IoT AND ARTIFICIAL INTELLIGENCE SUPPORTED SMART
SYSTEM DESIGN TO INCREASE PRODUCTIVITY IN
AGRICULTURE
Köksal GÜNDOĞDU 1 Ali ÇALHAN 2
schola @koksalgundogdu.com [email protected].
ORCID Kimliği: 0000-0001-7694-1841 ORCID Kimliği: 0000-0002-5798-3103
Mu aza CİCİOĞLU 3
mu [email protected].
ORCID Kimliği: 0000-0002-5657-7402
1Elec ical, Elec onic and Compu e Enginee ing, Düzce Uni e si y, Düzce, TÜRKİYE
2 Compu e Enginee ing, Düzce Uni e si y, Düzce, TÜRKİYE
3 Compu e Enginee ing, Bu sa Uludağ Uni e si y, Bu sa, TÜRKİYE
ABSTRACT
The ag icul u al sec o is inc easing i s s a egic impo ance daily due o he inc easing
popula ion wo ldwide and he dec easing labo o ce in u al a eas. In oday's wo ld, whe e
he demand o basic esou ces such as ood, wa e , and ene gy is apidly inc easing, he
inadequacy o adi ional ag icul u al me hods has made i necessa y o use new echnologies
e ec i ely in ag icul u e. The e o e, concep s such as Ag icul u e 4.0, a i icial in elligence,
In e ne o Things (IoT), senso echnologies, and digi al ans o ma ion ha e become he
main ools o inc ease p oduc i i y in ag icul u al p oduc ion and p e en esou ce was e.
This s udy de eloped a use - iendly, low-cos , modula IoT-based sys em ha does no
equi e echnical knowledge o add ess he p oblems encoun e ed in ag icul u e. The sys em
includes a mul i-senso s uc u e ha can measu e en i onmen al a iables such as soil
mois u e, ai empe a u e, humidi y, gas concen a ion, wind speed, and di ec ion, and i
p o ides eal- ime moni o ing and analysis by ans e ing hese da a o cloud-based se e s.
P ocessing he da a ob ained wi h a i icial in elligence-suppo ed algo i hms enables
applica ions such as an in eg a ed sma ag icul u al s a ion, clima e con ol in he g eenhouse,
and a plan heal h moni o . The sys em has a lexible s uc u e ha ag icul u al enginee s,
ag icul u al o ganiza ions, and a me s can adap o hei needs. In addi ion, i is designed
wi h he “plug-and-play” p inciple so ha non- echnical use s can easily ins all and use he
sys em. Digi al aining ma e ials encou age use s o de elop hei applica ions on he sys em.
This solu ion aims o elimina e ba ie s o digi aliza ion in ag icul u e, such as complexi y,
high cos , and he need o expe ise, and i o e s a p ac ical, accessible, and scalable pa h o
sus ainable ag icul u e.
Key Wo ds: Ag icul u e 4.0, A i icial In elligence, In e ne o Things, Digi al
T ans o ma ion, Sma Ag icul u e Sys em
I. INTRODUCTION
Ag icul u e is o c i ical impo ance in he wo ld. In o he wo ds, i is a s a egic p oduc [1].
The Eu opean Commission has emphasized he impo ance o Ag icul u e and p epa ed
a ious s a egic plans. The e a e app oxima ely 9.5 million a me s and 10.8 million holding-
le el ag icul u al companies in Eu ope [2]. As can be seen in Figu e 1, he sha e o
ag icul u al employmen is 17.7% in Tü kiye [3].
Figu e 1. Tu kish S a is ical Ins i u e 2018 Labo Fo ce and Employmen Table [3]
While he wo ld popula ion is inc easing, he numbe o people engaged in ag icul u e is
dec easing. Fo his eason, Ag icul u e 4.0 and he use o echnology in ag icul u e ha e
become ine i able. While echnology inc eases p oduc i i y in ag icul u e, i also educes he
equi ed labo o ce [4-13]. In he coming yea s, ewe esou ces o a g owing wo ld will
exis . Howe e , i will be necessa y o p oduce mo e ood, eed, uel, and ibe . Sma
ag icul u e and digi aliza ion in ag icul u e will be he mos impo an ac o in o e coming
hese challenges [14]. Realizing hese si ua ions globally, coun ies ha e in es ed hea ily in
sma ag icul u e and IOT applica ions [5]. The Eu opean In es men Bank has alloca ed 32
billion eu os mo e o sma ag icul u e be ween 2014 and 2018 [15]. The Uni ed Kingdom
alloca es a budge o 4.7 billion pounds o applica ions ha will be ealized as Ag icul u e
4.0, a i icial in elligence, and p ecision ag icul u e [16]. I is p edic ed ha he wo ld
popula ion will exceed 9 billion by 2050, and he u baniza ion a e will inc ease o 70% [7,
17]. I we do no use he esou ces we ha e e ec i ely, he wo ld will ine i ably ace hunge
soon. While he sha e o ag icul u al land in o al land in he wo ld is 38%, Tü kiye is in a
lucky posi ion wi h 51% [18]. Howe e , unconscious ag icul u e leads o a dec ease in he
p oduc i i y o ou ag icul u al lands and e en o he des uc ion o ag icul u al lands.
Excessi e i iga ion causes he soil po es o be illed and he soil o be ai less. I des oys
mine als use ul o he soil on sloping lands. I also causes plan oo s o o . Pump i iga ion
causes wa e loss up o 50% and unnecessa y ene gy consump ion [18, 19]. Unnecessa y and
un imely sp aying causes excessi e damage o plan s and soil by c ea ing pes icide esidues.
Unnecessa y e iliza ion makes he soil ine icien inancially and changes he mine al and
PH balances o he soil. One o he mos common p oblems is ag icul u al os . The main
cause o os is he cooling o he soil su ace as a esul o he loss o ene gy om he soil by
g ound adia ion in cloudless, clea , d y ai . A os e en can cause la ge ha es eco ds o
he a me s [20]. Al hough digi aliza ion is qui e common in ag icul u e, i is seen ha hese
sys ems, which ha e a s anda d s uc u e oday, ha e e y low yield a es in some cases, o
e en educe p oduc i i y, depending on he land condi ions o he p oduc in he ag icul u al
a ea. Challenges such as he complexi y o digi aliza ion, high cos s, elec onic in o ma ion
equi emen s, and lack o skills a e among he obs acles o he use o his echnology [4-9]. In
his s udy, a se has been de eloped o enable ag icul u al enginee s, au ho ized ins i u ions in
he ield o ag icul u e, o a me s o p oduce hei solu ions o he p oblems in ag icul u al
lands. Wi h his se , people wo king in he ield o ag icul u e can easily ead he senso s on
he de ice and w i e he esul s o he cloud se e . They will be able o p oduce solu ions by
e alua ing he w i en esul s hemsel es o wi h au ho ized o ganiza ions. The sys em is as
simple as a child assembling Legos and can be used wi hou equi ing elec onic knowledge.
In o he wo ds, i does no con ain echnical complexi y. I can be used wi h plug-and-play
logic. In addi ion, digi al cou se con en has been c ea ed o use s who wan o imp o e
hemsel es and make di e en a i icial in elligence-based applica ions on he se . In his way,
use s in he ield o ag icul u e will be able o c ea e hei own IOT sys ems and c ea e sma
ag icul u al sys ems sui able o hei lands.
II.SYSTEM DESIGN
A. SYSTEM OVERVIEW
As seen in Figu e 2, ou sys em consis s o ou main pa s. The i s pa is he
“Measu emen Uni ”, whe e he measu emen esul s a e aken using senso s in he
ag icul u al land. The second pa is he “E alua ion and Managemen Uni ”, whe e he
measu emen esul s a e e alua ed and he e alua ion esul s a e ans e ed o he wi eless
communica ion uni and w i en o he da abase. The hi d pa is he “Wi eless
Communica ion Uni ” ha enables he sys em o be used as an IOT de ice and o ansmi
da a wi elessly. The ou h pa is he “Da abase and Use In e ace”, which con ains he
da abase in which he da a is s o ed, and he da a in he da abase is displayed o he use ia
he Web and mobile de ices.
Figu e 2. Block Diag am o he Designed Sys em
B. MEASUREMENT UNIT
The Measu emen Uni has he senso s o he alues needed in he ag icul u al ield, and he
alues ha need o be measu ed a e loca ed. I con ains senso s o Soil Mois u e
Measu emen , Soil Tempe a u e Measu emen , Ambien Mois u e Measu emen , Ambien
Tempe a u e Measu emen , Ambien Ligh Amoun Measu emen , Ai Quali y Measu emen ,
Ai P essu e Measu emen , We ness Measu emen ( o lea es, ees, o simila objec s),
Rain all measu emen , Wind in ensi y measu emen , and Wind di ec ion measu emen . The
p obe o he ai quali y senso is shown in Figu e 3.
Figu e 3. Ai Quali y Senso P obe
As can be seen in Figu e 3, he senso s a e easy o use and plug-and-play. I does no equi e
any echnical knowledge and is designed o ag icul u al land. A mosphe e Senso is a senso
ha can measu e om a dep h o -500m o 9000m abo e sea le el. This senso can ope a e
be ween -40°C and +85°C empe a u e le els. Thanks o he senso , he a mosphe ic p essu e
o he ag icul u al land is measu ed. Soil Tempe a u e Senso can measu e soil empe a u e
be ween -55°C and +125°C. I wo ks acco ding o ±0.5°C deg ee o accu acy. The Soil
Mois u e Senso ope a es wi h 16-bi sensi i i y. I has an ope a ing cu en o app oxima ely
20mA. I wo ks acco ding o he p inciple o elec ical conduc i i y be ween wo p obes
plan ed in he soil. I he soil is mois , conduc i i y inc eases; i he soil is d y, conduc i i y
dec eases. The Ambien Tempe a u e Senso can measu e ambien empe a u e be ween -
40°C and +80°C. I wo ks acco ding o ±0.5°C accu acy. The Ambien Humidi y Senso
wo ks acco ding o a 0% ~ 100% RH ange. I has ±1% RH humidi y sensi i i y. The
Ambien Ligh Amoun Measu emen Senso is 540 nanome e s in wa eleng h. The ai
quali y measu emen senso is 10-100 ppm. P ecipi a ion amoun , wind in ensi y, and wind
di ec ion senso s can p oduce accu a e measu emen esul s be ween -40°C and +80°C.
C. EVALUATION AND MANAGEMENT UNIT
E alua ion and Managemen Uni ; as can be seen in Figu e 4, he e is an RP2040
mic ocon olle inside. I con ains a dual-co e ARM Co ex M0+ 133 MHZ p ocesso . I has
264KB SRAM and 2MB on-boa d memo y. I has low powe consump ion. I has 26 GPIO
uni s, 2xSPI, 2xI2C, 2xUART, 3x12-bi ADC, and 16x con ollable PWM channels. I
includes an in e nal empe a u e senso , a ime , and a clock module [21]. The e is a Mic o
USB on he boa d o p og am he boa d. I has a Buzze and SMD LEDs on i o gi e
wa nings. The e a e special socke inpu s o senso connec ions. The e a e special socke s o
connec wi eless communica ion uni s.
Figu e 4. E alua ion and Managemen Uni
D. WIRELESS COMMUNICATION UNIT
The Wi eless Communica ion Uni p o ides pins o he Wi i, GSM module, ZigBee, LoRa,
and Blue oo h modules. The e a e socke s in he E alua ion and Managemen Uni whe e
hese modules a e connec ed. The de aul wi eless communica ion me hod o he se is
wi eless da a ans e wi h a GSM module. A isual o he wi eless communica ion uni s
used on he boa d is shown in Figu e 5.
Figu e 5. Wi eless Communica ion Uni s
E. DATABASE AND USER INTERFACE
A Linux-based da abase was c ea ed o he sys em, and da abase ope a ions we e pe o med
wi h MYSQL. IOT da a is w i en o he da abase in his way. The da a ecei ed om he
da abase is displayed o he use as in Figu e 6.
Figu e 6. Use In e ace
The use in e ace is buil using PHP, HTML, MYSQL, and Ja aSc ip . Mobile compa ibili y
is p o ided wi h he Boo s ap me hod. In his way, i can be used on bo h web and mobile
de ices.
III. FINDINGS AND DISCUSSION
A. PLUG AND PLAY OPERATION
Figu e 7 shows he sys em in use on ag icul u al land. As can be seen, i is su icien o place
ou sys em on he ag icul u al land o be measu ed. Then, he da a is w i en o he da abase in
he cloud se e as soon as we ene gize i . The measu emen esul s a e sen a one-minu e
in e als.
Figu e 7. Physical S a e o he Sys em
When we plug he senso s o he alues o be measu ed in o he socke s shown in Figu e 8,
he senso s a s o wo k and au oma ically sends he measu emen esul s o he cloud se e .
In his way, i is enough o ene gize he sys em a e connec ing he senso s whose alues we
wan o obse e and e alua e. Fo example, i we wan o measu e and obse e Ai Quali y,
we should connec he senso S5 shown in Figu e 3 o he S5 socke shown in Figu e 8. Then,
when we ene gize he sys em, he measu emen esul s a e au oma ically sen o he cloud
se e .
Figu e 8. Connec ion Socke s
B. POWER CALCULATION
The sys em shown in Figu e 7 is a g een ene gy sys em [22]. The powe om he sola panel
is s o ed in a ba e y. The sys em is ed wi h powe om he ba e y, and a g een ene gy
sys em cycle is c ea ed. The block diag am o he sys em is shown in Figu e 9.
Figu e 9. Powe Block Diag am o he Sys em
The powe calcula ion based on he cu en and ol age alues o he sys em is as ollows.
Sola _Panel_P oduced_Powe = Vol ageValue* Cu en Value = 9V*0,033A ≈ 3W (1)
Ba e y_Powe = 12V*6A=72W (2)
Table 1. Calcula ion o Maximum Powe Consumed by he Sys em
Powe Consuming Uni Amoun o Powe Consumed (P=V*A) Amoun o Powe
Consumed (W)
Soil Mois u e Senso
5V*20mA=100mW 0.1W
Soil Tempe a u e Senso 5V*1.5mA=7.5mW 0.0075W
Ai Mois u e Senso 5V*0.5mA=2.5mW 0.0025W
Ai Tempe a u e Senso
Ambien Ligh Amoun
Senso
Day ime(1KΩ) I=5V/(1 KΩ +10 KΩ)=2.25mW
Nigh (1MΩ) I=5V/(1 MΩ +10 KΩ)=4.95µW 0.00225W
Ai Quali y Senso 5V*0.16A=0.8W 0.8W
Ai P essu e Senso 3.3V*0.000012A=0.0396mW 0.0000396W
Lea We ness Senso 5V*20mA=100mW 0.1W
P ecipi a ion Amoun Senso
5V*0.2mA=1mW 0.001W
Wind In ensi y Senso
Wind Di ec ion Senso
Mic ocon olle
3,3V*0.02A=66mW 0.066W
GSM Module ( o IOT)
4V*0.15A=0.6A 0.6A
O he s ≈ 40mA 0.04W
To al Maximum Powe Consumed ≈ 2.675W
While calcula ing he consumed powe shown in Table 1, he alues o he powe -consuming elemen s
in he sys em a he ime o maximum powe consump ion a e aken in o accoun , no a s eady s a e.
In addi ion, since he GSM module consumes he mos powe among he wi eless communica ion
modules in he sys em, he powe alues o he GSM module a e aken in o accoun .
As can be seen in Table 1;
To al Maximum Powe Consumed = 2.675W
As can be seen in equa ion (2);
Powe Gene a ed om Sola Panel = 3W
Residual powe a he sys em's maximum ene gy consump ion;
Remaining_Powe = Powe Gene a ed om Sola Panel - To al Maximum Powe Consumed (3)
Remaining_Powe = 3W – 2.675W = 0.325W
As can be seen in Equa ion (3), e en when he sys em consumes maximum powe in sunny
wea he , an ene gy su plus o 0.325 W is gene a ed. This ene gy su plus is s o ed using he
12V 6Ah ba e y in he sys em. Conside ing ha he sys em consumes maximum ene gy in 1
minu e, ope a es in sleep mode a o he imes, and ene gy consump ion dec eases om 1 o 1
in 10;
Sleep_mode_ene gy = Maximum_ene gy / 10 = 2.675W / 10 = 0.2675W (4)
Sleep_mode_ emaining_powe = Powe Gene a ed om Sola Panel - Sleep_mode_ene gy (5)
Sleep_mode_ emaining_powe = 3W – 0.2675W = 2.7325W
Ba e y_Wo king_Hou = Ba e y_Powe / To al Maximum Powe Consumed (6)
Ba e y_Wo king_Hou = 72W / 2.675W = 26.92 hou s
C. IOT OPERATIONS WITH ARTIFICIAL INTELLIGENCE ALGORITHMS
The main unc ion o he sys em is o be used in ag icul u e wi hou equi ing elec onic
knowledge, i.e., wi hou equi ing knowledge o echnical complexi y. I enables ag icul u al
enginee s, au ho ized o ganiza ions in ag icul u e, o a me s o p oduce hei own solu ions.
The sys em allows use s wi h so wa e knowledge o p og am and use i as hey wish. The
RP2040 mic ocon olle has a dual-co e ARM Co ex M0+ 133 MHZ p ocesso . This
p ocesso can be p og ammed wi h Mic oPy hon. In his way, ligh a i icial in elligence and
machine lea ning (ML) echniques o p e- ained and op imized hea y a i icial in elligence
models can be un on ou sys em. In addi ion, IOT p ojec s can be ealized by using Wi i,
GSM module, ZigBee, LoRa, and Blue oo h modules in he sys em.
A i icial in elligence models such as TinyML (Tiny Machine Lea ning), Rule-Based AI,
Decision T ees, and Simple Neu al Ne wo ks can be un in ou sys em. Some examples ha
can be done in ou sys em using ou p oposed Mic oPy hon and AI a e gi en in Table 2.
Table 2. Ou P oposed Mic oPy on+AI Ag icul u e P ojec Examples
Senso
Measu ed alue
Possible p ojec
Type o AI used
Explana ion
Soil Tempe a u e
Senso Soil Tempe a u e
Sma i iga ion and
plan heal h
moni o ing
Rule-based /
TinyML
I iga ion o disease
isk based on soil
empe a u e
Ai Mois u e Senso
+ Ai Tempe a u e
Senso
Ai Mois u e +
Tempe a u e
G eenhouse
clima e con ol and
i iga ion
scheduling
Rule-based /
Decision ee
High empe a u e +
low humidi y →
i iga ion
ecommenda ion
Ai P essu e Senso P essu e
Rain p edic ion and
wea he
moni o ing
TinyML (wea he
pa e n)
Rain p edic ion
based on p essu e
changes
Ambien Ligh
Amoun Senso Ligh Amoun
Pho osyn hesis
moni o ing and
i iga ion ime
scheduling
K-means /
Th esholding
Au oma ic con ol
based on
mo ning/e ening
ligh
Ai Quali y Senso Ai Quali y
Disease de ec ion
and g eenhouse ai
analysis
Anomaly de ec ion
(AI)
Risk en i onmen
de ec ion based on
gas inc eases
Soil Mois u e
Senso Soil Mois u e
Sma i iga ion
sys em and wa e
conse a ion
TinyML / KNN /
Rule-based
I he humidi y le el
d ops, AI calcula es
he i iga ion ime
