IoT-Edge Computing for Efficient and Effective Information Process on

Industrial Automation
2023 International Conference on Networking and Communications (ICNWC) | 979-8-3503-3600-9/23/$31.00 ©2023 IEEE | DOI: 10.1109/ICNWC57852.2023.10127492

M. Rudra Kumar1, B. Rupa Devi2, K. Rangaswamy3, M. Sangeetha4 and Korupalli V Rajesh Kumar5

Abstract— The Internet of Things (IoT) has transformed the key benefits of IoT in industrial automation include improved
way industrial systems operate by allowing physical devices, efficiency, increased safety, and cost savings [1].
machines, and systems to be connected and monitored over a One of the critical components of IoT in industrial au-
network. The Internet of Things has facilitated the development
of smart factories and industrial automation systems, in which tomation is Edge and Fog computing, which involves the
data from various sources is collected and analyzed in real-time processing of data at the edge of a network, rather than in a
to make informed decisions. IoT-enabled industrial automation centralized cloud. Edge and Fog computing reduce latency
systems can perform a variety of tasks, including predictive and increase reliability, as decisions can be made quickly
maintenance, process control, and energy management, among and locally, without the need for data to be transmitted
others. The main advantages of IoT in industrial automation
are increased efficiency, increased safety, and cost savings. The over long distances [2]. Another important aspect of IoT
generation of massive amounts of data from numerous sources in industrial automation is the use of sensors and other
or sensors is common in the IIoT world, but there are numerous connected devices to gather data. IoT sensors can be used
challenges. The goal of this study is to demonstrate how the to monitor various parameters such as temperature, pressure,
Internet of Things and AI enable Industrial Automation to vibration, and humidity, among others. This data can then be
develop in response to real-world issues.
analyzed and used to optimize industrial processes, improve
quality control, and reduce downtime.
I. INTRODUCTION IoT in industrial automation has the potential to revolu-
Industrial automation with Edge and Fog computing lever- tionize the way factories and other industrial systems operate,
ages the Internet of Things (IoT) to provide real-time data by providing a way to connect and monitor physical devices,
processing and decision-making. In this approach, Edge and machines, and systems over a network. IoT has enabled the
Fog devices are used to collect and process data from IoT development of smart factories and industrial automation
sensors and devices at the edge of a network, rather than in systems, providing benefits such as improved efficiency,
a centralized cloud. This reduces latency and increases relia- increased safety, and cost savings. With the continued growth
bility, as decisions can be made quickly and locally, without of IoT technology and the development of new applications
the need for data to be transmitted over long distances. and solutions, the possibilities for industrial automation are
Fog computing refers to the use of intermediate computing virtually limitless [3].
nodes in a network, between IoT devices and the cloud, to The combination of the Internet of Things (IoT), Edge
process and analyze data. Edge computing, on the other hand, computing, and Fog computing offer several advantages in
involves processing data directly on the IoT devices them- industrial automation, including:
selves, allowing for real-time decision-making and control. 1) Reduced Latency: By processing data locally at the
Together, Edge and Fog computing enable industrial automa- Edge or Fog nodes, the latency associated with trans-
tion systems to perform tasks such as predictive maintenance, mitting data over long distances to a central cloud is
real-time process control, and energy management, among reduced. This allows for real-time decision-making and
others. These systems also offer benefits such as improved control, improving efficiency and reducing downtime
efficiency, cost savings, and increased safety by allowing [3].
for real-time monitoring and control of industrial processes. 2) Increased Reliability: Edge and Fog computing reduce
Industrial automation systems that use IoT technology can the dependence on a centralized cloud, which can be
perform a variety of tasks such as predictive maintenance, a single point of failure. By processing data locally,
process control, and energy management, among others. The Edge and Fog computing increase reliability, ensuring
that critical processes continue even if the cloud is
1Professor, Department of Computer Science Engineering, GPCET,
unavailable.
Kurnool, A.P, India. 3) Improved Efficiency: By processing data locally, Edge
2Associate Professor, Department of Computer Science Engineering,
and Fog computing can reduce the amount of data
Annamacharya Institute of Technology and Sciences, Tirupati, A.P, India. transmitted to the cloud, improving overall efficiency
3Associate Professor, Department of CSE (DS), Rajeev Gandhi Memorial

college of Engineering and Technology, Nandyal, A.P India. and reducing bandwidth requirements.
4Associate Professor, Department of CSE, Fatima Michael College of
4) Better Security: Edge and Fog computing reduce the
engineering and technology, Madurai, TN, India amount of data transmitted to the cloud, reducing the
5 Assistant Professor, AI Research Centre, School of Business, Woxsen

University, Telangana, India. risk of data breaches and improving security. Edge and
If necessary, contact Korupalli V Rajesh Kumarat Fog nodes can also be secured with local firewalls,
[email protected] for the appropriate copyright notice.

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 26,2023 at 11:00:04 UTC from IEEE Xplore. Restrictions apply.
 encryption, and other security measures, providing an
additional layer of security.
5) Cost Savings: By reducing latency, increasing relia-
bility, and improving efficiency, Edge and Fog com-
puting can help reduce costs associated with indus-
trial automation systems. In addition, Edge and Fog
computing can reduce the need for expensive cloud
infrastructure, further reducing costs
II. METHODS AND REAL-TIME APPLICATIONS
A. IoT Enabled Industrial Automation
The Internet of Things (IoT) has numerous applications
in industrial automation, which can help improve efficiency,
reduce costs, and increase safety in industrial processes [4].
Some of the key applications include:
a) Predictive Maintenance: IoT sensors can be used
to monitor various parameters such as tempera-
ture, vibration, and pressure, among others. This
data can then be analyzed to detect potential prob-
lems and predict when maintenance is required,
reducing downtime and improving equipment re-
liability.
Fig. 1. Cloud, Fog, and Edge Computing Hierarchy
b) Process Control: IoT devices can be used to
monitor and control various industrial processes
in real-time. For example, IoT sensors can be B. Edge and Fog Computing in Industrial Automation
used to monitor the temperature, pressure, and
flow rate in a chemical plant, allowing for precise Edge and fog computing are distributed computing ar-
control and optimization of the process. chitectures that process and store data near their source.
c) Energy Management: IoT devices can be used Edge and Fog computing reduces latency, efficiency, and
to monitor and control energy consumption in dependability to overcome cloud computing’s drawbacks.
industrial processes, reducing energy waste and Edge computing involves processing data at the network’s
increasing energy efficiency. For example, IoT edge, usually near its source. Edge computing solves the
sensors can be used to monitor and control problem of sending big amounts of data to a central cloud
lighting, heating, and cooling systems, reducing over vast distances [6], [7].
energy consumption and improving energy effi- Fog computing enhances Edge computing by putting data
ciency. processing and storage closer to users. Even without cloud
d) Quality Control: IoT sensors can be used to mon- connectivity, fog computing allows real-time data processing
itor various parameters in real-time, providing and decision-making [5].
data that can be used to improve quality control. Well-known applications of Edge and Fog computing in
For example, IoT sensors can be used to monitor industries include:
the temperature, humidity, and pressure in a food • Industrial IoT: Edge and Fog computing can be used to
processing plant, ensuring that food is processed process and store large amounts of data generated by
under the right conditions. industrial IoT devices, reducing latency and improving
e) Supply Chain Management: IoT devices can be efficiency.
used to track products and components in real- • Predictive Maintenance: Edge and Fog computing can
time, providing visibility into the entire supply be used to process and analyze sensor data in real-time,
chain. For example, IoT sensors can be used to reducing downtime and improving equipment reliability.
monitor the location, temperature, and condition • Process Control: Edge and Fog computing can be used
of goods in transit, improving supply chain effi- to control and optimize industrial processes, reducing
ciency and reducing the risk of loss or damage. waste and improving efficiency.
f) Safety and Security: IoT devices can be used to • Energy Management: Edge and Fog computing can be
monitor and control various safety and security used to monitor and control energy consumption in in-
systems in industrial processes, reducing the risk dustrial processes, reducing energy waste and increasing
of accidents and improving overall safety. For efficiency.
example, IoT sensors can be used to monitor • Quality Control: Edge and Fog computing can be used
fire and gas detection systems, allowing for rapid to monitor various parameters in real-time, improving
response in case of an emergency [5]. quality control in industrial processes.

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 26,2023 at 11:00:04 UTC from IEEE Xplore. Restrictions apply.
 • Supply Chain Management: Edge and Fog computing
can be used to track products and components in real-
time, providing visibility into the entire supply chain.
• Safety and Security: Edge and Fog computing can be
used to monitor and control various safety and security
systems in industrial processes, reducing the risk of
accidents and improving overall safety.
By bringing data processing and storage closer to the
source of the data, Edge and Fog computing can pro-
vide significant benefits, improving efficiency, reducing
costs, and increasing safety in industrial processes.
III. IOT, EDGE AND FOG MODELS DEPLOYMENT -
INDUSTRIAL USE-CASE
IoT (Internet of Things), Edge Computing, and Fog Com-
puting are all related to the collection and processing of data
in various industries. Here the discussion is all about the Fig. 2. Seafood Processing Industry - States
manufacturing sector and how IoT, Edge, and Fog computing
technologies are helping industrial transformation.
Food Processing Industry: The IoT can monitor food safety 5) Packaging: Finished products are packaged to protect
by collecting and analyzing data from numerous production them from contamination and damage during storage
stages in the food processing industry. Edge computing and transport.
reduces latency and improves food processing decision- 6) Quality control: Finished products are inspected to
making by processing data closer to the source. To promote ensure they meet food safety and quality standards,
scalability and flexibility, fog computing distributes com- including testing for harmful bacteria and other con-
puting resources over multiple network nodes. In the food taminants.
processing business, fog computing can process sensor data 7) Distribution: Finished products are shipped to retail
and optimize operations [1]. stores, restaurants, and other customers.
8) Storage: Finished products may be stored in freezers,
A. Use Case: Sea Food Processing Industry - Conventional refrigerators, or other facilities until they are sold or
Processing Steps used.
India’s seafood business was worth $7 billion, with most Overall, the seafood processing industry is a complex
exports going to the US, Europe, and Japan. Shrimp, cut- and highly regulated industry that involves multiple steps
tlefish, squid, and fish are India’s main seafood exports. to ensure the safety and quality of the seafood products that
Millions of people, especially in coastal areas, work in the are produced.
industry, which boosts the economy. Raw fish is processed
into safe, edible items in the seafood processing sector. [8]. B. IoT, Edge, and Fog Computing Deployment at Seafood
The steps involved in seafood processing can vary depending Processing Industry
on the type of seafood being produced, but generally include
the following: Here, the main focus is on the Processing section. How
Seafood Processing - Conventional Model seafood is processed, and how the technology-driven ap-
Figure 2, shows the seafood import to export conditions proaches are helping to overcome the problem are explained
and steps, in this Processing - takes much time based on the in a detailed manner.
requirement and quantity of the seafood. The entire process is Processing — such as washing, chilling, skinning, gilling
explained with simplified abbreviations to understand much and gutting, filleting, shucking, boiling, smoking, salting,
clearer. drying, preserving, or canning — is performed on the product
1) Harvesting or sourcing raw materials: Seafood such (this step varies depending on which seafood species is
as fish, shellfish, and crustaceans are obtained from processed and if it is to be sold fresh or frozen). Storing.
fisheries, aquaculture farms, or other sources. In the Manual Processing center, a lot of manpower will
2) Sorting and grading: Raw seafood is sorted and graded be working to do the mentioned processes, shown in Figure
by species, size, and quality. 3.
3) Cleaning and filleting: Raw seafood is cleaned to Whereas Automated Seafood Processing Industries require
remove dirt, debris, and other impurities. Fish may also large machinery units to process the food, as shown in Figure
be filleted to remove bones and skin. 4.
4) Processing: Raw seafood is transformed into finished In this regard, our problem statement pointed out to
seafood products through a variety of processes, such bring innovation and integration into the Seafood Processing
as cooking, smoking, canning, freezing, and drying. Industry.

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 26,2023 at 11:00:04 UTC from IEEE Xplore. Restrictions apply.
 frequently. This approach is predicated on the examination
of sensory quality criteria mostly pertaining to the skin, eyes,
gills, and abdomen of raw seafood. This method, which
was designed for each species of seafood, evaluates factors
that undergo considerable change over the shelf life of the
seafood, and it is widely used to determine the quality
of seafood, even though it requires the support of other
methods.

D. Model Deployment - IoT Driven Models - with Sensors

Integration
• Biosensors: Biosensors are analytical devices that use
biological components such as enzymes, antibodies,
or microorganisms to detect and measure a specific
substance. In seafood quality testing, biosensors can be
used to detect pathogens, toxins, and other contaminants
in seafood. Common types of biosensors for seafood
quality testing include electrochemical biosensors, op-
tical biosensors, and DNA-based biosensors. These
Fig. 3. Manual Processing Section
biosensors offer advantages such as high sensitivity,
specificity, and rapid analysis, which can improve the
accuracy and efficiency of seafood quality testing.
• Computer Vision Applications: The effectiveness and
efficiency of the inspection process are dependent on
human inspectors because a significant portion of the
tasks involved is monotonous, boring, and extremely
time-consuming. This inspection work may also involve
automation through the use of computer vision. Because
of its many benefits, including much higher operating
speed, consistency, reliability, objectivity, and applica-
bility to industrial environments, computer vision is a
promising technology for use in applications related to
food safety and quality assurance. Although computer
vision systems are very simple to operate, they can per-
form a wide variety of labor-intensive jobs. The primary
responsibilities can be broken down into three cate-
gories: image acquisition, processing or analysis, and
recognition procedures. During the phase of analysis,
the various properties of the objects are extracted from
the images that were acquired, and during the phase of
Fig. 4. Seafood Processing Industry - States
recognition, the various techniques and algorithms that
are used in image processing are utilized to make the
final decisions.
C. Proposed Model
• IoT Devices - Edge and Fog Computing: IoT devices
Technology Driven Methods to overcome the Problems are connected devices that can communicate and share
at the Industry: In order to evaluate the quality of the data over the internet. Edge computing is a distributed
seafood and determine whether or not it has gone bad, computing paradigm where data processing and storage
reliable methodologies and techniques are required. The happen closer to the edge of the network, i.e., on the IoT
most important aspects of seafood quality are referred to device itself or on a nearby gateway. This approach can
as its sensory properties. These characteristics include the reduce latency and improve efficiency. Fog computing is
seafood’s appearance, colour, meat flexibility or texture, a similar concept but involves processing and storage on
aroma, and taste. A set of sensory characteristics, including a nearby network node or server, rather than directly on
as the brightness of appearance, meat colour, gill colour, the IoT device. Fog computing can provide additional
meat flexibility, and fragrance, could be used to classify resources and capabilities that may not be available on
freshness. These characteristics are based on changes in the the IoT device itself, such as more powerful processors
aforementioned factors. The quality index method (QIM) is and additional storage capacity. Both edge and fog com-
the approach to sensory evaluation that is used the most puting can improve the performance and efficiency of

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 26,2023 at 11:00:04 UTC from IEEE Xplore. Restrictions apply.
 IoT devices by reducing latency, increasing processing The responses from Biosensors are threshold values
speed, and reducing the amount of data that needs to based on the seafood quality. Biosensors are able to detect
be transmitted over the network. They can also improve volatile organic compounds, also known as VOCs, which are
security by keeping sensitive data closer to the source given off by seafood and are correlated with the freshness
and reducing the attack surface [9]. and quality of the seafood. The output of a biosensor that was
The Proposed model is shown in Figure 5. In the figure, used to test the smell quality of seafood would typically be a
B1, B2,.. indicated Biosensors, and C1, C2 indicates, cc signal indicating the presence or absence of specific volatile
cameras for computer vision applications. organic compounds (VOCs) that are indicative of freshness
or deterioration in the product. This signal may be presented
in the form of a digital readout, or it used to control other
devices or systems, such as alarms or refrigeration systems,
in order to assist in the preservation of the quality of seafood.
This data transmits to the IoT device and cloud for storage.
Figure 7 shows the biosensors’ responses on 20/01/2023
seafood processing states - 1 indicates quality was good, 0
indicates when the quality of the food is bad. Here, a group of
8 biosensors’ responses will be considered as 1 or 0. Edge
and Fog computing techniques were used to minimize the
data, and improve the effectiveness [8], [10].

Fig. 5. Proposed Model

IV. RESULTS AND DISCUSSIONS - PERFORMANCE

ANALYSIS
The proposed model was developed with Biosensors,
Computer vision, and IoT-Edge Fog Computing techniques
to make efficient and effective processing in the Seafood
processing industry.
Figure 6 shows the exact functional structure of the
proposed model. The same model was implemented at Fig. 7. Biosensors response - Sample Data
Navyandhra Fisheries, East Godavari Dist, Andhrapradesh,
and gathered the data and analyzed the responses. Computer vision techniques are used in seafood process-
ing to classify seafood types from images and find defective
seafood. This can be done using machine learning algorithms
that are trained on a large dataset of images of different
seafood types and levels of defects. The algorithms can then
be used to classify new images of seafood and identify any
defects. This can help improve the quality and safety of
seafood products, as well as increase efficiency and reduce
waste in the processing industry [11], [12].
Steps
• Image acquisition: Images of the seafood are acquired
using cameras or other imaging devices.
• Pre-processing: The acquired images are pre-processed
to remove any noise or unwanted information, such
as background or watermarks, using image processing
techniques like filtering, thresholding, and segmenta-
tion.
• Feature extraction: Features are extracted from the pre-
processed images to represent the seafood’s visual char-
acteristics. These features could include color, texture,
shape, or other relevant attributes.
Fig. 6. Proposed Model - Structure • Classification: The extracted features are used to clas-

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 26,2023 at 11:00:04 UTC from IEEE Xplore. Restrictions apply.
 sify the seafood into different categories using Convo- V. CONCLUSIONS
lution Neural Networks (CNN) model, shown in figure Transforming industrial systems with cutting-edge tech-
9. nology can result in innovation and streamlined operational
• Defect detection: In parallel, the pre-processed images changes in industrial automation. As part of the research,
are analyzed for defects such as bruises, deformities, or IoT Edge and Fog computing techniques, as well as Deep
parasites. Image processing techniques are used to high- Learning models, were used to bring innovation to the
light the defect regions and machine learning algorithms Seafood Processing Industry. The significance of the research
can be trained to identify these defects. is demonstrated by the findings. High-speed processing of
• Post-processing: The results of the classification and sensor data via IoT-connected edge and fog computing aids
defect detection can be combined and visualized to in better scenario processing. Industrial automation was
provide a comprehensive view of the seafood’s quality. aided by computer vision and deep learning-based object
Any false positives or negatives can be corrected using classification and segmentation.
post-processing techniques like threshold adjustment or
clustering. REFERENCES
[1] P. Nirmala, S. Ramesh, M. Tamilselvi, G. Ramkumar, and G. Anitha,
Computer vision applications provide an efficient and re- “An artificial intelligence enabled smart industrial automation sys-
liable way to classify seafood types and detect any defective tem based on internet of things assistance,” in 2022 International
seafood in seafood processing, improving food safety and Conference on Advances in Computing, Communication and Applied
Informatics (ACCAI). IEEE, 2022, pp. 1–6.
quality [13], [14], [15]. Figure 8 shows the response from the [2] G. Seeja, O. Reddy, K. V. R. Kumar, S. Mounika, and Others, “Internet
Computer vision model, classifying the Seafood categories of Things and Robotic Applications in the Industrial Automation
as Fishes - 1, Prawns - 2, and unknown - 0. Process,” in Innovations in the Industrial Internet of Things (IIoT)
and Smart Factory. IGI Global, 2021, pp. 50–64.
[3] S. D. Bhogaraju, K. V. R. Kumar, P. Anjaiah, J. H. Shaik, and Others,
“Advanced Predictive Analytics for Control of Industrial Automation
Process,” in Innovations in the Industrial Internet of Things (IIoT) and
Smart Factory. IGI Global, 2021, pp. 33–49.
[4] R. W. Coutinho and A. Boukerche, “Towards a novel architectural
design for iot-based smart marine aquaculture,” IEEE Internet of
Things Magazine, vol. 5, no. 3, pp. 174–179, 2022.
[5] K. V. R. Kumar, K. D. Kumar, R. K. Poluru, S. M. Basha, and M. P. K.
Reddy, “Internet of things and fog computing applications in intelligent
transportation systems,” in Architecture and Security Issues in Fog
Computing Applications. IGI Global, 2020, pp. 131–150.
[6] M. Adhikari, S. N. Srirama, and T. Amgoth, “Application offloading
strategy for hierarchical fog environment through swarm optimiza-
tion,” IEEE Internet of Things Journal, vol. 7, no. 5, pp. 4317–4328,
2019.
[7] A. A. Alli and M. M. Alam, “Secoff-fciot: Machine learning based
secure offloading in fog-cloud of things for smart city applications,”
Internet of Things, vol. 7, p. 100070, 2019.
[8] Z. Wang, S. Hirai, and S. Kawamura, “Challenges and opportunities
Fig. 8. Computer Vision-based Segmentation of Seafood in robotic food handling: A review,” Frontiers in Robotics and AI,
vol. 8, p. 433, 2022.
[9] S. D. Bhogaraju and V. R. K. Korupalli, “Design of Smart Roads-
A Vision on Indian Smart Infrastructure Development,” in 2020
International Conference on COMmunication Systems NETworkS
(COMSNETS). IEEE, 2020, pp. 773–778.
[10] D. Romanov, O. Korostynska, O. I. Lekang, and A. Mason, “Towards
human-robot collaboration in meat processing: Challenges and possi-
bilities,” Journal of Food Engineering, p. 111117, 2022.
[11] J. Zhang, H. Huang, G. Song, K. Huang, Y. Luo, Q. Liu, X. He, and
N. Cheng, “Intelligent biosensing strategies for rapid detection in food
safety: A review,” Biosensors and Bioelectronics, p. 114003, 2022.
[12] Z. Yang and Z. Ge, “On paradigm of industrial big data analytics: From
evolution to revolution,” IEEE Transactions on Industrial Informatics,
vol. 18, no. 12, pp. 8373–8388, 2022.
[13] C. Yu, X. Fan, Z. Hu, X. Xia, Y. Zhao, R. Li, and Y. Bai, “Segmen-
tation and measurement scheme for fish morphological features based
on mask r-cnn,” Information Processing in Agriculture, vol. 7, no. 4,
pp. 523–534, 2020.
[14] E. Prasetyo, N. Suciati, and C. Fatichah, “A comparison of yolo
and mask r-cnn for segmenting head and tail of fish,” in 2020 4th
International Conference on Informatics and Computational Sciences
(ICICoS). IEEE, 2020, pp. 1–6.
[15] P. Hridayami, I. K. G. D. Putra, and K. S. Wibawa, “Fish species
recognition using vgg16 deep convolutional neural network,” Journal
of Computing Science and Engineering, vol. 13, no. 3, pp. 124–130,
Fig. 9. CNN Based Seafood Classification 2019.

Authorized licensed use limited to: VIT University- Chennai Campus. Downloaded on May 26,2023 at 11:00:04 UTC from IEEE Xplore. Restrictions apply.