ROUTE OPTIMIZATION PROCEDURE

Description

[0001] The present invention relates to a method for route optimization according to the preamble of claim 1, a computer system for route optimization according to the preamble of claim 14, an agricultural machine according to the preamble of claim 16 and a computer program product according to claim 17.

[0002] When cultivating agricultural land, for example, a given parcel of land, an agricultural machine travels the entire area in several lanes one after the other. Turnarounds are necessary to switch between the parallel lanes, which, together with the distances within the lanes, form the entire route. For efficient cultivation, route planning and the selection of a driving strategy are necessary. Three different parameters or groups of parameters can be distinguished: the orientation of the (parallel) lanes, the positions of the lanes, and the order in which the lanes are to be traveled.

[0003] In purely human planning, the selection and definition of a driving strategy is based on the accumulated experience of the machine operator. In automated or assisted planning, the following elementary steps are carried out independently of one another and, in some cases, only individually. In a first step, an orientation is selected taking into account an optimization criterion (e.g., minimizing the number of turning maneuvers). This is achieved by solving a mathematical optimization problem. In a second step, the order in which lanes should be driven is determined, for example, with the goal of minimizing the distance traveled or the time required by the machine in the headland. This is achieved by solving a combinatorial optimization problem.Such automated planning and definition of a driving strategy is carried out either by a software component within a farm management information system or as a vehicle-related component in the form of a feature of a classic steering system.

[0004] EP 1 602 267 A2 discloses a route planning system for agricultural machines, wherein the agricultural machine is assigned a defined working width for generating travel routes within a territory, and wherein the route planning system comprises dynamic adjustment of the planned route. In particular, the planned route can be dynamically adjusted depending on the actual machine position and the actual machine orientation. Furthermore, the route can be formed from a plurality of routes determined according to optimization criteria, wherein the selection of a next route to be processed is determined by optimization criteria.

[0005] EP 3 591 488 A1 discloses a system for determining a driving route, comprising a route generation unit that generates planned driving routes containing a plurality of work routes along which a work vehicle is caused to perform an autonomous drive, a control unit that causes the work vehicle to perform an autonomous drive along each of the planned driving routes, an information collection unit that collects position information and orientation information about the work vehicle, and a determination unit that determines a candidate route for an autonomous drive on which the work vehicle is allowed to start an autonomous drive before the work vehicle starts an autonomous drive.

[0006] US 2019 / 0208695 A1 describes a method for automatic route planning to optimize a machine's route within an area. The parameters considered are the coordinates of a surface contour, a working width of the machine, and a distance between the surface contour and a peripheral, closed machine route within the surface contour. By taking this distance into account, an eroded surface contour is determined from the surface contour. An alignment of lanes within an interior of the eroded surface contour, referred to as inner lanes, is selected. By intersecting the inner lanes with the contour of the eroded area, a lane grid and a corresponding connected, undirected transition graph of the area are generated.Furthermore, a mathematical algorithm is performed to determine a shortest route taking into account at least the mentioned parameters, the eroded area and the provided lane grid.

[0007] DE 102016 121 523 A1 relates to a method for predictively generating data for controlling a route and an operating sequence for an agricultural vehicle and an agricultural machine. Vehicle and / or machine data are automatically acquired and stored via a sensor device arranged on individual vehicles and / or individual machines to generate a vehicle and machine model. The vehicle and machine model is mapped onto a three-dimensional predictive terrain model, and route control data for defining a route and / or machine control data for controlling machine components are calculated.

[0008] WO 2021 / 025108 A1 discloses an automatic driving system for a work vehicle. The driving system includes a measuring part that measures a position and orientation of the work vehicle, and an automatic driving control part that performs driving control according to parallel paths in a working state. A deviation detection part detects an angular deviation and a lateral deviation of the work vehicle relative to the parallel paths at the time the work vehicle reaches the entrance area. If a start condition of the driving control is not met, a position adjustment drive is executed through an operation that combines a forward-backward direction change and a steering operation of the work vehicle.

[0009] EP 3 508 045 A1 discloses an autonomous driving system for a work vehicle that moves around a worksite during work. The system includes a satellite positioning module, an area setting unit that sets an area to be processed at the worksite, a route management unit that calculates a travel route element set that is a concatenation of multiple travel route elements constituting a travel route, a travel route element selection unit that sequentially selects a next travel route element to be traveled from the travel route element set, and an autonomous travel control unit that causes the work vehicle to travel autonomously based on the next travel route element and the vehicle position.

[0010] The object of the invention is to enable optimized agricultural processing of a given processing area.

[0011] The object is achieved by a method having the features of independent patent claim 1. Advantageous embodiments can be found in the dependent claims.

[0012] For this purpose, a method is created for route optimization during field cultivation with at least one processing step, in which a predetermined processing area is traversed by an agricultural machine according to a route with a plurality of parallel lanes, wherein the route can be characterized by an alignment of the lanes, a positioning which determines the positions of all lanes, and an order in which the lanes are traversed, wherein: for at least one processing step, an automatic optimization of the sequence is carried out for each of a plurality of combinations of an orientation and a positioning, in that an optimal sequence is determined from a plurality of sequences for this combination in accordance with a defined optimization criterion for route optimization; for at least one processing step, an automatic optimization of the route is carried out by determining from the plurality of combinations, taking into account the optimal sequence determined for the respective combination, an optimal combination in accordance with the optimization criterion, which corresponds to an optimal route for this processing step; and control data for controlling at least one agricultural machine during a processing step are automatically generated, wherein the control data represent the optimal route.

[0013] The method is intended for route optimization during field cultivation, which field cultivation comprises a plurality of processing steps. Each processing step is carried out by an agricultural machine, whereby different processing steps are generally performed by different agricultural machines. However, several or all work steps can also be carried out by the same agricultural machine. The agricultural machine can also be referred to as an agricultural work machine. This expressly includes combinations comprising a tractor and at least one machine pulled by it and / or a trailer. The agricultural machine can in particular be a harvesting machine such as a forage harvester, a combine harvester, a baler or a loading wagon. However, it could also be, for example, a tractor with a tedder, a plough, a fertilizer spreader, a slurry tanker or the like.

[0014] During the cultivation step, the agricultural machine travels along a route comprising a plurality of parallel lanes within a cultivation area, for example, a field, a croft, or a portion thereof. For the sake of simplicity, the terms "field" and "crop" are used synonymously below. The actual cultivation, such as plowing, fertilizing, mowing, tedding, harvesting, or the like, typically takes place within this cultivation area. A lane, which could also be referred to as a cultivation lane, corresponds in this context to a portion of the agricultural machine's route within the cultivation area. Typically, the entire or at least predominant area of ​​the cultivation area is cultivated successively by driving along individual lanes. The lanes run parallel, meaning that the distance between two adjacent lanes remains constant along their entire length.In addition to the lanes within the cultivation area, the driving route also includes turning lanes, which are necessary for lane changes and / or turning paths when the agricultural machine moves from one lane to another, usually in a headland adjacent to the cultivation area and, more precisely, located at its edge. Such a headland can, for example, be located outside the field or it can be an edge area of ​​the field that is pre-cultivated, for example, harvested, at the beginning of cultivation or, in some cases, even afterward, which may also depend on the type of cultivation.

[0015] Overall, the route can be characterized by the alignment of the lanes, a positioning that determines the positions of all lanes, and the order in which the lanes are traveled. In the case of straight lanes, the alignment corresponds to a horizontal or azimuthal angle; one could also say, a cardinal direction in which all lanes run. It goes without saying that instead of an angle specified in degrees, other information could be used that allows for a clear assignment. However, the alignment does not determine the positions of the individual lanes. This is done by positioning. One possibility is that the positioning for each lane contains two-dimensional coordinates of a point through which the respective lane passes. Together with the alignment, the arrangement of the lane is thus completely determined.As a rule, however, the distance or lateral offset between adjacent lanes is identical for all lanes and corresponds to a lane width that can correspond to the effective working width of the agricultural machine, i.e. the width that can be effectively worked transversely to the direction of travel. In the case of a combine harvester or a forage harvester, for example, this would be the width of the harvesting header, in particular less a safety margin of a few centimeters. In the case of a baler or a loader wagon, the lane width usually corresponds to the working width of the previous process, such as raking or mowing with swath formation. Thus, with knowledge of the lane width, the positions of all lanes can be determined using a single point on a lane, the two-dimensional coordinates of which can be used to express the positioning.Finally, the driving route can be characterized by the order in which the lanes, which are spatially defined in particular by orientation and positioning, are traversed. This allows for various driving strategies to be represented, for example, the agricultural machine changing from one lane to the spatially next lane, or to the next but one, via the next but one, etc. For example, in irregularly shaped processing areas, it may also be useful for a change to the next lane in one sub-area, while a change to the next but one occurs in another sub-area, or similar. With regard to the data volume, the order of N lanes can be expressed by an N-tuple, for example, an N-dimensional vector.

[0016] The method comprises the following steps, although these do not necessarily have to be performed in the order listed. The chronological sequence of two steps may be reversed from the order in which they are listed. It is also possible for two steps to be performed entirely or partially in parallel.

[0017] In one method step, an automatic optimization of the sequence is carried out for at least one processing step for each of a plurality of combinations each comprising an orientation and a positioning. This is done by determining from a plurality of sequences an optimal sequence for this combination according to a defined optimization criterion for route optimization. This means that for the respective processing step, a plurality of possible combinations are considered, with each combination consisting of an orientation and a positioning or combining them. For each of these combinations, which in particular completely determines the spatial arrangement of the lanes, an optimization of the sequence is carried out by determining from a plurality of sequences an optimal sequence with regard to the optimization criterion.A number of sequences are considered, representing candidates for an optimal sequence, and based on the single optimization criterion for route optimization, it is investigated which of these sequences is optimal for the respective combination, particularly of orientation and positioning. The optimization is carried out automatically, i.e., mechanically or computer-assisted. Where the term "automatic" is used here and below, this includes in particular the possibility that the corresponding processes are performed entirely or partially by software implemented on suitable hardware. This step, as well as further steps of the method, can be performed, for example, by a farm management information system (FMIS).

[0018] In principle, all conceivable sequences could be examined for optimization—that is, with N lanes, N sequences! However, many of these theoretically conceivable sequences can be excluded from the outset, for example, a sequence with multiple lane changes into distant lanes, or the like, which significantly reduces the computational effort. This sequence optimization is performed for each of the multiple combinations, whereby different optimal sequences can generally result for different combinations. If this process step is performed for multiple processing steps, different combinations of alignment and positioning can be used for different processing steps.Even if this is not the case, different optimal sequences usually result for different processing steps, simply because different agricultural machines operate with different track widths, and thus one and the same positioning results in a different number of lanes. Furthermore, for example, the turning radius of different agricultural machines can differ, so that a certain type of lane change may be efficient for one agricultural machine, while inefficient or even impossible for another.

[0019] Furthermore, an automatic optimization of the route is carried out for at least one processing step by determining from the plurality of combinations, taking into account the optimal sequence determined for the respective combination, an optimal combination according to the optimization criterion, which corresponds to an optimal route for this processing step.

[0020] While sequence optimization can be viewed as optimization at a lower level, route optimization can be viewed as optimization at a higher or upper level, whereby here and in the following, one can also refer to "optimization levels" instead of "levels." This involves determining which of the majority of combinations of orientation and positioning, together with the optimal sequence determined for this, is considered optimal according to the optimization criterion. It is understood that the aforementioned step of determining the optimal sequence must have been performed for a specific combination before this combination can be compared with other combinations in this regard. However, it is not necessary to first determine the total set of all combinations to be considered.For example, it would be conceivable to initially cover the entire range of all combinations using a comparatively coarse grid, with, for example, the orientation being varied in 10° increments and the positioning in 50 cm increments. Subsequently, an area that has proven to be advantageous compared to other areas could be examined using a finer grid, with, for example, the orientation being varied in 1° increments and the positioning in 10 cm increments. The optimal combination thus determined, together with the associated optimal sequence, corresponds to an optimal route for this processing step. If multiple processing steps are performed, the optimal route usually differs between the different processing steps. The term "optimal route" should be understood to mean the best route found according to the optimization criterion.In some circumstances, a better route might actually exist, but it was not found, for example because not enough combinations of orientation and positioning were examined.

[0021] In any automatic optimization, different methods can be used to find the optimal solution, or to search for the optimal solution, in particular metaheuristic methods such as simulated annealing, genetic or evolutionary algorithms.

[0022] In a further method step, control data for controlling at least one agricultural machine during a processing step is automatically generated, which control data represents the optimal route. The control data always contains the information necessary to steer the agricultural machine along the optimal route. In the corresponding processing step, the agricultural machine can be controlled according to the control data. The term "control" here generally refers to any targeted influence on the orientation and / or movement state of the agricultural machine, for example, steering, accelerating, decelerating, etc. The agricultural machine can therefore be controlled using the control data so that it ideally follows the determined optimal route. The format of the control data and its content can be selected differently, particularly depending on the respective agricultural machine and, if applicable,other components used to implement the method. In particular, the route can be represented by a more or less dense sequence of waypoints. The control data can also contain explicit steering instructions, or merely position information for waypoints, with the agricultural machine determining the appropriate steering parameters to get from one waypoint to the next. It is understood that control data can be generated for a plurality of agricultural machines for one processing step each and / or for one agricultural machine for a plurality of processing steps. In the case of multiple processing steps, the optimal routes for individual processing steps can differ from one another.

[0023] According to the invention, the plurality of combinations contains different combinations of a single orientation with different positionings. This means that there is no clear association between the orientation and the positioning; rather, one and the same orientation is combined with different positionings, whereby the resulting different combinations can be compared with each other during optimization. This differs from prior art methods, in which different orientations are considered during optimization, but the respective orientation is combined with a specific positioning according to a specified criterion.Such a criterion is normally defined with regard to a specific lane, for example such that a side-by-side lane is positioned such that the agricultural machine, taking into account its individual working width, moves exactly along the edge of the cultivation area without protruding beyond the cultivation area. Even if such a definition is suitable for the individual lane, it can nevertheless lead to a suboptimal solution when considering all lanes as a whole. For example, this choice could result in an extremely short, and therefore definitely uneconomical, lane at the opposite edge of the cultivation area, which would not occur, for example, if one also accepted that the first-mentioned side-by-side lane is shifted in such a way that the agricultural machine protrudes laterally beyond the cultivation area.The method according to the invention is capable of comparing different possible positionings for one and the same orientation with regard to the optimization criterion and thus of determining among these the positioning which is optimal in combination with the orientation in question.

[0024] Preferably, the optimization criterion is based at least partially on an optimization of an optimization value determined by summing contributions from individual route sections of a travel route. In many cases, the optimization value can also be referred to as an effort value or cost value, in which case the optimization of the optimization value lies in a minimization. However, it is also conceivable that, depending on the type or definition of the optimization value, the optimization lies in a maximization of the same. In the simplest case, the optimization criterion consists in a minimization or maximization of the optimization value. This means that the travel route is optimal which optimizes the optimization value. As will be explained below, other variables, for example further optimization values, could also be taken into account, so that the optimization of one optimization value competes with other objectives.In each case, the optimization value is determined by summing the contributions of individual route sections along a route. Each route section is assigned a contribution, and if the route includes the route section, the contribution is added when determining the optimization value. For example, if the optimization value corresponds to a distance that must be traveled along the route, the contribution of a route section corresponds to its length. The optimization value. W can be calculated as follows: W = ∑ j W j where W j represents the contribution of the j-th route section and the sum over j runs over all route sections.

[0025] In the case of multiple processing steps, contributions from the routes of the processing steps can also be summed up, so that the optimization value can be calculated as follows: W = ∑ i W i = ∑ i , j W ij where W ij represents the contribution of the j-th route section of the i-th route and the sum over i runs over all processing steps.

[0026] It is preferred that the optimization criterion, for each optimization, considers both route sections of the lanes and route sections of turning lanes connecting the lanes, with the route sections of the turning lanes being considered depending on the respective sequence. This means that if, for example, a specific optimization value of the route is to be minimized or maximized, not only the contribution of the lanes to this optimization value is considered, but also the contribution of the turning lanes that must be traversed within the route when changing from one lane to the next, depending on the underlying sequence. The latter influences, for example, the distance to be covered for individual turning lanes, in particular depending on whether changing to an adjacent lane, the next lane, the lane after but one, etc., the required time, the fuel consumption, and / or other parameters.This means that the contribution taken into account is normally not constant for all possible sequences, but depends on the respective sequence, i.e., it is a function of the sequence. Of course, this does not preclude the possibility that the contribution may be the same for some sequences. This embodiment is typically combined with the above-mentioned embodiment, so that the optimization criterion is based at least in part on an optimization of an optimization value determined by summing contributions from route sections of the lanes and route sections of the turning paths of a travel route, wherein the contributions of the route sections of the turning paths depend on the sequence.

[0027] Advantageously, the optimization criterion considers a route section of a turning lane connecting two lanes for each optimization, at least depending on the relative position of these lanes. This means that, with regard to the influence of a turning lane, at least the position of the connected lanes relative to each other is taken into account. The relative position can be expressed, for example, by the distance between the two lanes perpendicular to the alignment. It is also possible that the two lanes, in particular their endpoints, are offset parallel to the alignment if the boundary of the processing area is not perpendicular to the alignment. In addition to the relative position, the absolute positions could also be considered, for example, because the soil conditions vary locally and affect the turning lane, because the space available for the turning lane varies locally, etc.

[0028] One possibility for optimising the route is to treat the orientation, which can be characterised in particular by a single angle specification, and the positioning, which can be characterised in particular by a single, in particular two-dimensional, point in the processing area, as equal variation parameters.

[0029] This allows you to compare routes that may differ in both orientation and positioning. Alternatively, you can first optimize one variation parameter, such as orientation or positioning, while keeping the other unchanged.

[0030] Such a design provides that, in order to optimize the route, for each of a plurality of orientations, an optimal positioning is determined from a plurality of positionings, and the optimal combination is determined from a plurality of combinations of an orientation and the optimal positioning determined for this. This means that a total of a plurality of orientations are considered. For each of these orientations, a plurality of positionings is considered, and from these, an optimal positioning is determined, which in turn implicitly includes the determination of the optimal sequence for each combination of the orientation with one of the positionings. From a hierarchical perspective, an optimal sequence for a combination of orientation and positioning is determined at the lower level.Then, at a middle level, an optimal positioning for a specific alignment is determined, based on the optimal sequence determined at the lower level. At an upper level, an optimal alignment is then determined, based on the optimal positioning and optimal sequence determined at the middle and lower levels. When searching for an optimal positioning, it is not necessary to consider the entire processing area. If the positioning is characterized by two-dimensional coordinates of a single point, as described above, it is sufficient to consider only points on a straight line perpendicular to the alignment, along a distance corresponding to the track width. Shifts in the direction of the alignment do not change the positioning.Likewise, a shift across the alignment by an integer multiple of the track width will result in identical positioning. For example, for an agricultural machine with a track width of 8 m and a north-south orientation, it is sufficient to examine positions along an 8 m long stretch in an east-west direction.

[0031] Regardless of whether only one processing step is planned or - as should generally be the case - several processing steps, it is possible that only the individual processing step is taken into account when optimising the route. In reality, however, there is a connection between the individual processing steps, in particular because the alignment of the tracks is the same for all processing steps. There may also be other connections, for example because each agricultural machine causes soil compaction as it passes through the processing area, which significantly impairs the affected area. In this respect, it is advantageous if an agricultural machine drives through the same area with its wheels as often as possible, as far as possible, instead of compacting a previously uncompacted area.For these and other reasons, a holistic consideration of several, in particular all, processing steps is advantageous. An advantageous embodiment of the method therefore provides that the optimization criterion takes each of a plurality of processing steps into account, so that the optimization of the route for the respective processing step takes place depending on all processing steps, with the orientation being the same for all processing steps. The travel routes in the individual processing steps are therefore linked by the fact that the orientation is the same in all processing steps. For example, it could be the case that a combination that is extremely advantageous in one processing step contains an orientation that inevitably leads to a highly disadvantageous combination in another processing step.These disadvantages can be avoided if multiple processing steps are considered when optimizing the route. The corresponding optimization criterion is formulated in such a way that a plurality of processing steps are taken into account. In this case, the individual route of an agricultural machine may be suboptimal when viewed in isolation, but when viewed in conjunction with the routes of other processing steps, it leads to an optimal overall result. One possible implementation of this embodiment is to calculate an optimization value, as described above, by summing the contributions from all processing steps and then minimizing or maximizing this optimization value.

[0032] As an alternative to the embodiment described above, in which the orientation is first recorded and the positioning is optimized, it is also possible to optimize the route by determining an optimal orientation from a plurality of orientations for each of a plurality of positionings, and to determine the optimal combination from a plurality of combinations of a respective positioning and the optimal orientation determined for this. This means that a total of a plurality of positionings are considered. For each of these positionings, a plurality of orientations is considered, and from these an optimal orientation is determined, which in turn implicitly includes determining the optimal sequence for each combination of the positioning with one of the orientations. Viewed hierarchically, an optimal sequence for a combination of orientation and positioning is found at the lower level.Then, at a middle level, an optimal alignment for a specific position is found, based on the optimal sequence determined at the lower level. At an upper level, an optimal positioning is then determined, based on the optimal alignment and optimal sequence determined at the middle and lower levels. When searching for an optimal positioning, it is not necessary to consider the entire processing area. Rather, it is sufficient to consider positions within a search circle whose diameter corresponds to the track width. For example, for an agricultural machine with a track width of 8 m, it is sufficient to examine positions within a circle with a diameter of 8 m.

[0033] According to one embodiment, the optimization criterion is based at least partially on minimizing a travel distance. "At least partially" in this context means that minimizing the travel distance does not have to be the sole goal, but that other values ​​should also be minimized or maximized, so that, for example, a compromise is reached that may differ from solely minimizing the travel distance. In particular, it may be intended to minimize the entire distance traveled on the route.

[0034] Alternatively, the total travel distance required for turning lanes can be minimized. This can be considered an "unproductive" travel distance. It goes without saying that the two criteria mentioned are equally important at the lower optimization level, i.e., when finding the optimal sequence, since the travel distance in the lanes is always the same.

[0035] Alternatively or additionally, the optimization criterion can be based at least in part on minimizing travel time. This can be based on the travel time for the entire route. However, the total travel time required for all lane changes and / or turning maneuvers can also be considered, which in turn represents "unproductive" travel time. Unlike the travel distance, the two criteria at the lower level are not necessarily synonymous. For example, the processing area could have a gradient that allows lanes to be traveled faster in one direction than in the opposite direction. The number and individual length of the total "downhill" or "uphill" lanes traveled can vary depending on the sequence, and thus also the total travel time required for the lanes.

[0036] Alternatively or additionally, the optimization criterion can be based at least partially on minimizing energy consumption. This typically considers the expected energy consumption for the entire route. This depends on the total distance traveled, but may also be influenced by other parameters. For example, energy consumption can also depend on the orientation, for example, if the agricultural machine has to negotiate a more or less steep incline within the lanes. The order can also influence energy consumption, since the number and individual length of the total "downhill" or "uphill" lanes traveled can vary, as already mentioned.

[0037] Under certain circumstances, the optimization criterion may consist of minimizing or maximizing a single parameter or optimization value, for example, minimizing the entire travel distance. Depending on the nature of the processing area, the type of processing operation to be performed, the performance data of the agricultural machine, and other factors, minimizing or maximizing one optimization value may, to a certain extent, compete with an equally desirable minimization or maximization of another optimization value. In this case, the isolated optimization of a single optimization value often does not represent a satisfactory solution. One embodiment therefore provides that the optimization criterion is based on the optimization of a weighted combination of optimization values.Instead of a weighted combination, one can usually also speak of a linear combination, although it is conceivable in principle for an optimization value to be non-linear, for example, quadratic. One optimization value could, for example, be the travel distance, while another optimization value is the travel time. The optimization criterion could then be the minimization of a sum, with one term proportional to the travel time and another term proportional to the travel distance. By selecting suitable weighting factors or normalization factors, the relative weight of the respective optimization value can be adjusted. The sum can also be referred to as the "total optimization value." W gesamt which is defined as follows: W gesamt = ∑ k a k W k where W k the k-th optimization value, for example the distance, travel time, etc. and a k is the respective weighting factor.

[0038] Alternatively, the optimization criterion can be based on a Pareto optimization of several optimization values. This means that – within a parameter range under investigation – a parameter set is sought that optimizes the optimization values ​​to the extent that no other parameter set improves one of the optimization values ​​without worsening another. For example, a parameter set consisting of orientation, positioning, and sequence could represent a Pareto optimum with respect to travel time and distance if no other parameter set delivers a shorter travel time without delivering a longer travel distance, and no other parameter set delivers a shorter travel distance without delivering a longer travel time.

[0039] It is conceivable that the agricultural machine has a computer system, for example an evaluation unit or a processing unit, which is capable of determining the optimal route based on sufficient information about the processing area. In many cases, however, it is more efficient if the optimal route is determined externally for at least one agricultural machine and the control data is generated externally and transmitted to the agricultural machine. This can particularly apply to all agricultural machines in the case of multiple processing steps. The determination of the optimal route and the generation of the control data are then carried out via a central system or a central evaluation unit, which can even be stationary in a building that does not even have to be near the processing area.Through wireless communication, the generated control data could be transmitted to the respective agricultural machine, which would then use it to follow the optimal route. Centralized, external processing requires only a few resources in terms of computing capacity and storage space on the agricultural machine side. This also makes it easier to add or remove individual agricultural machines and their associated processing steps as needed, compared to if the resources required for planning were localized within one of the agricultural machines.

[0040] Control data can be generated for at least one autonomously driving agricultural machine, which autonomously performs at least one processing step based on the control data. This means that, as long as the control data is available to this agricultural machine, it can follow the optimal route and perform the corresponding processing step without human intervention. The agricultural machine can use various internal and / or external sensors for navigation. Under certain circumstances, it can orient itself at least partially according to a structure of the processing area, for example, a structure of a crop that the processing area has, in particular crop boundaries, rows, etc. However, other structures can also be used without an existing crop, for example, furrows.Depending on the nature of the area being worked on, different sensors can be used, such as mechanical or optical sensors, active or passive sensors. In particular, the agricultural machine can use a GNSS receiver to determine its current actual position and compare it with a target position corresponding to the optimal route.

[0041] Alternatively or, particularly in the case of multiple processing steps, additionally, control data can be generated for at least one agricultural machine steered by a driver, so that control instructions for the driver can be generated based on the control data. The control data can be available within the agricultural machine and converted into control instructions. Alternatively, it would also be conceivable for control instructions to be created externally based on the control data and transmitted to the agricultural machine. It is also conceivable for one and the same agricultural machine to be steered autonomously for a time and by a driver for a time. The control instructions can be issued visually and / or acoustically. The control instructions could explicitly instruct the driver how to steer the agricultural machine, or, for example, a target driving line could be displayed on a screen, which the driver can use as a guide.

[0042] The object is further achieved by a system for route optimization during field processing with at least one processing step, in which a predetermined processing area is traversed by an agricultural machine according to a route with a plurality of parallel lanes, wherein the route can be characterized by an alignment of the lanes, a positioning which determines the positions of all lanes, and an order in which the lanes are traversed, wherein the system is configured to to carry out an automatic optimization of the sequence for each of a plurality of combinations of an orientation and a positioning for at least one processing step, by determining from a plurality of sequences an optimal sequence for this combination according to a defined optimization criterion for route optimization; to carry out an automatic optimization of the route for at least one processing step, by determining from the plurality of combinations, taking into account the optimal sequence determined for the respective combination, an optimal combination according to the optimization criterion, which corresponds to an optimal route for this processing step; and to automatically generate control data for controlling at least one agricultural machine during a processing step, wherein the control data represent the optimal route.

[0043] According to the invention, the plurality of combinations includes different combinations of a single orientation with different positionings.

[0044] The terms mentioned have already been explained above with reference to the method according to the invention and will therefore not be explained again. The computer system comprises at least one computer or a computer or a data processing unit. It can also comprise additional components, for example, wireless and / or wired interfaces for one-way or two-way communication with other devices. Advantageous embodiments of the computer system according to the invention correspond to those of the method according to the invention. In particular, the computer system can be a farm management information system that is arranged outside the at least one agricultural machine, for example, stationary within a building. The computer system could also be a mobile unit, for example, a laptop, tablet, smartphone, etc., which displays control instructions for the driver or transmits control data, in particular wirelessly, to the agricultural machine. More generally, the computer system can be designed externally with respect to the at least one agricultural machine and configured to generate the control data for transmission to the at least one agricultural machine. It can have an interface for data transmission to the at least one agricultural machine and be configured to transmit the control data, in particular wired and / or wirelessly, to the at least one agricultural machine.

[0045] Alternatively, the computer system can be integrated into an agricultural machine, meaning it can be part of the agricultural machine and located within it. This is particularly possible if only one processing step is performed. In any case, the computer system can be partially implemented in software. Regardless of whether the computer system is part of the agricultural machine or not, it can be configured to control the agricultural machine according to the control data.

[0046] According to the second alternative, the invention also provides an agricultural machine with a computer system for route optimization during field processing with at least one processing step, in which a predetermined processing area is traversed by an agricultural machine according to a route with a plurality of parallel lanes, wherein the route can be characterized by an alignment of the lanes, a positioning which determines the positions of all lanes, and an order in which the lanes are traversed, wherein the computer system is configured to to carry out an automatic optimization of the sequence for each of a plurality of combinations of an orientation and a positioning for at least one processing step, by determining from a plurality of sequences an optimal sequence for this combination according to a defined optimization criterion for route optimization; to carry out an automatic optimization of the route for at least one processing step, by determining from the plurality of combinations, taking into account the optimal sequence determined for the respective combination, an optimal combination according to the optimization criterion, which corresponds to an optimal route for this processing step; and to automatically generate control data for controlling at least one agricultural machine during a processing step, wherein the control data represent the optimal route.

[0047] According to the invention, the plurality of combinations includes different combinations of a single orientation with different positionings.

[0048] Again, preferred embodiments of the agricultural machine according to the invention correspond to those of the method according to the invention.

[0049] The invention further provides a computer program product with program code means that enable a computer system to execute the method according to the invention. The computer program product thus includes software that implements the method according to the invention on the hardware of the computer system. It can be in the form of a data carrier on which the software and / or the program code means are stored, in particular volatilely and / or non-volatilely. The data carrier can also be permanently integrated into the computer system or can be integrated into it.

[0050] The invention is described below with reference to figures. The figures are merely exemplary and do not limit the general concept of the invention. They show Fig. 1 is a plan view of a part of a parcel of land with an agricultural machine and a computer system according to the invention for route optimization; Fig. 2 is a plan view of the field from Fig.1 with a first driving route; Fig. 3 a plan view of the field Fig.1 with a second route; Fig. 4 a top view of the field Fig.1 with a third route; Fig. 5 a top view of the field from Fig.1 with a fourth route; Fig. 6 a plan view of the field from Fig.1 with a fifth route; Fig. 7 shows a flowchart of a first method according to the invention for route optimization; and Fig. 8 shows a flowchart of a second method according to the invention for route optimization.

[0051] Fig. 1 shows a plan view of part of a parcel of land 20 and an agricultural machine 10, for example, a forage harvester. A computer system 1 according to the invention for route optimization is shown in a highly schematic manner, which in this case is arranged outside the agricultural machine 10, for example, in a building that may be far away from the parcel of land 20. The computer system 1 may be formed by a farm management information system or may represent a part thereof. It has an interface (not shown individually here) for wireless data transmission to the agricultural machine 10. It is provided that the agricultural machine 10 performs a processing step in a processing area 21 of the parcel of land 20, for example, harvesting and chopping corn, with the agricultural machine 10 traveling along a plurality of parallel lanes.The processing area 21 is surrounded by a surrounding headland 22, which serves to enable the agricultural machine 10 to perform turning maneuvers between the individual lanes S 1 -S 5. Overall, a plurality of successive processing steps can be provided, which are generally carried out by different agricultural machines 10.

[0052] Before field processing, the computer system 1 carries out a method according to the invention for route optimization, which is described in the flow chart in Fig. 7 and based on the plan view of parcel 20 in Fig. 2 bis 6 The computer system 1 has various data relating to the parcel of land 20, in particular the geometric dimensions of the processing area 21 and, if applicable, those of the headland 22. Optionally, further data such as the local soil conditions or any existing gradient can be included. Furthermore, the computer system 1 has data relating to the agricultural machine 10, in particular its effective working width and its minimum turning radius. Furthermore, the performance data of the agricultural machine 10 can be known as a function of soil conditions, gradient, or other factors, for example, a speed dependent thereon, fuel consumption, or the like.

[0053] Route optimization serves to find an optimal route F opt for agricultural machine 10, which can be expressed by an optimal alignment A opt , an optimal positioning P opt , and an optimal sequence R opt . For this purpose, an optimization criterion is defined, and an optimization is performed based on the optimization criterion. This can be, for example, the minimization of the total travel distance, the minimization of the travel distance for turning maneuvers, the minimization of the total travel time, the minimization of fuel consumption, or the like. It is also possible to define several subcriteria, between which a certain degree of competition may exist. On the one hand, the minimization of a weighted combination of different optimization values, for example, a travel distance, a travel time, etc., can be aimed for, on the other hand, a Pareto optimization can also be carried out with regard to various optimization values.

[0054] Each driving route can be characterized by an orientation A 1 , A 2 , a positioning P 1 -P 4 and a sequence R 1 -R 5 . The orientation A 1 , A 2 refers to how the lanes S 1 -S 5 are oriented relative to a reference system, whereby the figures show a two-dimensional coordinate system with an X-axis and a Y-axis, whereby the X-axis can, for example, point east while the Y-axis points north. For the straight, parallel lanes S 1 -S 5 shown here, the orientation A 1 , A 2 in such a reference system can be represented by a single angle specification. The positioning P 1 -P 4 designates the positions of the lanes S 1 -S 5 , whereby, knowing a lane width b corresponding to the above-mentioned effective working width of the agricultural machine 10, it is sufficient to specify a two-dimensional coordinate point on one of the lanes S 1 -S 5 , as in Fig. 1 bis 6 To cover the entire processing area 21, depending on the orientation A 1 , A 2 , different numbers of lanes S 1 -S 5 may be necessary, in the example according to Fig. 2 There are fourteen lanes S 1 -S 5 , which are arranged according to a first orientation A 1 and a first positioning P 1 . These lanes S 1 -S 5 are traveled in a first sequence R 1 , which can be selected differently.

[0055] The optimization is carried out in a nested form, whereby, depending on the perspective, optimization is carried out on two or three different levels. One can say that at a higher level, at S100 in the flow chart of Fig. 7 , an optimal combination of an alignment A 1 , A 2 and a positioning P 1 -P 4 is determined, while on a lower level at S150 an optimal sequence R opt is determined. More precisely, one can say that with more than two levels on the upper level at S110 an optimal alignment A opt is determined, on a middle level at S130 an optimal positioning P opt is determined and on the lower level at S150 an optimal sequence R opt is determined, whereby the optimizations on the three levels are nested within each other, as graphically shown in Fig. 7 is clearly visible.

[0056] At the upper level, an alignment A 1 , A 2 is selected at S120. For this, a first positioning P 1 is selected in step S140, which forms a first step for determining the optimal positioning P opt. In step S160, a sequence R 1 -R 5 is again selected for the aforementioned alignment A 1 , A 2 and the positioning P 1 -P 4. In step S170, it is checked whether an optimal sequence R opt has already been found, which is generally negative for the first sequence R 1, so that the method returns to step S160, where a Fig. 3 shown second sequence R 2 is selected. For each sequence R 1 -R 5 it is checked whether this, in combination with the respectively selected alignment A 1 , A 2 and positioning P 1 -P 4, is optimal with regard to the optimization criterion, for example whether it minimizes an optimization value such as the total travel time, the total travel distance or the like. To calculate the optimization value, which can also be referred to as the cost value, the contributions of individual route sections of the travel route are summed up, which includes both the route sections of lanes S 1 -S 5 and the route sections of turning lanes W 1 -W 4. It is envisaged that the contributions of the turning lanes W 1 -W 4 are realistically taken into account in that they depend at least on the relative positions of the lanes S 1 -S 5 connected by the turning lane W 1 -W 4, but preferably depend explicitly on the connected lanes S 1 -S 5.The latter makes sense insofar as, for example, in . Fig. 2 the second turning path W 2 from the second lane S 2 to the third lane S 3 objectively has a longer distance and requires a longer travel time than, for example, the fourth turning path W 4 from the fourth lane S 4 to the fifth lane S 5 . In the case of multiple processing steps, the optimization value can be determined by summing the contributions of the travel routes of all processing steps. To find the optimal sequence R opt, all conceivable sequences R 1 -R 5 could be tested. This would be a reliable approach, but inefficient in terms of time and computational effort. Instead, various numerical methods, in particular metaheuristic methods, can be used.

[0057] If it is decided in step S170 that the optimal order R opt has been found, the optimization at the lower level is finished and it is checked in step S180 whether the optimal positioning P opt for the respective orientation A 1 , A 2 has already been found. If this is not the case, the method returns to step S140 where a new positioning P is selected, for example the one in Fig. 4 shown second positioning P 2 , in which the total number of lanes S 1 -S 5 increases to fifteen. For each positioning P 1 -P 4 , the optimization must be carried out on the lower level, which means that an optimal sequence R opt must be determined in each case. Since a shift of the positioning P 1 -P 4 in the direction of the alignment A 1 , A 2 does not change the actual position of the lanes, just as a shift transverse to the alignment A 1 , A 2 by an integer multiple of a lane width B does not change the actual position of the lanes, the search for the optimal positioning P opt can be limited to a Fig. 1 shown search line L, which runs transversely to the first alignment A 1 and whose length corresponds to the track width b.

[0058] If it is decided in step S180 that the optimal positioning P opt has been found, the optimization at the middle level is finished, whereby the optimal positioning P opt and optimal sequence R opt have been found for a specific orientation A 1 , A 2 . If several processing steps are considered, a check is carried out in step S190 as to whether the optimal positioning P opt has already been found for the last processing step. If not, the next processing step for the optimization of the positioning P 1 -P 4 is selected in step S200 and the method returns to S140. For different processing steps, different optimal positionings P opt and sequences R opt generally result. Fig.6 shows, for the first alignment A1, an example of a fourth positioning P4 for a different processing step, which is carried out with a different agricultural machine with a different track width. The result of the optimization of the positioning P1-P4 in the subsequent processing steps may, under certain circumstances, depend on the optimal positioning Popt for the first processing step, for example, if a compacted soil area is to be minimized, which is qualitatively possible if a following agricultural machine drives with its wheels in the tracks of the preceding agricultural machine.

[0059] If it is determined in step S190 that all processing steps have been considered, or in the case of a single processing step, the method continues with step S210. After finding the optimal positioning P opt , or positions for different processing steps, the alignment A is then determined on the upper level, which in combination with the associated optimal positioning P opt and optimal sequence R opt represents an optimal route F opt. In step S220, a check is made as to whether the optimal alignment A opt has already been found. If not, the method returns to step S120, where a new, in Fig. 5 shown alignment A is checked, for which in turn the corresponding optimal positioning P opt and optimal sequence R opt must be determined.

[0060] If it is determined in step S210 that the optimal alignment A opt has been found, optimal parameters for the alignment A 1 , A 2 , the positioning P 1 -P 4 and the sequence R 1 -R 5 are now determined, which correspond to an optimal travel route F opt. The computer system 1 then generates control data D for the agricultural machine 10 in step S220, which corresponds to the optimal travel route F opt, in particular the optimal alignment A opt, optimal positioning P opt and optimal sequence R opt. If the agricultural machine 10 is controlled by a driver, the control data D can correspond to instructions for the driver, based on which he can steer the agricultural machine 10 along the travel route. If the agricultural machine 10 is driving autonomously, the control data D can contain explicit driving commands and / or steering commands for the systems of the agricultural machine 10. In step S230, the control data D is sent wirelessly to the agricultural machine 10, as in Fig. 1 indicated.

[0061] In order to enable the computer system 1 to carry out the method shown, the necessary software can be provided as a computer program product, for example as a mobile or integrated data carrier, which has program code means and / or a program code that implements the method on the hardware of the computer system.

[0062] According to an alternative not shown, the computer system 1 can also be integrated into the agricultural machine 10. In this case, the control data D are available directly in the loading machine 10.

[0063] Fig. 8 shows a second embodiment of a method for route optimization, which in some aspects corresponds to the embodiment and is therefore not explained again. However, here, at the upper optimization level in a block S115, an optimal positioning P opt is determined, while at the middle optimization level in S135, an optimal orientation A opt for the respective positioning P is determined. Regardless of which orientation A is selected, shifts in the positioning P by an integer multiple of the lane width b lead to an identical arrangement of the lanes F, which is why the search for an optimal positioning P opt can be limited to the area of ​​a search circle K, the diameter of which corresponds to the lane width b, as in Fig. 1shown. In step S125, a positioning is selected, and in step S215, it is checked whether the optimal positioning P opt has been found. In step S145, an alignment is selected, and in step S185, it is checked whether the optimal alignment A opt for the respective positioning has been found. Among other things, different positionings are combined with a single alignment in order to test various combinations in this regard. Steps S160 and S170 do not differ from the first embodiment, nor do steps S220 and S230. The embodiment shown here only allows optimization with regard to a single processing step, so that steps S190 and S200 are omitted here without replacement. The reason for this is that the alignment is not selected and optimized at the upper level, but at the middle level.This same alignment must be adopted by the subsequent processing steps, if any, which generally leads to a suboptimal solution for the entirety of all processing steps.

Claims

1. A method for performing driving route optimization in field cultivation having at least one cultivation step in which a predefined cultivation area (21) is driven through by an agricultural machine in accordance with a driving route (F1-F5) containing a plurality of parallel lanes (S1-S5), wherein the driving route (F1-F5) is able to be characterized by an orientation (A1, A2) of the lanes (S1-S5), a positioning (P1-P4) that defines positions of all of the lanes (S1-S5), and an order (R1-R5) in which the lanes (S1-S5) are driven through, wherein, the method comprising the steps of: performing, as a cultivation step, an automatic optimization of the order (R1-R5) for each of a plurality of combinations of in each case an orientation (A1, A2) and a positioning (P1-P4) by ascertaining, from a plurality of orders (R1-R5), an order (Ropt) that is optimum for this combination in accordance with a defined optimization criterion for performing driving route optimization; performing, as a cultivation step, an automatic optimization of the driving route (F1-F5) by ascertaining, from the plurality of combinations, incorporating the optimum order (Ropt) ascertained for the respective combination, a combination that is optimum in accordance with the optimization criterion and that corresponds to an optimum driving route (Fopt) for this cultivation step; and generating control data (D), for controlling at least one agricultural machine during a cultivation step, automatically, wherein the control data represent the optimum driving route (Fopt), wherein the plurality of combinations contains different combinations of a single orientation (A1, A2) with different positionings (P1-P4).

2. The method as claimed in claim 1, wherein the optimization criterion is based at least partially on an optimization of an optimization value that is ascertained by summing contributions of individual route sections of a driving route (F1-F5).

3. The method as claimed in claim 1, wherein the optimization criterion in each optimization takes into consideration both route sections of the lanes (S1-S5) and route sections of turning paths (W1-W4) connecting the lanes (S1-S5), wherein the route sections of the turning paths (W1-W4) are taken into consideration on the basis of the respective order (R1-R5).

4. The method as claimed in claim 1, wherein the optimization criterion in each optimization takes into consideration a route section of a turning path (W1-W4) connecting two lanes (S1-S5) at least on the basis of a relative position of these lanes (S1-S5).

5. The method as claimed in claim 1, wherein, in order to optimize the driving route (F1-F5), an optimum positioning (Popt) from a plurality of positionings (P1-P4) is ascertained (S130) for each of a plurality of orientations (A1, A2), and the optimum combination from a plurality of combinations of in each case one orientation (A1, A2) and the optimum positioning (Popt) ascertained with respect thereto is ascertained.

6. The method as claimed in claim 1, wherein the optimization criterion takes into consideration each of a plurality of cultivation steps, such that the driving route (F1-F5) for the respective cultivation step is optimized on the basis of all cultivation steps, wherein the orientation (A1, A2) is the same for all cultivation steps.

7. The method as claimed in claim 1, wherein, in order to optimize the driving route (F1-F5), an optimum orientation (Aopt) from a plurality of orientations (A1, A2) is ascertained (S135) for each of a plurality of positionings (P1-P4), and the optimum combination from a plurality of combinations of in each case one positioning (P1-P4) and the optimum orientation (Aopt) ascertained with respect thereto is ascertained.

8. The method as claimed in claim 1, wherein the optimization criterion is based at least partially on minimizing a driving distance, minimizing a driving time and / or minimizing an energy consumption.

9. The method as claimed in claim 1, wherein the optimization criterion is based on optimizing a weighted combination of multiple optimization values.

10. The method as claimed in claim 1, wherein the optimization criterion is based on a Pareto optimization of multiple optimization values.

11. The method as claimed in claim 1, wherein the optimum driving route (Fopt) is ascertained externally for at least one agricultural machine and the control data (D) are generated externally and transmitted to the agricultural machine.

12. The method as claimed in claim 1, wherein control data (D) are generated for at least one self-driving agricultural machine, which performs at least one cultivation step autonomously based on the control data (D).

13. The method as claimed in claim 1, wherein control data (D) are generated for at least one agricultural machine steered by a driver, such that control instructions are able to be generated for the driver on the basis of the control data (D).

14. A computer system for performing driving route optimization in field cultivation having at least one cultivation step in which a predefined cultivation area is driven through by an agricultural machine in accordance with a driving route (F1-F5) containing a plurality of parallel lanes (S1-S5), wherein the driving route (F1-F5) is able to be characterized by an orientation (A1, A2) of the lanes (S1-S5), a positioning (P1-P4) that defines positions of all of the lanes (S1-S5), and an order (R1-R5) in which the lanes (S1-S5) are driven through, wherein the computer system is configured to perform the following steps: performing, as a cultivation step, an automatic optimization of the order (R1-R5) for each of a plurality of combinations of in each case an orientation (A1, A2) and a positioning (P1-P4) by ascertaining, from a plurality of orders (R1-R5), an order (Ropt) that is optimum for this combination in accordance with a defined optimization criterion for performing driving route optimization; performing, as a cultivation step, an automatic optimization of the driving route (F1-F5) by ascertaining, from the plurality of combinations, incorporating the optimum order (Ropt) ascertained for the respective combination, a combination that is optimum in accordance with the optimization criterion and that corresponds to an optimum driving route (Fopt) for this cultivation step; and automatically generating control data (D) for controlling at least one agricultural machine during a cultivation step, wherein the control data (D) represent the optimum driving route (Fopt), wherein the plurality of combinations contains different combinations of a single orientation (A1, A2) with different positionings (P1-P4).

15. The computer system as claimed in claim 14, wherein it is formed externally in relation to the at least one agricultural machine and is configured to generate the control data (D) for transmission to the at least one agricultural machine.

16. An agricultural machine having a computer system for performing driving route optimization in field cultivation having at least one cultivation step in which a predefined cultivation area is driven through by an agricultural machine in accordance with a driving route (F1-F5) containing a plurality of parallel lanes (S1-S5), wherein the driving route (F1-F5) is able to be characterized by an orientation (A1, A2) of the lanes (S1-S5), a positioning (P1-P4) that defines positions of all of the lanes (S1-S5), and an order (R1-R5) in which the lanes (S1-S5) are driven through, wherein the computer system (1) is configured, performing, as a cultivation step, an automatic optimization of the order (R1-R5) for each of a plurality of combinations of in each case an orientation (A1, A2) and a positioning (P1-P4) by ascertaining, from a plurality of orders (R1-R5), an order (Ropt) that is optimum for this combination in accordance with a defined optimization criterion for performing driving route optimization; performing, as a cultivation step, an automatic optimization of the driving route (F1-F5) by ascertaining, from the plurality of combinations, incorporating the optimum order (Ropt) ascertained for the respective combination, a combination that is optimum in accordance with the optimization criterion and that corresponds to an optimum driving route (Fopt) for this cultivation step; and automatically generating control data (D) for controlling at least one agricultural machine during a cultivation step, wherein the control data (D) represent the optimum driving route (Fopt), wherein the plurality of combinations contains different combinations of a single orientation (A1, A2) with different positionings (P1-P4).

17. A computer program product containing program code means that enable a computer system to carry out the method as claimed in claim 1.