Functional environment of a mobile work unit

Functional environment of a mobile work unit Sørensen, C.G. 1) ; Suomi, P. 2) ; Kaivosoja, J. 2) ; Pesonen, L 2) 1) University of Aarhus, Faculty of Agricultural Sciences, Department of Agricultural Engineering, 8700 Horsens, Denmark. 2) MTT Agrifood Research Finland Vakolantie 55, FI-03400 Vihti, Finland Project: InfoX T - User-centric mobile management in automated plant production 1. Introduction...2 2. Data flow analysis and technical requirements for the management system in automated plant production...3 2.1. Entity definition: identification of work units...4 2.1.1. Competences...5 2.1.2. Identification of elements of mobile work units...6 2.1.3. Selected work units...6 3. Information model...7 3.1. Entities involved with field operations...7 3.2. Managing field operations...8 3.3. Information model for spraying...9 3.3.1. Process decomposition...10 3.3.2. Data modeling...21 4. References...27 1

1. Introduction The planning and control architectures for mobile work units in the field include different layers of abstraction for handling both deliberation and reactivity (Chatila, 1995). In a hybrid architecture deliberation or mission planning focus on the predictable or goal-directing behaviour of the work units (e.g. route plan) while local reactive behaviour deals with the uncertainty of the environment and adaptation to local conditions during execution. A number of approaches to operation planning for agricultural machinery, ranging from manual planning systems to various degrees of automated planning involving parameterisation of the planned operation have been attempted (Stoll, 2003). Fig. 1 outlines the basic management processes which are identified within the agricultural plant production cycle for both manned and unmanned machinery items. The management activities concentrate on planning and controlling the execution of operations on some soil or crops (Sørensen, 1999). These operations include soil treatment, seedbed preparation, seeding, fertilising, plant care, harvesting and irrigation. Operation describes the agronomic purpose of an activity, while tasks describe the realisation of the operation involving relevant resources in terms of implements. The decomposition of processes is based on the management functions ranging from strategical to operational planning, execution control and evaluation, and a number of underlying processes and subprocesses. The operational plans are decomposed for formulation and control of the planned operations and tasks. Arable farming Strategic planning Tactical planning Operational planning Execution Evaluation Planned operations Operational crop planning Handling tasks Execute tasks Revision of Task task formulation in in terms terms of of instructions route planning for and vehicle/implement task scheduling formulate jobs formulate jobs formulate tasks formulate tasks modify task modify tasks required operations operations urgency operations specifications select task inspect task control task control operation control device Figure 1. Information and planning activities in agricultural operations management with the identification of the revised task formulation to be invoked in the case of autonomous vehicles (adapted from Goense & Hofstee, 1994) In relation to Figure 1, it should be noticed that the activity of observing and monitoring is also to be regarded as an operation, which can be planned, executed and controlled in the same way as for the traditional machine operations (Sørensen et al., 2002). In this way, the task of observing/monitoring can be formulated according to the actual needs of observing or monitoring of the systems states, the costs-benefits of acquiring a specific, etc. The functional environment of a mobile work unit within an automated plant production context consists of its internal and external interaction with an overall management system on the farm. The focal point of the management system is to sustain the planning and execution of farm operations. 2

When focusing on the management of field operation the task of optimised field data management becomes one of carrying out the following steps or procedures: creating planned field operations transferring or delivering the plans to the field with specified tasks setting up the mobile work units for executing the planned operation managing, controlling and recording the field operation documenting the executed field operation for recordkeeping and managerial purposes By specifying in detail the provided and the required for the handling processes in Figure 1, the design and functionalities of the individual system elements can be derived. That is the case both for on-board machinery systems as well as for support service systems. 2. Data flow analysis and technical requirements for the management system in automated plant production A detailed structuring and formalisation of physical entities and the, which surrounds the planning and control of efficient mobile working units in automated agricultural plant production systems is a decisive prerequisite for the development of comprehensive and effective ICT-system for task management on the farm. In this context, it is essential that requirements, communication protocols and common definitions of the exchanged are set up. Scheepens (1991) presented the concept of modelling as the basis for this important task. The basic idea is to model all the activities and decisions, which take place in a targeted production section and combine this modelling with all the relevant data. The formal description includes entity definition (in this case mobile work units), a process model (activities and decision processes) and a data model (data relating to the processes). The defined processes in the process model and the entities and attributes in the data model provide the basis for developing compatible systems. A corresponding modelling approach was developed in a Finnish research project focusing on analysis of user requirements for farm management systems (Nurkka et al. 2007, Pesonen et al. 2007). This project focused on malt barley production and used this process as a case to implement a usercentred approach to farm system development. The project team exploited experience of cognitive systems engineering methods and especially such that apply functional modelling of the work domain. The methodology was originally developed at Risö National Laboratory in Denmark by Jens Rasmussen (1986) and was later developed by Vicente (1999). The model has typically been used to analyse generic work domain control demands for the design of automation systems for complex industrial processes. At VTT Technical Research Centre of Finland the method has developed further in two respects: First, a method for deriving generic user work demands, labelled core-task demands, on the basis of the functional analysis was developed. Second, a tool for describing situational decision-making models was constructed which include analyses of available and operating possibilities. (Norros 2004). These tools were applied in the malt barley production process. The modelling exercise involved experts of agriculture, malt process experts, and farmers involved with farming methods. The acquired models were tested and developed further in interaction with actual malt barley growers in four farms. The models acquired based on experience of the present traditional farming process were in the next phase completed by research results concerning new needs and possibilities of ICT-based precision farming. The results of these several modelling phases were used to conceptualise the structure and user interface for farm management. The invoking of modelling provides a sound approach for specific applications. As part of the applicability of such modelling approaches, the formation of international standards is important. On-going work in this area include the ISO TC 23/SC 19/WG1, which has the purpose of setting 3

up an open interconnected on-board system, permitting electronic units to communicate, and to define the data exchange with the Farm Management Information System (FMIS: includes software, decision support system, etc. for farm management). 2.1. Entity definition: identification of work units In the primary agricultural production, the high degree of mechanization has increased the productivity considerably in the last decades. This development is in the process of being coupled with automation and extensively use of embedded ICT system. Control of field machinery (conventional, like tractor with implement, or autonomous vehicles, like robots) enables, by use of advanced ICT, collection of detailed sitespecific during operation execution. This contributes to minimized resource input and to an environmentally sound and quality optimized production via decision support systems or directly via on-line control (Martin-Clouaire & Rellier, 2000; Sørensen, 1999; Sørensen et al., 2002; Suomi, 2006). Planning and formulation of jobs will include indication of expected time schedule on the basis of the immediate crop development, weather forecasts, etc. The job descriptions will be transmitted to the tractor/implement for control and manual/automatic site-specific adjustment of implements. In cases of realized work results deviating from the planning, on-line corrective measures will be invoked. The final work result will be recorded and documented, and the obtained data will be stored for learning and use in connection with new loops of planning or control see Figure 2. Arbejdskraft Labour/ / technique teknik Mark: Field: - transient midlertidige attributter attributes Mark: Field: - permanente attributter attributes - ekspertviden Expert knowledge -processering processing Eksterne External databaser databases Maskine/ Machine/ redskab implement Implemented operation ø FMIS - database Adjustments Operation plan Information Operator ør Operat Instructions Beslutninger Decisions Experience/ Erfaring/ preferencer preferences Figure 2. Information handling in the field and on the farm Figure 2 gives an overview of the management necessary to implement an effective task management on mobile work units in plant production. Within this concept, the task management function will provide the farmer with a scheduling tool for planning and controlling the tasks relevant to the field operations. The principal output from the deliberation processes is the formulated operation and task plan, which will be downloaded to the machine/implement unit. Field operations maps are the instructions that both guide vehicle-based movements in the field and control concurrent agronomic operations. This task plan will be dynamic indicating that time will be dealt with explicitly and if knowledge on the environment increases or improves, it will be possible to reformulate the plan in order to continually uphold a timely and 4

cost-efficient operation. The planning and control system must be able to predict the evolution of system states (field and crop development, machine performance, etc.) and plan the actions of work units accordingly. Also, the system must be reactive and capable of reformulating plans based on observations and feedback from the actual implementation of the tasks. Figure 3 shows the planning and control loops of the general prescriptive task management functions together with its integration with the operator at various levels of the system. Figure 3. Planning and executive control loops for field machinery. The global model indicate the overall off-line planned operation specifications outlining the agronomic requirements, like dosage, working depth, prescriptive driving patterns, etc. The local model handles the reformulation of prescribed plans as a consequence of unexpected events during execution, user-induced alterations, etc. The local model together with the on-line control function keeps the machine and implements settings continuously on target. 2.1.1. Competences The requirement assessment, development and implementation of an management system for mobile work units require a number of competences and approaches to be employed. In terms of management efforts and technology assessment the following competences are included: - usability and applicability studies of the proposed technologies - analysis and decomposition of machine operations - resource optimization and decision support 5

As regards technology components, these include a mixture of sensor and communication elements: - on-machinery monitoring/display units - on-machinery control units - tractor-implement communication systems - wireless communication system between machines and support services (ex. internal/external databases) - sensor systems for continuously updating of system 2.1.2. Identification of elements of mobile work units The implementation of field operations requires working units comprising field machinery. Field machinery is traditionally based on tractor-implement combinations, or on large, self-propelled units. A working unit is, in this context, defined as an individual unit working on its own with a prescribed job. A fully integrated ICT for mobile work units in arable farming is seen as comprising three main communicating elements: the central unit being responsible for supervising the general job execution by sending specific tasks to the mobile main work unit, handling unexpected events, managing results of the performed job, and updating a farm database with work results the main work unit is an independent vehicle that is able to traverse a field and reach specific and planned locations. Coordinates of the locations are sent from the central unit as part of the task description the implement is a device along with its software connected to the main work unit. It is responsible for performing the planned agronomic operation according to the pre-set settings The embedded task management has the following workflow: planning field tasks and/or operations using software on a FMIS in the farmers or contractors office the task data produced by the planning software is converted to the data format required for the implement control units the task data is transferred wireless to the task controller of the mobile work unit the task controller uses the task data to transmit process data to the ECU on the implement. the task controller collects task data the collected data are transferred wireless to FMIS the collected data are evaluated by the FMIS 2.1.3. Selected work units By using the derived definition of a work unit a number of work unit configurations were selected for the InfoXT project to be included in the scenario construction. The selections were divided into 2 types of configurations: tractor-implement work unit: o equipped with an ISOBUS communication system. o equipped with a machine specific communication system. supporting network infra structure: o on-line wireless communication (PDA, black box terminal, GPRS, Wlan, etc.) o cell phone communication 6

The further specifications of the selected work units should all be finalised according to the functionalities described in the previous Sections. Requirements topics include rugged on-board computers (laptop, PDA ), user adapted interfaces in terms of interactions and displays, and type of wireless network (GPRS, Wlan, etc) 3. Information model 3.1. Entities involved with field operations On the arable farm, field operations are carried out in relation to a number of different objects, like field, crop, operators, machinery, etc. Figure 4 gives the objects involved in field operation management together with the main attributes describing the needed, as well as the interrelations between objects Operators funktion wage hours working Farm name address enterprise Machinery machinery capacity Land acreage location Field location length breath shape area Field characteristics Planned field period value Observed field tim e value Planned use period duration Realized use tim e duration Work method type activities machinery labour Planned operation period duration Im plem ented operation tim e Operation type characteristics Labour/machinery use operation labour/machinery Production product produced Product type Machinery characteristics Product characteri- stics Planned machinery period value Observed machinery observation value Planned period value Observed tim e value Cropping method method operations period implementation Crop type Observed weather tim tim e e condition value Observed Crop characteristics Observed Planned crop tim period e value Observed crop tim e value Observed Forecast period tim e characterivalue stics Figure 4. Objects involved with the planning and implementing of field operations (Sørensen, 1999). 7

The main components involve the farm and a number of fields hosting different types of crops. Production resources include operators and other types of work crews together with their use of machinery and equipments. Another component is the product or product mix identifying the output from the production process or operation (e.g. harvested yield). The focal element is the operation as depicting the operational activity performed by the resources on the fields and crops. The operations are carried out according to some specified work method describing, for example, the type of machinery items involved. The associated with the operational activities of the farm is described in the attributes of the elements listed in the above figure. The attributes specify the kind of knowledge and data relevant to the decision-making processes connected to the planning and implementation of the operations. These attributes include planned and observed of the entities, where the observation can be done directly by human observation or by the use of some monitoring device. It is important to understand the concept of an operation and, for example, a task as part of the operational activities on the farm (Sørensen, 1999). An formal definition of an operation is given by van Elderen (1977), who states that an operation is a technical coherent combination of treatments by which at a certain time a characteristic change of condition of an object (a field, a building, an equipment, a crop) is observed, realised or prevented. This definition extends operations beyond those for crop production to supporting enterprise functions like maintenance, repairs, observation, etc. An operation is generally seen as the link between some resources (e.g. labour and machinery), some materials processed, and some material produced (e.g. harvested crops, repaired machine, etc.). Table 1 gives the various definitions applying to farm operations.. Table 1. Structure of arable farm operations Cultural practise Cultural practises are seen as the human intervention in the agricultural crop production system. Examples include soil tillage, seeding, fertilising, plant care, and harvesting. Operation To realise the cultural practises the farm manager choose one or more operations, i.e. operations deals with what should be done to meet the objectives set by the cultural practises. Specifications like working speed, working depth etc. must be given for each operation. Working method The method by which the individual activities within and operation or chain of operations are carried out and co-ordinated with each other. Task A task defines one or more operations carried out by a group of workers and machinery items, a work-set, working physically together (e.g. operator1 + tractor + sprayer). Each work-set carry out one or more operations simultaneously or sequentially following defined specifications, a certain working method, and on a specified object. Attributes: - type - Attributes: - operation type - work method type - specification for execution - Attributes: - type of work method - sequence of activities - number and type of equipment - Attributes: - task type - operation type - 3.2. Managing field operations When field operations and tasks are planned and implemented, the central aspects of Figure 4 will evolve around farmers acquiring on the current or future states of the objects involved. The farmers will then make their planning efforts and subsequently revise their plans according to the observations made. This mechanism is shown in detail in Figure 5. 8

Tactical planning Formulation of of the the operational plan Operational plan Control and adjustment of the operational plan Work plan Scheduling Planned operation Observation Observation Observed Status to be observed Observed Decision making Required operation Required Implemented operation Operation to be carried out Implementation Figure 5. Task management for field work (Sørensen, 1999) The task management is the tool to plan and evaluate work in the field. The decision processes are centered on specifying what, where, how, by whom, and when the field work should be carried out. On the basis of a tactical plan, the formulation of an operational plan can be carried out. In the course of time this plan can be adjusted while following observations of the crop and the forecast of for instance the weather as well as the results of already executed operations. The implementation of the operational planning will follow the below scheme: a planned operation is reported to the decision making process. Based on some observed the decision-maker decides whether or not an operation is required a required operation is reported to the scheduling process. This process co-ordinates the required operation with the other operational activities on the farm. On the basis of the capacity, availability and priorities the labour and machinery process determines which operation to implement and at what time the implemented operation is reported to the control/adjustment process for further planning and evaluation. As will be noted, the operational planning process in agriculture is highly dynamic and interactive. That makes heavy demands on any proposed planning system, which will have to cover the creation of a schedule of work processes over a longer period (predictive scheduling) and the adaptation of an existing schedule due to actual events in the scheduling environment (reactive scheduling). When a task is formulated and transferred to the mobile unit for implementation, the modification and revising of the task, may be done as part of the operational planning or directly by the operator through a task controller interface (e.g. virtual terminal). 3.3. Information model for spraying The case of plant care or spraying application is selected for the derivation of data flow and handling. In the notion of Figure 4, the case depicts the spraying operation, where this operation combines machinery items (tractor + sprayer), a crop which is in the need of herbicide, and generates the product of a crop being in a new state cleared of weeds, fungicides, or insects. 9

3.3.1. Process decomposition The process model describes the flows that is interchanged between the different processes. Analyzing processes in an system is done independently of the possible technical solutions and possible types of architecture. In the case of spraying, the activities of the spraying operation are described in a targeted process model. A business like arable farming involving field operations is decomposed into functions and processes following the principle of Figure 6. Arable farming : Business process Strategic planning Tactical planning Operational planning Execution Evaluation : Functions Planned operations Operational crop planning Handling tasks Execute tasks : Processes Revision of task Task formulation in in terms terms of of instructions route planning for and vehicle/implement task scheduling formulate jobs formulate jobs formulate tasks formulate tasks modify task modify tasks required operations operations urgency operations specifications select task inspect task control task control operation control device : Sub-processes Figure 6. Process decomposition of arable farming as a business process Functions include strategical planning, tactical planning, operational planning, marketing, management of land, management of labour and machinery, evaluation, etc. These functions are further divided into possesses which is demonstrable and which has a clear starting and ending point. For example, in the case of the operational planning, the continuous natured function encompasses precisely defined decision activities concerned with formulating, controlling and adjusting the operational plan and implementation of specific field operations in the short term (day, week). Each process requires a process description and flow, which gives the definition of the process, what is necessary for the process to perform, and what is made available by the process after execution. In the following, the different functions, processes, and sub-processes making up the operational planning, the execution and the evaluation are defined: Function: Strategic planning The strategic planning function comprises activities concerned with determining/changing the organizational structure and physical entities (e.g. machinery items) of the farm production system. It concerns decisions which have consequences for a longer period than 1-2 cropping seasons. Formulating the spraying target Determining the overall spraying strategy of the farm based on internal and external preferences and possibilities. 10

Information flow: The formation of the spraying target of the farm is based on like new chemicals available on the market, current and new spraying requirements/restrictions formed by legislative bodies, new and changed possibilities in crop production and product prices, and new and changed demands from the market in terms of required documentation and traceability. The latter is especially important as it may be seen as the market feed-back to the production system affecting the technology requirements in terms of required sensor as an example. The output from the process is a determination of the technology requirements in order to fulfil the planned spraying target. Farm financial plan Creating the financial plan for the whole farm, based on book keeping and available technology costs. Information flow: Farm financial plan get s from companies and dealers concerning available technology. The realized economic is available from the database. The economic is the finance from the past years. Budget concerning the possible technology acquisition is an output from the decision process. Choosing spraying technology Selection the best adopted spraying technology based on the expected production possibilities and constraints for the coming years. Information flow: The primary input to the choosing of spraying technology process is the specified technology requirements and the financial possibilities. Also, the expected crop types and crop acreages together with the current and historical spraying performance are important input to this decision process. The output from the process is the actually selected spraying technology, which is then passed on to the tactical planning level. In order to derive the diagrams for the different planning levels in precision spraying a number of usage processes (decision making) and providers (internal and external entities) were identified see Figure 7. 11

Legend for Information Flow in Precision Spraying Process Information Actor Markets Adv. organization Information usage The that actor offers Information producer Different buyers and sellers in food industry Advising organization: agricultural expert organization Information flow Alternative flow route Legislative rules Governmental/EU/Environmental rules for the spraying process Weather service Local weather provider External service Agricultural service company Farmer, user Decision maker Farmer Actors Databases 1 n Task Controller Data warehouse The number of spraying units Communication device between working unit and external system (e.g. FMIS) Sprayer ECU Implement controller computer User Interface (VT) The Virtual Terminal (VT) is a common user interface for all ISOBUS compatible implements Tractive unit Tractor, self propeller sprayer or robot Internal sensors Sensors for the online controls External sensors Sensors for the monitoring field process, implement or environmental Company, dealer Agricultural machinery, hardware and product companies Figure 7. Legend for the Information Flow 12

Markets New demands for the spraying Change of product prizes Adv. organization New cropping practices Farm financial plan Budget Legislative rules New spraying requirements Weather service External service Farmer, user Formulating the spraying target Choosing spraying technology Databases Spraying history Farm economic Technology requirements Expected crop production plan Current spraying performance Selected sprayers 1 n Task Controller Sprayer ECU User Interface (VT) Tractive unit Internal sensors External sensors Company, dealer New chemicals Available technology and costs Figure 8. Strategic planning STRATEGIC PLANNING Time Function: Tactical planning Tactical planning comprises activities concerned with the planning of spraying tasks involved with the new cropping season Selecting functions Selecting the required functions of the sprayer or sensors Information flow: The input to this process is possible new technical from the sprayer dealer concerning sprayer functionalities. The selected comprising which sprayer functions to update is outputted to the sprayer dealer and/or database for further downloading to the task controller as input to the updated parameters of the sprayer equipment control unit (ECU). Sub-process: Updating Updating the sprayer and sensors 13

Information flow: The updating is carried out based on selected update and output is the resulting parameters of internal sensors and sprayer ECU. Sub-process: Recommend chemical use Recommend a farm specific list of usable chemicals Information flow: Input comprise the crop production plan, current provisions on chemical use, and the historical spraying practises of the farm. The output comprises a farm specific list of usable chemicals, which is input to the process of selecting planning. Selecting planning Recommend a farm specific list of usable chemicals Information flow: a) Input comprises the specific list of chemicals (produced by advising organization) and the selected sprayer. b) Input comprise the crop production plan, the selected sprayer, current provisions on chemical use, the historical spraying practises of the farm, restrictions on chemical uses, and a suitable list of chemicals to be used The output (a and b) comprises farm specific crop protection, which is input to the further planning process of the spraying work in the coming season. Spraying process planning Planning the expected use of external services, chemical acquisition, and the planned amount of spraying work Information flow: Input comprises the updated crop protection and current storage of chemicals. The output comprises the expected use of external sensing services (e.g. aerial imaging), the necessary additional chemical acquisition, and the expected amount of spraying work to be carried out, which is input to the operational planning process. Finally the chemical properties are updated to the database. 14

Markets Adv. organization Recommend chemical use Regional list of chemicals List of suitable chemicals Legislative rules Restrictions on chemical use Weather service External service Expected sensing services Planned spraying work Farmer, user Selecting functions Selecting planning Spraying process planning Databases Updating Expected crop cycle Spraying history Selected sprayers Updated crop protection Chemicals in storage Chemical properties 1 n Task Controller Updating Sprayer ECU Updated parameters User Interface (VT) Tractive unit Internal sensors Updated parameters External sensors Company, dealer New technical Selected update Updating Acquired chemicals Figure 9. Tactical planning TACTICAL PLANNING Time Function: Operational planning The function operational planning comprises activities concerned with formulating, controlling and adjusting the operational plan and planning implementation of the spraying work in the short term (day, week) Selecting final sensing services The final selecting and ordering of external sensing services Information flow: Input comprises documented field operations from the time of seeding and the previous planned spraying work. The output comprises the final amount of external sensing services to be ordered. Field inspection Dedicated sensing/measuring operations carried out for specified fields 15

Information flow: Input comprises the selected and specified sensing services and the documented field operations from the time of seeding. The output comprise farm and field specific spatial (e.g. aerial imaging on biomass) Formulate operational spraying plan Task formulation comprising activities for specifying and scheduling the spraying operation Information flow: Input comprises spatial field measurement/, historical weather as indicator of the average weather, updated local crop, updated norm application rates, etc., sprayer functionality, and possible additional field (e.g. latest observation, supplemental treatments). The primary output comprises a specified Task-file with specifications including VRA map or control settings for the operation of the sprayer unit together with an expected spraying schedule. Also, as out put from the task formulation process, continuously updated on the of the TASK-file creation and completeness is made available.. Markets Adv. organization Updated local crop Updated norm Legislative rules Weather service Historical weather External service Planned spraying work Selected sensing services Field inspection Formulate operational spraying plan Farmer, user Selecting final sensing service Additional field Status of TASK-file Databases Documented field Spatial field Sprayer TASK-file Expected spraying schedule 1 n Task Controller Sprayer ECU User Interface (VT) Tractive unit Internal sensors External sensors Company, dealer OPERATIONAL PLANNING Time Figure 10. Operational planning 16

Function: Execution The function execution comprises activities concerned with the initiating and controlling the execution of the planned task Field observation The timely observation of the current crop condition Information flow: Input comprises field identifying the relevant fields and external received on possible occurrence of plant diseases, weeds, etc. The output comprises the immediate of the crop in terms of actual disease and weed occurrence. Select formulated task for realization Select formulated TASK-files for execution and derive an updated spraying schedule Information flow: Input comprises the observed actual crop condition and the actual weather together with the weather forecast for the coming days. The output comprises the final specified and selected TASK-file as well as the updated spraying schedule indicating the timetable for the pending spraying operation. The specifications of the TASK-file (specifications on the spraying operation, like dosage, expected pressure levels, etc.) are downloaded to the task controller as default values. Based on these default values and actual weather conditions, recommended parameter or set values are generated by the sprayer ECU and show on the user interface for consideration. Also, the initializing parameters are generated for sensors that can control the spraying process. (e.g. Yara N-sensor). Selection of spraying parameters Selection of recommended parameter values Information flow: Input comprises the selected spraying schedule and the actual weather conditions as the basis for selecting the best adopted spraying parameters. The output comprises the selected spraying parameters at the time of initiating spraying execution as well as the continuously updated parameters and schedule due to changing weather conditions and forecasts. Inspecting and controlling the spraying task Supervision and controlling of the operations specifications Information flow: Input comprises the updated spraying parameters and schedule, the spraying unit, and overall task and operation. Internally within the sprayer ECU, a control process is running with inputs like process (e.g. online Yara sensor), and set values for the spraying operation, etc. Based on the set values and the actual spraying performance controls are realised. The realised controls are input to the realised spraying work which again is input into the continuously overall operation (this operation might include multiple spraying units). Based on the operation, the operator or some control systems might generate correcting which are fed into the task controller and the realised controls affecting the overall task performance. 17

The spraying equipment may be equipped with external sensors providing on various parts of the spraying performance. This monitoring and documentation is inputted into the task controller for later usage. Markets Adv. organization Legislative rules Weather service Actual weather and forecast External service Plant disease alarm Farmer, user Field observation Select formulated task for realization Selection of spraying parameters Inspecting controlling spraying task Databases Field Actual crop condition Selected TASK-file Selected spraying schedule Weather Forecast Operation 1 n Task Controller Default values Actual weather condition Forecast Raw data Realized spraying work Updated task Sprayer ECU Recommended parameter values Updated parameters Updated timetable Realized controls User Interface (VT) Information for tank filling and mixing Updated timetable and parameters Spraying Overall Task monitoring Tractive unit Status Internal sensors Initializing parameters Process External sensors Monitoring Company, dealer EXECUTION Time Figure 11. Execution Function: Evaluation Evaluation of the executed spraying operation and task Define processing needs Determine the amount and scope of data processing Information flow: Input comprises the realised spraying work and raw data from monitoring making up the documentation from the spraying operation. Also, input comprises the dedicated required registration set up by legislative bodies. The output comprises the selected services required by advising organization or external services in terms of data processing for external requirements or internal management purposes. 18

Sub-processes: Data handling Dedicated data and processing Information flow: Input comprises the determined and selected data handling services together with overall amount of documentation available in the database. Also the summarized spraying performances from other farms are used if evaluation processes are included. The output from the various data handling processes comprises processed documentation for the future planning of the spraying activities and processed usable for dedicated traceability and documentation requirements from the customer/market. Evaluation of spraying performance Determine the spraying performance and economics based on an evaluation of the planned task versus the realised task Information flow: Input comprises task documentation, summarised local weather during the spraying season, the selected planned TASK-file, and norm data from various sources on average spraying performance, effectiveness of chemicals, etc. The output comprises the realised spraying performance (e.g. spraying capacity, operation efficiency, chemical used, etc.) and the economics of the spraying operation (e.g. input costs, cost-benefits, etc.). 19

Markets Traceability Adv. organization Selected services Data handling Norm data 1 Spraying economy performance Farm financial plan Legislative rules Required registrations Weather service Crowing season weather External service Selected services Data handling Norm data 2 Farmer, user Define processing needs Evaluation of spraying performance Databases Documentation Selected TASK-file Summarized local weather Current spraying performance 1 n Task Controller Realized spraying work Raw data Sprayer ECU User Interface (VT) Tractive unit Internal sensors External sensors Company, dealer Figure 12. Evaluation EVALUATION Time 20

3.3.2. Data modeling Based on the identified flows associated with the management functions and controlling of the spraying operation, the data inherent in the flows is identified. Table 2. Strategic planning Entity Definition Attributes/data New chemicals - name of the agent Account of new chemicals on the - name of supplier market - chemical formula Available technologies and costs Farm economic New spraying requirements New cropping practices New spraying requirements New demands for the spraying Change of product prices Technology requirements Budget resources Crop production plan Current spraying performance Selected sprayers Spraying history Description of available technologies on the market and costs The economic and forecast data for the farm Description of new imposed legislative requirements New crops to be grown or new cropping practises New operational spraying requirements Description of new imposed legislative requirements Description of changed product prices Description of derived technology requirements Description of the expected revenues allocated to acquiring spraying equipment Description of expected crop production in the coming growing season Description of the historical spraying performance Description of selected sprayer for spraying Description of the historical spraying operations at the farm - type of spraying equipment - technical specifications (tank volume, boom width, pressure levels, etc.) - name of supplier - price - development plans for the farm - expected revenues - expected conditions for growing season - type of requirement - threshold values for chemical use - type of chemical allowed - type of crops - name - description - new types of infection/infestation - type of requirement - machine set values - type of requirement - threshold values for chemical use - type of chemical allowed - type of product - differentiated prices - type of requirement - quantified requirement - type of chemical allowed - expected investment amount - threshold values for chemical use - type of chemical allowed - production plants and areas - timetable draft - input draft - quality of work - costs of operation - type of sprayer - specifications (tank volume, boom width, etc.) - schedule of spraying operations - operations criteria for execution (weather conditions, etc.) - type of chemical - normative dosage - realised dosage 21

Table 3. Tactical planning Entity Definition Attributes/data New technical - type of Description of potential new technical - type of solution and solutions to be - specifications (pressure levels, updated boom height, nozzle, sensors, etc) Selected update Updated parameters Spraying history Restrictions on chemical use Regional list of chemicals List of suitable chemicals Updated crop protection Chemicals in storage Chemical properties Description of the selected update Description of the updated parameters Description of the historical spraying operations at the farm Description of the restrictions on chemical use Description of the potential regionally specific list of chemicals Description of the selected farm specific list of chemicals Description of the crop specific chemicals for the selected sprayers Description and the amount of instorage chemicals Description of the chemical properties of the selected chemicals - type of - type of solution - specifications (pressure levels, boom height, nozzle, sensors, etc) - type of parameter - setup values - updated files and programs - new hardware descriptions - schedule of spraying operations - operations criteria for execution (weather conditions, etc.) - type of chemical - normative dosage - realised dosage - legislative restrictions - operational restrictions - specifications (dosage, type of chemical, etc.) - type of chemical - recommended dosage - operational restrictions - specifications (nozzle type, pressure levels, mixture ratio, boom height, etc) - type of chemical - recommended dosage - operational restrictions - specifications (nozzle type, pressure levels, mixture ratio, boom height, etc) - type of chemical associated with type of crop - recommended dosage - operational restrictions - specifications nozzle type, pressure levels, mixture ratio, boom height, etc) - preliminary spraying schedule - type of chemical - amount of chemical - type of chemical - name of the agent - name of the supplier - chemical formula - active component - content active component - duration activity - residue tolerance - safety period 22

Acquired chemicals Description of the necessary - type of chemical purchased chemicals in the season - amount of chemical - type of sensing service (e.g. aerial Expected sensing services imaging) Description of the planned use of - planned use of the sensing service external sensing services - planned schedule for use of the sensing service Updating Specific settings and update for the sprayer - type of update - setup values - updated software - updated hardware and its Expected crop cycle Description of crop cycle - type of crops - preliminary spraying schedule Selected sprayers - type of sprayer Description of selected sprayers for - specifications (tank volume, boom spraying Planned spraying work Description of the expected and planned spraying work in the coming spraying season width, etc.) - field ID - type of chemical to be sprayed - planned dosage - workable weather conditions - operational specifications (nozzle type, pressure levels, mixture ratio, boom height, etc) Table 4. Data model for operational management Entity type Definition Attributes/data Planned spraying work Selected sensing services Documented field Spatial field Updated local crop Description of the expected and planned spraying work in the coming spraying season Description of the final selected sensing services Description of the prior to spraying executed field operations and analyses Description of sensed spatial field Description of updated local crop based on growing season conditions - field ID - type of chemical to be sprayed - planned dosage - workable weather conditions - operational specifications (nozzle type, pressure levels, mixture ratio, boom height, etc) - order ID - type of sensing service (e.g. aerial imaging) - planned use of the sensing service - planned schedule for use of the sensing service - field ID - type of operation (e.g. seeding, fertilizing) for the present growing season - actual date of execution - actual amount of application (e.g. VRA) - field specific (e.g. soil properties) - type of measured parameter - spatial distributed values of measured parameter (classified canopy map) - type of crop - values for updated crop (crop condition, expected development, etc.) 23

Updated norm Sprayer Additional field TASK-file Status of TASK-file Expected spraying schedule Historical weather Description of updated norm Description of the sprayer and its functionalities Description of the additional field Description of control settings for the sprayer Description of the completeness of the TASK-file Description of the planned spraying schedule Description of the realized crowing season weather - type of norm - updated set-values for specific norm (e.g. application rate, workable weather conditions) - type of functionality (e.g. pressure range, nozzle arrangement) - functionality values or ranges and set-points - type of (e.g. latest observation, supplemental treatments) - quantification of additional (e.g. spatially observed weeds) - field ID - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates, driving speed, boom height and width, documented parameters, variable rate application (VRA) map) - control settings value for the specified types of settings - level of finalizing the TASK-file - field ID - expected data of executing spraying operations - expected dosage to be applied - workable weather conditions - field ID - local weather Table 5. Execution Entity Definition Attributes/data Plant disease alarm External plant disease alarms - type of alarm (e.g. fungicide) - prognosis for occurrence - field ID - crop type Field Description of needed field - crowing - farming actions - known risks based on previous field observations on the farm Actual crop condition Current of the growth - field ID - current growth - type of observation (e.g. weeds or fungicides) - level of occurrence Selected TASK-file The selected TASK-file for execution - field ID - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates, driving speed, boom height and width, documented parameters, variable rate application (VRA) map) - control settings value for the specified types of settings 24

Actual weather and forecasts Selected spraying schedule Default values Recommended parameter values Initializing parameters Information for tank filling and mixing Updated parameters Updated timetable Operation Realized spraying work Updated task Current weather and short term forecasts The selected spraying schedule for implementation Downloaded default values in the TASK-file Selection of the best adopted spraying parameters Default values for external sensors (e.g. Yara sensor) Guidelines for filling and mixing chemicals in the sprayer Adopted ECU parameters Adopted spraying timetable Current operation Executed spraying work and documented supporting Revised and updated task specification - type of weather parameter - parameter value (e.g. temperature, wind, humidity, precipitation) - forecast probabilities - expected time of executing spraying operations - expected dosage to be applied - workable weather conditions - field ID - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates, driving speed, boom height and width, documented parameters) - control settings value for the specified types of settings - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates, driving speed, boom height and width, current weather) - control settings value for the specified types of settings - type of set parameters - setting values - type of chemicals - type of nozzles - ingredient rate - water dilution - mixing rates - driving speed - wind speed - humidity - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates) - control settings value for the specified types of settings - type of spraying - time for individual spraying operations - specified spraying unit - current capacity - current operation progress - remaining spraying work - of external system (e.g. dryer) - acreage sprayed - applied amount of chemicals - weather - process data (e.g. farmer notes, fuel consumption) - field ID - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates, driving speed, boom height and width, documented parameters, variable rate application (VRA) map) 25

Realized controls Raw data Process Monitoring Actual weather conditions Weather Forecast Updated timetable and parameters Spraying Status Overall task monitoring Invoked control parameters Raw monitoring of operational data Control from external sensors (e.g. Yara sensor) Supervision Current weather Current weather Short term weather forecasts Adopted spraying timetable and ECU parameters Description of spraying process Description of tractive unit Real-time and adjustments - control settings value for the specified types of settings - capacity compliance (for example, the need for additional capacity) - of external system (e.g. dryer) - type of realized control (e.g. spraying pressure, dose rate, driving speed, headland automation) - control values - diagnosed (e.g. faults) - undefined online operational parameter (raw data) - type of operational parameter - log of operational parameter - online measurements (e.g. spatially calculated biomass amount) - type of operational parameter - log of operational parameter - type of weather parameter - parameter value (e.g. temperature, wind, humidity, precipitation) - type of weather parameter - parameter value (e.g. temperature, wind, humidity, precipitation) - weather forecast probabilities - parameter value (e.g. temperature, wind, humidity, precipitation) - type of spraying - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates) - control settings value for the specified types of settings - time for individual spraying operations - current control values (e.g. spraying pressure, dose rate, driving speed, headland automation, alarms) - control values - type of tractive unit parameters parameter value (e.g. fuel consumption, driving speed, pto) - overall spraying task - capacity compliance - of external system (e.g. dryer) - spraying proses adjustments Table 6. Evaluation Entity Definition Attributes/data - acreage sprayed Realized spraying work - applied amount of chemicals Executed spraying work and - weather documented supporting - process data (e.g. farmer notes, fuel consumption) Raw data Raw monitoring of operational data - type of operational parameter 26

Documentation Required registrations Selected services Traceability Norm data1,2 Selected TASK-file Summarised local weather Current spraying performance Spraying economy performance Overall spraying documentation Selected compliance data Targeted services for data handling and modelling Selected tracing Norm data on spraying performance from various sources Executed TASK-file Weather occurrence during spraying operation Estimated spraying performance Costs of farm spraying operations - log of operational parameter - type of parameter - log of parameter - field ID - e.g. time of spraying - e.g. applied amount of chemical - e.g. type of applied chemical - - type of data handling - identification of service - field ID - e.g. time of spraying - e.g. applied amount of chemical - e.g. type of applied chemical - - average spraying capacity - effectiveness of chemical - threshold for applied amount of chemical - - field ID - type of setting (chemical 1 n, nozzle type, nominal dosage, mixture rates, driving speed, boom height and width, documented parameters, variable rate application (VRA) map) - control settings value for the specified types of settings - weather period - type of weather parameter - parameter value (e.g. temperature, humidity, precipitation) - average capacity (min-max) - field efficiency index - acreage sprayed - applied amount of chemicals - energy input for spraying - quality of spraying work - labour input - technology costs - capital costs - operational costs 4. References Chatila, R. 995. Deliberation and reactivity in autonomous mobile robots. Robotics and Autonomous Systems 16 (1995) 197-211 Goense, D. ; J. W. Hofstee; Van Bergejk.1996. AN INFORMATION MODEL TO DESCRIBE SYSTEMS FOR SPATIALLY VARIABLE FIELD OPERATIONS. Computers and electronics in agriculture 1996, vol. 14, no 2-3, pp. 197-214 Martin-Clouaire, R. & Rellier, J.P., 2000. Modelling needs in agricultural decision support systems. Proc. XIV Memorial CIGR World Congress 2000. 27

Norros, L., 2004. Acting under Uncertainty. The Core-Task Analysis in Ecological Study of Work. Publications 546. Espoo: VTT, Available also URL: http//www.vtt.fi/inf/pdf/. Nurkka, P., Norros, L., & Pesonen, L., 2007. Improving usability of and user acceptance of ICT systems in farming. Paper presented at the EFITA/WCCA Joint Conngress in IT in Agriculture, Glasgow. Pesonen, L., Nurkka, P., Norros, L., Taulavuori, T., Virolainen, V., Kaivosoja, J., Mattila, T., & Suutarinen, J.,2007. Kasvinviljelyn asianhallintajärjestelmän käyttäjäkeskeinen kehittäminen (Usercentred developoment of a farm management system) (97). Vihti: Maa- ja elintarviketalouden tutkimuskeskus (Agrifood Research Finland). Rasmussen, J. 1986. A Cognitive Engineering Approach to the Modelling of Decision Making and Its Organization, Risø National Laboratory, Denmark Scheepens, A.J., 1991. Information modelling for arable farming. PAGV Report, 133, 118 pp Sørensen, C.G., 1999. A Bayesian Network Based Decision Support System for the Management of Field Operations. Case: Harvesting Operations. Ph.D.-Thesis, Technical University of Denmark, 193 pp. Sørensen, C.G., Olsen, H.J., Ravn, A.P. & Makowski, P., 2002. Planning and Operation of an Autonomous Vehicle for Weed Inspection. ASAE Ann. Int. Meeting/CIGR XVth World Cong. Chicago, Illinois, USA, 2002. ASAE paper 021177, 9 p Stoll, A. 2003. Automatic operation planning for GPS-guided machinery. In: Proceedings of the 4th European Conference on Precision Agriculture, 15-19 June, 2003. Berlin. pp 657-664 Suomi, P.; T. Oksanen; L. Pesonen; J. Kaivosoja; H. Haapala; A. Visala 2006. Intelligent functions for crop production automation. Proc. Of XVI CIGR World Congress 2006, Bonn Van Elderen, E. 1977. Heuristic strategy for scheduling farm operations. Centre for Agricultural Publishing and Documentation, Wageningen, The Netherlands, 217 pp. Vicente, K. J. 1999. Cognitive Work Analysis. Toward a Safe, Productive, and Healthy Computer-Based Work. Mahwah, NJ: Lawrence Erlbaum Publishers. 28