Sunteți pe pagina 1din 2

2018 IEEE/ACM Third International Conference on Internet-of-Things Design and Implementation

IoT Platform for Engineering Education and Research


(IoT PEER)–Applications in Secure and Smart Manufacturing
Terry Guo Damon Khoo Michael Coultis Marbin Pazos-Revilla
Tennessee Tech University Tennessee Tech University Tennessee Tech University Tennessee Tech University
Cookeville, TN 38505 Cookeville, TN 38505 Cookeville, TN 38505 Cookeville, TN 38505
NGuo@tntech.edu dtkhoo42@students.tntech.edu mrcoultis42@students.tntech.edu MPazos@tntech.edu

Ambareen Siraj
Tennessee Tech University
Cookeville, TN 38505
ASiraj@tntech.edu
Abstract—A small-scale IoT testbed called IoT PEER (IoT The system
Platform for Engineering Education and Research) was built at contains typical
Tennessee Tech for both research and education. Currently, this IoT elements
testbed is being used in studying security and smart manufac-
turing related topics. In this poster we report two Industrial IoT including gateways,
(IIoT) case studies: 1) Machinery Health Monitoring (MHM), and personal-area and
2) Intrusion Detection in Industrial/Manufacturing Environment. local-area wireless
System design and architecture along with preliminary results are connections,
provided. end nodes with
Index Terms—IoT, IIoT, smart manufacturing, digital thread,
cyber security, Machinery Health Monitoring (MHM), Predictive sensors, actuators,
Maintenance (PdM), Industrial Control System (ICS), Intrusion data base, cloud
Detection System (IDS). and users, etc.
Additional
I. BACKGROUND elements, such
Internet of Things (IoT) continues to grow and is permeating as 3D printer,
to our daily lives. The International Data Corporation (IDC) Intrusion Detection
Fig. 1. IoT PEER testbed architecture.
forecasts worldwide spending on IoT to reach $772 Billion in System (IDS) and
2018 [1], and nearly $1.4 Trillion in 2021 [2]. According to Industrial Control System (ICS) emulator, are included for
Statista, the installed base of IoT devices is forecast to grow to study of IIoT security. The microcontroller-based End nodes
almost 31 billion worldwide in 2020 [3]. Despite IoT applica- play a list of functions including 1) interface with sensors, data
tions grow steadily, IoT faces many critical issues, such as how acquisition and packaging, and hosting wireless transceivers.
to strengthen security and how to optimize performance. To ZigBee (IEEE 802.11.4), a popular communication standard
tackle these issues, it is desired to have an IoT research testbed adopted by major players in multiple industries, is used for
to test ideas and verify solutions. Currently, supported by two the links between the end nodes and gateways, where the
centers at Tennessee Tech, i.e., the Center for Manufacturing ZigBee links may be extended from single-hop to multi-hop
Research (CMR) and Cybersecurity Education, Research and for extended range at the cost of increased transmission delay.
Outreach Center (CEROC), this IoT PEER testbed is being WiFi/Ethernet is used to connect the gateways to the campus
used in studying security and smart manufacturing related network infrastructure. Microsoft Azure cloud services are
topics. Smart manufacturing is a global trend and a future employed for data visualization and analytics.
growth engine. In particular, IoT/wireless platform is one of B. Design Consideration
the core technologies for smart manufacturing [4]. Compared
to the consumer IoT that mainly provides convenience for our The following three aspects are considered in designing and
daily life, criticality is expected of industrial IoT (IIoT). Any implementing the IoT platform:
‘thing’ communicating over the Internet is vulnerable to being (a) Purpose–research and education are the main use of the
hacked, and IIoT devices are in no way exempt from this platform, which means flexibility and easiness are major
[5]–[7]. Our research focus is on (a) securing the networked concerns and we dont worry about commercialization;
industrial system and (b) collecting, analyzing and visualizing (b) Features–the platform is friendly to study of smart manu-
the machinery data. facturing and IIoT security;
(c) Cost and development cycle–for cost effectiveness and
II. S YSTEM D ESCRIPTION fast deployment, the platform is developed by using low-
cost commercially available devices and taking advantage of
A. Major Elements
existing facilities on campus.

0-7695-6387-2/18/$31.00 ©2018 IEEE 277


DOI 10.1109/IoTDI.2018.00038
For instance, with limitations in end node’s processing
power, buffer size, data packet size and transmission data
rate, a burst mode data stream with 8-bit resolution and peak
sampling rate 8 kb/s can be used, which means it takes
longer time to transmit a chunk of data at a transmission rate
much lower than the sampling rate. In addition to sensing
and control, network security is another essential function
that can be fulfilled at three levels: encryption at end nodes,
security management using a proxying machine that can be
a fog device such as a Raspberry Pi 3 module, and Intrusion
Detection System (IDS).
Fig. 3. Setup for study of industrial control system intrusion detection.
III. C ASE S TUDY
multivariate analysis and, Hurst exponent based, and some
A. Machinery Health Monitoring (MHM) open-source intrusion detection and prevention software like
One of many fundamental manufacturing tasks is main- Snort will be utilized. In particular, the IDS takes advantage of
taining machinery health. This can be done efficiently and three important resources: 1) inputs generated from sensing the
economically with the help of sensors and communication ICS, 2) network probing to monitor the over-the-air data pack-
networks. As IoT and data science advance rapidly, detect- ets in complicated electromagnetic (EM) environment, and 3)
ing faults in real-time and even predicting when equipment the cloud computation and intelligence. An ICS emulator will
failure might occur (i.e., Predicative Maintenance, PdM) are be built to emulate typical ICS behaviors and identify useful
achievable. Indeed, this technology can reduce maintenance parameters for intrusion detection. The ICS emulator and IDS
costs and downtime as well as preserve assets simultaneously. as part of IoT PEER testbed are essential for this case study.
The experimental setup is designed to collect sensor data Their structures and the relation between the two elements can
from machines wirelessly, store and visualize the data at be seen in the above conceptual diagram.
Microsoft Azure cloud, and perform Artificial Intelligence
IV. F UTURE W ORK
(AI) to detect or predict failures. Shown in Fig.2 are a) a
vibration data acquisition frontend mounted at a 3D printer 1) Identify cyber security vulnerabilities based on the digital
and b) the data interpreted in time-frequency domain at MS thread in manufacturing, and develop and examine counter-
Azure cloud, where the data is placed on an Azure WebApp measures by considering physics parameters observed in ICSs.
for remote access. Moreover, Node-RED is used to send email 2) Collect sufficient volume of data and apply big data
alerts triggered by some pre-defined conditions like excessive analytics to MHM and intrusion detection, taking advantage
vibration and failures of 3-D printing. of available AI resources.
ACKNOWLEDGMENT
Microsoft Azure cloud services are kindly sponsored by
Data
Microsoft Corporation. This work is supported partially by US
Acquisition unit
with XBee
National Science Foundation under the grant number 1560434.
Accelerometer

3D printer
extruder
R EFERENCES
[1] IDC Forecasts Worldwide Spending on the Internet of Things to Reach
$772 Billion in 2018. [Online]. Available: https://www.idc.com/getdoc.
jsp?containerId=prUS43295217
(a) 3D printer vibration [2] Worldwide Spending on the Internet of Things Forecast to Reach Nearly
$1.4 Trillion in 2021, According to New IDC Spending Guide. [Online].
data acquisition. Available: https://www.idc.com/getdoc.jsp?containerId=prUS42799917
(b) Vibration data visualization in [3] Statista. Internet of Things (IoT) connected devices in-
the cloud (MS Azure). stalled base worldwide from 2015 to 2025 (in billions).
Fig. 2. Vibration data acquisition and visualization in the cloud (MS Azure). https://www.statista.com/statistics/471264/iot-number-of-connected-
devices-worldwide/.
[4] H. S. Kang, J. Y. Lee, S. Choi, H. Kim, J. H. Park, J. Y. Son, B. H. Kim,
B. Intrusion Detection in Industrial/Manufacturing Environ- and S. Do Noh, “Smart manufacturing: Past research, present findings,
ment and future directions,” International Journal of Precision Engineering and
Manufacturing-Green Technology, vol. 3, no. 1, pp. 111–128, 2016.
The IoT testbed can be extended to emulate different smart [5] A.-R. Sadeghi, C. Wachsmann, and M. Waidner, “Security and privacy
manufacturing scenarios and develop security countermea- challenges in industrial internet of things,” in Design Automation Con-
ference (DAC), 2015 52nd ACM/EDAC/IEEE. IEEE, 2015, pp. 1–6.
sures. One particular objective is to safeguard the Industrial [6] M. Mowbray, “Detecting security attacks on the enterprise internet of
Control System (ICS). The big multidimensional data collected things: An overview,” Hewlett Packard Labs, April 2017.
from experiments on the testbed will be used for studying [7] S. R. Chhetri, N. Rashid, S. Faezi, and M. A. Al Faruque, “Security
trends and advances in manufacturing systems in the era of industry 4.0,”
intrusion detection in different ways such as outlier detection in IEEE/ACM international conference on computer aided design, 2017.
(unsupervised learning), anomaly based (supervised learning),

278

S-ar putea să vă placă și