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; [email protected] Shizuoka Prefectural Research Institute of Agriculture and Forestry, 678-1 Tomioka, Iwata, Shizuoka 438-0803, Japan; [email protected] (J.I.); [email protected] (N.O.) National Institute of Information and Communications Technology, 4-2-1 Nukui-Kitamachi, Koganei 184-8795, Japan; [email protected] JST, PRESTO, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan Correspondence: [email protected] (H.I.); [email protected] (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) [1–4]. 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) [7–10] 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) 2405–2480 MHz (5 MHz interval 16ch) Based on IEEE 802.15.4 standard 250 kbps 10 mW 3-volt DC

CC430F5137 (Texas Instruments) 429.25–429.75 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, Plat’Home, 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, Plat’Home, 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 Plat’Home, 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 node’s 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 node’s 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 node’s 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 node’s 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 node’s 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 node’s 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. ˆ 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

0

0

Community 1 Inner

87426 / 87453

0

27

Community 2 Upper

84991 / 85110

0

119

Community 2 Inner

89687 / 89702

0

15

1

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 am–5 pm

80

5

Community 2

7 am–5 pm

50

5

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 am–5 pm 7 am–5 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, 351–353. Emmi, L.; Gonzalez-de-Soto, M.; Pajares, G.; Gonzalez-de-Santos, P. Integrating sensory/actuation systems in agricultural vehicles. Sensors 2014, 14, 4014–4049. [CrossRef] [PubMed] Pajares, G.; Peruzzi, A.; Gonzalez-de-Santos, P. Sensors in agriculture and forestry. Sensors 2013, 13, 12132–12139. [CrossRef] [PubMed] Mirebella, O.; Brischetto, M. A hybrid wired/wireless networking infrastructure for greenhouse management. IEEE Trans. Instrum. Meas. 2011, 60, 398–407. [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, 13–18 September 2011; pp. 2806–2807. 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, 25–26 July 2015; pp. 219–228. Jahnavi, V.S.; Ahamed, S.F. Smart Wireless Sensor Network for Automated Greenhouse. IETE J. Res. 2015, 61, 180–185. [CrossRef] Park, D.; Park, J.W. Wireless sensor network-based greenhouse environment monitoring and automatic control system for dew condensation prevention. Sensors 2011, 11, 3640–3651. [CrossRef] [PubMed] Erazo, M.; Rivas, D.; Pérez, 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, 17–19 February 2015; pp. 256–261. 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 June–2 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, 44–50. [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, 10–20. 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, 23–26 March 2003; pp. 143–150. 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, 2–6 October 2006; pp. 132–135. 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, 7–9 April 2014; pp. 98–103. 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, 480–493. [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, 2998–3006. [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, 243–247. [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. 331–338. 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, 191–197. [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, 8–11 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. 57–62. 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, 4–6 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, 12–14 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/).

A Reliable Wireless Control System for Tomato Hydroponics.

Agricultural systems using advanced information and communication (ICT) technology can produce high-quality crops in a stable environment while decrea...
5MB Sizes 2 Downloads 14 Views