From Industry 4.0 To Agriculture 4.0 Current Status, Enabling Technologies, and Research
From Industry 4.0 To Agriculture 4.0 Current Status, Enabling Technologies, and Research
From Industry 4.0 To Agriculture 4.0 Current Status, Enabling Technologies, and Research
fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
Abstract—The three previous industrial revolutions profoundly for seedbed preparation, sowing, irrigation, weeding, and harvesting.
transformed agriculture industry from indigenous farming to Mechanized agriculture greatly increased food production and reduced
mechanised farming and recent precision agriculture. Industrial
manual labors.
farming paradigm greatly improves productivity, but a number
of challenges have gradually emerged, which have exacerbated The second industrial revolution took place in the 20th century,
in recent years. Industry 4.0 is expected to reshape the referred to as Industry 2.0. On the one hand, the main energy source
agriculture industry once again and promote the fourth which was steam, was replaced by oil and gas. The new energy sources,
agricultural revolution. In this paper, first, we review the current together with innovations in the transportation industry, greatly contributed
status of indus trial agriculture along with lessons learned from
to the development of the agri-food supply chain, in which agricultural
industrialized agricultural production patterns, industrialized
productschain.
agricultural production processes, and the industrialized agri-food supply could be shipped to long distances. Consequently, new
Furthermore, five emerging technologies, namely, the Internet agricultural markets were created for farmers as previously isolated
of Things, robotics, Artificial Intelligence, big data analytics, and communities were connected together. On the other hand, the introduction
blockchain, toward Agriculture 4.0 are discussed. Specifically, of assembly line-based mass production significantly improved the
we focus on the key applications of these emerging technologies
efficiency and productivity. This mass production model in the
in the agricultural sector and corresponding research challenges.
This paper aims to open up new research opportunities for manufacturing industry was then applied to livestock production, where
readers, particularly industrial practitioners. traditional home-based animal husbandry was replaced by large-scale
intensive animal farming.
Index Terms—Industry 4.0, Agriculture 4.0, Industrial Agri
cultivation, Precision Agriculture.
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
Machine Translated by Google
2
IoT
HAVE big data
automationinformation technology
Industry
electrical energy 4.0
mass production Industry
3.0
Intelligence
water power Industry
steam engine 2.0 Electronics
Agriculture
Industry Electricity 4.0
Agriculture
1.0
3.0
Mechanization
yield variable rate
Agriculture autonomous farming
indigenous toolmanual work 2.0 monitoring guidance application trustworthy food supply
pesticide system ubiquitous agriculture sensing
fertilizer
Farming 1.0 animal power tractor
academic and industrial groups
green revolution formed for precision agriculture Industry 4.0 in Agriculture
1784 1870 1950 1969 1992 2011 2017 2030 2050 Industry 4.0
first mechanical loom first assembly line first programmable concept
logic controller
Fig. 1. The development roadmap of industrial revolutions [3] and agricultural revolutions.
The fourth industrial revolution (Industry 4.0) is ongoing, Generally speaking, several authors [3], [5]–[10] presented
and is characterized by a fusion of emerging technologies comprehensive literature reviews and in-depth discussions
such as the Internet of Things (IoT), robotics, Big Data, on how emerging technologies could be harnessed to
Artificial Intelligence (AI), and blockchain technology. At transform manufacturing, production, and supply chain
present, industrial production processes and supply chains management in modern industrial field. Interestingly, the
have become more autonomous and intelligent. application of Industry 4.0 technologies in fighting coronavirus
Correspondingly, the integration of Industry 4.0 and agriculture (COVID-19) pandemic was introduced in [11]. Different from
provides the opportunity to transform industrial agriculture the above discussed works, the motivation of this paper is to
into the next generation, namely, Agriculture 4.0 [4]. In this provide discussions on the adoption of Industry 4.0
context, sustainable and intelligent industrial agriculture approaches in the agriculture industry. Over the last three
would be achieved through real-time variable fine-grained years, many high quality surveys on this topic (ie, the
collection, processing, and analyzing of spatio-temporal data application of modern technologies in the agriculture industry)
in all aspects of the agricultural industry, from food production, [12]–[25] have been conducted, in which the roles of emerging
processing, distribution to consumer experience. Such an technologies such as IoT, robotics ( drone included), and AI
industrial agri culture ecosystem with real-time farm in smart agriculture are provided. The agriculture industry
management, a high degree of automation, and data-driven can be roughly divided into agri-food production and agri-food supply chai
intelligent decision making would greatly improve productivity, Different from existing surveys, this paper provides readers
withresources.
agri-food supply chain efficiency, food safety, and the use of natural the current status and lessons learned from industrial
agriculture in terms of industrialized agricultural production
C. Related Surveys and Our Contributions pattern, industrialized agricultural production process, and
A survey of related work was conducted to compare industrialized agri-food supply chain. Moreover, we aim at
existing state-of-the-art (SoA) papers with this study, as well providing readers with key applications and technical
as direct readers to more information. Literature was collected challenges for future research when Industry 4.0 meets agriculture.
across several academic research databases, including the
Web of Science (WoS), IEEE Xplore, ScienceDirect, and
Google Scholar. Since this study falls in the area of Industry
4.0 and Agriculture 4.0, the following search keywords were The rest of this paper is organized as follows. Section II
used: “Industry 4.0 + Survey/Review”, “Agriculture 4.0 + presents the current status and lessons learned from industrial
Survey/Review”, and “Industry 4.0 + Agriculture 4.0”. To agriculture. Subsequently, Section III discusses enabling
refine the search results, highly cited papers in the field, hot technologies toward Agriculture 4.0 in terms of key
papers in the field, and highly relevant survey reports were applications and research challenges. Finally, concluding
selected, which are summarized in Table I. remarks are high lighted in Section IV.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
TABLE I
SUMMARY OF STATE-OF-THE-ART SURVEYS RELATED TO INDUSTRY 4.0 AND AGRICULTURE 4.0
II. UNDERSTANDING TODAY'S INDUSTRIAL 2) Lessons Learned: Monoculture farming has shown its
AGRICULTURE advantage of increasing yields and boosting profits over many
years of practice, but it also resulted in many ecological
This section discusses the production pattern, production
problems [26]. The growing of a single crop on the same
process, and supply chain in industrial agriculture. More
farmland year by year excessively depletes the soil nutrients,
specifically, we focus on its current status and lessons learned.
which meanwhile cannot be supplemented from surroundings
The take-home messages are summarized in Table II.
due to the lack of biological diversity in monoculture farming.
Nutrient depletion and farmland soil erosion aggravate plant
A. Industrialized Agricultural Production Pattern diseases and pests. To protect crop plants, farmers have to use
the chemical fertilizers, herbicides, and insecticides, which, in
1) Current Status: Monoculture and intensive animal farming
turn, damage farmlands, pollute water, and become a serious
are the typical characteristics of agricultural production in threat to human health.
modern practices. Monoculture is an industrialized pattern of
food production, where a single type of cereal, vegetable, and Intensive animal farming also has negative impacts [27]
fruit is grown on the same farmland year after year. This on environmental pollution, climate change, public health,
pattern increases productivity during the entire agricultural and even animal welfare. Local farmlands are overwhelmed
cycle, as farmers only need to focus on a single plant species, by a million tons of manure produced by a huge number
and create a uniform plan for seeding, fertilization, irrigation, of animals in a concentrated space. Rather than providing
and harvesting. It is also profitable as a specialized set of farm nutrients for crop growth, the redundant waste pollutes soil
equipment for one specific crop is sufficient without investing and river. Moreover, industrial livestock production is one
in other types of expensive farm machinery. major cause of climate change due to the heavy greenhouse
Intensive animal farming means that thousands of chickens, gas emissions. Disease-causing organisms such as bacteria
pigs, or cattle are raised in indoor facilities at a high density and viruses are more easily generated in animal factories and
to produce food (eg, meat, eggs, and milk) more efficiently spread between the crowded animals, farmworkers, and their
and economics. This pattern helps feed the increasing urban communities. Lastly, the unfair treatment of livestock raises
population with few farmers because it not only increases food the animal ethical issue due to the serious harmful physical
production but also promotes the fall of food prices. and physiological effects on livestock.
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
Machine Translated by Google
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Transactions on Industrial Informatics
TABLE II
THE TAKE-HOME MESSAGES OF CURRENT STATUS AND LESSONS LEARNED ABOUT INDUSTRIAL AGRICULTURE
Lessons Learned • Ecological Problems ÿ soil • Lack of Digitization • Food Safety Issue
degradation ÿ lost ÿ pre-digital machinery ÿ ÿ food contamination
B. Industrialized Agricultural Production Process capability in agricultural information systems hinder multi scale
analysis, environment reproduction, and make it difficult to
1) Current Status: The mechanization revolution in discover biological knowledge. The use of
agricultural production has ended or is experiencing a rapid transition
big data in agriculture is challenging as agricultural data are
in most countries. Agricultural machinery and equipment are collected from individuals, research groups, and companies
now widely used during the entire production process, including using different types of devices, which causes multi-sources
land preparation, crop planting, fertilization, harvesting, and heterogeneity problems [32]. In addition, secure data
animal feeding, and food processing. Agricultural mechaniza sharing between multiple groups is an obstacle to the full
tion significantly reduces manual work and improves produc exploitation of agricultural big data. Consequently, the above
tivity, so that fewer farmers can provide more food to meet the discussed issues need to be addressed from all perspectives on
high global demand for food. More excitingly, the innovation standardization, technological innovation, and legislation.
of information and communication technology (ICT) and Last but not least, most of the developed mechanized agri
its integration with agricultural production is promoting the cultural machinery and systems are for large farms since large
digital farming transformation. Sensors are used to measure scale monoculture farming is their main production pattern.
the status of soil and plant leaves for precise microclimate However, different kinds of small farms are widely distributed
control [28]. Low-power wide area networks and wireless mesh across the world [33], especially in developing countries.
networks are deployed to report the data generated during In addition, the benefits of small farms are gradually recognized
agricultural production [29]. Large-scale farmland monitoring, [34], [35]. Distributed small farms would be a possible
crop identification, and yield forecasting are available through agricultural production pattern for sustainable agriculture in
remote sensing [30]. the era of Industry 4.0. Therefore, the industrialization of small
2) Lessons Learned: Compared to the advanced industrial farms also should be paid significant attention [36], [37].
production processes, the automation capability in agricultural
production process is limited due to the following two barriers: C. Industrialized Agri-food Supply Chain
lack of digitization and lack of intelligence in agriculture. 1) Current Status: It usually involves five stages from farm
Most of the agriculture machines in use are still pre-digital. to fork, namely, food production, industrial processing,
The analog equipment works independently and is manually distribution, retail, and consumption. A few types of vegetables,
operated by humans. To further improve the automation fruits, or animals are mass produced on each farm. After the
capability, full connectivity of agricultural machinery is vital harvesting and selection by farmers, these agricultural foods
during the whole production process. Precision agriculture is products are sent to rural cooperatives or food brokers for
still in its preliminary stage. Real-time monitoring of farmland, industrial processing such as sorting, grading, and packaging.
livestock, and other agricultural activities through wireless The food products are then aggregated to the top wholesalers
sensor networks (WSNs) are currently in small-scale and in each region through transportation and logistics services.
short-term due to the high deployment and maintenance cost. Thousands of food retailers such as supermarkets, farmers'
Further improvement is needed in the network bandwidth and markets, and grocery stores in the region need to place orders
delivery latency to ensure large-scale high-throughput plant for the food products from the top wholesalers to make food
phenotyping [31]. In rural areas, the lack of telecommunications products available for sale. Finally, consumers can purchase
infrastructure can only provide limited signal coverage, their favorite vegetables, fruits, and meats conveniently at retail
making it difficult to increase agricultural productivity. outlets. This model is commonly referred to as business-to-
Furthermore, weak data integration and element tracking business-to-consumer (B2B2C) [38].
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
Machine Translated by Google
5
In recent years, the integration of the Internet and agriculture Communications Satellite
has promoted new food supply models such as online-to
offline (O2O) [39], social commerce [40], and customer-to Sat Nav Remote Sensing Satellite
customer (C2C) [41]. These models greatly improve shopping
convenience, as prospective customers could search and
order preferred food products online. Then the products could
be either picked up in stores or delivered at home.
2) Lessons Learned: The three lessons learned in the agri Drone Swarm
food supply chain are food safety issue, asymmetric and
fragmented information, and an inefficient supply chain.
First, accidental and deliberate food contamination incidents
have occurred too often in the last decades. These incidents
usually result from the use of prohibited chemicals, unsafe raw
materials, unapproved food additives, as well as insufficient
poor storage and poor transportation conditions. Food fraud
and low food quality lead to significant negative impacts on Underground WSNs Underwater WSNs
public health and consumer confidence. Therefore, ensuring
transparency and traceability throughout the entire agri-food Fig. 2. Space-Air-Ground-Undersurface Integrated Network (SAGUIN) for
supply chain is essential to prevent food fraud, improve the ubiquitous agriculture sensing and networking.
quality of food, and finally boost consumer confidence.
Second, asymmetric information between food production
sectors and end consumers leads to an imbalance between A.Internet of Things
supply and demand, price volatility, and foods scarcity.
Fragmented information exchange or even data isolation 1) Key Applications of IoT in Agriculture: The IoT, as a core
across the agri-food supply chain causes inadequate technology in Industry 4.0, is transforming many aspects of
communication between multiple actors and the lack of food traceability.
our daily lives by creating a smart connected world. Some use
These, in turn, affect logistics efficiency, cost containment, risk cases are smart home, industrial internet, and connected
management, customer intimacy, and consumer confidence. vehicles. Accordingly, it is also expected to encourage the
Thus, it is urgent to realize a consistent interconnection be agricultural sector. The agricultural applications of IoT include
tween all the elements, so that data and discovered knowledge precision farming, livestock monitoring, smart greenhouse,
could be efficiently shared among the elements in the agri- fishery management, and weather tracking. As these
food supply chain. agricultural applications of IoT were comprehensively reviewed
Third, according to the Food and Agriculture Organization in [12], [17], [18], we herein discuss them at a general level,
of the United Nations, more than 30% of the world's food namely Space-Air-Ground-Undersurface Integrated Network
produced each year is not consumed due to food loss and (SAGUIN) for ubiquitous agriculture sensing and networking.
food waste. Food loss occurs at every stage of the agri-food
The paradigm of SAGUIN is demonstrated in Fig. 2. It would
supply chain, especially during food harvesting, sorting,
be realized through the combination of remote sensing,
grading, and processing. This is due to the lack of technology
drones, smart agricultural vehicles, WSNs, and mobile crowd
and adequate management, such as a mismatch in supply
sensing. Based on national and commercial space
and demand, poor food quality, and lack of processing
infrastructure, the establishment of remote sensing satellite
techniques, in the existing agri-food supply chain. Meanwhile,
constellation would be able to provide full coverage acquisition
food waste usually happens during the later stages in
of agricultural information, which is especially efficient for
distribution, retailing, and consumption. This is mainly because
crop production forecasting, yield modeling, pest identification, and so on.
the food is damaged or spoiled during the logistics process, In addition, the advanced unmanned aerial vehicles equipped
expires at retailers, or consumers end. Accordingly, it is
with hyperspectral sensors, multispectral cameras, and other
essential to further improve the technical and managerial
novel instruments can provide fast emergency responses and
capabilities throughout the agri food supply chain to prevent food loss and waste.
improve observation precision through high-throughput 3D
monitoring at different geographical areas. Finally, different
kinds of agricultural sensor nodes, autonomous farm vehicles,
III. INDUSTRY 4.0 TECHNOLOGIES IN AGRICULTURE and mobile crowdsensing are responsible for ground and un
dersurface perception. Thanks to advances in communication
Using emerging Industry 4.0 technologies in agricultural technologies, different types of wireless networks (eg, 5G,
production and agri-food supply chain management provides LoRa, NB-IoT, Sigfox, ZigBee, and WiFi) can be chosen to
an opportunity to address the limitations discussed above. meet the diverse service requirements in agricultural
Correspondingly, the blueprint of Agriculture 4.0 would also applications (such as real-time remote equipment control,
be realized. Five emerging technologies on key applications high throughput plant phenotyping) for better coverage,
for Agriculture 4.0 and research challenges are discussed below. connection density, bandwidth, and end-to-end latency.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
rest of the agricultural area is under threat. Plant factory [53] Agricultural Environment
looks promising to meet global food demand and sustainable
use of natural resources. 3D food printing [54] is becoming an
alternative to producing food in an automated additive agronomy horticulture forestry aquaculture livestock farming
manufacturing manner. In biodiverse farms, multiple crops are
grown simultaneously on a farmland. The crop diversity helps
to make soil nutrients more balance, and to protect against Stage 1: Stage 2: Stage 3:
Navigation Detection Stock
diseases and pests. With the help of agricultural robotics,
• path following • feature selection • object localization
limitations such as low productivity and intensive manual labor
• trajectory tracking • classification • task allocation
could be solved. It should be noted that biodiverse farming
• obstacle avoidance • human-robot
also increases the complexity in the application of agricultural
• collective motion interaction
robotics due to variation in detection, classification, harvesting,
and other agricultural activities. The technical challenges will
be discussed later on. Fig. 4. The core operation stages in agricultural robotics along with key
tasks [66].
On the other hand, RAS is also transforming the agricultural
production process, where the key applications include au
tonomous farming, aerial spraying and monitoring, automatic b) Accurate Detection: In the second stage, the two urgent
animal husbandry, etc. Smart tractors [55] would work collab tasks are suitable biological feature selection and robust object
oratively to generate operational routes and intelligently avoid classification for the following reasons. First, multi spectral
barriers in the field to ensure the safety for both farmland and image is commonly chosen as a feature in selective spraying
humans. Robotic weeders [56] can differentiate weeds from and mechanical weeding removal applications, so that the
crops through computer vision, and then precisely spray the agricultural robots can differentiate weeds from crops.
herbicide on weeds only or directly eradicate them without However, the dynamic and complex farm environment makes
harming the crops. Picking and harvesting robots [57] help it challenging to precisely classify plants because of the
farmers to collect tomatoes, citrus, apples, and strawberries changing appearance, growth stage, weather condition, object
more efficiently. They recognize fruits and vegetables in the overlapping, or even partial occlusion [67], [68]. Second, the
field, determine their location, and harvest the ripe ones in robots used for automatic harvesting should be capable of
boxes. Since operating speed is dozens of times faster than identifying the different types of fruits and vegetables, their
ground machines, aerial crop spraying and monitoring [58] has maturity, size, texture properties for precision picking, quality
attracted considerable attention recently. Finally, RAS also grading, and other purposes. Unfortunately, the dif ferent
contributes to animal husbandry such as automatic feeding, kinds of plants and unstructured orchard with dynamic
milking, and herding. environmental conditions lead to performance degradation
2) Research Challenges of RAS in Agriculture: We can see [69] when classical pattern recognition methods are applied to
that, from Fig. 4, three research challenges in agricultural RAS agricultural practices. Unreliable classification methods would
are autonomous navigation, accurate detection, and intelligent undoubtedly cause serious losses to agricultural production.
action. The detail discussions are presented as follows: Third, the large fields in breeding nurseries need to be rapidly
a) Autonomous Navigation: In the first stage, RAS should monitored in a short duration for the germinated crop detection
make sense of the target in the agricultural environment to so that the best plant seeds can be found for future breeding.
fulfill path following [59], trajectory tracking [60], obstacle This requires fast classification methods to distinguish the
avoidance [61], or collective motion [62]. Accurate guidance is good and bad crop seeds germination [70]. Therefore, there is
extremely important since it affects crop protection, human a need to solve these problems for the improvement in
safety, operation costs, and work efficiency. Currently, the detection accuracy and operational
global navigation satellite system (GNSS), camera vision, and effectiveness. c) Intelligent Action: Finally, the specific
laser imaging detection and ranging (LIDAR) scanner are three tasks such as robot-assisted plant phenotyping [71], fruit
popular techniques for autonomous driving. However, each counting and harvesting [72], leaf peeling [73], selective
one has limitations when used alone in agricultural spraying [74], and 3D mapping [58] are completed during internship action.
environments. For example, satellite signal attenuation and Three fundamental research concerns are object localization,
task allocation, and human-robot interaction. Accurate object
multipath effects frequently occur in orchards and greenhouses [63].
The performance of computer vision-based navigation system localization is the core requirement to ensure inerrant
is often degraded due to unstable light conditions in farm fields operations and minimize the external influences from the
and plant growth [64]. Although LIDAR scanner-based surroundings. Optimized task sequence planning of individual
navigation systems outperform the other two approaches in agricultural robot systems and coordination between multiple
orchard and grove scenarios, the overhanging tree branches, robots are critical to improving performance in terms of energy
moving objects, and other obstacles affect its performance on efficiency, time consumption, and reliability. Human-robot
tree row detection [65]. Therefore, novel guidance algorithms interaction has been widely studied in the industrial domain,
and control strategies are needed. Moreover, the combination but the progress in agriculture is still nascent. Proper interaction
of these techniques with data fusion is helpful to ensure between farmers and autonomous agricultural machinery are
navigation accuracy and robustness. extremely necessary to improve productivity, profitability, and
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
prevent potential accidents [75]. Last but not least, robotic arm
design is equally important as different types of grippers need
to be developed for enhancing the effectiveness and efficiency
following their specific applications.
above, AI plays an important role in RAS. With classification, • fruit picking • labor management • soil health analysis • price forecasting •
grading and sorting food supply chain • farm at hand production insights
logistical regression, association analysis, and decision-making
capabilities, AI is also being applied to other applications in Artificial Intelligence and Big Data in Agriculture
agriculture, such as agricultural decision support system
(ADSS), mobile agricultural expert system, and agricultural
Fig. 5. Key applications of AI and big data in agricultural industry.
predictive analytics.
Four use cases of ADSS are reviewed in [22], which
includes mission planning, water resources management, b) Gap between Farmers and AI Researchers: Seam
climate change adaptation, and food waste control. While a less translation from agricultural problems to decision-making
recent work [24] focused on the review of crop data models is crucial for the success of adopting AI in agriculture.
management, in which its decision-making capability enabled Farmers are often met with a lot of difficulties in the agri
by ADSS was presented. Moreover, intelligent animal health cultural production process, but AI researchers are not aware
monitoring systems is a major focus of today's research and of these agricultural problems, which could be fixed by AI
drive the growth of AI in agriculture markets. Animal technology. More importantly, AI is practical only after the
wearables, computer vision systems, and other sensing nature of these agricultural problems and solutions are fully
devices can capture the status of animals in real-time. Then, understood. However, AI researchers usually do not know
the intelligence engine helps analyze livestock health, animal much about professional agricultural knowledge. Thus, it is
welfare, production, etc. In recent years, many technological important to link farmers, agricultural professionals and AI
advances have been made in the area of intelligent animal researchers together to bridge the gap.
health monitoring. The advancement in biosensors was summarized inc)[76] and [77].Secure Machine Learning: To success
Distributed
Machine vision approaches for animal behavior detection was fully implement an artificial neural network, fuzzy control
reviewed in [78]. A recent study [79] systematically discussed system, and other AI engines in agriculture, a huge amount of
the ADSS for intelligent aquaponics. Also, a pilot study [80] data is needed to train AI models. However, the related data
shows that digital twin technology is helpful to prevent animal on agriculture is much less than that in the industrial field due
diseases in livestock.
to the lack of digitization in agriculture, inaccessible data, as
As shown in Fig. 5, mobile expert system is another key well as privacy protection issues. With the help of deploying
application of AI in agriculture. Farmers can now easily use 5G infrastructure in rural areas [82], SAGUIN, and novel
smartphones to identify plant pests and diseases [81]. sensors, massive agricultural data could be collected in the future.
They are also able to identify soil problems on their own with Novel ML techniques such as federated learning [83] are
the help of mobile apps. In addition, the progress of farmland expected to be explored for farm data protection in agricultural
can be tracked by satellite imagery, which is then analyzed by AI applications.
AI engine. Mobile apps simultaneously visualize the result so
that farmers could understand the status remotely.
Finally, AI-enabled agricultural predictive analytics and big D.Big Data Analytics
data technology are capable of forecasting weather conditions, 1) Key Applications of Big Data in Agriculture: IoT helps
predicting crop yields, modeling agricultural market volatility, collect data in every step of the agricultural production and
and performing price estimation. Based on these observations, agri-food supply chain management. Thus, it would also be
professional insights in production can be provided to farmers, beneficial to implement big data analytics during food
as well as to guide companies to optimize business resources. production, processing, logistics, and marketing. Data-driven
2) Research Challenges of AI in Agriculture: Although these agri cultural industry would profoundly transform the production
applications are promising, agriculture is one of the most and consumption behaviors [12]. For instance, ADSS, mobile
challenging practical areas for AI for the following reasons: agricultural expert system, and agricultural predictive analytics
a) Hard to Find Single Standard Solution: Unlike the all rely on the power of big data, which can provide smart
standardized industrial environment, the conditions in the recommendation to farmers toward precision farming [84].
agricultural fields are constantly changing, making statistical Accurate risk assessment could help farmers to better manage
quantification difficult. All the factors such as weather, soil agricultural risks in terms of production, market, institutional
quality, and visited pests have an impact on AI's performance. risk, along with personal and financial risk [85].
As a result, a machine learning (ML) algorithm that performs Moreover, big data plays a key role in agri-food supply chain
a task well in one field may fail in neighboring fields, not to management in addressing to solve food safety, imbal ance
mention in other regions. supply, as well as food loss and waste. For example, the
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
10
TABLE III
THE TAKE-HOME MESSAGES OF ENABLING TECHNOLOGIES TOWARD AGRICULTURE 4.0
of these approaches into agriculture and analyzed blockchain confidentiality, privacy, and integrity assurance. However, like
based privacy protection for agriculture applications. other popular network technologies, blockchain has become
2) Research Challenges of Blockchain in Agriculture: The a target for hackers, who can launch cyber-attack from the
adoption of blockchain during agri-food production and supply protocol side, implementation side, and logical defect. AT
chain management is an opportunity to improve traceability, detailed survey on cyber-attack in a blockchain network and
transparency, and trust among providers, farmers, suppliers, recent advances can be found in [99].
retailers, and consumers. However, many technical bottlenecks
need to be resolved to adopt blockchain technology in agri IV. CONCLUSION
culture. These bottlenecks are listed as follows: Monoculture and intensive animal farming are the main
a) Interoperability: Currently, thousands of blockchains agricultural production patterns in today's industrial agriculture.
projects are active. They are developed on a wide variety However, the damage to the ecological environment,
of platforms with different programming languages, protocol public health, and animal welfare are big limitations. Although
stacks, and security mechanisms. The lack of interoperability presently, the agricultural production process is mechanized
leads to network isolation problems, where the different types and computerized, the lack of digitization and intelligence
of blockchain networks can hardly communicate with each are major obstacles to improving the automation capability.
other. To integrate existing blockchain networks, it is necessary Moreover, the agri-food supply chain at the current stage is not
to propose novel architecture, interoperable communication intelligently managed. To address these issues, it is essential to
protocol, and improved middleware so that efficient cross integrate emerging Industry 4.0 technologies into agriculture.
blockchain communication [94], [95] is realized. Hence, in this study, we present detailed discussions on the key
b) Scalability: Unlike the Visa transaction network, applications and research challenges when these technologies
which is capable of processing thousands of transactions in meet agriculture. For the ease of readers, the take-home
a second, the transaction speed in blockchain-based smart messages are summarized in Tables II and III.
contract networks such as Bitcoin and Ethereum are less than
20 per second [96]. Fast response time and high network ACKNOWLEDGMENT
throughput are essential since innumerable transactions need The authors would like to thank the anonymous reviewers
to be processed when millions of actors in the entire supply for their constructive feedback. The authors also would like to
chain are involved in the integrated networks. thank the icon authors from Flaticon (www.flaticon.com). This
c) Energy Consumption: Apart from the huge amount of work was supported in part by the National Natural Science
energy expended by bitcoin mining, the complex consensus Foundation of China under Grant 61902188, the Fundamental
mechanisms in transaction validation also lead to high energy Research Funds for the Central Universities under Grant
consumption. Therefore, it is necessary to reduce the complex KJQN202047, and the China Postdoctoral Science Foundation
ity of blockchain networks for cost savings. For example, the under Grant 2019M651713.
concept of proof-of-stake (PoS) [97] has been proposed for
lightweight distributed consensus. REFERENCES
d) Security and Privacy: The immutable data storage,
[1] MA Rapela, “Fostering innovation for agriculture 4.0: A comprehen sive
decentralized network, and peer-to-peer communication are plant germplasm system,” Fostering Innovation for Agriculture 4.0,
three key features of blockchain technology that make it 2019.
promising to enhance security and privacy for IoT. For example, [2] F. Yandun Narvaez, G. Reina, M. Torres-Torriti, G. Kantor, and FA
Cheein, “A survey of ranging and imaging techniques for precision
[98] presented a systematic review of blockchain based security agriculture phenotyping,” IEEE/ASME Transactions on Mechatronics,
services solutions, including authentication, flight. 22, no. 6, p. 2428–2439, Dec 2017.
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
11
´
[3] G. Aceto, V. Persico, and A. Pescape, “A survey on information and [26] SB Brodt, NM Fontana, and LF Archer, “Feasibility and sustainability of
communication technologies for industry 4.0: State-of-the-art, tax agroforestry in temperate industrialized agriculture: preliminary insights
onomies, perspectives, and challenges,” IEEE Communications Surveys from california,” Renewable Agriculture and Food Systems, pp. 1–9,
Tutorials, flight. 21, no. 4, p. 3467–3501, Fourthquarter 2019. 2019.
[4] M. De Clercq, A. Vats, and A. Biel, “Agriculture 4.0: The future of farming [27] LA Smit and D. Heederik, “Impacts of intensive livestock production on
technology,” Proceedings of the World Government Summit, Dubai, human health in densely populated regions,” GeoHealth, vol. 1, no. 7, p.
UAE, pp. 11–13, 2018. 272–277, 2017.
[5] Y. Lu, “Industry 4.0: A survey on technologies, applications and open [28] W. Chen, Y. Lin, Y. Lin, R. Chen, J. Liao, F. Ng, Y. Chan, Y. Liu, C. Wang,
research issues,” Journal of Industrial Information Integration, vol. 6, p. C. Chiu, and T. Yen, “Agritalk : Iot for precision soil farming of turmeric
1–10, 2017. cultivation,” IEEE Internet of Things Journal, vol. 6, no. 3, p. 5209–5223,
[6] Q. Qi and F. Tao, “Digital twin and big data towards smart manufacturing June 2019.
and industry 4.0: 360 degree comparison,” IEEE Access, vol. 6, p. 3585– [29] N. Ahmed, D. De, and I. Hussain, “Internet of things (iot) for smart
3593, 2018. precision agriculture and farming in rural areas,” IEEE Internet of Things
[7] LD Xu, EL Xu, and L. Li, “Industry 4.0: state of the art and future trends,” Journal, vol. 5, no. 6, p. 4890–4899, Dec 2018.
International Journal of Production Research, vol. 56, no. 8, p. 2941– [30] M. Weiss, F. Jacob, and G. Duveiller, “Remote sensing for agricultural
2962, April 2018. applications: A meta-review,” Remote Sensing of Environment, vol. 236,
[8] AG Frank, LS Dalenogare, and NF Ayala, “Industry 4.0 technologies: p. 111402, 2020.
Implementation patterns in manufacturing companies,” International [31] F. Tardieu, L. Cabrera-Bosquet, T. Pridmore, and M. Bennett, “Plant
Journal of Production Economics, vol. 210, p. 15–26, 2019. phenomics, from sensors to knowledge,” Current Biology, vol. 27, no.
[9] RD Raut, A. Gotmare, BE Narkhede, UH Govindarajan, and SU Bokade, 15, p. 770–783, 2017. ´
“Enabling technologies for industry 4.0 manufacturing and supply chain: [32] JM Costa, J. Marques da Silva, C. Pinheiro, M. Baron, P. Mylona, M.
concepts, current status, and adoption challenges,” Centritto, M. Haworth, F. Loreto, B. Uzilday, I. Turkan, and MM
IEEE Engineering Management Review, p. 1–1, 2020. Oliveira, “Opportunities and limitations of crop phenotyping in southern
[10] E. Oztemel and S. Gursev, “Literature review of Industry 4.0 and related european countries,” Frontiers in Plant Science, vol. 10, p. 1125, 2019.
technologies,” Journal of Intelligent Manufacturing, vol. 31, no. 1, p. [33] N. Guiomar, S. Godinho, T. Pinto-Correia, M. Almeida, and F. Bartolini et
127–182, January 2020. al., “Typology and distribution of small farms in europe: Towards a better
[11] M. Javaid, A. Haleem, R. Vaishya, S. Bahl, R. Suman, and A. Vaish, picture,” Land Use Policy, vol . 75, p. 784–798, 2018.
“Industry 4.0 technologies and their applications in fighting covid 19 [34] JW Mellor and SJ Malik, “The impact of growth in small commercial farm
pandemic,” Diabetes & Metabolic Syndrome: Clinical Research & productivity on rural poverty reduction,” World Development, vol. 91, p.
Reviews, vol. 14, no. 4, p. 419–422, 2020. 1–10, 2017.
[35] M. Gautam and M. Ahmed, “Too small to be beautiful? the farm size and
[12] O. Elijah, TA Rahman, I. Orikumhi, CY Leow, and MN Hindia, “An
overview of internet of things (iot) and data analytics in agriculture: productivity relationship in bangladesh,” Food Policy, vol. 84, p. 165–
175, 2019.
Benefits and challenges,” IEEE Internet of Things Journal, vol. 5, no. 5,
p. 3758–3773, Oct 2018. [36] DF Larson, R. Muraoka, and K. Otsuka, “Why african rural development
strategies must depend on small farms,” Global Food Security, vol. 10,
[13] AT Braun, E. Colangelo, and T. Steckel, “Farming in the era of industry
p. 39–51, 2016.
4.0,” in 51st CIRP Conference on Manufacturing Systems, vol. 72, 2018,
[37] GP Paudel, H. Gartaula, DB Rahut, and P. Craufurd, “Gender differ
p. 979–984.
´ entiated small-scale farm mechanization in nepal hills: An application of
[14] I. Kovacs and I. Husti, “The role of digitalization in the agricultural 4.0 –
exogenous switching treatment regression,” Technology in Society, vol.
how to connect the industry 4.0 to agriculture?” Hungarian Agricultural
61, p. 101250, 2020.
Engineering, p. 38–42, 2018.
[38] S. Iankova, I. Davies, C. Archer-Brown, B. Marder, and A. Yau, “A
[15] JP Belaud, N. Prioux, C. Vialle, and C. Sablayrolles, “Big data for agri-
comparison of social media marketing between b2b, b2c and mixed
food 4.0: Application to sustainability management for by-products supply
business models,” Industrial Marketing Management, vol. 81, p. 169–
chain,” Computers in Industry, vol. 111, p. 41–50, 2019.
179, 2019.
[16] I. Zambon, M. Cecchini, G. Egidi, MG Saporito, and A. Colantoni,
[39] CW Shen, M. Chen, and CC Wang, “Analyzing the trend of o2o commerce
“Revolution 4.0: Industry vs. agriculture in a future development for
by bilingual text mining on social media,” Computers in Human Behavior,
smes,” Processes, vol. 7, no. 1, p. 36, 2019.
vol. 101, p. 474–483, 2019.
[17] M. Ayaz, M. Ammad-Uddin, Z. Sharif, A. Mansour, and EM Aggoune, [40] H. Han, H. Xu, and H. Chen, “Social commerce: A systematic review and
“Internet-of-things (iot)-based smart agriculture: Toward making the fields data synthesis,” Electronic Commerce Research and Applications, vol.
talk,” IEEE Access, vol . 7, p. 129,551–129,583, 2019. 30, p. 38–50, 2018.
[18] MS Farooq, S. Riaz, A. Abid, K. Abid, and MA Naeem, “A survey on the [41] JW Peltier, AJ Dahl, and EL Swan, “Digital information flows across a b2c/
role of iot in agriculture for the implementation of smart farming,” c2c continuum and technological innovations in service ecosystems: A
IEEE Access, vol. 7, p. 156,237–156,271, 2019. service-dominant logic perspective,” Journal of Business Research, 2020.
[19] J. Kim, S. Kim, C. Ju, and HI Son, “Unmanned aerial vehicles in agri
culture: A review of perspective of platform, control, and applications,” [42] J. Ding and R. Chandra, “Towards low cost soil sensing using wi-fi,” in
IEEE Access, vol. 7, p. 105,100–105,115, 2019. The 25th Annual International Conference on Mobile Computing and
[20] J. Ruan, H. Jiang, C. Zhu, X. Hu, Y. Shi, T. Liu, W. Rao, and FTS Networking, ser. MobiCom '19. New York, NY, USA: Association for
Chan, “Agriculture iot: Emerging trends, cooperation networks, and Computing Machinery, 2019.
outlook,” IEEE Wireless Communications, vol. 26, no. 6, p. 56–63, [43] F. Yang, L. Shu, Y. Liu, K. Li, K. Huang, Y. Zhang, and Y. Sun, “Poster:
December 2019.
Photovoltaic agricultural internet of things the next generation of smart
[21] J. Miranda, P. Ponce, A. Molina, and P. Wright, “Sensing, smart and farming,” in Proceedings of the 2019 International Conference on
sustainable technologies for agri-food 4.0,” Computers in Industry, vol. Embedded Wireless Systems and Networks, ser. EWSN '19. USA:
108, p. 21–36, 2019. Junction Publishing, 2019, p. 236–237.
[22] Z. Zhai, JF Mart´ÿnez, V. Beltran, and NL Mart´ÿnez, “Decision support [44] Q. Chen, Y. Liu, G. Liu, Q. Yang, X. Shi, H. Gao, L. Su, and Q. Li, “Harvest
systems for agriculture 4.0: Survey and challenges,” Computers and energy from the water: A self-sustained wireless water quality sensing
Electronics in Agriculture, vol. 170, p. 105256, 2020. system,” ACM Trans. Embed. Computer. System, vol. 17, no. 1, Sep.
[23] MA Ferrag, L. Shu, X. Yang, A. Derhab, and L. Maglaras, “Security and 2017.
¨ ¨
privacy for green iot-based agriculture: Review, blockchain solutions, [45] Y. Liu, T. Voigt, N. Wirstrom, and J. Hoglund, “Ecovibe: On-demand
and challenges,” IEEE Access, pp. 1–1, 2020. sensing for railway bridge structural health monitoring,” IEEE Internet of
´
[24] V. Saiz-Rubio and F. Rovira-Mas, “From smart farming towards agriculture Things Journal, vol. 6, no. 1, p. 1068–1078, Feb 2019.
5.0: A review on crop data management,” Agronomy, vol. 10, no. 2, 2020. [46] L. Sigrist, N. Stricker, D. Bernath, J. Beutel, and L. Thiele, “Thermo
electric energy harvesting from gradients in the earth surface,” IEEE
[25] M. Lezoche, JE Hernandez, M. del Mar Eva Alemany D´ÿaz, H. Panetto, Transactions on Industrial Electronics, pp. 1–1, 2019.
and J. Kacprzyk, “Agri-food 4.0: A survey of the supply chains and [47] F. Tonolini and F. Adib, “Networking across boundaries: Enabling wireless
technologies for the future agriculture,” Computers in Industry, vol. 117, communication through the water-air interface,” in Proceedings of the
p. 103187, 2020. 2018 Conference of the ACM Special Interest Group on Data
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
12
Communication, ser. SIGCOMM '18. New York, NY, USA: Association images and mav for smart farming,” IEEE Robotics and Automation
for Computing Machinery, 2018, p. 1175–131. Letters, vol. 3, no. 1, p. 588–595, Jan 2018.
[48] Y. He, J. Guo, and X. Zheng, “From surveillance to digital twin: Challenges [69] I. Sa, C. Lehnert, A. English, C. McCool, F. Dayoub, B. Upcroft, and T.
and recent advances of signal processing for industrial internet of Perez, “Peduncle detection of sweet pepper for autonomous crop
things,” IEEE Signal Processing Magazine, vol. 35, no. 5, p. 120–129, harvesting—combined color and 3-d information ,” IEEE Robotics and
Sep. 2018. ˜ Automation Letters, vol. 2, no. 2, p. 765–772, April 2017.
¨
[49] CA Boano,¨ M. Zu´niga, J. Brown, U. Roedig, C. Keppitiyagama, and K. [70] J. Valente, L. Kooistra, and S. Mucher, “Fast classification of large
Romer, “Templab: A testbed infrastructure to study the impact of germinated fields via high-resolution uav imagery,” IEEE Robotics and
temperature on wireless sensor networks,” in Proceedings of the 13th Automation Letters, vol. 4, no. 4, p. 3216–3223, Oct 2019.
International Symposium on Information Processing in Sensor Networks, [71] C. Wu, R. Zeng, J. Pan, CCL Wang, and Y. Liu, “Plant phenotyping by
ser. IPSN '14. IEEE Press, 2014, p. 95–106. ¨ deep-learning-based planner for multi-robots,” IEEE Robotics and
[50] CA Boano, M. Cattani, and K. Romer, “Impact of temperature variations Automation Letters, vol. 4, no. 4, p. 3113–3120, Oct 2019.
on the reliability of lora,” in Proceedings of the 7th International [72] E. Bellocchio, TA Ciarfuglia, G. Costante, and P. Valigi, “Weakly
Conference on Sensor Networks. SCITEPRESS-Science and Technology supervised fruit counting for yield estimation using spatial consistency,”
Publications, Lda, 2018, p. 39–50. IEEE Robotics and Automation Letters, vol. 4, no. 3, p. 2348–2355, July
[51] J. Luomala and I. Hakala, “Effects of temperature and humidity on radio 2019.
signal strength in outdoor wireless sensor networks,” in 2015 Federated [73] J. Hughes, L. Scimeca, I. Ifrim, P. Maiolino, and F. Iida, “Achieving
Conference on Computer Science and Information Systems (FedCSIS), robotically peeled lettuce,” IEEE Robotics and Automation Letters, vol.
Sep. 2015, p. 1247–1255. 3, no. 4, p. 4337–4342, Oct 2018.
[52] M. Zimmerling, L. Mottola, and S. Santini, “Synchronous transmissions in [74] R. Berenstein and Y. Edan, “Automatic adjustable spraying device for site-
low-power wireless: A survey of communication protocols and network specific agricultural application,” IEEE Transactions on Automation
services,” arXiv preprint arXiv:2001.08557, 2020. Science and Engineering, vol. 15, no. 2, p. 641–650, April 2018.
[53] Y. Fujimoto, S. Murakami, N. Kaneko, H. Fuchikami, T. Hattori, and Y. [75] JP Vasconez, GA Kantor, and FAA Cheein, “Human–robot interaction in
Hayashi, “Machine learning approach for graphical model-based analysis agriculture: A survey and current challenges,” Biosystems Engineering,
of energy-aware growth control in plant factories,” IEEE Access, vol. 7, vol. 179, p. 35–48, 2019.
p. 32,183–32,196, 2019. [76] S. Neethirajan, “Recent advances in wearable sensors for animal health
[54] JI Lipton, M. Cutler, F. Nigl, D. Cohen, and H. Lipson, “Additive man management,” Sensing and Bio-Sensing Research, vol. 12, p. 15–29,
ufacturing for the food industry,” Trends in Food Science & Technology, 2017.
vol. 43, no. 1, p. 114–123, 2015.
[77] S. Neethirajan, SK Tuteja, ST Huang, and D. Kelton, “Recent advancement
[55] FA Auat Cheein and R. Carelli, “Agricultural robotics: Unmanned robotic
in biosensors technology for animal and livestock health management,”
service units in agricultural tasks,” IEEE Industrial Electronics Magazine,
Biosensors and Bioelectronics, vol. 98, p. 398–407, 2017.
vol. 7, no. 3, p. 48–58, Sep. 2013.
[56] P. Lottes, J. Behley, A. Milioto, and C. Stachniss, “Fully convolutional
[78] A. Nasirahmadi, SA Edwards, and B. Sturm, “Implementation of machine
networks with sequential information for robust crop and weed detection
vision for detecting behavior of cattle and pigs,” Livestock Science, vol.
in precision farming,” IEEE Robotics and Automation Letters, vol. 3, no.
202, p. 25–38, 2017.
4, p. 2870–2877, Oct 2018.
[79] AR Yanes, P. Martinez, and R. Ahmad, “Towards automated aquaponics:
[57] L. Zhang, J. Jia, G. Gui, X. Hao, W. Gao, and M. Wang, “Deep learning
A review on monitoring, iot, and smart systems,” Journal of Cleaner
based improved classification system for designing tomato harvesting
Production, vol. 263, p. 121571, 2020.
robot,” IEEE Access, vol. 6, p. 67,940–67,950, 2018.
[80] S. Jo, D. Park, H. Park, and S. Kim, “Smart livestock farms using digital
[58] C. Potena, R. Khanna, J. Nieto, R. Siegwart, D. Nardi, and A. Pretto,
twin: Feasibility study,” in 2018 International Conference on Information
“Agricolmap: Aerial-ground collaborative 3d mapping for precision
and Communication Technology Convergence (ICTC), Oct 2018, pp .
farming,” IEEE Robotics and Automation Letters, vol. 4, no. 2, p.
1461–1463.
1085–1092, April 2019.
[59] B. Fernandez, PJ Herrera, and JA Cerrada, “A simplified optimal path [81] S. Prasad, SK Peddoju, and D. Ghosh, “Energy efficient mobile vision
following controller for an agricultural skid-steering robot,” IEEE Access, system for plant leaf disease identification,” in 2014 IEEE Wireless
vol. 7, p. 95,932–95,940, 2019. Communications and Networking Conference (WCNC), April 2014, pp.
3314–3319.
[60] S. Li, H. Xu, Y. Ji, R. Cao, M. Zhang, and H. Li, “Development of a
following agricultural machinery automatic navigation system,” [82] R. Ma, KH Teo, S. Shinjo, K. Yamanaka, and PM Asbeck, “A gan pa for
Computers and Electronics in Agriculture, vol. 158, p. 335–344, 2019. 4g lte-advanced and 5g: Meeting the telecommunication needs of
[61] D. Ball, P. Ross, A. English, P. Milani, D. Richards, A. Bate, B. Up croft, various vertical sectors including automobiles, robotics, health care,
G. Wyeth, and P. Corke, “Farm workers of the future: Vision based factory automation, agriculture, education, and more,” IEEE Microwave
robotics for broad-acre agriculture,” IEEE Robotics Automation Magazine, Magazine, vol. 18, no. 7, p. 77–85, Nov 2017.
vol. 24, no. 3, p. 97–107, Sep. 2017. [83] Q. Yang, Y. Liu, T. Chen, and Y. Tong, “Federated machine learning:
[62] C. Ju and HI Son, “Modeling and control of heterogeneous agricultural Concept and applications,” ACM Trans. Intel. System Technol., vol. 10,
field robots based on ramadge–wonham theory,” IEEE Robotics and no. 2, p. 12:1–12:19, 2019.
Automation Letters, vol. 5, no. 1, p. 48–55, Jan 2020. [84] S. Wolfert, L. Ge, C. Verdouw, and MJ Bogaardt, “Big data in smart
[63] CP Baillie, CR Lobsey, DL Antille, CL McCarthy, and JA farming – a review,” Agricultural Systems, vol. 153, p. 69–80, 2017.
Thomasson, “A review of the state of the art in agricultural automation. [85] AM Komarek, AD Pinto, and VH Smith, “A review of types of risks in
part iii: Agricultural machinery navigation systems,” in 2018 ASABE agriculture: What we know and what we need to know,”
Annual International Meeting. American Society of Agricultural and Agricultural Systems, vol. 178, p. 102738, 2020.
Biological Engineers, 2018, p. 1. [86] A. Weersink, E. Fraser, D. Pannell, E. Duncan, and S. Rotz, “Opportunities
[64] S. Bonadies and SA Gadsden, “An overview of autonomous crop row and challenges for big data in agricultural and environmental analysis,”
navigation strategies for unmanned ground vehicles,” Engineering in Annual Review of Resource Economics, vol. 10, no. 1, p. 19–37, 2018.
Agriculture, Environment and Food, vol. 12, no. 1, p. 24–31, 2019. `
[65] PM Blok, K. van Boheemen, FK van Evert, J. Ijsselmuiden, and GH [87] P. Neveu, A. Tireau, N. Hilgert, V. Negre, J. Mineau-Cesari, N. Brichet,
Kim, “Robot navigation in orchards with localization based on particle R. Chapuis, I. Sanchez, C. Pommier, B. Charnomordic, F. Tardieu, and
filter and kalman filter,” Computers and Electronics in Agriculture, vol. L. Cabrera-Bosquet, “Dealing with multi-source and multi-scale
157, p. 261–269, 2019. information in plant phenomics: the ontology-driven phenotyping hybrid
[66] Images from Pixabay, that can be used for free for commercial and information system,” New Phytologist, vol. 221, no. 1, p. 588–601, 2019.
noncommercial use. [Online]. Available:
https://pixabay.com/service/license/ [88] MP Caro, MS Ali, M. Vecchio, and R. Giaffreda, “Blockchain-based
[67] P. Bosilj, T. Duckett, and G. Cielniak, “Analysis of morphology-based traceability in agri-food supply chain management: A practical
features for classification of crop and weeds in precision agriculture,” implementation,” in 2018 IoT Vertical and Topical Summit on Agriculture
IEEE Robotics and Automation Letters, vol. 3, no. 4, p. 2950–2956, Oct - Tuscany (IOT Tuscany), May 2018, pp. 1–4.
2018. [89] K. Salah, N. Nizamuddin, R. Jayaraman, and M. Omar, “Blockchain
´
[68] I. Sa, Z. Chen, M. Popovic, R. Khanna, F. Liebisch, J. Nieto, and R. based soybean traceability in agricultural supply chain,” IEEE Access,
Siegwart, “weednet: Dense semantic weed classification using multispectral vol. 7, p. 73,295–73,305, 2019.
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TII.2020.3003910, IEEE
Machine Translated by Google Transactions on Industrial Informatics
13
1551-3203(c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Macquarie University. Downloaded on June 24,2020 at 06:41:17 UTC from IEEE Xplore. Restrictions apply.