IoT-Enabled Agroecology: Advancing Sustainable Smart Farming

Through Knowledge-Based Reasoning

Nicolas Chollet

1 a

, Naila Bouchemal

1 b

and Amar Ramdane-Cherif

2 c

ECE Paris, 10 rue Sextius Michel, 75015, Paris, France

Universit

e Paris-Saclay, UVSQ, LISV (Laboratoire d’Ing enierie des Syst

emes de Versailles), 78140,

Velizy-Villacoublay, France

Keywords:

Ontology, Knowledge Base, IoT, Agroecology, Smart Farming.

Abstract:

The global increase in population necessitates enhanced food security, yet current agricultural practices are in-

adequate in feeding everyone and are detrimental to the environment. Consequently, agriculture faces the task

of increasing production while minimizing resource usage and prioritizing sustainability. To assist farmers,

new technological tools using AI, Robotics and IoT have been developed in a new ﬁeld called Smart Farming.

Unfortunately, these tools are primarily employed in unsustainable farming practices, such as mono-cropping.

However, sustainable methods like Agroecology exist, which involve observing how plants interact with their

environment to devise crop management strategies that work harmoniously with nature, requiring minimal re-

sources and ensuring sustainability. In this paper, we propose an Internet of Things (IoT) platform that utilizes

an ontology and a set of rules to provide farmers with recommendations for optimizing crop development

while adhering to agroecology principles. This platform employs Knowledge-based reasoning to correlate

crop requirements with local environmental data obtained through a wireless sensor network deployed on the

farm. It can suggest crop layouts, crop calendars, detect relevant events, and manage irrigation. Our sys-

tem has been tested in a simulated environment and yielded promising results, leaving ample room for future

improvements and developments.

1 INTRODUCTION

1.1 Context

By 2050, it is projected that the global population

will grow to approximately 9 billion, according to

estimates. In order to ensure food security world-

wide, the Food and Agriculture Organization (FAO)

of the United Nations suggests that food production

needs to increase by approximately 60% by that time

(Ranganathan et al., 2018). In response to this grow-

ing concern, the agricultural industry is being trans-

formed by Smart Farming (SF) technologies, also

known as Precision Agriculture (PA), with the aim

of enhancing productivity and sustainability. These

SF tools utilize information and communication tech-

nologies (ICT) like Artiﬁcial Intelligence (AI), Inter-

net of Things (IoT) platforms, and Robotics to of-

https://orcid.org/0009-0004-1605-756X

https://orcid.org/0000-0001-8294-9276

https://orcid.org/0000-0001-8289-747X

fer modern and sustainable solutions (Walter et al.,

2017).

IoT plays a crucial role in revolutionizing agricul-

tural practices. IoT is a network of physical objects

or devices that are embedded with sensors, actuators

and connectivity capabilities, allowing them to col-

lect and exchange data with other devices and systems

over the internet. After the perception of data, they

are processed with different algorithms using big data

analytics methods and AI to make a decision about

actions to implement. This interconnected system en-

ables real-time monitoring, control, and automation

of various processes and tasks (Elijah et al., 2018).

One use case example involves the deployment of IoT

sensors in crop ﬁelds. These sensors can collect data

on soil moisture levels, temperature, humidity, and

other environmental parameters. The collected data

is transmitted to a central platform or system via the

Internet. Farmers and agricultural experts can then ac-

cess this data remotely and make informed decisions

about irrigation, fertilization, and pest control. For ex-

ample, an IoT platform can decide to open or not an

190

Chollet, N., Bouchemal, N. and Ramdane-Cherif, A.

IoT-Enabled Agroecology: Advancing Sustainable Smart Farming Through Knowledge-Based Reasoning.

DOI: 10.5220/0012183500003598

In Proceedings of the 15th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2023) - Volume 2: KEOD, pages 190-199

ISBN: 978-989-758-671-2; ISSN: 2184-3228

Figure 1: Simple IoT platform for Smart Farming.

irrigation valve regarding the value of the soil mois-

ture sensor attached to it as depicted in ﬁgure 1. By

leveraging IoT in this manner, farmers can optimize

resource usage, improve crop yields, and reduce en-

vironmental impact through precise and data-driven

farming techniques (Misra et al., 2020). IoT plat-

forms commonly handle more substantial quantities

and diverse arrays of data, encompassing locally ac-

quired data from sensors and remotely acquired data,

such as satellite observations. It is also notewor-

thy to acknowledge the increasing usage of emerging

smart sensors that incorporate AI analysis. These ad-

vancements enable the direct interpretation of com-

plex data, such as sound or images, at the source of

data collection (Chollet et al., 2022).

Regrettably, IoT, similar to numerous Smart

Farming techniques, is predominantly employed to

enhance production using fewer resources while dis-

regarding the ecological consequences of inherently

unsustainable farming practices (Bronson, 2018).

Hopefully, true sustainable farming method exist such

as Agroecology. Agroecology is an agricultural tech-

nique that relies on careful observation of the crop-

growing environment to collaborate with it rather than

oppose it (Altieri, 2018). In particular, it encom-

passes various techniques such as crop rotation to

safeguard against soil nutrient depletion (Ball et al.,

2005), intercropping to encourage beneﬁcial plant in-

teractions and deter pest and insect development (Ge-

bru, 2015), and the utilization of genetically diverse

crop varieties that are well-suited to particular cli-

matic conditions (Hajjar et al., 2008). An actual illus-

tration of Agroecology can be seen in the ancient agri-

cultural practice known as the Three Sisters farming

strategy, which was discovered by Native Americans

Figure 2: Three sisters planting method.

thousands of years ago. Remarkably, this method

continues to be employed by a signiﬁcant number

of rural small-holder farmers in South America to-

day (Lopez-Ridaura et al., 2021). It involves inter-

cropping three main crops: corn (maize), beans, and

squash. These crops are grown together in a mutu-

ally beneﬁcial way. The corn provides a trellis for the

beans to climb, while the beans enrich the soil with

nitrogen, a fertilizer, through nitrogen ﬁxation. The

squash serves as a ground cover, shading the soil and

reducing weed growth while retaining moisture. The

process is depicted in Figure 2. This interdependence

among the three crops creates a sustainable and pro-

ductive farming system. Therefore, the core princi-

ples of agroecology are observation and interpretation

through knowledge of the environment and plant bi-

ology. Those are difﬁcult processes to implement in

real-life scenarios for farmers who wish to transition

to sustainable farming as they require an important

amount of time and workload.

1.2 Proposal

To facilitate the transition of farmers towards agroe-

cology, our research endeavors to deploy an Internet

of Things (IoT) platform to acquire an extensive range

of environmental data from various sources. This en-

compasses the utilization of both conventional and in-

telligent sensors for local data acquisition, as well as

accessing global weather databases for remote data

retrieval. Subsequently, the IoT platform undertakes

the task of interpreting these data in accordance with

agroecological principles, thereby proposing action-

able measures to farmers for effective farm manage-

ment. For example, sensor management (layout and

maintenance), crops spatial distribution, rotation, pro-

visional schedule, irrigation procedure, and event de-

tection (pest development). In this paper, we take

IoT-Enabled Agroecology: Advancing Sustainable Smart Farming Through Knowledge-Based Reasoning

191

Figure 3: Overall view of the IoT platform.

a focus on the decision process. For the system to

make the best decision, it needs to have knowledge

about agroecology, crops, and sensors while manag-

ing a vast amount of data. Knowledge is deﬁned as

the explicit functional associations between informa-

tion and/or data items. To allow computer systems

to understand agroecology, we need to use knowl-

edge engineering, which is the process of develop-

ing knowledge-based systems in a ﬁeld (Kendal and

Creen, 2007). Knowledge Engineering, also known

as Knowledge Representation and Reasoning (KRR),

proposes numerous tools to achieve the desired goals,

such as ontology. Ontologies have demonstrated their

effectiveness in handling organized and structured

data. Consequently, the primary objective of this

study revolves around the establishment of a knowl-

edge base (KB) utilizing an ontology for Agroecology

IoT platforms. The performance of this knowledge

base will be assessed through the examination of var-

ious case-study scenarios. The overall view of the IoT

platform is depicted in ﬁgure 3.

1.3 Structure of the Paper

Section 2 of this paper is dedicated to the deﬁnition

of knowledge engineering and a brief state of the Arts

for Knowledge-base regarding agriculture and IoT.

The third section describes the architecture of our sys-

tem and the ontology-building process. Afterward,

we’ll use our ontology in different use-case scenar-

ios described in section 4. Finally, we will share our

conclusion and future work orientation in section 5.

2 RELATED WORK

2.1 Deﬁnition

Knowledge Engineering refers to the process of de-

signing, creating, and managing knowledge-based

systems. It involves capturing, representing, orga-

nizing, and reasoning with knowledge to develop in-

telligent systems that can perform tasks, make deci-

sions, and solve problems. Knowledge is base on data

and information as depicted in ﬁgure 4 (Kendal and

Creen, 2007).

Figure 4: Data, information, and Knowledge.

Ontology, on the other hand, plays a signiﬁcant

role in Knowledge Engineering as a key component

of knowledge representation. An ontology is a formal

and explicit speciﬁcation of concepts, relationships,

and properties within a particular domain. It serves

as a shared vocabulary or conceptual framework that

enables effective communication and understanding

among humans and computer systems. In Knowledge

Engineering, ontologies provide a structured and stan-

dardized way to represent and organize knowledge.

They deﬁne the entities, their attributes, and the rela-

tionships between them, allowing for clear conceptual

modeling and knowledge representation. By deﬁning

a common vocabulary and formal semantics, ontolo-

gies facilitate knowledge sharing, integration, and in-

teroperability between different systems and domains.

A fully developed and populated ontology, containing

a comprehensive collection of individuals, rules, and

properties, is commonly referred to as a knowledge

base. In technical terms, the knowledge base con-

sists of two components: the Tbox (Terminological

Box) and the Abox (Assertion Box). The Tbox repre-

sents the ontology itself, where information is stored,

while the Abox encompasses the rules and properties

associated with it. Ontologies enable intelligent rea-

soning and inference over the represented knowledge.

They allow for the application of logical rules and au-

tomated reasoning techniques to derive new knowl-

edge or make deductions based on existing knowledge

KEOD 2023 - 15th International Conference on Knowledge Engineering and Ontology Development

192

through the use of a reasoner. A reasoner is a tool that

enables the deduction of logical conclusions based on

a given set of facts, thereby facilitating the classiﬁ-

cation process within an ontology. For instance, if

we deﬁne an instance V as a Car, and the class Car

is a subclass of Vehicle, the reasoner can infer that

V is also a Vehicle. In more intricate scenarios, cer-

tain reasoners can incorporate SWRL rules (Semantic

Web Rule Language) (O’connor et al., 2005). SWRL

is a logic description language that allows the com-

bination of diverse rules to construct more intricate

axioms. The ofﬁcial documentation provides a basic

example to deﬁne the syntax: : hasParent(?x1,?x2)

ˆ:hasBrother(?x2,?x3) then :hasUncle(?x1,?x3) . By

combining the two axioms, namely, :hasParent and

:hasBrother, the hasUncle relationship can be applied

to individuals, thus establishing the individual X1 as

the child of X2 and the nephew of X3. This reasoning

capability helps in solving complex problems, mak-

ing intelligent decisions, and generating logical out-

puts from input information (Staab and Studer, 2010).

2.2 State of the Art

In this section, we will highlight various Knowledge

Bases that mainly utilize ontologies and are speciﬁ-

cally designed for the domains of IoT, agriculture, or

both.

2.2.1 Knowledge Base for IoT

The widespread presence of smart devices in Inter-

net of Things (IoT) applications has led to a signiﬁ-

cant challenge in achieving interoperability due to the

diverse and heterogeneous nature of these ”things.”

The other issue with IoT is the vast amount of data

that have to be fusion together to make an effective

decision. Both of these issues can be solved with

Knowledge Engineering and ontologies. This area

of research has been widely studied over time, and

good surveys have been done by the authors in (Szi-

lagyi and Wira, 2016), (Bajaj et al., 2017), (Graf

et al., 2019).Two notable observations can be drawn

from the literature: Firstly, ontologies developed in

the IoT domain often focus on speciﬁc application

areas, addressing the unique requirements and char-

acteristics of those domains. Secondly, a signiﬁ-

cant portion of these ontologies is built upon the Se-

mantic Sensor Network recommendation proposed by

the World Wide Web Consortium (W3C) (Neuhaus

and Compton, 2009). Brieﬂy, in a traditional sen-

sor network, sensors collect raw data and transmit

it to a central processing unit for analysis. How-

ever, in a semantic sensor network, additional meta-

data and semantic annotations are associated with the

sensor data. This metadata provides contextual infor-

mation about the data, such as the sensor type, lo-

cation, time of measurement, and the observed phe-

nomenon but also maintenance data such as battery

level or ﬁrmware version. By incorporating seman-

tic annotations, SSNs enable enhanced data interpre-

tation, discovery, and integration.

2.2.2 Knowledge Base for Agriculture

Knowledge in Agriculture has been a widely explored

area of research. Their main aim is to model botanical

knowledge about cultivated crops, and environmen-

tal knowledge about various ecosystems and farm-

ing practices. Firstly there is general purposes on-

tology: The Food and Agriculture Organization of

the United Nations (FAO) developed AGROVOC,

which is recognized as the most extensive and com-

prehensive semantic resource in the agricultural do-

main (Caracciolo et al., 2013). AGROVOC serves as

a controlled vocabulary, encompassing 35,000 con-

cepts and 40,000 terms. Its scope extends beyond

agriculture to include domains such as food, nutri-

tion, ﬁsheries, forestry, and the environment, aligning

with the FAO’s wide-ranging purview. AGROVOC

is a multilingual resource, available in 27 languages,

including English, Arabic, and Chinese. Furthermore,

AGROVOC adheres to the Linked Open Data Schema

(LOS), ensuring compatibility with modern data inte-

gration practices. Despite being utilized in numerous

case studies, AGROVOC exhibits certain limitations.

For instance, it lacks the ability to identify speciﬁc

types of fertilizers, diagnose crop diseases, or clas-

sify soil types, which restricts its coverage across var-

ious domains. Furthermore, while AGROVOC serves

as a vocabulary system or thesaurus, it falls short of

being a complete ontology. Other noteworthy gen-

eral purposes agricultural ontologies are the Crop on-

tology (Shrestha et al., 2011) , CIARD Ring (Pesce

et al., 2010), Planteome (Cooper et al., 2018). To

explore this domain further, the authors in (Drury

et al., 2019) have proposed a comprehensive survey.

Apart from general purpose ontology, a large num-

ber of ontologies have been constructed for speciﬁc

domains within the Agricultural scope, like the pota-

toes ontology which focuses on the cultivation of this

vegetable from seeding to distribution (Haverkort and

Top, 2011) or the wheat ontology to model its pheno-

type (N

edellec et al., 2020). Other speciﬁc ontologies

focus on agricultural practices like vertical farming

(Sivamani et al., 2014), or aquaponics (Abbasi et al.,

2022) or even general organic farming in (Abayomi-

Alli et al., 2021). Most of those speciﬁc ontologies,

can be found in a large repository called Agroportal

(Jonquet, 2017).

IoT-Enabled Agroecology: Advancing Sustainable Smart Farming Through Knowledge-Based Reasoning

193

2.2.3 Knowledge Base for Agricultural IoT

Platforms

The objective of knowledge bases in the context of

IoT and agriculture is to establish connections be-

tween information relative to IoT device hardware

and data, and agricultural knowledge necessary for

effective farm management. Two notable studies,

conducted by the authors in (Bhuyan et al., 2022)

and (Ngo et al., 2018), present comprehensive sur-

veys of such ontologies. One prominent ontology

widely used in Agricultural IoT is AgOnt (Hu et al.,

2011), which introduced a foundational model for

this domain. Additionally, the recently developed

OAK system proposes a comprehensive knowledge

base, demonstrating promising results but highlight-

ing scalability limitations (Ngo et al., 2020). Fur-

thermore, numerous use-case-speciﬁc IoT and agri-

cultural knowledge bases exist, catering to speciﬁc

needs such as managing orchid farms (Kaewboonma

et al., 2020) or coffee plantations (Suarez et al., 2022).

2.3 Limitation

Knowledge-base systems are extensively employed in

the ﬁelds of IoT and agriculture primarily for sensor

data management. These systems facilitate the fusion

of diverse data sources and enable the control of fun-

damental systems like irrigation or ventilation. How-

ever, there is a notable scarcity of applications that

combine plant phenotype ontologies with sensor data

to make informed decisions. In the meantime, our

research indicates that only one ontology, proposed

by the authors in (Soulignac et al., 2019), focuses

on agroecology principles. Regrettably, we could not

identify an ontology that speciﬁcally integrates both

agroecology and IoT aspects. Hence the purpose of

this paper.

3 ARCHITECTURE PROPOSAL

3.1 Scenario

A farmer desires to initiate the implementation of

Agroecology in one of his ﬁelds. In order to do so,

he must cultivate crops that align with the ecosys-

tem and climate of the ﬁeld and its surroundings. To

make the best decision about what to plant, where to

plant and when to plant, the farmer interrogate our

Knowledge Base named Permonto, which focuses on

Agroecology. By providing the ﬁeld’s location, di-

mensions, and optionally the desired crop types, the

Figure 5: Permonto platform.

farmer queries the system. Subsequently, the sys-

tem retrieves climate data for the speciﬁed location

over a certain period of time, as well as information

about the soil type in the region on different server

over the internet. Based on these details, the system

suggests an intercropping arrangement that adheres to

Agroecology principles by proposing the best asso-

ciations. Simultaneously, the system recommends a

layout for the hardware, meaning the irrigation valves

and the sensors to gather additional data about the lo-

cal ecosystem. Once the crops are planted and the

necessary hardware is set up, the system manages

irrigation precisely in accordance with the climate

and the speciﬁc requirements of each plant during

their growth. Additionally, intelligent sensors detect

various events and propose corresponding measures.

Leveraging the available data, the system also predicts

the optimal harvest time for each crop type. Finally,

at the conclusion of a farming cycle, the system pro-

poses a new layout for crops and hardware. This pro-

posal takes into account both remote data and the lo-

cally measured data, and previous crops layout, en-

abling the system to continuously improve itself with

each iteration of the growing period.

KEOD 2023 - 15th International Conference on Knowledge Engineering and Ontology Development

194

3.2 Environment Development

Our work revolves solely around software, and the

hardware components are regarded as theoretical in-

stances for now. We used a computer with 16 GB

RAM and a I7 8th generation processor running

Ubuntu 20.04. To construct our ontology, we used

Prot

e 5.5 Software. Prot

e software offers a com-

prehensive and adaptable platform for ontology de-

velopment. Its user-friendly interface, ﬂexibility, ro-

bust features, collaborative capabilities, open-source

availability, and seamless integration make it the ideal

choice for projects of any size or complexity (Musen,

2015). We used the Ontology Web langage (OWL)

with SWRL for our ontology. You can ﬁnd our ontol-

ogy on the Agroportal database. To interact with our

Ontology we developped a simple application using

Python and notably Owlready2 library. It is a module

for ontology-oriented programming in Python 3. For

more comprehensive details about our project and on-

tology, including aspects that may not be explicitly

covered here, you can visit our GitHub page. There,

you will ﬁnd additional sources and information that

provide a deeper understanding of our system.

3.3 Structure of the Ontology

In this part we will describe the Ontology we cre-

ated. It is depicted in ﬁgure 6. Like any ontology,

it is an explicit but partial representation of a targeted

conceptualization. The assumptions made for our on-

tology may therefore be valid in certain contexts and

not in others where the ontology is not intended to

be used. This is why we refer to ontological com-

mitment to indicate the assumptions made in an on-

tology and the implicit adherence to these assump-

tions by the users of that ontology (Kendal and Creen,

2007). Our goal in this endeavor is to develop an

ontology that enables the inference of relationships

between plants based on agroecology principles and

data gathered from sensors. This ontology is centered

around three primary components: the farm and its

constituent growing ﬁelds, the plants, and the IoT de-

vices.

3.3.1 Farm

The class farm regroup the farmers, the ﬁeld, and

the agricultural procedure. Farmers are regarded as

the key agents responsible for executing an action ef-

fectively. The agricultural procedure class describes

the fundamental actions that farmers have at their

disposal, such as planting, harvesting, and apply-

ing countermeasures in response to detected events.

The ﬁeld class ultimately symbolizes an individual

Figure 6: Global representation of Permonto and its main

classes.

plot of land within a farm. A ﬁeld possesses a location

indicated by GPS coordinates and is associated with

various environmental data. This includes local data

obtained from sensors, such as temperature, humid-

ity, wind speed, rain level, soil type, sun exposure,

ground temperature, and ground humidity. Addition-

ally, remote data pertaining to the same parameters

is retrieved from a global weather database as long

as predicted weather data. Furthermore, the ﬁeld en-

compasses the crops and IoT devices described in the

remaining sections of the ontology.

3.3.2 Plants

Within the plant class, there are four subdivi-

sions: vegetable, aromatic, ﬂower, and Agrocecol-

ogy interaction. The ﬂowers, aromatics, and vegeta-

bles subclasses provide detailed information about the

biological traits of plants. This includes their respec-

tive families, water requirements, temperature range

(minimum and maximum), preferred soil types, and

other signiﬁcant characteristics. These characteris-

tics, provided as examples, have already been incor-

porated into numerous ontologies related to plant bi-

ology, such as AGROVOC, and are readily accessi-

ble for reference. Our primary emphasis was on the

Agroecology interaction class, where we aimed to

model the advantages and disadvantages of each plant

in relation to various ecosystem parameters. This was

done to enable our system to identify the most fa-

vorable plant interactions. One speciﬁc interaction

we focused on was the ”three sisters” associations, as

mentioned in the paper’s introduction. By assigning

properties of need and provide to the plants, we can

depict their positive inﬂuence on the surrounding en-

vironmental factors. By implementing this approach

across a wide range of plant species, our system can

independently uncover valuable plant interactions on

a large scale. The interaction of three sisters is mod-

eled in Figure 7.

Another instance of interaction we modeled was

the pest protection mechanism. For instance, tomato

IoT-Enabled Agroecology: Advancing Sustainable Smart Farming Through Knowledge-Based Reasoning

195

Figure 7: Three sisters modelisation in the ontology.

plants are highly susceptible to aphids, which are

small insects that feed on their leaves. However,

marigold ﬂowers act as a natural repellent for aphids.

By representing these facts through properties such

as menaced by and protect from on the Aphids class,

as depicted in Figure 8, our system can deduce that

planting tomatoes alongside marigold would offer

protection against aphids.

Figure 8: Aphids relations.

3.3.3 IoT Devices

The ﬁnal major class in our ontology pertains to

IoT device. We have incorporated four device types:

ground sensor, weather station, smart sensor, and ir-

rigation valve. Each device is associated with a spe-

ciﬁc location within the farm and operates on battery

power. The weather stations provide information on

air temperature, air humidity, wind speed, and overall

farm rainfall levels. The ground sensors measure sun

exposure, soil temperature, and ground temperature

within a 5-square-meter radius. Each ground sensor

is linked to speciﬁc crops based on its location in the

ﬁeld. The smart sensors observe crops within a desig-

nated 10m radius and can identify the development

of pests (such as aphids or slugs) or diseases (like

mildew or Botrytis). The smart sensor also has two

extra parameters, the ﬁrmware version and the accu-

racy of the inferences. Indeed, as explained by the au-

thors in (Chollet, 2022), Smart sensors need ﬁrmware

over-the-air update when their accuracy drops below

a certain point. The irrigation valve also has an effec-

tive 20m radius and is similarly connected to speciﬁc

crops.

3.4 Data Fusion

Once our ontology is populates with individuals sen-

sors plants and ﬁeld, the value from the sensor will be

stored in the different classes. Doing so we perform

Data fusion. It involves combining data from multiple

sources to create a comprehensive and accurate repre-

sentation of information. By integrating diverse data,

such as sensor readings or database information, data

fusion improves accuracy, completeness, situational

awareness, and robustness. It enables the discovery

of hidden patterns and correlations, leading to better

decision-making (Khaleghi et al., 2016).

3.5 Rules Implementation

Once the Agroecology model, which qualiﬁes the

data, has been incorporated into the ontology, it be-

comes necessary to establish rules to standardize the

ontology. These rules serve as the basis for automat-

ing the farming procedure. An example of a rule we

implemented pertains to the irrigation process. In

this rule, plants and sensors are linked to an irriga-

tion valve based on their location. If a sensor de-

tects that the soil’s water level is below the required

amount for a speciﬁc plant, the valve will open until

the sensor detects the appropriate water level. Ad-

ditionally, an extra layer of reasoning is included to

consider the rain prediction parameter sourced from a

weather database for the next 24 hours. If the predic-

tion indicates a high likelihood of rain, the irrigation

valve will not be opened. Another example of rule we

implemented was the ones regarding the smart sen-

sors. When a pest development is detected, on one

plant, the farmer is requested to apply a counter mea-

sure on it. On the hardware side, if the battery of a

sensors drops below a certain threshold, farmer is re-

quested to replace it. The rules we have written do not

rely on exact values but rather on fuzzy values, which

are based on the principles of fuzzy logic (Chen et al.,

2001). This means that parameter values are qualiﬁed

using fuzzy notions such as ”strong,” ”very strong,” ”a

little bit,” ”little,” etc. These fuzzy values are utilized

to infer knowledge from the written rules, allowing

for a more nuanced and ﬂexible approach to reason-

ing.

KEOD 2023 - 15th International Conference on Knowledge Engineering and Ontology Development

196

4 USE-CASE

EXPERIMENTATION

4.1 Simulator Building

To validate our model, we created a straightforward

command-line interface (CLI) tool. This tool allows

us to input data and obtain query results regarding var-

ious procedures. By utilizing this CLI tool, we can as-

sess the effectiveness of our model based on the gen-

erated outputs.

4.2 Irrigation Management

For this experiment, we selected two ﬁelds: one

planted with tomatoes and the other with potatoes.

Each ﬁeld is equipped with a single irrigation valve

responsible for managing the water supply to all the

plants within it. Additionally, there are four ground

sensors in each ﬁeld, strategically placed in quarters.

The layout can be visualized in the provided ﬁgure9.

Figure 9: Irrigation procedure.

In the ﬁrst scenario, we observed that the mea-

sured ground soil moisture level on all the sensors

was low, and the weather prediction for the next 24

hours indicated very low precipitation. As a result,

we noticed that only the irrigation valve in the toma-

toes ﬁeld opened. This behavior is because toma-

toes require water when the available moisture level is

low, while potatoes necessitate water when the level

is classiﬁed as very low.

In a different conﬁguration, we adjusted half of

the ground sensors in the potato ﬁeld to a very low

moisture level, while the other half remained at a low.

As expected, the irrigation valve only opened for the

corresponding half of the potato ﬁeld.

Lastly, we set all the ground sensors to a very low

moisture level, but the ”rain prediction” parameter

was set to very high. As expected, no irrigation valve

opens as they wait for raining.

4.3 Pest Development

In this scenario, we deployed a smart sensor camera

to monitor a lettuce ﬁeld and another one positioned

in front of a tomato ﬁeld. During the initial test, we

simulated a pest detection event speciﬁcally for slugs

in the lettuce ﬁeld. Our system’s response consisted

of two actions: ﬁrst, it prompted the farmer to remove

the slugs and apply organic slug poison, and secondly,

it recommended planting Lavender around both the

lettuce and tomato ﬁelds. This recommendation was

based on the high susceptibility of both crops to slug

infestations, and Lavender’s natural repellent proper-

ties against these insects. This process is described in

ﬁgure 10.

Figure 10: Pest development.

4.4 Firmware Update over the Air for

Smart Sensors

We also performed a test to evaluate the accuracy of

the smart sensors. In this test, we intentionally set the

”accuracy” parameter of a smart sensor to an ”insuf-

ﬁcient” level. As a result, the system automatically

initiated a Firmware Update Over The Air (FUOTA)

procedure, as previously explained.

4.5 Calendar Estimation

To assess the capability of our system in determining

the optimal planting periods for different crops based

on farm location, we conducted simulated tests in two

cities: Marseille in the south of France and Lille in

the north. We queried the system to determine the

ideal timing for planting tomatoes in each location.

The system retrieved temperature data from previous

years for both cities and compared them against the

minimum temperature requirement for tomato plants.

Based on these parameters, the system advised initi-

ating tomato planting from early April in Marseilles

while recommending waiting until the end of May in

Lille.

IoT-Enabled Agroecology: Advancing Sustainable Smart Farming Through Knowledge-Based Reasoning

197

4.6 Crops Layout

In our ﬁnal scenario, we aimed to demonstrate the

system’s ability to identify known beneﬁcial plant in-

teractions based on the provided information. We re-

quested the system to propose a crop layout for a ﬁeld

measuring 2m by 8m, specifying our desire to plant

corn, squash, peas, tomatoes, and basil. Utilizing the

range parameter of each plant, which represents the

space it requires on the ground, the system gener-

ated a layout map. It strategically positioned squash,

corn, and peas together to foster beneﬁcial interac-

tions among them. Simultaneously, the system sepa-

rated the tomato ﬁeld from the corn to avoid potential

competition. Instead, it recommended planting basil

alongside the tomatoes to provide shade for the soil.

By considering these factors and employing the range

parameter, the system successfully proposed a crop

arrangement that optimizes beneﬁcial plant interac-

tions within the given ﬁeld. Each plant was given by

coordinates, we represented the layout in ﬁgure 11.

Figure 11: Crops Layout.

5 CONCLUSIONS

The proposed knowledge base demonstrates promis-

ing outcomes, enabling farmers to effectively man-

age essential stages of the farming process. These

include crop layout, care tasks (such as irrigation

and pest detection), and optimal harvest timing, all

in concordance with agroecology principles. Addi-

tionally, the knowledge base facilitates the seamless

integration of IoT devices, allowing farmers to har-

ness the generated data effortlessly. Moreover, the

model possesses the capability to identify advanta-

geous crop interactions based on agroecology prin-

ciples. To advance this research, it is crucial to in-

corporate further knowledge from farmer experiences

into the ontology. Furthermore, aligning our ontol-

ogy with existing plant-related knowledge sources

like AGROVOC would enable the discovery of novel

beneﬁcial relationships between plants, speciﬁc to

their respective environmental contexts. In conclu-

sion, this knowledge-based system holds great poten-

tial for farmers seeking to transition towards sustain-

able farming practices with the aid of IoT technolo-

gies.

REFERENCES

Abayomi-Alli, A. A., Misra, S., Akala, M. O., Ikotun,

A. M., Ojokoh, B. A., et al. (2021). An ontology-

based information extraction system for organic farm-

ing. International Journal on Semantic Web and In-

formation Systems (IJSWIS), 17(2):79–99.

Abbasi, R., Martinez, P., and Ahmad, R. (2022). An on-

tology model to represent aquaponics 4.0 system’s

knowledge. Information Processing in Agriculture,

9(4):514–532.

Altieri, M. A. (2018). Agroecology: the science of sustain-

able agriculture. CRC Press.

Bajaj, G., Agarwal, R., Singh, P., Georgantas, N., and Is-

sarny, V. (2017). A study of existing ontologies in the

iot-domain. arXiv preprint arXiv:1707.00112.

Ball, B., Bingham, I., Rees, R., Watson, C., and Litterick,

A. (2005). The role of crop rotations in determining

soil structure and crop growth conditions. Canadian

Journal of Soil Science, 85(5):557–577.

Bhuyan, B. P., Tomar, R., and Cherif, A. R. (2022). A

systematic review of knowledge representation tech-

niques in smart agriculture (urban). Sustainability,

14(22):15249.

Bronson, K. (2018). Smart farming: including rights hold-

ers for responsible agricultural innovation. Technol-

ogy Innovation Management Review, 8(2):7–14.

Caracciolo, C., Stellato, A., Morshed, A., Johannsen, G.,

Rajbhandari, S., Jaques, Y., and Keizer, J. (2013). The

agrovoc linked dataset. Semantic Web, 4(3):341–348.

Chen, G., Pham, T. T., and Boustany, N. (2001). Introduc-

tion to fuzzy sets, fuzzy logic, and fuzzy control sys-

tems. Applied Mechanics Reviews, 54(6):B102–B103.

Chollet, Naila Bouchemal, A. R.-C. (2022). Energy efﬁ-

cient ﬁrmware over the air update for tinyml models in

lorawan agricultural networks. In 2022 32nd Interna-

tional Telecommunication Networks and Applications

Conference (ITNAC), pages 21–27. IEEE.

Chollet, C., Naila, B., and Amar, R.-C. (2022). Tinyml

smart sensor for energy saving in internet of things

precision agriculture platform. In 2022 Thirteenth

International Conference on Ubiquitous and Future

Networks (ICUFN), pages 256–259. IEEE.

Cooper, L., Meier, A., Laporte, M.-A., Elser, J. L., Mungall,

C., Sinn, B. T., Cavaliere, D., Carbon, S., Dunn, N. A.,

Smith, B., et al. (2018). The planteome database:

an integrated resource for reference ontologies, plant

genomics and phenomics. Nucleic acids research,

46(D1):D1168–D1180.

Drury, B., Fernandes, R., Moura, M.-F., and de An-

drade Lopes, A. (2019). A survey of semantic web

technology for agriculture. Information Processing in

Agriculture, 6(4):487–501.

Elijah, O., Rahman, T. A., Orikumhi, I., Leow, C. Y., and

Hindia, M. N. (2018). An overview of internet of

things (iot) and data analytics in agriculture: Bene-

ﬁts and challenges. IEEE Internet of things Journal,

5(5):3758–3773.

Gebru, H. (2015). A review on the comparative advantages

KEOD 2023 - 15th International Conference on Knowledge Engineering and Ontology Development

198

of intercropping to mono-cropping system. Journal of

Biology, Agriculture and Healthcare, 5(9):1–13.

Graf, D., Kapsammer, E., Retschitzegger, W., Schwinger,

W., and Baumgartner, N. (2019). Cutting a path

through the iot ontology jungle-a meta-survey. In

2019 IEEE International Conference on Internet of

Things and Intelligence System (IoTaIS), pages 1–7.

IEEE.

Hajjar, R., Jarvis, D. I., and Gemmill-Herren, B. (2008).

The utility of crop genetic diversity in maintaining

ecosystem services. Agriculture, Ecosystems & En-

vironment, 123(4):261–270.

Haverkort, A. and Top, J. (2011). The potato ontology:

delimitation of the domain, modelling concepts, and

prospects of performance. Potato Research, 54:119–

136.

Hu, S., Wang, H., She, C., and Wang, J. (2011). Agont:

ontology for agriculture internet of things. In Com-

puter and Computing Technologies in Agriculture IV:

4th IFIP TC 12 Conference, CCTA 2010, Nanchang,

China, October 22-25, 2010, Selected Papers, Part I

4, pages 131–137. Springer.

Jonquet, C. (2017). Agroportal: an ontology repository for

agronomy. In EFITA WCCA CONGRESS.

Kaewboonma, N., Chansanam, W., and Buranarach, M.

(2020). Ontology-based big data analysis for orchid

smart farming. LIBRES: Library & In-formation Sci-

ence Research Electronic Journal, 29:292–296.

Kendal, S. L. and Creen, M. (2007). An introduction to

knowledge engineering. Springer.

Khaleghi, B., Khamis, A., Karray, F. O., and Razavi, S. N.

(2016). Multisensor data fusion: A review of the state-

of-the-art. Information fusion, 14(1):28–44.

Lopez-Ridaura, S., Barba-Escoto, L., Reyna-Ramirez,

C. A., Sum, C., Palacios-Rojas, N., and Gerard, B.

(2021). Maize intercropping in the milpa system. di-

versity, extent and importance for nutritional security

in the western highlands of guatemala. Scientiﬁc Re-

ports, 11(1):1–10.

Misra, N., Dixit, Y., Al-Mallahi, A., Bhullar, M. S., Upad-

hyay, R., and Martynenko, A. (2020). Iot, big data,

and artiﬁcial intelligence in agriculture and food in-

dustry. IEEE Internet of things Journal, 9(9):6305–

6324.

Musen, M. A. (2015). The prot

e project: a look back and

a look forward. AI matters, 1(4):4–12.

edellec, C., Ibanescu, L., Bossy, R., and Sourdille, P.

(2020). Wto, an ontology for wheat traits and phe-

notypes in scientiﬁc publications. Genomics & Infor-

matics, 18(2).

Neuhaus, H. and Compton, M. (2009). The semantic sen-

sor network ontology. In AGILE workshop on chal-

lenges in geospatial data harmonisation, Hannover,

Germany, pages 1–33.

Ngo, Q. H., Kechadi, T., and Le-Khac, N.-A. (2020).

Oak: Ontology-based knowledge map model for digi-

tal agriculture. In Future Data and Security Engineer-

ing: 7th International Conference, FDSE 2020, Quy

Nhon, Vietnam, November 25–27, 2020, Proceedings

7, pages 245–259. Springer.

Ngo, Q. H., Le-Khac, N.-A., and Kechadi, T. (2018). On-

tology based approach for precision agriculture. In

Multi-disciplinary Trends in Artiﬁcial Intelligence:

12th International Conference, MIWAI 2018, Hanoi,

Vietnam, November 18–20, 2018, Proceedings 12,

pages 175–186. Springer.

O’connor, M., Knublauch, H., Tu, S., Grosof, B., Dean,

M., Grosso, W., and Musen, M. (2005). Supporting

rule system interoperability on the semantic web with

swrl. In The Semantic Web–ISWC 2005: 4th Inter-

national Semantic Web Conference, ISWC 2005, Gal-

way, Ireland, November 6-10, 2005. Proceedings 4,

pages 974–986. Springer.

Pesce, V., Maru, A., and Keizer, J. (2010). The ciard ring,

an infrastructure for interoperability of agricultural re-

search information services.

Ranganathan, J., Waite, R., Searchinger, T., and Hanson, C.

(2018). How to sustainably feed 10 billion people by

2050, in 21 charts.

Shrestha, R., Davenport, G. F., Bruskiewich, R., and Ar-

naud, E. (2011). Development of crop ontology for

sharing crop phenotypic information. drought pheno-

typing in crops: from theory to practice. part i: Plant

phenotyping methodology. Technical report.

Sivamani, S., Bae, N.-J., Shin, C.-S., Park, J.-W., and Cho,

Y.-Y. (2014). An owl-based ontology model for intel-

ligent service in vertical farm. In Advances in Com-

puter Science and its Applications: CSA 2013, pages

327–332. Springer.

Soulignac, V., Pinet, F., Lambert, E., Guichard, L., Trouche,

L., and Aubin, S. (2019). Geco, the french web-based

application for knowledge management in agroecol-

ogy. Computers and Electronics in Agriculture,

162:1050–1056.

Staab, S. and Studer, R. (2010). Handbook on ontologies.

Springer Science & Business Media.

Suarez, C., Griol, D., Figueroa, C., Corrales, J. C., and Cor-

rales, D. C. (2022). Rustont: An ontology to explain

weather favorable conditions of the coffee rust. Sen-

sors, 22(24):9598.

Szilagyi, I. and Wira, P. (2016). Ontologies and seman-

tic web for the internet of things-a survey. In IECON

2016-42nd Annual Conference of the IEEE Industrial

Electronics Society, pages 6949–6954. IEEE.

Walter, A., Finger, R., Huber, R., and Buchmann, N. (2017).

Smart farming is key to developing sustainable agri-

culture. Proceedings of the National Academy of Sci-

ences, 114(24):6148–6150.

IoT-Enabled Agroecology: Advancing Sustainable Smart Farming Through Knowledge-Based Reasoning

199