Content uploaded by Walid El-Shafai
Author content
All content in this area was uploaded by Walid El-Shafai on Aug 28, 2022
Content may be subject to copyright.
Received: 28 October 2021 Revised: 9 June 2022 Accepted: 28 June 2022
DOI: 10.1002/ett.4607
SURVEY ARTICLE
Climate-smart agriculture using intelligent techniques,
blockchain and Internet of Things: Concepts, challenges,
and opportunities
Rania A. Ahmed1,2 Ezz El-Din Hemdan3Walid El-Shafai2,4 Zeinab A. Ahmed1
El-Sayed M. El-Rabaie2Fathi E. Abd El-Samie2
1Climate Change Information Center &
Renewable Energy & Expert System, Giza,
Egypt
2Department of Electronics and Electrical
Communications Engineering, Faculty of
Electronic Engineering, Menoufia
University, Menouf, Egypt
3Department of Computer Science and
Engineering, Faculty of Electronic
Engineering, Menoufia University,
Menouf, Egypt
4Security Engineering Laboratory,
Department of Computer Science, Prince
Sultan University, Riyadh, Saudi Arabia
Correspondence
Walid El-Shafai, Department of
Electronics and Electrical
Communications Engineering, Faculty of
Electronic Engineering, Menoufia
University, Menouf 32952, Egypt.
Email: eng.waled.elshafai@gmail.com
Abstract
The Internet of Things (IoT) is an important technology that provides efficient
and dependable solutions in a variety of domains, such as smart agriculture and
climatic change. It integrates billions of smart devices that can communicate
with one another and gives solutions to automatically maintain and monitor
smart agricultural and environmental fields. The combination of IoT, Artifi-
cial Intelligence (AI), and blockchain technology will allow us to transform
smart agriculture into the Internet of smart agriculture, providing greater con-
trol, management, and security in supply-chain networks. This paper presents
an overview of the technologies in the domains of IoT, Climate-Smart Agricul-
ture (CSA), AI, Machine Learning (ML), and blockchain. In addition, the paper
presents several approaches for integrating IoT with CSA data analysis. Both AI
and blockchain are adopted for efficient CSA systems. The paper is concerned
with the combination of three recent technologies: IoT, ML, and blockchain
to serve the CSA applications. The challenges and opportunities of combining
these technologies to serve CSA are also discussed in the paper.
1INTRODUCTION
Climate change increases uncertain and extraordinary weather that affects farming. Farmers regularly face crop damage
due to climate change. This change affects the quantity and quality of plants. Hence, the crucial issue is developing a
monitoring system that helps to solve this problem. Climate monitoring through sustainable land and water management
can also increase productivity and enhance crops and the quality of plants.
Optimization of sustainable agricultural processes is an important issue. The CSA technology could be the key to
improved and increased productivity, while decreasing the natural footprint. Researchers and policymakers can be driven
by climate change, the global population increase, the food security needs, and the decrease of human labor in agribusi-
ness. The IoT helps ranchers improve productivity, quality, and efficiency by providing data access, decision-making
assistance, and the use of sensors and actuators that integrate products, knowledge, and services that provide efficient
production. Furthermore, smart agriculture with IoT improves geospatial data and real-time events, and it is the driving
strength towards the agriculture sector supportability. This paper presents a comprehensive survey of several technologies
that can be used to improve CSA and enhance agriculture as shown in Table 1.
Trans Emerging Tel Tech. 2022;e4607. wileyonlinelibrary.com/journal/ett © 2022 John Wiley & Sons, Ltd. 1of27
https://doi.org/10.1002/ett.4607
2of27 AHMED .
TABLE 1 Related work on IoT, ML, and blockchain in agriculture
Ref paper Year Field Contribution
[1] 2015 IoT This paper provides an overview of the IoT enabling technologies, various applications,
different protocols, and IoT challenges.
[2]2018 IoT, ML, andsmart city This paper presents a taxonomy of ML algorithms, challenges of ML for IoT data
analytics and an application of SVM to the smart city.
[3] 2020 IoT This paper presents a taxonomy for classifying the architecture and protocols used in
IoT, in addition to technical challenges such as security, privacy, scalability, and
energy efficiency. It covers the recent research work on such challenges.
[4]2019 IoT, and smart farming This paper covers the components of IoT, network architecture and layers, network
technologies such as cloud computing, big data storage used in smart farming,
security issues in IoT, challenges and policies made by several countries to
standardize IoT-based agriculture.
[5] 2009 WSN, RFID, and
agri-food sector This paper reviews the technical and scientific state-of-the-art WSN technologies and
standards for wireless communications. It presents the different systems such as
ZigBee-based WSN and passive, semi-passive and active RFID in the agri-food sector.
[6]2020 IoT, ML, AI, and
blockchain This paper presents an overview of the IoT technology and its application areas. It covers
the primary security issues of IoT through ML, AI, and blockchain technologies.
[7] 2016 ML, and big data This paper presents the basics of ML and data mining and names some tools, and
platforms for dealing with big data in ML.
[8]2020 ML,WSN,IoT,and
precision agriculture This paper provides a review of the applications of various ML algorithms in WSN data
analytics within the agricultural ecosystem. It gives a case study on an IoT-based
data-driven smart farm prototype system.
[9] 2019 ML, and IoT This paper provides an overview of types of ML, methods, and applications of ML in IoT
[10]2020 Blockchain, and IoT This paper presents a survey explaining different blockchain-based consensus methods
implemented in IoT networks. It covers the needs to reduce the computational
power, time for the consensus methods and the different types of blockchain.
[11] 2019 Blockchain, and IoT This paper presents a classification of blockchain technologies into four layers and
presents a study of the consensus strategies, network, and applications of blockchain
in IoT.
[12]2018 Blockchain, smart
cities This paper gives a comprehensive survey of the literature work involving blockchain
technology applied to smart cities. It also discusses background knowledge, related
works and challenges.
[13] 2017 Blockchain This paper provides concepts, structure, types, and protocols of blockchain. It also
discusses some applications, security issues and challenges.
[14]2019 Blockchain, and IoT This paper presents a comprehensive survey of blockchain technologies with IoT
applications. It provides limitations, characteristics and security analysis on IoT. It
also covers blockchain technologies, structure and security analysis. It provides
applications of blockchain for IoT and future research directions.
[15] 2018 Blockchain This paper provides the blockchain taxonomy, consensus algorithms and applications.
It also discusses technical challenges and opportunities with future directions in the
blockchain technology.
[16]2019 IoT, and healthcare This paper presents a review that determines the main application areas of IoT in
healthcare, components of IoT architecture, security, privacy issues in IoT, and
finally challenges of IoT in healthcare.
Presented
survey IoT, ML, AI,
blockchain, and CSA Our paper presents an overview of the technologies in the domains of IoT, CSA, AI, ML,
and blockchain. It also presents several approaches for integrating the IoT withCSA,
and data analysis that uses AI and blockchain for efficient CSA systems. It provides a
combination of three recent technologies: IoT, ML, and blockchain to serve the CSA
applications. The challenges opportunities, and future works are also discussed.
AHMED . 3of27
1.1 Objective and contributions
The main objective of this paper is studying and adopting advanced technologies such as IoT, AI, ML, and blockchain
that can be applied in CSA. This can improve the agriculture to realize the automation of agricultural technology in
order to provide sustainability and productivity without affecting the environment. In addition, the paper covers the CSA
concepts and applications. It provides a vision to combine IoT, ML and blockchain to enhance CSA. This gives a roadmap
to farmers and experts who need to learn these smart agricultural technologies that can be used in many fields, such as
water management and pest control, to meet the challenges of sustainability and climate change.
The main contributions of this paper can be summarized as follows:
•It provides an overview of the IoT, AI, and blockchain for developing smart CSA systems.
•It presents and explores approaches for integrating the IoT with CSA.
•It explores the climate data analysis using intelligent methods for smart CSA systems.
•It introduces a discussion of the use of blockchain for efficient CSA systems.
•It presents a security study of IoT-enabled smart agriculture.
•It presents a combination of blockchain technology with IoT devices and ML in CSA applications.
•Finally, it introduces the challenges and opportunities for smart agriculture.
1.2 Paper organization
The structure of this paper is summarized as follows. Section 2provides an overview of the IoT, while Section 3
presents the use of AI in agriculture. Section 4provides the concept of blockchain, while the concepts of CSA sys-
tems are presented in Section 5. Section 6provides the IoT use for the smart CSA systems, while Section 7covers
the data analysis using intelligent methods for smart CSA systems. The importance of blockchain for developing
smart CSA systems is presented in Section 8. Section 9provides a discussion of the security in IoT-enabled smart
agriculture. The combination of blockchain technology with IoT devices and ML is presented in Section 10.Thechal-
lenges and opportunities for smart agriculture are introduced in Section 11. The paper conclusion is presented in
Section 12.
2INTERNET OF THINGS
Internet of Things (IoT) is an important technology that offers an efficient and reliable solution for large applications
and several domains, especially in CSA. It also develops and integrates multiple devices and embedded technologies and
produces a smart environment that saves time, energy, and money. IoT includes four main components: sensing devices,
processing networks, data analysis and control and monitoring systems. They cooperate and work with each other to get
the best and most efficient performance1as follows (Figure 1):
•Sensing devices: They allow collecting data from the environment and transferring it to a local database server or
remotely to the cloud-database server. The main function of a sensing device depends on the sensor used for gathering
data and the actuator used to control the system.
•Processing networks: These networks use devices that can transfer messages or information. There are several meth-
ods that depend on devices to communicate, such as Bluetooth, Long Range Wireless Area Network (LoRaWAN),
Long-Term Evolution (LTE), Wi-Fi, and ZigBee. The main function of these networks depends on devices like routers,
repeaters, access points, and gateways to transmit data in an accurate and secure way.
•Dataanalysis: Data analysis depends on several types of servers with large memory and high speed to make the analysis
process of large data received from different kinds of resources in order to get the best decision for the application of
interest in a minimum time. The algorithms process, calculate and handle data according to its characteristics. IoT
works on a variety of sources of data in real time. This data can be analyzed by different algorithms. In addition, the
data needs to be available on a dynamic website, to allow real-time data analytics.
4of27 AHMED .
FIGURE 1 Components of IoT structure
•Control and monitoring systems: The control and monitoring systems cooperate and work with previous components
to get the best and most efficient performance and get the best decisions for applications. After that, the control system
usesthe decisiontogive the commandtocontrol the application.Thiscan bemadeby sending feedbacktosome devices
in the system as actuators, motors, and valves to control the system.
The communication in IoT can be performed between sensor nodes in Wireless Sensor Network (WSNs) called
Device-to-Device (D2D) networks. In the second communication, the communication device sends all collected data to
the servers in the so-called Device-to-Server (D2S) connection.2The last communication type is represented in servers
transmitting data between each other in a Server-to-Server (S2S) manner.1
2.1 Architecture of the IoT system
The IoT needs a flexible layered architecture to be able to connect a large number of heterogeneous devices through the
Internet. In this part, we discuss the seven-layer architecture of IoT17 as follows (Figure 2):
•Physical devices and controllers layer: A layer of recognition or equipment gathers and sends data from the physical
world to the other layer. This layer incorporates physical objects and sensors to detect and record natural data such
as humidity, temperature, motion, water quality, and so forth. In addition, this layer converts analog signals collected
from the sensor to transmit them as digital signals to the next layer.
•Connectivity layer: This layer is utilized to connect with different IoT components by means of interconnection frame-
works such as switches and routers. In addition, it exchanges data collected from the sensors to be processed in the top
layerby usingdifferentnetworking technologies suchas Wi-Fi, ZigBeeand so on.It incorporatesinformation exchange
from physical gadgets to the cloud or other devices.
•Edge connecting layer: This layer is used for edge computing. Layer 3 requires data from the connectivity phase and
makes it suitable for storing and processing information at a higher level. It also helps to reduce system latency by
applying some analysis on data in edge devices before transmission to the cloud and reducing the data volume in real
time.
AHMED . 5of27
FIGURE 2 Architecture of IoT systems
•Data accumulation layer: It is critical where data is put away. So, most data will need to be conveyed to the cloud. The
data accumulation stage serves as a link between event-driven data generation and query-driven data consumption.
Enormous machines with high efficiency and computing power are used to analyze these huge amounts of data.
•Data abstraction layer: This layer takes data from various sources and converts it to an application-specific format in a
comprehensibleand efficient manner.Inaddition, itmakes data aggregationin aplace foreasy accessby theapplication
layer.
•Application layer: It is applicable to information derived from many IoT applications by using different technology
stacks and operating systems.
The information is analyzed to give solutions using ML. This covers multiple IoT applications such as therapeutic
care, smart agriculture, smart city, smart buildings, smart networks, associated cars, and so forth.
•Collaboration and processes layer: The prepared data within the lower levels are combined for commercial purposes at
this stage, and application handling and collaboration are displayed to clients. Collaboration, individuals, businesses,
and decision-making forms based on IoT-derived information are all included.
2.2 Requirements for IoT
The IoT confronts several difficulties such as security and privacy, device heterogeneity, data management, interoperabil-
ity, and energy efficiency3as follows:
•Identification and scalability: The IoT has a considerably broader scope than the Internet of computers because of the
large number of items and devices that are linked together. This needs scalability issues at different levels.
•Self-organizing capability: Configuring smart objects, organizing, arranging, and adjusting of circumstances of com-
puters should be implemented without the need for human mediation. The complexity and large scale of IoT can
be met by devices that monitor themselves without external intermediation, where smart distributed objects have
self-capabilities, including service discovery automatically to adapt protocol behavior.
•Interoperability: In the IoT, different types of smart things have distinctive capabilities in terms of computation, com-
munication, transfer speed, energy availability, and so on. A common standard is essential to enable communication
and participation between these different types.
•Data management: Sensor systems create expansive sums of data in an IoT environment, which can be put away on
central hubs or servers. The server must efficiently provide the needed service to any device.
6of27 AHMED .
•Security and personal privacy: Security must be taken into consideration in the engineering design of IoT. Due to
restricted transmission capacity, security calculations must be simple with the least number of message trades.
•Energy efficiency: Energy efficiency is an important issue that needs to be addressed in IoT. It has three phases: power
harvesting, power saving, and power consumption. Hence, minimizing energy consumption for IoT devices that have
limited battery power is important.
2.3 IoT communication protocols
Agricultural data is collected and encapsulated via communication protocols. Farmers can communicate more effectively
using these protocols, to get suitable decisions for smart agriculture, and enhance the growth of crops and the qual-
ity of production. Several protocols use Internet-related technologies. These protocols include Bluetooth, IEEE 802.11
WiFi, LoRaWAN, SigFox, WiMAX, and ZigBee to enable long-distance communication. The taxonomy of various wireless
communication protocols is shown in Table 2. The different protocols can be summarized as follows:
•IEEE 802.11 WiFi: It is a set of WLAN communication standards that includes 802.11ac, 802.11n, 802.11g, 802.11b,
and 802.11a. There are different bandwidths for all these standards. All these benchmarks work with totally different
transmission capacities that are 5, 2.2, 2.4/5, and 60GHz, respectively. Data exchange provides data rates from 1 Mb/s
to 7 Gb/s. The communication extends from 20 (indoor) to 100m (outdoor).18
•LoRaWAN: It is a protocol that supports long-distance communication, and the primary reason for employing this
protocol is to guarantee interoperability among different users. The objective is to enhance agricultural efficiency and
anticipate the problems.4
•WiMa AX: It offers broadband multi-access connectivity, including portable and mobile communication via wireless
or wired connectivity.
•LR-WPAN: It is a Wireless Personal Area Network (WPAN) with a low rate that specifies high-level communication
standard specifications. Data exchange rate is 40–250kb/s. This standard provides communication services with low
cost and low speed. LR-WPAN is generally utilized for indoor farming as for domestic plants or in small ranches.19
•ZigBee: It is a set of specifications at the top of the IEEE 802 standards. These specifications are used to exchange data
from device to device in the network with low power data rates.20
TABLE 2 The taxonomy of various wireless communication protocols
Wireless
technology Network type Operating
frequency Max. range Max data rate Power Security
IEEE 802.11
WIFI WLAN 5, 2.2, 2.4/5 GHz 20–100 m 1 Mb/s to 7 Gb/s High power
consumption -
LoraWAN Star Mesh Network 868 MHz (EU)
900 MHz (USA) Short range
10–100m 0.3–50 kbps Low power
consumption AES
WiMAX Metropolitan Area
Network (MAN) 2–66 GHz <50 km 1 Mb/s to 1 Gb/s Medium power
consumption -
LR-WPAN WPAN 868MHz (EU)
900MHz,
2.4 GHz
10–20m 40–250 kbps Low power
consumption -
ZigBee Star, Mesh Cluster
Network 2.4 GHz Short range
10–100m 250 kbps 30mA Low power
consumption AES
SigFox Start Network 868 MHz (EU)
902 MHz (USA) Long range 10 km
(URBAN) 50km
(RURAL)
100 bps (UL),
600 bps (DL) 10–100 mW Partially
addressed
Bluetooth Star–Bus Network 2.4 GHz Short range
∼15–30m 1Mbps 30mALowpower
consumption E0 Stream
AES-128
AHMED . 7of27
•SigFox: It is an ultra-narrowband standard with a low data rate. It is also suitable for IoT applications. A geolocation
system can be built using the SigFox network that localizes animals pasture. The system helps farmers to find the
locations of their cattle and increase their efficiency.21
•Bluetooth: Bluetooth is a Personal Area Network (PAN) with low power and low range. Bluetooth powers numerous
IoT devices that are found in agricultural applications; such as color sensors working with bluetooth technique.5
2.4 Classification of computing technologies in IoT
Cooperation between IoT and cloud computing provides a widespread access to shared resources. There are several types
of computing that are used in IoT. We discuss four main types of computing as shown in Table 3:
•Edge computing: The first type of computing is edge computing, which is used in sensors, controllers, actuators, tags,
tag readers, communication components, gateways, and physical devices. Edge devices can store local data, filter and
clean data, and enhance security.
•Fog computing: The second type is fog computing, which has a middle position between cloud and edge and connects
the cloud and edge resources. The limiting computing, storage, and network services are provided by fog computing
with more data filtering and smart logic processing for data centers.22
•Cloud computing: Cloud computing is essential for processing and data analysis. Cloud computing has a high latency
with a high balancing load, which aids in manipulating IoT data, because most manipulations should be performed at
high speed.23 Cloud computing has picked up much considerations to support various services, including Software as
a Service (SaaS), Platform as a Service (PaaS), and Infrastructure as a Service (IaaS).2
•Distributed computing: The last type is intended for high-volume data processing in the case of generating big data by
sensors. It is called distributed computing, and it is used in some applications.24
2.5 Analytic techniques for IoT
There are various types of techniques to analyze data collected and used in IoT applications.
•Predictive analytics: This type of analytics depends on chronicled data and employs ML techniques or complex calcu-
lations to produce a model that describes the behavior or pattern and aids in predicting patterns that are possible in
the future. Classification-based models and regression-based models are the two main types of predictive models.25
TABLE 3 Taxonomy of various IoT cloud platforms
No. IoT cloud
platforms Cloud service
type Application
development Monitoring
management Visualization Cost
1 Arrayant Connect TM SaaS √√√Low
2Ubodots Public √ √ √ Free
3 Phytech Private (IaaS) √√√Pay per access
4Xively Public (IoTaaS) √ √ √ Free
5 ThingSpeak Public √√√Free
6ThingsBoard Public √ √ √ Low
7 Nimbits Hybrid √√√Free
8Google server Public √ √ √ Pay per access
9 SeeControl IoT Private (PaaS) ××√Low
10 Oracle IoT cloud Public (PaaS) × × √High
11 KAA Public √√√Free
12 Plotly Public Public √ √ √ Free
13 Thethings.io Public √√√Free
8of27 AHMED .
•Descriptive analytics: Any IoT system can assemble data from many to thousands of intelligent devices and send it
to the cloud. Using advanced ML techniques based on historical data, detailed information about past events can be
obtained.26
•Prescriptive analytics: Based on the investigation and analysis of the data, this type of analytics suggests how to respond
to any future events. This type of analysis forecasts the future state and provides recommendations for making use of
theoutcome. Itis similarto afuture analysisscenario that gathers both descriptive and predictive analyticadvantages.27
•Adaptive analytics: This analysis adjusts the predictive analytic outcome obtained during actual implementation. The
adjustment or optimization of the outcome is based on the data recent history and data correlations. This improves the
performance of the model and reduces errors.
2.6 IoT security
Securing the IoT is an exceedingly essential and complex task that lies in the top layer of its devices, networks, and
applications. So, security design must be done right by considering factors in the planning phase. Numerous IoT devices
have security difficulties due to the very small and low power of these devices. IoT security requirements are:
•Confidentiality, where the data is accessible only by authorized users.
•Integrity, where the content of data that is received from another device or stored to the device is the same without
change by anyone.
•Data freshness, where it is vital to guarantee that each message is new within the network.
•Non-repudiation, which implies that any node prevents denial of sending a message that was sentprior.
•Authentication, which means that any device that communicates should have an identity of communication.
The IoT sensors and devices gather highly detailed data, but this could negatively affect the stolen or compromised
data due to insecurity. In addition, the complexity of IoT environments expands the IoT system capabilities, but this
needs more cost because of the wider attack area. Centralizing IoT systems and gathering data by the main database
from thousands of devices can reduce the cost of the system, but produce a risk of wide attack connected to a single
root. Therefore, the developers, researchers, manufacturers, and companies should understand IoT attack areas to deploy
IoT in their organizations in a safe way and prevent the data and systems from attacks. Security attacks on the IoT have
different types as follows:
•Eavesdropping/sniffing: The attacker tries to get useful data such as usernames, passwords, and node identifiers. RFID
devices have tendencies to be attacked through eavesdropping.
•HELLO flood: The attacker converts every node to be as a parent to another node by marking itself. So, the nodes that
have been attacked redirect the packets to each other. This ensures that a large number of nodes will be out of range,
and this causes a lot of packets to be lost.
•Denial-of-service (DoS) attack:It isa commonassault utilizedin IoT applications, such as batterydraining by an attacker
sending lots of packets. The low-end devices are the weakest and most suitable target for this attack. The attacker
uses a data traffic stream to access a device connected to the network or a huge volume of network packets through
infrastructure and can stop the normal performance of the device or refuse to route messages, or redirect them to
unwanted locations.28
•Intrusion detection system (IDS): It is a system by which the aggressor exerts control over network activity. There are
several types of IDS attacks, including abuse detection, irregularity detection, network-based IDS, and host-based
IDS.29
•Malicious node: Assault can be possible, because smart devices have a heterogeneous nature in IoT networks. Within
the network, this attack generates virtuoso nodes or fake nodes that are difficult to be known.30
•Jamming attack: It is a set of DoS attacks in which the aggressor attempts to influence the communication channel
creating radio interference that exhausts IoT devices.6
AHMED . 9of27
•Power analysis: This assault prevents the fundamental cryptographic algorithm from executing in the node by gaining
the computational power of this node. This prevents trust-building among the nodes in an IoT network and prevents
the node from having its privacy.31
•Wormhole attack: This assault produces a tunnel between two nodes connected to each other at the 6LoWPAN layer.
The attacker could record the packets at one location, and then tunnel them to a different location.32
•Distributed DoS attack: It is the method, where the server cannot serve the smart nodes in the network because of the
inaccessibility of the server. This attack not only disables the network but also prevents it from being accessible to a
very large network.33
•Man-in-the-middle attack: This attack alters the communication between two parties that communicate directly with
each other. Its process is to transfer modified messages by an assailant during transmission over the unreliable
channel.34
•Sybil attack: The attacker can be in many places at once by using a single node presenting multiple identities to other
nodes in the network
•Routing attacks: By using many types of attacks such as Sybil, HELLO flood, and wormhole attacks, the attacker can
misdirect, redirect, or drop packets.
3ARTIFICIAL INTELLIGENCE
Smart agriculture can be managed and controlled using various tools, including IoT, wireless communications, WSN,
Deep Learning (DL), image processing, robotics, ML, Convolutional Neural Networks (CNNs), and Artificial Intelligence
Neural Networks (AINNs).
The AI can help farmers arrive at efficient land utilization with sustainable resources. This happens by analyzing
water utilization, soil conditions, temperature, energy usage, and climate conditions collected from cultivated farms. The
analyzed data can be used to predict yield and detect diseases to get better decisions and management. The AI is defined
as the science and engineering of creating intelligent machines, particularly intelligent computer programs. There are
many ways of implementing AI, as shown in Figure 3. The first branch is ML that is selected to be discussed in detail in
this paper.
The AI contains many subsets. ML is one of these subsets with rapid and critical advancements.35 ML offers assistance
to solve complex issues for people to unravel by moving the burden of decision making to the algorithm.36 There are
three main categories of ML algorithms that are classified based on the properties, learning style, and data usage of these
algorithms. They include reinforcement learning, unsupervised and supervised algorithms.7The ML main categories are
shown in Figure 4.
The supervised and unsupervised learning are widely applied in IoT smart data analysis.37
•Supervised learning: The algorithm is trained using the labeled input dataset (training data). This makes suitable fore-
casts on the input data and improves its estimates using the ground truth and emphasizing the algorithm outcomes
to a desired level of precision. The supervised learning goal is to learn how to foresee the fitting output vector for a
given input vector.37 Supervised learning also has common applications, such as classification and regression tech-
niques. Support Vector Machines (SVMs) are prevalent supervised learning models, primarily utilized for binary and
multi-class classification.
•Unsupervised learning: Unsupervised learning does not require labels for the training set, and it is hard to define the
objective of unsupervised learning. Unsupervised approaches attempt to find some form of trends or some structure
in the training data. Clustering is a common example of unsupervised learning, which uses the input data to identify
sensible clusters of similar samples.38
•Reinforcement learning: Reinforcement learning addresses the problem of determining the suitable action or combi-
nation of actions to take in given circumstances to maximize payoff. Reinforcement learning algorithms have been
utilized for designing routing protocols in IoT systems and WSN networks to reduce Q-Learning power consumption.
Likewise, there are several functions for ML, such as:
•Classification: Classification in ML refers to a predictive modeling problem in which class naming is expected for
a particular instance from the input data. It classifies the output variable into categories. The model produced by
10 of 27 AHMED .
FIGURE 3 Artificial intelligence branches
Machine
Learning
Supervised
Learning
k-nearest
Neighbor (k-NN)
Decision Tree (DT)
Bayesian Interface
Support Vector
Machine (SVM)
Reinforcement
Learning
Unsupervised
Learning
Principal
Component
Analysis (PCA)
Self-Organizing
Maps (SOM)
k-means
Clustering
FIGURE 4 Machine learning main categories
classification training dataset is used to estimate the best map examples of input data to determine class labels. In
addition, the classification model attempts to draw a few conclusions from monitored values. Finally, the classification
modelpredicts the value of one or moreresults basedon the inputs provided. Thereare severalclassification techniques
such as SVM, Random Forest (RF), Decision Tree (DT), K-Nearest Neighbors (K-NN), and Naïve Bayes (NB).
AHMED . 11 of 27
Gathering
Data Model
Building
Cleaning
Data
Visualizing
Information
Evaluating
Output
Gaining
Insights
FIGURE 5 Machine learning workflow
•Regression: Regression is a technique that predicts continuous responses such as time-series sensor data, and fluctu-
ations in measurements by using supervised ML. Regression algorithms have two main types, which are linear and
nonlinear regression. The linear relationship between independent and dependent variables is the foundation of linear
models. The examples of linear and nonlinear regression include SVM regression, Gaussian Process Regression Model
(GPRM), Generalized Linear Model (GLM), and ensemble methods.8
•Clustering: It is the foremost common unsupervised learning model, and its applications expand to exploratory data
analysis.
The workflow for ML begins with gathering data from different resources or sensors. The second step is cleaning
data from errors and correlation for generating homogenous data. The model generation is an important process that is
performed with correct calculations to get a suitable model. Gaining insights from the outcomes is the next step after
arriving at the model that must be evaluated and checked. The last step is to look at the information as a whole and arrive
at a suitable decision.9The ML workflow is shown in Figure 5.
4BLOCKCHAIN
The concept of blockchain refers to a decentralized, distributed ledger that stores time-stamped transactions between
devices in a peer-to-peer network. The blockchain design can be a growing stack of records, called blocks. The first block
in the chain is named a genesis block, which has no parent block. Each block has a unique hash value that identifies the
block identity.Any block has a header and a body, and all transactions that have been affirmed and approved are found
in the block body. The components of the block header can be as follows:
•Block version indicates which consensus protocol can be used and can be defined as a software version number.
•Markle tree root hash is a hash code of a binary tree utilized to confirm the hash code that recognizes all block
transactions.
•Timestamp is utilized to track block creation and statically update it to ensure block integrity.
•N-bits point to the valid block by identifying the target threshold of the hash code.
•Nonce is a randomly generated number that is used once in a cryptographic communication. It contains a four-byte
field that mainly starts with “0s” and grows with each hash computation.
•Parent block is a hash code with 256 bits, and it points to the previous block.
4.1 Classification of blockchain
Blockchain topology can be categorized into three groups: public, private, and consortium.10,11 Each category has method
of accessing the ledger by the user and allowing share in the consensus.12 The consensus protocol is important for the
stability of ledger and distributed access in any blockchain (Table 4).
•Public blockchain: It is a distributed database that is impervious to tampering and does not rely on a single entity for
control, and also all users can share and access it. The existing blocks can add new records as long as all within the
network affirm the new block. In addition, it is not feasible to adjust or delete blocks once they are recorded.13 This
type has advantages. It is easy to access, and the reading and writing operations are open for all participants on the
network, but this needs complex security that requires large energy and time.
12 of 27 AHMED .
TABLE 4 Comparison of blockchain types
Blockchain type Advantages Disadvantages Application examples
Public blockchain Easy access, reading and writing
operations,
Open for all participants
Needs complex security,
Needs large energy and time Supply Chain,
Precision Agriculture,
Farm Overseeing
Private blockchain Fast network response,
Low energy Restricted access for specific
applications Smart Greenhouse,
Water Management,
Food Safety
Consortium blockchain Fast network response,
Low energy Restricted access to several
institutions Supply Chain,
Land Registration
•Private blockchain: Different companies design blockchains with restricted public access for specific applications.
The third party manages the permission to access these blockchains or participate in their consensus proto-
col.14 The private blockchain advantages are faster network response time and lower computational requirements
necessitating low energy. These advantages make private blockchain more desirable for IoT applications and also
for CSA.
•Consortium blockchain: Consortium blockchains are administered by several institutions with mutual interests, all of
which cooperate in the consensus protocol.13 These institutions have authority over the rules that govern how the
users can participate in the consensus protocol. They also indicate the ledge accessibility level for various users. This
type of blockchain has low computational cost, and it needs low energy because it uses multiparty-to-vote consensus
protocols. This makes it suitable to apply for CSA with IoT.
4.2 Blockchain features
The blockchain has many features:
•Decentralization: Blockchain handles and validates transactions through a distributed ledger based on a peer-to-peer
network.
•Persistency: When an identified transaction is added to a block, it cannot be deleted or reversed on the blockchain.
•Anonymity: A virtual identity code is generated for each member to communicate with the blockchain, which covers
the genuine character of the member.
•Auditability: It is indicated that every block is safely connected to the past block. So, the transactions in this design are
easily confirmed and followed.
4.3 Blockchain protocol
The blockchain protocol enables all nodes in the peer-to-peer network to collaborate and work together in the way of
security and verification by including the blocks in the chain that agrees to the set of mining rules. Three rules are defined
by the blockchain protocol: generating data blocks, confirming data blocks, and validating data blocks and resolving block
conflicts in the chain.
Blockchain has consensus methods, which are the fundamental portion of any blockchain network that agrees with
different nodes on adding a new block to the blocks of the blockchain. The consensus strategy enables blockchain net-
works to function in a distributed fashion. Different blockchain protocols, such as Proof of Work (PoW), are used to verify
and aggregate blocks in various ways.15 For example, in the PoW method, different nodes attempt to solve a cryptographic
hash function. The mining method target is to find the following block by solving the hash problem, and the miners are
the users who perform this task.
Proof of Stake (PoS),39 Proof of Authority (PoA),40 and Proof of Burn (PoB)41 are protocols that expend more power
and have a lower speed during preparation for mining.These protocols work on new security technologies such as Intel
AHMED . 13 of 27
SoftwareGuardExtension (SGX) in Proofof luck (Pol)and greenmining technology in Proof of Authority system(PoAs).42
This innovative technology provides a safer and more reliable environment for block mining and saves more energy in
the process of block generation and verification.
4.4 Structure of blockchain-based IoT
The IoT network is data-centric, with a large number of terminal devices loading data. Therefore, data integrity and
privacy are important. Blockchain is considered to be the key to resolving IoT security, reliability issues, and data integrity.
Blockchaintechnology solves the security risks of sensordata and terminalequipment. Accordingto the various functions
of IoT devices, two different structures can be applied in IoT blockchain applications, namely:
•IoT involved blockchain: Devices of IoT can be a part of core blockchain functions by connecting these devices to the
blockchain network.43 These functions include verifying mining blocks and transactions and generating raw sensor
data transactions.
•Blockchain as a service for IoT: Blockchain provides a service layer that coordinates and integrates with traditional IoT
architectures, such as the four-tier architecture.44
Through the blockchain design model, IoT peers collaborate and build trust in IoT device communication.45
•Monitoring Agent (MA) is integrated into every IoT truck through related IoT decentralized Applications (IoTdApp).
The logs generated as a portion of the transaction process can be examined by the monitoring agent.
•A Log Collection Engine (LCE) controls the IoT truck flow log information and realizes it for further transaction
processing.
•The Elastic Nodes Cluster (ENC) processes a huge sum of log data, organizes and indexes it into matching IoT data
records,andtheserecordsaresharedandstoredascopies.
•A Visualization -Platform (VP) uses ENC to organize IoT transaction data and provide efficient information on
blockchain nodes and network statistics.
5CLIMATE-SMART AGRICULTURE
Due to the unpredictable climate change, the agricultural industry requires innovative ways to increase crop yields. Cli-
mate change disrupts the growing season of plants, turns farmlands into deserts, and causes seawater floods, which affect
fertile deltas. Adaptation of climate change and lower emissions can achieve food security goals and agricultural devel-
opment. The world population will increase by one-third by 2050. This needs more contribution between sustainable
agricultural development and climate change to achieve food security and make enough food available to everyone and
anywhere.
The CSA is defined as agriculture that sustainably increases productivity, increases resilience, reduces/eliminates
greenhouse gas emissions, and improves the achievements of national development and food security goals. The CSA
employs a combination of climate change pillars (adaptation, flexibility, and mitigation) as well as new smart technology
insights to reach these objectives.46
The main goal of CSA is to increase agricultural productivity and income in a sustainable manner, adapt and
develop adaptability to climate change, and reduce or eliminate greenhouse gas emissions.47 The CSA has many
aspects that help to optimize sustainable agricultural processes.46 These processes are shown in Figure 6, and they
include:
•Food security: It implies that all individuals can obtain sufficient, safe, and nutritious food from a material, social and
economic point of view to satisfy their needs for an active and healthy life without affecting the environment.
•Production: It is required to improve food security and efficiency of the production system (productivity) by reducing
gas emissions and improving sustainability.
14 of 27 AHMED .
FIGURE 6 Climate-Smart Agriculture aspects
•Capacitybuilding: This activity helps to formulate adaptation measures,increase productivityand improvethe capacity
to adapt to the pressures of climate change.
•Emission reduction: The operation of reducing or eliminating greenhouse gas emissions as much as possible is con-
cerned with emissions per kilogram of food, fiber, and fuel produced, avoiding agricultural logging, and managing soil
and trees in a way that increases their carbon sink potential, thereby improving the atmosphere.48
•New technology: Ultramodern technologies and forms are adopted to boost cultivation efficiency and earnings, whereas
expanding the farm and farmers’ capacity to oversee climate change through GreenHouse Gase (GHG) outflow
lessening.
•Transformation: These processes promote food security by changing agricultural systems under current climate
change.
•Vulnerability reduction: A farming system should reduce the vulnerability of climate change and create improved
productivity profits, while lowering emissions.
5.1 Impacts of climate change on agriculture
Climate change will have a significant impact on agricultural production. In some areas, yields will decline, and yield
volatility will increase. So, major changes may be required in the geographic regions, where crops are grown.49 As already
developing regions are more severely affected, the gap between developing and developed countries is expected to widen
due to climate change. Average temperature increase; changes in rain patterns; variability in rainfall patterns; changes
in the available water; sea-level rise and salinization are all factors that have a profound impact on fisheries, forestry,
and agriculture.50 Climate change has effects on nutrition. All of the nutritional effects of climate change on animals,
vegetables, and wild foods must be better captured.51
5.2 Agriculture impact on climate change
Carbon dioxide (CO2) is not the only direct source of greenhouse gas emissions in the agricultural sector. Nitrous oxide
(N2O)is producedin agriculture and is emitted from soil and fertilization,and methane (CH4) emissions arefrom animals
and rice cultivation. On the other hand, through effective management of the forestry sector, agriculture may lead to
storage and biological capture of biomass and soil carbon. Forest management has the potential to play a significant role
in climate change mitigation,50 especially in the long term.52
AHMED . 15 of 27
5.3 Efficient and resilient systems
When the system uses fewer inputs, land, and water for more food production in a sustainable manner and become more
efficient and resilient, most of the greenhouse gas emissions from the agricultural sector will be reduced. This can be
achievedby improving the resource efficiencyof plant production,improving the resource efficiency of animal husbandry
production, integrating planting and animal husbandry systems, and reducing food loss and waste.
5.4 CSA and sustainable development and green economy
Agriculture and rural development are the engines of sustainable economic development and have effectively promoted
national economic growth. Agricultural development has increased agricultural productivity, reduced food shortages,
increased food surplus, and increased income at the community level. Improvements in agricultural production provide
opportunities to reduce poverty and food insecurity in a sustainable manner, thereby improving livelihoods. In addition,
CSA can help achieve the goal of sustainable development by addressing food security and climate issues and helping to
improve resource efficiency and resilience.53-55
5.5 Climatic smart agricultural applications
Climate change has significant impacts on many agricultural productions. This effect can be in a direct or indirect way.
So, the CSA can be used in many applications that increase the production of agriculture and reduce the impact of climate
change. The graphical representation of CSA and the applications affected by this change are presented in Figure 7.
•Water management
Water plays a vital role in the production of crops and animals (including fish), and changes in the water cycle are
expected to have an impact on agriculture as a result of climate change. Climate change affects every component of the
FIGURE 7 Classification of climatic smart agricultural applications
16 of 27 AHMED .
water cycle, where it increases atmospheric temperatures that affect both rain, river runoff and groundwater. In addition,
irrigated agriculture is predicted to affect the agricultural productivity.
The agricultural sector, especially the irrigation sector, consumes most of the freshwater. Due to the lack of
cost-effective smart water systems, compared to developed countries, developing countries consume more water to
achieve the same performance.55 As a result, the water sources for irrigation are doubled although the amount of land
required to produce crops has decreased since 1961. Irrigated land expansion is expected to continue in the future. So, we
need more control over irrigation by monitoring the climatic change.
•Soil management
Soils are fashioned over lengthy durations of time. Different proportions form them. Due to geology, topography,
climate, plants, and human management variations, soils are tremendously variable. The land provides nutrients and the
water absorbed by the roots of plants and helps regulate atmospheric gases and water. The ability of the soil to provide
basic services to support plant growth determines its health.
With global warming, a decrease in precipitation and an increase in evaporation and transpiration rates are expected
in many places. These changes reduce the soil moisture required for plant growth. In addition, due to high temperatures,
the decomposition (mineralization) rate of Soil Organic Matter (SOM) increases, particularly close to the soil surface. Soil
management can help farmers adapt to increasing weather variability and the effects of climate change. The soil proper-
ties and features are essential to climate change. These properies include soil structure and texture, pH, soil organisms,
nutrients, and organic matter content. There are some operations that can be performed on soil for climate change adap-
tation and mitigation, such as increasing soil water storage, reducing soil erosion, using organic matter to improve soil
structure, and increasing nutrient management. These operations can be managed and controlled using several artificial
technologies such as AI and IoT.
•Climate-smart crop production system
Climate change has a global impact on plant production, which is critical for global food security. In the com-
ing decades, especially in developing countries, billions of people will face food and water shortages and increased
risks to health and life due to climate change. Environmental factors affect the successes and failures of crops due
to weather events facing the earth, such as heavy rainfall, rising coastal waters, geographic changes in the storm, dry
season, and rising temperatures.56 Southern Africa is expected to suffer significant crop losses (with a 30% decrease
in corn production by 2030) as a result of climate change and also South Asia (with a 10% decrease in staple foods
like rice).57
Improved agricultural sector adaptation to the negative effects of climate change is critical to protect and improve the
livelihoods of the poor and ensure food security.58 By improving constitutive components, the agricultural sector overall
efficiency,ability to recover and adapt, and potential to help mitigate the effects of climate change can be improved. Plant
fabric is directly affected by moisture and dry cold. The crop depends on thermal resistance, as temperature can destroy
internal cells and cause irreversible damage or totally destroy the plant.59,60
•Climate-smart livestock
Climate change has a major impact on the natural resources. When the temperature increases, the amount of rain and
changes in precipitation models modify this indirect and modified influence on ecosystems. This leads to changes in the
yields, quality, and type of crops. All these factors contribute to an increased competition for resources and an increase
in animal diseases.
In addition, due to high greenhouse gas emissions, grazing systems are expected to be the hardest hit, because they
are highly dependent on the natural resource base and climatic conditions and have limited adaptability.60 In addition,
non-grazing systems are affected indirectly by higher energy prices, lower crop yields, and feed scarcity.
The livestock industry contributes significantly to climate change and causes large amounts of N2O, CH4,andCO
2
emissions.61 In addition, a large amount of food is wasted before it reaches the consumer. The fundamental strategy for
increasing the environmental sustainability is to improve the efficiency of feed-to-food conversion in animal production
systems. The customization options provide behavioral modifications, managerial options, and technological options as
possible adaptive responses.
AHMED . 17 of 27
•Climate-smart forestry
Sustainable Forest Management (SFM) helps mitigate and adapt to climate change, while also contributing to food
security in various ways. Climate-smart forestry requires a broader application of SFM principles. The services of forest
ecosystems and trees are often overlooked and are largely underestimated, but they are increasingly important for human
adaptation to climate change. Forests and trees provide a variety of forest products as well as ecosystem services. Forest
areas are very effective administrative areas that can help to capture natural carbon. They are the biomass store and
lands that contribute as sinks. Thus, their management using AI and IoT systems can play a vital role in climate change
management.62
•Climate-smart fisheries and aquaculture
Fisheries and aquaculture have provided a source of income and a way of life for 660.82 million people. The growing
population will increase the demand for aquatic food and increase the importance of fishery resources and production
systems.The aquaculturesector mayneed toincrease the output by 70%–100%over the next 20 years from current levels to
meet this demand. Many physical phenomena are linked to the effects of greenhouse gas accumulation in the atmosphere
and water, including rising sea levels, changes in ocean currents, acidification of water bodies, and gradual changes in
water temperature. These physical changes have an impact on the ecological functions of aquatic systems as well as the
frequency, intensity, and location of extreme weather events.63,64 Wider changes in hydrological conditions and seasonal
changes in pH, salinity, temperature, and ecosystem health are expected to affect aquaculture productivity and increase
risks. Some production systems with intelligence technologies may need to be repositioned to accommodate for these
changes.
6IOT FOR SMART CSA
The IoT serves a range of applications in the area of digital agriculture and climatic smart agriculture, including soil and
crop control, animal production, plant growth analysis and distribution, precision agricultural production, assistance in
irrigation evaluations, greenhouse monitoring systems, food supply chain monitoring and so on.
Paper65 presented a system that controls water resources and allocates water for farming under climate change condi-
tions. IoT platform monitors various parameters such as temperature, pH, and moisture, and then collects the data with
sensors and transmits the data over Arduino to computer/machine. A search algorithm can allow suitable irrigation of
crops for a particular soil.
The article66 introduced a new sensor node topology that depends on cheap and efficient components, such as
precipitation sensors, humidity, temperature, soil moisture, and water level. For the low-cost and effective intelligent agri-
cultural system in Vietnam, a LoRa module transfer solution was introduced. It has several advantages that can lower the
costs, increase agricultural productivity, save energy, improve efficiency and realize inter-farm connections to improve
performance and quality. It has the disadvantage that it is difficult to store large amounts of data.
In the paper,67 the authors developed and tested a decision support system for late blight based on cloud IoT. It enables
the use of weather data from specific locations to control disease forecasting and mechanical models of late blight to
provide real-time and seasonal guidance on late blight prevention and control. The sensors used in the environment are
designed to monitor and collect microclimate information, such as temperature, pressure, and humidity. Then, the data
is stored and processed in the cloud. The paper presented an early warning system based on risk factor calculations. If
the disease occurs early, thefarmer is notified by an SMS.
In the paper,68 the water system is monitored using IoT. It examines temperature parameters, soil dampness substance
and temperature. The system also detects the information generated by several sensors to select the water demand system
and determine the crop to make a decision about the irrigation method. On the other hand, it uses smart compounds
consisting of pulses in the farm to keep animals away from the farm and harvest to avoid any harmful effects.
The article,69 shows that the quality and cost of agriculture can be controlled using technologies. The accuracy of the
agricultural sensor monitoring network is primarily used to capture information on agriculture, such as soil nutrition
levels, soil pH, water level, temperature, and climate change. The primary goal is to collect real-time data on agricul-
ture production environments that agricultural professionals can easily access to provide advice on weather patterns
and crops.
18 of 27 AHMED .
In the paper,70 the design aims to use data collected in the greenhouse, allowing multiple sensors to use the data for
processing to improve the greenhouse growth. These data can be used to control the greenhouse climate using mega
Arduino. A data sensor with multiple inputs can simply detect temperature, humidity, and carbon dioxide level. In addi-
tion, it can be used to measure the soil moisture needed by the irrigation system, and lighting intensity required for the
greenhouse.
The paper71 introduced a development of IoT applications for crop protection to prevent animals from entering fields.
A defense and repelling system was provided to prevent potential agricultural damage from wild animal attacks and
weatherconditions. The paper presented an ultrasound repellent system that neither causesphysical orbiological damage
toanimals nor produces noise audible to humans. In addition, the system collects and transmits weatherdata ona periodic
or an on-demand basis. It contributes to creating a database for historical data about the territory micro-climate.
In Reference 72, the presented system uses IoT and WSN to monitor water quality. Collecting information about the
reproduction of aquatic organisms is useful for understanding the condition that affects production, preventing envi-
ronmental disasters, and optimizing resources for pond care. Furthermore, the biological, physical, and chemical pond
water parameters can be monitored and shared between different breeders to predict and control negative aquaculture
conditions.
We can conclude from the previous papers that IoT system is used to monitor and collect important data of climate
changes and data relatedto the studied applications such as crop management, watermanagement, wild animal detection
and animal monitoring. The collected data is used to enhance and control the agriculture system by different devices to
get a suitable decision and achieve high production.
7DATA ANALYSIS FOR SMART CSA SYSTEM
Because of their advantages in traditional systems, AI and ML have been widely adopted in the agricultural sector. For
example, IoT micro weather stations are used to collect real-time data and generate a weather map. In addition, sev-
eral applications depend on the analysis and visualization of gathered data with the help of ML. These applications are
integrated with IoT to make the system more intelligent and get the best decisions.
The paper73 introduced image processing, genetic algorithms, and soft computing techniques to detect plant leaf dis-
eases. The classification and detection of plant leaf diseases depend on the segmentation of the collected images. The
classification was performed using k-means clustering with an accuracy of 86.54%. On the other hand, genetic algorithms
achieved an accuracy of classification of 93.63%, and the SVM classifier achieved an accuracy of 95.71%.
The presented model in the paper74 is for an intelligent irrigation system that uses ML algorithms to predict crop
water needs. Humidity and temperature are the most important parameters in determining the amount of water required
in any agricultural area. Hence, the system consists of temperature, and humidity sensors in farmland. They send data
by a microprocessor to the cloud. In order to predict the results efficiently, the DT algorithm which is an efficient ML
algorithm, is applied on the data sensed from the field.
In Reference 75, a new technology is used to aid in the development of reliable fruit detection systems for automatic
collection. A new DL model based on single-stage architectures was proposed to optimize memory and detection speed
for the task of mango fruit detection. The performance of this new model was compared with those of six existing DL
models.
The paper76 offered an IoT-based platform for monitoring of animal behavior. It includes a local IoT network and
a cloud platform that extracts information from sensor data and uses ML characteristics for data analysis. In addition,
various ML algorithms were also evaluated to assess the platform power.
In the study,77 Soil Moisture Content (SMC) is predicted on an irrigated maize field using a Microscopic Cellular
Automated (MCA) model combined with a Deep Belief Network (DBN). The paper showed that the DBN-MCA model
is superior to the Multi-Layer Perceptron MLP-MCA model, and is a more powerful tool for predicting SMC in highly
nonlinear forms.
This study78 has developed methods to estimate crop yields by using DL to extract traits that significantly affect plant
growth. It also makes use of a Neural Network (NN), Normalized Vegetation Index (NDVI), Absorbed Photosynthetically
Active Radiation (APAR), and input data such as canopy surface temperature, water stress index, and environmental
factors.
In Reference 79, the authors discussed a novel model for recognizing plant diseases. This model depends on deep
CNN and leaf image classification. The developed model can distinguish 13 different diseases.
AHMED . 19 of 27
The paper80 focused on applying DL to aquaculture, including identifying live fish, classifying species, analyzing
behavior, making feeding decisions, estimating size or biomass, and predicting water quality. It also analyzes the technical
details of the DL method applied to smart fish farming, including data, algorithms, and performance.
We can conclude from previous papers that ML and AI systems are used to analyze and process climate change data
and important data related to the studied application. This data is collected from different resources such as WSN, IoT,
and remote sensors to get suitable decisions on crop management, water management, wild animal detection and animal
monitoring.
8BLOCKCHAIN FOR CSA SYSTEMS
IoT applications produce a tremendous volume of information. Before completing the data analysis, the data need to be
verified to avoid harmful or redundant data. Blockchain and IoT can work together to apply the verification of data and
provide more security for transaction and data transmission.81
The authors of the paper82 checked security and threats in IoT applications. Another method that depends on edge
computing, fog computing, ML, and blockchain was presented. In Reference 83, IoT transactions between devices are
transmitted in the network. The blockchain protocol manages, confirms, fixes IoT transactions and helps to specify the
complete picture to manage traffic, events, and transactions between IoT devices. To monitor a variety of IoT transaction
models, the form of blockchain must be modular.
The paper84 presented a traceability system for food safety that depends on blockchain an manages data architecture
on-chain and off-chain. This prevents data explosion problems that may happen in the IoT system and blockchain.
The paper85 aims to better understand the possibilities, benefits, and applications of distributed ledger technology in
the food industry. It also identifiestechnical limits and possible institutional barriers of implementation. Blockchain can
be used to face large challenges in precision agriculture, particularly when combined with IoT systems.
The paper45 discussed building smart system applications to monitor and digitize farms using IoT sensors such as
light, humidity, harvest time, and temperature sensors. In addition, the farmer and stakeholders can use blockchain to
monitor plant storage technology being used to avoid post-harvest losses.
We can conclude from the previous papers that blockchain can be integrated with IoT systems and then used to
manageand secureimportant dataof climatechanges andthe data related tothe studiedapplications. Thedata iscollected
from different resources and the transfer of data is performed to manage and control the agriculture systems that need to
be secure and trusted by the blockchain system (Table 5).
9SECURITY IN IOT-ENABLED SMART AGRICULTURE
IoT and smart devices are used to enhance and increase production in smart agriculture and precision agriculture, but
this needs more security for several smart devices that are used in the IoT systems to protect them from potential attacks.
In this part, we discuss how to improve IoT security in smart agriculture.
The paper86 investigated the security scenarios and analysis of possible attacks and threats and provided a detailed
direction of the progress in farming security sub-areas, and also the number of available testbeds for agriculture based on
IoT. It also introduced an architecture for smart farming and the requirements of security to generalize the application
scenario and layer-wise security framework.
In Reference 87, the authors provided a different literature review that contains the security goals for using blockchain
technology in smart agriculture. The study also included a comparative analysis of schemes with advantages, drawbacks,
application areas and detailed cost analysis. In addition, the study has led to a new avenue for future research that can
depened on AI.
The paper88 introduced AKMS-AgriIoT and a blockchain-based implementation that supports better security,
low cost and more functionality features. The blockchain-based authentication is based on gathering the data in a
secure way from drones. A Ground Station Server (GSS) sends the list of encrypted transactions with their signa-
tures to the cloud server in the Blockchain Center (BC) after transactions are formed and data is collected in a
secure way.
The article89 introduced a secure smart farming environment through designing a new authenticated key agree-
ment mechanism based on blockchain. This environment is produced in two ways, first between IoT devices and
20 of 27 AHMED .
TABLE 5 State-of-the-art research works related to the paper subjects
Research Technology Field Contribution Positive aspects
(Advantages) Negative aspects
(Disadvantages)
[65] IoT Water control Introducing a proposed system
that controls water resources
under all conditions of climatic
change.
•Preventing water waste
•Reducing cost
•Achieving full automation
•Need to apply in large area
•Need for more data analysis
•Environment that is not
friendly to farmer
[66]IoT Smart agriculture Presenting a solution using LoRa
module transmission for an
inexpensive and effective
intelligence agriculture system.
•Improving yield and pro-
ductivity
•Achieving low cost and low
energy
•Need for large storage
database
[67] Cloud-IoT Crop disease
detection Allowing disease forecasters to
use location-specific data,
consideration of weather data
and development of a
mechanistic model for late
blight disease detection.
•Wide accessibility and low
cost
•Friendly environment
•Utilization of SMS to alarm
farmers
•Not fully-automated system
•Need for more datasets to
detect diseases
•No comparison with other
algorithms
[68]IoT Water control and
soil
management
IoT assignment of the water
system, and verification of
parameters, ground
temperature and soil moisture
to make decisions on how to
cultivate lands and perform
irrigation.
•Improving crop production
•Avoidance of animal dam-
age by alarms
•Full automation of the sys-
tem
•Low accuracy
•Neglectance of cost analysis
[69] IoT Weather pattern
and crop
protection
Collecting real-time data from the
agricultural production
environment, and providing
convenient access and weather
pattern recommendations for
agriculture.
•Energy saving
•Low cost of wireless net-
work
•Sending alerts to farmers
•No full automation
•Need for a large area
[70]IoT Greenhouse Creating a data acquisition
system in the greenhouse with
multiple sensors.
•Full automation
•Sutability for small farms
•Low cost and power
•Need for a friendly frame-
work for farmers
[71] IoT Crop protection
and animal
monitoring
Monitoring of repelled wild
animals to prevent crop
damages, and also weather
monitoring.
•Low power
•Low cost
•Sending notifications to
farmers
•Limitation to small farms
•Neglectance of work latency
[72]IoT Water quality in
aquaculture IoT-based monitoring of water
quality as well as physical,
chemical, and biological
parameters of pool water.
•Water quality monitoring
•Low production cost
•Small system latency
•Friendly framework
•Neglectance of ML in analy-
sis and prediction
•Neglectance of dangerous
condition
[73] Genetic
algorithms and
image
processing
Plant leaf disease
detection Classification and detection of
plant leaf diseases using an
image segmentation algorithm.
•Full automation
•Utilization of a friendly
environment
•Need for a large number of
training samples
•Need for a large database to
enhance accuracy
•Ability to detect few diseases
[74]Decision tree
algorithm Water control Utilization of a model for an
intelligent irrigation system
that uses ML algorithms to
predict the water needs of
plants.
•Saving time and cost
•Allowing high accuracy
•Increasing productivity
•Applicability on small areas
•Neglectance of comparison
with other ML algorithms
[75] Deep learning Reliable fruit
detection Utilization of a deep learning
algorithm to improve the
detection and estimation of
fruit yield and allow automated
harvesting.
•Good detection of objects
•Enhancing speed and accu-
racy
•High complexity
•Need for a friendly frame-
work
(Continues)
AHMED . 21 of 27
TABLE 5 (Continued)
Research Technology Field Contribution Positive aspects
(Advantages) Negative aspects
(Disadvantages)
[76] IoT and ML Animal
monitoring Utilization of a platform for
monitoring of animal behavior
depending on IoT and ML
algorithms to assess the
platform power.
•Utilization of a large
database
•Utilization of a friendly
framework
•Utilization of a manage-
ment and alarm system
•No full automation
•Need for large energy
•Utilization of a small dataset
[77]NN, multi-layer
perceptron and
deep belief
network
Soil management Predicting the SMC for the
cornfield, and employment of a
combination of a DBN model
with macroscopic cellular
automata.
•Low sensitivity to missing
data
•Suitability for learning
nonlinear and complex
behaviors
•Need for more samples for
prediction
•Negelctance of comparison
with other models
[78] NN and NVI Crop protection Utilization of a crop yield
estimation methodology based
on DL to extract features that
have a significant impact on
crop growth.
•Ability to estimate crop
yield in a large area •Needformoredatato
enhance the model accuracy
•Limitted applicability of the
model
[79]Deep CNN Plant disease
detection Utilization of a deep CNN model
for plant disease detection
depending on leaf image
classification.
•Ability of the model to
detect 13 diseases
•Introduction of a new plant
disease image database
•Need for a friendly frame-
work
•Need for a large database
•Need for large area for proper
application
•Need for high hardware
requirements
[80] Deep learning Water quality and
fish monitoring Ability of DL application in
aquaculture including
identifying live fish, classifying
species, analyzing behavior,
making decisions about
feeding, estimating size or
biomass, and predicting water
quality.
•Ability of the model to
detect unusual behaviors
•Ability to predict water
changes
•Monitoring of fish in real
time
•Need for a large dataset to
improve accuracy
•Need for high hardware
requirements
•High cost
[82]Edge computing,
fog computing,
blockchain, IoT,
and ML
Security in
application Introduction of a novel approach
based on edge computing, fog
computing, blockchain, and
ML to solve security and threat
problems in IoT applications.
•Consideration of security
for all IoT layers
•Introduction of a solution
for security problems using
different technologies
•Need for large storage
•High cost
•Need for huge data and large
database
[84] IoT, and
Blockchain Traceability
system Proposal and development of a
traceability system for food
safety depending on blockchain
and EPC information services.
•High degree of decentral-
ization
•High storage
•High security and reliabil-
ity
•Limited performance
•Low speed of data uploading
[85]IoT, and
blockchain Monitoring of
crops Creation of intelligent farms
based on IoT and blockchain
sensors to monitor crop storage
to prevent post-harvest losses.
•Decentralization of trans-
actions
•Good immutability and
traceability
•Need to devolve for data pro-
tection
•Low network availability in
some areas
•Complexity of ledgers
[45] IoT, and
blockchain Precision
agriculture Proposal of blockchain models to
solve the major problems in
IoT-based precision
agricultural systems.
•Utilization of blockchain to
solve IoT security problems
•Consideration of different
applications of precision
agriculture
•Need for more storage capac-
ity for scalability
•Need for more time and
speed for transaction
•Need for high hardware
requirements
22 of 27 AHMED .
the -GateWay Node (GWN), and second between Device-to-Gateway (D2G) and Device-to-Device (D2D). The pre-
sented Smart Contract-Based Blockchain envisioned Authentication Scheme (SCBAS) provides functionality features and
superior security compared to the existing authentication protocols.
InReferences90,91, anovel remoteuser authentication scheme using WSNs foragriculture monitoring waspresented.
Burrows-Abadi-Needham (BAN) model is used and a random oracle model is also used in the scheme to analyze security.
The analysis shows the resistance to various kinds of malicious attacks, when a remote user authentication protocol
with six phases is applied. In addition, the informal security analysis shows that the scheme is efficient for real-time
applications.
Paper92 presented a scheme based on the recognized automated validation of Internet security protocols and also
supporting several features including user credentials change phase, user mobile device revocation phase, and dynamic
IoT smart device phase. These phases help to protect against different attacks with a secure and accepted random oracle
model.The AVISPA software tool is used for verification, and it shows that the presented scheme is immune against replay
and man-in-the-middle attacks because of the utilization of controller nodes and IoT smart devices. It also preserves the
anonymity of users (Table 5).
10 COMBINATION OF BLOCKCHAIN TECHNOLOGY WITH IOT DEVICES
AND ML
Combination of blockchain technology with IoT devices and ML can be adopted in different applications, and some
studies are focused on this combination. In the paper,93 the combination of the three different technologies, IoT,
ML, and blockchain, is discussed to allow new business models and apply on autonomous agents. This combina-
tion helps in the digital transformation of industrial corporations and data management. In References 16,94,the
environmental monitoring of health within the health system current technologies such as ML, IoT, and blockchain
is considered. The IoT helps for monitoring of patients outside the hospital, and the large collected databases are ana-
lyzed using ML algorithms. Blockchain helps to solve interoperability and data sharing issues and enhance decision
making.
In our suggestions in this paper, IoT operates across heterogeneous domains, and it is equipped with sensing, actuat-
ing, and different network devices that help for data acquisition to monitor, manage, and control applications that need to
be improved in CSA. IoT is combined with AI and ML to provide better Quality of Service (QoS), analysis, and calculation
of data transmitted over IoT networks.
The ML allows classification, clustering, and prediction to control the CSA applications by using IoT actuators. This
helps to get the best decisions and reduce cost and time. There is a high importance for preserving the integrity of data and
security.This protection is added by using blockchain technology adopted in the IoT domain. Blockchain is also used for
manipulation, load distribution, decentralization, especially with huge amounts of collected data. This can help in CSA
applicationsin largeareas. Inaddition, thisadds additional features tothe IoT network andprovides CSAwith the security
and privacy needed for all network devices and operations. Figure 8shows a combination of blockchain technology with
IoT devices and ML with CSA. We see that the three technologies complement each other, and new advantages are added
to enhance CSA applications.
11 CHALLENGES AND OPPORTUNITIES
Any network depends on elements and technologies that need to be standard to communicate between different devices
and help devices sharing data. Unfortunately, IoT networks have a lack of technical standards, and the changes in
hardware and software have led to inconsistent technology ecosystems. Various technologies may also cause system
maintenance and scalability issues.95,96
Middleware is another component of IoT that poses unique security challenges, because it communicates and pro-
cesses the data between network and application layers. Hence, confidentiality and secure data storage are required for
middleware layer security.97
The high cost of equipment used in the development of smart agriculture products and monitoring of climatic
change are major concerns. In addition, blockchain needs high costs for computing, maintenance overheads, and storage
requirements. By using AI with IoT and blockchain, the overall cost of the infrastructure is reduced.
AHMED . 23 of 27
FIGURE 8 Combination of blockchain technology with IoT devices and ML
Power consumption is high for sensors and devices used for monitoring, controlling, and storage in the CSA systems,
especially when we combine the three technologies. Hence, we need smart grids and energy management to reduce
energy consumption. In addition, the use of renewable energy sources can help to solve the problem.
Wireless communication is most commonly used in agricultural applications. As we all know, several areas do not
have fast and reliable connectivity. The environment is also one of the main contributors to the poor quality of the wireless
connection due to the multipath effect and its contribution to background noise. The challenges of the agricultural envi-
ronment are how to produce reliable wireless communication and improve the performanceof the limited bandwidth in
a complex and variable environment.
The large data generated by IoT devices and the transaction of blockchain is a challenge because this needs more data
management, speed, real-time communication in transmission, security, and storage. The processing and computing that
are performed at the network edge can reduce transmission and reduce attacks.
Various complex technologies can be a problem for farmers and administrators. Hence, there is a need for more expe-
rience and specific skills to deal with these technologies. That is why we need to develop user-friendly systems to solve
this problem.
12 CONCLUSIONS
In this article, we have presented a survey of IoT in climatic smart agriculture. We talked about IoT components, agri-
cultural network architecture, and communication protocols that help farmers gain access to IoT and improve crop
productivity. The ML and blockchain are briefly discussed in the context of AI. Furthermore, this article outlined the
current and ongoing advances in CSA with IoT and many innovative technologies such as AI, ML, and blockchain. This
study considered several IoT agricultural challenges and opportunities to better understand IoT when applied in smart
farming. It is also clear that many large organizations have begun investing in and developing new farm management sys-
tems. It is expected that IoT in smart agriculture and automation will soon replace traditional farming methods. Finally,
this comprehensive survey is expected to yield useful information for analysts, students, experts, and farmers working in
IoT and agricultural technologies.
12.1 Future works
For future research, the following items can be considered:
24 of 27 AHMED .
•Researchers can apply blockchain technology to improve some applications in CSA, such as enhancing security of
collected data in smart farming, and optimizing irrigation and precision agriculture system based on IoT technology.
•The blockchain technology can be used for managing data and producing load balance, especially in case of big
collected data and large systems.
•Experts and farmers can apply IoT in monitoring of climate change to overcome the effects of change on plants and
crops by analyzing collected data using several ML algorithms to get suitable decisions.
•The process of water management can also be monitored and controlled to get the suitable needs for crops in order to
increase productivity with optimum water needs.
•The agriculture land images collected from remote sensing can be processed and analyzed using ML algorithms to
serve for the optimization of sustainable agricultural processes.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.
ORCID
Walid El-Shafai https://orcid.org/0000-0001-7509-2120
REFERENCES
1. Al-Fuqaha A, Guizani M, Mohammadi M, Aledhari M, Ayyash M. Internet of things: a survey on enabling technologies, protocols, and
applications. IEEE Commun Surv Tutorials. 2015;17(4):2347-2376. doi:10.1109/COMST.2015.2444095
2. MahdavinejadMS,RezvanM, Barekatain M, Adibi P, Barnaghi P,Sheth AP. Machine learning for internet of things data analysis: a survey.
arXiv. 2018;4:161-175. doi:10.1016/j.dcan.2017.10.002
3. Sobin CC. A survey on architecture, protocols and challenges in IoT. Wirel Pers Commun. 2020;112(3):1383-1429. doi:10.1007/s11277-020-
07108-5
4. Farooq MS, Riaz S, Abid A, Abid K, Naeem MA. A survey on the role of IoT in agriculture for the implementation of smart farming. IEEE
Access. 2019;7(October):156237-156271. doi:10.1109/ACCESS.2019.2949703
5. Ruiz-Garcia L, Lunadei L, Barreiro P, Robla JI. A review of wireless sensor technologies and applications in agriculture and food industry:
state of the art and current trends. Sensors. 2009;9(6):4728-4750. doi:10.3390/s90604728
6. MohantaBK, Jena D, Satapathy U, PatnaikS.Survey on IoT security: challenges and solution using machine learning, artificial intelligence
and blockchain technology. Internet Things. 2020;11:100227. doi:10.1016/j.iot.2020.100227
7. Gissella MVE. A quick view on current techniques and machine learning algorithms for big data analytics. 18th International Conference
on Transparent Optical Networks (ICTON). IEEE; 2016:1-4. doi:10.1109/ICTON.2016.7550517
8. Mekonnen Y, Namuduri S, Burton L, Sarwat A, Bhansali S. Review—machine learning techniques in wireless sensor network based
precision agriculture. J Electrochem Soc. 2020;167(3):037522. doi:10.1149/2.0222003jes
9. Rana AK, Salau A, Gupta S, Arora S. A survey of machine learning methods for IoT and their future applications. Amity J Comput Sci.
2019;2(2):1-5. https://hal.archives-ouvertes.fr/hal-01983429
10. Salimitari M, Chatterjee M, Fallah YP. A survey on consensus methods in blockchain for resource-constrained IoT networks. Internet
Things. 2020;11:100212. doi:10.1016/j.iot.2020.100212
11. Wu M, Wang K, Cai X, Guo S, Guo M, Rong C. A comprehensive survey of blockchain: from theory to IoT applications and beyond. IEEE
Internet Things J. 2019;6(5):8114-8154. doi:10.1109/JIOT.2019.2922538
12. Xie J, Tang H, Huang T, et al. A survey of blockchain technology applied to smart cities: research issues and challenges. IEEE Commun
Surv Tutorials. 2019;21(3):2794-2830. doi:10.1109/COMST.2019.2899617
13. Lin IC, Liao TC. A survey of blockchain security issues and challenges. Int J Netw Secur. 2017;19(5):653-659. doi:10.6633/IJNS.201709.
19(5).01
14. Wang X, Zha X, Ni W, et al. Survey on blockchain for internet of things. Comput Commun. 2019;136(September):10-29. doi:10.1016/j.
comcom.2019.01.006
15. Wang H, Zheng Z, Xie S, Dai HN, Chen X. Blockchain challenges and opportunities: a survey. IntJWebGridServ. 2018;14(4):352. doi:10.
1504/ijwgs.2018.10016848
16. Bublitz FM, Oetomo A, Sahu KS, et al. Disruptive technologies for environment and health research: an overview of artificial intelligence,
blockchain, and internet of things. Int J Environ Res Public Health. 2019;16(20):1-24. doi:10.3390/ijerph16203847
17. Alanm Mansaf KS, Kashish AS. Internet of Things (IoT) Concepts and Applications. Switzerland: Springer; 2020. doi:10.1007/978-3-030-
42520-3_31
18. Ahmed N, De D, Hussain I. Internet of things (IoT) for smart precision agriculture and farming in rural areas. IEEE Internet Things J.
2018;5(6):4890-4899. doi:10.1109/JIOT.2018.2879579
19. Kodali RK, Sarjerao BS. A low cost smart irrigation system using MQTT protocol. In: TENSYMP 2017 – IEEE Int. Symp. Technol. Smart
Cities, October, 2017:1-5. doi:10.1109/TENCONSpring.2017.8070095
AHMED . 25 of 27
20. Al-Sarawi S, Anbar M, Alieyan K, Alzubaidi M. Internet of things (IoT) communication protocols: review. In: ICIT 2017 – 8th Int. Conf.
Inf. Technol. Proc. October, 2017:685-690. doi:10.1109/ICITECH.2017.8079928
21. Pitì A, Verticale G, Rottondi C, Capone A, Lo Schiavo L. The role of smart meters in enabling real-time energy services for households:
the Italian case. Energies. 2017;10(2):1-25. doi:10.3390/en10020199
22. Papageorgiou A, Zahn M, Kovacs E. Efficient auto-configuration of energy-related parameters in cloud-based IoT platforms. In: 2014 IEEE
3rd Int. Conf. Cloud Networking, CloudNet 2014; 2014:236-241. doi: 10.1109/CloudNet.2014.6968998
23. ShiY,DingG,WangH,EduardoRomanH,LuS.ThefogcomputingserviceforHealthcare.River Publ Ser Commun. 2015;60(51):70-74.
doi:10.1109/Ubi‐HealthTech.2015.7203325
24. López M, Peinado A, Ortiz A. An extensive validation of a SIR epidemic model to study the propagation of jamming attacks against IoT
wireless networks. Comput Networks. 2019;165:106945. doi:10.1016/j.comnet.2019.106945
25. Adi E, Anwar A, Baig Z, Zeadally S. Machine learning and data analytics for the IoT. Neural Comput Appl. 2020;32(20):16205-16233.
doi:10.1007/s00521-020-04874-y
26. Vu DL, Nguyen TK, Nguyen TV, Nguyen TN, Massacci F, Phung PH. HIT4Mal: hybrid image transformation for malware classification.
Trans Emerg Telecommun Technol. 2019;31(11):1-15. doi:10.1002/ett.3789
27. Wang L, Ranjan R. Processing distributed internet of things data. Researchgate.Net. 2015;2(1):76-80. doi:10.1109/MCC.2015.14
28. Baig ZA, Sanguanpong S, Firdous SN, Vo VN, Nguyen TG, So-In C. Averaged dependence estimators for DoS attack detection in IoT
networks. Futur Gener Comput Syst. 2020;102:198-209. doi:10.1016/j.future.2019.08.007
29. Almiani M, AbuGhazleh A, Al-Rahayfeh A, Atiewi S, Razaque A. Deep recurrent neural network for IoT intrusion detection system. Simul
Model Pract Theory. 2020;101:102031. doi:10.1016/j.simpat.2019.102031
30. LiuL,Ma Z, Meng W. Detectionofmultiple-mix-attack malicious nodes using perceptron-based trust inIoT networks. Futur Gener Comput
Syst. 2019;101:865-879. doi:10.1016/j.future.2019.07.021
31. Moon J, Jung IY, Park JH. IoT application protection against power analysis attack. Comput Electr Eng. 2018;67:566-578. doi:10.1016/j.
compeleceng.2018.02.030
32. Deshmukh-Bhosale S, Sonavane SS. A real-time intrusion detection system for wormhole attack in the RPL based internet of things. Proc
Manuf. 2019;32:840-847. doi:10.1016/j.promfg.2019.02.292
33. Yin D, Zhang L, Yang K. A DDoS attack detection and mitigation with software-defined internet of things framework. IEEE Access.
2018;6:24694-24705. doi:10.1109/ACCESS.2018.2831284
34. Li C, Qin Z, Novak E, Li Q. Securing SDN infrastructure of IoT-fog networks from MitM attacks. IEEE Internet Things J.
2017;4(5):1156-1164. doi:10.1109/JIOT.2017.2685596
35. Kakkad V, Patel M, Shah M. Biometric authentication and image encryption for image security in cloud framework. Multiscale Multidiscip
Model Exp Des. 2019;2(4):233-248. doi:10.1007/s41939-019-00049-y
36. Shah D, Dixit R, Shah A, Shah P, Shah M. A comprehensive analysis regarding several breakthroughs based on computer intelligence
targeting various syndromes. Augment Hum Res. 2020;5(1):1-12. doi:10.1007/s41133-020-00033-z
37. Barber D. Bayesian Reasoning and Machine Learning. Cambridge: Cambridge University Press; 2012. doi:10.1017/cbo9780511804779
38. Braga-Neto U. Fundamentals of Pattern Recognition and Machine Learning. Switzerland: Springer; 2020. doi:10.1007/978-3-030-27656-0
39. Kiayias A. PPCoin: peer-to-peer crypto-currency with proof-of-stake. In: Proc. 2016 ACM SIGSAC Conf. Comput. Commun.
Secur. – CCS’16, Vol. 1919, January, 2017:1-27. http://connections-qj.org/article/ppcoin-peer-peer-crypto-currency-proof-stake
40. Bentov I, Lee C, Mizrahi A, Rosenfeld M. Proof of activity: extending Bitcoin’s proof of work via proof of stake. ACM SIGMETRICS Perform
Eval Rev. 2014;42(3):34-37. doi:10.1145/2695533.2695545
41. Karantias K, Kiayias A, Zindros D. Financial Cryptography and Data Security 2013. Berlin: Springer; 2020. doi:10.1007/978-3-030-51280-4
42. Milutinovic M, He W, Wu H, Kanwal M. Proof of luck: an efficient blockchain consensus protocol. arXiv. 2016;1-6. doi:10.1145/3007788.
3007790
43. Maymounkov P, Mazières D. Kademlia: a peer-to-peer information system based on the XOR metric. Lect Notes Comput Sci.
2002;2429:53-65. doi:10.1007/3-540-45748-8_5
44. Dorri A, Kanhere SS, Jurdak R. Towards an optimized blockchain for IoT. In: Proc. – 2017 IEEE/ACM 2nd Int. Conf. Internet-of-Things
Des. Implementation, IoTDI 2017 (Part CPS Week), April, 2017:173-178. doi:10.1145/3054977.3055003
45. Torky M, Hassanein AE. Integrating blockchain and the internet of things in precision agriculture: analysis, opportunities, and challenges.
Comput Electron Agric. 2020;178(May):105476. doi:10.1016/j.compag.2020.105476
46. Adesipo A, Fadeyi O, Kuca K, et al. Smart and climate-smart agricultural trends as core aspects of smart village functions. Sensors.
2020;20(21):1-22. doi:10.3390/s20215977
47. Garnett T. Where are the best opportunities for reducing greenhouse gas emissions in the food system (including the food chain)? Food
Policy. 2011;36(SUPPL. 1):S23-S32. doi:10.1016/j.foodpol.2010.10.010
48. Thornton PK, Whitbread A, Baedeker T, et al. A framework for priority-setting in climate smart agriculture research. Agr Syst.
2018;167(October):161-175. doi:10.1016/j.agsy.2018.09.009
49. Metz B, Davidson O. Climate Change 2007 Mitigation. Cambridge: Cambridge University Press; 2007.
50. Beddington JR, Asaduzzaman M, Clark ME, et al. The role for scientists in tackling food insecurity and climate change. Agric Food Secur.
2012;1(1):1-9. doi:10.1186/2048-7010-1-10
51. Lidon FJC. Climate change and food security. Emirates J Food Agric. 2018;30(6):428. doi:10.4324/9781315272085-3
52. Nunes LJR, Meireles CIR, Gomes CJP, Ribeiro NMCA. Forest management and climate change mitigation: a review on carbon cycle flow
models for the sustainability of resources. Sustainability. 2019;11(19):1-10. doi:10.3390/su11195276
26 of 27 AHMED .
53. Bruinsma J. The Resource Outlook to 2050. Rome: Food Agric. Organ. United Nations; 2009:1075. doi:10.1007/978-3-642-41714-6_180309
54. Me
˛trak M, Pokarowski P, Sulwi
nski M, Gantumur A, Suska-Malawska M. Plant response to N availability in permafrost-affected alpine
wetlands in arid and semi-arid climate zones. Sci Total Environ. 2020;721:137791. doi:10.1016/j.scitotenv.2020.137791
55. Goap A, Sharma D, Shukla AK, Rama Krishna C. An IoT based smart irrigation management system using machine learning and open
source technologies. Comput Electron Agric. 2018;155(May):41-49. doi:10.1016/j.compag.2018.09.040
56. Dokken D, Field CB, Barros V, et al. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation; 2012.
https://www.ipcc.ch/report/managing-the-risks-of-extreme-events-and-disasters-to-advance-climate-change-adaptation/
57. Lobell DB, Burke MB, Tebaldi C, Mastrandrea MD, Falcon WP, Naylor RL. Prioritizing climate change adaptation needs for food security
in 2030. Science (80-). 2008;319(5863):607-610. doi:10.1126/science.1152339
58. FAO. The State of Food and Agriculture 2012. Rome: Food and Agriculture Organization of the United Nations (FAO); 2012.
59. Neuner G, Huber B, Plangger A, Pohlin JM, Walde J. Low temperatures at higher elevations require plants to exhibit increased freezing
resistance throughout the summer months. Environ Exp Bot. 2020;169:103882. doi:10.1016/j.envexpbot.2019.103882
60. Aydinalp C, Cresser MS, Cresser MS. The effects of global climate change on agriculture. Agric Environ Sci. 2008;3(5):672-676. https://
www.researchgate.net/publication/238091112
61. FAO. Climate Smart Agriculture. Rome: FAO; 2013.
62. Amin A, Mubeen M, Mohkum Hammad H, Nasim Jatoi W. Climate smart agriculture: an approach for sustainable food security. Agric
Res Commun. 2015;2(3):13-21. www.aspublisher.com
63. FAO. The State of World Fisheries and Aquaculture 2012. Rome: Food and Agriculture Organization of the United Nations; 2012.
64. McclanahanT, Allison EH, Cinner JE. Managing fisheries for human and food security. Fish Fish. 2015;16(1):78-103. doi:10.1111/faf.12045
65. Pathak A, Amazuddin M, Abedin J, Andersson K, Mustafa R, Hossain MS. IOT based smart system to support agricultural parameters: a
IoT based smart system to support agricultural parameters: a case study case. Procedia Comput Sci. 2019;155:648-653. doi:10.1016/j.procs.
2019.08.092
66. Khoa TA, Man MM, Nguyen T, Nguyen V. Smart agriculture using IoT multi-sensors: a novel watering management system. Sens Actuat
Networks. 2019;8(45):1-22. doi:10.3390/jsan8030045
67. Foughali K. Cloud-IOT based decision support system for potato pest prevention. Proc Comput Sci. 2019;160:616-623. doi:10.1016/j.procs.
2019.11.038
68. Nagpure S, Ingale S, Pahurkar S, Bobade AM, Ghosal M. Smart agriculture using IOT. Helix. 2019;9(3):5081-5083. doi:10.29042/2019-
5081-5083
69. Singh A, Kumar S, Nagar H. A smart agricultural model by integrating IoT. Vol 6. IEEE; 2018:1-5. www.ijraset.com
70. Shirsath PDO, Kamble P, Mane R, Kolap A, More PRS. IOT based smart greenhouse automation. Int J Innov Res Comput Sci Technol.
2017;5(2):1-6. doi:10.21276/ijircst.2017.5.2.4
71. Giordano S, Seitanidis I, Ojo M, Adami D, Vignoli F. IoT solutions for crop protection against wild animal attacks. In: 2018 IEEE Int. Conf.
Environ. Eng.; 2018:1-5. doi:10.1109/EE1.2018.8385275
72. Encinas C, Ruiz E, Cortez J, Espinoza A. Design and implementation of a distributed IoT system for the monitoring of water quality in
aquaculture. In: 2017 Wireless Telecommunications Symposium (WTS); 2017:1-7. doi:10.1109/WTS.2017.7943540
73. Singh V, Misra AK. Detection of plant leaf diseases using image segmentation and soft computing techniques. Inf Process Agric.
2017;4(1):41-49. doi:10.1016/j.inpa.2016.10.005
74. Blasi AH, Abbadi MA, Al-Huweimel R. Machine learning approach for an automatic irrigation system in southern Jordan Valley. Eng
Technol Appl Sci Res. 2021;11(1):6609-6613. doi:10.48084/etasr.3944
75. Walsh AKKB, Mccarthy ZWC. Deep learning for real-time fruit detection and orchard fruit load estimation: benchmarking of ‘MangoY-
OLO’. Precis Agric. 2019;20(6):1107-1135. doi:10.1007/s11119-019-09642-0
76. NóbregaL,Gonçalves P. Animal monitoringbased on IoT technologies. In: 2018 IoT Vertical and TopicalSummit on Agriculture – Tuscany
(IOT Tuscany); 2018:9-13. doi:10.1109/IOT-TUSCANY.2018.8373045
77. Xiaodong S, Ganlin Z, Feng LIU, Decheng LI, Yuguo Z, Jinling Y. Modeling spatio-temporal distribution of soil moisture by deep
learning-based cellular automata model. JAridLand. 2016;8:734-748. doi:10.1007/s40333-016-0049-0
78. Kuwata K, Shibasaki R. Estimating crop yields with deep learning and remotely sensed data. In: 2015 IEEE International Geoscience and
Remote Sensing Symposium (IGARSS); 2015:858-861. doi:10.1109/IGARSS.2015.7325900
79. Sladojevic S, Arsenovic M, Anderla A, Culibrk D, Stefanovic D. Deep neural networks based recognition of plant diseases by leaf image
classification. Comput Intell Neurosci. 2016;2016:1-11. doi:10.1155/2016/3289801
80. Yang X, Zhang S, Liu J, Gao Q, Dong S, Zhou C. Deep learning for smart fish farming: applications, opportunities and challenges. Rev
Aquac. 2021;13(1):66-90. doi:10.1111/raq.12464
81. Moonen H, Van Sinderen M. Towards Self-sovereign Identity using Blockchain Technology. University of Twente; 2016. http://purl.utwente.
nl/essays/71274
82. Hassija V, Chamola V, Saxena V, Jain D. A survey on IoT security: application areas, security threats, and solution architectures. IEEE
Access. 2019;7:82721-82743. doi:10.1109/ACCESS.2019.2924045
83. Blockchain U. Management and monitoring of IoT devices using blockchain. Sensors. 2019;19(856):1-12. doi:10.3390/s19040856
84. LinQ, WangH, Pei X, Wang J. Food safety traceability system based on blockchainandEPCIS.IEEEAccess.2019;7:1. doi:10.1109/ACCESS.
2019.2897792
85. Schmidhuber J. Emerging Opportunities for the Application of Blockchain in the Agri-Food Industry. Switzerland: FAO; 2018.
AHMED . 27 of 27
86. Vangala A, Das AK, Kumar N, Alazab M. Smart secure sensing for IoT-based agriculture: blockchain perspective. IEEE Sens J.
2021;21(16):17591-17607. doi:10.1109/JSEN.2020.3012294
87. Vangala A, Kumar A, Vinay D, Valery C. Security in IoT-enabled smart agriculture: architecture, security solutions and challenges. Cluster
Comput. 2022;7:1. doi:10.1007/s10586-022-03566-7
88. Bera B, Vangala A, Das AK, Lorenz P, Khan MK. Private blockchain-envisioned drones-assisted authentication scheme in IoT-enabled
agricultural environment. Comput Stand Interfaces. 2022;80(August):1-18. doi:10.1016/j.csi.2021.103567
89. VangalaA,SutralaAK,DasAK,JoM.Smartcontract-basedblockchain-envisionedauthenticationschemefor smart farming. IEEE Internet
Things J. 2021;8(13):10792-10806. doi:10.1109/JIOT.2021.3050676
90. Ali R, Pal AK, Kumari S, Karuppiah M, Conti M. A secure user authentication and key-agreement scheme using wireless sensor networks
for agriculture monitoring. Futur Gener Comput Syst. 2018;84:200-215. doi:10.1016/j.future.2017.06.018
91. Rangwani D, Sadhukhan D, Ray S, Khan MK, Dasgupta M. An improved privacy preserving remote user authentication scheme for
agricultural wireless sensor network. Trans Emerg Telecommun Technol. 2021;32(3):1-31. doi:10.1002/ett.4218
92. Vangala A, Das AK, Lee JH. Provably secure signature-based anonymous user authentication protocol in an internet of things-enabled
intelligent precision agricultural environment. Concurr Comput. 2021;(September):1-27. doi:10.1002/cpe.6187
93. Sandner P, Gross J, Richter R. Convergence of blockchain, IoT, and AI. Front Blockchain. 2020;3(September):1-5. doi:10.3389/fbloc.2020.
522600
94. Kamilaris A, Prenafeta-boldú FX. Deep learning in agriculture: a survey. Comput Electron Agric. 2018;147(July 2017):70-90. doi:10.1016/
j.compag.2018.02.016
95. Ahmadi H, Arji G, Shahmoradi L, Safdari R, Nilashi M, Alizadeh M. The application of internet of things in healthcare: a systematic
literature review and classification. Univers Access Inf Soc. 2019;18(4):837-869. doi:10.1007/s10209-018-0618-4
96. S. Government. IoT and AI for smart government: a research agenda. Gov. Inform. Quart. 2019;36:304-309. doi:10.1016/j.giq.2019.02.003
97. Tzounis A, Katsoulas N, Bartzanas T, Kittas C. Internet of things in agriculture, recent advances and future challenges. Biosyst Eng.
2017;164(November):31-48. doi:10.1016/j.biosystemseng.2017.09.007
How to cite this article: Ahmed RA, Hemdan EE-D, El-Shafai W, Ahmed ZA, El-Rabaie E-SM, Abd
El-Samie FE. Climate-smart agriculture using intelligent techniques, blockchain and Internet of Things:
Concepts, challenges, and opportunities. Trans Emerging Tel Tech. 2022;e4607. doi: 10.1002/ett.4607
A preview of this full-text is provided by Wiley.
Content available from Transactions on Emerging Telecommunications Technologies
This content is subject to copyright. Terms and conditions apply.
