most effective way to address such an issue is making use of data fusion techniques. Within the context of IoT, data fusion can be performed in-situ, in-network and in-cloud, where IoT devices may represent a first line of fusion for data acquired from their sensing hardware or coming from their neighbors, while more complex data fusion tasks are conducted within the cloud. The literature on in-situ data processing/fusion strongly depends on the theory of signal processing and descriptive statistics [27], while in-cloud data fusion is quite prolific [42]. Also, the literature of in-network data fusion, mainly fucused in sensor network scenarios, has a long history and several solutions have been proposed, so that an interested reader can find a valid survey on it from [35]. Despite the high number of solutions available and the efforts spent in dealing with them, all these three aspects still present several open issues that research has to consider and deal with. In this work, in-network data analysis has been taken into consideration, and several issues have been identified, essentially concerning the need of having the same schema for all the data, and the same weight for all the sources involved in the data fusion process. This usually does not hold within the IoT context and consequently we present a novel solution for automatic schema matching so as to flexibly fuse data with heterogeneous structure together with a way to manage trust among sensors and encompassing it within the data fusion process. To be effective, data fusion requires a suitable communication facility to convey data from the sensors’ locations to where data fusion takes place, i.e., edge gateways. Such a communication is characterized by a data-centric and decoupled nature, which can be only supported by a proper publish/subscribe service for the seamless integration of sensors and gateways. The current services available in the literature are not tailored on the sensors’ needs in terms of efficient use of resources and energy consumption. In a previous work we presented a protocol for event-based communication based on beaconing and without the use of intermediaries [14], that has been expanded in this work by adding the flexibility required for the seamless integration of sensors.

This approach also represents an evolution of our previous works on cloud service selection with theory of evidence [16] and trust management [13], specifically applied to the context of sensor data fusion. In detail, the main contributions of this work consist in:

  • a event-based data fusion, where heterogeneous devices exchange notifications of events among each other;

  • a solution for stream processing;

  • a new approach for sensor communications integrated in the publish/subscribe service framework that we have designed and simulated within the context of a sensing infrastructure.

The remaining of this article is organized as follows. Section 2 introduces the most important aspects of data fusion and event-based communications within the context of IoT, and highlights the many issues available in the current literature on these topics. Section 3 presents the proposed approach and is organized in two sub-sections: one devoted to present the event-based communication protocol, and another for the flexible and trust-based sensor data fusion. The following Section 4 describe an assessment of the presented approach so as to prove its quality. We close this paper with Section 5 by drawing some final remarks and some final plans with the directions for future work on the addressed topics.

Section snippets

Background and related work

As above mentioned the vision of IoT is characterized by two key elements: a sensor network and a set of cloud-based services. Specifically, we have a set of tiny sensors able to monitor the surrounding environment by measuring specific metrics of interest, such as temperature, pressure, humidity, luminescence, and so on. Each sensor is able to make simple data processing, such as the basic signal processing tasks (e.g., analog-to-digital conversion, denoising or ), sampling and/or aggregation,

Approach

We propose a publish/subscribe service to support the seamless data-centric interconnection of the IoT nodes, providing automatic schema matching and a set of combinatorial rules based on the theory of evidence and trust management. Specifically, our approach is illustrated in Fig. 2, where the publish/subscribe service supports the event-based decoupled interconnection of sensors and gateways thanks to a beaconing-based communication protocol without the need of intermediaries. Built upon this

Assessment

The scope of this section is to prove the effectiveness of the described solution in realizing efficient event-driven communications and in-network data fusion in the IoT scenario, by meeting its peculiar needs and limitations. To this aim, we have realized a prototype of our solution by using an event-based network simulator called OMNET++,1 where the sensory hardware and the wireless communication protocols have been simulated by using the off-the-shelf

Conclusions

This paper presented the data fusion characteristics and known issues within the context of IoT environments. We have based our work on the consideration that the communication and the data fusion are characterized by an event-driven perspective and on the lack in the literature of proper event-based communication protocols tailored to the sensors’ needs, as well as of data fusion approaches that are flexible to tolerate possible heterogeneities among data sources and an estimation of their

Acknowledgment

The research presented in this paper is supported by the following projects: MONROE - Toff project (H2020 No 644399), NETIO - ForestMon ( 53/05.09.2016, cod SMIS2014+ 105976), SPERO (PN-III-P2-2.1-SOL-2016-03-0046, 3Sol/2017) and ROBIN (PN-III-P1-1.2-PCCDI-2017-0734).

This article is based upon work from COST Action CA15127 (Improving Applicability of Nature-Inspired Optimisation by Joining Theory and Practice – ImAppNIO) supported by COST (European Cooperation in Science and Technology).

We

Christian Esposito is adjunct professor at the University of Napoli “Federico II”, and at the University of Salerno, where he is also a research fellow. He received the graduate-degree in computer engineering in 2006, and the Ph.D. degree in 2009 from the University of Naples “Federico II”. His main interests include mobile computing, benchmarking, aspects of publish/subscribe services, and security and reliability strategies for data dissemination in large-scale critical systems.

References (53)

  • Al-FuqahaA. et al.

    Internet of things: A survey on enabling technologies, protocols, and applications

    IEEE Commun. Surv. Tutor.

    (2015)
  • Al-FuqahaA. et al.

    Toward better horizontal integration among IoT services

    IEEE Commun. Mag.

    (2015)
  • AMQP Consortium, OASIS Advanced Message Queuing Protocol (AMQP) Version 1.0., 2012. OASIS Standard, Available At...
  • M. Bauer, et al., Final Architectural Reference Model for the IoT, 2013. http://www.iot-a.eu/public. (Accessed...
  • BehzadianmM. et al.

    A state-of the-art survey of TOPSIS applications

    Expert Syst. Appl.

    (2012)
  • BormannC. et al.

    CoAP: An application protocol for billions of tiny internet nodes

    IEEE Internet Comput.

    (2012)
  • ChangV. et al.

    Towards achieving data security with the cloud computing adoption framework

    IEEE Trans. Serv. Comput.

    (2016)
  • ChilipireaC. et al.

    Energy efficiency and robustness for IoT: Building a smart home security system

  • CugolaG. et al.

    Introducing uncertainty in complex event processing: Model, implementation, and validation

    Computing

    (2015)
  • EspositoC. et al.

    Trust management for distributed heterogeneous systems by using linguistic term sets and hierarchies, aggregation operators and mechanism design

    Future Gener. Comput. Syst.

    (2015)
  • C. Esposito, A. Castiglione, F. Palmieri, M. Ficco, K.K.R. Choo, A publish/subscribe protocol for event-driven...
  • EspositoC. et al.

    An investigation on flexible communications in publish/subscribe services, software technologies for embedded and ubiquitous systems

    Lecture Notes Comput. Sci.

    (2011)
  • EspositoC. et al.

    Smart cloud storage service selection based on fuzzy logic, theory of evidence and game theory

    IEEE Trans. Comput.

    (2016)
  • EugsterP.Th. et al.

    The many faces of publish/subscribe

    ACM Comput. Surv.

    (2003)
  • FilipI.D. et al.

    Microservices scheduling model over heterogeneous cloud-edge environments as support for IoT applications

    IEEE Internet of Things J.

    (2018)
  • GanzF. et al.

    Automated semantic knowledge acquisition from sensor data

    IEEE Syst. J.

    (2016)
  • Cited by (46)

    • Geospatial complex event processing in smart city applications

      2023, Simulation Modelling Practice and Theory
    • Blockchain-based authentication and authorization for smart city applications

      2021, Information Processing and Management
      Citation Excerpt :

      Such technology is used to guarantee consistency of replicated data in a distributed environment, even if being asynchronous, thanks to the used consensus algorithms, and to provide data immutability from intentional and unintentional manumissions thanks to the used crypto primitives. The blockchain has started to be used within the context of the Internet of Things and the smart city for the protection of data (Biswas & Muthukkumarasamy, 2016; Esposito et al., 2018; Wang et al., 2019), in many different applications, in particular to device security-related services, such as access control and authentication. Within smart cities, the available academic publications are mainly theoretical and at the architecture level, with few concrete implementations.

    • Smart architectures for evaluating the autonomy and behaviors of people with autism spectrum disorder in smart homes

      2021, Neural Engineering Techniques for Autism Spectrum Disorder: Volume 1: Imaging and Signal Analysis
    • Online heart monitoring systems on the internet of health things environments: A survey, a reference model and an outlook

      2020, Information Fusion
      Citation Excerpt :

      Moreover, methods of knowledge base theory, which include intelligent aggregation methods, such as RNA, genetic algorithms, and fuzzy logic. Finally, there are methods of evidence reasoning, which include Dempster–Shafer [75], theory of evidence and recursive operators. Feature fusion strategies provide an excellent means to combine heterogeneous sensor data using machine learning algorithms.

    View all citing articles on Scopus

    Christian Esposito is adjunct professor at the University of Napoli “Federico II”, and at the University of Salerno, where he is also a research fellow. He received the graduate-degree in computer engineering in 2006, and the Ph.D. degree in 2009 from the University of Naples “Federico II”. His main interests include mobile computing, benchmarking, aspects of publish/subscribe services, and security and reliability strategies for data dissemination in large-scale critical systems.

    Aniello Castiglione received the degree in computer science and the Ph.D. degree in computer science from the same university. He joined the Department of Computer Science of the University of Salerno in 2006. He serves as a reviewer for several international journals and has been a member of international conference committees. He has been involved in forensic investigations, collaborating with several law enforcement agencies as a consultant. His research interests include security, communication networks, information forensics, security, and cryptography. He is a member of various associations, including the IEEE and the ACM.

    Francesco Palmieri received the M.S. degree and the Ph.D. degree in computer science from the University of Salerno. He was an assistant professor at the Second University of Napoli. Currently he is an associate professor at the University of Salerno. His research interests include advanced networking protocols and architectures and network security. He has been the director of the Networking Division of the Federico II University of Napoli and contributed to the development of the Internet in Italy as a senior member of the Technical-Scientific Advisory Committee and of the CSIRT of the Italian NREN GARR. He serves as the editor-in-chief of an international journal and participates to the editorial board of other ones.

    Massimo Ficco received the degree in computer engineering from the University of Naples Federico II (IT) in 2000 and the Ph.D. degree in information engineering from the University of Parthenope in 2010. He is an assistant professor in the Department of “Ingegneria Industriale e dell’Informazione” at the Second University of Naples (SUN). From 2000 to 2010, he was a senior researcher at the Italian University Consortium for Computer Science (CINI). His current research interests include software engineering architecture, cloud computing, security aspects of critical infrastructure, and mobile computing.

    Ciprian Dobre Ph.D., has scientific and scholarly contributions in the field of large scale distributed systems concerning monitoring (MonALISA), data services (PRO, DataCloud@Work), high-speed networking (VINCI, FDT), large scale application development (EGEE III, SEE-GRID-SCI), evaluation using modeling and simulation (MONARC 2, VNSim). Ciprian Dobre was awarded a Ph.D. scholarship from California Institute of Technology (Caltech, USA), and another one from Oracle. His results received two CENIC Awards, and three Best Paper Awards, and were published in 6 books, 10 articles in major international peer-reviewed journal, and over 60 articles in well-established international conferences and workshops (these articles received more than 150 citations). He is local project coordinator for national projects CAPIM - Context-Aware Platform using Integrated Mobile Services, and TRANSYS Models and Techniques for Traffic Optimizing in Urban Environments.

    George Valentin Iordache, Ph.D. student, has scientific and scholarly contributions in the field of large scale distributed systems more specifically in the field of Grid scheduling (DIOGENESDistributed near-Optimal Genetic Algorithm for Grid Applications project) and Cloud scheduling (more specifically scheduling in the Cloud under Service Level Agreement Constraints). George Valentin Iordache received a Master of Science in Computer Science from the Stony Brook University, New York State, USA and was awarded both a Ph.D. fellowship -Renaissance Fellowship- from the Computer Science Department at Stony Brook University, New York, USA and a Ph.D. scholarship from Politehnica University of Bucharest). He was a Ph.D. student researcher at the Computer Science Department at Denmark Technical University for several months. He also worked for almost six years in the industry where he got proficient in different Microsoft technologies. His research results were presented in a book chapter, one article in a major international peer-reviewed journal, and 3 articles in well-established international conferences and workshops (these articles received around 60citations).

    Florin Pop Professor, Ph.D., Habil., received his Ph.D. in Computer Science at the University Politehnica of Bucharest in 2008 with “Magna cum laude” distinction. His main research interests are in the field of large scale distributed systems concerning scheduling and resource management (decentralized techniques, re-scheduling), adaptive and autonomous methods, multi-criteria optimization methods, Grid middleware tools and applications development (satellite image processing an environmental data analysis), prediction methods, self-organizing systems, data retrieval and ranking techniques, contextualized services in distributes systems, evaluation using modeling and simulation (MTS2). Florin Pop was awarded with two Prizes for Excellence from IBM and Oracle, three Best Paper Awards (in 2013, 2012, and 2010), and one IBM Faculty Award. He is involved in many national projects and international research projects (6 as project leader). He is active reviewer for several Journals (TPDS, FGCS, ASOC, Soft Computing, Information Sciences, etc.) and he has acted as Guest Editor for several special issues in FGCS and Soft Computing. The results were published in 4 books, more than 10 chapters in edited books, over 30 articles in major international peer-reviewed journal, and over 100 articles in well-established international conferences and workshops. He has long-running collaborations with several institutes from EU and around the world, including INRIA Rennes (KerData team), VU Amsterdam (The Netherlands), and University Marie and Pierre Curie Paris 6 (France).

    View full text