Smart Agricultural Technology 8 (2024) 100446

Contents lists available at ScienceDirect

Smart Agricultural Technology

journal homepage: www.journals.elsevier.com/smart-agricultural-technology

Application of Fuzzy logic and IoT in a small-scale Smart

Greenhouse System
Vasileios Thomopoulos a, Fotios Tolis b, Taxiarchis-Foivos Blounas b, Dimitrios Tsipianitis b,
Angeliki Kavga c, *
a
Department of Computer Engineering and Informatics, University of Patras Campus, Rion EL26504, Greece
b
Department of Electrical and Computer Engineering, University of Patras Campus, Rion EL26504, Greece
c
University of Patras, Department of Agriculture, University of Patras Campus, Rion EL26504, Greece

A R T I C L E I N F O A B S T R A C T

Keywords: This research aims to address the critical challenges caused by global population increase, climate change, and
Greenhouse technology urbanization on food production by means of a Fuzzy Logic controller in a small-scale greenhouse and improving
Fuzzy control agricultural efficiency and sustainability.
Pi controller
The growing global population has stressed the significance of implementing efficient food production tech­
Actuator systems
niques. Furthermore, the impact of climate change and extreme weather events complicates the production
Autonomous plant cultivation
Sustainability process. The goal of this study is to combine an innovative small greenhouse design with a fuzzy logic-based
control system in order to develop an IoT-enabled self-regulating environment for autonomous plant growing.
Continuous advances in technology have been focusing on automating greenhouses to provide self-regulating
microclimates and optimal developmental conditions. Current trends focus on automating greenhouses to pro­
vide self-regulating microclimates. Looking ahead, the strategy entails combining automated approaches with
precisely planned structures via IoT.
This study’s key findings include the successful integration of a fuzzy logic control system with greenhouse
infrastructure, allowing for real-time monitoring and adjustment of environmental parameters, as well as the
development of a pilot IoT ecosystem demonstrating the feasibility of self-regulating microclimates for improved
plant growth.

1. Introduction refining the design of greenhouse structures, materials, and control

methods [3,4].
Recently, there has been an extensive pursuit of efficient food pro­ Many environmental factors significantly influence the production
duction methods due to the global population surge [1]. The agricultural process, such as changes in temperature, variations in light intensity,
sector deals with unprecedented challenges that consist of several phe­ levels of relative humidity, soil moisture content, and the presence of
nomena including climate change, extreme weather events, and carbon dioxide (CO2), among other factors [5]. More traditional ap­
encroachment of urbanization on arable lands. These challenges not proaches to greenhouse design have focused on monitoring these factors
only impede the production process but also sever the connection be­ via sensors which are made to detect their states and via actuators which
tween modern humans and crop cultivation, endangering are made to drive them toward desired conditions [6]. Currently, the
self-sustainability, especially during periods of production shortages Internet of Things (IoT) has revolutionized this paradigm, by intro­
[2]. ducing a new type of scheme that enables remote monitoring of various
To give prominence to these multidisciplinary challenges, emphasis microclimate parameters [7].
has started to grow mainly on the development of modern agricultural The works of Khan et al. [19], Hernández-Morales et al. [16], Huynh
technology which aims to create optimal growth conditions and effec­ et al. [17], Chakraborty et al. [11], Kumar et al. [21], Olalla et al. [23],
tive production methods, with particular attention directed towards Bersani et al. [10], and Elashiri and Shawky [13] establish important
modern greenhouse technology. Development of this technology entails strides in this domain. Khan et al. [19], for instance, concentrate on

* Corresponding author.
E-mail address: [email protected] (A. Kavga).

https://doi.org/10.1016/j.atech.2024.100446
Received 8 February 2024; Received in revised form 26 March 2024; Accepted 30 March 2024
Available online 31 March 2024
2772-3755/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Table 1 Table 3
Sensor Information. Inference Rules.
Sensor Type Sensor Description Additional Information e \ Δe NL NS ZE PS PL

Air Temperature, Sensor: OEM The sensor offers dual NL NL NL NL NS ZE

Relative DHT22 measurement capabilities, compact NS NL NL NS ZE PS
Humidity dimensions, low power ZE NL NS ZE PS PL
consumption, and an extensive PS NS ZE PS PL PL
range of transmission. PL ZE PS PL PL PL
Soil Moisture Sensor: Waveshare The sensor functions by detecting
Soil Moisture Sensor soil moisture levels using exposed
pads that act as variable resistors. aimed to address the various challenges that global population growth
As the soil moisture increases, the and climate change posed. Huynh et al. [17] made a significant contri­
resistance decreases. Additionally,
the sensor measures environmental
bution to the smart greenhouse and IoT research by successfully
luminosity, with its resistance designing, constructing, and testing a 30 m2 smart greenhouse tailored
decreasing as the light intensity for cultivating Brassica Juncea. With their research, they also integrated
increases. Multiple Light diverse IoT technologies, a web interface for real-time monitoring and
Dependent Resistors (LDRs) are
control, as well as a mobile application for on-the-go access (Figs. 3, 4, 6,
strategically placed throughout the
greenhouse, and a conversion rule 7, 11, 12, 19-25 and Tables 1-3).
is applied to translate voltage Furthermore, Chakraborty et al. [11] proposed to solve the agricul­
readings into physical luminosity tural challenges by integrating IoT into the greenhouse environment for
values. monitoring and by using a smart irrigation system. They also utilized the
Luminosity OEM 5 mm Light The sensor measures
Dependent Resistor environmental luminosity, with its
BLYNK application to monitor real-time data. Previously, Kumar et al.
(LDR) resistance decreasing as light [21] devised an IoT-based system to monitor and manage a greenhouse
intensity increases. Multiple Light powered by solar energy. They utilized diverse methods including GSM
Dependent Resistors (LDRs) are technology, Arduino, and various sensors to transmit real-time data to
strategically positioned within the
mobile devices. Olalla et al. [23] applied fuzzy control methods to hy­
greenhouse, and a conversion rule
is applied to translate voltage droponic strawberry crops in greenhouses, demonstrating improved
readings into physical luminosity crop quality and resource optimization. Bersani et al. [10] delved into
values. IoT methodologies for monitoring and controlling smart greenhouses
within the framework of Industry 4.0. They emphasize real-time data
collection and control facilitated by information and communication
Table 2 infrastructure. Finally, Elashiri and Shawky [13] proposed a fuzzy
Actuator Information. computational algorithm for IoT applications in agriculture, that was
Actuator Type Actuator Description Additional Information specially made for crop tracking in smart greenhouses, exploiting sen­
sors, actuators, and microcontrollers for decision-making.
Temperature Corsair 120 mm 12 V DC fan Mounted for cooling, a servo-
controlled window on the
Hence, this research seeks to propel smart greenhouse technology
opposite side provides forward by integrating a distinctive small greenhouse design with a
adaptable air circulation. fuzzy logic-based control system. This fusion is aimed at creating an IoT-
Air Humidity Combined operation of fan and Regulates air humidity. enabled, self-regulating environment conducive to autonomous plant
servo-controlled window
cultivation, thereby improving accessibility and simplifying agricultural
Soil Moisture OEM Peristaltic Liquid Pump Utilized for irrigation
with Silicone TubingUtilized purposes, the peristaltic liquid operations.
for irrigation purposes, the pump with silicone tubing
peristaltic liquid pump with delivers water from a 2. State of the Art
silicone tubing delivers water designated tank or alternative
from a designated tank or water source.
alternative water source.
Recent research has focused on regulating microclimate parameters
Luminosity OEM cool white LED Strips Serve as artificial light sources like temperature, relative humidity, soil moisture, and luminosity. This
to supplement sunlight. These is achieved through the deployment of cooling fans, heating or lumi­
strips are arranged on the top nosity lamps, and LED tapes tailored to meet specific luminosity re­
of the structure.
quirements (da [12]).In another study, it was used a core piece of an
Side Window Waveshare MG90S servo motor Mounted onto the plexiglass
Mode structure through a custom electric valve to compare the irrigation with a water-filled tank [30],
3D-printed design, the MG90S while Ullah et al., [29] utilized a water pump and sprayers for irrigation
servo motor enables the purposes to elevate humidity levels. The implementation of a
creation of an adaptable servo-controlled window, designed to remain open during rainfall, was
airway for regulating
temperature and humidity
investigated to utilize rainwater effectively [26]. Additionally, CO2
levels. By responding to PWM levels were considered, and two fans were employed to ventilate the
signals, the servo adjusts the greenhouse [8]. Recent studies have emphasized the significance of
window opening to achieve energy efficiency by incorporating Solar PV Panels and a Rechargeable
the desired effect.
Battery to achieve system energy autonomy [24]. Moreover, the benefits
of smart environments were validated through growth comparisons with
integrating automation into a small-scale smart greenhouse using IoT, outdoor conditions [20].
with a particular focus on monitoring environmental conditions, man­ Numerous designs utilize microcontrollers such as Arduino, mainly
aging water irrigation, and predicting plant diseases. Addressing for tasks like data processing and network connectivity. IoT imple­
broader challenges, Hernández-Morales et al. [16] decided to design an mentation frequently leverages established platforms like Blynk [30] or
IoT-based monitoring system with the ability to incorporate greenhouse Arduino Cloud. Nevertheless, some researchers still choose to develop
micro-climate monitoring, with low-power wide area network (LPWAN) their web applications [27]. While basic on-off control algorithms are
technology, and with a predictive model for efficient control. They common in many studies, more advanced methodologies like the fuzzy

2
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 1. Various greenhouse roof shapes: (a) Gable roof, (b) Arch roof, (c) Sawtooth roof, (d) Trussed roof, (e) Gable frame (even span), (f) Gothic house, (g) Hoop
house (Quonset), (h) Lean-to house [32].

Fig. 2. The developed greenhouse structure designed in Autodesk Fusion 360. It consists of a main compartment for crop cultivation with dimensions 500 × 350 ×
500 mm, and a secondary one for electronic components storage with dimensions 200 × 350 × 200 mm.

PID controller [15] and fuzzy controllers [27] have shown promising (Quonset), and lean-to house, are depicted in Fig. 1 [32].
outcomes. Additionally, larger-scale applications of fuzzy logic control For this project, designing an adaptable structure that is suitable for
have emphasized the link between energy optimization and indoor urban crop cultivation is this research’s main concern given that it will
temperature humidity [9]. Other control techniques include Neural address various concerns mainly related to weather fluctuations while at
Networks (NNs) [33], Machine Learning (ML), Particle Swarm Optimi­ the same time, it will maximize the exposure to solar radiation. Given
zation (PSO) [34], and Model Predictive Control (MPC). The application the greenhouse’s proximity to urban areas, it is intended to leverage the
of ML techniques, such as utilizing the VGG-trained model for early heat emitted by nearby walls to improve the microclimate. Furthermore,
plant disease detection [14], has significantly enhanced control in addressing solar radiation and internal air temperature concerns, this
strategies. study departed from traditional designs and opted to implement an
The subsequent sections of this paper are organized as follows: uneven-span structure. Given these considerations, the lean-to-house
Section 3 outlines the greenhouse design process, providing detailed form emerges as a fitting option. An indicative 3D design of the devel­
insights into the sensors, actuators, and Printed Circuit Board (PCB). oped structure is depicted in Fig. 2.
Section 4 delves into the fundamentals of fuzzy control scheme and
fuzzy logic. Section 5 explores the implementation of IoT technology. 3.2. Material selection
Finally, Section 6 presents the findings of the study, while Section 7
offers concluding remarks. Choosing the right material for the greenhouse is essential as it
directly influences the quality of light inside the microclimate. Typi­
3. The greenhouse structure cally, materials commonly used include a mix of glass and robust
polymers such as polycarbonate (plexiglass) [25,31]. To prioritize user
3.1. Shape selection safety and provide better protection against ultraviolet light exposure
([18]; La [22]), plexiglass was decided to be utilized, which has a
The design of the greenhouse is crucial for enhancing production thickness of 0.5 mm.
efficiency as it ensures the stability of the microclimate and shields crops
from external elements like rain, wind, and debris. A prevalent design
choice is the straight sidewall and gable roof configuration, as outlined 3.3. Sensors
by Yildiz [32]. Various greenhouse shapes, such as gable, arch,
sawtooth, trussed, even-span gable frame, gothic house, hoop house The monitoring of the microclimate is done by sensors trained to
track system parameters. Specifically, each sensor monitors a distinct

3
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 3. The sensors utilized in the greenhouse system: (a) DHT22 Air Temperature and Relative Humidity Sensor, (b) Waveshare Moisture Sensor, (c) 5 mm OEM
Light Dependent Resistor (LDR).

Fig. 4. 3D Sketch of the structure and the positions of the sensors.

Fig. 5. Sensors indicative circuit diagram.

separate physical property such as luminosity or temperature, etc. These 3.3.1. Temperature and air humidity
sensors provide real-time data on the state of each parameter, facili­ Sensor DHT 22 is known for its dual measurement capabilities. It is
tating the implementation of the subsequent closed-loop control system. small, with low power consumption, and a long transmission range.
To ensure the proper function of the electrical components a 12 V DC Being dual consists of two sensors, one sensor is placed inside the
power source was used. The power source could be changed with a 12 V structure, while the other is placed outside. This setup enables the
battery in order to increase flexibility. monitoring of both internal and external temperatures and relative

4
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 6. Actuators: (a) Corsair 12 V DC Fan, (b) OEM 12 V Liquid Pump, (c) Waveshare MG90S Servo, (d) OEM LED Strip.

humidities, as well as the measurement of their difference. This fact can These readings are then converted into physical luminosity values using
be useful to avoid over-actuation, making the system more power a predefined conversion rule applied to voltage measurements. To
efficient. obtain a single measurement of luminosity, a weighted average of
multiple readings is calculated.
3.3.2. Soil moisture
The Waveshare sensor is used to detect soil moisture by exploiting 3.4. Actuators
the pads that have been exposed and function as variable resistors. The
resistance of the sensor is by analogy opposite to the proportion of the Actuators are used to highlight deviations from desired system pa­
soil moisture level. This sensor is directly inserted into the soil to extract rameters. This project operates under the assumption that the effect of
accurate measurements. each actuator is locally bound with the designated parameter of the
microclimate, like the sensors mentioned above.
3.3.3. Luminosity
A 5 mm Light Dependent Resistor (LDR) is employed to gage envi­ 3.4.1. Temperature
ronmental luminosity. The LDR’s resistance varies inversely with light A Corsair 120 mm 12 V DC fan is mounted for cooling, and a servo-
intensity, serving as a measure thereof. The LDR resistance has a low controlled window on the opposite side provides adaptable air circula­
value when the environment is light and a high value when the envi­ tion. Considering that no heating or cooling elements are utilized in the
ronment is dark. Since the sensor is read through measuring the voltage, implementation, the temperature inside the greenhouse can only
simply connecting the LDR to a voltage source would provide low values decrease to the level of the ambient air temperature. This characteristic
in light and high values in darkness, due to Ohm’s law. This is coun­ results in relatively low power requirements for the greenhouse, a
terintuitive as the luminosity measurement should be higher the lighter realistic scenario consistent with methods employed in large-scale
an environment is. To solve this issue the LDR is connected in a voltage greenhouses [28].
divider, as shown in Fig. 5. The measured voltage is given by the
following equation: 3.4.2. Air humidity
Humidity regulation within the greenhouse is achieved through the
R
Vout = Vin coordinated operation of the fan and the servo-controlled window.
R + RLDR
While the fan manages airflow, the servo-controlled window regulates
where R is a resistor value chosen at 10 kΩ since it provided the best ventilation, ensuring optimal humidity levels.
results. The measured voltage is close to the input voltage when the
environment is light and close to zero when it is dark. So by normalizing 3.4.3. Soil moisture
the measurements to a range of [0, 1] an estimation of a fuzzy boolean To facilitate irrigation, an OEM Peristaltic Liquid Pump equipped
measurement of its luminosity. Its slope depends on the value of the with Silicone Tubing is utilized. Despite its low-pressure flow, this pump
resistance R. It is obvious that such a measurement is not measured in a effectively delivers water from a tank or alternative source. Its suitability
formal quantity, such as Lux or Solar Irradiance, since it is considered to for this application lies in its minimal power consumption, making it an
be beyond the scope of this paper. Multiple LDRs are strategically efficient choice for our purposes.
located within the greenhouse to capture light levels comprehensively.

5
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 7. Actuators indicative circuit diagram.

Fig. 8. Design of the PCB.

3.4.4. Luminosity 3.5. Combined circuit

OEM cool white LED Strips are utilized as artificial light sources to
complement sunlight. These strips are positioned on the top side to To optimize system efficiency, we’ve devised a PCB layout for both
provide supplementary lighting. the sensors and actuators circuits (Fig. 8). This PCB design facilitates
seamless integration with the Arduino Nano RP 2040 Connect (Fig. 9),
3.4.5. Side window mode leveraging a microcontroller compatible with cloud connectivity and
This functionality involves employing a Waveshare MG90 S servo data uploading for IoT applications.
motor, which is affixed to the plexiglass structure through a tailored 3D
printed design. This setup enables the creation of a flexible airway,
facilitating the regulation of temperature and humidity levels. The servo
responds to PWM signals, thereby adjusting the window opening to
achieve the desired effect.

6
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 9. Arduino Nano RP2040 Connect.

Fig. 10. Crisp and Fuzzy set comparison.

Fig. 11. Block diagram of the Fuzzy Inference System.

4. The control scheme The fundamental components of fuzzy logic are fuzzy sets, which
lack a well-defined boundary and are characterized by a curve that maps
4.1. Fuzzy logic preliminaries the input space to a value representing the degree of membership:
A = (x, μA (x) ∨ x ∈ X) (1)
Fuzzy logic operates on approximations rather than exact values,
accommodating uncertainties and unknown system models, much like A curve like μA(x) is termed a membership function and is typically
human reasoning. Unlike Boolean logic, which relies on clear-cut dis­ determined subjectively. When comparing a fuzzy set with a crisp
tinctions where statements are either entirely true or false, fuzzy logic Boolean set, the linguistic expression "The temperature is high" can be
allows for intermediate truth values between 0 and 1. Fuzzy logic rec­ seen why is often employed in goals related to microclimate regulation.
ognizes that statements can have varying degrees of membership in a In Fig. 10, when considering a temperature value of 24 ◦ C, the Crisp
set, thus expanding the concept to include multiple possible truth values set categorizes the statement as False (0). However, even a slight
ranging from 0 to 1. adjustment, such as increasing the temperature to 25 ◦ C, leads to a

7
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 12. Fuzzification.

Fig. 13. Output Fuzzy Sets.

Fig. 14. Defuzzification for the crisp input Temperature = 39.1 ◦ C, implemented using MATLAB R2021a.

significant change in the classification, now rendering the statement as interpretable rules to map crisp inputs to crisp outputs.
True (1). In contrast, the Fuzzy Set (FS) assigns a range of participation To delve deeper, let’s explore how a Cooling Fan functions based on
to the scenario of 24 ◦ C, indicating a "level of Truth" within that state­ temperature. During the Fuzzification stage, the discourse is divided into
ment, which amounts to 0.4256 (42.56 % True). At a higher temperature different fuzzy sets by selecting suitable linguistic variables and mem­
of 50 ◦ C, the participation level would be 1, indicating complete truth­ bership functions.
fulness in the statement. As the Cooling Fan progresses to the Inference/Rules stage, it takes
Fuzzy logic finds application in designing Fuzzy Inference Systems into account linguistic rules derived from expert knowledge and data.
(FIS). Unlike abstract mathematical approaches, FIS employs human- The fuzzy sets of input variables established during the fuzzification

8
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 15. Control block diagram.

Algorithm 1
Control Algorithm.
1: Set the reference values of the microclimate parameters
2: Get the actual values from the sensors
3: Produce the PWM signals for the actuators based on the deviation from the reference value.
4: Let the system evolve for an adaptation period, under the produced control signals
5: Get the new values from the sensors
6: Go to 3

Fig. 16. The Fuzzy PI controller.

stage are then mapped to the output sets. Below is a summary of an output and the output fuzzy sets are synthesized using various methods,
example rule set illustration, while Fig. 13 visually represents the output as depicted in Fig. 14. This illustration showcases the defuzzification
of fuzzy sets: process utilizing the Centroid method for the crisp inputs Temperature =
39.1 ◦ C
• If (Temperature is Low) then (Fan is Off)
• If (Temperature is Desired) then (Fan is Off) 4.2. The control scheme
• If (Temperature is Medium) then (Fan is Low)
• If (Temperature is Medium High) then (Fan is Medium) The fuzzy controller, outlined as described in the preceding section,
• If (Temperature is High) then (Fan is High) constitutes a FIS. However, the microclimate encompasses various pa­
rameters, leading to diverse inputs and outputs for the controller.
To finalize the process, during the Defuzzification stage, the crisp Attempting to encompass all of these within a unified inference system

9
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 17. Input membership functions.

Fig. 18. Output membership functions.

Fig. 19. IoT scheme.

would result in an excessively intricate ruleset. However, dedicated Fig. 15. The defuzzified outputs are translated into PWM signals, serving
actuators can directly influence various parameters, simplifying this as inputs to the actuators. Initially, reference values for the microclimate
mapping. Consequently, the initial system can be decomposed into parameters are established to ensure optimal plant growth, informed by
multiple subsystems and design individual single-input FIS controllers expert data. Subsequently, each measured value from the designated
for each. sensor is compared with these reference values, and any disparity serves
A system block diagram, partitioned into subsystems, is illustrated in as input to the controller. The controller generates commands for the

10
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 20. Temperature monitoring allows users to set desired reference values and observe both internal and external temperatures, along with their evolution over
time via a chart. Additionally, users can monitor the operating percentage of the fan and manually intervene if necessary.

the control input (Δu), maintaining the integrated Δu. Fig. 16 illustrates
the controller’s operation.
Both inputs will be assigned normalized values within the range [− 1,
1], helping to prevent integral windup. Additionally, the output will be
limited to the same range [− 1, 1], enabling insightful observations
regarding the ruleset and the resulting PWM control signal. Fig. 17 and
Fig. 18 display the membership functions for each input and the output,
respectively. Table 3 outlines the inference rules, where:

• NS = Small negative value

• NL= Large negative value
• PS = Small positive value
• PL = Large positive value
• ZE = Zero or near zero value

The FIS’s software implementation is achieved using a lookup table.

Therefore, the two versions of the input are connected to the nearest
value that matches one of the standardized outputs. This method proves
that the Fuzzification, Inference, and Defuzzification stages are
computationally more efficient as a matching process when applied by
microcontrollers.
It’s important to note that while the PWM signals control the electric
pump regarding the operation intensity of the LEDs and the Fan, the
PWM signals determine the duration of the operation. This allows for a
more accurate calculation of the amount of provided water. In the end,
the window receives an on-off signal based on the operational state of
the fan.

5. IoT implementation

IoT technology enables the remote monitoring of greenhouse con­

ditions. Sensors gather data, which is then transmitted to the micro­
controller in the physical layer and subsequently uploaded to the cloud.
This data is accessible from the IoT cloud platform by any device within
the network, facilitating processing using the designated control
Fig. 21. The app dashboard offers functionality similar to temperature moni­ scheme. Following data processing, corresponding commands are
toring as depicted in Fig.20, allowing users to monitor the LEDs through the relayed back to the device layer to govern the behavior of the actuators.
mobile application. In this research project, the Arduino cloud as the IoT platform was
employed, resulting in the development of a Dashboard monitoring
actuator to correct any deviations from the desired values. Afterwards, a environment. This Dashboard is accessible via mobile phones through
period of update inactivity is implemented to allow the controller’s the IoT Remote Arduino app. Examples of the Dashboard interface are
actions to influence the environment. In the final stage, the new value is depicted in Fig. 20 and Fig. 21.
captured by the sensor and used to regulate the microclimate once again, IoT technology has been integrated to enable remote monitoring of
if necessary. Algorithm 1 below summarizes the complete algorithm of the greenhouse environment. Sensors gather data, which is then trans­
the control process. mitted to the cloud via the microcontroller. In this study, the Arduino
In each subsystem, a fuzzy PI controller will be integrated, composed cloud is utilized as the IoT platform, where this data undergoes pro­
of a FIS with two inputs and one output. The inputs comprise the error cessing using the control scheme. Subsequently, commands are dis­
(e) and the error change (Δe), while the output represents the change in patched back to the actuators based on the processed data. The Arduino

11
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 22. The Cooling Subsystem’s response to a sudden increase in the microclimate temperature is depicted as follows: Initially, the dashboard displays the outside
temperature, followed by the evolution of the inside temperature.

Fig. 23. Irrigation Subsystem Response.

cloud also offers a Dashboard monitoring environment, accessible to 6. Results

other devices. Fig. 20 illustrates an example of this Dashboard, show­
casing features such as temperature monitoring and LED control. The greenhouse’s overall structure is depicted in Fig. 24, while the
Additionally, Fig. 21 demonstrates a mobile app Dashboard with similar design specifics of the window opening are outlined in Fig. 25. To gain
functionalities. insights into system responses, this research delves into the performance
of the Cooling and Irrigation Subsystems. In the context of the Cooling
Subsystem (Fig. 22), its reaction to internal temperature is scrutinized,

12
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

Fig. 24. The GH structure.

Funding

This paper was funded in the framework of “Pilot application of

greenhouse simulation of indigenous climatic conditions for the culti­
vation of Aloe (Aloe Vera) for optimal production of Acemannan poly­
saccharide”. M16SYN2-00282. Sub-measure 16.1–16.2 ”Establishment
and operation of Operational Groups of the European Innovation Part­
nership for agricultural productivity and sustainability”, Action 2, of the
Rural Development Programme 2014–2020 (RDP 2014–2020), Region
of Western Greece, December 2022 – 2024

Ethics statement

Not applicable: This manuscript does not include human or animal

research.

CRediT authorship contribution statement

Vasileios Thomopoulos: Writing – original draft, Validation, Soft­

ware, Resources, Data curation. Fotios Tolis: Writing – original draft,
Visualization, Software, Resources, Investigation, Formal analysis,
Conceptualization. Taxiarchis-Foivos Blounas: Writing – original
draft, Visualization, Software, Resources, Methodology, Investigation,
Fig. 25. The window opening design. Conceptualization. Dimitrios Tsipianitis: Writing – review & editing,
Supervision, Methodology, Formal analysis. Angeliki Kavga: Writing –
aiming for a setpoint of 30.75 ◦ C. Leveraging external temperature data, review & editing, Supervision, Project administration, Funding acqui­
the system demonstrates stability by maintaining values below the sition, Conceptualization.
desired internal temperature, allowing the Fan to introduce cooler air
into the microclimate. In Fig. 23, the Irrigation Subsystem’s behavior is Declaration of competing interest
observed, illustrating the process of attaining moisture equilibrium in
relatively dry soil to synchronize with a plant watering cycle. The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
7. Conclusions the work reported in this paper.

In this study, an original design for a small-scale greenhouse is pre­ Data availability
sented, covering various aspects including structure, materials, sensors,
and actuators. The inclusion of a fuzzy PI controller plays a crucial role Data will be made available on request.
in enabling self-regulation of the microclimate, while the integration of
IoT technology allows for remote monitoring. Future research endeavors
will focus on evaluating the effectiveness of the proposed scheme across References
different challenging environmental conditions to identify any potential
structural or monitoring shortcomings. Furthermore, exploring energy [1] H.C.J. Godfray, J.R. Beddington, I.R. Crute, L. Haddad, D. Lawrence, J.F. Muir,
J. Pretty, S. Robinson, S.M. Thomas, C. Toulmin, Food security: the challenge of
consumption patterns and optimization strategies, as well as integrating feeding 9 billion people, In, in: Science, 327, American Association for the
machine learning algorithms for growth stage monitoring and disease Advancement of Science (AAAS), 2010, pp. 812–818, https://doi.org/10.1126/
detection, represent important areas for further investigation. Lastly, the science.1185383. VolIssuepp.
[2] Ray, D.K., West, P.C., Clark, M., Gerber, J.S., Prishchepov, A.V., & Chatterjee, S.
insights gained from this design could be valuable for educational pur­ (2019). Climate change has likely already affected global food production. In Y. H.
poses, providing students with exposure to new technologies and envi­ Jung (Ed.), PLOS ONE (Vol. 14, Issue 5, p. e0217148). Public Library of Science
ronmental concepts. (PLoS). https://doi.org/10.1371/journal.pone.0217148.
[3] C. Maraveas, C.S. Karavas, D. Loukatos, T. Bartzanas, K.G. Arvanitis, E. Symeonaki,
Agricultural greenhouses: resource management technologies and perspectives for
zero greenhouse gas emissions, Agriculture 13 (2023) 1464, https://doi.org/
10.3390/agriculture13071464.

13
V. Thomopoulos et al. Smart Agricultural Technology 8 (2024) 100446

[4] A. Kavga, V. Thomopoulos, P. Barouchas, N. Stefanakis, A. Liopa-Tsakalidi, engineer optical functions in next-generation light management, Adv. Mater. 34 (6)
Research on innovative training on smart greenhouse technologies for economic (2022) 2104473.
and environmental sustainability, Sustainability. 13 (2021) 10536, https://doi. [19] Fazeel Ahmed Khan, Adamu Abubakar Ibrahim, Akram M. Zeki, Environmental
org/10.3390/su131910536. monitoring and disease detection of plants in smart greenhouse using internet of
[5] A. Çaylı, A. Akyüz, A.N. Baytorun, S. Boyacı, S. Üstün, F.B. Kozak, Sera Çevre things, J. Phys.Commun. 4 (2020), https://doi.org/10.1088/2399-6528/ab90c1,
Koşullarının Nesnelerin İnterneti Tabanlı İzleme ve Analiz Sistemi ile 5ISSN 23996528.
Denetlenmesi, in: Turkish journal of agriculture - food science and technology, 5, [20] Pinar Kirci, Erdinc Ozturk, Yavuz Celik, A novel approach for monitoring of smart
Turkish Science and Technology Publishing (TURSTEP), 2017, pp. 1279–1289, greenhouse and flowerpot parameters and detection of plant growth with sensors,
https://doi.org/10.24925/turjaf.v5i11.1279-1289.1282. Agriculture 12 (10) (2022) 1705.
[6] A. Bhujel, J.K. Basak, F. Khan, et al., Sensor systems for greenhouse microclimate [21] Abhishek Kumar, Vikrant Singh, Saurabh Kumar, Shiva Pujan Jaiswal, Vikas
monitoring and control: a review, J. Biosyst. Eng. 45 (2020) 341–361, https://doi. Singh Bhadoria, Iot enabled system to monitor and control greenhouse, 49, Elsevier
org/10.1007/s42853-020-00075-6. Ltd, 2020, pp. 3137–3141, https://doi.org/10.1016/j.matpr.2020.36911.040,
[7] B.B. Sinha, R. Dhanalakshmi, Recent advancements and challenges of internet of pages.
things in smart agriculture: a survey, In, in: Future Generation Computer Systems, [22] Luca La Notte, Lorena Giordano, Emanuele Calabrò, Roberto Bedini,
126, Elsevier BV, 2022, pp. 169–184, https://doi.org/10.1016/j. Giuseppe Colla, Giovanni Puglisi, Andrea Reale, Hybrid and organic photovoltaics
future.2021.08.006. for greenhouse applications, Appl. Energy 278 (2020) 115582.
[8] Khaldun I Arif, Hind Fadhil Abbas, Design and implementation a smart greenhouse, [23] Edgar Maya Olalla, Andres Lopez Flores, Marcelo Zambrano, Mauricio
Internat. J. Comput. Sci. Mobile Comput. 4 (8) (2015) 335–347. Domínguez Limaico, Henry Diaz Iza, Carlos Vasquez Ayala, Fuzzy control
[9] Maher Azaza, Camel Tanougast, Enrico Fabrizio, A. Mami, Smart greenhouse fuzzy application to an irrigation system of hydroponic crops under greenhouse: case
logic based control system enhanced with wireless data monitoring, ISA Trans. 61 cultivation of strawberries (fragaria vesca), Sensors 23 (2023), https://doi.org/
(2016) 297–307. 10.3390/s23084088.378, 4ISSN 14248220.
[10] Chiara Bersani, Carmelina Ruggiero, Roberto Sacile, Abdellatif Soussi, Enrico Zero, [24] T. Porselvi, et al., Automatic control and monitoring of greenhouse system using
Internet of things approaches for monitoring and control of smart greenhouses in IoT, Turkish J. Comput. Math. Edu. (TURCOMAT) 12 (11) (2021) 2708–2715.
industry 4.0, 2022, 5ISSN 19961073. [25] P. Parvatha Reddy, Sustainable crop protection under protected cultivation.
[11] Arindom Chakraborty, Monirul Islam, Animesh Dhar, Mohammad Shahadat Number BOOK, Springer, 2016.
Hossain, Iot Based greenhouse environment monitoring and smart irrigation [26] Omkar Sakpal, Swapnil Shetye, Shreyas Patil, Shubham Surve, and Anuradha
system for precision farming technology, Institute of Electrical and Electronics Jadiya. Iot based smart greenhouse. Available at SSRN 4108448, 2022.
Engineers Inc., 2022, pp. 123–128, https://doi.org/10.1109/ [27] Zaidon Faisal Shenan, Ali Fadhil Marhoon, Abbas A Jasim, Iot based intelligent
ICISET54810.2022.9775852, pagesISBN 9781665483971. greenhouse monitoring and control system, Basrah J. Eng. Sci. 17 (1) (2017)
[12] Deisiane Santos da Cruz, Caio Castro Rodrigues, Otavio A Chase, Denmora Gomes 61–69.
de Araujo, Jose Felipe Souza de Almeida, IoT-based smart mini greenhouse, 7, In [28] Meriem Soussi, Mohamed Thameur Chaibi, Martin Buchholz, Zahia Saghrouni,
Federal Rural University of Amazonia, 2019. Comprehensive review on climate control and cooling systems in greenhouses
[13] Mohamed A. Elashiri, Ahmed T. Shawky, Fuzzy smart greenhouses using Iot, IEEE, under hot and arid conditions, Agronomy 12 (3) (2022) 626, https://doi.org/
2018, pp. 1–4, https://doi.org/10.1109/ICCIC. 2018.8782307, pages12ISBN 978- 10.3390/agronomy12030626. MarchISSN 2073-4395.
1-5386-1508-9URL, https://ieeexplore.ieee.org/document/8782307/. [29] Mohammad Woli Ullah, Mohammad Golam Mortuza, Md Humayun Kabir, Zia
[14] Nafees Akhter Farooqui, Amit Kumar Mishra, Ritika Mehra, Iot based automated Uddin Ahmed, Sovan Kumar Dey Supta, Partho Das, and Syed Mohammad Didar
greenhouse using machine learning approach, Internat. J. Intell. Syst. and App. Hossain. Internet of things based smart greenhouse: remote monitoring and
Eng. 10 (2) (2022) 226–231. automatic control, DEStech Trans. Environ. Energy Earth Sci. (2018).
[15] Carlos Andrés González-Amarillo, Juan Carlos Corrales-Muñoz, Miguel [30] Amantur Umarov, Murat Kunelbayev, Maksat Satymbekov, Gulzat Turken, Bagila
Ángel Mendoza-Moreno, Ahmed Faeq Hussein, N. Arunkumar, Gustavo Ramirez- Alimbayeva, Kulbarchin Imanzhanova, and Laura Duisembayeva. Micro climate
González, An iot-based traceability system for greenhouse seedling crops, IEEe monitoring system for a home greenhouse as part of esp32, Test Eng. Manage. 82
Access. 6 (2018) 67528–67535. (1–2) (2020) 4564–4573.
[16] Carlos A. Hernández-Morales, J.M. Luna-Rivera, Rafael Perez-Jimenez, Design and [31] Demetres Briassoulis Von Elsner, Dries Waaijenberg, Antonis Mistriotis, Chr
deployment of a practical iot-based monitoring system for protected cultivations, Von Zabeltitz, Jean Gratraud, Giovanni Russo, Ricardo Suay Cortes, Review of
Comput. Commun 186 (2022) 51–64, https://doi.org/10.1016/j. structural and functional characteristics of greenhouses in european union
comcom.2022.01.009, 3ISSN 1873703X. countries: part i, design requirements, J. Agricult. Eng. Res. 75 (1) (2000) 1–16.
[17] Hiep Xuan Huynh, Linh Nhut Tran, Nghia Duong-Trung, Smart greenhouse [32] Ilhami Yildiz, Greenhouse engineering: integrated energy management, CRC Press,
construction and irrigation control system for optimal brassica juncea 2021.
development, PLoS. One 18 (2023), https://doi.org/10.1371/journal. [33] Israr Ullah, et al., ANN based learning to Kalman filter algorithm for indoor
pone.0292971, 10ISSN 19326203. environment prediction in smart greenhouse, IEEe Access. 8 (2020)
[18] Joice Jaqueline Kaschuk, Yazan Al Haj, Orlando J Rojas, Kati Miettunen, 159371–159388.
Tiffany Abitbol, Jaana Vapaavuori, Plant-based structures as an opportunity to [34] Medhat A. Tawfeek, Saad Alanazi, A.A. Abd El-Aziz, Smart greenhouse based on
ANN and IOT, Processes 10 (11) (2022) 2402.

14