An unmanned agricultural machine obstacle-avoiding operation path optimization method and device in an agricultural field scene

Abstract

The application provides an unmanned agricultural machine obstacle-avoiding operation path optimization method and device in a farmland scene, relates to the fields of automatic driving and path planning, and comprises the following steps: finding a collision risk strip by detecting whether an intersection exists between an unmanned agricultural machine operation strip in a farmland and an obstacle inflation range; based on key parameters such as the size of an agricultural machine and a minimum turning radius, the minimum influence range of repulsive force of an obstacle is constrained to ensure that the obstacle-avoiding path meets the actual obstacle-avoiding requirements of the agricultural machine and the mounted machine tool in space; local obstacle-avoiding path planning is carried out based on the artificial potential field method, the distance relationship between a sampling point on the operation strip and the mass center of the obstacle is introduced, and the deflection angle of the repulsive force direction is adjusted in combination with the influence radius of the obstacle, so that the potential field distribution is dynamically optimized. The application can take into account the operation requirements, path smoothness, path continuity and operation efficiency, and improve the application value of the unmanned agricultural machine in a complex environment in a farmland.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automatic driving and path planning, in particular to a method and device for optimizing the obstacle-avoiding path of unmanned agricultural machinery in a farmland scene. BACKGROUND

[0002] With the continuous development of agricultural machinery automatic driving technology and intelligent implements, unmanned agricultural machinery gradually realizes the whole process of plowing, management and harvesting in farmland, greatly improves the efficiency and quality of agricultural production, and significantly reduces labor costs and resource consumption.

[0003] Unmanned agricultural machinery obstacle-avoiding path planning is an important research direction in the field of intelligent agriculture technology, and its research and application value is increasingly prominent with the rise of large-scale unmanned farms and the continuous improvement of agricultural mechanization and intelligence. In the vast working land of large-scale unmanned farms, there are often static obstacles such as water wells and base stations. Therefore, how to realize safe and efficient obstacle-avoiding path planning has become one of the key technologies to ensure the smooth operation of unmanned agricultural machinery. In the path planning process, the physical size and dynamic characteristics of the agricultural machinery and the implements carried must be fully considered, and the path efficiency, environmental constraints and driving safety must be optimized comprehensively to ensure the continuity and reliability of the agricultural machinery in the automatic driving process and improve its operation performance in complex farmland environment.

[0004] Under the background of pursuing economic benefits, the intelligent terminal of unmanned agricultural equipment usually uses a processor with low computing power, which is difficult to process complex image or laser scanning data in real time to support real-time obstacle avoidance schemes based on laser perception or camera. In addition, the non-deterministic performance of crops in the natural environment, such as sway caused by wind and changes in morphology during growth, brings additional challenges to obstacle identification and classification. These factors make it difficult to directly apply real-time obstacle avoidance technology to agricultural scenarios.

[0005] Under this background, prior path planning based on prior known environmental map and obstacle information can greatly reduce the computational burden. Traditional obstacle-avoiding path planning technology takes "safe obstacle avoidance" as the main goal, but in the application scenario of farmland, unmanned agricultural machinery not only needs to realize obstacle avoidance, but also needs to consider operation efficiency and operation area coverage. In addition, agricultural machinery equipped with different operation implements has a series of dynamic constraints, such as turning radius and physical size of operation implements, etc., to avoid damage to the machine or affect the operation effect due to sharp turning, which requires better curvature continuity and smoothness of the obstacle-avoiding trajectory, so the artificial potential field method is adopted for path optimization.

[0006] However, the existing path optimization method based on artificial potential field method still has the problems of path falling into local minimum value, insufficient path smoothness, insufficient consideration of coverage rate and operation efficiency in the complex agricultural operation scene. SUMMARY

[0007] To solve the above technical problems, the present application provides a method and device for optimizing the obstacle-avoiding operation path of unmanned agricultural machinery in a farmland scene, which further optimizes the full-coverage operation path planning scheme of unmanned agricultural machinery in a farmland scene. The present application can obtain high-precision farmland information and obstacle information in the field in advance through remote sensing technology. First, the full-coverage operation planning of the farmland is carried out based on the farmland information and the parameters of the agricultural machinery, and a series of operation strips are obtained. Then, the size of the obstacles is expanded using the size parameters of the agricultural machinery, and then it is detected whether each operation strip has a collision risk with the obstacles in the field. The strips with collision risks are locally optimized, and then the full-coverage operation path of the agricultural machinery is corrected. The present application can balance the operation demand, path smoothness, path continuity and operation efficiency, and improve the application value of unmanned agricultural machinery in complex environments in the field.

[0008] To achieve the above purpose, the present application adopts the following technical scheme:

[0009] A method for optimizing the obstacle-avoiding operation path of unmanned agricultural machinery in a farmland scene, comprising the following steps:

[0010] Step 1: Collecting the farmland map and obstacle information in the farmland, and obtaining the basic full-coverage operation path for the operation path in the field;

[0011] Step 2: Finding the collision risk strips by detecting whether the unmanned agricultural machinery operation strips in the field and the expanded range of the obstacles have intersection points;

[0012] Step 3: Based on the size of the unmanned agricultural machinery and the minimum turning radius, the minimum influence range of the repulsive force of the obstacles is constrained to ensure that the obstacle-avoiding path meets the actual obstacle-avoiding requirements of the unmanned agricultural machinery and the mounted implements in space;

[0013] Step 4: Carrying out local obstacle-avoiding path planning based on the artificial potential field method, introducing the distance relationship between the sampling points on the operation strips and the center of mass of the obstacles, and adjusting the deflection angle of the repulsive force direction combined with the influence radius of the obstacles, so as to dynamically optimize the potential field distribution and realize effective obstacle avoidance.

[0014] The present application also provides a device for optimizing the obstacle-avoiding operation path of unmanned agricultural machinery in a farmland scene, comprising the following modules:

[0015] The acquisition module collects the farmland map and obstacle information in the farmland, and obtains the basic full-coverage operation path for the operation path in the field;

[0016] The detection module finds the collision risk strips by detecting whether the unmanned agricultural machinery operation strips in the field and the expanded range of the obstacles have intersection points;

[0017] The constraint module constrains the minimum influence range of the repulsive force of the obstacle based on the implement size of the unmanned agricultural machine and the minimum turning radius, and ensures that the obstacle-avoiding path meets the actual obstacle-avoiding requirements of the unmanned agricultural machine and the mounted implement in space.

[0018] The obstacle-avoiding module performs local obstacle-avoiding path planning based on the artificial potential field method, introduces the distance relationship between the sampling points on the operation strip and the obstacle center of mass, and adjusts the deflection angle of the repulsive force direction in combination with the influence radius of the obstacle, so as to dynamically optimize the potential field distribution and realize effective obstacle avoidance.

[0019] The application further provides an electronic device, including a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the above-mentioned unmanned agricultural machine obstacle-avoiding operation path optimization method in the farmland scene.

[0020] The application further provides a non-transitory computer-readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the above-mentioned unmanned agricultural machine obstacle-avoiding operation path optimization method in the farmland scene.

[0021] Advantages:

[0022] (1) The application determines the repulsive force influence range and the repulsive force coefficient of the obstacle in the artificial potential field method based on the curvature constraint of the agricultural machine, so as to ensure that the agricultural machine can realize smooth and safe obstacle avoidance.

[0023] (2) The application combines the repulsive force deflection angle mechanism and the attractive force of the path sampling point in the already operated area on the basis of the traditional artificial potential field method, effectively avoids the problem of falling into a local optimal solution in the path planning process, reduces the interference to the crops in the unoperated area, guarantees the operation continuity and reduces the influence on the unoperated area. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 The flowchart of the unmanned agricultural machine obstacle-avoiding operation path optimization method in the farmland scene of the application;

[0025] Figure 2 The principle diagram of the unmanned agricultural machine obstacle-avoiding operation path optimization method in the farmland scene of the application;

[0026] Figure 3 The final operation path schematic diagram;

[0027] Figure 4 The schematic diagram of the unmanned agricultural machine obstacle-avoiding operation path optimization device in the farmland scene of the application. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and not to limit the present application. In addition, the technical features involved in each embodiment of the present application described below can be combined with each other as long as they do not conflict with each other.

[0029] As shown in Figure 1 The present application provides a method for optimizing the obstacle-avoiding operation path of unmanned agricultural machinery in a farmland scene. The method optimizes the obstacle-avoiding operation of the unmanned agricultural machinery in advance, ensures the obstacle-avoiding operation, and greatly reduces the calculation amount of the intelligent terminal. The method includes the following steps:

[0030] Step 1: Information collection: The information of the farmland map and the obstacles in the farmland is collected in advance by remote sensing and other means. The basic full-coverage operation path is obtained by AB line operation and other methods for the in-field operation path, which prepares for the subsequent local path optimization in the obstacle-avoiding area.

[0031] Step 2: Intersection detection: The collision risk strip is found by detecting whether there is an intersection between the unmanned agricultural machinery operation strip and the obstacle expansion interval in the farmland.

[0032] Step 3: Constraint construction: Based on the key parameters such as the size of the agricultural machinery and the minimum turning radius, the minimum influence range of the repulsive force of the obstacle is constrained to ensure that the obstacle-avoiding path meets the actual obstacle-avoiding requirements of the agricultural machinery and the mounted implements in space, and to avoid affecting the safety and operation effect due to insufficient turning space or too close to the obstacle.

[0033] Step 4: Effective obstacle avoidance: The local obstacle-avoiding path planning is carried out based on the artificial potential field method (APF), the distance relationship between the sampling points on the operation strip and the obstacle centroid is introduced, and the repulsive force direction is adjusted by the deflection angle in combination with the influence radius of the obstacle, so as to dynamically optimize the potential field distribution and realize effective obstacle avoidance.

[0034] Specifically, the step 1 includes:

[0035] Data collection is performed on the target farmland area to obtain complete geographic information and obstacle distribution information. The position and size information of the static obstacles should be included in the farmland map. Based on the known farmland area and obstacle information, the traditional AB line operation method (i.e. the parallel operation trajectory generation method defined by two points A and B) is used to divide the path according to the agricultural operation requirements, combined with the shape of the farmland boundary and the terrain characteristics, to generate a basic full-coverage operation path covering the entire operation area. The basic full-coverage operation path does not consider the influence of obstacles and is used as a reference for the ideal operation path for subsequent path adjustment and local optimization in the obstacle area.

[0036] Specifically, the step 2 comprises:

[0037] The common obstacles in farmland are further constructed into two types: circular-like obstacles and convex polygonal obstacles. The two types of obstacles are inflated based on the implement size and minimum turning radius of the agricultural machine. The collision risk strip is the working strip intersecting with the space after the obstacle inflation.

[0038] To meet the turning constraints of the agricultural machine, the obstacle inflation width L is expressed as:

[0039] (1)

[0040] wherein, is the maximum curvature when the unmanned agricultural machine turns, is the working width of the agricultural machine with an implement, and max represents the maximum value. Whether the working strip belongs to the collision risk strip is determined by detecting whether the sampling points on the working strip are located in the obstacle collision interval.

[0041] Specifically, the step 3 comprises:

[0042] The artificial potential field method with increased detour rules is used to develop local path optimization for the collision risk strip. The artificial potential field method is to obtain the resultant force of the attractive force of the target point to the current point of the agricultural machine and the repulsive force of multiple obstacles to the current point, so as to obtain the resultant force vector. The attractive potential field function and the repulsive potential field function of the conventional distance-based artificial potential field method can be expressed as:

[0043] (2)

[0044] (3)

[0045] (4)

[0046] wherein, are the attractive gain coefficient and the repulsive gain coefficient, respectively; is a vector, representing the Euclidean distance between the current point of the agricultural machine and the target point of the agricultural machine, which is obtained by formula (3), , are the longitudinal coordinate and the transverse coordinate of the target point, , are the longitudinal coordinate and the transverse coordinate of the current point, and the vector direction is from the current point to the target point; is the obstacle influence radius, is the Euclidean distance between the current point and the obstacle.

[0047] To meet the needs of agricultural machinery steering, the obstacle influence radius Can be defined as:

[0048] (5)

[0049] The target point to the current point of the attractive force And the repulsive force of the obstacle Respectively, the negative gradient of the attractive potential field function and the repulsive potential field function, the expression is:

[0050] (6)

[0051] (7)

[0052] In formula (7) Indicates the partial derivative of repulsion.

[0053] Specifically, the step 4 comprises:

[0054] The artificial potential field method may have a situation where the combined force of the attractive force and the repulsive force is zero, resulting in a local minimum value to stop the path point exploration or infinite circulation at a certain place. The present application introduces the distance relationship between the sampling point on the work strip and the obstacle centroid, and adjusts the deflection angle of the repulsive force direction in combination with the obstacle influence radius, thereby dynamically optimizing the potential field distribution and avoiding the problem of falling into a local minimum value.

[0055] Assume that the point on the strip (i.e. the current point) , the target point And the obstacle centroid coordinates ( Indicate the x direction and y direction coordinates of the obstacle centroid) are known, the slope expressions of the current point and the target point and the obstacle centroid and the target point are respectively:

[0056] (8)

[0057] (9)

[0058] Wherein, , Respectively represent the slope of the line connecting the target point and the obstacle, and the slope obtained by connecting the target point and the strip sampling point (the current point).

[0059] When , it is realized from the right side of the local obstacle. At this time The x-axis component force , the y-axis component force Are:

[0060] (10)

[0061] (11)

[0062] (12)

[0063] wherein, represents the deflection angle, is the right positive proportional angle coefficient, is the maximum distance of the current agricultural machine not affected by the obstacle.

[0064] Similarly, when , the local obstacle is realized from the left side.

[0065] At this time the component force on the x-axis , the component force on the y-axis is:

[0066] (13)

[0067] (14)

[0068] wherein, is the left positive proportional angle coefficient.

[0069] As shown in Figure 2 , given the obstacle coordinates in the farmland and the information of each work strip in the field, the irregular obstacle is first pre-processed and inflated according to its maximum circumscribed circle to simplify the obstacle collision detection; according to formula (1), the collision inflation is carried out according to the working width of the agricultural machine and the turning radius to realize the optimization of the obstacle. It is judged whether the work strip collides with the obstacle collision area. If there is an intersection point, it means that there is a collision risk between the work strip and the obstacle in the field. For the collision risk strip, the local path optimization (APF local obstacle path optimization) based on artificial potential field method is carried out according to steps 3 and 4.

[0070] The minimum turning radius of the unmanned agricultural machine is 5m, and the working width of the mounted implement is 2.93m. For a piece of farmland for unmanned operation in the farmland, in the Universal Transverse Mercator projection coordinate system under the World Geodetic System 1984 standard, the vertex coordinates are: [[571532.36, 4089678], [571492.36, 4089679], [571525.61, 4089406], [571485.61, 4089406]]; There is an obstacle, the center coordinates of which are [571504, 4089575], and the radius of which is 3m.

[0071] Through the collision detection of the work strips by the method, 4 work strips with collision risks are identified. Further, the local path planning algorithm is used to reconstruct the path in the collision risk area, and the final work path is generated, as shown in Figure 3 Figure 3 In the formula, x is the east-west direction coordinate of the Universal Transverse Mercator projection under the World Geodetic System 1984 standard, and y is the south-north direction coordinate of the Universal Transverse Mercator projection under the World Geodetic System 1984 standard.

[0072] Through actual work verification, the obstacle-avoiding path generated by the method can effectively avoid the collision between the agricultural machine and the obstacle under the premise of ensuring the efficiency of the agricultural machine, and has good practicability and robustness.

[0073] As shown in Figure 4 The application further provides an unmanned agricultural machine obstacle-avoiding work path optimization device in a farmland scene, which is used to implement the above method and comprises the following modules:

[0074] The acquisition module is used to complete the acquisition of the farmland map and the obstacle information in the farmland, and obtain a basic full-coverage work path for the work path in the field.

[0075] The detection module is used to find the collision risk strip by detecting whether there is an intersection between the unmanned agricultural machine work strip and the obstacle expansion range in the farmland.

[0076] The constraint module is used to constrain the minimum influence range of the repulsive force of the obstacle based on the size of the machine tool of the unmanned agricultural machine and the minimum turning radius, so as to ensure that the obstacle-avoiding path meets the actual obstacle-avoiding requirements of the unmanned agricultural machine and the machine tool mounted thereon in space.

[0077] The obstacle-avoiding module is used to carry out local obstacle-avoiding path planning based on the artificial potential field method, introduce the distance relationship between the sampling points on the work strip and the obstacle center of mass, and adjust the deflection angle of the repulsive force direction in combination with the influence radius of the obstacle, so as to dynamically optimize the potential field distribution and realize effective obstacle avoidance.

[0078] The application further provides an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the steps of the above-mentioned unmanned agricultural machine obstacle-avoiding work path optimization method when executing the program.

[0079] The application further provides a non-transitory computer readable storage medium, which stores a computer program, and the computer program implements the steps of the above-mentioned unmanned agricultural machine obstacle-avoiding work path optimization method when executed by a processor.

[0080] ​Those skilled in the art will appreciate that embodiments of the present application can be devised for a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer-readable program code thereon for use by or in connection with an instruction execution system. The program code can comprise one or more instructions enabling a computer system to carry out one or more of the methods described herein. The program code can be in any form suitable for use in the implementation and adapted for the computer system where the application is practiced.

Claims

1. A method for optimizing the path of an unmanned agricultural machine's obstacle avoidance operation in a farmland scenario, characterized in that: The steps include: Step 1: Complete the collection of farm maps and obstacle information within the farm, and determine the basic full-coverage operation path for the field operation path; Step 2: Find the collision risk strip by detecting whether there is an intersection between the unmanned agricultural machinery operation strip in the farmland and the obstacle expansion range; Step 3: Based on the size and minimum turning radius of the UAV, constrain the minimum impact range of the obstacle repulsion force to ensure that the obstacle avoidance path spatially meets the actual obstacle avoidance requirements of the UAV and its mounted implements. The artificial potential field method with added detour rules is used to optimize the local path of the collision risk strip. The artificial potential field method uses the distance between the current point of the agricultural machine and the obstacle to calculate the resultant force of the target point on the agricultural machine and the repulsive force of multiple obstacles on this point, thereby obtaining the resultant force vector. The gravitational potential field function of the distance-based artificial potential field method is and the repulsive potential field function Expressed as: （2） （3） （4） in, are the attraction gain coefficient and the repulsion gain coefficient respectively; Is a vector, indicating the current point of the agricultural machinery Target point with agricultural machinery The Euclidean distance of is obtained by formula (3): 、 are the ordinate and abscissa of the target point, 、 The ordinate and abscissa of the current point, and the direction of the vector is from the current point to the target point; is the obstacle influence radius, is the Euclidean distance between the current point and the obstacle; To meet the steering requirements of agricultural machinery, the obstacle impact radius here is Defined as: （5） Where L is the expansion width of the obstacle, is the maximum curvature of the unmanned agricultural machinery when turning, The working width of agricultural machinery for mounting implements; The gravitational force of the target point on the current point Repulsion from obstacles are the negative gradients of the gravitational potential field function and the repulsive potential field function, respectively, and their expressions are: （6） （7） In formula (7) represents the partial derivative of the repulsive force; Step 4: Conduct local obstacle avoidance path planning based on the artificial potential field method. Introduce the distance relationship between the sampling point on the operation strip and the center of mass of the obstacle. Combined with the obstacle influence radius, adjust the deflection angle of the repulsive force direction to dynamically optimize the potential field distribution and achieve effective obstacle avoidance.

2. The method for optimizing the path of an unmanned agricultural machine avoiding obstacles in a farmland scene according to claim 1 is characterized in that: In step 2, the obstacles in the farmland are classified into quasi-circular obstacles and convex polygonal obstacles, and the obstacles in the farmland are expanded based on the machine size and minimum turning radius of the unmanned agricultural machinery to determine the collision risk strip.

3. The method for optimizing the path of an unmanned agricultural machine avoiding obstacles in a farmland scene according to claim 1 is characterized in that: In step 2, whether the operation strip belongs to a collision risk strip is determined by detecting whether the sampling point on the operation strip is located in the obstacle collision interval, thereby determining the operation strip that requires local obstacle avoidance path optimization.

4. The method for optimizing the path of an unmanned agricultural machine avoiding obstacles in a farmland scene according to claim 3 is characterized in that: In step 4, when the local obstacle avoidance is achieved from the right side of the unmanned agricultural machine, the components of the obstacle repulsion force on the x and y axes are determined.

5. The method for optimizing the path of an unmanned agricultural machine's obstacle avoidance operation in a farmland scene according to claim 3 is characterized in that: In step 4, when the local obstacle avoidance is achieved from the left side, the components of the obstacle repulsive force on the x and y axes are determined.

6. A device for optimizing the path of an unmanned agricultural machine's obstacle avoidance operation in a farmland scene, characterized in that: Includes the following modules: The acquisition module collects farm maps and obstacle information within the farm, and determines the basic full-coverage operation path for the field operation path; The detection module locates the collision risk zone by detecting whether there is an intersection between the unmanned agricultural machinery operation zone and the obstacle expansion range in the farmland; The constraint module constrains the minimum impact range of the obstacle repulsion force based on the size and minimum turning radius of the unmanned agricultural machinery, ensuring that the obstacle avoidance path spatially meets the actual obstacle avoidance requirements of the unmanned agricultural machinery and its mounted machinery. The artificial potential field method with added detour rules is used to optimize the local path of the collision risk strip. The artificial potential field method uses the distance between the current point of the agricultural machine and the obstacle to calculate the resultant force of the target point on the agricultural machine and the repulsive force of multiple obstacles on this point, thereby obtaining the resultant force vector. The gravitational potential field function of the distance-based artificial potential field method is and the repulsive potential field function Expressed as: （2） （3） （4） in, are the attraction gain coefficient and the repulsion gain coefficient respectively; Is a vector, indicating the current point of the agricultural machinery Target point with agricultural machinery The Euclidean distance of is obtained by formula (3): 、 are the ordinate and abscissa of the target point, 、 The ordinate and abscissa of the current point, and the direction of the vector is from the current point to the target point; is the obstacle influence radius, is the Euclidean distance between the current point and the obstacle; To meet the steering requirements of agricultural machinery, the obstacle impact radius here is Defined as: （5） Where L is the expansion width of the obstacle, is the maximum curvature of the unmanned agricultural machinery when turning, The working width of agricultural machinery for mounting implements; The gravitational force of the target point on the current point Repulsion from obstacles are the negative gradients of the gravitational potential field function and the repulsive potential field function, respectively, and their expressions are: （6） （7） In formula (7) represents the partial derivative of the repulsive force; The obstacle avoidance module conducts local obstacle avoidance path planning based on the artificial potential field method. It introduces the distance relationship between the sampling point on the operation strip and the center of mass of the obstacle, and adjusts the deflection angle of the repulsive force direction based on the obstacle influence radius, thereby dynamically optimizing the potential field distribution and achieving effective obstacle avoidance.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method for optimizing the obstacle avoidance operation path of an unmanned agricultural machinery in a farmland scenario as described in one of claims 1-5 are implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a method for optimizing the obstacle avoidance operation path of an unmanned agricultural machinery in a farmland scenario as described in one of claims 1 to 5 are implemented.

Images