A method for semi-automatically determining a path for at least one agricultural robot to cultivate farmland.

The semi-automatic path planning method for agricultural robots optimizes path configurations on complex farmland by detecting non-optimal zones and recalculating paths, enhancing efficiency and safety through user interaction.

JP2026510248APending Publication Date: 2026-04-02クーン エスアーエス
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2024-03-14
Publication Date
2026-04-02

AI Technical Summary

Technical Problem

Existing agricultural robot path planning methods are not optimized for complex farmland shapes with obstacles and hazards, and lack user discretion and verification, leading to inefficient and potentially unsafe operations.

Method used

A semi-automatic method for determining agricultural robot paths that involves defining an initial reference direction, detecting non-optimal zones, and recalculating paths based on multiple reference directions to optimize coverage and safety, with user verification and interaction.

Benefits of technology

Enables efficient and safe path planning for agricultural robots on complex farmland by optimizing path configurations and allowing user verification, improving operational efficiency and safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2026510248000001_ABST
    Figure 2026510248000001_ABST
Patent Text Reader

Abstract

A method for semi-automatically determining the path of at least one agricultural robot working in a given plot. The present invention relates to a path to be performed by at least one agricultural robot (1) to work in a given plot (P) and a method for semi-automatically determining the path. The present invention method comprises calculating a path parallel to an initial reference direction (OR1), determining one or more non-optimal zones (Z2), defining another reference direction (ORi), calculating a path (Ti) parallel to each of these, repeating these operations as necessary, and visualizing at least one proposed configuration of the planned operation for the entire plot.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of agricultural machinery, and more particularly, to a method of cultivating the soil or plants of agricultural land by at least one autonomously operating agricultural machine or implement (engin ou machine), i.e., an agricultural machine or implement that cultivates agricultural land without operator assistance except when necessary and preferably without control from a control center. This type of machine is generally called an agricultural robot and has at least one tool (integrated, mounted, semi-mounted or towed), integrated means (moyens integres) for control, communication, tracking, etc., and is a mobile, powered agricultural machine or implement that can autonomously cultivate agricultural land in cooperation with at least one other machine that may or may not perform the same type of work as needed.

[0002] In this regard, the present invention relates to a method for semi-automatically determining the passages and trajectories that at least one agricultural robot should follow to work across a given section.

Background Art

[0003] This type of agricultural robot can work alone to work on a section or as part of a fleet of at least two robots assigned to that section. This robot can operate completely independently after being set and programmed (either directly by the on-site user or remotely by a central control management system) before starting work in the field. Each robot in the fleet (flotte) may or may not communicate with each other and is assigned to work on a given part of the section as needed.

[0004] Typically, farm work is performed by moving back and forth within a single plot of land. In agricultural robotics, the machine's path is often generated in the form of a pre-calculated planned path or passage (trajets ou passages prevus precaleules). This path is generally a straight line parallel to the edge of the field (called the "reference edge") (at least within the cultivated area of ​​the land), and is chosen to be straight whenever possible, but does not necessarily have to be straight. These straight lines (paths or passages) exist at intervals equal to the working width of the agricultural robot's tool in question.

[0005] This solution is easy to implement and suitable for farmland of basic geometric shapes (e.g., rectangles, squares, quadrilaterals) or farmland without concave or non-linear edges (bords). However, when the plots have more complex shapes, include obstacles, and / or have certain relief or hazard zones (as is the case in many real-world situations), or when all paths are in a single direction or parallel to a single reference edge (see, for example, Figure 1A), the paths are not necessarily optimized (in terms of number, length, safety during operation, unworked movement, etc.).

[0006] [Patent Document 1] (U.S. Patent No. 10,459,447) discloses a path planning method for an agricultural robot to create a windrow, which includes defining one or more partitions in a single plot, determining windrow path sets by multiple angle increments, calculating the difference between the longest and shortest path lengths for each set, and retaining the set of paths with the smallest difference. A method for determining a vehicle's planned route disclosed in [Patent Document 2] (European Patent No. EP2,446,725) involves defining a section using line segments and concave connecting nodes, identifying the concave surface of the section, subdividing the section into sectors using lines associated with the nodes, determining the reference direction of the route, and defining work routes within sectors and interconnecting routes between sectors.

[0007] These known methods are systematically and fully automated, relying on complex methodologies for dividing the relevant sections, and cannot be partially executed at the user's discretion. There is no final verification method by the user, nor can the user choose one. [Prior art documents] [Patent Documents]

[0008] [Patent Document 1] U.S. Patent No. 10,459,447 [Patent Document 2] European Patent No. EP2,446,725 [Overview of the project] [Problems that the invention aims to solve]

[0009] The object of the present invention is to overcome at least the major drawbacks of the conventional methods described above. [Means for solving the problem]

[0010] For the purposes described above, the present invention provides a method for semi-automatically determining the paths and routes that at least one agricultural robot should traverse (execute) to work through a given area, comprising the following steps: a) Automatically or by the user, define an initial reference direction (orientation de reference initiale) according to a reference edge (un bord de reference) of the section, b) Calculate paths parallel to the initial reference direction above so as to cover the entire area, c) Automatically detect one or more zones within the above section called non-optimal zones (zones(s) non-optimale(s)) (where i varies from 2 to n, and n is a value of 2 or greater), in which case the evaluation of the planned paths (l'evaluation des trajets prevus) in the non-optimal zones according to the initial reference direction and the resulting predicted behavior (configuration resultante des deplacements projetes) for at least one predefined performance criterion (un crit·re de performance) for paths and / or operations does not exceed a threshold or reach a predetermined optimal value. d) Automatically define at least one other reference orientation for each non-optimal zone, e) For each non-optimal zone, calculate a path parallel to at least one other reference direction defined for the non-optimal zone so as to cover the entire non-optimal zone, f) If necessary, define a different reference direction each time, and repeat steps d) and e) for a given non-optimal zone until at least one path is obtained for the non-optimal zone and the evaluation against at least one performance criterion reaches an optimal value. g) Visualize the above area with at least one configuration of predicted deplacement projects for agricultural robots in different zones, each having its own path direction, and cover the entire area with the above different zones (i varies from 1 to n).

[0011] The present invention will be described below with reference to the attached schematic diagrams and preferred embodiments shown as non-limiting examples. The present invention will be better understood from the following description. [Brief explanation of the drawing]

[0012] [Figure 1A]It is a conceptual plan view of a section in which the planned path of an agricultural robot is determined according to a single direction with a single reference edge as a reference.

[0013] [Figure 1B] and [Figure 1C] is a diagram showing a path determined according to two reference edges or directions in the same section as [Figure 1A] by the method of the present invention.

[0014] [Figure 2]

[0015] [Figure 3] and

[0016] [Figure 4] are conceptual plan views of two other sections in which the planned path of the agricultural robot is determined according to at least three reference edges or directions by the method of the present invention.

Embodiments for Carrying Out the Invention

[0017] [Figure 1B], [Figure 1C] and [Figure 2] to [Figure 4] show a method for semi-automatically determining the passages and paths (T) to be executed by at least one agricultural robot (1) to work (cultivate) the entire predetermined section (P) according to the present invention when applied to sections (P) having different geometric shapes and topographies.

[0018] The method of the present invention essentially includes the following steps: a) Automatically or by the user (U), preferably according to the reference edge (bord de reference) (BR1) of the target section (P), define an initial reference direction (orientation de reference) (OR1). b) Calculate a path (T1) parallel to the initial reference direction (OR1) so as to cover the entire section (P). c) Automatically detect one or more zones (Zi) in the above section (P) called non-optimal zones (Zi) (where i varies from 2 to n and n takes a value of 2 or greater), in which case the evaluation of the planned route (T1) according to the initial reference direction (OR1) and the resulting planned movement configuration (configuration) against at least one predefined performance criterion (un crit·re de performance) for the route and / or work does not exceed a threshold or reach a predetermined optimal value. d) Automatically define at least one other reference direction (ORi) for each zone (Z2) or (Z2 to Zn) of the above non-optimal zone (Zi), e) For each non-optimal zone (Zi), calculate a path (Ti) parallel to at least one other reference direction (ORi) defined in the non-optimal zone (Zi) in question, so as to cover the entire non-optimal zone (Zi) each time. f) If necessary, repeat steps d) and e) for a given non-optimal zone (Zi), defining a different reference direction (ORi) each time, and repeating until at least one path proposal (Ti) is obtained for that non-optimal zone (Zi) and the evaluation of that path proposal against at least one performance criterion (critere de performance) reaches an optimal value. g) Visualize the above parcel (P) with a proposed configuration of at least one configuration of predictive behavior of an agricultural robot (1) in different zones (Zi) (i varies from 1 to n) where each zone has its own path (Ti) direction (ORi), covering the entire parcel with the above different zones (i varies from 1 to n).

[0019] The present invention makes it possible to determine the optimal configuration of the path for performing work on the entire section (P) under consideration by combining the features identified above. The initial direction (OR1) can be automatically determined by appropriate software (particularly according to mapping or geometric analysis of the section, e.g., OR1 = the direction of the longest straight line that can be drawn in the section) or can be manually defined by the user (U). When defining this initial direction, knowledge gained during previous work campaigns can also be taken into consideration.

[0020] Furthermore, by using the above progressive guided approach to determine the direction and work path of the non-optimal zone, it becomes possible to explore and evaluate almost all possible options for the proposed layout configuration (of course, the T1 path is retained as the optimal path configuration for the rest of the partition (=partition - non-optimal zone)).

[0021] The same direction (ORi) can be applied to multiple unrelated different configuration zones within a partition (see Figure 4, where OR3 is applied to Z3 and Z4).

[0022] In a preferred embodiment, the visualization of proposed configurations for the planned operation of the agricultural robot (1) on a section (P) divided into n zones (Zi) is performed by an interactive graphical interface (2), characterized in that, after a proposal selected from multiple proposals is justified, it is subjected to verification by the user (U) via the agricultural robot's (1) central management system as needed before being sent to the agricultural robot (1). Thus, if necessary, the user can regain control after the initial direction is defined (step a) and determine the configuration to be executed.

[0023] The definition of the initial reference direction (OR1) is preferable to be performed using at least one evaluation criterion selected from the following, after considering the topographic and geometric characteristics of the parcel (P), especially when it is done automatically: A predetermined number of predicted paths, for example, the minimum number of predicted paths (T1); A predetermined number, for example, the minimum number of U-turns; A continuous, edge-to-edge path (T) that is as long as possible and, where appropriate, straight; At least one statistical parameter optimized for the population length of the predicted path (T1), e.g., variance, mean, and / or median; However, if at least two criteria (crit·res) are used, a weighting of the criteria (ponderation) can be applied.

[0024] Preferably, with respect to the initial path (T1) prediction, the evaluation criteria used to detect zones (Zi) where the likelihood of efficient movement and / or work is sufficiently low are selected from the following: The topography and / or shape of the parcel (P); The average value of the largest possible path lengths; The fewest possible number of U-turns; At least one statistical parameter optimized for the population of the predicted path length (T1), e.g., variance, mean, and / or median; However, if at least two criteria are used, a weighting of the criteria can be applied.

[0025] When using multiple evaluation criteria, it is possible to assign appropriate weights to prioritize one or more criteria that are considered more important than others.

[0026] Planned paths (Ti) in two adjacent zones (Zi) tend to overlap or intersect at the junction (RJ) region between these two zones. To avoid this interference, the configuration of paths in these regions can be planned to be managed automatically or through user intervention.

[0027] That is, as shown in Figures 1C and 2-4, the present invention method involves defining a boundary (L) in a junction region (RJ) between two adjacent zones (Zi) among n different zones (Zi) covering the entire surface of a partition (P), with a preference, as a criterion, for the minimum number of paths (Ti) covering the entire partition (P) (where i varies from 1 to n) and / or the minimum number of U-turns within the partition (P).

[0028] Preferably, each boundary (L) in the junction region (RJ) between two adjacent zones (Zi) (where i varies from 1 to n) is defined by connecting two opposing points (P1, P2) on the periphery surrounding a partition (P), or by extending a path (Ti) of one zone (Zi) from its outer edge to the intersection with a longer path (Ti) of another zone (Zi). This boundary definition is performed automatically or manually by the user (U) using an interactive graphical interface (2).

[0029] The boundary (L) may or may not be a straight line, and may consist, for example, of at least part or all of a planned boundary path (Ti) in one of two adjacent zones (Zi) that actually exist. However, this boundary (L) may also be of a hypothetical nature, defining a virtual boundary between two adjacent regions (Figure 4).

[0030] To better consider the actual field conditions, particularly factors that hinder the uniformity and continuity of work in the work area, steps a), b), d), and e) may be planned to take into account one or more obstacles (O) and / or one or more hazardous areas (ZR) that may be present within the area (P), and / or the direction of movement planned during subsequent farm work, based on extended map data of the area (P), when defining the initial reference orientation (OR1) and at least one other possible reference orientation (ORi), and when calculating the subsequent path (T1, Ti) (see Figures 2, 3, and 4).

[0031] Obstacles (O) can themselves determine the reference orientation or edge for calculating the path, particularly depending on their shape. Furthermore, in the case of a hazardous area (ZR), the orientation specific to this area (ORR) can be determined, and measures can be taken to generate a path (TRR) that minimizes the risk that may occur (Figure 4).

[0032] As a supplement or alternative, and to account for the undulations of the work area, three-dimensional terrain data of the area under consideration (P) can be taken to identify zones (ZDj) that may be on a slope and whose slope is greater than a predetermined threshold, and measures can be taken to impose a reference orientation (ODj) for the direction (j≧0) of the path (TDj) calculated for each agricultural robot (1) within each zone (ZDj) (see Figure 3).

[0033] To enable the user (U) to make a final selection from the submitted proposals, and in particular to inform the user of safety risks, the user (U) may be notified, if necessary, if one or at least one of the proposed configurations of the planned operation for the agricultural robot (1) is dangerous due to slope or other identified risks, or is not optimal for subsequent agricultural work.

[0034] As illustrated in Figure 3, the present invention may include generating paths (TC) parallel to the edge portion (PC) and defining a zone (ZC) when at least one portion (PC) is a non-linear partition edge (P). The extension of each path (TC) is limited by its intersection with the edge of the partition (P) and / or a straight path (Ti).

[0035] More generally, if a reference direction (OR1) and / or another later reference direction (ORi) are associated with a nonlinear reference edge (BR1, BRi), the calculated path (T1, Ti) is parallel to the edge path and therefore calculated according to the nonlinear trace.

[0036] The present invention further relates to a method for processing a plot (P) using at least one agricultural robot (1), characterized by including a pre-parameterization and programming step, which includes the aforementioned at least one method for semi-automatically determining the passages and paths (T) to be executed.

[0037] Following automatic or semi-automatic definition and final selection by the user, the selected route configuration proposal (Ti) for implementation is sent to the robot (1), which then performs the programmed work on the section (P). The calculation operations and configuration transmission are performed, for example, by a monitoring system or a portable end-end available to the user.

[0038] The actual work on the section thereafter can be carried out, for example, according to the methods described in the applicant's Patent Documents 3 to 7 below. [Patent Document 3] French Patent No. FR3119508 [Patent Document 4] French Patent No. FR3114218 [Patent Document 5] French Patent No. FR3114217 [Patent Document 6] French Patent No. FR3119507 [Patent Document 7] French Patent No. FR3122063

[0039] As a practical example of the present invention, the following flow is conceivable: The user connects to a "web" application via a computer / tablet / smartphone. Next, the user accesses a list of existing parcels. New parcels can be added as needed. The external contours of the parcels are obtained (via data measured in the field) or drawn by hand. Subsequently, each step of the present invention described above is performed. The result of this process is a file containing the user-generated shapes (line structures). This file can be distributed to a robot, which executes this file via monitoring software that generates control commands for the robot.

[0040] Existing software available for use in this invention includes known software (e.g., AGCO's "Geo-Bird" or Lacos Computer Service's "FieldPlanner") that, either directly or by adapting its implementation to the needs of the invention, imports and / or plots maps of plots or fields from satellite imagery, generates optimized parallel guidance lines, and exports them to edge ends for later use by an autonomous guidance tractor, in order to perform specific tasks or operations within the framework of the practical implementation of the method described above. In particular, these software programs need to be adapted or implemented according to multiple reference edges or directions. Also available is software that enables the generation of a series of parallel lines along a reference line (e.g., Terminal CCI's CCI.Command and Parallel Tracking functions).

[0041] Of course, the present invention is not limited to the embodiments described and illustrated in the accompanying drawings. Various modifications are possible without departing from the scope of protection of the present invention, particularly the configuration of each element and the substitution of technical equivalents.

Claims

1. A method for semi-automatically determining pathways and routes (T) for at least one agricultural robot (1) to work across a given area (P), comprising the following steps: a) The initial reference direction (OR1) is defined automatically or by the user (U), preferably according to the reference edge (BR1) of the section (P), b) Calculate a path (T1) parallel to the initial reference direction (OR1) above, so as to cover the entire area (P), c) Automatically detect one or more zones (Zi) of a section (P) called a non-optimal zone (Zi) (where i varies from 2 to n, and n takes a value of 2 or greater), in which case the evaluation of the planned path (T1) according to the initial reference direction (OR1) and the resulting predicted operation configuration against at least one performance criterion predefined for the above path and / or operation does not exceed a threshold or reach a predetermined optimal value. d) Automatically define at least one other reference direction (Ori) for the above zone (Z2) or each (Z2 to Zn) of the non-optimal zone (Zi), e) For the target non-optimal zone (Zi), calculate a path (Ti) parallel to at least one other reference direction (ORi) defined in that non-optimal zone (Zi), such that it covers the entire non-optimal zone (Zi) each time. f) If necessary, define a different reference direction (ORi) each time for a specific non-optimal zone (Zi), and repeat steps d) and e) until at least one path proposal (Ti) is obtained for that non-optimal zone (Zi) and the evaluation with respect to at least one performance criterion reaches an optimal value. g) Visualize the parcel (P) in at least one proposed path for the configuration of the predicted movement of the agricultural robot (1) in each of the distinct zones (Zi) (where i varies from 1 to n) as a whole covers the entire surface of the parcel (P) 9.

2. The method according to claim 1, characterized in that the visualization of the above-mentioned predicted movement configurations proposed for an agricultural robot (1) in a section (P) divided into n zones (Zi) is performed using an interactive graphical interface (interface graphique interactive) (2), and the proposal selected from among several based on valid reasons is subjected to validation by a user (U) via the agricultural robot's centralized management system as necessary before being sent to the agricultural robot (1).

3. The method according to 1 or 2, characterized in that the initial reference direction (OR1) is defined using at least one evaluation criterion selected from the following, taking into account the topographic and geometric characteristics of the parcel (P): A predetermined number, for example, a minimum number of predicted paths (T1); A predetermined number, for example, the minimum number of U-turns; A continuous, edge-to-edge path (T) that is as long as possible and, if necessary, straight; At least one statistical parameter optimized for the population of the predicted path length (T1), e.g., variance, mean, and / or median; Here, if at least two criteria are used, a weighting of the criteria can be applied.

4. The method according to any one of claims 1 to 3, characterized in that the evaluation criteria used to detect zones (Zi) that are unlikely to be moved and / or worked on sufficiently efficiently with respect to the initial path prediction (T1) are selected from the following: The topography and / or shape of the parcel (P); The average value of the longest possible path length; The fewest number of U-turns possible; At least one statistical parameter optimized for the population of predicted path length (T1), e.g., variance, mean, and / or median; Here, if at least two criteria are used, a weighting of the criteria can be applied.

5. The method according to any one of claims 1 to 4, comprising the step of defining a boundary (L) of a junction region (RJ) between two adjacent zones (RJ) in n distinct zones (Zi) covering the entire surface of a partition (P), characterized in that, as criteria, the minimum number of paths (Ti) and / or the minimum number of U-turns in each of the n zones (Zi) (where i varies from 1 to n) covering the entire partition (P) is preferred.

6. The method according to 5, characterized in that each limit (L) in the junction region (RJ) between two adjacent zones (Zi) (where i varies from 1 to n) is defined by connecting two opposite points (P1, P2) on the periphery surrounding a partition (P), or by extending a path (Ti) of a zone (Zi) from the periphery to the intersection with another longer path (Ti) of another zone (Zi), wherein the definition of this limit is performed automatically or manually by a user (U) using an interactive graphical interface (2).

7. The method according to any one of claims 1 to 6, characterized in that in steps a), b), d), and e), at least one other possible reference direction (ORi) and subsequent route (T1, Ti), one or more obstacles (O) that may be present in the parcel (P), and / or one or more hazardous areas (ZR) that may be present in the parcel (P), and / or directions of travel planned during subsequent agricultural work, in order to define an initial reference direction (OR1) based on extended map data of the parcel (P).

8. The method according to any one of claims 1 to 7, characterized in that, taking into consideration the three-dimensional terrain data of the target area (P), zones (ZDj) that may be on a slope and whose slope is greater than a predetermined threshold are identified, and a reference direction (ODj) is given to the direction of the path (TDj) calculated for each agricultural robot (1) in each of these zones (ZDj) (j≧0).

9. The method according to any one of claims 1 to 8, characterized in that, where appropriate, the user (U) is notified that at least one of the proposed configurations of the planned movements for the agricultural robot (1) is dangerous due to incline or other identified risks or is not optimal for subsequent agricultural work.

10. The method according to any one of claims 1 to 9, characterized in that a zone (ZC) is defined by generating a path (TC) parallel to at least one non-linear edge portion (PC) of a section (P), wherein the extension of each path (TC) is limited by intersection with the edge and / or linear path (Ti) of the section (P).

11. A method for working in a plot (P) using at least one agricultural robot (1), comprising a parameter pre-setting and programming step, comprising at least one method for semi-automatically determining a path (T) to be executed according to the method of any one of claims 1 to 10.

Citation Information

Patent Citations

  • US10,459,447

  • EP2,446,725