WO2025163074A1

WO2025163074A1 - Method and apparatus for generating a monitoring and/or control condition for an agricultural spraying application

Info

Publication number: WO2025163074A1
Application number: PCT/EP2025/052412
Authority: WO
Inventors: John MANOBIANCO; Caitlin Isobel JONES; Davide VODOLA
Original assignee: BASF SE
Current assignee: BASF SE
Priority date: 2024-01-30
Filing date: 2025-01-30
Publication date: 2025-08-07
Anticipated expiration: 2026-07-30

Abstract

The present invention relates to a method for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field and comprises the steps: - receiving a weather data set associated with at least one weather condition at the target agricultural field, - retrieving from a database comprising a plurality of reference weather data sets associated with a plurality of reference weather conditions at a plurality of reference agricultural fields, respectively, a subset of reference weather data sets based on the at least one weather condition at the target agricultural field as provided by the weather data set, and - generating for at least one reference weather data set from the subset of reference weather data sets, the at least one monitoring and/or control condition for the spraying application at the target agricultural field, using the subset of reference weather data sets.

Description

Method and apparatus for generating a monitoring and/or control condition for an agricultural spraying application

FIELD OF THE INVENTION

The present invention generally pertains to the technical field of agricultural spraying applications. In particular, the present invention relates to a method and to an apparatus for generating a monitoring and/or control condition for a spraying application to be performed on an agricultural field. Furthermore, the present invention relates to a method for updating a database that can be used for generating a monitoring and/or control condition for a spraying application. The present invention also relates to a use of a monitoring and/or control condition for controlling a spraying application, e.g., performed with an agricultural device such as spraying device. In particular, the present invention may be useful for predicting a spray drift of a chemical, e.g., in relation to an agricultural field. The present invention may contribute to preventing or at least reducing spray drift of a chemical outside a certain area, e.g., an agricultural field. Thereby, the present invention may contribute to protecting flora and fauna in landscapes comprising agricultural fields.

BACKGROUND OF THE INVENTION

Nowadays, in most countries around the world, there are a plethora of regulations that set boundaries to the application of chemicals onto agricultural fields. In order to comply with such regulations, often models for pesticide risk assessment are employed in advance before actually performing spraying applications on an agricultural field. These models serve to assess the risk of a pesticide to human health or the environment and often consider the toxicity of the pesticide as well as the amount of pesticide to which a person or the envir- onments may be exposed. For example, aquatic models, terrestrial models, spray drift models and health effects models are known and frequently employed.

SUMMARY OF THE INVENTION

The present invention is based on the objective of providing an improved method and apparatus for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field. Furthermore, the present invention is also based on the objective of providing an improved method for updating a database method for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field. The present invention also aims for providing an improved use of a monitoring and/or control condition for controlling a spraying application to be performed on a target agricultural field.

According to the present invention, a method for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field is proposed. In particular, the proposed method may be a computer-implemented method for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field. The method comprises the steps of:

- receiving a weather data set associated with at least one weather condition at the target agricultural field,

- retrieving from a database comprising a plurality of reference weather data sets associated with a plurality of reference weather conditions at a plurality of reference agricultural fields, respectively, a subset of reference weather data sets based on the at least one weather condition at the target agricultural field as provided by the weather data set, and

- generating for at least one reference weather data set from the subset of reference weather data sets, the at least one monitoring and/or control condition for the spraying application at the target agricultural field, using the subset of reference weather data sets.

The present invention is based on the recognition that triggered by an increasing awareness of protecting flora and fauna, there is an ongoing demand of providing further improved methods and apparatuses for agricultural spraying applications of chemicals. In particular, these methods and apparatuses should enable agricultural spraying applications that comply with the national regulations and are safe for flora and fauna next to agricultural fields. For assessing whether a chemical will be, is or was applied by means of spraying onto an agricultural field in accordance with established regulations, it is beneficial to generate a monitoring and/or control condition, e.g., a spray drift, associated with the spraying application. The monitoring and/or control condition should be as accurate and reliable as possible in order to be suitable for assessing whether spraying a chemical will be, is or was applied by means of spraying onto an agricultural field in accordance with established regulations. In particular, a monitoring and/or control condition for a spraying application should be generated such that when actually applying a chemical onto the target agricultural field based on the generated monitoring and/or control condition, the spraying application should be carried out in accordance with such regulations.

The present invention includes the further recognition that a particular focus should be set on possible spray drifts by wind present above the target agricultural field. Spray drift is generally dependent on the wind speed and wind direction, i.e., the velocity field above the ground. Typically, it is thus of great importance to also consider spray drifts as one of the monitoring and/or control conditions when assessing whether a chemical is or was applied by means of spraying onto an agricultural field in accordance with established regulations.

A monitoring and/or control condition associated with a spraying application can be generated based on weather data, e.g., using models for pesticide risk assessment. One spray drift model that can be employed is based on the AGDISP™ model that was designed to optimize agricultural spraying operations and comprises algorithms for characterizing the release, dispersion, and deposition over and downwind of the application area, see, e.g., Teske, M., Thistle, H., Fritz, B.K, 2019, “Modeling aerially applied sprays: An update to AGDISP model development’, Transactions of the ASABE, 62(2):343-354, ht- tps://doi.org/10.13031 /trans.13129. For example, the AGDISP™ model can be used in estimating downwind deposition of spray drift from aerial and ground boom applications.

Based on the AGDISP™ model, it is possible to determine based on weather data the direction of a spray drift that may result from spraying a pesticide, a herbicide, and/or a fertilizer on an agricultural field. The determined spray drift can be used for assessing risks and establishing safe spraying areas between the spraying application and people, animals, water bodies and non-target crops. Weather data may, e.g., be measured in real-time with a sensor such as a weather station. Such measured weather data can also be called real weather data. Weather data can also be provided, e.g., forecasted, as synthetic weather data using a model such as computational fluid dynamics (CFD) model. Weather data, i.e., real weather data and/or synthetic weather data, can be stored in a database as used for generating the at least one monitoring and/or control condition according to the present invention.

Generally, for generating the at least one monitoring and/or control condition associated with a spraying application using, e.g., a spray drift model such as the Weather Research and Forecast (WRF) system, Skamarock W C, Klemp J B, Dudhia J, Gill D O, Barker D M, Duda M G, Huang X Y, Wang W and Powers J G, 2008, 1-113, URL http://www2.mmm.ucar.edu/wrf/users/docs/arwv3.pdf, a physically consistent hyperlocal database with weather data can be used. For obtaining meaningful and reliable results for the at least one monitoring and/or control condition based on such models, a comparatively complete database of reference weather data sets should be provided that comprises reference weather data sets associated with reference weather conditions at reference agricultural fields over a rather long time span, e.g., a year or more. This is because, e.g., a large physically consistent hyperlocal database with reference weather data sets typically contains sufficient three-dimensional information associated with various different agricultural fields based on which a monitoring and/or control condition for a spraying application can be accurately and reliably determined. In case such a physically consistent hyperlocal database with reference weather data sets is available, it generally is possible to determine various different monitoring and/or control conditions such as a chemical drift accurately and to avoid unnecessarily large no-spraying zones.

Yet, generating such large physically consistent hyperlocal databases in real-time, e.g., each time when needed for generating a monitoring and/or control condition can be challenging due to large amounts of time and computational resources generally required. Therefore, in the method according to the present invention, the database is preferably not generated each time, a monitoring and/or control condition shall be generated. The database of reference weather data sets, preferably, is created initially. Once established, the database can then be accessed for conducing the method for generating at least one monitoring and/or control condition. For creating the database, classical methods such as based on a CFD model and/or high performance computing techniques, e.g., based on quantum computing and/or three dimensional chip designs, can be employed. Since the database of reference data sets is preferably not created each time a monitoring and/or control condition shall be generated, it is possible to reliably generate a monitoring and/or control condition in comparatively short times and with comparatively less computational resources. This allows generating a monitoring and/or control condition also with devices with comparatively small computational resources, such as hand held device like smart phones or the like. A user may, for example, generate a monitoring and/or control condition with a hand held device directly at the target agricultural field within short times. Thereby, a mon- itoring and/or control condition becomes available when needed and may be used for controlling a spraying application even in real-time.

With the method according to the present invention, it is thus possible to generate at least one monitoring and/or control condition for a spraying application at a target agricultural field in a comparatively reliable and accurate manner such that the at least one monitoring and/or control condition may be useful for assessing whether a chemical will be, is or was applied by means of spraying onto a target agricultural field in accordance with certain regulations. This is achieved with the proposed method in that first, weather data associated with at least one weather condition at the target agricultural field are received. These weather data are indicative of at least one weather condition at the target agricultural field. The received weather data are used for retrieving a subset of reference weather data sets from a database using the received weather data. The subset of reference weather data sets is thus associated with the received weather data. More specifically, the reference weather data sets of the provided subsets are associated with the received weather data. The subset of reference weather data sets is a group of reference weather data sets from the database that is provided using the received weather data. For generating the at least one monitoring and/or control condition, the provided subset of reference weather data sets is used. The present method thus does not require that the complete database of reference weather data sets is used but instead only the subset of reference weather data sets is used that is associated with the received weather data. In particular, when generating a subset of reference weather data sets using the received weather data set, it is possible to focus on those reference weather data sets from the database that are of higher relevance. Those reference weather data sets from the database that are of less relevance can be omitted right from the beginning. For example, reference weather data sets from the database can be of higher relevance because they represent reference weather conditions that are to a certain degree similar to the weather condition at the target agricultural field. It is also possible reference weather data sets from the database can be of higher relevance because they are associated with a reference agricultural field that has a certain degree of similarity with the target agricultural field such as a similar flora, fauna and/or topography.

Since the subset of reference weather data sets is retrieved using the received weather data, it is thus possible to only use those reference weather data sets from the database that have a certain degree of relevance for the received weather data and the spraying application at the target agricultural field. It is thus possible to select those reference weather data sets from the database to form the subset that may be of higher relevance for the received weather data and the targeted spraying application. Since those reference weather data sets that may be of increased relevance for the received weather data and the target agricultural field can be provided as the subset, it is possible to generate a monitoring and/or control condition in an accurate and reliable manner that is meaningful for the received weather data and the target agricultural field.

Since not the complete database needs to be created and/or considered but only the generated subset of reference weather data sets, for generating the monitoring and/or control condition, the computational effort can be significantly reduced and a meaningful monitoring and/or control condition can be generated in a comparatively small time while being significant for the received weather data and the spraying application at the target agricultural field. In particular, the computational effort can be reduced because it is possible to populate the database ahead of time and to only extract the most relevant reference weather data sets in the subset. In case, less computational resources are required, also the consumed energy needed for generating the monitoring and/or control condition can be comparatively less for the proposed method.

As a result, with the method according to the present invention, it is possible to generate in an accurate and reliable manner at least one monitoring and/or control condition such as a spray drift that allows a more reliable and safer spraying application of a chemical at a target agricultural field. In particular, when considering spray drift as one of the generated monitoring and/or control conditions, with the method according to the present invention, it is possible to enable a spraying application that prevents an applied chemical from drifting off the target agricultural field and into surrounding landscapes. The method according to the present invention thus bridges the gap to enable a safer application of a chemical considering potential spray drifts in dependence on velocity filed above the ground, i.e., the method according to the present invention contributes to controlling spray drifts in a reliable manner.

With the method according to the present invention, it is also possible to more reliably perform a spraying application in line with current regulations. The method according to the present invention can also be used to assess whether a spraying application conducted in the past has been in line with regulations, e.g., when it comes to regulatory challenges. That is, the present invention may even contribute to avoiding costs for regulatory challenges in the first place. Moreover, the present invention may protect users from wasting products that may drift too far and not achieve the correct dosage rates given less accurate weather conditions.

Preferably, the weather data set is provided at a first time and the at least one monitoring and/or control condition for the at least one of the reference weather data set is provided at a second time that is different from the first time. The second time can be prior to, equal to, or after the first time. It is thus possible to determine the at least one monitoring and/or control condition for the spraying application for a future event. Alternatively, it is also possible to determine the at least one monitoring and/or control condition for the spraying application for a past event. For example, it is possible to reconstruct the at least one monitoring and/or control condition for a spraying application that has been performed in the past, e.g., in case of a regulatory challenge. Often, in a regulatory challenges for cases involving chemicals, it is necessary to study release and drift scenarios based on weather data. The method can thus be used to assess rapidly and accurately the fate of chemicals based on the received weather data set that may be historical weather data and the application parameters such as time of day and amount of release, and the like. On the other hand, in case the at least one monitoring and/or control condition for the spraying application is generated for a real-time or future event, it is possible to use the method, e.g., in a real-time digital application such that an end user can efficiently evaluate when it is safe and most optimal to spray chemicals based on real-time weather conditions. The monitoring and/or control condition can also be used to control a spraying device in real-time based on the monitoring and/or control condition.

Preferably, for retrieving the subset of reference weather data sets from the database, at least one statistical model is employed. Preferably, the statistical model is configured for receiving the weather data set associated with the target agricultural field as input. Preferably, the statistical model is configured for providing the subset of reference weather data sets from the database as output using one or more predictor variables. The statistical model may be obtained using a trained artificial neural network. The artificial neural network may be an autoencoder or a convolutional neural network (CNN) just to mention a few. In particular, such an artificial neural network can be trained to receive a training weather data set as input and to provide a statistical model as output that is configured to provide a subset of reference weather data sets from a database of reference weather data sets using a received weather data set. In particular, the statistical model provided by a trained artificial neural network may be configured to use one or more predictor variables for retrieving the subset of reference weather data sets from the database. Generally speaking, in statistical modelling such predictor variables can be used to build a model that describes the relationship between the predictors and the dependent variable. In particular, in the method, predictor variables such as a wind speed or wind direction can be used that may provide sufficient categorical distinction and a sufficiently large subset. In case a statistical model is used for retrieving the subset of reference weather data sets by means of one or more predictor variables, the subset of reference weather data sets may be provided based on one or more statistical relations that map the received weather data set to a specific subset of reference weather data sets from the database. Based on such one or more statistical relations, the database can be searched in an efficient manner for reference weather data sets associated with the weather condition at the target agricultural field as represented by the received weather data. For example, the reference weather data sets provided as a subset using a statistical model can be a group or ensemble of reference weather data sets, e.g., historical weather data, that most closely match the weather condition at the target agricultural field as represented by the received weather data, for example, based on categorical distinctions such as similar wind direction or speed and the like.

Preferably, the method comprises assigning a score to at least some of the reference weather data sets of the subset of reference weather data sets. Preferably, the score is indicative of a relationship of the reference weather conditions of the at least some weather data sets to the at least one weather condition at the target agricultural field. For example, a relationship can be a degree of similarity. A degree of similarity can be expressed quantitatively, e.g., by a statistical distance between reference weather conditions of the at least some weather data sets to the at least one weather condition at the target agricultural field. A statistical distance can be a Euclidean and/or cosine distance.

The score, preferably, represents the determined degree of similarity found by means of a statistical distance. For example, the higher the determined degree of similarity the higher may be the assigned score and, accordingly, the lower the determined degree of similarity the lower may be the assigned score. By means of the score, it is thus possible to see how well a certain reference weather data set of the subset matches the weather condition at the target agricultural field as represented by the received weather data. Assigning a score to at least some of the reference weather data sets of the subset of reference weather data sets therefore enables weighting and/or categorizing the reference weather data sets in relation to the received weather data. For example, it is possible to filter the subset based on the assigned scores to only consider those reference weather data sets of the subset that have a score above a predefined threshold value. Thereby, it is possible to only use those reference weather data sets of the subset that have a predefined minimum degree of similarity as given by a threshold value to generate the at least one monitoring and/or control condition. Assigning a score to at least some of the reference weather data sets of the subset of reference weather data sets may thus contribute to generating the at least one monitoring and/or control condition having a further improved accuracy and reliability.

Preferably, the method comprises using a predefined metric associated with the reference weather data sets to obtain a group of reference weather data sets from the subset that fulfils the metric. The metric may define a threshold value for a criterion that is used for filtering reference weather data sets of the subset. When using a metric, it is thus possible to further reduce the number of reference weather data sets of the subset to those reference weather data sets that fulfil the metric. A criterion that can be used for filtering the reference weather data sets of the subset using the metric can be the score assigned to at least some of the reference weather data sets of the subset of reference weather data sets as explained before and/or a weather condition at the target agricultural field. It is thus preferred that the metric is associated with the score as explained before and is used to filter those reference weather data sets of the subset of reference weather data sets that fulfil the condition as provided by the metric, e.g., of lying above a predefined threshold score. The metric can be defined such that only those reference weather data sets of the subset are used for generating the at least one monitoring and/or control score that fulfil the metric. Additionally or alternatively, it is possible to define the metric in relation to a weather condition such as a temperature, a wind direction or a wind speed. For example, a metric may be used to filter those reference weather data sets from the subset that represent a target weather condition that does not deviate from the weather condition at the target agricultural field by a certain amount. It is thus possible to define a threshold interval relative to the weather condition at the target agricultural field as the metric and to filter the subset of reference weather data sets forthose reference weather data sets that represent a reference weather condition that is within the threshold interval, i.e., within the metric. Taking the example of a temperature, the metric may be +/- 5 °C in relation to the temperature at the target agricultural field that may be, e.g., 15 °C. Using the metric, only those reference weather data sets from the subset are filtered that have a temperature of 10 °C to 20 °C.

Preferably, for at least two reference weather data sets from the subset of reference weather data sets, at least one monitoring and/or control condition for the spraying application at the target agricultural field are generated, respectively. Generally, it is possible to generate an individual monitoring and/or control condition for several or even for all reference weather data sets of the subset of reference weather data sets, respectively. In case, the reference weather data sets from the subset are further filtered, e.g., based on an assigned score or weather condition or the like, e.g., by using a metric, for several or all of the filtered reference weather data sets, at least one monitoring and/or control condition for the spraying application at the target agricultural field may be generated, respectively. As a result, a set of monitoring and/or control conditions associated with at least two reference weather data sets from the subset of reference weather data sets, respectively, can be generated. From the set of monitoring and/or control conditions, one or more monitoring and/or control conditions can be selected and further processed or presented to a user. For generating the at least one monitoring and/or control condition, one or more reference weather data sets from the subset of reference weather data sets can be provided as input to a model for pesticide risk assessment such as a spray drift model, e.g., based on the AGDISP™ model mentioned before. For one or more of the reference weather data sets from the subset, at least one monitoring and/or control condition can be generated. For example, using an spray drift model that sometimes is also referred to as spray drift model, the one or more reference weather data sets of the subset can be provided as input to the model, and for each of the one or more reference weather data sets, a respective at least one monitoring and/or control condition forthe spraying application at the target agricultural field may be generated.

Preferably, the at least one generated monitoring and/or control condition for the spraying application is used for generating at least one confidence interval. Preferably, the confidence interval represents a likelihood for a spatial distribution of the spraying application at the target agricultural field, for example, in relation to a digital representation of the target agricultural field. In particular, the confidence interval may be indicative of an area over which the spraying application is expected to be distributed with a certain likelihood. Since the monitoring and/or control condition is generated from one or more reference weather data sets from the subset of reference weather data sets and since the subset of reference weather data sets is provided using the received weather data, the confidence interval thus represents a spatial distribution of the spraying application at the target agricultural field associated with the received weather data set. The confidence interval can be an extrapolation from the received weather data set and thus may be indicative of an area over which the spraying application is distributed with a certain likelihood at a second time that lies in the future. It is also possible that the confidence interval represents a likelihood for a distribution of a spraying application at the target agricultural field in real-time, e.g., at the first time, or before the first time associated with the received weather data. The confidence interval can thus be used for real-time applications, for future planning for spraying applications and also for assessing the spatial distribution of spraying applications performed in the past, e.g., in case of regulatory challenges or the like. In the method, it is also possible that at least two confidence intervals are generated. Each of the at least two confidence intervals may represent a different likelihood for a corresponding spatial distribution of the spraying application relative to a digital representation of the target agricultural field.

A likelihood for a certain spatial distribution of the spraying application at the target agricultural field based on the received weather data can be determined based the at least one generated monitoring and/or control condition. For example, if several monitoring and/or control conditions are generated, these several monitoring and/or control conditions can be combined to find areas with an increased likelihood for a presence of a chemical, e.g., of 95 %. Moreover, a combination of several monitoring and/or control conditions may indicate that a spraying application is expected to be distributed within certain areas with a lower likelihood of, e.g., 90 % or 80 % for a distribution of a spraying application in those areas. For example, statistics may be performed on the several monitoring and/or control conditions to obtain the likelihoods for a distribution of a spraying application within different areas that may have overlapping parts. Each area associated with a certain likelihood for a spatial distribution of the spraying application at the target agricultural field can be defined as an individual confidence interval. For example, the several generated monitoring and/or control conditions can be superimposed to find overlapping areas for spatial distribution of the spraying application. The more monitoring and/or control conditions indicate that a spraying application is present within a certain area, the larger may be the likelihood associated with the corresponding generated confidence interval. For example, if most of the monitoring and/or control conditions indicate that a spraying application is present within a centre of the target agricultural field, a confidence interval may be generated that represents a comparatively larger likelihood for a spraying application being present in this centre area. If only some monitoring and/or control conditions indicate that a spraying application is present within a certain area of the target agricultural field, a corresponding confidence interval may be generated that represents a comparatively lower likelihood for a spraying application being present in this area. The confidence intervals may also be output by a model for pesticide risk assessment such as a spray drift model. With the method, it is thus possible to provide uncertainty estimates from the subset of reference weather data sets.

Preferably, at least some of the reference data sets of the database from which the subset of reference weather data sets is retrieved, comprise a three-dimensional distribution of a reference weather condition at a reference agricultural field at a specific time within one or more years. For generating the database of reference weather data sets, a computational fluid dynamics (CFD) model configured for atmospheric applications can be employed to generate high resolution, three-dimensional velocity fields. For example, high resolution, three-dimensional velocity fields can be generated up to a certain distance from the ground, e.g., the lowest 1 km above ground, with a spacing between horizontal points as fine as 100 m or smaller. The CFD model may also be used to generate three-dimensional ambient atmospheric temperature, humidity, pressure, and other parameters that are physically consistent with the velocity fields. Additionally, three-dimensional CFD data can be used to estimate ultra-high-resolution temperature and evapotranspiration distributions across active growing sites. These parameters are relevant components of crop yield and growth stage modelling. The reference data sets of the database may be associated with a plurality of reference weather conditions at a plurality of reference agricultural fields. In particular, the database may be established to include a sufficiently large amount of reference data sets so that for the received weather data set, it is possible to find a subset of reference data sets that matches closely at least one weather condition at the target agricultural field represented by the received weather data set. In particular, it should be possible to retrieve a subset of reference weather data from the database that is representative of weather conditions at the target agricultural field and that are physically consistent with the velocity fields. A CFD model can thus be used to create a massive archive of cases at different reference agricultural fields based on days of the year, atmospheric conditions, and other parameters. Once created, the archive may be stored in the database. It is possible that the database is updated as users register further agricultural fields. A reference weather data set of the database may include for a reference target field the reference weather conditions at a certain time in the past. It is a particular advantage of the database that the reference weather data sets are associated with a reference agricultural field. This allows to find reference weather data sets that represent reference weather conditions at a reference agricultural field that are to a predefined degree similar to the weather conditions at the target agricultural field. The reference weather data sets at a scale of metres are thus different to weather data from weather stations that are generally at a scale of kilometres. A further difference between the reference weather data sets, e.g., as generated using a CFD model and/or high performance computing and common weather data from weather stations is that the reference weather data sets may represent a plurality of different weather conditions consistent with a velocity field at a reference agricultural field in a comparatively densely sampled manner. A reference weather data set may generally represent historical weather data. Reference weather conditions may include wind regimes, temperature profiles, humidity, moisture, and the like. For example, the database may include reference weather data sets representing a time span of one or more years. It is possible that the reference weather data sets represent the weather conditions at different reference agricultural fields in intervals of one or more hours for the one or more years. Further information may be included in the reference weather data sets such as the time in a year, the location, what is growing on the fields, information about neighbouring environment, and the like.

For establishing the database, high performance computing can be employed, e.g., including parallelization and/or cloud computing. The reference weather data sets can be created using one or more CFD models. Additionally or alternatively, hybrid quantum-classical computing can be used for obtaining the reference weather data sets of the database. The input to such a quantum-classical hybrid computation model may be the map coordinates of the target areas and current and/or historical weather measures of the temperature, humidity, pressure and/or wind velocity of the reference agricultural fields. The output of such a quantum-classical hybrid computation model may be a forecast of the wind velocity for the reference agricultural fields, i.e., a reference weather data set. A quantum-classical hybrid computation model can be trained using quantum machine learning. A computationally efficient way of obtaining reference weather data sets, for example, using high performance computing such based on a quantum-classical hybrid computation may have the advantage that the time taken to create or update the database of reference weather data sets can be further reduced leading potentially also to reduced costs of running a computer system.

The database of reference weather data sets as used herein may provide a more complete three-dimensional, accurate, and representative collection of weathers data than it may be available from public or even private weather data sources for spray and drift applications. The database thereby may enable both real-time and post-event analyses making it safer and reliable for decision support.

Preferably, the monitoring and/or control condition associated with the spraying application is a spray drift of a chemical to be applied at the target agricultural field. A spray drift refers to the movement of pesticide dust or droplets through the air at the time of application or soon after, e.g., to any site other than the area intended. Pesticide droplets are produced by spray nozzles, e.g., of a spraying device. In the method, the monitoring and/or control condition being a spray drift can be generated based on the AGDISP™ model mentioned before. The AGDISP™ model allows for the mathematical modelling of the fate and transport of agricultural sprays applied by aerial and ground boom spray systems. Other spray drift that can be employed for generating the monitoring and/or control condition may be the AgDRIFT® model, the PERFUM (Probabilistic Exposure and Risk Model for FUMigants model), the SOFEA (SOil Fumigant Exposure Assessment) model, and FEMS (Fumigant Exposure Modeling System) model just to name a few.

Preferably, the method comprises generating a control signal that is suitable for controlling a spraying device based on the provided at least one monitoring and/or control condition. For example, real-time controlling of a spraying device based on the monitoring and/or control condition is possible. A spraying device may be any machine that is capable of performing an agricultural spraying application such as a tractor, a robot, a flying drone or the like. A spraying device may be configured for performing the agricultural spraying fully autonomously. A spraying device that is configured for performing the spraying application fully autonomously, preferably, is configured for sensing its environment and moving and performing a spraying application with little or no human input. A spraying device may be configured for performing the spraying application primarily upon human input, i.e., the spraying device is primarily controlled by a human being. In this case, the driver or user of the spraying device may be instructed via the control signal of when and/or how to perform a spraying application at the target agricultural field.

The present invention also relates to a computer program for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field, the computer program including instructions for executing the steps of the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field as described herein, when run on a computer. Furthermore, the present invention relates to a non-trans- itory computer readable data medium storing the computer program for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field.

According to the present invention, also an apparatus for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field is proposed. The apparatus comprises a receiving unit, a subset providing unit and a monitoring and/or control condition providing unit. The receiving unit is configured for receiving a weather data set associated with at least one weather condition at the target agricultural field. The subset retrieving unit is configured for retrieving from a database comprising a plurality of reference weather data sets associated with a plurality of reference weather conditions at a plurality of reference agricultural fields, respectively, a subset of reference weather data sets based on the at least one weather condition at the target agricultural field as provided by the weather data set. Moreover, the monitoring and/or control condition providing unit is configured for providing for at least one of the reference weather data sets from the subset of reference weather data sets, the at least one monitoring and/or control condition forthe spraying application at the target agricultural field, using the subset of reference weather data sets.

The apparatus for generating at least one monitoring and/or control condition for a spraying application can be used for conducting the method or generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field as described herein.

Furthermore, the present invention relates to a use of at least one monitoring and/or control condition provided by the apparatus for generating at least one monitoring and/or control condition as described herein for controlling a spraying application to be performed on a target agricultural field. For example, the spraying application may be performed by spraying device that can be controlled based on the at least one monitoring and/or control condition.

The present invention relates to a spraying device that is configured for applying a chemical to a target agricultural field by means of spraying. The spraying device comprises a control interface that is configured to receive a control signal suitable for controlling the spraying device. For example, the control interface can be configured for receiving a control signal wirelessly. To this end, the control interface can be connected to a server or the like for receiving the control signal. Preferably, the control signal is indicative of the monitoring and/or control condition. Preferably, upon receiving the control signal via the control interface, the spraying device can be controlled based on the monitoring and/or control condition as generated with a method for generating at least one monitoring and/or control condition described herein.

Accordingly, the present invention relates to a spraying device configured for performing a spraying application and having a control interface connected to a server or the like for receiving a monitoring and/or control condition from the server or the like as generated with the method for generating at least one monitoring and/or control condition described herein.

According to the present invention, also a method for updating a database as used in the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field described herein is proposed. In particular, the proposed method may be a computer-implemented method for updating a database as used in the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field. The method comprises the steps of:

- receiving a request for providing one or more reference weather data sets associated with a plurality of reference weather conditions at a plurality of reference agricultural fields for a weather data set associated with at least one weather condition at a target agricultural field, and

- generating and storing the one or more reference weather data sets for the weather data set in the database. Preferably, the updated database is used for generating a monitoring and/or control condition as described herein. That is, from the updated database a subset of reference weather data sets may be retrieved and used for generating a monitoring and/or control condition as described herein.

The request may be provided by a user or may be generated automatically, e.g., in response to a received weather data set. For example, in case a user receives the at least one monitoring and/or control condition which may be associated with a comparatively high uncertainty, e.g., expressed by a corresponding confidence interval, the user may send a request for updating the database. The aim is to run the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field again but with an improved subset of reference weather data sets to generate a monitoring and/or control condition which is associated with a higher degree of certainty. Yet, the improved subset of reference weather data sets may need to be generated first of all. So updating the database is necessary in this case.

It is possible that the method comprises updating the database with a reference weather data set associated with a reference weather condition measured with a sensor at a reference agricultural field. It is thus possible that the database is updated with reference weather data sets that are measured with a sensor, e.g., with a weather station. The database may thus include synthetic reference weather data sets generated with a model, e.g., a CFD model, and may also include measured reference weather data sets measured with a sensor at a reference agricultural field.

In the method for updating a database, high performance computing can be employed such as based on a hybrid quantum-classical computing model, for example, that is trained by quantum machine learning. The hybrid quantum-classical computing model preferably is configured, e.g., trained, for generating reference weather data sets for establishing and/or updating the database of reference weather data sets. For example, a hybrid quantum- classical computing model can be trained for receiving coordinates of the target agricultural field as well as weather data as input and for providing a reference weather data set as output.

The method for updating a database may comprise providing at least one of the reference weather data sets from the updated database for generating at least one monitoring and/or control condition for a spraying application at a target agricultural field. For generating at least one monitoring and/or control condition for a spraying application at a target agricul- tural field, the method for generating at least one monitoring and/or control condition for a spraying application at a target agricultural field described herein can be employed.

The present invention also relates to a computer program for updating a database as used in the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field described herein, the computer program including instructions for executing the steps of the method for updating a database as described herein, when run on a computer. Furthermore, the present invention relates to a non-transitory computer readable data medium storing the computer program for updating a database as used in the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field described herein.

Furthermore, the present invention also relates to an apparatus for updating a database as used in the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field described herein. The apparatus for updating a database comprises a request receiving unit and a reference weather data sets generating unit. The request receiving unit is configured for receiving a request for providing one or more reference weather data sets associated with a plurality of reference weather conditions at a plurality of reference agricultural fields for a weather data set associated with at least one weather condition at a target agricultural field. The reference weather data sets generating unit is configured for generating and storing the one or more reference weather data sets for the weather data set in the database. The apparatus for updating a database may be used for carrying out the method for updating a database as used in the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field described herein.

Further explanations of terms:

As used herein, a monitoring and/or control condition, preferably, is a spray drift of a chemical to be applied onto the target agricultural field. However, a monitoring and/or control condition may also be an amount of release, a droplet size, a spraying pattern, a deposition rate, a chemical to be used or the like. The monitoring and/or control condition preferably, may be suitable so that a spraying device may be controlled based on the monitoring and/or control condition. Preferably, a spraying device may be controlled in real-time based on the monitoring and/or control condition. As used herein, a spraying application relates to the application of a chemical at an agricultural field by means of spraying, e.g., with a spraying device.

As used herein, an agricultural field is an area of land, enclosed or otherwise, used for agricultural purposes such as cultivating crops. Crops may include, for example, wheat, corn, soybeans, hay, or any other crop type. Typically, the crop is seeded on a field at a first point in time in order to grow and to be harvested at a later point in time. Crops may also include fruits of plants, such as fruit trees, which are seeded at a first point in time and grow over time such that they may provide fruits over time, such as over several years.

As used herein, a weather data set may include a measured wind speed, wind direction, air temperature and/or a soil temperature, e.g., a temperature profile of the last two years or an expected temperature profile, e.g., for the upcoming ten days as a weather forecast. The weather data set may alternatively, or additionally, include a measured amount of precipitation. The weather data set may also include a weather forecast, e.g., a ten-day weather forecast.

As used herein, a weather condition may be indicative of at least one of a wind speed, an ambient temperature, a humidity or a pressure. Preferably, the at least one weather condition is measured with a sensor, e.g., with a weather station at or close to the target agricultural field.

As used herein, a reference weather data set is stored in the database. A reference weather data set of the database can be measured reference weather data set that has been measured with a sensor. A reference weather data set can also be synthetic reference weather data set that has been generated with a model such as a CFD model. The database may comprise measured reference weather data sets and/or reference weather data sets generated with a model, i.e., synthetic reference weather data sets. The database may be stored on one or more external servers that can be accessed for carrying out the methods described herein. For example, external servers may include a weather satellite network or weather station network. The external servers may receive reference weather data sets from sensors sensing weather information including, for example, wind speed, wind direction, precipitation, temperature, and humidity. The reference weather data sets may, for example, include soil temperature, air temperature, humidity, and precipitation. The external servers may gather the reference weather data sets for storing them in the database. The reference weather data sets may comprise historical climate data, e.g., including weather information of the last one or more years, such as the last two, the last three, the last four, or the last five years. As used herein, a reference weather condition may be indicative of at least one of a reference wind speed, a reference ambient temperature, a reference humidity or a reference pressure.

As used herein, a subset of reference weather data sets refers to a group or ensemble of reference weather data sets from the database. Preferably, the reference weather data sets of the subset are analogues to the weather condition at the target agricultural field as represented by the weather data obtained from measured or real-time field observations. The subset preferably provides an ensemble of historical weather data that most closely match current weather conditions at the target agricultural field based on categorical distinctions, e.g., similar wind direction, and may be used to quantify highly accurate scenarios for spray and drift application.

It shall be understood that the aspects described above, and specifically the method of claim 1 , the method of claim 12, the apparatus of claim 14 and the use of claim 15, have similar and/or identical preferred embodiments, in particular as defined in the dependent claims.

It shall be understood that a preferred embodiment of the invention can also be any combination of the dependent claims or above embodiments with the respective independent claim.

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 shows a flowchart diagram representing a method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field;

Fig. 2 schematically and exemplary shows an apparatus for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field;

Fig. 3 shows a flowchart diagram representing a method for updating a database as used generating at least one monitoring and/or control condition as- sociated with a spraying application to be performed on a target agricultural field;

Fig. 4 shows a flowchart diagram representing a method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field in which several confidence intervals are generated;

Fig. 5 shows a flowchart diagram representing a method for generating a spray drift for a spraying application to be performed on a target agricultural field;

Fig. 6 shows a flowchart diagram representing a workflow of a weather prediction model that includes a hybrid quantum-classical prediction model;

Fig. 7 shows a flowchart diagram representing a method for training a model which given an input at time step t gives a solution at t + 1 ; and

Fig. 8 shows a flowchart diagram representing a method for using a trained model which given an input at time step t gives a solution at t + 1 .

DETAILED DESCRIPTION OF EMBODIMENTS

Figure 1 shows a flowchart diagram representing a method for generating at least one monitoring and/or control condition such as a spray drift associated with a spraying application to be performed on a target agricultural field. In the method, initially, a weather condition is measured at a target agricultural field using a weather station. A weather data set representing the weather condition at the target agricultural field are then fed into a statistical model (step S1). Furthermore, a database of reference weather data sets is provided. The reference weather data sets each represent a reference weather condition at a reference agricultural field, respectively.

The statistical model comprises statistical relations that map the received weather data set to a subset of reference weather data sets of the database (step S2). In particular, the statistical model is configured to extract from the database a subset of reference by the data sets that according to the statistical relations most closely match the weather condition at their target agricultural field represented by the weather data set. Furthermore, the statistical model is configured to assign a score to each of the reference by the data sets of the subset. The score is indicative of a degree of similarity of a respective one of the reference weather data sets of the subset in comparison to the received weather data set. For example, the score can be related to a certain weather condition of the received weather data set. The score may be related to a wind speed, wind direction or air temperature. As an example, the score can be determined based on a statistical distance between, e.g., a weather condition such as the wind speed of the received weather data set to a reference weather condition of the reference weather data set. For example, the higher the degree of similarity, the higher can be the assigned score.

The obtained subset of reference weather data sets is then retrieved from the database (step S3) and further filtered using a metric (step S4). The metric is defined with respect to the scores assigned to the reference weather data sets of the subset and provides a threshold score that is used for further filtering the reference weather data sets of the subset. In particular, when applying the metric to the reference weather data sets of the subset, only those reference weather data sets will be considered further that have a score assigned to it that lies above the threshold score defined by the metric. The further filtered reference weather data sets which now is a reduced group of reference weather data sets of the subset that fulfil the metric, is then provided to a spray drift model (step S5).

For example, the spray drift model can be based on the AGDISP™ model. The spray drift model generates for each reference weather data set that fulfils the metric at least one monitoring and/or control condition (step S6). In particular, each of the reference weather data sets that fulfils the metric is fed separately one after the other into the spray drift model and for each reference weather data set, the spray drift model generates at least one individual monitoring and/or control condition such as a spray drift.

The set of individual monitoring and/or control conditions associated with the reference weather data sets that fulfil the metric can be provided to a user end device to inform a user about the result. Additionally or alternatively, the obtained set of individual monitoring and/or control conditions can be used for generating a control signal that can be provided to a spraying device that is supposed to perform the spraying application at the target agricultural field. For example, the monitoring and/or control condition being a spray drift may be indicative of a spray drift of a specific length oriented along a certain direction. For example, the monitoring and/or control condition may be indicative of a spray drift of 3 m in the direction of southwest. The monitoring and/or control condition may then used to control a spraying device to perform the spraying application such that drift of chemicals outside the target agricultural field is prevented. Employing the method, it may be possible to assess more rapidly and accurately the fate of chemicals based on reference weather data sets that closely match the weather conditions at the target agricultural field as represented by the received weather data set. By linking the method to a real-time digital application, end users can efficiently evaluate whether it is safe and most optimal to spray chemicals based on real-time weather conditions.

Figure 2 schematically and exemplary shows an apparatus 200 that is configured for generating at least one monitoring and/or control condition for spraying application to be performed on the target agricultural field. The apparatus 200 can be used for conducting the steps of receiving a weather data set, retrieving a subset of reference weather data sets from the database and of generating at least one monitoring and/or control condition as performed in a method for generating at least one monitoring and/or control condition to be performed on a target agricultural field as described with reference to figure 1 .

To this end the apparatus 200 comprises a receiving unit 202 that is configured for receiving a weather data set 201 , for example, from a weather station. The weather data set 201 represents a weather condition at the target agricultural field.

Furthermore, the apparatus 200 comprises a subset retrieving unit 204. The subset retrieving unit 204 is configured for retrieving a subset of reference weather data sets 203 that has been extracted from the database of reference weather data sets, for example, using the weather data set 201 and a statistical model that is configured for mapping the received weather condition to reference weather data sets of the database using statistical relations.

The subset retrieving unit 204 can be configured to itself generate the subset of reference weather data sets 203 from the database. For example, the subset retrieving unit 204 may comprise a statistical model and access the database for providing the subset of reference weather data sets. It is also possible, that the subset of reference weather data sets 203 is generated outside the apparatus 200, for example, on an external server. In this case, the subset retrieving unit 204 may be configured for receiving the subset of reference weather data sets 203 from an external device and for retrieving the subset of reference weather data sets 203 for processing by the apparatus 200. Retrieving the subset of reference weather data sets 203 by the subset retrieving unit 204 thus includes at least one of the generation of the subset 203 by the apparatus 200 and the receiving of the subset 203 from an external device by the apparatus 200. Moreover, the apparatus 200 comprises a monitoring and/or control condition generating unit 206 that is configured for generating for one or more reference weather data sets 203 of the subset at least one monitoring and/or control condition 205, respectively. The at least one monitoring and/or control condition 205 may, preferably, be a spray drift. Accordingly, the at least one monitoring and/or control condition may be indicative of a spray drift at the target agricultural field that is associated with the weather condition of the received weather data set.

Figure 3 shows a flowchart diagram representing a method for updating a database as used in the method for generating at least one monitoring and/or control condition described with reference to figure 1 or by the apparatus 200 described with reference to figure 2.

In the method for updating a database, a request is received, e.g., from an external device, to provide one or more reference weather data sets (step T1). For example, the request may be to provide one or more reference weather data sets that at least closely match a weather condition at a target agricultural field represented by a weather data set. It is possible that the request may be received in form of or together with a weather data set. For example, the receiving of a weather data set itself may already be processed as a request for providing one or more reference weather data sets that, preferably, at least closely match the weather condition represented by the provided weather data set. Additionally or alternatively, a request may be to provide for a weather data set another subset of reference weather data sets that has a better match to the weather data set.

Upon receiving the request for providing one or more reference weather data sets, in the method, one or more reference weather data sets are generated and stored in the database (step T2). Thereby, the database is updated. The generating of the one or more reference weather data sets may be carried out using a CFD model and/or high performance computing such as based on a hybrid quantum-classical prediction model or the like.

One or more of the newly generated reference weather data sets may be provided in or as a subset of reference weather data sets to the sender device of the request in order to fulfil the request. Using the retrieved subset of reference weather data sets, at least one monitoring and/or control condition for a spraying application at a target agricultural field may then be generated (step T3), for example, using the method as described with reference to figure 1 or the apparatus 200 described with reference to figure 2. Figure 4 shows a flowchart diagram representing a method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field in which also several confidence intervals are generated. In the method, a database 400 comprising a plurality of reference weather data sets is provided that is also referred to as training data set. This is because the database 400 is also used for training an artificial neural network to provide a statistical model 404 that is configured for retrieving a subset 405 of reference weather data sets from the database 400. The reference weather data sets have been generated using a dynamical model such as a CFD model 402. Additionally or alternatively, a hybrid quantum-classical prediction model may be employed for generating reference weather data sets for the database 400. The reference weather data sets each represent reference weather conditions at a reference agricultural field at a certain time within a year. As exemplary depicted in figure 4, the reference weather data sets represent reference weather conditions over more than one year, e.g., here 651 days although it may be more or less days. For each of these days, the reference weather data sets represent reference weather conditions for multiple hours per day. Accordingly, the database provides a rather complete picture about the weather conditions during more than one year. The reference weather conditions may include wind regimes, temperature profiles and further weather parameters. Furthermore, the reference weather conditions may include information about adjacent environment, what is growing on the respective reference agricultural field, the spraying activities such as the type of chemicals used and a droplet size.

As mentioned before, by means of the database 400, an artificial neural network can be trained to provide the statistical model 404. The statistical model 404 can then be used in the method for retrieving subsets 405 of reference weather data sets from the database 400. To this end, the statistical model 404 uses one or more predictor variables, such as a wind speed or wind direction, that provide good categorical distinction and sufficiently large subsets 405 of reference weather data sets. In figure 4, a subset 405 of reference weather data sets is also referred to as analogue ensemble subset. This is because the subset 405 of reference weather data sets is preferably retrieved by the statistical model 404 in response to receiving a weather data set and at least closely match the provided weather condition at the target agricultural field.

Once the database 400 and the statistical model 404 are established, the method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field in which also several confidence intervals are generated may be carried out. Initially, a weather data set is provided that includes information about a weather condition such as a wind speed at a target agricultural field (step M1). The weather data set can represent a current weather condition directly at the target agricultural field or can be provided from a near-by weather station or from several weather stations.

For the weather data set, a metric is provided (step M2) that defines a threshold score useable for selecting or filtering reference weather data sets to be used for generating the at least one monitoring and/or control condition. In particular, a metric may be generated to select similar cases from a provided subset 405 with regard to the weather condition represented by the received weather data set.

The weather data set is then used as input for the statistical model to retrieve the subset 405 of reference weather data sets that preferably closely match the weather condition of the weather data set. The retrieved subset 405 of reference weather data sets is then further filtered using the metric (step M3). To this end, the statistical model 404 is configured to assign a score to at least some of the reference weather data sets of the subset 405. The score can be provided based on a statistical distance between the reference weather conditions of the reference weather data sets to the weather condition at the target agricultural field. For example, the statistical distance can be determined for a wind speed at the target agricultural field and reference wind speeds at the reference agricultural fields. By filtering the subset 405 using the metric, a further reduced group of reference weather data sets is provided that even more closely match the weather condition at the target agricultural field.

The group of reference weather data sets is then input into a spray drift model (step M4). For example, the AGDISP™ can be employed. With the spray drift model, for each reference weather data set of the group of reference weather data sets, at least one monitoring and/or control condition is generated, respectively. Here, the monitoring and/or control condition is a spray drift. Based on the set of monitoring and/or control conditions associated with those reference weather data sets that fulfil the metric, three confidence intervals 406, 408, 410 are generated by the spray drift model (step M5). The three confidence intervals 406, 408, 410 are to be understood as exemplary and more or less than three confidence intervals can be generated such as one, two, four, five six, or more confidence intervals. Each of the three confidence intervals 406, 408, 410 represents a different likelihood for a spatial distribution of the spraying application at the target agricultural field. That is, the confidence interval 406 represents a likelihood of 99 % for a spray drift over a first area. The confidence interval 408 represents a likelihood of 90 % that the spraying application will drift within a second area. Moreover, the confidence interval 410 represents a likelihood of 80 % that the spraying application will spread over a third area due to spray drift. Figure 5 shows a flowchart diagram representing a method for generating a spray drift for a spraying application to be performed at a target agricultural field. In the method, a decision support system that can approve a spray application or disapprove a spraying application is employed. The decision support system can be accessed via a computer program such as an application software in real-time or for regulatory challenges. The decision support system receives a weather data set indicative of at least one weather condition at the target agricultural field, e.g., from a weather station or from a larger-scale forecast (step N1). For generating at least one monitoring and/or control condition for a spraying application, the decision support system provides the weather data set as input for a statistical model. The statistical model comprises statistical relations that map the weather data set to a subset of reference weather data sets that at least closely match to the weather condition at the target agricultural field. To this end, the statistical model accesses a database of reference weather data sets (step N2). The reference weather data sets have been generated using a CFD model with optional quantum computing scale up. The reference weather data sets span multiple years and multiple hours and contain weather parameters such as wind speed, wind direction, temperature and the like and comprise fine-scale three- dimensional representations of reference weather conditions at reference agricultural fields.

Using the weather data set as input, the statistical model filters a subset of reference weather data sets from the database using statistical relations (step N3). The statistical model with the statistical relations can be obtained with an artificial neural network that is trained, e.g., using the database of reference weather data sets, for providing a statistical model that is configured to map a weather data set to a subset of reference weather data sets of the database. For example, 10 to 20 representative reference weather data sets can be provided as the subset that closely matches the weather condition at the target agricultural field.

The obtained subset of reference weather data sets is then provided as input to a spray drift model to generate at least one monitoring and/or control condition for a spraying application at the target agricultural field (step N4). Furthermore, the subset of reference weather data sets can be used as analogues to generate a probability, e.g., in form of a confidence interval, that a product sprayed at given time and location will drift off the target agricultural field.

Figure 6 shows a flowchart diagram representing a workflow of a weather prediction model that includes a hybrid quantum-classical prediction model. The hybrid quantum-classical prediction model can be used for creating or updating a database of reference weather data sets. In particular, the hybrid quantum-classical prediction model is configured to solve a partial differential equation describing the weather condition. It could be a quantum machine learning model which first would need to be trained before it can make a forecast. In the workflow, map coordinates of the target agricultural field and a weather data set comprising local weather measurements are provided as input to the hybrid quantum-classical prediction model. The hybrid quantum-classical prediction model is trained, e.g., by quantum machine learning, to provide a forecast of a monitoring and/or control condition such as of a wind speed at the target agricultural field as well as an uncertainty estimate of the forecast, e.g., in form of a confidence interval. The use of a hybrid quantum-classical prediction model may provide an increase in the accuracy and precision of the wind forecast that in turn may lead to better control of a spraying application. Thereby, it is possible to provide an improved application of a chemical, e.g., with reduced spray drifts, and better regulatory compliance.

Additionally or alternatively, it is also possible to employ to a hybrid quantum-classical prediction model to solve the Navier Stokes equations. For example, the Navier Stokes equation can be solved and used to predict a wind speed for a target agricultural field. The Navier Stokes equation could be adapted specifically for use in a spraying application. Thereby, since both space and time are discretized, the Navier Stokes equation outputs the wind speed on a set of grid points for a time t.

Figure 7 shows a flowchart diagram representing a method fortraining a model which given an input at time step t gives a solution at t + 1 . Initially, a parameterized quantum circuit is provided as the model-to-be-trained (step V1). As training data, the quantum circuit receives a vorticity and streamline at each grid point at a time t as well as boundary conditions at the edge of the grid (step V2). The vorticity and streamline may be found directly from physical measurements or from a previous forecast.

The quantum circuit then provides a proposed solution at time t+1 (step V3). Using classical computation, an error of the outputs from the quantum circuit is minimized using a cost function (step V4). For minimizing the outputs of the quantum circuit, historical weather data as well as a differential equation describing physical weather relations can be employed (step V5). Based on the result provided by the cost function, an optimisation of parameters can be provided and used for adapting the weights of the quantum circuit (step V6). Thereby, an error of an output of the quantum circuit can be decreased in a step-wise manner to obtain a trained parameterised quantum circuit. Figure 8 shows a flowchart diagram representing a method for using a trained model which given an input at time step t gives a solution at t + 1 . The trained model is a parameterised quantum circuit. The training of the quantum circuit can be carried out as described with reference to figure 7. The quantum circuit is trained to receive a vorticity and streamline at each grid point at time t as well as boundary conditions at the edge of the grid as input. Based on the input, the trained model provides a vorticity and streamline at each grid point at time t+1 . The trained model can be used for generating reference weather data sets for a database. Based on the reference weather data sets, a monitoring and/or control condition for a spraying application can be generated.

Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.

A single unit or device may fulfil the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

Procedures like receiving a weather data, retrieving a subset of reference weather data sets and generating for at least one reference weather data set from the subset of reference weather data sets, the at least one monitoring and/or control condition, etc. performed by one or several units or devices can be performed by any other number of units or devices. These procedures can be implemented as program code means of a computer program and/or as dedicated hardware.

A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium, supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

Any units described herein may be processing units that are part of a classical computing system. Processing units may include a general-purpose processor and may also include a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or any other specialized circuit. Any memory may be a physical system memory, which may be volatile, non-volatile, or some combination of the two. The term “memory” may include any computer-readable storage media such as a non-volatile mass storage. If the computing system is distributed, the processing and/or memory capability may be distributed as well. The computing system may include multiple structures as “executable components”. The term “executable component” is a structure well understood in the field of computing as being a structure that can be software, hardware, or a combination thereof. For instance, when implemented in software, one of ordinary skill in the art would understand that the structure of an executable component may include software objects, routines, methods, and so forth, that may be executed on the computing system. This may include both an executable component in the heap of a computing system, or on computer- readable storage media. The structure of the executable component may exist on a computer-readable medium such that, when interpreted by one or more processors of a computing system, e.g., by a processor thread, the computing system is caused to perform a function. Such structure may be computer readable directly by the processors, for instance, as is the case if the executable component were binary, or it may be structured to be interpretable and/or compiled, for instance, whether in a single stage or in multiple stages, so as to generate such binary that is directly interpretable by the processors. In other instances, structures may be hard coded or hard wired logic gates, that are implemented exclusively or near-exclusively in hardware, such as within a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or any other specialized circuit. Accordingly, the term “executable component” is a term for a structure that is well understood by those of ordinary skill in the art of computing, whether implemented in software, hardware, or a combination. Any embodiments herein are described with reference to acts that are performed by one or more processing units of the computing system. If such acts are implemented in software, one or more processors direct the operation of the computing system in response to having executed computer-executable instructions that constitute an executable component. Computing system may also contain communication channels that allow the computing system to communicate with other computing systems over, for example, network. A “network” is defined as one or more data links that enable the transport of electronic data between computing systems and/or modules and/or other electronic devices. When information is transferred or provided over a network or another communications connection, for example, either hardwired, wireless, or a combination of hardwired or wireless, to a computing system, the computing system properly views the connection as a transmission medium. Transmission media can include a network and/or data links which can be used to carry desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general-purpose or special-purpose computing system or combinations. While not all computing systems require a user interface, in some embodiments, the computing system includes a user inter- face system for use in interfacing with a user. User interfaces act as input or output mechanism to users for instance via displays.

Those skilled in the art will appreciate that at least parts of the invention may be practiced in network computing environments with many types of computing system configurations, including, personal computers, desktop computers, laptop computers, message processors, hand-held devices, multi-processor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, main-frame computers, mobile telephones, PDAs, pagers, routers, switches, data centres, wearables, such as glasses, and the like. The invention may also be practiced in distributed system environments where local and remote computing system, which are linked, for example, either by hardwired data links, wireless data links, or by a combination of hardwired and wireless data links, through a network, both perform tasks. In a distributed system environment, program modules may be located in both local and remote memory storage devices.

Those skilled in the art will also appreciate that at least parts of the invention may be practiced in a cloud computing environment. Cloud computing environments may be distributed, although this is not required. When distributed, cloud computing environments may be distributed internally within an organization and/or have components possessed across multiple organizations. In this description and the following claims, “cloud computing” is defined as a model for enabling on-demand network access to a shared pool of configurable computing resources, e.g., networks, servers, storage, applications, and services. The definition of “cloud computing” is not limited to any of the other numerous advantages that can be obtained from such a model when deployed. The computing systems of the figures include various components or functional blocks that may implement the various embodiments disclosed herein as explained. The various components or functional blocks may be implemented on a local computing system or may be implemented on a distributed computing system that includes elements resident in the cloud or that implement aspects of cloud computing. The various components or functional blocks may be implemented as software, hardware, or a combination of software and hardware. The computing systems shown in the figures may include more or less than the components illustrated in the figures and some of the components may be combined as circumstances warrant.

Any reference signs in the claims should not be construed as limiting the scope.

Claims

CLAIMS:

1 . A method for generating at least one monitoring and/or control condition associated with a spraying application to be performed on a target agricultural field, the method comprising the steps of: receiving a weather data set associated with at least one weather condition at the target agricultural field, retrieving from a database comprising a plurality of reference weather data sets associated with a plurality of reference weather conditions at a plurality of reference agricultural fields, respectively, a subset of reference weather data sets based on the at least one weather condition at the target agricultural field as provided by the weather data set, and generating for at least one reference weather data set from the subset of reference weather data sets, the at least one monitoring and/or control condition for the spraying application at the target agricultural field, using the subset of reference weather data sets.

2. The method of claim 1 , wherein the weather data set is provided at a first time and the at least one monitoring and/or control condition for the at least one of the reference weather data set is generated at a second time that is different from the first time.

3. The method of at least one of the preceding claims, wherein for retrieving the subset of reference weather data sets from the database, at least one statistical model is employed that is configured for receiving the weather data set associated with the target agricultural field as input and for retrieving the subset of reference weather data sets from the database as output using one or more predictor variables.

4. The method of at least one of the preceding claims, comprising assigning a score to at least some of the reference weather data sets of the subset of reference weather data sets, the score being indicative of a relationship of the reference weather conditions of the at least some weather data sets to the at least one weather condition at the target agricultural field.

5. The method of at least one of the preceding claims, comprising using a predefined metric associated with the reference weather data sets to obtain a group of reference weather data sets from the subset that fulfils the metric.

6. The method of at least one of the preceding claims, wherein for at least two reference weather data sets from the subset of reference weather data sets, at least one monitoring and/or control condition for the spraying application at the target agricultural field is generated, respectively.

7. The method of at least one of the preceding claims, wherein the at least one monitoring and/or control condition for the spraying application is used for generating at least one confidence interval, the confidence interval representing a likelihood for a spatial distribution of the spraying application at the target agricultural field.

8. The method of claim 7, wherein the at least two confidence intervals are generated, each of the at least two confidence intervals representing a different likelihood for a corresponding spatial distribution of the spraying application relative to a digital representation of the target agricultural field.

9. The method of at least one of the preceding claims, wherein at least some of the reference data sets of the database from which the subset of reference weather data sets is retrieved, comprise a three-dimensional distribution of a reference weather condition at a reference agricultural field at a specific time within one or more years.

10. The method of at least one of the preceding claims, wherein the monitoring and/or control condition associated with the spraying application is a spray drift of a chemical to be applied at the target agricultural field.

11. The method of at least one of the preceding claims, comprising generating a control signal that is suitable for controlling a spraying device based on the generated at least one monitoring and/or control condition.

12. A method for updating a database as used in at least one of the preceding claims, the method comprising the steps of: receiving a request for providing one or more reference weather data sets associated with a plurality of reference weather conditions at a plurality of ref- erence agricultural fields for a weather data set associated with at least one weather condition at a target agricultural field, and generating and storing the one or more reference weather data sets for the weather data set in the database.

13. The method of claim 12, comprising providing at least one of the reference weather data sets from the updated database for generating at least one monitoring and/or control condition for a spraying application at a target agricultural field.

14. An apparatus for generating at least one monitoring and/or control condition for a spraying application to be performed on a target agricultural field, the apparatus comprising: a receiving unit that is configured for receiving a weather data set associated with at least one weather condition at the target agricultural field, a subset retrieving unit that is configured for retrieving from a database comprising a plurality of reference weather data sets associated with a plurality of reference weather conditions at a plurality of reference agricultural fields, respectively, a subset of reference weather data sets based on the at least one weather condition at the target agricultural field as provided by the weather data set, and a monitoring and/or control condition generating unit that is configured for generating for at least one of the reference weather data set from the subset of reference weather data sets, the at least one monitoring and/or control condition for the spraying application at the target agricultural field, using the subset of reference weather data sets.

15. A use of at least one monitoring and/or control condition provided by the apparatus according to claim 14 for controlling a spraying application to be performed on a target agricultural field.