Intelligent path planning method for inspection robots integrating GPR and DWA algorithms

Abstract

This invention discloses an intelligent path planning method for inspection robots that integrates the GPR and DWA algorithms. The method involves acquiring overall plant area point cloud data, partitioning local maps, and conducting global and local path planning. Within the plant environment, a simulated plant growth algorithm and a dynamic window method are introduced into the robot's local path planning to evaluate the optimal obstacle avoidance path for the mobile robot, improving the efficiency and safety of path planning in dynamic environments. The invention divides the grid map based on inspection nodes and employs block-based path planning. This algorithm has a short runtime and can quickly plan the global optimal path. The integration of the simulated plant growth algorithm and the dynamic window method enables intelligent local path planning in complex dynamic environments, effectively addressing scenarios with obstacles.

Description

Technical Field

[0001] The present invention belongs to the technical field of robot path planning, and relates to an inspection robot path intelligent planning method that integrates a plant growth simulation (GPR) algorithm and a dynamic windowing algorithm (DWA). Background Art

[0002] With the development and maturity of artificial intelligence (AI) technology, as well as the popularization of the Internet of Things (IoT), cloud computing, and 5G technologies, intelligent inspection robots, designed to replace manual inspections, have been applied to an increasing number of fields. Intelligent inspection robots are now present in factory production lines, IDC rooms, power lines, and many other production environments that require regular and frequent inspections. Regarding path planning, the mobile robot's environment is unknown and dynamic in most scenarios, and some areas may experience environmental changes. The motion of dynamic obstacles is difficult to accurately predict, and it is impossible to rely on static maps to plan a safe and reliable path. This undoubtedly poses a significant challenge to the path planning module. Therefore, improving the accuracy of robot path planning and improving the timeliness of inspections are urgent issues that those skilled in the art need to address. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent path planning method for an inspection robot that integrates the GPR and DWA algorithms. The method includes the acquisition of overall point cloud data of the factory area, local map division, global path planning and local path planning. The simulated plant growth algorithm and dynamic window method are introduced into the local path planning of the robot in the factory environment to evaluate the optimal obstacle avoidance path of the mobile robot, thereby improving the path planning and obstacle avoidance efficiency and safety of the mobile robot in a dynamic environment.

[0004] An intelligent path planning method for an inspection robot that integrates GPR and DWA algorithms includes the following steps:

[0005] Step 1: The robot is equipped with a laser radar to scan the factory area. The robot is remotely controlled to fly around the factory area to obtain the overall point cloud data of the factory area.

[0006] Step 2: Create a grid map of the entire factory area based on the point cloud data obtained in step 1;

[0007] Step 3: Mark the inspection nodes, starting points, and end points on the overall grid map of the factory area established in step 2 based on inspection requirements;

[0008] Step 4: According to the inspection plan, the detection nodes marked in step 3 are numbered and sorted, and the overall grid map of the factory area is divided into blocks according to the inspection order. Some exits are closed and path planning is performed in blocks to reduce the amount of calculation;

[0009] Step 5. Calculate light intensity:

[0010] Calculate the distance from the 8 grids around the robot to the target grid, where the 8 grids are recorded as up, down, left, right, upper left, upper right, lower left and lower right. When calculating the distance, calculate the distance from the current grid to the target grid on the x and y axes respectively. The light intensity is inversely proportional to the distance length. The light intensity OA from the current grid to the target grid in the x-axis direction is x And the light intensity OA from the current grid to the target point in the y-axis direction y The calculation formulas are:

[0011]

[0012] Where, O x is the distance from the current grid to the target grid in the x-axis direction, O y The distance from the current grid to the target point in the y-axis direction;

[0013] Step 6: Calculate the shading intensity:

[0014] Step 61: Set the unit distance of each grid in the four directions of up, down, left, and right to 1, and the unit distance of each grid in the four directions of upper left, upper right, lower left, and lower right to

[0015] Step 62: When encountering an obstacle or boundary, stop moving forward and record the number N of all grids in each direction of up, down, left, right, upper left, upper right, lower left, and lower right. Then multiply it by the corresponding unit distance to get the final distance. The distance length is inversely proportional to the shading intensity. The calculation formula of the shading intensity S is as follows:

[0016]

[0017] Where N is the number of grids required from the eight grids around the robot to the obstacle or boundary; S is the shading intensity;

[0018] Step 7. Calculate the total light intensity:

[0019] Calculate the total light intensity of the 8 grids around the robot respectively. The calculation formula for the total light intensity is:

[0020] S x =S×cosθ

[0021] S y =S×sinθ

[0022] R x =OA x -S x

[0023] R=OA-S

[0024]

[0025] Where θ is the angle between the eight grids around the robot and the current position (for example, the grid in the right direction has an angle of 0°, and the grid in the upper right direction has an angle of 45°); S x is the component of the shading intensity in the x-axis direction; S y is the component of the shading intensity in the y-axis direction; R x is the component of the total light intensity on the x-axis; R y is the component of the total light intensity on the y-axis; R is the total light intensity; OA x is the light intensity in the x-axis direction from the current grid to the target grid, OA y The light intensity in the y-axis direction from the current grid to the target grid;

[0026] Step 8: Determine the direction of movement:

[0027] Based on the total light intensity obtained in step 7, the direction with the largest total light intensity is selected as the robot movement direction;

[0028] Step 9: Record the path coordinates. According to the moving direction of each step of the robot, record the robot trajectory coordinates to obtain the robot's global path (x j ,y j );

[0029] Step 10: Determine the speed space:

[0030] Due to the limitations of the robot itself, surrounding obstacles, and motor performance, the robot's linear and angular velocities will operate within a certain range. The specific limitations are as follows:

[0031] When the robot speed is limited by its own speed, its linear speed V m and angular velocity ω m The restrictions are:

[0032]

[0033] Where V min 、V max are the minimum and maximum speeds that the robot can reach respectively; ω min 、ω max are the minimum and maximum angular velocities that the robot can achieve respectively;

[0034] When the robot is limited by the motor performance, its linear speed V a , and angular velocity ω, a The restrictions are:

[0035]

[0036] Where V c 、ω care the current linear velocity and angular velocity of the robot respectively; V a 、ω a are the absolute value of the robot's linear acceleration and the absolute value of its maximum angular acceleration; Δt is the unit time;

[0037] When the robot is restricted by obstacles, its linear velocity V s and angular velocity ω s The restrictions are:

[0038]

[0039] Where d is the distance between the obstacle and the robot;

[0040] Based on the above three conditions, a speed space can be jointly determined. The speed activity space is composed of V m 、V a , and V s Determined together; similarly, the acceleration activity space is determined by ω m 、ω, a and ω s Determine together, randomly sample the linear velocity V and angular velocity ω in the velocity activity space and acceleration activity space respectively, and form a series of velocity groups [V i ,ω i ];

[0041] Step 11: Speed ​​space evaluation:

[0042] An evaluation function based on three factors, light intensity, shading intensity, and proximity to the global path, is constructed to evaluate the trajectory represented by each speed group in the speed space and select the optimal speed group. The evaluation function formula is as follows:

[0043] G(V,ω)=σ(β·B(V i ,ω i )+α·S max (V i ,ω i )+γ·P(V i ,ω i ))

[0044] In the formula, α, β, and γ are the coefficients of the evaluation function, σ represents normalization, and the formulas for each sub-item are as follows:

[0045]

[0046] Where, d is the distance between the obstacle and the robot, d is the distance between the robot and the end point, (x a ,y a ) is the coordinate of the robot's trajectory end point under the current speed group, (x j ,yj ) is the global path coordinate of the robot, n is the number of global path trajectory coordinates; m is the number of random samples in the velocity group;

[0047] Step 12: Determine the optimal trajectory:

[0048] The speed group that minimizes G(V,ω) is selected as the optimal speed group, thereby obtaining the local path trajectory and completing the global path planning and local path planning tasks.

[0049] Compared with the prior art, the present invention has the following advantages:

[0050] The present invention divides the grid map based on inspection nodes and adopts block path planning. The algorithm has a short running time and can quickly plan the global optimal path. The integration of the simulated plant growth algorithm and the dynamic window method realizes the intelligent planning of local paths in complex dynamic environments, which can effectively deal with scenarios with obstacles. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is an overall flow chart of the path planning method of the present invention;

[0052] Figure 2 Divide the map;

[0053] Figure 3 Schematic diagram of the eight grids around the robot;

[0054] Figure 4 Calculation result of the grid light intensity around the robot;

[0055] Figure 5 The calculation result of the grid shading intensity around the robot;

[0056] Figure 6 The calculation result of the total light intensity of the grid around the robot;

[0057] Figure 7 The direction of robot movement;

[0058] Figure 8 is the overall path of the robot;

[0059] Figure 9 is the local path of the robot. DETAILED DESCRIPTION

[0060] The technical solution of the present invention is further described below with reference to the accompanying drawings, but is not limited thereto. Any modification or equivalent replacement of the technical solution of the present invention that does not depart from the spirit and scope of the technical solution of the present invention should be included in the scope of protection of the present invention.

[0061] The present invention provides a path planning method that simulates the fusion of GPR algorithm and DWA, such as Figure 1 As shown, the method includes the following steps:

[0062] Step 1: The robot is equipped with a laser radar to scan the factory area. The robot is remotely controlled to circle the factory area to obtain the overall point cloud data of the factory area.

[0063] Step 2: Create an overall grid map of the factory area (boundaries, obstacles) based on the point cloud data obtained in step 1.

[0064] Step 3: Based on the inspection requirements, mark the inspection nodes, starting points, and end points in the overall grid map of the factory area established in step 2.

[0065] Step 4: According to the inspection plan, the detection nodes marked in step 3 are numbered and sorted, and the overall grid map of the factory area is divided into blocks according to the inspection order, some exits are closed, and path planning is performed in blocks to reduce the amount of calculation, such as Figure 2 shown.

[0066] Step 5. Calculate light intensity:

[0067] The light intensity is expressed in terms of distance, and the eight grids around the robot are calculated (respectively marked as up, down, left, right, upper left, upper right, lower left and lower right, as shown in the figure below). Figure 3 When calculating the distance, the distance from the current grid to the target grid is calculated on the X and Y axes respectively. The distance from the current grid to the target grid in the x-axis direction is 0 x , the distance from the current grid to the target point in the y-axis direction is O y The light intensity is inversely proportional to the distance. The farther the current grid is from the target grid, the weaker the light intensity. Therefore, the light intensity in the x-axis direction from the current grid to the target grid is OA. x , the light intensity in the y-axis direction is OA y ,like Figure 4 shown.

[0068]

[0069] Step 6: Calculate the shading intensity:

[0070] Set the unit distance of each grid in the four directions of up, down, left and right to 1, and the unit distance of each grid in the four directions of upper left, upper right, lower left and lower right to like Figure 5 shown.

[0071] When encountering an obstacle or boundary, stop moving forward and record the number of all grids N in each direction (up, down, left, right, upper left, upper right, lower left, lower right), and then multiply it by the corresponding unit distance to get the final distance. Because the distance length is inversely proportional to the shading intensity, the shading intensity S is calculated as follows:

[0072]

[0073] Step 7. Calculate the total light intensity:

[0074] Calculate the total light intensity of the eight grids around the robot (up, down, left, right, upper left, upper right, lower left and lower right), as follows: Figure 6 As shown, the total light intensity calculation formula is:

[0075] S x =S×cosθ

[0076] S y =S×sinθ

[0077] R x =OA x -S x

[0078] R y =OA y -S y

[0079]

[0080] Step 8: Determine the direction of movement:

[0081] Based on the total light intensity obtained in step 7, the direction with the largest total light intensity is selected as the robot movement direction, such as Figure 7 shown.

[0082] Step 9: Record the path coordinates. According to the moving direction of each step of the robot, record the robot trajectory coordinates to obtain the robot's global path (x j ,y j ),like Figure 8 shown.

[0083] The robot's operating environment is constantly changing, and people or objects often cross its path. In these situations, local path planning is necessary to avoid obstacles. The following describes local path planning for situations where unexpected obstacles appear.

[0084] Step 10: Determine the speed space:

[0085] Due to the limitations of the robot itself, surrounding obstacles, motor performance, etc., the linear speed and angular speed of the car will move within a range.

[0086] The robot speed cannot be increased indefinitely, it is limited by its own speed, and the linear speed V m and angular velocity ω m The restrictions are:

[0087]

[0088] Where V min 、V max are the minimum and maximum speeds that the robot can reach respectively; ω min 、ω max are the minimum and maximum angular velocities that the robot can achieve respectively;

[0089] The robot is limited by the motor performance, the linear speed V a , and angular velocity ω, a The restrictions are:

[0090]

[0091] Where V c 、ω c are the current linear velocity and angular velocity of the robot, V a 、ω a are the absolute value of the robot's linear acceleration and the absolute value of its maximum angular acceleration, respectively.

[0092] The robot is restricted by obstacles, and its linear speed V s and angular velocity ω s The restrictions are:

[0093]

[0094] Where d is the distance between the obstacle and the robot.

[0095] Based on the above three conditions, a velocity space can be jointly determined. In the velocity space, the linear velocity and angular velocity are randomly sampled to form a series of velocity groups [V i ,ω i ].

[0096] Step 11: Velocity space evaluation: After obtaining the velocity group, the corresponding trajectory can be inferred based on the kinematic model.

[0097] An evaluation function based on light intensity, shading intensity, and proximity to the global path is constructed to evaluate the trajectory represented by each speed group in the speed space and select the optimal speed group. The evaluation function formula is as follows:

[0098] G(V,ω)=σ(β·B(V i ,ω i )+α·S max (Vi ,ω i )+γ·P(V i ,ω i ))

[0099] In the formula, α, β, and γ are the coefficients of the evaluation function, σ represents normalization, and the formulas for each sub-item are as follows:

[0100]

[0101] Where d is the distance between the obstacle and the robot.

[0102]

[0103] Where d is the distance between the robot and the end point.

[0104]

[0105] In the formula, (x a ,y a ) is the coordinate of the robot's trajectory end point under the current speed group, (x j ,y j ) is the global path coordinate of the robot, and n is the number of global path trajectory coordinates.

[0106] Since the local path planning process requires the collection of multiple sensors, the collected information cannot be continuous, which will lead to large differences after evaluation. Normalization processing (smoothing processing) is performed, where σ represents normalization.

[0107]

[0108] Where m is the number of random samples in the speed group.

[0109] Step 12: Determine the optimal trajectory:

[0110] The speed group that minimizes G(V,ω) is selected as the optimal speed group, thereby obtaining the local path trajectory and completing the global path planning and obstacle avoidance and other local path planning tasks, such as Figure 9 shown.

Claims

1. An intelligent path planning method for inspection robots that integrates GPR and DWA algorithms, characterized by The method comprises the following steps: Step 1: The robot is equipped with a laser radar to scan the factory area. The robot is remotely controlled to fly around the factory area to obtain the overall point cloud data of the factory area. Step 2: Create a grid map of the entire factory area based on the point cloud data obtained in step 1; Step 3: Mark the inspection nodes, starting points, and end points on the overall grid map of the factory area established in step 2 based on inspection requirements; Step 4: According to the inspection plan, the detection nodes marked in step 3 are numbered and sorted, and the overall grid map of the factory area is divided into blocks according to the inspection order. Some exits are closed and path planning is performed in blocks to reduce the amount of calculation; Step 5. Calculate light intensity: Calculate the distance from the 8 grids around the robot to the target grid, where the 8 grids are recorded as up, down, left, right, upper left, upper right, lower left and lower right. When calculating the distance, calculate the distance from the current grid to the target grid on the x and y axes respectively. The light intensity is inversely proportional to the distance length. The light intensity from the current grid to the target grid in the x-axis direction is And the light intensity from the current grid to the target point in the y-axis direction The calculation formulas are: Where, O x is the distance from the current grid to the target grid in the x-axis direction, O y The distance from the current grid to the target point in the y-axis direction; Step 6: Calculate the shading intensity: Step 61: Set the unit distance of each grid in the four directions of up, down, left, and right to 1, and the unit distance of each grid in the four directions of upper left, upper right, lower left, and lower right to ; Step 62: When encountering an obstacle or boundary, stop moving forward and record the number N of all grids in each direction of up, down, left, right, upper left, upper right, lower left, and lower right. Then multiply it by the corresponding unit distance to get the final distance. The distance length is inversely proportional to the shading intensity. The calculation formula of the shading intensity S is as follows: Where N is the number of grids required from the eight grids around the robot to the obstacle or boundary; S is the shading intensity; Step 7. Calculate the total light intensity: Calculate the total light intensity of the 8 grids around the robot respectively. The calculation formula for the total light intensity is: Where θ is the angle between the eight grids around the robot and the current position; S x is the component of the shading intensity in the x-axis direction; S y is the component of the shading intensity in the y-axis direction; R x is the component of the total light intensity on the x-axis; R y is the component of the total light intensity on the y-axis; R is the total light intensity; OA x is the light intensity in the x-axis direction from the current grid to the target grid, OA y The light intensity in the y-axis direction from the current grid to the target grid; Step 8: Determine the direction of movement: Based on the total light intensity obtained in step 7, the direction with the largest total light intensity is selected as the robot movement direction; Step 9. Record the path coordinates. According to the movement direction of each step of the robot, record the robot trajectory coordinates to obtain the robot's global path. ; Step 10: Determine the speed space: Due to the limitations of the robot itself, surrounding obstacles, and motor performance, the robot's linear and angular velocities will operate within a certain range. The specific limitations are as follows: When the robot speed is limited by its own speed, its linear speed and angular velocity The restrictions are: Where, 、 are the minimum and maximum speeds that the robot can achieve respectively; 、 are the minimum and maximum angular velocities that the robot can achieve respectively; When the robot is limited by the motor performance, its linear speed and angular velocity The restrictions are: Where, 、 are the current linear velocity and angular velocity of the robot respectively; 、 are the absolute value of the robot's linear acceleration and the absolute value of its maximum angular acceleration respectively; is the unit time; When the robot is restricted by obstacles, its linear speed and angular velocity The restrictions are: Where d is the distance between the obstacle and the robot; Based on the above three conditions, a speed space can be jointly determined. The speed activity space is composed of 、 and Determined together; similarly, the acceleration activity space is determined by 、 and Determine together, in the speed activity space and acceleration activity space respectively the linear velocity and angular velocity Random sampling to form a series of speed groups[ , ]; Step 11: Speed ​​space evaluation: An evaluation function based on light intensity, shading intensity, and proximity to the global path is constructed to evaluate the trajectory represented by each speed group in the speed space and select the optimal speed group. The formula of the evaluation function is as follows: In the formula, α, β and γ are the coefficients of the evaluation function, σ represents normalization, is the distance between the obstacle and the robot, d is the distance between the robot and the end point, is the coordinate of the robot's trajectory end point under the current speed group, is the global path coordinate of the robot, n is the number of global path trajectory coordinates; m is the number of random samples in the velocity group; Step 12: Determine the optimal trajectory: Select The smallest speed group is used as the optimal speed group to obtain the local path trajectory and complete the global path planning and local path planning tasks.

Images