2ª Conferência Brasileira de Sistemas Embarcados Críticos – CBSEC 2012
Cover Sheet Title: Use of agricultural vehicles with embedded networks and Wireless Sensor Network in crop dusting process; Student: Ivairton Monteiro Santos; Supervisor: Carlos Eduardo Cugnasca; Degree: Doctorate; Graduate Program: “Sistemas Digitais, Depto. De Eng. De Computação e Sistemas Digitais, Escola Politécnica, USP”; Email contact for the student:
[email protected]; Contact email of the supervisor: carlos.cugnasca @ poli.usp.br Year of joining the program: 2009; Expected time of completion: 2013; Date of approval of the proposed thesis / dissertation (qualification): March, 2011; Summary in Portuguese or English: 1 Introdução 1.1 Objetivo 1.2 Justificativa 1.3 Organização da tese 2 Redes de Sensores Sem Fio 2.1 Propriedades e características das RSSF 2.2 RSSF móveis 2.3 Aplicações 2.4 Protocolos de roteamento em RSSF (estática e móvel) 3 Tecnologia adaptativa 3.1 Autômato adaptativo 3.2 Tabelas de decisão adaptativas 4 Protocolo de roteamento para RSSF móvel 4.1 Definição do contexto e do problema de roteamento 4.2 Aspectos gerais do protocolo de roteamento 4.3 Simulação 4.3.1 Análise do desempenho do protocolo de roteamento 4.3.2 Resultados e discussões 5 Conclusões
Keywords: crop dusting, embedded sensors, routing data, wireless sensor network, unmanned aerial vehicles.
1
2ª Conferência Brasileira de Sistemas Embarcados Críticos – CBSEC 2012
1
Use of agricultural vehicles with embedded networks and Wireless Sensor Network in crop dusting process Ivairton M. Santos, and Carlos E. Cugnasca
Abstract— In large-scale agriculture, a common procedure is crop dusting to control pests and diseases. With this process, the main effect to be avoided is the pesticide drift. Pesticide drift is the horizontal displacement of pesticide. It can contaminate neighboring regions to the area to be treated, such as pastures, other cultures, populated regions, and rivers. Wireless Sensor Networks have potential for different applications, especially for environment monitoring and in agriculture. They can be used in pesticide drift problem, working as decision support system. In this context, they can be distributed in the area to be treated, gathering data about temperature, humidity, and wind velocity and direction. The sink node will be embedded in the crop dusting vehicle, to get data from the sensors network. Dusting process could be done by air or ground vehicles, manned or unmanned. Thus, it is possible to monitor the environmental parameters in order to identify unfavorable conditions for crop dusting. Furthermore, the operator can adjust the route according to information received from the sensors network that can measure the wind conditions. The Wireless Sensor Network can be an important tool to improve the crop dusting process efficiency, in order to minimize the pesticide drift, and contamination of soil and water.
Index Terms— crop dusting, embedded sensors, routing data, wireless sensor network, unmanned aerial vehicles.
I. INTRODUCTION
P
Agriculture (PA) is the main mechanism for the rational use of natural resources in agricultural crops, aiming at the improvement of inputs utilization, reducing costs and environmental impact, and improving productivity [1], [2]. Several technologies in information, automation and communication areas have been used in farm equipment, and more recently the use of Wireless Sensor Networks (WSN). WSN are composed by several sensors distributed in a certain area, constituting a special type of ad hoc network, with capacity to collect and process information autonomously [3], [4], [5]. Efficient crop dusting, with low cost, and without RECISION
I. M. Santos is PhD student in Polytechnic School of University of São Paulo, Brazil (corresponding author e-mail: ivairton @ usp.br). C. E. Cugnasca is with Polytechnic School of University of São Paulo, Brazil (e-mail: carlos.cugnasca @ poli.usp.br).
contaminating the environment is challenge in agricultural production. The main effect to be avoided during application is the pesticide drift. To minimize this effect, it is important to know environmental conditions, such as direction and intensity of the wind, temperature, and humidity at the application time. WSNs can be used to minimize pesticide drift. The crop area can be monitored more efficiently with the integration of two technologies: embedded networks in agricultural vehicles and WSN for environment monitoring. These vehicles work such as mobile gateways, using the ISO 11783 standard (ISOBUS) [6], it specifies a serial data network for control and communications on agricultural tractors and implements. The sensors data are collected directly by the vehicle on board computer, at the moment it enters in the signal area of the sensor network. In crop dusting, the environmental conditions are determinant to a good work. Bad conditions cause pesticide drift, waste of product and money, an inefficient dusting, and environmental pollution (air, soil and water). Pesticide drift is the horizontal movement of the drops to be launched until reaching the soil or plants [7]. The drift always occurs, but it should be avoided as much as possible. It becomes intolerable when the distance traveled by droplets is large enough for them to leave the crop area to be treated. This problem is more serious when the pesticide drifts cause damage to neighboring crops, cities, peoples, animals or rivers [8]. The main factors that affect the pesticide drift occurrence are: the size and weight of the drop, direction and intensity of the wind, temperature, environment humidity, and the drops launch height. The expectation of this work is that WSNs can be used as support tools in crop dusting processes and pesticide drift mapping. The aim is to gain greater efficiency in crop dusting process, with lower occurrence of pesticide drift and, consequently, lower product waste and environment contamination around the crop. The second section of this work proposes the functionalities for WSN in crop dusting process. Three strategies for the implementation of sensors network in pesticide drift are shown in the third section. The fourth section discusses the main challenges to be confronted with the use of WSN in crop dusting problem, and finally, the last section concludes the
2ª Conferência Brasileira de Sistemas Embarcados Críticos – CBSEC 2012
2
work with some comments about advantages, expectations and challenges.
II. FUNCTIONALITY OF WIRELESS SENSOR NETWORK IN CROP DUSTING PROCESS
Considering the ability of WSN to monitor the environmental parameters, such as temperature, humidity and wind, some strategies in the crop dusting process can be developed in order to minimize the occurrence of pesticide drift. Thereby, it is expected that the system is an assistant tool to farmers, automating process and providing decision support. Initially, two functionalities for WSN in the context of crop dusting are considered: environmental conditions evaluation and support in crop dusting route definition. A. Environment conditions evaluation The simplest implementation using a WSN is to monitor and to evaluate the environmental conditions, such as temperature, humidity, wind velocity and direction. This monitoring happens constantly, before and during spraying. The WSN will work as a decision system, checking if environmental conditions are favorable or not for dusting. For example, if the environmental temperature is over the limit determined by experts, then the system informs the farmer about this problem. Another example is, if during the pesticide application, the system detects wind gusts, and then it will warn the operator that he should stop the spraying. B. Support in crop dusting route definition A WSN based system can used as a support tool, assisting in route definition, especially in aerial crop dusting. In this context, the crop dusting vehicle (tractor or airplane) has an embedded sensor node, working as a data gateway sensor network, integrated with the vehicle Global Positioning System (GPS). While going over the crop area to be treated, the vehicle will receive the sensor network recorded data. WSN is expected to monitor the wind direction and speed, and through this information it can interact with the vehicle changing/suggesting route path. Considering the event of tolerable winds, the aim is to use this information for the crop dusting process, providing a uniform spray and avoiding pesticide drift over neighborhood regions, outside the farmer interest area. The vehicle route change should be timely happen on time, supporting the operator. Figure 1 shows an example of airplane route change. In the simulation, the WSN is monitoring the environmental conditions of crop area to be dusted. During spraying, when a wind variation occurs, the sensor network detects the change and interacts with the airplane embedded system, suggesting the correction on its route. The aim is to propose an adjustment in the route to the airplane pilot, keeping him the system control; it is his decision to follow suggested change or not. However, to
Fig. 1. Mobile gateway route changing based on WSN monitoring.
imagine the use of Unmanned Aerial Vehicles (UAV’s), this strategy can be implemented with the UAV control system, to adjust the path of the vehicle automatically.
III. STRATEGIES FOR THE IMPLEMENTATION OF WIRELESS SENSOR NETWORK IN PESTICIDE DRIFT MONITORING The use of WSN for pesticide drift monitoring, poses some the challenges. The main one is the routing data problem. The speed of the crop dusting vehicle (with embedded gateway node) is determinant in this process. If the vehicle motion is slow, such as tractor, the problem is simplified. Therefore, if the vehicle moves at high speed, such as an airplane, than routing data is more difficult. This work proposes three strategies to be used to allow the integration between WSN and crop dusting vehicle (mobile gateway). A. Gateway route representation in Wireless Sensor Network The first strategy is representing the mobile node route on WSN. It is expected that all WSN nodes know the nearest location where the mobile gateway will move. This strategy can help the routing data process, especially in case of fast motion of the gateway node. To implement this route representation on sensor network, a pre-processing will be necessary to represent the mobile gateway route path over the crop area, and send this information to the sensor network. Considering the mobile gateway route defined, some nodes near the route path will be selected. These nodes will be called “reference node”. The other sensor nodes will indicate them, in order to know where the mobile gateway will move. Figure 2 shows the route representation over the crop area to be treated, the selected reference nodes and the organization of the other nodes in relation to the reference nodes. B. Organization of network nodes into clusters. The strategy of grouping the sensor network nodes in clusters, aims to organize them hierarchically and make energy savings [9]. Figure 3 shows an example of an organized WSN in
2ª Conferência Brasileira de Sistemas Embarcados Críticos – CBSEC 2012
3
Fig. 4. Routing of data using the interception strategy.
IV. CHALLENGES Fig. 2. Representation of mobile gateway route path, selected reference nodes and organization of other sensor nodes.
clusters. Each sensor node is in a cluster and each cluster has a “cluster head”. In the figure, the cluster heads are highlighted. In the context of this work, the set of cluster head node will be the same set of referenced nodes established in the mobile gateway route representation process. Therefore, there will be as many clusters as reference nodes. The other cluster nodes should send their data to their cluster heads.
Fig. 3. Organized WSN in clusters with its cluster head highlighted.
C. Data routing based on Interception Deliver monitored data by the WSN to the mobile gateway can not be an easy task, especially if the gateway is moving at high speed, even though the mobile node route is represented in the sensor network and nodes arranged in clusters. Considering the high speed mobility, such as embedded gateway on crop dusting airplane, we propose the use of data routing based on interception [10]. The interception data routing consist in monitoring the gateway movement. At the time it approaches, the sensor nodes calculate the gateway speed and report it to the neighborhood sensors. Then, the data sensors are routed and sent to the mobile gateway, so as to intercept it ahead. Figure 4 shows an example of routing data using this strategy. When the mobile gateway enters the monitored area, in t1 time, the sensor network indentifies its approach. The sensor nodes calculate the gateway speed and send the recorded data to intercept the mobile node in the next referenced node in gateway route in t2 time.
The proposed strategies results in some challenges to be confronted. The first aspect is determining the adequate number of referenced nodes in gateway route representation. Determining a low number may represent an overhead of transferring data and possibly cause failures in the data delivery process. Nevertheless, a large number of referenced nodes can make the intercept routing data difficult, setting a large number of clusters and increasing the energy expenditure of sensor nodes. Therefore, we must specify an optimal number, balanced, that properly fits the different aspects associated with reference nodes number definition. Another challenge is associated with the cluster head redefinition. This function switching between network nodes is needed to reduce power consumption and to extend the sensor network life time. However, for the purpose of this work, the gateway node route is the main factor for the cluster heads determination, which limits the selection of only those nodes close to the gateway path route. Optionally, the system would provide new routes, as an alternative to the initial route. Therefore, it is possible to diversify the cluster head nodes (referenced nodes) and consequently increase the sensor network lifetime. However, this strategy can conflict with the user system interests that may not have the flexibility in route changing. Finally, the main challenge is the data routing problem. Although the interception strategy has been demonstrated by computational simulations, the route has always been to linear route. This work proposes an interception routing for any route, from its representations in the sensor network. The continuity of this work is to verify the presented strategies, using computational simulations, in the context of the crop dusting.
V. CONCLUSIONS The application of WSNs in agricultural can perform an important role in many activities, such as in pesticide drift monitoring and mapping problem caused in crop dusting process. The use of WSNs in this context can amplify the crop dusting efficiency, minimizing product wastage, the financial loss of the farmer and especially reducing the probability of contamination of neighborhood areas, such as pasture, other
2ª Conferência Brasileira de Sistemas Embarcados Críticos – CBSEC 2012 crop areas, communities and river or lakes. This work proposed the use of WSNs for two pesticide drift monitoring functions. The first consists of a decision system, which evaluates the climatic data and determines if the environmental conditions are favorable or not to crop dusting. The second function is to assist the operator during spraying. In this case, the WSN evaluate the wind data (direction and speed) and suggests corrections on the crop dusting vehicle route. This function aims to optimize the dusting process and to minimize the contamination of neighborhood areas. The utilization of WSN in pesticide drift monitoring poses some technical challenges. This work proposed some approaches to solve these challenges. These strategies will be verified in the wake of the research through computational simulations.
ACKNOWLEDGMENT The authors acknowledge The State of Mato Grosso Research Foundation – FAPEMAT – for supporting this work via research projects Nº 465450/2009 and the Coordination for the Improvement of Higher Education Personnel - CAPES.
REFERENCES [1]
N. Zhang, M. Wang, N. Wang, “Precision agriculture-A worldwide overview”, Computers and Electronics in Agriculture, 36, 2002, pp. 113–132. [2] A. M. Adrian, S. H. Norwood, P. L. Mask, “Producers’ perceptions and attitudes toward precision agriculture technologies”, Computers and Electronics in Agriculture, 48, 2005, PP. 256–271. [3] I. F. Akyildiz, W. Su, Y. Sankarasubramaniam, E. Cayirci, “Wireless sensor networks: a survey”, Computer networks, 38, 2002, pp.393-422. [4] M. Tubaishat, S. Madria, “Sensor networks: an overview”, IEEE Potentials, 22 (2), 2003. [5] P. Gajbhiye, A. Mahajan, “A survey of architecture and node deployment in Wireless Sensor Network”, Applications of Digital Information and Web Technologies-ICADIWT, 2008, pp.426-430. [6] M. Darr, K. Hudson, “Standardization of electronics in machinery systems”, Engineering and Technology for Sustainable World, 11(10), 2004, pp. 13-14. [7] A. Chaim, Manual de Tecnologia de Aplicação de Agrotóxico, Embrapa, 2009. [8] A. J. Hewitt, “Spray drift: impact of requirements to protect the environment”, Crop Protection, 19, 2000, pp. 623–627. [9] N. Vlajic, D. Xia, “Wireless sensor networks: to cluster or not to cluster?”, in International Symposium World of Wireless, Mobile and Multimedia Networks-WoWMoM, 2006, pp.–268. [10] H. A. B. F. Oliveira, R. S. Barreto, A. L. Fontao, A. A. F. Loureiro, E. F. Nakamura, “A Novel Greedy Forward Algorithm for Routing Data toward a High Speed Sink in Wireless Sensor Networks”, in International Conference Computer Communications and NetworksICCCN, 2010, pp. 1–7.
4
Ivairton Monteiro Santos é graduado em Ciência da Computação pela Universidade Federal de Mato Grosso (2002) e mestre em Ciência da Computação pela Universidade Federal Fluminense (2005). Atualmente é aluno de doutorado da Escola Politécnica da USP, sob a orientação do Prof. Dr. Carlos Eduardo Cugnasca. É membro do Laboratório de Automação Agrícola da Escola Politécnica da USP e professor da Universidade Federal de Mato Grosso, no Campus Universitário do Araguaia. Tem experiência na área de Ciência da Computação e seus interesses em pesquisa concentram-se em Redes de Sensores Sem Fio e otimização combinatória.
Carlos Eduardo Cugnasca é graduado em Engenharia de Eletricidade (1980), mestre em Engenharia Elétrica (1988) e doutor em Engenharia Elétrica (1993). É livre-docente (2002) pela Escola Politécnica da Universidade de São Paulo (EPUSP). Atualmente, é professor associado da Escola Politécnica da Universidade de São Paulo, e pesquisador do LAA Laboratório de Automação Agrícola do PCS - Departamento de Engenharia de Computação e Sistemas Digitais da EPUSP. Tem experiência na área de Supervisão e Controle de Processos e Instrumentação, aplicadas a processos agrícolas e Agricultura de Precisão, atuando principalmente nos seguintes temas: instrumentação inteligente, sistemas embarcados em máquinas agrícolas, monitoração e controle de ambientes protegidos, redes de controle baseados nos padrões CAN, ISO11783 e LonWorks, Redes de Sensores Sem Fio e computação pervasiva. É editor da Revista Brasileira de Agroinformática (RBIAgro).
