(version 21feb22)

This document describes the AgGateway charter 'Closed Loop Spraying'.
The aim of this initiative is to support, as a producer of chemical crop protection products, the contractor and the farmer in the best possible way in the use and application of crop protection products.
This meets the industry's claim to "Help growers avoid mistakes", the claim to Compliance and traceability and the claim about good "Stewardship of product usage".
To be clear, with "Closed Loop Spray" we mean "end-to-end documentation and resource identification" and not "real-time feedback-driven control".

Objectives

• Realize a closed loop in data exchange concerning: product identification -> product master data -> to be applied maps -> as-applied maps to improve the optimal use of chemical crop protection products:
  • to help growers avoid mistakes.
  • to ensure compliance and enable traceability.
  • to exercise good stewardship of product usage.

Scope

The solution should be applicable at a global level.
Technical components involved are: the FMIS', databases with crop protection products reference data (catalog data), spraying machinery and equipment, platforms to communicate the data.

In scope:

• Managing buffer zones is in scope.

Out of scope:

• Interference with the the sprayer's control system itself, as it runs on the machine, is out of scope.

Start situation

• There are existing AgGateway standards for catalog, work orders and work records, that are relevant. This initiative provides the opportunity to tie these together - fill gaps on discovery, protocol negotiation (like OAuth orchestration) and archival.
• There are existing work order and work record data models that can be leveraged.
• There has been a request for the next stage of the product catalog to include application and safety data, although nothing has been committed yet.

Start principles

• Make use of available global standards.
• Make use of existing AgGateway standards and solutions:
  • e.g. the existing standard work order and work record design.
  • e.g. AgGateway's Product Catalog.
• Make use of existing databases with crop protection product reference data such as the Homologa database (Lexagri).
• Make use of existing AEF standards.

The challenge

• To create a system/protocol that ensures compliance and minimizes mistakes in the use of crop protection products – by creating a way for the spraying equipment, the farm management information system (FMIS) and the input product to "talk to each other".
• This includes:
  1. Product Discovery – Say when an input is loaded on the machine, that the system enable unambiguous product identification discovers which product has just been loaded – either through an RFID or a GS1 QR code or something similar.
     This means "that the system enable unambiguous product identification" rather than "the machine discovers", which would be too close to solutioning.
  2. Label Data Compliant Work Order – The machine and the FMIS are able to query a catalog data source for this product to find the agronomic constraints (wind speed, buffer zone, crop, etc.). The FMIS is able to generate work orders, taking the constraints into consideration, and the machine is able to enforce these constraints.
  3. Verifiable Work Record – The spray as carried out is recorded in a verifiable way that can be used for regulatory compliance, or traceability purposes.
• To be clear, the operator of the system – the farmer or the contractor that drives the tractor or the sprayer – always remains in control and takes the final decision. Its is not aimed that the machinery takes over the decision making form the operator.
  The idea is that the system warns the operator in case a serious risk is detected for the operator to make a mistake.
  The decision to adhere to the presented restrictions – buffer zone etc. – is up to the applicator and not "stopped" by the equipment based on the rule

State of art and what is to be done

Product Discovery

Say when an input is loaded on the machine, the operator or the machine discovers which product is to be loaded – either through an RFID or a GS1 QR code or something similar.

• Concerning the state-of-art including a QR code or 2D-matrix code on the packaging labels:
  • In Europe, the Cristal recommendations for applying the 2D matrix code on the labels is being implemented in the AgroCloSer project.
  • The 2D matrix code contains: the GTIN, the batch number, the production date and (optional) the serialization identifier. This makes it possible to automatically identify the product when using the product in the field, by scanning the code. The scanned GTIN can than be used to retrieve product related data from any data source.
  • Distinguish the carrier (QR code, 2D matrix code, bar code, RFID) from the identifier (the payload with the GTIN or another type of identifier).
  • It is important to align with the CropLife work that is being done in this field. In other regions than Europe, other types of identifiers might be used; this should not be a problem, as long as it is clear what type of identifier is used on the label (there is room for this information in the 2D matrix code) and that that same identifier can be used to retrieve data from external data sources.
  • The Closed Loop Spray project should contribute to:
    • Promote the CropLife Cristal recommendations in implementing the 2D matrix code on labels.
    • Promote the use of GTIN’s as the standard identifier for products and articles.
  • Equipment, FMIS and input product should talk to each other. Communication between equipment and FMIS is already established through ISOXML. How to get the input product included? The RFID or QR code is the “how”, a technical solution that probably shouldn’t be proposed in the charter yet. Is the reader on the machine or is it an app on the phone and the data gets processed in the cloud and the machine just receives the Rx map online?
  • Future implementations need to support the use of different types of identifiers in different global regions; in Europe the use of GTINs is recommended, In North America a different type of identifier might be recommended. 
  • A catalog data source would be very valuable and help every FMIS system to improve planning.
• Align with the AgroCloSer initiative:
  • The AgroCloSer project is supported by the main producers (BASF, Bayer, Syngenta, Adama, Certis, a.o.) and distributors of crop protection products.
  • Key objectives of the AgroCloSer initiative are:
    • Provide all packaging labels with the Cristal 2D matrix code so inbound and outbound product flows can be automatically scanned and identified.
    • Make use of standard Order – DespatchAdvice – Invoice messages to exchange data between the supply chain parties; so the implementation will not only enable track & trace but will also support a more efficient supply chain management.
    • Make use of GS1 identifiers: GLN, GTIN, SSCC.
    • Connect to the ERP and WMS systems that are used by producers and distributor.
    • Make use of an existing platform (e.g. Proagrica) to exchange the data; it supports different protocols and syntaxes to get connected.
    • Look into the Label Data Compliant Work Order.

Label Data Compliant Work Order

The machine and the FMIS are able to query a catalog data source for this product to find the agronomic constraints (wind speed, buffer zone, crop, etc.). The FMIS is able to generate work orders, taking the constraints into consideration, and the machine is able to enforce these constraints.

• Once the product is identified, a catalog data source can be queried for information / constraints about the application of the product.
• Taking buffer zones into account: 
  • In addition to label compliance, there are regional requirements that go beyond the label. We need to survey, for instance, the regional buffer requirements. We need to have an awareness of GIS layers that mark exclusion zones. 
• Before starting executing the spraying, it is important to check the weather conditions and forecasts; weather condition are important restrictions for spraying.
• Questions to be answered:
  • What data stores are available in Europe and in the US to provide this information?
  • What challenges are there to overcome in connecting to catalog data sources?
  • Does the exchange of work orders and work records need extra attention?
• There are data stores in place in Europe and in the US to provide information / constraints about the application of crop protection products:
  • In Europe Lexagri's Homologa database is leading; it provides information about the authorized use (agronomic data included) of crop protection product throughout 80 different countries.
  • In the US,  AGIIS and the Product Catalog are used to provide reference and master data about crop protection products. The Product Catalog provides commercial and logistics data of the products and not agronomic data.
  • Other relevant types of data sources for the machine and the FMIS to query for agronomic constraints (wind speed, buffer zone, crop, etc.) are weather data, soil maps, crop observations, etc. Further harmonizing the interfacing to these data source is something to look into.
  • It is recommended not to start the process in the field, when filling the machine and identifying the product that is to be use, but to start the process when the farmer or contractor is preparing the crop protection tasks that needs to be carried out; this might result in a work order or task map that is fed into the terminal of the tractor or the spraying machine.
• Should the step that is checking for wind speed, buffer zone happen earlier in the process that is when the application map is being generated? Once product is loaded in the sprayer, these checks could serve as audit but not for decision making:
  • Take advantage of the Product Reference work done. For weather data, leverage PAIL project Observations and Management work.
• Pay special attention to managing buffer zones. 
• Look into the results of AgGateway’s SPADE initiative, concerning the planning – preparation – execution of plant protection work orders.
• Look into the results of AgGateway's OK-To-Spray initiative. Get Solinftec involved.
• For data modeling, make use of existing data models (reference model CROP, the ADAPT data model).

Capturing Work Records

The spray as carried out is recorded in a verifiable way, via the work record that can be used for regulatory compliance, or traceability purposes.

• There is little support by farmers and contractors to make reporting work records to third parties completely tamper proof; an achievable aim should be to make the reported work records to some extend verifiable.
• There is little support for developing a Block Chain application word reporting work records.
  Block-Chain Technology is mature and available, if requested for a specific transparency use case, to be used for this purpose. One has to keep in mind that a block-chain needs to be fed by existing systems like ERPs, FMIS’.
  NIST (US) is working in a document addressing the application of BCT in all types of industries. Many are ambivalent on the use of block-chain.
• Start with the sharing of metadata records (ala Dublin Core)? 

Initiatives to align with

• The CropLife work-stream: To define the digital labels and provide recommendations to the growers to remain compliant while planning their spray, and finally thinking about compliance check/audit of the data. Also align with the CropLife 'Closed Transfer System'  initiative.
• Reach out to Daniel Martini (KTBL) about his work on managing data flows concerning crop protection.

Critical success factors

• This initiative needs the support and the input from:
  • Global manufacturers of crop protection products.
  • Global manufactures of spray equipment.
  • The AEF (the Agricultural Industry Electronics Foundation).
• A challenge is getting the cooperation of crop protection companies to put some sort of AIDC label on their products. 

Project organization, governance

• AgGateway Europe takes the lead in organizing the meetings and in supporting the work groups:
  • organize meetings, draft agenda’s, make minutes
  • for each work package the participants choose a chairman to lead the meeting
  • most meeting will be virtual meetings (teleconferences), if necessary and possible there will be face-to-face meetings as well
• AgGateway’s Wiki will be used to publish agenda’s, minutes, draft results.
• Final deliverables will be published by AgGateway’s Digital Resource Center.
• Participants need to be an AgGateway member or be represented by an AgGateway Associate Member (like Agro-EDI-Europe, AgroConnect, AEF, a.o.).

Step wise approach, short sprints

• Work in relatively short sprints for each work package (3 to max 6 months).
• For each work package different experts might be involved.
• Each sprint starts with specifying, as concrete as possible, the desired deliverable(s).
• Each sprint ends with the delivery of concrete results.
• Final deliverables will be published by AgGateway’s Digital Resource Center.

Work packages

• WP1: Identification
  
  • What type of identifier is to be used? A GTIN code for Europe? What type of identifier is used in the US?
  • What type of data carriers are in use and recommended: QR, 2D-matrix code, RFID?
• WP2: Reference data
  • Once the identifier of the product is read, a reference data store can be queried for information / constraints concerning the application of the product.
  • What data stores are available in Europe and the US to provide this information?
    In Europe Lexagri's Homologa database is leading.
• WP3: Generate work orders
  • Get manufactures of tractors, spray equipment and FMIS' engaged to implement generating work orders and to transfer these work orders to the spray machine.
• WP4: Verification work order execution and data generation
  • Verify transferring the FMIS work order to the machine and getting the as-applied-map back into the FMIS.
• WP5: Process the as applied maps
  • Receive and process the as-applied maps. Generate a digital report message about the application of the crop protection product.

<to be worked out in further detail>

Required input work group members

• During a sprint, the work group meets every two weeks.
• Each meeting will take up to a maximum of 2 hours.
• In between the meetings the work group members are asked to do their "homework" as agreed in the previous meeting.
• Work group members make their efforts at their own cost; there is no additional budget to finance the efforts of work group members.

Next steps

(dd February 21, 2022)

• Process the results of the meet-ups.
• Go – no go. As AgGateway, decide if there is enough support to proceed with this initiative.
• If to proceed:
  
  • Break down the work that needs to be done in workable parts (work packages) that can be addressed in short sprints.
  • Specify the task description for the first workgroup to address work package 1.
  • Populate the workgroup for the first work package.
  • Organize the first meeting of a workgroup to start working on the first work package.