Robot path planning method, system and device, medium and program product

Abstract

The invention provides a path planning method, system and device of a robot, a medium and a program product. The path planning method comprises the following steps: acquiring a first starting point and a first ending point marked in a grid map; based on the grid map, the first starting point, the first ending point and an ant colony algorithm, a first advancing path of the robot is obtained, and the first advancing path and the grid map are input into a target vector machine model; and in response to an output result of the target vector machine model that there is no barrier on the first travel path, planning the first travel path as a target travel path of the robot. A first advancing path from a first starting point to a first ending point can be found in a complex grid map through an ant colony algorithm; the first advancing path output by the ant colony algorithm is verified through the target vector machine model, it is ensured that the first advancing path output by the ant colony algorithm is free of obstacles, and therefore the accuracy and reliability of the obtained target advancing path can be further improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and particularly to a path planning method, system, device, medium and program product for a robot. Background Art

[0002] In an industrial scenario, a path planning method for a robot needs to consider various factors, such as obstacles, path length, etc. Manually planning the travel path of the robot not only takes a lot of time and effort, but also has poor accuracy and effectiveness. Therefore, the path planning method for the robot still needs to be improved. Summary of the Invention

[0003] The technical problem to be solved by the present disclosure is to overcome the defect that manual path planning in the prior art not only takes a lot of time and effort, but also has poor accuracy and effectiveness, and to provide a path planning method, system, device, medium and program product for a robot.

[0004] The present disclosure solves the above technical problem through the following technical solutions:

[0005] In a first aspect, a path planning method for a robot is provided, and the path planning method includes:

[0006] Obtain a first starting point and a first ending point marked on a grid map;

[0007] Based on the grid map, the first starting point, the first ending point and an ant colony algorithm, obtain a first travel path of the robot, where the first travel path includes the first starting point and the first ending point;

[0008] Input the first travel path and the grid map into a target vector machine model;

[0009] In response to the output result of the target vector machine model being that there is no obstacle on the first travel path, plan the first travel path as the target travel path of the robot.

[0010] Optionally, the obtaining the first travel path of the robot based on the grid map, the first starting point, the first ending point and the ant colony algorithm includes;

[0011] Calculate the clustering radius of grid nodes in the grid map;

[0012] Calculate a guiding probability according to the clustering radius, the first starting point and the first ending point; where the guiding probability is used to determine the travel direction of the robot;

[0013] Based on the first starting point, the first ending point, the guiding probability, and the ant colony algorithm, the first traveling path is obtained.

[0014] Optionally, the guiding probability is calculated by the following formula:

[0015]

[0016] where P ij is the guiding probability, R is the clustering radius; n represents the number of grid nodes; x ik is the k-th pixel point in the x i of the first starting point; x jk is the k-th pixel point in the x j of the first ending point, and p k is the weighted value in the Euclidean distance of the k-th pixel point.

[0017] Optionally, the target vector machine model is obtained by the following steps:

[0018] Obtain each second starting point marked in the grid map and each second ending point corresponding to each second starting point;

[0019] Based on the grid map, each second starting point, each second ending point, and the ant colony algorithm, obtain the second traveling path of the robot; wherein, each second traveling path includes its corresponding second starting point and second ending point;

[0020] Construct training samples according to each second traveling path and the annotation data corresponding to each second traveling path; wherein, the annotation data includes whether there are obstacles on the second traveling path;

[0021] Use the training samples to train the initial vector machine model to obtain the target vector machine model.

[0022] Optionally, using the training samples to train the initial vector machine model to obtain the target vector machine model includes:

[0023] Obtain the optimization objective function of the initial vector machine model;

[0024] Based on the true value of the annotation data and the predicted value output by the initial vector machine model, calculate the minimum value of the optimization objective function;

[0025] Estimate the target parameters of the target vector machine model based on the minimum value;

[0026] Generate the target vector machine model based on the target parameters.

[0027] Optionally, before the step of obtaining the optimization objective function of the initial vector machine model, the following steps are included:

[0028] Construct the optimization objective function based on the insensitive loss function; wherein, the insensitive loss function is used to quantify the error between the predicted value and the true value.

[0029] In a second aspect, a path planning system for a robot is provided. The path planning system includes:

[0030] An acquisition module, configured to acquire a first starting point and a first ending point marked on a grid map;

[0031] A obtaining module, configured to obtain a first traveling path of the robot based on the grid map, the first starting point, the first ending point, and the ant colony algorithm, wherein the first traveling path includes the first starting point and the first ending point;

[0032] An input module, configured to input the first traveling path and the grid map into a target vector machine model;

[0033] A response module, in response to the output result of the target vector machine model indicating that there is no obstacle on the first traveling path, plans the first traveling path as the target traveling path of the robot.

[0034] Optionally, the obtaining module includes:

[0035] A first calculation unit, configured to calculate the clustering radius of grid nodes in the grid map;

[0036] A second calculation unit, configured to calculate a guiding probability according to the clustering radius, the first starting point, and the first ending point; wherein the guiding probability is used to determine the traveling direction of the robot;

[0037] A first obtaining unit, configured to obtain the first traveling path based on the first starting point, the first ending point, the guiding probability, and the ant colony algorithm.

[0038] Optionally, the calculation formula of the second calculation unit is:

[0039]

[0040] wherein, P ij is the guiding probability, R is the clustering radius; n represents the number of grid nodes; x ik is the k-th pixel point in the first starting point x i ; x jk is the k-th pixel point in the first ending point x j ; p kIt is the weighted value in the Euclidean distance of the k-th pixel point.

[0041] Optionally, the input module further includes:

[0042] An acquisition unit, configured to acquire each second starting point marked in the grid map and each second ending point corresponding to each second starting point;

[0043] A second obtaining unit, configured to obtain a second travel path of the robot based on the grid map, each second starting point, each second ending point, and the ant colony algorithm; wherein, each second travel path includes its corresponding second starting point and second ending point;

[0044] A construction unit, configured to construct a training sample according to each second travel path and the annotation data corresponding to each second travel path; wherein, the annotation data includes whether there are obstacles on the second travel path;

[0045] A third obtaining unit, configured to train an initial support vector machine model using the training sample to obtain the target support vector machine model.

[0046] Optionally, the third obtaining unit includes:

[0047] An acquisition component, configured to acquire an optimization objective function of the initial support vector machine model;

[0048] A calculation component, configured to calculate the minimum value of the optimization objective function based on the true value of the annotation data and the predicted value output by the initial support vector machine model;

[0049] An estimation component, configured to estimate the target parameters of the target support vector machine model based on the minimum value;

[0050] A generation component, configured to generate the target support vector machine model based on the target parameters.

[0051] Optionally, the third obtaining unit further includes:

[0052] A construction component, configured to construct the optimization objective function based on an insensitive loss function; wherein, the insensitive loss function is used to quantify the error between the predicted value and the true value.

[0053] In a third aspect, there is provided an electronic device, including a memory, a processor, and a computer program stored on the memory and configured to run on the processor, where when the processor executes the computer program, it implements the path planning method of the robot described in any one of the above.

[0054] Fourthly, a computer-readable storage medium is provided, on which a computer program is stored. It is characterized in that when the computer program is executed by a processor, the path planning method of the robot described in any one of the above is implemented.

[0055] Fifthly, a computer program product is provided, including a computer program. It is characterized in that when the computer program is executed by a processor, the path planning method of the robot described in any one of the above is implemented.

[0056] On the basis of conforming to the common knowledge in the art, the above preferred conditions can be combined arbitrarily, that is, the preferred examples of the present disclosure can be obtained.

[0057] The positive and progressive effects of the present disclosure are as follows: By simulating the behavior of ants releasing pheromones through the ant colony algorithm, it is possible to gradually converge to the global optimal solution or an approximate optimal solution and avoid falling into the local optimal solution. This enables finding the first travel path from the first starting point to the first ending point in a complex grid map; by verifying the first travel path output by the ant colony algorithm through the target vector machine model, it is ensured that there are no obstacles on the first travel path output by the ant colony algorithm, which can further improve the accuracy and reliability of obtaining the target travel path. Description of the Drawings

[0058] Figure 1 It is a flowchart of a path planning method for a robot provided in Embodiment 1 of the present disclosure;

[0059] Figure 2 It is a schematic module diagram of a path planning system for a robot provided in Embodiment 2 of the present disclosure;

[0060] Figure 3 It is a schematic structural diagram of an electronic device shown in Embodiment 3 of the present disclosure. Detailed Embodiments

[0061] The present disclosure will be further described below by way of embodiments, but the present disclosure is not limited to the scope of the described embodiments.

[0062] In the embodiments of the present disclosure, prefix words such as "first" and "second" are only used to distinguish different described objects, and have no limiting effect on the position, order, priority, quantity or content of the described objects, etc. The use of ordinal words and other prefix words for distinguishing described objects in the embodiments of the present disclosure does not constitute a limitation on the described objects. The description of the described objects refers to the description in the context of the embodiments, and should not constitute an unnecessary limitation because of the use of such prefix words. In addition, in the description of this embodiment, unless otherwise specified, the meaning of "a plurality" is two or more.

[0063] Embodiment 1

[0064] To overcome the defects that manual path planning not only takes a lot of time and effort, but also has poor accuracy and effectiveness, Embodiment 1 of the present disclosure provides a path planning method for a robot. Figure 1 The flowchart of a path planning method for a robot provided by Embodiment 1 of the present disclosure, the path planning method includes the following steps:

[0065] Step 101, obtain the first starting point and the first ending point marked on the grid map.

[0066] The first starting point and the first ending point are marked on the grid map according to the required path planning situation, and both the first starting point and the first ending point include areas of several pixel points.

[0067] Step 102, based on the grid map, the first starting point, the first ending point and the ant colony algorithm, obtain the first traveling path of the robot.

[0068] Among them, the first traveling path includes the first starting point and the first ending point. The first traveling path obtained by the ant algorithm is an approximately optimal traveling path, that is, there may be obstacles or dangerous areas on the approximately optimal traveling path.

[0069] Step 103, input the first traveling path and the grid map into the target vector machine model.

[0070] Step 104, in response to the output result of the target vector machine model being that there is no obstacle on the first traveling path, plan the first traveling path as the target traveling path of the robot.

[0071] The target support vector machine is used to predict whether there are dangerous areas (i.e., obstacles) in the path planning process. If so, a new traveling path is planned again using the ant colony algorithm, and the new traveling path is input to the target support vector machine to re-predict whether there are obstacles on the new traveling path; if not, the path planning is completed.

[0072] In this embodiment, by simulating the behavior of ants releasing pheromones through the ant colony algorithm, it is possible to gradually converge to the global optimal solution or an approximately optimal solution, avoiding falling into the local optimal solution. This enables finding the first traveling path from the first starting point to the first ending point in a complex grid map; by verifying the first traveling path output by the ant colony algorithm through the target vector machine model, it is ensured that there are no obstacles on the first traveling path output by the ant colony algorithm, which can further improve the accuracy and reliability of obtaining the target traveling path.

[0073] In one embodiment, based on the grid map, the first starting point, the first ending point and the ant colony algorithm, obtaining the first traveling path of the robot includes:

[0074] S1: Calculate the clustering radius of the grid nodes in the grid map.

[0075] For the specific method of the clustering radius of grid nodes in the grid map, please refer to the relevant technical description and will not be elaborated here.

[0076] S2: Calculate the guiding probability according to the clustering radius, the first starting point and the first ending point; wherein, the guiding probability is used to determine the traveling direction of the robot.

[0077] S3: Based on the first starting point, the first ending point, the guiding probability and the ant colony algorithm, obtain the first traveling path.

[0078] In this embodiment, by calculating the clustering radius of grid nodes, the obstacle distribution and spatial structure in the grid map can be understood more accurately. The clustering radius provides the density and range of obstacles, which helps the robot to better perceive the environment during traveling. Furthermore, the reliability and safety of path planning are improved. Guiding the traveling direction of the robot according to the calculated guiding probability can improve the flexibility and adaptability of path planning, and thus reduce the collision risk. By combining the ant colony algorithm with the guiding probability, the shortest path from the first starting point to the first ending point can be efficiently found, thereby reducing the path planning time. In addition, by calculating the clustering radius and the guiding probability, the path planning process of the robot can be automated, reducing the need for manual intervention, thus improving the efficiency of path planning and reducing the labor cost.

[0079] In one embodiment, the guiding probability is calculated by the following formula:

[0080]

[0081] wherein, P ij is the guiding probability, R is the clustering radius; n represents the number of grid nodes; x ik is the k-th pixel point in the first starting point x i ; x jk is the k-th pixel point in the first ending point x j , and p k is the weighted value in the Euclidean distance of the k-th pixel point.

[0082] In this embodiment, using this formula can calculate the guiding probability more accurately, and thus make the obtained first traveling path more accurate.

[0083] In one embodiment, a further description is given on how to obtain the first traveling path of the robot based on the ant colony algorithm in the Rviz visualization tool:

[0084] RViz (Robot Visualization) is a powerful visualization tool in ROS (Robot Operating System), which is used to display the robot's perception data, status information, motion trajectories, etc. in real time. In path planning, RViz plays a role in graphically displaying the path planning results to users, enabling them to intuitively see the planned path of the robot from the starting point to the ending point. It has broad application prospects and important practical value. By continuously optimizing algorithms and adapting to the requirements of specific scenarios, more efficient and accurate path planning can be achieved, improving the automation level and efficiency of industrial production.

[0085] S1: Robot model format conversion. First, the STEP file needs to be converted into a format that ROS can handle. Import the 3D model into URDF (Unified Robot Description Format) and configure the URDF file to ensure that the configured URDF file can correctly describe all the links, joints, linkages, and sensors of the robot, etc.

[0086] S2: Open the terminal to start the Rviz visualization tool. Start the Rviz visualization tool by entering the Rviz command in the terminal.

[0087] S3: After starting the Rviz visualization tool, set a fixed coordinate system to represent the reference frame of the world coordinate system. This is usually map or world. In the interface of the Rviz visualization tool, find the option to set the fixed coordinate system.

[0088] It should be noted that the grid map is a form of the map.

[0089] S4: Add different types of display items, including point cloud and robot status. Through the "Add" button in the lower left corner of the interface of the Rviz visualization tool, select the display type to be added and assign a unique name to it. Configure the attributes of the display item as needed, such as color, transparency, etc.

[0090] It should be noted that the point cloud is the grid node mentioned above.

[0091] S5: Compile the algorithm configuration function package of the robot in the working space of the Rviz visualization tool. Input the grid nodes of the Rviz visualization tool into the ant colony algorithm to construct a mathematical model of the ant colony search process. Write a topic publishing program, start the topic publishing program and the Rviz visualization tool, and use the Rviz visualization tool to listen to the topics of the lidar point cloud and the robot. To display the URDF file, a launch file needs to be created to load the robot model, and this file will start the Rviz visualization tool and enable the robot model to be seen in the opened window.

[0092] S6: The first starting point X marked on the grid map i The set of all pixel points in it is x i , that is, the existing set is x i =(x i1 , x i2 , x i3 …x ip ), where p represents the number of pixel points. This set can be used as the basis for the ant's travel path to determine the robot's first travel path. D ij is the first starting point X i The first ending point X j The distance of, where the first ending point X j The set of all pixel points in it is x j , that is, the existing set is x j =(x j1 , x j2 , x j3 …x jp ), where p represents the number of pixel points. Therefore, the mathematical model of the ant's crawling is:

[0093]

[0094] Among them, n represents the number of grid nodes in the grid map; p k is the weighted value in the Euclidean distance, and the size of this value depends on the clustering influence degree of the grid node on the grid map. x ik is the kth pixel point in the first starting point x i ; x jk is the kth pixel point in the first ending point x j , and the distance between the first starting point X i and the first ending point X j can be obtained through the ant algorithm model.

[0095] S7: Obtain the first travel path between the first starting point X i and the first ending point X j .

[0096] In order to obtain the ant's travel route (this ant's travel route is the first travel path), first, the guiding probability p ij , p ij The guiding probability value of is calculated by the following formula, that is:

[0097]

[0098] Among them, among them, P ij is the guiding probability, R is the clustering radius; n represents the number of grid nodes; x ik is the first starting point x ithe k-th pixel in; x jk is the first end point x j the k-th pixel in, p k is the weighted value in the Euclidean distance of the k-th pixel, where k represents the ant number.

[0099] Divide the clustering radius by D ij distance to obtain the guiding probability p ij , that is, the guiding probability p ij The value depends on the value of the clustering radius, that is, the probability of the ant choosing this path. Initialize a set of random routes of ants in the image, calculate the gradient values of the pixels around each ant, and use them as the information perceived by the ants. Each ant selects the next moving direction according to the perceived information, updates the position of the ant, and updates the pheromone concentration according to the evaporation of the pheromone. Repeat the step of "each ant selects the next moving direction according to the perceived information, updates the position of the ant, and updates the pheromone concentration according to the evaporation of the pheromone" until the stop condition is reached. The stop conditions include that the number of iterations reaches the target number. According to the guiding probability value p ij of the finally obtained guiding function, the traveling route of the ant is determined, that is, the first traveling path of the robot.

[0100] S8: In the terminal, after running the written launch file and algorithm file, open two software, gazebo (an open-source 3D robot simulation software) and the Rviz visualization tool. Control the robot model to traverse the entire environment in gazebo. The robot will draw a grid map in real time in the Rviz visualization tool and obtain the first traveling path between the first starting point and the first ending point through the ant colony algorithm. Among them: the 'i' key can control the robot to move forward, the ',' key can control the robot to move backward, the 'j' key controls the robot to turn left, and the 'l' key controls the robot to turn right. The 'u', 'o','m', '.' four keys respectively control the robot to move in the front left, front right, rear left, and rear right directions. The 'k' key or any other key can stop the robot. When controlling the robot to move, the terminal running the keyboard control program is the currently selected terminal, and the first traveling path of the robot is found.

[0101] S9: Enter the command in this terminal to save the drawn grid map in the root directory of the workspace. Find the file in the ros_workspace path and close all programs. The search for the first traveling path of the robot is completed.

[0102] In one embodiment, the target vector machine model is obtained by the following steps:

[0103] S1: Obtain each second starting point marked in the grid map and each second ending point corresponding to each second starting point.

[0104] S2: Based on the grid map, each second starting point, each second ending point, and the ant colony algorithm, obtain the second travel path of the robot; wherein, each second travel path includes its corresponding second starting point and second ending point.

[0105] S3: Construct training samples according to each second travel path and the annotation data corresponding to each second travel path; wherein, the annotation data includes whether there are obstacles on the second travel path.

[0106] The training sample can be: The second travel path between the second starting point A and the second ending point B has obstacles. Wherein, having obstacles is the annotation data.

[0107] S4: Use the training samples to train the initial support vector machine model to obtain the target support vector machine model.

[0108] In this embodiment, the second travel path that can be found by the ant colony algorithm is the shortest path from each starting point to each ending point. Using these shortest paths and whether there are obstacles on the shortest paths as training data helps the initial support vector machine model to more accurately learn the distribution pattern of obstacles, and further enables the obtained target support vector machine model to have generalization ability, and further enables the obtained target support vector machine model to predict the input travel path more accurately.

[0109] In one embodiment, using the training samples to train the initial support vector machine model to obtain the target support vector machine model includes:

[0110] S1: Obtain the optimization objective function of the initial support vector machine model.

[0111] This optimization objective function can enable the initial support vector machine model to have a clear goal during the training process.

[0112] S2: Based on the true value of the annotation data and the predicted value output by the initial support vector machine model, calculate the minimum value of the optimization objective function.

[0113] S3: Estimate the target parameters of the target support vector machine model based on the minimum value.

[0114] The target parameters can be the weight vector w and bias b of the target support vector machine model.

[0115] S4: Generate the target support vector machine model based on the target parameters.

[0116] In this embodiment, by optimizing the objective function, the initial support vector machine model can be prevented from blindly optimizing during the training process, thereby improving the efficiency and effectiveness of training the initial support vector machine model. By calculating the minimum value of the optimized objective function, the error between the predicted value and the true value of the initial support vector machine model can be quantified. The effective error quantification enables the initial support vector machine model to more accurately evaluate its own performance during the training process, thereby gradually reducing the error during the training process to improve the prediction accuracy of the obtained target support vector machine model. By minimizing the optimized objective function, the target parameters of the target support vector machine model can be accurately estimated. The target support vector machine model generated based on the target parameters has the smallest prediction error on the training data, and while controlling the complexity, has good generalization ability. Furthermore, the generated target support vector machine model can accurately predict on new data, improving the reliability and effectiveness of the target support vector machine model in practical applications.

[0117] In one embodiment, before the step of obtaining the optimized objective function of the initial support vector machine model, it includes:

[0118] Construct an optimized objective function based on the insensitive loss function. Among them, the insensitive loss function is used to quantify the error between the predicted value and the true value.

[0119] In this embodiment, since the insensitive loss function allows the obtained target support vector machine model not to be penalized when the error between the predicted value and the true value is less than or equal to ∈ (∈ is the parameter value in the insensitive loss function). Therefore, the optimized objective function constructed based on the insensitive loss function and the target support vector machine model obtained based on this objective function can tolerate small errors within a certain range, thereby reducing overfitting to noise data and outliers. Furthermore, the robustness of the target support vector machine model can be improved, enabling the target support vector machine model to make more stable predictions when facing imperfect data and reducing the risk of overfitting.

[0120] In one embodiment, specifically illustrate how to obtain the target support vector machine model.

[0121] S1: Select a suitable kernel function according to parameter characteristics such as point cloud filtering concentration and point cloud noise distribution to obtain the initial support vector machine model.

[0122] It should be noted that: The point cloud is the point cloud (i.e., raster node) added in the above Rviz visualization tool. The selection of the kernel function is determined according to the characteristics of the point cloud (such as filtering concentration, noise distribution, etc.). The role of the kernel function is to define the relationship between adjacent points in the point cloud, so as to achieve the filtering operation. For example: If the noise distribution of the point cloud is relatively uniform and smoothing processing is required, a Gaussian kernel can be selected. The Gaussian kernel performs weighted averaging on the points in the neighborhood through weight attenuation, and can effectively smooth the noise. If the noise is relatively sparse and outliers need to be removed, a kernel function based on statistical analysis, such as the Statistical Outlier Removal (SOR) algorithm, can be used. SOR determines whether a point is a noise point by calculating the average distance between each point and its neighboring points.

[0123] Create a preset regression function for the initial vector machine model, the basic idea of which is through a certain non-linear mapping relationship Map the sample space to a high-dimensional space, thereby transforming the original low-dimensional non-linear problem into a high-dimensional linear problem to complete the regression. Given x i =(x i1 , x i2 , x i3 …x ip ), where p represents the number of pixel points and is the feature vector of the input factor. Among them, x i ∈R p , y i ∈R, the preset regression function of the initial support vector machine is:

[0124] The preset regression function of this initial vector machine model is:

[0125]

[0126] Among them, is the non-linear mapping relationship, w and b respectively represent the complexity and bias of the function, and γ represents the kernel function parameter of the preset regression function.

[0127] S2: Use the training samples according to the ant colony algorithm and the raster map data generated by the Rviz visualization tool to train the initial vector machine model to obtain the target vector machine model.

[0128] The value of the coefficient w can be estimated by the minimum value of the following formula:

[0129]

[0130] Among them, L ε is the insensitive loss function; ε is the parameter in the insensitive loss function; C is the penalty factor; n is the number of training samples; f(x i ) is the predicted value; yi is the true value of the labeled data on the training samples; i is the i-th sample.

[0131] Introduce the slack variable ξ i and With the insensitive loss function L ε As the structural minimization risk problem, the optimization objective can be transformed into a minimum value solving problem, and this minimum value function is:

[0132]

[0133] The constraint condition s.t. of this minimum value function is:

[0134]

[0135] Introduce the Lagrangian equation to take partial derivatives with respect to w, b, ξ respectively i 、 and set them to 0, then its dual problem is a maximum value solving problem, and this maximum value function is:

[0136]

[0137] The constraint condition s.t. of this maximum value function is:

[0138]

[0139] Among them, a i 、 a j 、 represent the Lagrange multiplier coefficients.

[0140] Solving the above problem finally obtains the regression function of the target support vector machine:

[0141]

[0142] In the above formula, K(x i , x j ) is the kernel function, x i and x j represent the second starting point and the second ending point of the second travel path, and γ is the kernel function parameter of the regression function. Select its form as the Gaussian radial basis function:

[0143] K(x i , x j ) = exp(-‖x i - x j ‖ 2 / 2γ 2 )

[0144] Example 2

[0145] Corresponding to the embodiment of the path planning method of the foregoing robot, the present disclosure also provides an embodiment of a path planning system of the robot. Figure 2 FIG. is a schematic diagram of modules of a path planning system of a robot provided for Exemplary Embodiment 2 of the present disclosure. The path planning system includes:

[0146] An obtaining module 21, configured to obtain a first starting point and a first ending point marked in a grid map;

[0147] A obtaining module 22, configured to obtain a first traveling path of the robot based on the grid map, the first starting point, the first ending point, and an ant colony algorithm, where the first traveling path includes the first starting point and the first ending point;

[0148] An input module 23, configured to input the first traveling path and the grid map into a target vector machine model;

[0149] A response module 24, in response to the output result of the target vector machine model being that there is no obstacle on the first traveling path, plans the first traveling path as the target traveling path of the robot.

[0150] Optionally, the obtaining module 22 includes:

[0151] A first calculation unit, configured to calculate a clustering radius of grid nodes in the grid map;

[0152] A second calculation unit, configured to calculate a guiding probability according to the clustering radius, the first starting point, and the first ending point; where the guiding probability is used to determine the traveling direction of the robot;

[0153] A first obtaining unit, obtains the first traveling path based on the first starting point, the first ending point, the guiding probability, and the ant colony algorithm.

[0154] Optionally, the calculation formula of the second calculation unit is:

[0155]

[0156] Where P ij is the guiding probability, R is the clustering radius; n represents the number of grid nodes; x ik is the k-th pixel point in the first starting point x i ; x jk is the k-th pixel point in the first ending point x j ; p k is the weighted value in the Euclidean distance of the k-th pixel point.

[0157] Optionally, the input module 23 further includes:

[0158] An obtaining unit, configured to obtain each second starting point marked in the grid map and each second ending point corresponding to each second starting point;

[0159] A second obtaining unit, configured to obtain a second traveling path of the robot based on a grid map, each second starting point, each second ending point, and an ant colony algorithm; wherein, each second traveling path includes its corresponding second starting point and second ending point;

[0160] A constructing unit, configured to construct a training sample according to each second traveling path and the annotation data corresponding to each second traveling path; wherein, the annotation data includes whether there are obstacles on the second traveling path;

[0161] A third obtaining unit, configured to train an initial support vector machine model using the training sample to obtain a target support vector machine model.

[0162] Optionally, the third obtaining unit includes:

[0163] An obtaining component, configured to obtain an optimization objective function of the initial support vector machine model;

[0164] A calculating component, configured to calculate the minimum value of the optimization objective function based on the true value of the annotation data and the predicted value output by the initial support vector machine model;

[0165] An estimating component, configured to estimate the target parameters of the target support vector machine model based on the minimum value;

[0166] A generating component, configured to generate a target support vector machine model based on the target parameters.

[0167] Optionally, the third obtaining unit further includes:

[0168] A constructing component, configured to construct an optimization objective function based on an insensitive loss function; wherein, the insensitive loss function is used to quantify the error between the predicted value and the true value.

[0169] For the system embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The system embodiment described above is only illustrative. The units described as separate components may or may not be physically separated. The components as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present disclosure solution.

[0170] Embodiment 3

[0171] Figure 3A schematic structural diagram of an electronic device shown in an exemplary embodiment of the present disclosure. The electronic device includes a memory, a processor, and a computer program stored in the memory and configured to run on the processor. When the processor executes the computer program, it implements the path planning method of the robot described in any of the above embodiments. Figure 3 The illustrated electronic device 30 is merely an example and should not impose any limitation on the functions and usage scope of the embodiments of the present disclosure.

[0172] As Figure 3 shown, the electronic device 30 may be presented in the form of a general-purpose computing device, for example, it may be a server device. The components of the electronic device 30 may include, but are not limited to: at least one of the above-mentioned processors 31, at least one of the above-mentioned memories 32, and a bus 33 connecting different system components (including the memory 32 and the processor 31).

[0173] The bus 33 includes a data bus, an address bus, and a control bus.

[0174] The memory 32 may include volatile memory, such as random access memory (RAM) 321 and / or cache memory 322, and may further include read-only memory (ROM) 323.

[0175] The memory 32 may further include a program tool 325 (or utility) having a set of (at least one) program modules 324. Such program modules 324 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0176] The processor 31 executes various functional applications and data processing by running the computer program stored in the memory 32, such as the path planning method of the robot provided in any of the above embodiments.

[0177] The electronic device 30 may also communicate with one or more external devices 34 (such as a keyboard, a pointing device, etc.). Such communication may be performed through an input / output (I / O) interface 35. And, the electronic device 30 may further communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 36. As shown in the figure, the network adapter 36 communicates with other modules of the electronic device 30 through the bus 33. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in combination with the electronic device 30, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems, etc.

[0178] It should be noted that although several units / modules or sub-units / modules of the electronic device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / modules. Conversely, the features and functions of one unit / modules described above can be further divided and embodied by multiple units / modules.

[0179] Embodiment 4

[0180] The embodiments of the present disclosure also provide a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the path planning method of the robot provided in any one of the above embodiments.

[0181] Among them, the more specific computer-readable storage medium that can be adopted may include, but is not limited to: portable disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0182] Embodiment 5

[0183] The embodiments of the present disclosure also provide a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the path planning method of the robot described in any one of the above.

[0184] Among them, the program code for executing the computer program product of the present disclosure can be written in any combination of one or more programming languages, and the program code can be executed entirely on the user device, partially on the user device, executed as an independent software package, partially on the user device and partially on a remote device, or entirely on a remote device.

[0185] Although the specific embodiments of the present disclosure are described above, those skilled in the art should understand that this is only for illustration, and the protection scope of the present disclosure is defined by the appended claims. Without departing from the principles and essence of the present disclosure, those skilled in the art can make various changes or modifications to these embodiments, but these changes and modifications all fall within the protection scope of the present disclosure.

Claims

1. A robot path planning method, characterized in that: The path planning method comprises: Get the first starting point and the first end point marked in the grid map; Based on the grid map, the first starting point, the first end point and the ant colony algorithm, a first travel path of the robot is obtained, wherein the first travel path includes the first starting point and the first end point; inputting the first travel path and the grid map into a target vector machine model; In response to the output result of the target vector machine model being that there is no obstacle on the first travel path, the first travel path is planned as the target travel path of the robot.

2. The path planning method according to claim 1, characterized in that: The obtaining of the first travel path of the robot based on the grid map, the first starting point, the first end point and the ant colony algorithm comprises: Calculating the clustering radius of grid nodes in the grid map; Calculating a guidance probability according to the cluster radius, the first starting point, and the first end point; wherein the guidance probability is used to determine the traveling direction of the robot; The first travel path is obtained based on the first starting point, the first end point, the guidance probability and the ant colony algorithm.

3. The path planning method according to claim 2, characterized in that: The bootstrap probability is calculated by the following formula: Among them, P ij is the guide probability, R is the cluster radius; n represents the number of grid nodes; x ik is the first starting point x i The kth pixel in x jk The first end point x j The kth pixel in p k is the weighted value in the Euclidean distance of the k-th pixel.

4. The path planning method according to claim 1, characterized in that: The target vector machine model is obtained by the following steps: Acquire each second starting point marked in the grid map and each second end point corresponding to the each second starting point; Based on the grid map, each second starting point, each second end point and the ant colony algorithm, a second travel path of the robot is obtained; wherein each second travel path includes its corresponding second starting point and second end point; Constructing a training sample according to each of the second travel paths and the annotation data corresponding to each of the second travel paths; wherein the annotation data includes whether there is an obstacle on the second travel path; The training samples are used to train an initial vector machine model to obtain the target vector machine model.

5. The path planning method according to claim 4, characterized in that: The step of training the initial vector machine model using the training samples to obtain the target vector machine model includes: Obtaining an optimization objective function of the initial vector machine model; Calculating the minimum value of the optimization objective function based on the true value of the labeled data and the predicted value output by the initial vector machine model; estimating a target parameter of the target vector machine model based on the minimum value; Based on the target parameters, the target vector machine model is generated.

6. The path planning method according to claim 5, characterized in that: The step of obtaining the optimization objective function of the initial vector machine model includes: The optimization objective function is constructed based on an insensitive loss function; wherein the insensitive loss function is used to quantify the error between the predicted value and the true value.

7. A robot path planning system, characterized in that: The path planning system comprises: An acquisition module, used to acquire a first starting point and a first end point marked in a grid map; An obtaining module, configured to obtain a first travel path of the robot based on the grid map, the first starting point, the first end point, and an ant colony algorithm, wherein the first travel path includes the first starting point and the first end point; An input module, used for inputting the first travel path and the grid map into a target vector machine model; The response module plans the first travel path as the target travel path of the robot in response to an output result of the target vector machine model indicating that there is no obstacle on the first travel path.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and used to run on the processor, characterized in that: When the processor executes the computer program, the robot path planning method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the path planning method for a robot according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed by a processor as the path planning method for a robot according to any one of claims 1 to 6.

Images