Fuzzy Logic Controlled Motor Speed in Rotating

Aquaponics Based on Lactuca sativa Leaf Chlorosis

and Necrosis and Environment Temperature
2021 IEEE 13th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) | 978-1-6654-0167-8/21/$31.00 ©2021 IEEE | DOI: 10.1109/HNICEM54116.2021.9731945

Heinrick L. Aquino Ronnie S. Concepcion II Argel A. Bandala Christan Hail Mendigoria

Department of Electronics and Department of Manufacturing Department of Electronics and Department of Electronics and
Computer Engineering Engineering and Management Computer Engineering Computer Engineering
De La Salle University De La Salle University De La Salle University De La Salle University
Manila, Philippines Manila, Philippines Manila, Philippines Manila, Philippines
[email protected] [email protected] [email protected] [email protected]

Oliver John Alajas Elmer P. Dadios Joel Cuello

Department of Electronics and Department of Manufacturing Agricultural and Biosystems
Computer Engineering Engineering and Management Engineering Department
De La Salle University De La Salle University The University of Arizona
Manila, Philippines Manila, Philippines Arizona, USA
[email protected] [email protected] [email protected]

Abstract—Necrosis and chlorosis are some of the leaf for herbs and leafy crops driven by a motor for its rotating
conditions that contribute to losses in crop production, which is mechanism. The crops rotate in a wheel-like structure and the
the browning and yellowing of leaves caused by improper speed of the motor indicates the frequency of watering of
irrigation and fertigation. A rotating aquaponics, crop crops in the system [2]. The physical appearance of the plant,
cultivation conceptualized initially by NASA to save space and
particularly the leaves, mimics essential aspects that are
grow their food on space stations is the inspiration of application
of this study. A fuzzy logic controlled (FLC) DC motor speed quickly modified by harmful viruses, bacteria, and fungus,
controller was developed using the Mamdani system to automate intruding weeds, and abiotic variables such as heat and water
the rotating mechanism that is responsible for watering turns of stresses [3–7]. Hence, it is imperative to maintain ideal values
lettuce crops. Using CIE L*, a*, color space, and environment of parameters such as the water intake of crops and
temperature as input linguistic values, the motor speed will temperature to avoid necrosis and chlorosis caused by
adjust depending on the status of the physical pigment condition underwatering (browning) and overwatering (yellowing)
of leaves as well as the intensity of temperature. The generated respectively.
fuzzy logic controller has four triangular membership functions Machine learning over the years has been utilized in the
with 16 rules on each of the inputs. This resulted in four possible
myriad field of applications such as agriculture that catalyze
outputs of rotating aquaponics motor speed measured in rpm:
very slow (0.25), slow (0.5), slightly fast (1.25), and fast (2). The smart farming solutions [8–12]. For detecting healthy and
modeled FLC was simulated in a Simulink environment in damaged corn leaves and estimating the leaf damaged
MATLABR2021 software and had a 10-step size and manifested percentage attributed by maize Cercospora leaf spot, a hybrid
essentially accurate results with 100% correct outputs based on gaussian quantum-behaved particle swarm and recurrent
input characteristics and rules developed. This developed FLC neural network (RNN) were used with R2 (0.949), RMSE
model is a substantial contribution to mitigating losses on lettuce (6.290), and inference time (3 seconds) results [13]. Hybrid
crops grown under rotating aquaponics by automating the water linear discriminant analysis and decision tree application were
absorption frequency depending on the status of the crop and the able to classify healthy and diseased grape leaves that is
temperature of its environment.
97.79% more accurate than support vector machine
Keywords—rotating aquaponics, fuzzy logic, motor, chlorosis classification, naïve Bayes, and k-nearest neighbors’
and necrosis, controlled environment engineering algorithm as well as quantify the infected percentage on
diseased leaves using regression tree (R2 = 0.943) [14]. A DC
I. INTRODUCTION motor fuzzy logic controller (FLC) is developed under the
Aquaponics, through the years, has been an emerging Mamdani system with two inputs: speed error and change in
practice in the agriculture field to maximize resources for the speed error. The parameters of a PID controller are chosen
more food production. It is the integration of aquaculture or using a trial-and-error process. Simulations using the
fostering of marine products as well as hydroponics which is Simulink program were used to examine FLC and
the soilless cultivation of crops. The growing demand and proportional-integral-derivative (PID) which has a more
losses of food supply are some of the reasons researchers complicated design [15]. In terms of settling time, rising time,
continue to find solutions on how to optimize sustainable peak time, and % overshoot, a speed controller for a DC series
aquaponics production [1]. wound motor utilizing FLC outperforms a DC motor without
Rotating hydroponics, an original idea for National a controller [16]. By adjusting the armature voltage in the
Aeronautics and Space Administration (NASA) space constant torque zone and the field current in the constant
stations’ application to grow their food, is a cultivation system power region, FLC was designed to control the speed of a DC

978-1-6654-0167-8/21/$31.00 ©2021 IEEE

Authorized licensed use limited to: INTERNATIONAL ISLAMIC UNIVERSITY. Downloaded on July 15,2023 at 08:43:52 UTC from IEEE Xplore. Restrictions apply.
motor utilizing combined armature voltage and field current conditions have the same b* pigment range. Fig. 2 shows the
[17]. These mentioned related studies show the apparent rise segmentation of necrotic and chlorotic areas on lettuce crops.
of machine learning applications specifically fuzzy logic (FL)
in control systems. However, FLC application on rotary
aquaponics is still an open research gap.
This paper aims to provide a DC motor speed controller
for rotating aquaponics for lettuce cultivation based on three
factors: temperature (T, °C), L*, and a* CIELab color space
components. Fuzzy logic was implemented in modeling a
motor speed controller under the Mamdani system and
assessed in Simulink in MATLAB program. A total of four Fig. 2. Entire leaf and damaged area segmentation [18].
output speeds quantified in revolution per minute (rpm) was
developed based on the three inputs that contribute to the TABLE I. CIE L*A*B* COLOR SPACE LETTUCE SEGMENTATION VALUES
necrosis and chlorosis build-up on lettuce [18]. A Simulink
functional block model of DC motor speed controller was Category L a* b*
generated. This is to determine the performance of the Necrotic 42.55 99.71 1.434 37.2 -1.06 58.78
modeled FLC. This approach will be helpful for the automated Chlorotic 89.47 99.71 -21.5 5.005 -1.06 58.78
variation of motor speed that is synonymous with the
frequency of watering of lettuce crops depending on the B. Fuzzy Logic Control Modeling
necessity and sufficiency of the input parameters. Zadeh (1965) proposed fuzzy logic as a multi-valued logic
to cope with ambiguous and uncertain concepts. Unlike
II. MATERIALS AND METHODS computers that recognize binary values only, FL mimics
The DC motor rotation speed is predicted using three input human cognition and judgment, giving an interpretation to
predictors: temperature (°C), L*, and a* color space values. words like "sometimes," "colder," or "hotter” [20]. In this
These are the values that would determine the needed research, 4 membership function with 16 rules was designed
frequency of water absorption in rotating aquaponics. Fig. 1 as FLC for each of the three inputs. A Mamdani controller was
shows the process of the FLC model development built from utilized, and linguistic variables were set of which have titles,
these signatures. MATLAB R2021a is the software utilized defined domain, range of values, and an interpretation of
for FLC modeling, FL rule setting, and prediction of motor values. Originally, Mamdani fuzzy systems were created to
rotation speed. mimic the actions of manual operations in charge of
supervising manufacturing processes [21].
Temperature (°C), L*, a* parameter values

Fuzzy logic control system modeling

Fuzzy logic rule designation

Fig. 3. Development framework of the motor speed FLC-based model.

The fuzzy logic Mamdani design is shown in Fig. 3. Three

Simulink functional block simulation main linguistic variables were declared: Temperature, L*, and
a* and the output is the motor speed.
DC motor rotation speed TABLE II. INPUT VARIABLE L* DETAILS
Linguistic Value Notation Min. Max.
Fig. 1. Development framework of the motor speed FLC-based model.
White W 81.85 99.71
Slightly Brown SB 67.56 88.99
A. Linguistic Variables Value Source Brown BR 53.27 74.70
The value considerations for three linguistic variables Blackish BL 42.55 60.41
(temperature, L*, and a*) are based on [18-19]. For a
minimum of two months of optimum daytime temperatures of TABLE III. INPUT VARIABLE A* DETAILS
17 to 28°C and nighttime temperatures of no more than 15°C, Linguistic Value Notation Min. Max.
lettuce cultivation is most successful. Considering the Greeny G -21.51 -3.17
temperature in a tropical country such as the Philippines, the Slightly Green SG -10.50 11.51
set range of values for temperature is from (12 to 36°C). Slightly Red SR 4.17 26.19
Redish R 18.85 37.2
Table 1 shows the range of necrotic and chlorotic indicator
values in the CIE L*a* b* color space. It determines the
pigmentation status of lettuce crops; the range of values are
L* (42.547 to 99.71) and a* (-21.512 to 37.197). The b*
component was not considered since necrotic and chlorotic

Authorized licensed use limited to: INTERNATIONAL ISLAMIC UNIVERSITY. Downloaded on July 15,2023 at 08:43:52 UTC from IEEE Xplore. Restrictions apply.
 TABLE IV. INPUT VARIABLE TEMPERATURE DETAILS
Linguistic Value Notation Min. (°C) Max.
(°C)
Cold C 12 19.5
Normal N 16.5 25.5
Slightly Hot SH 22.5 31.5
Hot H 28.5 36

TABLE V. OUTPUT VARIABLE MOTOR SPEED DETAILS

Linguistic Value Notation RPM
Very Slow VS 0.25
Slow S 0.5 Fig. 7. Membership functions set for output rpm of the motor.
Slightly Fast SF 1.25
C. Fuzzy Logic Rule Designation
Fast F 2
There is a total of 64 rule combinations for the three input
Tables 2, 3, and 4 show the four membership functions linguistic values. This is based on what watering frequency,
that were developed on each linguistic variable and triangular rotating aquaponics needs depending on temperature, L*, and
membership functions that resulted in 16 rules each. Table 5 a*. If there is a slightly hot or hot temperature, the motor speed
shows the rotation speed of a motor results in 0.25, 0.5, 1.25, will be slightly fast or fast. If there is cold or normal
2 rpm in the triangular membership function based on the temperature, the motor will run on very slow or slow rotation.
input linguistic values. For better understanding, the Concerning the CIE L*a* values, the chlorotic range will turn
membership functions’ configurations are shown in Fig. 4, 5, the rpm into a slow range due to possible overwatering, this is
6, and 7. to have the crops more time to dry out. For necrotic range, it
will do the opposite due to underwatering. The motor will
increase the rpm speed for more frequent watering of crops.
Fig. 8 shows the excerpt of the 64 rule combinations in the
FLC model.

Fig. 4. Membership functions set for L*.

Fig. 5. Membership functions set for a*.

Fig. 8. Excerpt of fuzzy rules set for DC motor speed control

D. Simulink Functional Block Simulation and Evaluation

The modeled fuzzy logic-based motor controller was
simulated on the Simulink package in the MATLAB R2021a
environment. In this output simulation, the performance of the
algorithm was tested based on the three linguistic values in the
input.
III. RESULTS AND DISCUSSIONS
A. Simulink Functional Block Simulation
Fig. 6. Membership function set for temperature.
The three input linguistic variables (temperature, L*, and
a*) are set to simulate sine wave, with time-based sine type
with 2.75, 1.75, and 1 rad/sec frequency respectively. Fig. 9
shows you the simulation block with three sine wave input,

Authorized licensed use limited to: INTERNATIONAL ISLAMIC UNIVERSITY. Downloaded on July 15,2023 at 08:43:52 UTC from IEEE Xplore. Restrictions apply.
FLC, display, and two scopes (one in input channel and one
on output) to determine the readings for every step size.

Motor Speed (rpm)

Step Sample
Fig. 11. Simulink Functionality Block Wave Results Comparison of expected
Fig. 9. Simulink fuzzy functional block model for DC motor speed control DC motor speed (orange dot lines) and the developed fuzzy logic-
generated values (blue line).
The simulation time is from 0 to 10 samples, with fixed-
Table 6 shows the values of input variables L*, a*, and
step solver selection type, solver of ode3 (Bogacki-
temp and their corresponding results in every 1 step size until
Shampine), and 0.001 fixed step size. Fig. 10 shows the sine
10 samples. Based on these results, it appears to be an accurate
wave input simulation values for the developed FLC for DC
fuzzy logic controller that can be used on the proposed
motor speed. Varied frequencies were implemented to show
rotating aquaponics cultivation system in the prevention of
the close to actual simulations since the inputs are not
chlorosis and necrosis conditions on lettuce crop productions.
necessarily in the same phase in actual experimentation. Fig.
11 however shows the corresponding wave results with In comparison to previous studies [15–17], a fuzzy logic
observed noises, along step time based on the inputs. In the controller under the Mamdani system indeed produces a
one-second, linguistic values manifested L* = 83, a* = flexible, decent, and reliable model. In all ten-step trials, the
36.8147, temp = 34.0227, and a result of 2 rpm which is ideal created FLC model for motor speed control significantly
since all three inputs fall under necrotic indicator values resulted in good rpm outputs based on the rules set. For future
especially the temperature that triggers drought to lettuce studies, this current simulation can be integrated with [22, 23]
crops. On the 3 seconds step time, input values were L* =
to increase agricultural productivity.
97.7302, a* = -17.248, temp = 25.9006. These L* and a*
values fall under chlorotic values that is why the rpm speed TABLE VI. SUMMARY OF SIMULATION LINGUISITIC VALUES AND RESULTS
lowered from 2 rpm into 0.5 rpm where the temperature is on
FLC-generated
the normal level. This indicates overwatering on crops; thus, Sample L* a* T (°C)
Motor Speed (rpm)
a slower rotation is recommended to slowdown turns in water
absorption of crops. On the 5 second drop, linguistic values 1 83 36.8147 34.0227 2
were 97.7302, 25.9006, and 12.4484 for L*, a*, and temp 2 50.7745 -2.5266 35.038 2
respectively. These indicate chlorotic signs on L*, while the
3 97.7302 -17.248 25.9006 0.5
necrotic sign on a* parameter, but since there is a significantly
low recorded temperature, a 0.25 rpm speed was suggested by 4 42.6524 27.1697 14.9866 0.25
the FLC. On the 9 seconds step time, there are L* = 60.4195, 5 97.7302 25.9006 12.4484 0.25
a* = 6.6107, temp = 28.9464 observed values, of which goes
6 51.2821 -17.756 20.3167 0.25
to the category of necrotic, thus rpm speed was increased to
be on 1.25 value. 7 82.2475 -1.2576 31.7384 2
8 71.5873 37.0685 35.7994 2
9 60.4195 6.6107 28.9464 1.25

10 91.3849 -20.04 17.2709 0.25

L*, a* and Temperature Values

IV. CONCLUSION
A fuzzy logic-controlled DC motor speed controller is
presented in this paper. Under the Mamdani system, four
membership functions having 16 rules were created for each
of the three linguistic values which are L*, a*, and
temperature. Triangular membership function was utilized for
the inputs and results in a four-membership output which is
the speed of the DC motor rotation indicated as very slow for
0.25 rpm, slow for 0.5 rpm, slightly fast for 1.25 rpm, and fast
for 2 rpm. The developed fuzzy logic controller then
Step Sample underwent simulation to evaluate its performance. A ten-
Fig. 10. Input sampled based steps of linguistic variables: L* (blue line), T second step time was implemented, and all step time
(light blue line), and a* (orange line) on the y-axis.

Authorized licensed use limited to: INTERNATIONAL ISLAMIC UNIVERSITY. Downloaded on July 15,2023 at 08:43:52 UTC from IEEE Xplore. Restrictions apply.
manifested essentially accurate results based on input [13] R. Concepcion, E. Dadios, J. Alejandrino, C. H. Mendigoria, H.
Aquino, and O. J. Alajas, “Diseased Surface Assessment of Maize
characteristics.
Cercospora Leaf Spot Using Hybrid Gaussian Quantum-Behaved
This developed FLC model is a substantial contribution to Particle Swarm and Recurrent Neural Network,” Apr. 2021. doi:
mitigating losses on lettuce crops grown under rotating 10.1109/IEMTRONICS52119.2021.9422515.
aquaponics by automating the water absorption frequency [14] O. J. Alajas, R. Concepcion, E. Dadios, E. Sybingco, C. H.
Mendigoria, and H. Aquino, “Prediction of Grape Leaf Black Rot
depending on the status of the crop and the temperature of its
Damaged Surface Percentage Using Hybrid Linear Discriminant
environment. In the future, it is recommended to consider Analysis and Decision Tree,” Jun. 2021. doi:
other additional inputs to have a better prediction and 10.1109/CONIT51480.2021.9498518.
observance of how chlorosis and necrosis occur on crops. [15] Y. A. Almatheel and A. Abdelrahman, “Speed control of DC motor
using Fuzzy Logic Controller,” Jan. 2017. doi:
ACKNOWLEDGMENT 10.1109/ICCCCEE.2017.7867673.
[16] N. L. Ismail, K. A. Zakaria, N. S. M. Nazar, M. Syaripuddin, A. S.
The researchers acknowledge the assistance given by De N. Mokhtar, and S. Thanakodi, “DC motor speed control using
La Salle University – Manila’s Intelligent System Laboratory fuzzy logic controller,” 2018. doi: 10.1063/1.5022920.
and the Department of Science and Technology of the [17] A. A. Sadiq, G. A. Bakare, E. C. Anene, and H. B. Mamman, “A
Fuzzy-Based Speed Control of DC Motor Using Combined
Engineering Research and Development for Technology in the Armature Voltage and Field Current,” IFAC Proceedings Volumes,
Philippines vol. 46, no. 20, 2013, doi: 10.3182/20130902-3-CN-3020.00146.
[18] R. Concepcion II and E. Dadios, “Lettuce Leaf Necrotic and
REFERENCES Chlorotic Surface Defect Assessment Using RNN Optimized by
[1] M. Rivera-Ferre, M. Ortega-Cerdà, and J. Baumgärtner, EM,” Beijing, 2020.
“Rethinking Study and Management of Agricultural Systems for [19] G. Rogers, “Impacts By Crop: Lettuce,” Applied Horticultural
Policy Design,” Sustainability, vol. 5, no. 9, Sep. 2013, doi: Research, 2021.
10.3390/su5093858. [20] M. Shamim, S. Enam, U. Qidwai, and S. Godil, “Fuzzy logic: A
[2] C. Kim, “The Green Wheel,” The Green Wheel, Jan. 19, 2013. ‘simple’ solution for complexities in neurosciences?,” Surgical
https://www.barrons.com/articles/SB50001424052748703596604 Neurology International, vol. 2, no. 1, 2011, doi: 10.4103/2152-
578235652208700088 (accessed Feb. 11, 2021). 7806.77177.
[3] Hadad, Luria, Smith, Sela, Lachman, and Dombrovsky, “Lettuce [21] C. Moraga, “Introduction to fuzzy logic,” Facta universitatis -
Chlorosis Virus Disease: A New Threat to Cannabis Production,” series: Electronics and Energetics, vol. 18, no. 2, 2005, doi:
Viruses, vol. 11, no. 9, Aug. 2019, doi: 10.3390/v11090802. 10.2298/FUEE0502319M.
[4] R. S. Concepcion, E. P. Dadios, J. N. Carpio, A. A. Bandala, and [22] R. Concepcion II, E. Dadios, J. Cuello, and B. Duarte, “Thermo-
E. Sybingco, “Phytopigments Profiling of Lactuca Sativa Leaf gas dynamics affect the leaf canopy shape and moisture content of
Chloroplast Photosystems via Vision-based Planar aquaponic lettuce in a modified partially diffused microclimatic
Chromatography,” Dec. 2020. doi: chamber,” Scientia Horticulturae, vol. 292, 2021. doi:
10.1109/HNICEM51456.2020.9400156. https://doi.org/10.1016/j.scienta.2021.110649.
[5] R. S. Concepcion II, E. P. Dadios, and J. Cuello, “Non-destructive [23] C. H. Mendigoria, H. Aquino, J. Alajas, R. Concepcion II, E.
in Situ Measurement of Aquaponic Lettuce Leaf Photosynthetic Dadios, E. Sybingco, A. Bandala, and R.R. Vicerra, “Varietal
Pigments and Nutrient Concentration Using Hybrid Genetic Classification of Lactuca Sativa Seeds Using an Adaptive Neuro-
Programming,” AGRIVITA Journal of Agricultural Science, vol. Fuzzy Inference System Based on Morphological Phenes,” Journal
43, no. 3, Oct. 2021, doi: 10.17503/agrivita.v43i3.2961. of Advanced Computational Intelligence and Intelligent
[6] R. Concepcion II, S. Lauguico, V. J. Almero, E. Dadios, A. Informatics, vol. 25, no. 5, pp. 618-624, 2021. doi:
Bandala, and E. Sybingco, “Lettuce Leaf Water Stress Estimation https://doi.org/10.20965/jaciii.2021.p0618
Based on Thermo-Visible Signatures Using Recurrent Neural
Network Optimized by Evolutionary Strategy,” Dec. 2020. doi:
10.1109/R10-HTC49770.2020.9356963.
[7] R. Concepcion et al., “Lactuca sativa leaf extract concentration
optimization using evolutionary strategy as photosensitizer for
TiO2-filmed Grӓtzel cell,” Optik, vol. 242, Sep. 2021, doi:
10.1016/j.ijleo.2021.166931.
[8] H. L. Aquino, R. S. Concepcion, E. P. Dadios, E. Sybingco, and A.
A. Bandala, “Trend Forecasting of Computer Vision Application in
Aquaponic Cropping Systems Industry,” Dec. 2020. doi:
10.1109/HNICEM51456.2020.9400070.
[9] N. M. Tiglao, M. Alipio, J. V. Balanay, E. Saldivar, and J. L.
Tiston, “Agrinex: A low-cost wireless mesh-based smart irrigation
system,” Measurement, vol. 161, Sep. 2020, doi:
10.1016/j.measurement.2020.107874.
[10] R. Concepcion, S. Lauguico, R. R. Tobias, E. Dadios, A. Bandala,
and E. Sybingco, “Genetic Algorithm-Based Visible Band
Tetrahedron Greenness Index Modeling for Lettuce Biophysical
Signature Estimation,” Nov. 2020. doi:
10.1109/TENCON50793.2020.9293916.
[11] R. Concepcion, S. Lauguico, E. Dadios, A. Bandala, E. Sybingco,
and J. Alejandrino, “Tomato Septoria Leaf Spot Necrotic and
Chlorotic Regions Computational Assessment Using Artificial Bee
Colony-Optimized Leaf Disease Index,” Nov. 2020. doi:
10.1109/TENCON50793.2020.9293743.
[12] S. C. Lauguico, R. S. Concepcion II, J. D. Alejandrino, R. R.
Tobias, D. D. Macasaet, and E. P. Dadios, “A Comparative
Analysis of Machine Learning Algorithms Modeled from Machine
Vision-Based Lettuce Growth Stage Classification in Smart
Aquaponics,” International Journal of Environmental Science and
Development, vol. 11, no. 9, 2020, doi:
10.18178/ijesd.2020.11.9.1288.

Authorized licensed use limited to: INTERNATIONAL ISLAMIC UNIVERSITY. Downloaded on July 15,2023 at 08:43:52 UTC from IEEE Xplore. Restrictions apply.