A GSM-based remote wireless automatic monitoring system for field
c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/compag A GSM-based remote wireless automatic monitoring system for ﬁeld information: A case study for ecological monitoring of the oriental fruit ﬂy, Bactrocera dorsalis (Hendel) Joe-Air Jiang a , Chwan-Lu Tseng b , Fu-Ming Lu a , En-Cheng Yang c,∗ , Zong-Siou Wu a , Chia-Pang Chen a , Shih-Hsiang Lin a , Kuang-Chang Lin b , Chih-Sheng Liao b a b c Department of Bio-Industrial Mechatronics Engineering, National Taiwan University, Taipei 106, Taiwan Department of Electrical Engineering, National Taipei University of Technology, Taipei 106, Taiwan Department of Entomology, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 106, Taiwan a r t i c l e i n f o a b s t r a c t Article history: Monitoring ﬁeld conditions is the foundation of modern agricultural management. In order Received 1 October 2007 to improve the efﬁciency of the data collection procedure, and to improve the precision with Received in revised form which agricultural operations are managed, it is necessary that we have an automated sys- 7 January 2008 tem that collects environmental data, especially to record long-term and up-to-the-minute Accepted 8 January 2008 environmental ﬂuctuations. The purpose of this study was to design a remote pest monitoring system based on wireless communication technology. This system automatically reports Keywords: environmental conditions and traps pest in real-time. The data we acquired was integrated Precision agriculture into a database for census and further analysis. The system consists of two components, a Environmental parameter remote monitoring platform (RMP) and a host control platform (HCP). Furthermore, based monitoring on the bio-characteristics of the oriental fruit ﬂy, a high precision automated trapping and Wireless communication counting device was designed. This device counts the number of trapped ﬂies and then technology sends the information back to the RMP. The RMP is in charge of acquiring the environmen- Mechatronics technology tal data and the number of trapped ﬂies, and it sends all the data back to the HCP in the form Bactrocera dorsalis (Hendel) of a short cell phone message through the wireless Global System of Mobile Communica- Ecological environment tion (GSM). Our system then transmits the data via a commercial base station. The system can work properly based on the effective coverage of base stations, no matter the distance from RMP to HCP. The function of the HCP is to receive and store, display, and analyze the database on line. It also provides functions like inquiries, early warning, and announcements. The system was ﬁeld tested over a 1-year period (March 2006 to July 2007), and the experimental results demonstrated that it can monitor the environmental parameters and population dynamics of the oriental fruit ﬂy in real-time. Based on the long-term monitoring database acquired by our system, the relationship between the population dynamics of the fruit ﬂy and the environmental changes can be easily analyzed. With the help of this system, researchers can judge the correlation of the occurrence of the oriental fruit ﬂy and climate conditions. Since the long-term database provides us with the details of the population dynamics of the fruit ﬂy, the system allows us to control the pest in time and reduce agricultural losses. The experimental results demonstrate that large scale, long distance, ∗ Corresponding author at: Department of Entomology, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 106, Taiwan. Tel.: +886 2 3366 9640; fax: +886 2 3365 2092. E-mail address: [email protected] (E.-C. Yang). 0168-1699/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.compag.2008.01.005 244 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 and long-term monitoring for agricultural information can be achieved by using our proposed monitoring system. Much improved spatial resolution and temporal resolution is obtained compared to traditional methods for monitoring the data of the oriental fruit ﬂy based on environmental changes. © 2008 Elsevier B.V. All rights reserved. 1. Introduction Integrated pest management (IPM) relies on the accuracy of the pest population monitoring technique. Without gathering information of population dynamics together with the related ecological factors it is almost impossible to execute the appropriate pest control at the right time in the right place. However, most insect pests are spread across large areas and across many boundaries. Like other insect pests, with their strong reproductive ability the oriental fruit ﬂy, Bactrocera dorsalis (Hendel), is one of the top major pests in the Asia-Paciﬁc region, causing serious fruit damage and agricultural losses throughout the whole of Taiwan, year after year. In this article, we report for the ﬁrst time a Global System of Mobile Communication (GSM)-based remote wireless automated system for monitoring the population dynamics of the oriental fruit ﬂy by means of modern wireless communication technology. By integrating the traditional trapping method with modern communication technology, our system is able to provide real-time information on the ﬁeld conditions and the dynamics of the pest population at different monitoring sites. This makes the system especially helpful for large scale monitoring in mountainous areas, which is a major obstacle for the traditional monitoring methods. 1.1. Agricultural ecology and pest management Ecological factors in the environment can be classiﬁed as physical (e.g. temperature and humidity), chemical (e.g. chemical composition of the soil), and biological factors (e.g. pathogens and pests). As far as cultivation and management of the land is concerned, the ecological factors are crucial to the quality and productivity of the crop. Among these factors, pests are those insects that directly damage the crop, and pest control has always been considered the most difﬁcult challenge to overcome. To reduce the loss of agricultural products caused by insect pests, thousands of methods, including physical and chemical ones, have been developed over the years. Each of these methods needs to be applied in the right place and the right time, and many ecological models have been developed speciﬁcally for this purpose. However, without accurate ﬁeld information, no theoretical estimations or strategies for pest control will be able to hit the target and reduce the enormous costs to both agriculture and ecology. Since all ecological factors have a dynamic connection with time, and each of these factors may interact with others, traditional monitoring techniques which rely on manpower to collect data from trap to trap, point to point are no longer efﬁcient enough for modern pest management (Shen, 2003). There are two main drawbacks with traditional pest monitoring: (1) it is labor intensive and therefore costly and (2) all monitoring points cannot be synchronized to measure the variables, not even between two recording instruments at a single site. Given that the traditional monitoring techniques are labor intensive with a poor temporal property, the dynamics of pest population density in the ﬁeld cannot be accurately monitored. Consequently, a proper estimation for a target pest population will be limited to a long-term scale, such as pest numbers per 10 days or per 30 days. An example is, the oriental fruit ﬂy (B. dorsalis (Hendel)), which is one of the major pests of many fruits and many commercial crops in the Asiapaciﬁc region. The methyl eugenol trapping method has been used for more than 30 years, yet we still know very little about the dispersion of the insect in the ﬁeld. The oriental fruit ﬂy is highly fecund, 8–9 generations per year, and is active throughout the entire year in the tropical and sub-tropical countries (TACTRI/COA, 2006). Previous studies have indicated that the oriental fruit ﬂies attack 173 different fruits and vegetables (Metcalf and Metcalf, 1992), and several of them are economically important. By stinging the fruit for oviposition, the ﬂy infests the fruit in the early stage of fruit development. The larvae hatching inside the fruit causes the fruit to rot, or to make the fruit ripen and drop before maturation (On News Report, 2004; Lin et al., 2006). Since the population growth and activity of the oriental fruit ﬂy can be inﬂuenced by many ecological parameters, e.g. temperature, rainfall, wind speed and diurnal rhythm (TAPHIQ/COA, 2006), an automated system for accurately integrating real-time monitoring information is needed to better know the related parameters and the dynamic population in different loci. 1.2. Research purpose The rapid improvement of today’s micro-fabrication technology and embedded systems allows for a tiny electronic sensor to integrate multiple functions, like precise sensation and calculation (Gschwind et al., 2001; Ali et al., 2004; Joshua et al., 2005). By integrating the latest in sensory chips and microfabrication technology with traditional sensor modules and the pest trapping device, an automated pest and environment monitoring system can be established, together with powerful processing ability. Furthermore, it will reduce the labor cost of collecting environmental data by combining the automated monitoring system with wireless telecommunication technology. By having real-time environmental data available to us we can better understand the variation among cropland areas, increase the effectiveness and the precision of cultivation management and establish a highly reliable pest forecast system. As a result, the automated, wireless pest monitoring system can serve as the major information source for our agriculture development and insure our competitiveness. Given that GSM wireless communication has been well established over all of Taiwan, both in the low-lying areas and the mountains (He, 2003), transporting data and signals using the GSM framework will be much more economical than any other communication technique, especially in the rural c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 areas (He, 2003; Jiang et al., 2003). Tseng et al. (2006) demonstrated the practicability of transporting data by mechatronics and Global System of Mobile Communication and Short Message Service (GSM–SMS) technologies (ETSI, 1999). Integrating the sensor modules and GSM wireless communication system will extend the transmission of the recorded environmental data from hundreds of kilometers away. In addition, an integrated system will facilitate the accumulation of longterm data and determination of the relationship between these environmental variations and pest occurrences, and hence may be the critical reference source needed for effective pest management. The trend of modern agriculture is to introduce, sensor techniques, long-distance wireless communication (like GSM–SMS), or global positioning systems (GPSs) and to proceed with automation, precision, and IT. Therefore, the main purpose of this study is to construct a wireless and automatic monitoring system that will be useful for monitoring pest population dynamics and related environmental variations. 1.3. Research background 1.3.1. Pest monitoring methods Many devices have been developed for trapping speciﬁc insect pests, based on the biological characteristics of the pests (Yang, 1988; Reynolds and Riley, 2002). Today’s research in the methodology of ﬁeld surveys for various pests also uses modern technologies, such as radar technologies for monitoring pest migration or ﬂight (Chapman et al., 2002; Riley and Smith, 2002), video equipment to observe ﬂying insects in the ﬁeld (Riley, 1993), thermal infrared imaging for the monitoring of rainfall in relation to the control of migrant pests (Milford and Dugdale, 1990), chemiluminescent tags for tracking insect movement in darkness (Spencer et al., 1997), electric systems for detecting moths (Hendricks, 1989), the micro-bar-code system for monitoring honeybee behavior (Sasaki, 1989), the remote sensing technique used to detect the effects of insects on their host plants (usually damage to crops or forests) or to monitor environmental factors (Riley, 1989; Hay et al., 1997), GPS for wildlife telemetry and habitat mapping (Hurlbert and French, 2001), the high-frequency echo-sounding method to detect the movements of larvae (Eckmann, 1998). However, the ﬁnancial burden for building the fundamental hardware of these high-tech facilities may only be affordable to governments or very large agricultural corporations. At the same time, real-time communication between data collection and control terminals may not be easy to achieve since the automatic data may be collected in remote areas and the communication coverage may be restricted. Therefore, constructing a monitoring system based on an automatic and real-time communication platform is urgently needed for modern agriculture. 1.3.2. 245 Chu, 1986). The history of ﬁghting the oriental fruit ﬂy in Taiwan can be traced back for 50 years. In 1955, the poisonous protein hydrolysate was airdropped in the ﬁelds to attract the ﬂy; in 1975, sterile males which were exposed to 60 Co irradiation were released, coupled with poisonous methyl eugenol as an integrated pest management (Liu, 1981). In addition, several other methods have been used to control the oriental fruit ﬂy since then, e.g. chemical control, yellow sticky paper trapping, fruit bagging, etc. (Chu and Chu, 1987; Chiu, 1990; Wu and Chu, 1990; Chang, 1994; Liu and Hwang, 2000; Chen et al., 2002; Liu, 2002). Ho (2003) reported that traps containing methyl eugenol as attractant can trap the most male oriental fruit ﬂies. Therefore, in the present study we use the male oriental fruit ﬂy as the subject of auto-monitoring and methyl eugenol as the male-speciﬁc attractant. Traditionally, the population dynamics of the oriental fruit ﬂy in Taiwan is monitored by yellow plastic traps containing methyl eugenol. The distributing, hanging, and collecting of the traps in the ﬁeld are labor intensive and the trapped amount of the oriental fruit ﬂy is then counted manually, which makes it even more costly and time consuming. The survey results of the population dynamics in Taiwan are therefore measured on a 10-day scale. It is evident that the traditional method is not economical, nor very efﬁcient considering the limited resolution in space and time scales, and the lack of environmental parameters. Even though the detection, attraction, and biological characteristics of the oriental fruit ﬂy are very well known (Liu, 2002; Chen et al., 2002), nevertheless we still lack a precise pest detection mechanism. The wireless auto-monitoring system can readily cope with this problem. Our previous investigation on the automonitoring of the Diamondback moth, Plutella xylostella, one of the major pests of Brassicaceae (Cruciferae) convincingly demonstrated the application of an automatic sensory system on pest control (Jiang et al., 2003; Lu et al., 2004; Tseng et al., 2004, 2006). The moth is attracted by a synthetic sex pheromone, detected and counted with sensors made of electric grid. In this study, with this experience in mind, we propose an automated and wireless system which simultaneously records both the environmental variations and the pest population dynamics. This technology reduces the costs of both labor and resources and at the same time monitors the population dynamics with an adjustable temporal resolution in a large space scale. In addition to the collecting of ﬁeld data, the developed system is also able to communicate, integrate, and analysis data in real-time. All the information is received and processed at the server end, and thus a pest forecast system is constructed with temporal and spatial precision. 2. Materials for experiment and system mapping Control history of the oriental fruit ﬂy in Taiwan The subject insect pest of this research is the oriental fruit ﬂy, B. dorsalis (Hendel) (Diptera: Tephritidae). It is the major pest for various fruits in Taiwan and is also the target insect of a quarantine law. The females lay their eggs into fruits by punching their ovipositor through the fruit’s skin, thereby greatly reducing the quality of the fruit (Liu, 1981; Chiu and The software and hardware used to develop our proposed GSM-based remote wireless automatic monitoring system are listed as follows: Software: MSP GCC, PHP, LabVIEW, MySQL, Apache server. 246 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Fig. 1 – Structural diagram of our proposed remote wireless automatic monitoring system. Fig. 2 – The information ﬂow of our proposed remote wireless automatic monitoring system. Hardware: sensors for measuring wind speed, temperature, and humidity, microcontroller (TI MSP430F449 chip), GSM module, GPS receiver, PC. The basic make-up of our system can be divided into two major parts: the Remote Monitoring Platform (RMP) and the Host Control Platform (HCP) for monitoring and statistical analysis of ﬁeld information, respectively. The RMPs are located at measuring places, and each RMP is equipped with an anemograph, temperature and humidity sensors, GPS receiver, and GSM module to measure the environmental variants, as well as a pest trap to determine the number of the insect pest locus in quo. The commercial GSM network system is adopted for data transmission to the HCP in short message (SM) format. The GSM module enables the HCP to receive and transmit ﬁeld data to a PC for further analysis. The format design for short message packets used in this study follows our previous work (Tseng et al., 2006). The graphics user interfaces (GUIs) of the HCP are programmed and integrated on a Laboratory Virtual Instrument Engineering Workbench (LabVIEW) development platform (Bitter et al., 2001; Ritter, 2002). The HCP receives ﬁeld data from the RMPs and stores the information in the database designed by MySQL (Koﬂer, 2001). Via the Worldwide Internet, the user can then explore the environmental variants and the collected insect pest numbers from the homepage of the RMPs which is programmed in PHP (Brown, 2002) and stored in the Apache (Melanie, 2001) server. 2.1. The structure of the RMP The structure of the RMP is described as follows: the MSP430F449 microcontroller developed by Texas Instruments, Inc. (MSP430F4xx Manual, Texas Instruments, Inc., 2006) is adopted as the core-processing chip of the RMP. This chip serves to create data packets, transmit, and dispatch control commands among the modules used in the RMP. Based on mechatronics technology, the RMP is designed in a modular fashion. In addition, the RMP also includes an anemograph, temperature and humidity meters, a pest-detecting trap, a GPS receiver, and a GPS module. The anemograph used here Fig. 3 – System architecture of the remote wireless monitoring platform. c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 is the type of AM-4203 with an accuracy of ±2.0% made by the Lutron Company. The AM-4203 possesses an RS-232 serial port and provides 16 bits string output function such that it can communicate easily with the MSP chip. The sensor used in the present study for both temperature and humidity measurements is a Sensirion SHT75, which has a built-in microcontroller to measure the temperature and the relative humidity simultaneously. The SHT75 provides long-term stability and high measurement accuracies of ±2.0% for relative humidity and ±0.4 ◦ C for temperature. The GSM module (model no.: FASTRACK M1203A) created by WAVECOM Corpo- 247 ration is used in both the RMP and the HCP. The GSM module meets both GSM900 and GSM1800 speciﬁcations, and has the same basic capability with the cell phone. The GPS receiver (GM44, San Jose Navigation Corporation) with an RS-232 interface is used in our system. The GPS receiver characterizes a 15 m positioning accuracy, and mainly provides the geographic information and correct time tag for the RMP. Using serial communication ports, the MSP microcontroller collects each module’s data including environmental information, number of pests, and the RMPs geographical location. The monitoring data is then packed and becomes a short mes- Fig. 4 – Construction of remote monitoring platform and ﬁnished product: (a) circuit layout, (b) PCB, (c) assembly of RMP components, and (d) ﬁnished product. 248 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Fig. 4 – (Continued ). sage which is transmitted by GSM module to HCP. As a result, the system is capable of collecting and transmitting the ﬁeld information automatically. 2.2. or the statistical data can be graphically displayed. This is particularly useful for future work to establish an identiﬁcation model for the quantity growth of the oriental fruit ﬂy. The structure of the HCP In addition to the data transmission of the GSM module, the HCP contains GUIs written in LabVIEW (National Instruments, Inc.) to control the data transmission from the GSM module (Bitter et al., 2001; Ritter, 2002). The GUIs are linked to a MySQL database (Koﬂer, 2001) and can save the ﬁeld data received from the RMP into the database of the HCP. The HCP also contains a dynamic homepage so that the users can explore the ﬁeld data collected from the ﬁeld in real-time as well as the statistical analysis of historical records via the Internet. The server adopted in the HCP was established by Apache (Melanie, 2001), PHP (Brown, 2002), and MySQL. Thus, we can perform remote monitoring and collect the number of pests and the environmental parameters, and then save this ﬁeld data in the MySQL database. The user and the superintendent of the system can quickly search data from the PHP dynamic homepage and each item of ﬁeld information 2.3. System integration and implementation 2.3.1. System architecture Combining the GSM network with the technologies of the Internet, mechatronics, and wireless communication, this study constructs a wireless automatic pest monitoring system. The structural diagram of the system is shown in Fig. 1. The system can be divided into two major parts: the RMP and the HCP. The RMP uses MSP430F449 as the core-processing chip and the GPS module for positioning information. The data recorded by sensors, including the number of trapped ﬂies, is transferred to the HCP through the commercial GSM network. The user will be able to monitor the ﬁeld information via the Internet. Users can obtain the historical records from the database of the HCP if necessary. Fig. 2 shows the information ﬂows of the designed remote wireless automatic monitoring system. c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 2.3.2. Implementation of the RMP The system architecture of the RMP is shown in Fig. 3. The MSP430F449 is the core-processing chip of the RMP for integrating the GPS and the GSM modules, as well as the environmental information sensory modules. The environmental information, including temperature, humidity and wind speed, and the number of trapped ﬂies are the parameters for measurement. Several serial ports are reserved for future applications. A liquid crystal display (LCD) is also installed in RMP. Via LCD, users can monitor the collected information in locus in quo, which is particularly useful for maintaining the system. Fig. 4 shows the construction of the RMP. Fig. 4(a) is the circuit layout which is designed with software Protel 99SE. Fig. 4(b) shows the printed circuit board, including MSP430F449 receptacle, the temperature/humidity sensory circuit, the circuits for USART voltage level conversion and channel switching, and several output ports which can connect with the anemometer, the LCD, the GPS and GSM modules. In addition, it also provides a voltage output to supply the electrical power for other circuits. The completed assembly of components is shown in Fig. 4(c). The photograph of the ﬁnished RMP is shown in Fig. 4(d). The outer covering of the RMP is made of acrylic plates, and the electric circuits of the RMP are ﬁxed in the box. The whole box is set on an angle-iron support. The GPS module is the white disc on the top right corner of the RMP. The anemograph and humidity/temperature sensor are mounted at the bottom of the box away from the ground. 2.3.3. Implementation of the HCP The HCP is set up in a personal computer (PC). The function ﬂow of the HCP in this monitoring system is shown in Fig. 5. After receiving the monitoring data by GSM module, the HCP analyzes the received data and the information regarding the RMP operational status. If there is an error, the HCP immediately sends a request command for re-transmission, and the RMP will send the data again. If the trapped pest number is over the pre-set threshold, the HCP will alert authorized per- 249 sonnel. The main program and the associated GUIs are based on the planned functions and are developed using LabVIEW. At the same time, a PHP-programmed dynamic homepage is developed so that users can retrieve the historical monitoring data via the Internet. The homepage is very user-friendly. 2.3.4. GUIs design of the monitoring system According to the system operation requirements, the designed GUIs are described as follows. 126.96.36.199. Login page. When the program is activated, the login page appears. The user is then requested to enter his/her account and password. If the account and password are correct, the user is then allowed to start the HCP to execute the setting function. 188.8.131.52. Setting and testing page. Fig. 6 shows the setting and testing window of the monitoring system. In this page the user can set up the connection port of the man–machine interface to the GSM module of the HCP. After choosing the proper connection port, the system will automatically test the connecting status of the GSM module to the HCP. If the GSM module works, then the indicator for the module connecting status will be turned on. Also, the power strength of the received signal is shown to tell the user the connecting condition of the GSM module. If the GSM signal is weak, then a weak indication will remind the user to adjust the antenna for a stronger signal or change the HCPs location. After the GSM module is connected, the user starts to set the RMPs. After choosing the RMP number and entering its phone number, one can click the “set” button followed by the “start” button; the system is then automatically activated. By clicking the “start” button, the HCP will send a ready performance ﬁle (RPF) as the starting message to the RMP for the initial handshake. As the HCP receives the acknowledgement conﬁrmed by the selected RMP, the start indicator of the RMP will be turned on to inform the user that the RMP is ready to use. An instant information frame is located below this Fig. 5 – Function ﬂow diagram of the host control platform. 250 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Fig. 6 – The setting and testing window for the GSM module and the RMP. window, and will list all the information about any action performed by the HCP and its relevant time. 184.108.40.206. Control interface window. After setting the RMP, clicking the “control interface” button (Fig. 6) the user can enter the control interface window as shown in Fig. 7(a). On the upper left side of the window are the indicators of the system status. There are three kinds of system-status indication, start regular transmission (Tx), start broadcasting, and fault alarming. When the HCP receives an error message sent from the RMP, the fault alarm is turned on to inform the user, and the system will send an error message to the maintenance staff. When the maintenance is ﬁnished, one can press the “fault reset” button on the window and turn off the fault alarm. In the middle above the window shown in Fig. 7(a) is the latest information in an SM transferred back from the ﬁeld. It contains the RMP information, including the RMP number, the location coordinates of the RMP, temperature, humidity, wind speed, pest number, and time of data transmitted back. All the ﬁeld information will be saved in the database for further inquiry and analysis. The partial portion of the VI ﬁle for performing the “save data” function is shown in Fig. 7(b). Located in the center of the window is the real-time command line for requesting the RMP to send the current information immediately or to set the counter to zero. Once the expected data was lost, the fault alarming module is triggered and the supervisor will be noticed to request the RMP to retransmit the data promptly or inform the worker locus in quo to ﬁx the malfunction of the RMP. Any information about the executed action is displayed on the message board in the lower part of the window (Fig. 7(a)). Fig. 7 – The control interface window for the operating RMP at the HCP: (a) GUI and (b) partial portion of VI ﬁle for setting the “broadcast” function in the control interface window. particularly useful for long-term monitoring. The broadcast setting function includes the information broadcast, the fault alarm broadcast, and pest alarm broadcast. The portion of the VI ﬁle for setting the “broadcast” function is shown in Fig. 8(b). The information broadcast can send the collected ﬁeld information to the user after receiving the sensing data, and thus the user can keep informed and does not need to stay in front of the computer monitor. If a RMP fault occurs, the fault alarm broadcast will automatically notify the maintenance staff. In addition, when the number of trapped pests is larger than the threshold, which is pre-set by the user, information will be sent to the user by the pest alarm broadcast to reduce the possible pest damage. From the perspective of agricultural applications, Tseng et al. (2006) reported that the system, which uses GSM–SMS to transmit data, can ensure that the remote data arrives on the user’s screen in 30–60 s without any error found in the data content. Therefore, the GSM–SMS technology used in this study is feasible and reliable. 2.4. 220.127.116.11. Advanced setting Trapping facility for the oriental fruit ﬂy window. Fig. 8(a) shows the advanced setting window which contains the regular transmission (Tx) setting and the broadcast setting functions. The function of the regular transmission setting allows users to set the time interval between transmissions and keep the RMP automatically transmitting data. This function is Traditionally a mixture of chemical attractant and insecticide is being used to attract and kill the fruit ﬂy for long-term monitoring of its distribution. The attractant placed in the tube is for attracting male ﬂies (Steiner, 1952). The trapping device we adopted in our proposed system is a modiﬁcation of the c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 251 Fig. 8 – Advanced setting window for operating the RMP at the HCP: (a) GUI and (b) partial portion of VI ﬁle for setting the “broadcast” function in the advanced setting window. traditional fruit ﬂy trapping tube, with an automatic counting module placed in the tube. The trapping device in our system uses a double-counting mechanism (Lin et al., 2006). The number of trapped ﬂies is counted as they cross the infrared interruption sensor. In general, this kind of counter needs a gate or an inhaler to avoid counting the same ﬂy more than once. To reduce the complexity as well as the cost of the device, and at the same time obtain a correct count, a double-counting solution was designed. This system consists of a set of optical sensors installed along the trap pathway the ﬂy will pass through. Fig. 9 shows the doublecounting optical sensors which are indicated by the arrows. The correct counting can be obtained by a microprocessor which takes the signal interruption into account only when the ﬂy crosses both optical sensors sequentially. The ﬂy which hovers around is then excluded by the double counting mechanism. From our observations we found that there are four possible actions for the ﬂy crossing the double-counting optical sensors, thus we set our counting rules based on these conditions: 1. If a ﬂy enters the trap and hovers around the ﬁrst sensor or even stays on it, the processor then holds the ﬁrst sensor in interrupting status until the ﬂy crawls across the second sensor. 2. If the ﬂy crosses the ﬁrst sensor and has not yet arrived at the second sensor, the processor then holds the ﬁrst sensor in interrupting status until the second sensor is interrupted. 3. When the ﬂy arrives at the second sensor and both sensors are in interrupting status, the processor will increase the count and reset the status of both sensors. Fig. 9 – The double-counting optical sensors deployed along the pathway. 252 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Fig. 11 – Test for GPS positioning accuracy of the GPS receiver of the RMP. The test experiments were conducted at the top balcony of the BIME building, NTU between 19 April 2005 and 26 April 2005. Fig. 10 – Flowchart of double-counting mechanism for the designed ﬂy trapping tube. 4. If the ﬂy hovers around the second sensor, the processor only increases the count once and clears the redundant signal caused by the hovering. Because the ﬁrst sensor is not in the interrupting status, the processor will ignore the second sensor’s signal even it is interrupted continuously. This will avoid the ﬂy from crawling back to interrupt the ﬁrst sensor again. Based on the above-mentioned ﬂy behavior, Fig. 10 shows a ﬂow chart of the double-counting mechanism. Only the occurrence of two consecutive interruptions triggers the counter to add one. By using this mechanism, compared to the singlecounting method, the double-counting optical sensory set can be more accurate in counting the trapped ﬂies. Note that in order to reduce the device complexity, this study ignores the following conditions: (1) the ﬂy crawls back to the entrance of the pathway passing through the optical interruption sensors 1 and 2 and (2) the distance between two ﬂies entering the pathway is less than 2 cm. In this paper, the pathway is designed to reduce the occurrence of these conditions. cultivated land, we analyzed the longitude and latitude information collected by the RMP. The RMP was mounted at a ﬁxed location on the top balcony of the building of the Department of Bio-Mechanical Engineering (BIME), National Taiwan University (NTU), as shown in Fig. 11. There was no canopy over the top balcony of the BIME building and no other surrounding buildings to affect the reception of the satellite signal. The GPS receiver of the RMP received the longitudinal and latitudinal data from the GPS signal every 30 min and delivered it to the HCP. The system collected 522 data during this operation lasted about 10 days, and the distribution of the data is shown in Fig. 12. The red point in Fig. 12 indicates the average of all data points, which represents north latitude 25.011007◦ , or equally 25◦ 0 39.63 , and east longitude 121.325819◦ or equally 121◦ 19 32.95 . During test, the RMP with its single GPS receiver was located in the ﬁxed location mentioned above. Due to the fact 3. Experimental results of system performance evaluation 3.1. Testing the GPS positioning accuracy According to the GM-44 manual of the GPS receiver (On Web of GM-44 Manual, 2006), its positioning accuracy is 15 m. To conﬁrm if the actual positioning accuracy of a GPS receiver can provide accurate geographic information for monitoring Fig. 12 – The distribution diagram of the received longitude and latitude data for the positioning accuracy test of the GPS receiver of the RMP between 19 April 2005 and 26 April 2005. c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 253 Fig. 13 – Distribution diagram of the positioning error margin for the received GPS signals. that there is no information available on the exact longitude and latitude of the location of the RMP, we could only take the average longitude and latitude data as a reference coordinate for judging the positioning error margin of the GPS receiver. Comparing every received longitude/latitude data with the mentioned reference coordinate, the average positioning error margin was 3.01 m, the maximum error was 22.22 m, the minimum error was 0.13 m, and the standard positioning error was 2.61 m, respectively. The distribution diagram of the positioning error margin of the GPS receiver is shown in Fig. 13. About 90% of the GPS received signals providing positioning data had a margin of error that fell within 5 m in longitude and latitude compared with the average data. This implies that the positioning accuracy of the adopted GPS receiver was quite reliable. Within the scope of this margin of error, if we treat the average longitude and latitude values as deﬁnite ones for a single RMP, it is acceptable. But, if we consider monitoring many points, we should use a more sophisticated GPS technique or incorporate the present technique with an electronic map to ﬁx the RMPs’ position exactly. Then we can effectively build up a distribution diagram for monitoring the ﬁeld information over a wide area. 3.2. tube Testing the counting accuracy of the ﬂy trapping The purpose of this experiment was to determine the counting accuracy of the ﬂy trapping tube we designed. The experiment was done outdoors, and we compared the number of ﬂies counted by the system and also counted them manually. Fig. 14 shows the OrCAD layout and a photograph of the actual ﬁnished product of the ﬂattened pathway of the ﬂy trapping tube applied in this paper. The cross-section of the entrance is made larger in order to distribute the attractant and to permit the ﬂies to enter the trapping tube easily. The height of the back pathway is lowered to increase the effect of the signal interruption when a ﬂy passes by, thereby further increasing the counting accuracy. The ﬁnished product with the double-counting device installed in the ﬂy trapping tube is shown in Fig. 15. The oriental fruit ﬂy is lured by methyl eugenol and crawls into the pathway via the entrance and passes by the double-counting optical sensors mechanism, resulting in the ﬂy being automat- Fig. 14 – The cross-section of the ﬂy entering pathway in the designed ﬂy trapping tube: (a) OrCAD layout and (b) entity photograph of ﬁnished product. ically detected and counted. Fig. 16 shows the enlarged picture of the RMP equipped with the ﬂy trapping tube. To prevent sunlight inﬂuencing the counting accuracy, the upper half of the ﬂy trapping tube was coated with a dark paint. Fly trapping tubes equipped with electronic counting circuits were placed under the trees outside of the BIME building on NTU campus in May 2006. The experimental results are summarized in Table 1. The average counting accuracy of the ﬂy trapping tube with two ﬂattened pathways was around 78.1% in this test. From the experimental results, the counting accuracy of the 6th experiment was only 56.5%. After careful examination, we found that when the oriental fruit ﬂies crawl over different pathways, some ﬂies might hide in the corner inside pathway or retrace the pathway towards the entrance. This fact also causes the variation in the counting accuracy. To cope with this problem, we reﬁned the dimension of the designed pathway and tuned the alignment of the optical sensors. The optimal dimension of the pest-entering pathway is as shown in Fig. 14(a). The average accuracy is 80.2% without the 6th experiment since the alignment of optical sensors was missed during the maintenance work. The accuracy is 254 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Table 1 – The experimental results for the counting accuracy of the ﬂy-trapping tube equipped with two ﬂattened pathways Evaluation items Counted number Actual number Accuracy (%) Experimental course Total Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 11 12 91.7 53 74 71.6 30 35 85.7 18 25 72.0 35 43 81.4 13 23 56.5 36 39 92.3 Average accuracy (%) 196 251 78.1 further conﬁrmed in accordance with the long-term experimental results of the entire monitoring system described in the next section. With the features mentioned above, the automatic detecting device in this study can attract male oriental fruit ﬂies and effectively count their number. 3.3. Field test Originally two prototypes of RMPs were produced and set up in the farm on the NTU campus to test their performance, as shown in Fig. 17(a). Later on, to ensure that the numbers of trapped ﬂies and the monitored data collected by the system were matched, the two RMPs were moved to outside the department building of the BIME for outdoor testing as shown in Fig. 17(b). The HCP of the monitoring system was placed in our Lab on the 3rd ﬂoor of the BIME building, and the distances between the HCP and two RMPs were about 20 and 30 m, respectively. Since the experimental location was in a bosket on the NTU campus, the quantity of oriental fruit ﬂies was smaller than in the farm ﬁeld. For monitoring ﬁeld information and trapping oriental ﬂies, however, this fact did not actually affect the performance evaluation of our system. The RMP system includes the following items: the core-processing system integrated by the MSP430F449 chip, the temperature/humidity sensing module, the GPS receiver module, the GSM module, Fig. 16 – Our proposed automatic counting trap for the oriental fruit ﬂy installed in the RMP. Fig. 15 – The photograph of the double-counting device installed in the ﬂy trapping tube. and the anemometer. The ﬂy trapping tube, the anemometer, and the temperature/humidity sensing module were put in the bottom of the box so as to be shaded and so decelerate any possible damage of these modules due to severe environmental variation. This arrangement of the sensing modules can effectively measure the variation of the environmental parameters. The whole system has been tested outdoors as of 1st March 2006. At the beginning, the number of ﬂies counted by the ﬂy trapping tube did not match the number we counted manually, so we changed the standard ﬂy trapping tube with our improved ﬂy trapping tube in May 2006. After that, the number counted by ﬂy trapping tube became more accurate and reliable. The temperature, humidity, wind speed, and the number of trapped ﬂies were then recorded, and the information was statistically analyzed in the HCP. Fig. 18 shows the monitoring data for a single day as collected by RMP #2 on 4 August 2006. The ﬂy trapping tubes were manually emptied, and the c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Fig. 17 – Deployment test of the developed remote monitoring system for the oriental fruit ﬂy: (a) pretest of the RMP conducted at the farm ﬁeld on the NTU campus and (b) long-term test of the RMP conducted at the bosket near the BIME department building, NTU. 255 number of ﬂies trapped in the ﬂy trapping tube was also manually counted at 10 p.m. each day to verify the accuracy. Except for the days we performed maintained on the instruments, we have tested and veriﬁed the entire system for 304 days, from 1st July 2006 to 30th April 2007. During these 304 days, we deleted the data for 90 days because no ﬂies were trapped due to rain or extreme low temperature or the number of ﬂies trapped was lower than 5. The latter recordings were excluded because in such cases the counting accuracy of the trapping ﬂy tube varied severely, some were very high and some were very low to the point that no statistical contribution for long-term performance evaluation could be obtained. Under these circumstances, the mean accuracies of monitoring oriental fruit ﬂies by the two RMPs (#1 and #2) deployed on the NTU campus were 84% and 76%, respectively. Our proposed system provides different time resolutions for recording ﬁeld information, so it is possible for users to set the time interval for sending the data from the RMPs to the HCP as required. An interval of 30 min was set in the present study. For example the data monitored by RMP #2 on 4th August 2006 (Fig. 18), was sent to the HCP every 30 min. The information included the number of trapped oriental fruit ﬂies, temperature, humidity, and the mean of the wind speed. The humidity (%RH), temperature (◦ C), wind speed (cm/s), and the accumulative number of the trapped oriental fruit ﬂies are depicted by blue bars, red line, green line, and black line, respectively. Fig. 18 also shows the trends in humidity and temperature during the recording time. The highest temperature and the lowest humidity were observed at noon. As it was getting dark, the temperature became and the humidity increased, as expected. Based on the monitored information, we could estimate the local daily activity of the oriental fruit ﬂy. The HCP system we developed provides various inquiry functions for ﬁeld information. After logging in to the PHP webpage of our monitoring system, users can set the inquiry functions at their PC. The daily, weekly, and monthly averaged Fig. 18 – Monitoring data for a single day collected by RMP #2 in 4 August 2006. 256 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Fig. 19 – The average recordings per day from July to August in 2006 monitored by RMP #1. data to individual/complex ﬁeld information can be inquired from our system. For example, as shown in Fig. 19, the average recordings of every day from July to August in 2006 monitored by RMP #1 can be displayed. The system can provide us with long-term observations on both the environmental variations and the ﬂy population dynamics in a local area by calculating the information for each day. Thus the effects of environmental variations on the ecological pattern of oriental fruit ﬂy can be investigated in a good temporal and spatial resolution if more RMPs are deployed. In this study we planned and designed the HCPs database of the monitoring system in detail so that the database can offer many kinds of statistical analysis functions for the long-term recordings of ecological information and various inquiry services for complex data. Thus, if we inquire the data per week, the inﬂuence of more long-term environmental changes on the population dynamics of the oriental fruit ﬂy can be readily observed via our system. When obtaining the weekly data points, the correlation between the environmental parameters and the population dynamics of the fruit ﬂy can be observed. For example, in Taiwan the daily temperature rises from June until September. Taiwan lies in the subtropical zone in the northwestern Paciﬁc Ocean. In the mentioned 4month period, this area is subject to frequent thunderstorms Fig. 20 – The average environmental data and the total population of ﬂies per week monitored by RMP #1 on the NTU campus. c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 257 Fig. 21 – The hourly records averaged from each day by RMP #1 during the period of 1 June 2006 to 31 December 2006. in the afternoon, as well as the occasional typhoon. Therefore, the temperature is not evenly distributed in the mentioned period. As a result, the number of oriental fruit ﬂies trapped is closely related to the changes in climate. In northern Taiwan, July–August is the peak period of the oriental fruit ﬂy activity. However, their activity reduces gradually when the weather turns cooler after September. Fig. 20 shows the average environmental data and the total population of ﬂies per week detected by RMP #1 of our monitoring system on the NTU campus during the period from June to November, 2006. According to the data of our longterm recordings, it rained nearly every day between 1 June 2006 and 17 June 2006. In addition, there were four typhoons during 1 July 2006 to 12 July 2006 and 26 July 2006 to 17 August 2006. The climate records from the Taiwan Weather Bureau also conﬁrmed the recordings from our system. Fig. 20 clearly indicates that the ﬂy population was quite low during June and the ﬁrst 2 weeks of July due to rainy weather. Also, the population curve of trapped ﬂies lowered signiﬁcantly since the end of July due to a typhoon over North Taiwan. The ﬂy population increased slowly after around the third week of August, but went down following the temperature after September as shown in Fig. 20. This experiment demonstrates that a longterm and a good time resolution can be achieved by our system in the monitoring of the population of the oriental fruit ﬂy in correlation to the environmental changes. With the long-term recording and high temporal resolution provided by our system, the population dynamics of the oriental fruit ﬂy correlated to climatic parameters can be further investigated. Since this system offers a huge database of long-term recordings from each of monitoring site, the information can be extracted from the database with any time interval, be it hourly, daily, weekly or monthly. For example, we can extract the average data of each hour to analyze the dynamics of the number of trapped ﬂies and its relationship to the recorded environmental physical parameters within a day. Fig. 21 shows the information (collected by RMP #1) of the averaged hourly records from each day for the period of 1 June 2006 to 31 December 2006. It is clearly shows that the oriental fruit ﬂy’s daily peak activity is between 06:00 and 10:00 a.m. This result matches the previous hourly physically observations by Jiang (1986), indicating that our system is capable of providing high temporal resolution for ﬁeld information using an economical recording technology. 3.4. Statistical analysis of the ﬂy counting accuracy provided by the proposed monitoring system To demonstrate and conﬁrm the reliability and accuracy of our system, the data collected from both RMPs (#1 and #2) for more than 1 year (RMP #1: 20 July 2006 to 31 July 2007; RMP #2: 1 June 2006 to 31 July 2007) was statistically analyzed. Faulty data due to system maintenance and zero trapped ﬂies were eliminated, resulting in 222 and 313 data points being used in this analysis for RMP #1 and RMP #2, respectively. Fig. 22 shows the linear regression of the counting results from RMP #1 (a) and RMP #2 (b). Each data point in the ﬁgures indicates the numbers being counted manually and automatically with reference to the abscissa and the ordinate, respectively. The slope of the regression line indicates the average accuracy of the ﬂy counting of the RMP, and R2 represents the reliability of the ﬂy counting accuracy. The average accuracy of RMP #1 is 81.42% and that of RMP #2 is 73.71%, both with very high reliability, i.e., R2 = 0.951 for RMP #1 and 0.952 for RMP #2, respectively. While the selected data in this statistical analysis is in a period different from that mentioned in Section 3.3, the average counting accuracy of RMP #1 is quite similar to previous one (i.e., 84%). This fact once again demonstrates the stability and the reliability of performance of our system. 258 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 agricultural management. Also, using our system will be very helpful for monitoring invading species. In addition, when combined with a biological identiﬁcation model and/or ecological theories, the ability to collect ﬁeld data and have access to the long-term recording database of this system will be one of the most useful references and tools for integrative pest management. For example, through the long-term information collected in the ﬁeld, including temperature, humidity and wind speed, and the population dynamics of the pest, one can create a dynamic population model in a particular application area. This is very useful when making decisions on a pest control strategy—which depends on applying the right method at the right time in the right place. Therefore, compared to the traditional monitoring methods, our system can effectively reduce the cost and increase the effectiveness of pest control. Fig. 22 – The linear regression analysis of the counting accuracy resulting from both RMPs: (a) RMP #1 and (b) RMP #2. 4. Conclusions This study developed a complete and automatic monitoring system for remote ﬁeld information. The purpose of developing this monitoring system was to achieve the remote wireless measurements of environmental parameters and population dynamics of the oriental fruit ﬂy in real-time. The monitoring system is designed in a modular fashion and consists of two parts, i.e., RMPs and an HCP. The RMP is equipped with a MSP430F449 core-processing chip. This chip can package the sensory data of temperature, humidity, wind speed, and the number of trapped ﬂies into a short message, and then send that message to the HCP at a pre-set time interval by GSM module. The HCP then writes the sensory data into the MySQL database under the control of a program written in LabVIEW. By using the PHP website development software, users can search the remote ecological information of the oriental fruit ﬂy on the Internet. In addition, any researcher or farmer can search the historical data on the website, and use the website to analyze the data. The environmental variations can be monitored in real-time by our system. The long-term monitoring accuracies for the oriental fruit ﬂy achieved by the two RMP we designed in this study, RMPs #1 and #2, were 81.42% and 73.71%, respectively. According to our results, the reliability of our automatic ﬂy trapping system is about 95%. Moreover, since the RMPs automatically sent the ﬁeld information back to the HCP every 30 min as scheduled, the daily activity model of the oriental fruit ﬂy was available and the model matched the previous study, indicating that our proposed system provides reliable data. The major contribution of this study is that it makes large scale, long distance, and long-term monitoring for agricultural information achievable. High spatial and temporal resolutions for monitoring the data of the oriental fruit ﬂies with respect to environmental changes can also be achieved. Via our system one can easily gather ﬁeld information in real-time, suggesting a possible application for a pest alarm system in the ﬁeld of Acknowledgements This work was supported in part by the Council of Agriculture of the Executive Yuan, Taiwan, under contracts: 92AS-1.1.6FD-Z1, 93AS-1.1.6-FD-Z1, and 94AS-1.3.6-FD-Z1. The authors would also like to thank the National Science Council of Taiwan, for their ﬁnancial supporting under contract no.: NSC 95-2218-E-002-073. references Ali, L., Sidek, R., Aris, I., Ali, A., Suparjo, B.S., 2004. Design of a micro-UART for SoC application. Comput. Electric. Eng. 30, 257–268. Bitter, R., Mohiuddin, T., Na, M., 2001. LabVIEW Advanced Programming Techniques. CRC Press, Boca Raton, FL, USA. Brown, M.C., 2002. XML Processing with Perl, Python, and PHP. Sybex, San Francisco, USA. Chang, T.Y. 1994. Attraction of natural plants to oriental fruit ﬂy, Dacus dorsalis (Hendel) (Diptera: Tephritidae). Master thesis. National Chung Hsing University, Taichung. 81 pp. Chapman, J.W., Smith, A.D., Woiwod, I.P., Reynolds, D.R., Riley, J.R., 2002. Development of vertical-looking radar technology for monitoring insect migration. Comput. Electron. Agric. 35, 95–110. Chen, W.S., Chen, S.K., Chang, H.Y., 2002. Study on the population dynamics and control tactics of the oriental fruit ﬂy (Bactrocera dorsalis (Hendel)). Plant Prot. Bull. 44, 267–278. Chiu, H.T., Chu, Y.I., 1986. The occurrence and injury of the oriental fruit ﬂy in the southern Taiwan. Plant Prot. Bull. (Taiwan, R.O.C.) 28, 313–321 (in Chinese with English Abstract). Chiu, H.T., 1990. Ethyl benzoate: an impact ovipositional attractant of the oriental fruit ﬂy, Dacus Dorsalis Hendel. Chin. J. Entomol. 10, 375–387. Chu, H.T., Chu, Y.I., 1987. Oviposition behavior of Dacus Dorsalis Hendel (I) on the oviposition potential. Chin. J. Entomol. 7, 119–126. Eckmann, R., 1998. Allocation of echo integrator output to small larval insect (Chaoborus sp.) and medium-sized (juvenile ﬁsh) targets. Fish. Res. 5, 107–113. ETSI, 1999. GSM 07.07. Technical Speciﬁcations. European Telecommunications Standards Institute, Sophia-Antipolis, France. c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 2 ( 2 0 0 8 ) 243–259 Gschwind, M., Salapura, V., Maurer, D., 2001. FPGA prototyping of a RISC processor core for embedded applications. IEEE Trans. VLSI Syst. 9 (2), 241–250. Hay, S.I., Packer, M.J., Rogers, D.J., 1997. The impact of remote sensing on the study and control of invertebrate intermediate hosts and vectors for disease. Int. J. Rem. Sens. 18, 2899–2930. He, R.G., 2003. Pushing the wireless scientiﬁc island forward-wireless communication. Sci. Dev. 371, 20–27 (in Chinese). Hendricks, D.E., 1989. Development of an electric system for detecting Heliothis sp. moths (Lepidoptera: Noctuidae) and transferring incident information from the ﬁeld to a computer. J. Econ. Entomol. 82, 672–684. Ho, K.Y., 2003. Evaluation of two new traps for the population management of the oriental fruit ﬂy, Bactrocera dorsalis (Hendel). Ph.D. Dissertation. Department and Institute of Entomology, National Taiwan University (in Chinese). Hurlbert, I.A.R., French, J., 2001. The accuracy of GPS for wildlife telemetry and habitat mapping. J. Appl. Ecol. 38, 869–878. Jiang, J.A., Tseng, C.L., Lee, R.G., Chang, C.H., Lu, F.M., 2003. A study on agricultural data collection using the GSM–SMS technology. In: Proceedings of 2003 Symposium on Applications of Information Technology in Agriculture (CD), Taipei, Taiwan, December 1–2, 2003, pp. 204–209 (in Chinese with English Abstract). Jiang, P.L., 1986. Study on reaction in oriental fruit ﬂies, Bactrocera dorsalis (Hendel) (Diptera: Tephritidae) to the methyl eugenol attractant. M.A. Thesis. Department of Plant Pathology and Entomology, National Taiwan University (in Chinese). Joshua, M.J., Gerardo, O.V., Rodrigo, C.M., Eusebio, V.R., Alfredo, Z.G., Gilberto, H.R., 2005. FPGA-based real-time remote monitoring system. Comput. Electron. Agric. 49, 272–285. Koﬂer, M., (D. Kramer, Trans.), 2001. MySQL, Apress, Berkeley, CA, USA. Lin, K.C., Wu, Z.S., Liao, J.S., Lin, S.H., Chen, J.P., Tseng, C.L., Lu, F.M., Jiang, J.A., 2006. Design and implementation of the automatic trap for oriental fruit ﬂy. In: Proceedings of the 3rd International Symposium on Machinery and Mechatronics for Agriculture and Bio-systems Engineering (ISMAB2006) (CD), Seoul, Korea, November 23(25, 2006, pp. 602–609. Liu, Y.C., 1981. A review on studies of the Oriental fruit ﬂy, Dacus dorsalis (Hendel) in Taiwan. Bull. Soc. Entomol. (NCHU) 16, 9–26 (in Chinese with English Abstract). Liu, Y.C., Hwang, R.H., 2000. Preliminary study on the attractiveness of volatile constituents of host fruits to Bactrocera Dorsalis (Hendel). Plant Prot. Bull. 42, 147–158. Liu, Y.C., 2002. A review of studies and controls of the oriental fruit ﬂy (Bactrocera dorsalis (Hendel)) and the melon ﬂy (B. cucurbitae Coquillett) in Taiwan (Diptera: Tephritidae). In: Wang, S.C., Ho, C.C., Chen, C.C., Chen, W.H. (Eds.), Proceedings of the Symposium on Insect Ecology and Fruit Fly Management, Special Issue, No. 4. Plant Protection Society, Taichung, ROC, pp. 1–40. Lu, F.M., Jiang, J.A., Chang, C.H., Chen, Y.S., Ouyang, C.S., Tseng, C.L., 2004. Performance analysis of agricultural information communication using GSM–SMS technology. In: Proceedings of Machinery and Mechatronics for Agriculture and Bio-industrial Engineering (CD), Ilan, Taiwan, September 2(3, 2004, pp. 273–274 (in Chinese). Melanie, H., 2001. Installing and conﬁguring Web servers using Apache. John Wiley & Sons, New York, USA. Metcalf, R.L., Metcalf, E.R., 1992. Fruit ﬂies of the family Tephritidae. In: Metcalf, R.L., Metcalf, E.R. (Eds.), Plant 259 Kairomones in Insect Ecology and Control. Chapman & Hall, New York, USA, pp. 109–152. Milford, J.R., Dugdale, G., 1990. Monitoring of rainfall in relation to the control of migrant pests. Phil. Trans. R. Soc. Lond. B 328, 705–717. On Web of MSP430F4xx Manual, Texas Instruments, Inc., 2006. http://focus.ti.com/lit/ug/slau056f/slau056f.pdf. On News Report, 2004. http://www.ag007.com.tw/cgibin/industry-info/article.pl?category=5&sn=5. On Web of GM-44 Manual, 2006. http://www.wavecom.com/media/ﬁles/support/Hard platforms/Modems/Fastrack M1206B/User manual/M1206B-ONuser%20guide.pdf. Reynolds, D.R., Riley, J.R., 2002. Remote-sensing, telemetric and computer-based technologies for investigating insect movement: a survey of existing and potential techniques. Comput. Electron. Agric. 35, 271–307. Riley, J.R., 1989. Remote sensing in entomology. Ann. Rev. Entomol. 34, 247–271. Riley, J.R., 1993. Flying insects in the ﬁeld. In: Wratten, S.D. (Ed.), Video Techniques in Animal Ecology and Behaviour. Chapman & Hall, London, pp. 1–15. Ritter, D.J., 2002. LabVIEW GUI: Essential Techniques. McGraw-Hill, New York, USA. Riley, J.R., Smith, A.D., 2002. Design considerations for a harmonic radar to investigate the ﬂight of insects at low altitude. Comput. Electron. Agric. 35, 151–169. Sasaki, M., 1989. A trial of the micro-bar-code system for monitoring honeybee behaviour. Honeybee Sci. (Mitsubachi-Kagaku) 10, 182–183. Shen, Y., 2003. The technology and application of the great spatial variation between the condition of crop growing and the quality of soil. In: Proceedings of Conference on Precision Agriculture, Agriculture Association of China, Taipei, Taiwan, pp. 23–41 (in Chinese). Spencer, J.L., Gewax, L.J., Keller, J.E., Miller, J.R., 1997. Chemiluminescent tags for tracking insect movement in darkness: application to moth photo-orientation. Great Lakes Entomol. 30, 33–43. Steiner, F., 1952. Methyl eugenol as an attractant for oriental fruit ﬂy. J. Econ. Entomol. 75, 173–178. TACTRI/COA, 2006. On Web of Plant Protection Manual, http://www.tactri.gov.tw/htdocs/ppmtable/frgu-07.pdf. TAPHIQ/COA, 2006. On Cordon Web, of Bactrocera dorsalis (Hendel). http://off.entomol.ntu.edu.tw/a3ﬂy/01about ﬂy/about03.html. Tseng, C.L., Lee, R.G., Jiang, J.A., Lu, F.M., Chang, C.H., Chen, Y.S., Ouyang, C.S., Chao, C.F., 2004. An automatic Diamondback moths monitoring system. In: Proceedings of the 2nd International Symposium on Machinery and Mechatronics for Agriculture and Bio-systems Engineering (ISMAB2004) (CD), S1, Kobe, Japan, September 21(23, 2004, pp. 153–156. Tseng, C.L., Jiang, J.A., Lee, R.G., Lu, F.M., Ouyang, C.S., Chen, Y.S., Chang, C.H., 2006. Feasibility study on application of GSM–SMS technology to ﬁeld data acquisition. Comput. Electron. Agric. 53, 45–59. Wu, H.H., Chu, Y.I., 1990. Inﬂuence of methyl eugenol on the mating ability of the male fruit ﬂy (Dacus Dorsalis Hendel). Chin. J. Entomol. 10, 69–78. Yang, L.Z., 1988. Mating behavior and evaluation of the effect of communication disruption with sex pheromone on diamond back moth, Plutella xylostella (L.). Ph.D. Dissertation. Department of Plant Pathology and Microbiology, National Taiwan University (in Chinese).
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
Knight’s RMP Digital with remote control introduces the newest in... industry’s most reliable platform. The high performance warewash system...