You are on page 1of 15

sensors

Article

A Reliable Wireless Control System for


Tomato Hydroponics
Hirofumi Ibayashi 1, *, Yukimasa Kaneda 2 , Jungo Imahara 3 , Naoki Oishi 3 , Masahiro Kuroda 4
and Hiroshi Mineno 1,5, *
1
2
3
4
5

Graduate School of Informatics, Shizuoka University, 3-5-1 Johoku, Naka-ku, Hamamatsu,


Shizuoka 432-8011, Japan
Graduate School of Integrated Science and Technology, Shizuoka University, 3-5-1 Johoku, Naka-ku,
Hamamatsu, Shizuoka 432-8011, Japan; kaneda@minelab.jp
Shizuoka Prefectural Research Institute of Agriculture and Forestry, 678-1 Tomioka, Iwata,
Shizuoka 438-0803, Japan; jungo1_imahara@pref.shizuoka.lg.jp (J.I.); naoki1.ohishi@gmail.com (N.O.)
National Institute of Information and Communications Technology, 4-2-1 Nukui-Kitamachi,
Koganei 184-8795, Japan; marsh@nict.go.jp
JST, PRESTO, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan
Correspondence: ibayashi@minelab.jp (H.I.); mineno@inf.shizuoka.ac.jp (H.M.); Tel.: +81-53-478-1491 (H.M.)

Academic Editor: Simon X. Yang


Received: 29 February 2016; Accepted: 30 April 2016; Published: 5 May 2016

Abstract: Agricultural systems using advanced information and communication (ICT) technology
can produce high-quality crops in a stable environment while decreasing the need for manual labor.
The system collects a wide variety of environmental data and provides the precise cultivation control
needed to produce high value-added crops; however, there are the problems of packet transmission
errors in wireless sensor networks or system failure due to having the equipment in a hot and
humid environment. In this paper, we propose a reliable wireless control system for hydroponic
tomato cultivation using the 400 MHz wireless band and the IEEE 802.15.6 standard. The 400 MHz
band, which is lower than the 2.4 GHz band, has good obstacle diffraction, and zero-data-loss
communication is realized using the guaranteed time-slot method supported by the IEEE 802.15.6
standard. In addition, this system has fault tolerance and a self-healing function to recover from
faults such as packet transmission failures due to deterioration of the wireless communication quality.
In our basic experiments, the 400 MHz band wireless communication was not affected by the plants
growth, and the packet error rate was less than that of the 2.4 GHz band. In summary, we achieved a
real-time hydroponic liquid supply control with no data loss by applying a 400 MHz band WSN to
hydroponic tomato cultivation.
Keywords: sensor network; actuator control; highly reliable wireless communication; agricultural
system; 400 MHz band

1. Introduction
Nowadays, numerous kinds of sensors are being developed for many different applications,
and many sensor-based control systems for agriculture have been implemented taking advantage
of advances in information and communication technology (ICT) [14]. These control systems work
automatically to control the growth of the plants instead of requiring the constant attention of farmers.
Recent control systems aim to produce high-quality vegetables and fruits by precise control. In Japan,
there is already a greenhouse environmental control system [5] using an ubiquitous environmental
control system (UECS) [6]. This system can collect a variety of environmental data by adding data
from any sensor conforming to UECS standards. However, wired communications is the mainstream
form of communication between sensors and sink node when communicating using UECS standards.
Sensors 2016, 16, 644; doi:10.3390/s16050644

www.mdpi.com/journal/sensors

Sensors 2016, 16, 644

2 of 15

Thus, there is the problem of the high cost burden on farmers due to the installation costs for items
such as wiring when installing a large number of sensors, although the communication reliability
is guaranteed.
Meanwhile, research on sensing the agricultural environments using wireless sensor networks
(WSNs) [710] is also being conducted. For example, there are some studies about sensing soil
moisture using a WSN [11] or irrigation control by sensing soil moisture and temperature, and air
temperature [12]. In particular, when installing a WSN in a horticulture environment, packet losses
may occur due to a decrease in receiving power caused not only by obstacles such as metal pipes and
the plants themselves, but also the installation position or characteristics of the radio antenna [13].
Moreover, the 2.4 GHz band has the property of being easily absorbed in water [14]. A reliable
environment monitoring system using a WSN aiming for at least 99.99% system availability has also
been promoted [15], but it is necessary to determine the reliability of the actuator control as well as
the sensing.
WSNs for agricultural applications generally use ZigBee [16], which is based on the layers
specified in the IEEE 802.15.4 wireless communication standard. This is because ZigBee has low power
consumption and is easy to operate. A study on improving the packet arrival rate by introducing
time-slotted or frequency division multiplexing [17] has been carried out, and its ideas have been
combined with studies on packet collision avoidance or the traffic distribution mechanism [18].
However, these studies mostly evaluate the results by simulation, and thus they do not take into
account the real environment where the wireless communication environment changes depending on
the amount of vegetation, climate, and installed products.
In recent years, there has been a focus on the use of lower frequency bands. By lowering the
frequency, the wavelength becomes longer, and the diffraction by obstacles is better. There is research
on potato field sensing using the 433 MHz band, where the radio propagation was good, even in high
humidity environments [19]. In Japan, there is the 429 MHz band, which is used for medical telemetry,
and there is the IEEE802.15.6 standard for using this frequency band [20]. The IEEE 802.15.6 standard
has reserved transmission slot (using time division multiple access or TDMA) and retransmission
processing functions. Therefore, it may be possible to build a highly reliable WSN using the 400 MHz
band and IEEE 802.15.6 standard, even when there are many obstacles.
In this paper, we propose a highly reliable wireless control system for hydroponic tomato
cultivation. We implemented a system prototype that can be reliably controlled using the 400 MHz
band and IEEE 802.15.6 standard. In addition, this system has high fault tolerance and self-healing
capability in response to faults. We operated the implemented system in a greenhouse environment
and confirmed whether the system can run without actuator control failure during the cultivation
period. Finally, we considered whether the wireless communication is affected by the plants.
2. Overview of the Tomato Hydroponics Control System
2.1. Overview of the Tomato Hydroponics Control System
The tomato hydroponics control system is a system prototype of a highly reliable wireless
environmental control system. Hydroponics is a kind of plant cultivation method without soil.
Hydroponic cultivation usually produces yield increases [21], but it needs precise control to produce
high quality crops. For example, there is a possibility that the size of the crop will be small or the
crop will die if the system supplies too little hydroponic liquid. On the other hand, if the system
supplies too much, it may produce unusually shaped crops. Therefore, reliable system operation and
hydroponic supply control is required.
An overview of our system is shown in Figure 1. Our tomato hydroponics control system
estimates the condition of the tomato plants in real-time using sensors, and the system then supplies
the appropriate amount of hydroponic liquid for the tomatoes. The amount of hydroponic liquid
supplied is determined by the evapotranspiration, which is calculated from the vapor pressure deficit

Sensors 2016,
2016, 16,
16, 644
Sensors
644

of 15
15
33 of

supplied is determined by the evapotranspiration, which is calculated from the vapor pressure
(VPD)
relative
light intensity
(RLI). Evapotranspiration
means means
the lossthe
of water
plant
deficit and
(VPD)
and relative
light intensity
(RLI). Evapotranspiration
loss ofby
water
byleaves.
plant
The
VPD
is
calculated
from
the
temperature
and
relative
humidity.
RLI
is
the
ratio
of
the
amount
of
leaves. The VPD is calculated from the temperature and relative humidity. RLI is the ratio of the
scattered
at the light
top and
bottom
of the
tomato
By installingBy
scattered
light
sensor
amount oflight
scattered
at the
top and
bottom
ofcommunities.
the tomato communities.
installing
scattered
nodes
at
the
top
and
bottom
of
the
tomatoes,
the
amount
of
transpiration
can
be
estimated
with
light sensor nodes at the top and bottom of the tomatoes, the amount of transpiration can an
be
error
of lesswith
thanan3%
fromofthe
of from
photosynthesis
and
using
data collected
by
estimated
error
lessamount
than 3%
the amount
of transpiration
photosynthesis
andthe
transpiration
using
the scattered
light by
sensors
[22].
data collected
the scattered
light sensors [22].

control system.
system.
Figure 1. Sensor layout of the tomato hydroponics control

2.2. System
Requirement Analysis
Analysis
2.2.
System Requirement
Our proposed
proposed system
system is
an environmental
environmental control
control system
Our
is an
system for
for the
the growth
growth control
control of
of tomatoes.
tomatoes.
The
architecture
of
our
proposed
system
is
shown
in
Figure
2.
Sensor
nodes
collect
greenhouse
The architecture of our proposed system is shown in Figure 2. Sensor nodes collect greenhouse
environmentdata
dataon
onthe
thestate
stateofofthe
thecrops.
crops.Then,
Then,
these
data
sent
to the
sensor
gateway
through
environment
these
data
areare
sent
to the
sensor
gateway
through
the
the
sink
node
that
aggregates
the
data
from
each
sensor
node.
The
sensor
gateway
determines
sink node that aggregates the data from each sensor node. The sensor gateway determines whether
whether
sensor
data is
correct,
then
sends data
the sensor
data to in
thethe
storage
thethe
cloud
and
the
sensorthe
data
is correct,
and
then itand
sends
theitsensor
to the storage
cloud in
and
actuator
the actuator
control
program.
Finally,
the system
controlsthat
thefeed
actuators
that feedliquid
the hydroponic
control
program.
Finally,
the system
controls
the actuators
the hydroponic
according
liquid
according
to
the
control
information.
We
use
wireless
sensor
nodes
to
facilitate
the
to the control information. We use wireless sensor nodes to facilitate the installation of the system
installation
of
the
system
and
reduce
wiring
costs.
The
control
information
is
input
by
the
user
who
and reduce wiring costs. The control information is input by the user who receives the environmental
receives
environmental
system
allows the user
toand
view
the control
environmental
datafrom
and
data.
Ourthe
system
allows the data.
user toOur
view
the environmental
data
input
information
input
control
information
from
anywhere
through
a
Web
browser.
anywhere through a Web browser.
Many growth
setset
in in
advance.
However,
the
Many
growth control
control systems
systemsare
arecontrolled
controlledbybytimes
timesthat
thatareare
advance.
However,
production
of
high-quality
crops
requires
a
control
system
based
on
real-time
environmental
the production of high-quality crops requires a control system based on real-time environmental
information because
because the
state of
of the
the greenhouse
greenhouse environment
environment and
and the
the growth
growth of
of plants
plants change
change with
with
information
the state
time. For
Forexample,
example,
evapotranspiration
per minute
conditions
several
time.
evapotranspiration
per minute
underunder
sunny sunny
daytimedaytime
conditions
is several is
milliliters,
milliliters,
so
the
system
has
to
be
able
to
irrigate
plants
based
on
real-time
evapotranspiration
the
so the system has to be able to irrigate plants based on real-time evapotranspiration of the plants of
every
plants
every
minute.
To
produce
high
quality
tomatoes,
it
is
necessary
to
manage
the
hydroponic
minute. To produce high quality tomatoes, it is necessary to manage the hydroponic liquid reliably
liquid
reliably Ifand
If the
and
control
the hydroponic
liquid supply,
and
delicately.
thedelicately.
system stops
andsystem
cannotstops
control
thecannot
hydroponic
liquid
supply, the tomato
harvest
the
tomato
harvest
and
quality
are
reduced.
As
the
proposed
system
is
a
feedback
control
system
and quality are reduced. As the proposed system is a feedback control system using sensor data,
if
using
sensor
data,
if
sensor
data
or
control
data
loss
occurs,
the
system
cannot
control
the
nutrient
sensor data or control data loss occurs, the system cannot control the nutrient supply, so it is necessary
supply,
it is necessary
reduce
packet losses as much as possible.
to
reducesopacket
losses asto
much
as possible.
In
crop
control
systems
it
is
necessary
to operate
the system
continuously
if the
user
In crop control systems it is necessary
to operate
the system
continuously
even ifeven
the user
cannot
cannot monitor
always monitor
theoperation
system operation
status
suchoperating
as when atoperating
at night.
Whenfails
the
always
the system
status such
as when
night. When
the system
system
fails
and
cannot
control
the
actuator
normally,
the
quality
of
crops
may
decrease.
Therefore,
and cannot control the actuator normally, the quality of crops may decrease. Therefore, it is necessary
it is
necessary
to operate
the systemduring
continuously
during the
cultivation period.
to
operate
the system
continuously
the cultivation
period.
C40
The greenhouse
greenhouseenvironment
environmentisishot
hot
and
humid;
maximum
temperature
can
or
The
and
humid;
the the
maximum
temperature
can be
40be
or C
more,
more,
and
the
relative
humidity
is
over
90%.
Therefore,
it
is
necessary
for
electronic
equipment
and the relative humidity is over 90%. Therefore, it is necessary for electronic equipment to haveto
a
have a waterproof
heat protection,
even
it is not
expected
to be used
outdoors.
waterproof
layer andlayer
heat and
protection,
even though
it isthough
not expected
to be
used outdoors.
Furthermore,
Furthermore,
the control
system the
cannot
control
the
actuator if the
environmental
lost because
the
system cannot
actuator
if the
environmental
data
is lost becausedata
this is
system
controlsthis
the
system controls the actuator in real time by using the sensor data. Therefore, for reliable actuator
control, it is necessary to suppress the data loss caused by not only wireless communication errors

Sensors 2016, 16, 644

4 of 15

Sensors 2016,
644time by using the sensor data. Therefore, for reliable actuator control, it is necessary
4 of to
15
actuator
in 16,
real
suppress the data loss caused by not only wireless communication errors in the WSN, but also buffer
in the WSN,
butreal-time
also buffer
real-time
data analysis
tier dueof
toprocessed
an increase
in per
the
overflow
in the
dataoverflow
analysis in
tierthe
due
to an increase
in the number
data
number
of
processed
data
per
unit
time.
unit time.

Figure 2. System architecture.

On the other hand, it is assumed that the users set the control information for the actuators
On the other hand, it is assumed that the users set the control information for the actuators
based on the collected environmental data. The users who can connect to the Internet can view the
based on the collected environmental data. The users who can connect to the Internet can view the
environmental data and change control information from any location. Therefore, users
environmental data and change control information from any location. Therefore, users convenience
convenience may be improved by managing these environmental data and actuator control
may be improved by managing these environmental data and actuator control information in a cloud
information in a cloud server. However, when the Internet fails, control signals do not reach the
server. However, when the Internet fails, control signals do not reach the control program, and the
control program, and the system cannot control the actuators. Also, if the environmental data is
system cannot control the actuators. Also, if the environmental data is only stored in the cloud server
only stored in the cloud server database management system (DBMS), data loss will occur when the
database management system (DBMS), data loss will occur when the system cannot connect to the
system cannot connect to the Internet. Note that the cloud server is in our laboratory for easy
Internet. Note that the cloud server is in our laboratory for easy maintenance at this time, and the server
maintenance at this time, and the server is accessed through the Internet by using the portable
is accessed through the Internet by using the portable 3G/4G wireless router at the greenhouse located
3G/4G wireless router at the greenhouse located about 10 km away from our campus. Therefore, it
about 10 km away from our campus. Therefore, it is necessary that the data not be lost even if failure
is necessary that the data not be lost even if failure occurs in the uplink and downlink direction of
occurs in the uplink and downlink direction of the communication path and cloud servers to achieve
the communication path and cloud servers to achieve reliable environmental control by backup
reliable environmental control by backup storage in the local network. The required specifications for
storage in the local network. The required specifications for a reliable wireless environmental
a reliable wireless environmental control system in this study are summarized below:
control system in this study are summarized below:
1.
1.

2.
2.

3.

By using
number
of sensors
to collect
data in thedata
greenhouse
By
usinga large
a large
number
of sensors
to environmental
collect environmental
in the environment,
greenhouse
the collected environmental data is stored in the cloud server, and it is possible to browse the
environment, the collected environmental data is stored in the cloud server, and it is possible
data from any location through a visualization interface.
to browse the data from any location through a visualization interface.
The environmental control system in the field environment can be configured by system users via
The environmental control system in the field environment can be configured by system users
the control information setting interface from any location, and actuator control can be performed
via the control information setting interface from any location, and actuator control can be
based on the control information set.
performed based on the control information set.
The system requires that a system failure hardly ever occurs, and it has a self-healing function
in the event of failure to prevent the crop being affected by the system stop.

Sensors 2016, 16, 644

5 of 15

Sensors
Sensors2016,
2016,16,
16,644
644

5 5ofof1515

3.
The system requires that a system failure hardly ever occurs, and it has a self-healing function in
2.3.
Reliable
Wireless
Sensor
2.3.AA400
400
MHzHighly
Highly
Reliable
Wireless
Sensor
Network
theMHz
event
of failure
to prevent
the
cropNetwork
being
affected by the system stop.

The
The400
400MHz
MHzhighly
highlyreliable
reliablewireless
wirelesssensor
sensornetwork
networkisiscomposed
composedofofsome
somesensor
sensornodes
nodesand
anda a
2.3. A 400 MHz Highly Reliable Wireless Sensor Network
sink
sinknode.
node.This
Thissatisfies
satisfiesrequirements
requirements1 1and
and3.3.InInthe
thetomato
tomatohydroponics
hydroponicscontrol
controlsystem,
system,we
weused
used
The
400
MHz
highly
reliable
wireless
sensor
network
is
composed
of
some
sensor
nodes
temperature,
temperature,humidity,
humidity,and
andlight
lightintensity
intensitysensors
sensorstotoestimate
estimatehydroponic
hydroponicliquid
liquidsupply.
supply.The
The
and a of
sink
node. This
satisfies
requirements
1 and
Inbasis
the of
tomato
hydroponics control
system,
amount
liquid
supply
isisdetermined
on
the
and
the
amount
ofhydroponic
hydroponic
liquid
supply
determined
on3.
the
basis
ofthe
theevapotranspiration,
evapotranspiration,
and
the
we used temperature,
humidity,from
and the
light
intensity
sensors
to estimateby
hydroponic
liquid supply.
evapotranspiration
isiscalculated
isisdetermined
the
and
evapotranspiration
calculated
from
theVPD,
VPD,which
which
determined
by
thetemperature
temperature
and
The amount
of hydroponic
liquid
supply
is
determined
on
the
basis of lighting.
the
evapotranspiration,
the
relative
humidity,
and
which
isisdetermined
by
the
scattered
Therefore,
used
relative
humidity,
andthe
theRLI,
RLI,
which
determined
by
the
scattered
lighting.
Therefore,we
weand
used
evapotranspiration
is nodes,
calculated
from
theinin
VPD,
which
determined
by the
temperature data.
and
relative
scattered
light
Figure
3a,
collect
environmental
The
scattered
lightsensor
sensor
nodes,asasshown
shown
Figure
3a,istoto
collectthese
these
environmental
data.
The
humidity,light
and the
RLI, which
determined
the scattered
lighting.
Therefore,
we used scattered
light
scattered
sensor
node
has
two
photodiode
sensors
(S1133,
Hamamatsu
Photonics,
scattered
light
sensor
nodeis
has
two SiSi by
photodiode
sensors
(S1133,
Hamamatsu
Photonics,
sensor
nodes,
as
shown
in
Figure
3a,
to
collect
these
environmental
data.
The
scattered
light
sensor
Hamamatsu,
Hamamatsu, Japan),
Japan), a a SiSi PIN
PIN photo
photo diode
diode sensor
sensor (S2506,
(S2506, Hamamatsu
Hamamatsu Photonics),
Photonics), and
and a a
node has two
Sihumidity
photodiode
sensors
(S1133,
Hamamatsu
Photonics,
Hamamatsu,
Japan),
a Si3b.
PIN
temperature
and
sensor
(SHT21,
Sensirion,
Zurich,
Switzerland)
asasshown
ininFigure
temperature
and
humidity
sensor
(SHT21,
Sensirion,
Zurich,
Switzerland)
shown
Figure
3b.
photo
diode
sensor
(S2506,
Hamamatsu
Photonics),
and
a
temperature
and
humidity
sensor
(SHT21,
One
Oneofofthe
theSiSiphotodiode
photodiodesensors
sensorsisisattached
attachedtotothe
thetop
topofofthe
thebox
boxand
andmeasures
measuresthe
theamount
amountofof
Sensirion,
Zurich,
Switzerland)
as shown
in Figuresensor
3b.
One
the Si photodiode
sensors
is attached
outdoor
solar
radiation.
The
SiSiphotodiode
isisof
attached
totothe
ofofthe
box
outdoor
solar
radiation.
Theother
other
photodiode
sensor
attached
theinside
inside
the
boxand
andto
the top ofthe
thestrength
box
and measures
amountinof
outdoor
radiation.
The
other aSiawider
photodiode
sensor
measures
ofoflight
scattered
Moreover,
totomeasure
ofof
measures
the
strength
lightthe
scattered
inthe
theair.
air.solar
Moreover,
measure
widerrange
range
is
attached
to
the
inside
of
the
box
and
measures
the
strength
of
light
scattered
in
the
air.
Moreover,
wavelength,
wavelength,we
weused
useda aSiSiPIN
PINphotodiode
photodiodesensor
sensor[23].
[23].The
TheSiSiphotodiode
photodiodesensors
sensorscan
canmeasure
measureto
measure a wider
of
wavelength,
we while
used
a the
Sithe
PIN
sensor [23].
Thecan
Si photodiode
wavelengths
from
320
toto 1000
SiSiphotodiode
PIN
sensor
measure
wavelengths
fromrange
320 nm
nm
1000 nm,
nm,
while
PIN photodiode
photodiode
sensor
can
measure
sensors canfrom
measure
wavelengths
from 320 nm to 1000 nm, while the Si PIN photodiode sensor can
wavelengths
wavelengths
from760
760nm
nmtoto1100
1100nm.
nm.
measure wavelengths from 760 nm to 1100 nm.

(b)
(b)

(a)
(a)

Figure
Figure3.3.Scattered
Scatteredlight
lightsensor
sensornode.
node.(a)
(a)photograph;
photograph;(b)
(b)overview.
overview.
Figure 3. Scattered light sensor node. (a) photograph; (b) overview.

InInaddition,
addition,we
weprepared
prepareda atemperature
temperatureand
andhumidity
humiditysensor
sensornode
nodethat
thatcan
canchange
changethe
the
Inmodule
addition,
weevaluation.
prepared aAlthough
temperature
and
humidity
node
that
can
change
the
wireless
wireless
sensor
node
isisnot
ititisis
used
for
wireless
modulefor
for
evaluation.
Althoughthis
this
sensor
nodesensor
notused
usedfor
forcontrol,
control,
used
for
module
for
evaluation.
Although
this
sensor
node
is
not
used
for
control,
it
is
used
for
comparison
comparison
comparisonofofthe
theradio
radiofrequency
frequencyand
andcommunication
communicationsystem.
system.This
Thissensor
sensornode
nodeisisattached
attachedtotoof
the
radio
frequency
and
communication
system.
This
sensor
node
is
attached
to
a
wireless
module
a a wireless
wireless module
module (Figure
(Figure 4a)
4a) using
using the
the 2.4
2.4 GHz
GHz band
band and
and IEEE
IEEE 802.15.4-2006
802.15.4-2006 standard
standard
(Figure
4a)
using
the
2.4
GHz
band
and
IEEE
802.15.4-2006
standard
(Figure
4b)
[24].
(Figure
(Figure4b)
4b)[24].
[24].

(a)
(a)

(b)
(b)

(c)
(c)

Figure
Figure4.4.(a)
(a)Temperature
Temperatureand
andhumidity
humiditysensor
sensornode,
node,(b)
(b)2.4
2.4GHz
GHzband
bandwireless
wirelessmodule,
module,and
and
Figure 4. (a) Temperature and humidity sensor node, (b) 2.4 GHz band wireless module, and (c) 400
(c)(c)400
400MHz
MHzband
bandwireless
wirelessmodule.
module.
MHz band wireless module.

This
Thiswireless
wirelessmodule
moduleisisconnected
connectedtotothe
thesensor
sensornode
nodewith
witha ageneric
genericconnector,
connector,and
andthe
the
This
wireless
module
is
connected
to
the
sensor
node
with
a
generic
connector,
and
the
communication
communicationsystem
systembetween
betweenthe
thesensor
sensornode
nodeand
andthe
thewireless
wirelessmodule
moduleuses
usesthe
theUART
UARTprotocol.
protocol.
communication
system abetween
thethat
sensor
and the wireless
module
uses
UARTmodule
protocol.
Therefore,
we
can
bebeconnected
totothe
MHz
wireless
Therefore,
wedeveloped
developed
asubstrate
substrate
that
cannode
connected
the400
400
MHzband
bandthe
wireless
module
Therefore,
we
developed
a
substrate
that
can
be
connected
to
the
400
MHz
band
wireless
module
(Figure
GHz
(Figure4c).
4c).The
Thespecifications
specificationsofofeach
eachwireless
wirelessmodule
moduleare
aregiven
givenininTable
Table1.1.IfIfwe
wecould
coulduse
use2.4
2.4
GHz
wireless
wirelessmodules
modulesimplemented
implementedbased
basedon
onthe
thesame
samestandard
standardasas400
400MHz
MHzwireless
wirelessmodules,
modules,we
we
could
couldonly
onlyevaluate
evaluatethe
theeffect
effectdue
duetotothe
thedifference
differenceininthe
thefrequency.
frequency.

Sensors 2016, 16, 644

6 of 15

(Figure 4c). The specifications of each wireless module are given in Table 1. If we could use 2.4 GHz
wireless modules implemented based on the same standard as 400 MHz wireless modules, we could
only evaluate the effect due to the difference in the frequency.
Table 1. Specifications of the wireless module.
Specification
Chip
Frequency
(Bandwidth)
Protocol
Bit rate
Transmit power
Power supply

2.4 GHz Band Wireless Module

400 MHz Band Wireless Module

CC2420
(Texas Instruments)
24052480 MHz
(5 MHz interval 16ch)
Based on IEEE 802.15.4 standard
250 kbps
10 mW
3-volt DC

CC430F5137
(Texas Instruments)
429.25429.75 MHz
(12.5 kHz interval 40ch)
Based on IEEE 802.15.6 standard
7200 bps
10 mW
3-volt DC

Since the proposed system is a feedback control system using sensor data, it is necessary to collect
data without losses to ensure actuator control. Therefore, we introduced the 400 MHz band under
the ARIB T67 radio standard [25] for highly reliable wireless communication. The 400 MHz band has
high diffraction from obstacles because the wavelength of the 400 MHz band is longer than that of
the 2.4 GHz band. Thus, it assumes that reliable wireless communication can be realized even if the
sensor nodes are in the plants. On the other hand, there is the disadvantage that the bit rate is low
because the bandwidth of the 400 MHz band is narrower than that of the 2.4 GHz band. Hence, it is
not suitable for the transfer of large size data such as pictures and videos, but we think that it is not a
big problem because we assumed the system would deal with only a few bytes of sensor data.
Additionally, this wireless module is provided with two channel access mechanisms, random
access with carrier sense multiple access/collision avoidance (CSMA/CA) and slotted access with
TDMA, to accommodate many sensor nodes quickly and ensure reliable packet delivery. The state
transition diagram of a sensor node is shown in Figure 5a, and a sink node is shown in Figure 5b [26].
First, the sink node is in the Random access accept state when it is powered on. When a sensor node
is powered on and joined into a WSN, it uses CSMA/CA in the Random access state. If the channel
is busy or a collision occurs, the node waits for a constant time. Then all sensor nodes are joined into
the WSN, and each sensor node sends a packet with TDMA in the Slotted access state. As all sensor
nodes send or receive the packets in assigned time slots, the data transmission is guaranteed to be
collision-free in a star topology. Our system does not assume the situation in which each sensor node
communicates with the others. Therefore, we used the star topology with TDMA for easy maintenance.
Although there are wireless modules with time slotted channel hopping (TSCH) mesh networking
technologies, it was very difficult to analyze the details of characteristics compared with our 400 MHz
wireless modules.
If the packet does not reach the sink node or an ACK does not reach the sensor node, the sensor
node transitions to the Transfer error state. In the Transfer error state, the sensor node resends the
packet up to nine times in our prototype implementation. If retransmission is successful, the sensor
node goes back to the Slotted access state. If the packet does not succeed after nine retransmissions,
the sensor nodes transition to the Random access state via the Slotted access recovery state and
Ready to join state. At the same time, the sink node does not receive the packet from a sensor node
for a certain time; in other words, if after nine retransmissions the message does not reach the sink
node, the sink node regards the sensor node as a breakaway node based on the configuration file.
Then, if the number of the breakaway nodes exceeds the threshold of the configuration file, the sink
node transitions to the Random access accept state via the Ready for random access state. Finally,
when the sensor nodes and the sink node transition to the Random access state at the same time, the
WSN is rebuilt. This failure recovery is performed automatically on the WSN side. For this reason, the
system user does not need to recover the system.

Sensors 2016, 16, 644

7 of 15

Sensors 2016, 16, 644


Sensors 2016, 16, 644

7 of 15
7 of 15

IDLE
Power on

IDLE

Rejoin
Rejoin
Receive
time sync.
Receive
time sync.

Random
Power on
access accept
state
Random
Ready for slotted access
access(>connectionWaiteNodeCount)
accept
state
Receive
Ready for slotted
access
Slotted
data
(>connectionWaiteNodeCount)
access state
Few nodes are joined

Receive
into WSN
Slotted
data
nodes
access(Current
state slotted access
< resetThreshouldNodeCount)
Few nodes are joined

Ready
into WSN
for random
(Current slotted access nodes
access< resetThreshouldNodeCount)
Delay wait
Ready
(Delay time
for random
< connectionWaitTimeout)
access

SinkDelay
node wait

(a)

(Delay time
< connectionWaitTimeout)

(b)

Sink node

Figure(a)
5. State diagram: (a) sensor node; (b) sink(b)
node.
Figure 5. State diagram: (a) sensor node; (b) sink node.
2.4. Real-Time Data Analysis
Figure 5. State diagram: (a) sensor node; (b) sink node.

The real-time
data
analysis tier includes two main programs: the sensor gateway program and
2.4. Real-Time
DataAnalysis
Analysis
2.4.
Data
the Real-Time
actuator control
program. Input and output between each program are performed by Internet
Thereal-time
real-time
data
analysis
tierincludes
includesTherefore,
twomain
mainprograms:
programs:
thesensor
sensor
gatewayprogram
program
and
protocol
suite ordata
serial
communication.
by increasing
the gateway
number
of modules
The
analysis
tier
two
the
and
the
actuator
control
program.
Input
and
output
between
each
program
are
performed
by
Internet
independent
of each
program,
it isand
possible
easily deal
with sensor
network changes
and
the
actuator control
program.
Input
outputtobetween
each program
are performed
by Internet
protocol
suite
or
serial
communication.
Therefore,
by
increasing
the
number
of
modules
control algorithms.
protocol
suite or serial communication. Therefore, by increasing the number of modules independent
program,
it is deal
possible
to easily
deal
with
sensor
network
changes
and
A program,
sensor of
gateway
verifies
whether
the
packet
from
the sensor
nodes
has been
broken and
ofindependent
each
iteach
is possible
to easily
with
sensor
network
changes
and control
algorithms.
control
algorithms.
registers
the sensor
data
in the
storagethe
onpacket
the cloud.
It also
sends
thehas
data
to broken
the actuator
control
A sensor
gateway
verifies
whether
from the
sensor
nodes
been
and registers
A
sensor
gateway
verifies
whether
the
packet
from
the
sensor
nodes
has
been
broken
and
program
user
(UDP).ItThe
control
program
analyzes
the program
sensor data
the
sensorby
data
in datagram
the storageprotocol
on the cloud.
alsoactuator
sends the
data to
the actuator
control
by
registers
the
sensor
data
in
the
storage
on
the
cloud.
It
also
sends
the
data
to
the
actuator
control
and datagram
estimates the
current
stateThe
of the
plant.control
In this program
system, the
actuator
program
estimates
user
protocol
(UDP).
actuator
analyzes
thecontrol
sensor data
and estimates
program
userof
datagram
protocol
The
actuator
control
program
analyzes
the sensor
data
the
plantsbystate
evapotranspiration
minute.
Ifthe
theactuator
accumulated
evapotranspiration
exceeds
a
the
current
the plant.
Inper
this(UDP).
system,
control
program
estimates
the
plants
and
estimates
the
current
state
of
the
plant.
In
this
system,
the
actuator
control
program
estimates
threshold set by the
theIfactuator
control program
sends a signal
to athe
infra-red
evapotranspiration
perusers,
minute.
the accumulated
evapotranspiration
exceeds
threshold
setremote
by the
the
plants
evapotranspiration
per
minute.
If
the
accumulated
evapotranspiration
exceeds
control
to
run
the
pump
that
supplies
the
hydroponic
liquid.
This
algorithm
is
shown
in
Figure
6.a
users, the actuator control program sends a signal to the infra-red remote control to run the pump that
threshold
by the users,
actuator
control
program
sends6.aThis
signal
to therequirement
infra-red remote
This
satisfies
requirement
2. theThis
supplies
theset
hydroponic
liquid.
algorithm
is shown
in Figure
satisfies
2.
control to run the pump that supplies the hydroponic liquid. This algorithm is shown in Figure 6.
This satisfies requirement 2.

Figure 6.
6. Algorithm
Algorithm for
for controlling
controlling the
the hydroponic
hydroponic liquid
liquid supply.
supply.
Figure
Figure 6. Algorithm for controlling the hydroponic liquid supply.

A real-time data analysis tier is put in the greenhouse environment. This environment will be
hot and humid to help the plants to grow. Therefore, we adopted OpenBlocks A7 (OBSA7P/J7,
PlatHome, Tokyo, Japan), guaranteed to operate up to 55 C as a microserver to run both programs.
Furthermore, the sensor gateway program and control decision program not only run with an
executable format, but also minimize other unnecessary processes in system operation and reduce
Sensors 2016, 16, 644
8 of 15
CPU and memory usage to suppress the occurrence of thermal runaways and reduce the risk
of failure.

ASensors
real-time
data analysis tier is put in the greenhouse environment. This environment will 8beof hot
2016,
16, 644
15
3. System
Evaluation
and humid to help the plants to grow. Therefore, we adopted OpenBlocks A7 (OBSA7P/J7, PlatHome,
Tokyo, Japan),
guaranteed
operate
55in Cthe
asgreenhouse
a microserver
to run bothThis
programs.
Furthermore,
A Experiment
real-time
data to
analysis
tierup
is to
put
environment.
environment
will be
3.1. Basic
hot and
humid
to help and
the control
plants to
grow. program
Therefore,not
weonly
adopted
OpenBlocks
A7 (OBSA7P/J7,
the sensor
gateway
program
decision
run with
an executable
format, but
First, we
performed
a basic
experiment
to compare
We placed
temperature
PlatHome,
Tokyo,
Japan),
guaranteed
to
up
to wireless
55 Cand
as modules.
areduce
microserver
to
run
both
programs.
also minimize
other
unnecessary
processes
in operate
system
operation
CPU
and
memory
usage to
and
humidity sensor
nodes
in a program
greenhouse
atcontrol
Shizuoka
Prefectural
Research
Institute
ofan
Furthermore,
the
sensor
gateway
and
decision
program
not
only
run
with
suppress the occurrence of thermal runaways and reduce the risk of failure.
Agriculture
Forestry
Japan). The
nodes were
placedinatsystem
four different
distances
at
executableand
format,
but (Iwata,
also minimize
othersensor
unnecessary
processes
operation
and reduce
three
heights,
and
they
were
respectively
equipped
with
two
wireless
modules
in
each
place:
the
CPU Evaluation
and memory usage to suppress the occurrence of thermal runaways and reduce the risk
3. System
2.4ofGHz
band and the 400 MHz band as shown in Figure 7.
failure.
installed the sensor nodes on four places: the nearest place from the sink nodes antenna
3.1. BasicWe
Experiment
(A),
the
middle
of each community (B) and (C) where the scattered light sensor nodes are installed
3. System Evaluation
First, we
experiment
compare
modules.
placed(D).The
temperature
during
theperformed
operation,a basic
and the
farthest to
place
fromwireless
the sink
nodes We
antenna
packetand
humidity
sensor
nodes in
a greenhouse
Shizuokaperiod
Prefectural
Research
Institute
3.1. Basic
Experiment
transmission
interval
was
60 s, and the at
evaluation
was from
20 March
2015 toof
10Agriculture
June 2015. Inand
Forestry
(Iwata,
Japan).
The
sensor
nodes
were
placed
at
four
different
distances
at
three
heights,inand
this period, there were plants in the greenhouse from 20 March 2015 to 31 May 2015 as shown
First, we performed a basic experiment to compare wireless modules. We placed temperature
theyFigure
were respectively
equipped
with
two
wireless
modules
in
each
place:
the
2.4
GHz
band
8a, while there were no plants from 1 June 2015 to 10 June 2015 as shown in Figureand
8b. the
and humidity sensor nodes in a greenhouse at Shizuoka Prefectural Research Institute of
400 Sensor
MHz band
shown
Figure 7.
nodesasare
put inin
radiation
shields to avoid failure during pesticide spraying.
Agriculture and Forestry (Iwata, Japan). The sensor nodes were placed at four different distances at
three heights, and they were respectively equipped with two wireless modules in each place: the
2.4 GHz band and the 400 MHz band as shown in Figure 7.
We installed the sensor nodes on four places: the nearest place from the sink nodes antenna
(A), the middle of each community (B) and (C) where the scattered light sensor nodes are installed
during the operation, and the farthest place from the sink nodes antenna (D).The packet
transmission interval was 60 s, and the evaluation period was from 20 March 2015 to 10 June 2015. In
this period, there were plants in the greenhouse from 20 March 2015 to 31 May 2015 as shown in
Figure 8a, while there were no plants from 1 June 2015 to 10 June 2015 as shown in Figure 8b.
Sensor nodes are put in radiation shields to avoid failure during pesticide spraying.
Figure 7. Sensor position for the basic experiment.

Figure 7. Sensor position for the basic experiment.

We installed the sensor nodes on four places: the nearest place from the sink nodes antenna (A),
the middle of each community (B) and (C) where the scattered light sensor nodes are installed during
the operation, and the farthest place from the sink nodes antenna (D).The packet transmission interval
was 60 s, and the evaluation period was from 20 March 2015 to 10 June 2015. In this period, there were
plants in the greenhouse from 20 March 2015 to 31 May 2015 as shown in Figure 8a, while there were
no plants from 1 June 2015 to 10 June 2015 as shown in Figure 8b. Sensor nodes are put in radiation
shields to avoid failure during pesticide
spraying.
Figure 7. Sensor
position for the basic experiment.

(a)

(b)

Figure 8. (a) Greenhouse environment with lush tomatoes; (b) Greenhouse environment without
lush tomatoes.

(a)

(b)

FigureFigure
8. (a)8.Greenhouse
environment
with
lush
tomatoes;
without
(a) Greenhouse
environment
with
lush
tomatoes;(b)
(b)Greenhouse
Greenhouse environment
environment without
lush tomatoes.
lush tomatoes.

following equation:
PER [%] =

Number of packet errors


100
Number of received packets + Number of packet errors

(1)

The 400 MHz band wireless module has a retransmission function, but the 2.4 GHz band
9 of 15
wireless module has no retransmission function. Therefore, for the sake of fairness packet errors
include packet retransmission, which means that the packet was not received at once in this
experiment.
The metrics of this experiment uses the packet error rate (PER). The PER is calculated by the
Figure
9 shows the wireless module comparison in a basic experiment. The blue bars mean the
following
equation:
2.4 GHz band wireless module, and the red bars mean the 400 MHz band wireless module. Besides,
Number of packet errors
the full bars mean
there
This graph shows
PER r%s
are plants, and the empty bars mean there are no plants.
100
(1)
Number
of
received
` Number
of packet
that the 2.4 GHz band wireless module with packets
lush tomatoes
is where
the errors
most packet loss occurred.
EachThe
sensor
node band
sent wireless
approximately
million
packets during
the but
cultivation
period,
about
400 MHz
module11
has
a retransmission
function,
the 2.4 GHz
bandbut
wireless
1000
to
3000
packet
units
are
wireless
communication
errors.
This
means
that
the
PER
is
1.0
to
module has no retransmission function. Therefore, for the sake of fairness packet errors include packet
3.0
percent
with
the
2.4
GHz
band
wireless
module.
retransmission, which means that the packet was not received at once in this experiment.
On the9other
the 400 module
MHz band
wireless in
module
fewer packets
evenbars
though
Figure
showshand,
the wireless
comparison
a basichad
experiment.
The blue
meanthere
the
were
plants.
The
number
of
packets
lost
was
four,
even
in
the
sensor
node
with
the
most
packet
2.4 GHz band wireless module, and the red bars mean the 400 MHz band wireless module. Besides,
errors. Only the B-1 node of the 400 MHz band without plants showed an increased packet error;
the full bars mean there are plants, and the empty bars mean there are no plants. This graph shows
however, the other nodes had fewer packet errors than when using the 2.4 GHz band even when
that the 2.4 GHz band wireless module with lush tomatoes is where the most packet loss occurred.
considering packet resend. The PER of the 400 MHz band with the method of IEEE 802.15.6
Each sensor node sent approximately 11 million packets during the cultivation period, but about 1000
standard is 0.2 percent at maximum. From this result, the 400 MHz band with the method of IEEE
to 3000 packet units are wireless communication errors. This means that the PER is 1.0 to 3.0 percent
802.15.6 standard is suitable in a greenhouse where hydroponics methods are used.
with the 2.4 GHz band wireless module.
Sensors 2016, 16, 644

3.5
3.0

PER [%]

2.5
2.0
1.5
1.0
0.5
0.0
A-1

A-2

A-3

B-1

B-2

B-3

C-1

C-2

C-3

D-1

D-2

D-3

Sensor Position
2.4 GHz band wireless module with lush tomatoes (3/20-5/31)
2.4 GHz band wireless module without lush tomatoes (6/1-6/10)
400 MHz band wireless module with lush tomatoes (3/20-5/31)
400 MHz band wireless module without lush tomatoes (6/1-6/10)

Figure 9.
9. PER
PER result
result for
for each
each sensor.
sensor.
Figure

We analyzed the packet error at the 2.4 GHz band wireless module in detail. Figure 10a shows
On the other hand, the 400 MHz band wireless module had fewer packets even though there were
the PER, and Figure 10b shows the leaf area index (LAI) transition period-to-period for each
plants. The number of packets lost was four, even in the sensor node with the most packet errors. Only
module. The LAI is an index indicating the degree to which plants grow leaves and can be
the B-1 node of the 400 MHz band without plants showed an increased packet error; however, the
indirectly estimated based on RLI because the LAI is strongly inversely proportional to the RLI.
other nodes had fewer packet errors than when using the 2.4 GHz band even when considering packet
Therefore, the lusher the leaves of the plants are the larger the LAI is, and then the line of sight
resend. The PER of the 400 MHz band with the method of IEEE 802.15.6 standard is 0.2 percent at
(LoS) is not ensured. In the PER graph in Figure 10a, for all sensor nodes except A-2, the most
maximum. From this result, the 400 MHz band with the method of IEEE 802.15.6 standard is suitable
packet loss occurred from 21 May 2015 to 31 May 2015 when the PER is over 4 percent. Looking at
in a greenhouse where hydroponics methods are used.
We analyzed the packet error at the 2.4 GHz band wireless module in detail. Figure 10a shows
the PER, and Figure 10b shows the leaf area index (LAI) transition period-to-period for each module.
The LAI is an index indicating the degree to which plants grow leaves and can be indirectly estimated
based on RLI because the LAI is strongly inversely proportional to the RLI. Therefore, the lusher the
leaves of the plants are the larger the LAI is, and then the line of sight (LoS) is not ensured. In the
PER graph in Figure 10a, for all sensor nodes except A-2, the most packet loss occurred from 21 May

Sensors 2016, 16, 644

10 of 15

Sensors 2016, 16, 644

10 of 15

2015 to 31 May 2015 when the PER is over 4 percent. Looking at this period, the LAI has not changed
compared
to the
On thecompared
other hand,
increased
or decreased
during
the the
period
this period,
theprevious
LAI has period.
not changed
to the
thePER
previous
period.
On the other
hand,
PER
ofincreased
changingor
LAI
(20 March
2015 the
to 21
May 2015).
decreased
during
period
of changing LAI (20 March 2015 to 21 May 2015).

(a)

(b)
Figure10.
10.(a)
(a)PER
PERtransition
transitionperiod-to-period;
period-to-period;(b)
(b)LAI
LAItransition
transitionperiod-to-period.
period-to-period.
Figure

It is assumed that the environment changes in the path of moving leaves, and the link quality
It is assumed that the environment changes in the path of moving leaves, and the link quality of
of wireless communication is also changed when the LAI increases. When the leaves move due to
wireless communication is also changed when the LAI increases. When the leaves move due to growth,
growth, it is considered that the radio wave absorption or the radio wave reflection changes. On the
it is considered that the radio wave absorption or the radio wave reflection changes. On the other
other hand, when the LAI is not changed because plant growth was completed, the environment
hand, when the LAI is not changed because plant growth was completed, the environment does not
does not change due to the effect of the leaves. Since the period of 21 May 2015 to 31 May 2015 has
change
due toLAI
the and
effectthe
of LoS
the leaves.
Since theinperiod
of 21itMay
2015 to that
31 May
has
the highest
the highest
is not ensured
the path,
is thought
this2015
is the
reason
for the
LAI
and
the
LoS
is
not
ensured
in
the
path,
it
is
thought
that
this
is
the
reason
for
the
occurrence
of
the
occurrence of the largest number of packet errors.
largestInnumber
of
packet
errors.
addition, we had a detailed analysis of the received packets from sensor nodes and packet
In
addition,
had a detailed
analysis
of thepacket
received
packets
from sensor
nodes
and packet
errors at sensorwe
position
D-3 where
the most
errors
occurred
from the
received
signal
errors
at sensor
position
D-3data.
where
the most
packet
errors
from the
received
signal
strength
strength
indicator
(RSSI)
Figure
11a is
the 2.4
GHzoccurred
band wireless
module,
and
Figure
11b is
indicator
Figure
11aisis lower
the 2.4 in
GHz
module,
and
Figure 11bthe
is the
MHz
the 400 (RSSI)
MHz data.
one. The
RSSI
theband
400 wireless
MHz band
when
comparing
two400
wireless
one.
The RSSI
is lower
the 400 MHz
band when
comparing
whichwhether
is the
modules,
which
is theindifference
in antenna
gain.
Therefore,the
in two
this wireless
section, modules,
we compared
difference
in plants
antenna
Therefore,
in this section, we compared whether there were plants in each
there were
in gain.
each wireless
module.
wireless
module.
The average RSSI of 2.4 GHz increased after the plants were removed. Besides, the change in
The
of 2.4
2.4 GHz
GHz band
increased
aftermodule
the plants
werewhen
removed.
Besides,
the
in
width ofaverage
the RSSIRSSI
in the
wireless
is large
there are
plants.
Inchange
particular,
width
of
the
RSSI
in
the
2.4
GHz
band
wireless
module
is
large
when
there
are
plants.
In
particular,
a lot of packet errors occurred when the RSSI fell to 90 dBm. Since such a RSSI decline is not seen
a when
lot of packet
errors
when
the means
RSSI fell
to 90
dBm.
Since
such
a RSSImodule
declinewas
is not
seen
there were
nooccurred
plants, this
result
that
the 2.4
GHz
band
wireless
affected
when
there
were no plants, this result means that the 2.4 GHz band wireless module was affected by
by the
plants.
the plants.
However, the RSSI of the 400 MHz band wireless module did not change much irrespective of
whether there were any plants. The packet error (retransmission) in this period is twice per 30 min
at most, despite this node being the farthest node from the sink node. Accordingly, it was found
that the 400 MHz band is suitable in greenhouse environments with many plants.

2.4 GHz wireless modules, the antenna is deployed on the outside of the wireless module shown in
Figure 4b. On the other hand, the antenna of 429 MHz wireless modules is a chip antenna located
on the wireless module. Therefore, we considered that the antenna gain of our 2.4 GHz wireless
module is higher than that of the 429 MHz wireless module. We think our experimental results show
the
usefulness of 429 MHz wireless communication because 429 MHz wireless communications11have
Sensors 2016, 16, 644
of 15
higher diffraction performance despite their lower antenna gain.

(a)

(b)

Figure 11. RSSI variation and packet loss with or without plants: (a) 2.4 GHz band; (b) 400 MHz band.
Figure 11. RSSI variation and packet loss with or without plants: (a) 2.4 GHz band; (b) 400 MHz band.

3.2. System Evaluation


However, the RSSI of the 400 MHz band wireless module did not change much irrespective of
As confirmed
the
effectThe
of the
400 error
MHz (retransmission)
band with the IEEE
802.15.6
method,
whether
there wereby
any
plants.
packet
in this
periodstandard
is twice per
30 minwe
at
have
conducted
an
evaluation
experiment
on
the
tomato
hydroponics
system.
A
set
of
scattered
most, despite this node being the farthest node from the sink node. Accordingly, it was found that the
light
sensor
nodes
was placed
in each plant
community
as shown
in Figure 12. We run this system
400 MHz
band
is suitable
in greenhouse
environments
with
many plants.
during
the
tomato
cultivation
period
of
about
3
months.
In Figure 11b, it can be seen that variation in RSSI occurred with or without plants. This is
considered to give more stable wireless communication to investigate the cause of the RSSI variation,
which could be the weather, daily work tasks, and so on. Regarding the antenna of 2.4 GHz wireless
modules, the antenna is deployed on the outside of the wireless module shown in Figure 4b. On the
other hand, the antenna of 429 MHz wireless modules is a chip antenna located on the wireless
module. Therefore, we considered that the antenna gain of our 2.4 GHz wireless module is higher
than that of the 429 MHz wireless module. We think our experimental results show the usefulness of
429 MHz wireless communication because 429 MHz wireless communications have higher diffraction
performance despite their lower antenna gain.
3.2. System Evaluation
As confirmed by the effect of the 400 MHz band with the IEEE 802.15.6 standard method, we
have conducted an evaluation experiment on the tomato hydroponics system. A set of scattered light
sensor nodes was placed in each plant community as shown in Figure 12. We run this system during
the tomato cultivation period of about 3 months.
Figure 12. Sensor position for system evaluation.
The results of data losses and packet resends for each scattered light sensor node are given in
Table 2.
Except for with the community 1 upper sensor node, some packet resends occurred, but all of
the sensor nodes were able to run without data loss. Therefore, although some retransmission was
necessary, we achieved reliable data collection without data loss. Figure 13a,b show the hydroponic
liquid control records of each community. Supply control is performed according to the control settings
in Table 3. There is also a sensor for measuring the power consumption at the pump supplying the
hydroponic liquid, and we were thus able to confirm whether the pump was on or off.
Figure 13 shows that the pump runs immediately when the cumulative evapotranspiration reaches
the threshold value in the control period. Due to the successful realization of wireless communication

(a)

(b)

Figure 11. RSSI variation and packet loss with or without plants: (a) 2.4 GHz band; (b) 400 MHz band.

3.2. System Evaluation


Sensors 2016, 16, 644

12 of 15

As confirmed by the effect of the 400 MHz band with the IEEE 802.15.6 standard method, we
have conducted an evaluation experiment on the tomato hydroponics system. A set of scattered
without
datanodes
loss, we
precise
control
for tomato
on a 12.
minute
scale,
which
is
light sensor
wasachieved
placed in
each plant
community
ashydroponics
shown in Figure
We run
this
system
essential
for
high-quality
tomato
cultivation.
during
the
tomato
cultivation
period
of
about
3
months.
Sensors 2016, 16, 644
12 of 15
The results of data losses and packet resends for each scattered light sensor node are given in
Table 2.
Table 2. Data losses and packet resends of system prototype.
Sensor Placement
Community 1 Upper

Receive/Send 1

Data Losses

Packet Resends

86475 / 86475

Community 1 Inner

87426 / 87453

27

Community 2 Upper

84991 / 85110

119

Community 2 Inner

89687 / 89702

15

The difference in the number of sent packets is caused by artificial factors, for example, battery replacement.

Except for with the community 1 upper sensor node, some packet resends occurred, but all of
the sensor nodes were able to
run 12.
without
loss.for
Therefore,
although some retransmission was
Figure
12. Sensor
Sensordata
position
for
system evaluation.
evaluation.
Figure
position
system
necessary, we achieved reliable data collection without data loss. Figure 13a,b show the hydroponic
liquid control records of
each
community.
controlofissystem
performed
according to the control
Table
2. Data
losses andSupply
packet resends
prototype.
settings in Table 3. There is also a sensor for measuring the power consumption at the pump
supplying Sensor
the hydroponic
and we were1 thus ableData
to Losses
confirm whether
the
pump was on
Placement liquid, Receive/Send
Packet
Resends
or off. Community 1 Upper
86475 / 86475
0
0
Figure
13 shows1 Inner
that the pump
runs
immediately when0 the cumulative evapotranspiration
Community
87426
/ 87453
27
Community
2 Upper
/ 85110
0 successful realization
119
reaches the
threshold
value in the84991
control
period. Due to the
of wireless
Community
2 Inner
89687
/ 89702precise control 0for tomato hydroponics
15 on a minute
communication
without
data loss, we
achieved
1 The difference in the number of sent packets is caused by artificial factors, for example, battery replacement.
scale, which
is essential for high-quality tomato cultivation.

(a)

(b)
Figure 13. Result of hydroponic liquid control (a) community 1; (b) community 2.
Figure 13. Result of hydroponic liquid control (a) community 1; (b) community 2.
Table 3. Control settings of each tomato community.
Placement

Control Period

Threshold

Supply Time

Community 1

7 am5 pm

80

Community 2

7 am5 pm

50

Sensors 2016, 16, 644

13 of 15

Table 3. Control settings of each tomato community.


Placement

Control Period

Threshold

Supply Time

Community 1
Community 2

7 am5 pm
7 am5 pm

80
50

5
5

4. Conclusions
In this paper, we proposed a highly reliable wireless control system for tomato hydroponics.
To address the challenge of using wireless sensor networks in an agricultural field, we adopted
the 400 MHz band and the IEEE 802.15.6 standard method for a highly reliable WSN. Besides, we
performed software implementation of a selection of fault-tolerant or interoperable hardware in a
high temperature and high humidity environment. Then we installed the tomato hydroponics control
system and operated it in a greenhouse. The experiments show the application effect of the 400 MHz
band and the method of IEEE 802.15.6 standard as follows:

The 400 MHz band is less affected by plants compared to the 2.4 GHz band.
Although there are some packet retransmissions, the 400 MHz band and the IEEE 802.15.6 standard
could reduce packet errors and data losses.
The scattered light sensor node with 400 MHz band could control hydroponics on a minute scale.

In our current prototype system, we focused on whether a reliable WSN could be built using the
429 MHz band wireless communications based on the IEEE 802.15.6 standard. If we could use 2.4 GHz
wireless modules implemented based on the same standard as the 400 MHz wireless modules, we
could evaluate the effect due to the frequency difference. In future work, we would like to evaluate
the effect due to the difference in the frequency in more detail and the power saving performance of
our developed reliable wireless control system. Moreover, a remaining issue is evaluation of the effect
of increasing the number of sensor nodes up to 40 and verification whether the whole farm can be
covered by one sink node. Besides, we have to show our developed system is useful for hydroponics
methods under other conditions and in other environments.
Acknowledgments: This study was partially supported by 2013-2014 SCOPE, Ministry of Internal Affairs and
Communications and JST PRESTO.
Author Contributions: Hirofumi Ibayashi contributed significantly to the development and experiments of the
proposed system. Yukimasa Kaneda supported to debug the system and contributed significantly to the evaluation
of it. Jungo Imahara provided useful suggestions of the proposed system and the experimental environment.
Naoki Oishi contributed significantly to the developing the scattered light sensor nodes and analyzing the LAI.
Masahiro Kuroda contributed significantly to the design of the proposed wireless system as well as the analysis of
the experimental results. Hiroshi Mineno contributed significantly to the conception of the reported research as
well as the analyzing the experimental results with useful discussions.
Conflicts of Interest: The authors declare no conflict of interest.

References
1.
2.
3.
4.

Pande, A.B.; Sheikh, N.S.; Upare, K.; Raut, S.; Ali, S. Wireless Sensor Network Based Greenhouse Monitoring
System. Int. J. Res. 2016, 3, 351353.
Emmi, L.; Gonzalez-de-Soto, M.; Pajares, G.; Gonzalez-de-Santos, P. Integrating sensory/actuation systems
in agricultural vehicles. Sensors 2014, 14, 40144049. [CrossRef] [PubMed]
Pajares, G.; Peruzzi, A.; Gonzalez-de-Santos, P. Sensors in agriculture and forestry. Sensors 2013, 13,
1213212139. [CrossRef] [PubMed]
Mirebella, O.; Brischetto, M. A hybrid wired/wireless networking infrastructure for greenhouse management.
IEEE Trans. Instrum. Meas. 2011, 60, 398407. [CrossRef]

Sensors 2016, 16, 644

5.
6.

7.

8.
9.
10.

11.
12.
13.

14.

15.

16.

17.

18.
19.
20.
21.
22.

23.
24.

14 of 15

FUJITSU Intelligent Society Solution Akisai. Available online: http://jp.fujitsu.com/solutions/cloud/agri/


uecs/ (accessed on 30 January 2016).
Yasuba, K.; Kurosaki, H.; Takaichi, M.; Suzuki, K. Development of an automatic data collection and
communication system for greenhouse in Japan. In Proceedings of SICE Annual Conference (SICE) 2011,
Tokyo, Japan, 1318 September 2011; pp. 28062807.
Jiang, S.; Wang, W.; Hu, Y.; Ye, Y. Design of Wireless Monitoring System for Environment Monitoring in
Greenhouse Cultivation. In Proceedings of the 6th International Asia Conference on Industrial Engineering
and Management Innovation, Tianjin, China, 2526 July 2015; pp. 219228.
Jahnavi, V.S.; Ahamed, S.F. Smart Wireless Sensor Network for Automated Greenhouse. IETE J. Res. 2015, 61,
180185. [CrossRef]
Park, D.; Park, J.W. Wireless sensor network-based greenhouse environment monitoring and automatic
control system for dew condensation prevention. Sensors 2011, 11, 36403651. [CrossRef] [PubMed]
Erazo, M.; Rivas, D.; Prez, M.; Galarza, O.; Bautista, V.; Huerta, M.; Rojo, J.L. Design and implementation
of a wireless sensor network for rose greenhouses monitoring. In Proceedings of the 6th International
Conference on Automation, Robotics and Applications (ICARA) 2015, Queenstown, New Zealand, 1719
February 2015; pp. 256261.
Huang, J.; Kumar, R.; Kamal, A.E.; Weber, R.J. Development of a wireless soil sensor network. In Proceedings
of ASABE Annual International Meeting, Rhode Island, RI, USA, 29 June2 July 2008.
Vellidis, G.; Tucker, M.; Perry, C.; Kvien, C.; Bednarz, C. A real-time wireless smart sensor array for scheduling
irrigation. Comput. Electron. Agric. 2008, 61, 4450. [CrossRef]
Ibayashi, H.; Kaneda, Y.; Li, P.; Suzuki, Y.; Imahara, J.; Oishi, N.; Kuroda, M.; Mineno, H. Study on Highly
Reliable Wireless Environmental Control System for Horticulture Environment. Inf. Process. Soc. Jpn Consum.
Device Syst. 2015, 5, 1020.
Youssef, M.A.; Agrawala, A.; Udaya, S.A. WLAN location determination via clustering and probability
distributions. In Proceedings of the First IEEE International Conference on Pervasive Computing and
Communications (PerCom) 2003, Fort Worth, TX, USA, 2326 March 2003; pp. 143150.
Doherty, L.; Teasdale, D.A. Towards 100% reliability in wireless monitoring networks. In Proceedings of the
3rd ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor and Ubiquitous
Networks, Torremolinos, Spain, 26 October 2006; pp. 132135.
Baviskar, J.; Mulla, A.; Baviskar, A.; Ashtekar, S.; Chintawar, A. Real time monitoring and control system
for green house based on 802.15. 4 wireless sensor network. In Proceedings of the Fourth International
Conference of Communication Systems and Network Technologies (CSNT) 2014, Bhopal, India, 79 April
2014; pp. 98103.
Watteyne, T.; Vilajosana, X.; Kerkez, B.; Chraim, F.; Weekly, K.; Wang, Q.; Glaser, S.; Pister, K. OpenWSN: a
standards-based low-power wireless development environment. Trans. Emerg. Telecommun. Technol. 2012, 23,
480493. [CrossRef]
Kim, Y.D.; Kang, W.S.; Cho, K.; Kim, D. RMRP: A reliable MAC and routing protocol for congestion in IEEE
802.15. 4 based wireless sensor networks. IEICE Trans. Commun. 2013, 96, 29983006. [CrossRef]
Song, W.; Zhou, L.; Yang, C.; Cao, X.; Zhang, L.; Liu, X. Tomato Fusarium wilt and its chemical control
strategies in a hydroponic system. Crop Prot. 2004, 23, 243247. [CrossRef]
Thelen, J.; Goense, D.; Langgendoen, K. Radio wave propagation in potato fields. In Proceedings of 1st
Workshop on Wireless Network Measurements, Trentino, Italy, 3 April 2005; pp. 331338.
Siddiqi, M.Y.; Malhotra, B.; Min, X.; Glass, A.D.M. Effects of ammonium and inorganic carbon enrichment
on growth and yield of a hydroponic tomato crop. J. Plant Nutr. Soil Sci. 2002, 165, 191197. [CrossRef]
Oishi, N. Development of the scattered light multi-sensor unit to get greenhouse environment and plant
ecology information. In Proceedings of Japanese Society of Agricultural, Biological, and Environmental
Engineers and Scientists 2014, Tokyo, Japan, 811 September 2014.
Okai, Z. Effects of the Near-Infrared-Ultraviolet Light is Applied to the Growth of the Plant; Fukui University:
Fukui, Japan, 2004; pp. 5762.
Matsuno, T.; Kushioka, S.; Imahara, J.; Fukuda, M.; Mizuno, T.; Mineno, H. Development of Agri-Environmental
Visualization and Control System using Spatial Interpolation. In Proceedings of Multimedia, Distributed,
Cooperative, and Mobile Symposium (DICOMO) 2012, Ishikawa, Japan, 46 July 2012.

Sensors 2016, 16, 644

25.
26.

15 of 15

ARIB STANDARD ARIB STD-T67. Available online: http://www.arib.or.jp/english/html/overview/doc/


5-STD-T67v1_1-E.pdf (accessed on 30 January 2016).
Kuroda, M.; Ibayashi, H.; Mineno, H. Affordable 400 MHz long-haul sensor network for greenhouse
horticulture. In Proceedings of the International Conference on Information Networking (ICOIN) 2015,
Siem Reap, Cambodia, 1214 January 2015.
2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

You might also like