A high-precision safety path planning method for a patrol robot in a chemical plant

Abstract

The application discloses a kind of high-precision safety path planning methods of inspection robot under chemical scene, first based on laser point cloud data constructs two-dimensional grid map, and the obstacle in map is treated with swelling, based on this generates the safe travel corridor of robot;After improvement is carried out to ant colony algorithm, introduce height heuristic function, visibility heuristic function and space safety coefficient improve the transfer rule of ant colony algorithm, finally route search is carried out, and the optimal safety path is obtained.The scheme of the application is applicable to complex inspection environment such as chemical industry park, uses the improved obstacle swelling method and the map constructed by safe travel corridor, so that the volume of obstacle is swollen according to the size of robot, speed, collision risk degree and task related, generates a high-precision map considering safety and efficiency, combined with the ant colony algorithm of introducing space safety coefficient, height heuristic function and visibility heuristic function, so that robot can plan a safe and feasible efficient path in map.

Description

Technical Field

[0001] This invention belongs to the field of inspection robots, specifically relating to a high-precision safe path planning method for inspection robots in chemical industry scenarios. Background Technology

[0002] Improper storage of chemical products within chemical industrial parks can lead to fires and explosions, causing significant economic losses and casualties. Manual inspections are prone to fatigue and negligence, resulting in missed or incorrect inspections. To ensure the safe and orderly operation of equipment in the plant area and to prevent secondary accidents, chemical inspection robots have become a hot research topic for experts and scholars both domestically and internationally.

[0003] Path planning is one of the most important aspects of inspection robot research and a fundamental guarantee for the robot to perform various tasks safely and reliably. Currently, commonly used algorithms include genetic algorithms, artificial potential field methods, and particle swarm optimization algorithms.

[0004] Chemical environments are complex, with narrow passages and suspended pipes. If a robot makes a mistake while moving, the consequences can be extremely dangerous and serious. Therefore, the focus of path planning is safety and feasibility. It is urgent to study a high-precision and safe path planning method for inspection robots in chemical environments. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to provide a high-precision and safe path planning method for inspection robots in chemical engineering scenarios, which can still achieve safe and feasible path planning even in complex inspection environments.

[0006] The specific technical solution for achieving the objective of this invention is as follows:

[0007] A high-precision safety path planning method for inspection robots in chemical engineering scenarios includes the following steps:

[0008] Step 1: Construct a two-dimensional grid map based on laser point cloud data, and perform puffing processing on obstacles in the map;

[0009] Step 2: Generate a safe driving corridor for the robot based on a 2D grid map and expanded obstacles;

[0010] Step 3: Improve the ant colony algorithm by introducing a height heuristic function, a visibility heuristic function, and a spatial safety factor to improve the transition rules of the ant colony algorithm;

[0011] Step 4: Based on the improved ant colony algorithm and the safe driving corridor generated in Step 2, perform route search to obtain the optimal safe path.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0013] This invention is based on an obstacle expansion method related to robot size, speed, collision risk level, and task requirements. It also constructs a safe driving corridor based on these expanded obstacles, forming a high-precision grid map. Within the safe driving corridor, an ant colony algorithm incorporating spatial safety factors, height heuristics, and visibility heuristics is used for global path planning on the map. This method is suitable for complex inspection environments such as chemical industrial parks. Using an improved obstacle expansion method and a map constructed from the safe driving corridor, the obstacle volume is expanded according to robot size, speed, collision risk level, and task requirements, generating a high-precision map that balances safety and efficiency. Combined with the ant colony algorithm incorporating spatial safety factors, height heuristics, and visibility heuristics, the robot can plan a safe, feasible, and efficient path within the map.

[0014] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the high-precision safe path planning method for inspection robots in chemical industry scenarios according to the present invention.

[0016] Figure 2 This is a schematic diagram of the security corridor generation process of the present invention.

[0017] Figure 3 This is a schematic diagram illustrating the generation of a safe driving corridor in an embodiment of the present invention.

[0018] Figure 4 This is a schematic diagram illustrating the process of searching for the optimal safe path based on an ant colony algorithm with improved transfer rules in an embodiment of the present invention. Detailed Implementation

[0019] A high-precision safety path planning method for inspection robots in chemical engineering scenarios includes the following steps:

[0020] Step 1: Construct a 2D raster map based on laser point cloud data, and perform puffing processing on obstacles in the map:

[0021] Step 1-1: Obtain laser point cloud data and determine the degree of influence of laser on the grid according to the two-dimensional normal distribution formula. Obtain the influence degree of all laser points in each grid, remove point cloud data with influence degree less than the set threshold, and store the remaining point cloud data into the grid.

[0022] Steps 1-2: Based on the distribution of laser point clouds in the vertical direction within the grid, determine the average height of each grid, complete the matching of laser point clouds and grids, and obtain a two-dimensional grid map containing the height information of each grid.

[0023] Steps 1-3: To assess the collision risk level of the robot path, a hazard function is constructed to evaluate the collision risk level of the robot path.

[0024]

[0025]

[0026] Where M is the number of obstacles within the affected area, ρ(x i ,x m ) is the path point x i To obstacle x m The straight-line distance, d0 is the range of influence of the obstacle, which is related to the size of the robot's body, and c and d are coefficients that are proportional to the robot's velocity vector and the material of the obstacle, respectively, and determine the degree of danger caused to the robot by the collision.

[0027] Steps 1-4: Extract obstacle information from the 2D grid map and puff the edges of the obstacles:

[0028] Traditional obstacle expansion methods involve expanding the obstacle outward along its outermost contour to the robot's maximum radius. However, this approach doesn't adequately balance accuracy, safety, and speed. Therefore, this paper proposes an obstacle expansion method that considers multiple constraints, expanding the obstacle outward along its outermost contour by L:

[0029] L=αr+βF s +γT i +ωv i

[0030]

[0031] In the formula, r is the maximum radius of the robot, and F s It represents the collision risk level along the robot's path, v i It is the robot's speed, T i ρ(x) represents the robot's current task execution degree. i ,x des ) represents the path point x i The shortest distance to the target point, d1 represents the inspection range of the target point.

[0032] Step 2: Generate a safe driving corridor for the robot based on a 2D grid map and expanded obstacles:

[0033] Step 2-1: After the obstacle is expanded, the robot itself is treated as a point mass. Therefore, based on the two-dimensional grid map, an A* search is performed on the rear axis center of the robot to obtain the coarse trajectory of the rear axis center, that is, to determine the coarse trajectory of the point mass, and to obtain the total time m for path planning completed by A* search.

[0034] Step 2-2: Using the local rectangular region generation method, uniformly sample m to extract uniform trajectory points. Expand each sampling point outward in a circle. Determine temporary boundaries based on collision detection between the corresponding circle and obstacles. Expand outward by Δs units along the four sides of the rectangle in either counterclockwise or clockwise directions based on the trajectory points.

[0035] Δs=ρ(x i ,x m )-c1F s -c2v i

[0036] If Δs is too large, the rectangular area will be too small, making it easy to miss the optimal or near-optimal trajectory; if Δs is too small, the computational load will be too large, and the path planning speed will be slow. The rectangle's length and width must not exceed the set threshold S until it overlaps with an obstacle, to prevent excessive computation on a large empty area.

[0037] Step 2-3: Connect the intersection of the temporary boundary points formed in Step 2-2 with the rectangles on multiple trajectory points to form a safe driving corridor.

[0038] Step 3: Improve the ant colony algorithm by introducing a height heuristic function, a visibility heuristic function, and a spatial safety factor to refine the transition rules of the ant colony algorithm:

[0039] Step 3-1: Introduce a height heuristic function. To improve the accuracy of path planning, the missing height information is used as the height heuristic function H at the current moment of path selection. j (t), because the target point has a certain height, and when projected onto the two-dimensional environment, the heights of the starting point, the target point, and obstacles are all masked, so the height difference between the candidate node and the target point is used as a factor affecting the heuristic function. We prioritize nodes that are closer to the target point in height, making the height selection directional, which to some extent makes up for the problem of missing height information:

[0040]

[0041]

[0042] Since the height difference may be zero, a height intensity coefficient W is added, which is a constant. For example, when the height difference between the candidate grid and the target grid is 0, it indicates that height has little effect on path selection. j h represents the height of the grid cells to be selected. j0 (t) represents the updated raster height, h des Indicates the height of the target point;

[0043] Step 3-2: Change the visibility of adjacent grid cells to the visibility of the target grid cell using a pheromone-carrying node method. Introduce a visibility heuristic function to change the visibility of adjacent grid cells to the visibility of the target grid cell using a pheromone-carrying node method.

[0044]

[0045] d j (t)=ρ(x j ,x des )

[0046] Wherein, ρ(x) j ,x des ) is the path point x j The straight-line distance to the target point, x des Indicates the target point;

[0047] Step 3-3: Introduce a security heuristic function:

[0048]

[0049]

[0050] Where M represents the number of obstacles at the current location;

[0051] Steps 3-4: Due to the impact of adding a safety factor on path planning, it is necessary to establish new evaluation criteria. Based on the safety heuristic function, determine the path safety evaluation index function:

[0052]

[0053] Among them, S i (t) represents the safety factor of the i-th grid on the current path, and n represents the number of all grids on the current search path;

[0054] Before path planning, a path safety evaluation index V is set. Paths with a V value greater than the preset value are considered feasible. For each path in the ant search, only the pheromones on paths that meet the V evaluation criteria are updated to ensure that the optimal path meets the evaluation criteria.

[0055] Steps 3-5: Determine the transition rules for the improved ant colony algorithm:

[0056] Assuming the ant is in the i-th grid cell, the transfer rule for moving from grid cell i to the candidate grid cell j at time t is as follows:

[0057]

[0058] Where, τ j(t) represents the pheromone concentration of the candidate grid, where α represents the weight of the candidate grid pheromone concentration in the transition rule. A larger value indicates that the ant relies more on pheromones to select a waypoint. β represents the weight of the visibility heuristic function in the transition rule; a larger value indicates that the ant is more likely to choose a candidate node closer to the target point. λ represents the weight of the height heuristic function in the transition rule; a larger value indicates that the ant is more likely to choose a candidate node with a smaller height difference from the target point. ω represents the weight of the safety heuristic function in the transition rule; a larger value indicates a greater tendency to choose a node with fewer surrounding obstacles.

[0059] The improved visibility heuristic function, height heuristic function, and safety heuristic function work together to influence the ant's choice of nodes, guiding it towards the target node. The heuristic function guarantees the mechanism by which the ant moves towards the target node, while the heuristic factor indicates the importance of the heuristic function in the entire equation.

[0060] Step 4: Based on the improved ant colony algorithm and the safe driving corridor generated in Step 2, perform route search to obtain the optimal safe path:

[0061] Step 4-1: Based on the safe driving corridor determined in Step 2, determine the robot's starting point and inspection target point, initialize the relevant parameters of the ant colony algorithm, and set the threshold of the path safety evaluation index V.

[0062] Step 4-2: Place all ants at the starting point and build a table to store the path nodes that the ants have already visited. If the next node to be visited by the ant already exists in this table, then the node is marked as a node that cannot be visited again.

[0063] Step 4-3: Based on the improved transfer rules determined in Step 3, determine the next path node that the ant should go to, and put the visited nodes into the taboo list. When the ant reaches the target node, a search is completed.

[0064] Step 4-4: For each path in the ant search, only update the pheromone on the path where the path safety evaluation index V is greater than the set threshold to ensure that the optimal path meets the evaluation criteria. Compare the optimal paths in each iteration to find the current optimal path and determine whether the set termination condition is met. If it is met, output the optimal safe path; otherwise, return to step 4-2 and continue to the next iteration until the termination condition is met.

[0065] Steps 4-5: After arriving at the inspection point, calculate the distance between the robot and the inspection target. Once the inspection threshold is reached, perform the inspection work. After the inspection is completed, execute the path planning from the current point to the next inspection point.

[0066] A high-precision safety path planning system for inspection robots in chemical industry scenarios includes the following modules:

[0067] Map building module: Used to build a two-dimensional raster map based on laser point cloud data and to enlarge obstacles in the map;

[0068] Safe Driving Corridor Generation Module: Used to generate a safe driving corridor for the robot based on a 2D grid map and expanded obstacles;

[0069] Transfer rule determination module: Used to improve the ant colony algorithm by introducing height heuristic function, visibility heuristic function and spatial safety coefficient to improve the transfer rules of the ant colony algorithm;

[0070] Optimal safe path generation module: used to search for routes based on the improved ant colony algorithm and the generated safe driving corridors to obtain the optimal safe path.

[0071] A computer device includes a memory, a processor, and a computer program stored in the memory and operable on the processor, which, when executed by the processor, can perform the following steps:

[0072] Step 1: Construct a two-dimensional grid map based on laser point cloud data, and perform puffing processing on obstacles in the map;

[0073] Step 2: Generate a safe driving corridor for the robot based on a 2D grid map and expanded obstacles;

[0074] Step 3: Improve the ant colony algorithm by introducing a height heuristic function, a visibility heuristic function, and a spatial safety factor to improve the transition rules of the ant colony algorithm;

[0075] Step 4: Based on the improved ant colony algorithm and the safe driving corridor generated in Step 2, perform route search to obtain the optimal safe path.

[0076] A computer-storeable medium storing a computer program on which a processor performs the following steps:

[0077] Step 1: Construct a two-dimensional grid map based on laser point cloud data, and perform puffing processing on obstacles in the map;

[0078] Step 2: Generate a safe driving corridor for the robot based on a 2D grid map and expanded obstacles;

[0079] Step 3: Improve the ant colony algorithm by introducing a height heuristic function, a visibility heuristic function, and a spatial safety factor to improve the transition rules of the ant colony algorithm;

[0080] Step 4: Based on the improved ant colony algorithm and the safe driving corridor generated in Step 2, perform route search to obtain the optimal safe path:

[0081] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0082] Example

[0083] Combination Figure 1 A high-precision safety path planning method for inspection robots in chemical industry scenarios includes the following steps:

[0084] Step 1: Construct a 2D raster map based on laser point cloud data, and perform puffing processing on obstacles in the map:

[0085] Step 1-1: Obtain laser point cloud data and determine the degree of influence of laser on the grid according to the two-dimensional normal distribution formula. Obtain the influence degree of all laser points in each grid, remove point cloud data with influence degree less than the set threshold, and store the remaining point cloud data into the grid.

[0086] Steps 1-2: Based on the distribution of laser point clouds in the vertical direction within the grid, determine the average height of each grid, complete the matching of laser point clouds and grids, and obtain a two-dimensional grid map containing the height information of each grid.

[0087] Steps 1-3: To assess the collision risk level of the robot path, a hazard function is constructed to evaluate the collision risk level of the robot path.

[0088]

[0089]

[0090] Where M is the number of obstacles within the affected area, ρ(x i ,x m ) is the path point x i To obstacle x m The straight-line distance, d0 is the range of influence of the obstacle, which is related to the size of the robot's body, and c and d are coefficients that are proportional to the robot's velocity vector and the material of the obstacle, respectively, and determine the degree of danger caused to the robot by the collision.

[0091] Steps 1-4: Extract obstacle information from the 2D grid map and puff the edges of the obstacles:

[0092] Traditional obstacle expansion methods involve expanding the obstacle outward along its outermost contour to the robot's maximum radius. However, this approach doesn't adequately balance accuracy, safety, and speed. Therefore, this paper proposes an obstacle expansion method that considers multiple constraints, expanding the obstacle outward along its outermost contour by L:

[0093] L=αr+βF s +γT i +ωv i

[0094]

[0095] In the formula, r is the maximum radius of the robot, and F s It represents the collision risk level along the robot's path, v i It is the robot's speed, T i ρ(x) represents the robot's current task execution degree. i ,x des ) represents the path point x i The shortest distance to the target point, d1 represents the inspection range of the target point.

[0096] Step 2, Combining Figure 2 and Figure 3 A safe driving corridor for robots is generated based on a 2D grid map and expanded obstacles:

[0097] Step 2-1: After the obstacle is expanded, the robot itself is treated as a point mass. Therefore, based on the two-dimensional grid map, an A* search is performed on the rear axis center of the robot to obtain the coarse trajectory of the rear axis center, that is, to determine the coarse trajectory of the point mass, and to obtain the total time m for path planning completed by A* search.

[0098] Step 2-2: Using the local rectangular region generation method, uniformly sample m to extract uniform trajectory points. Expand each sampling point outward in a circle. Determine temporary boundaries based on collision detection between the corresponding circle and obstacles. Expand outward by Δs units along the four sides of the rectangle in either counterclockwise or clockwise directions based on the trajectory points.

[0099] Δs=ρ(x i ,x m )-c1F s -c2v i

[0100] If Δs is too large, the rectangular area will be too small, making it easy to miss the optimal or near-optimal trajectory; if Δs is too small, the computational load will be too large, and the path planning speed will be slow. The rectangle's length and width must not exceed the set threshold S until it overlaps with an obstacle, to prevent excessive computation on a large empty area.

[0101] Step 2-3: Connect the intersection of the temporary boundary points formed in Step 2-2 with the rectangles on multiple trajectory points to form a safe driving corridor.

[0102] Step 3: Improve the ant colony algorithm by introducing a height heuristic function, a visibility heuristic function, and a spatial safety factor to refine the transition rules of the ant colony algorithm:

[0103] Step 3-1: Introduce a height heuristic function. To improve the accuracy of path planning, the missing height information is used as the height heuristic function H at the current moment of path selection. j (t), because the target point has a certain height, and when projected onto the two-dimensional environment, the heights of the starting point, the target point, and obstacles are all masked, so the height difference between the candidate node and the target point is used as a factor affecting the heuristic function. We prioritize nodes that are closer to the target point in height, making the height selection directional, which to some extent makes up for the problem of missing height information:

[0104]

[0105]

[0106] Since the height difference may be zero, a height intensity coefficient W is added, which is a constant. For example, when the height difference between the candidate grid and the target grid is 0, it indicates that height has little effect on path selection. j h represents the height of the grid cells to be selected. j0 (t) represents the updated raster height, h des Indicates the height of the target point;

[0107] Step 3-2: Change the visibility of adjacent grid cells to the visibility of the target grid cell using a pheromone-carrying node method. Introduce a visibility heuristic function to change the visibility of adjacent grid cells to the visibility of the target grid cell using a pheromone-carrying node method.

[0108]

[0109] d j (t)=ρ(x j ,x des )

[0110] Wherein, ρ(x) j ,x des ) is the path point x j The straight-line distance to the target point, x des Indicates the target point;

[0111] Step 3-3: Introduce a security heuristic function:

[0112]

[0113]

[0114] Where M represents the number of obstacles at the current location;

[0115] Steps 3-4: Due to the impact of adding a safety factor on path planning, it is necessary to establish new evaluation criteria. Based on the safety heuristic function, determine the path safety evaluation index function:

[0116]

[0117] Among them, S i (t) represents the safety factor of the i-th grid on the current path, and n represents the number of all grids on the current search path;

[0118] Before path planning, a path safety evaluation index V is set. Paths with a V value greater than the preset value are considered feasible. For each path in the ant search, only the pheromones on paths that meet the V evaluation criteria are updated to ensure that the optimal path meets the evaluation criteria.

[0119] Steps 3-5: Determine the transition rules for the improved ant colony algorithm:

[0120] Assuming the ant is in the i-th grid cell, the transfer rule for moving from grid cell i to the candidate grid cell j at time t is as follows:

[0121]

[0122] Where, τ j (t) represents the pheromone concentration of the candidate grid, where α represents the weight of the candidate grid pheromone concentration in the transition rule. A larger value indicates that the ant relies more on pheromones to select a waypoint. β represents the weight of the visibility heuristic function in the transition rule; a larger value indicates that the ant is more likely to choose a candidate node closer to the target point. λ represents the weight of the height heuristic function in the transition rule; a larger value indicates that the ant is more likely to choose a candidate node with a smaller height difference from the target point. ω represents the weight of the safety heuristic function in the transition rule; a larger value indicates a greater tendency to choose a node with fewer surrounding obstacles.

[0123] The improved visibility heuristic function, height heuristic function, and safety heuristic function work together to influence the ant's choice of nodes, guiding it towards the target node. The heuristic function guarantees the mechanism by which the ant moves towards the target node, while the heuristic factor indicates the importance of the heuristic function in the entire equation.

[0124] Step 4, Combining Figure 4 Based on the improved ant colony algorithm and the safe driving corridor generated in step 2, a route search is performed to obtain the optimal safe path:

[0125] Step 4-1: Based on the safe driving corridor determined in Step 2, determine the robot's starting point and inspection target point, initialize the relevant parameters of the ant colony algorithm, and set the threshold of the path safety evaluation index V.

[0126] Step 4-2: Place all ants at the starting point and build a table to store the path nodes that the ants have already visited. If the next node to be visited by the ant already exists in this table, then the node is marked as a node that cannot be visited again.

[0127] Step 4-3: Based on the improved transfer rules determined in Step 3, determine the next path node that the ant should go to, and put the visited nodes into the taboo list. When the ant reaches the target node, a search is completed.

[0128] Step 4-4: For each path in the ant search, only update the pheromone on the path where the path safety evaluation index V is greater than the set threshold to ensure that the optimal path meets the evaluation criteria. Compare the optimal paths in each iteration to find the current optimal path and determine whether the set termination condition is met. If it is met, output the optimal safe path; otherwise, return to step 4-2 and continue to the next iteration until the termination condition is met.

[0129] Steps 4-5: After arriving at the inspection point, calculate the distance between the robot and the inspection target. Once the inspection threshold is reached, perform the inspection work. After the inspection is completed, execute the path planning from the current point to the next inspection point.

[0130] The above embodiments illustrate and describe the basic principles and main features of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the present invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed.

Claims

1. A high-precision safe path planning method for inspection robots in chemical industry scenarios, characterized in that, Includes the following steps: Step 1: Construct a 2D raster map based on laser point cloud data, and perform puffing processing on obstacles in the map: Step 1-1: Obtain laser point cloud data and determine the degree of influence of laser on the grid according to the two-dimensional normal distribution formula. Obtain the influence degree of all laser points in each grid, remove point cloud data with influence degree less than the set threshold, and store the remaining point cloud data into the grid. Steps 1-2: Based on the distribution of laser point clouds in the vertical direction within the grid, determine the average height of each grid, complete the matching of laser point clouds and grids, and obtain a two-dimensional grid map containing the height information of each grid. Steps 1-3: Construct a hazard function to evaluate the collision risk level of the robot's path: ； ； in, The number of obstacles within the affected area. It is a path point To the obstacle The straight-line distance It refers to the range of influence of the obstacle, which is related to the size of the robot. These are coefficients that are proportional to the robot's velocity vector and the material of the obstacle, respectively, and determine the degree of danger posed to the robot by the collision; Steps 1-4: Extract obstacle information from the 2D grid map and puff the edges of the obstacles: Expand the obstacle outward along its outermost contour by L: ； ； In the formula, It is the maximum radius of the robot's dimensions. It refers to the degree of collision risk along the robot's path. It's the robot's speed. This indicates the robot's current task completion status; Representing path points The shortest distance to the target point Indicates the inspection range of the target point; Step 2: Generate a safe driving corridor for the robot based on a 2D grid map and expanded obstacles; Step 3: Improve the ant colony algorithm by introducing a height heuristic function, a visibility heuristic function, and a spatial safety factor to improve the transition rules of the ant colony algorithm; Step 4: Based on the improved ant colony algorithm and the safe driving corridor generated in Step 2, perform route search to obtain the optimal safe path.

2. The high-precision safety path planning method for inspection robots in chemical industry scenarios according to claim 1, characterized in that, The generation of the safe driving corridor in step 2 specifically involves: Step 2-1: Based on the two-dimensional grid map, perform A* search on the rear axis center of the robot to obtain the coarse trajectory of the rear axis center, that is, determine the coarse trajectory of the mass point, and obtain the total time m of the path planning completed by A* search. Step 2-2: Using the local rectangular region generation method, uniformly sample m to extract uniform trajectory points. Expand each sampling point outward in a circle. Determine temporary boundaries based on collision detection between the corresponding circle and obstacles. Expand outward continuously along the four sides of the rectangle in either counterclockwise or clockwise directions based on the trajectory points. Units: ； Until the rectangle overlaps with an obstacle, the length and width of the rectangle do not exceed the set threshold S; Step 2-3: Connect the intersection of the temporary boundary points formed in Step 2-2 with the rectangles on multiple trajectory points to form a safe driving corridor.

3. The high-precision safety path planning method for inspection robots in chemical industry scenarios according to claim 1, characterized in that, The improvement of the ant colony algorithm's transfer rules in step 3 specifically includes: Step 3-1: Introduce a height-based heuristic function: ； ； in, Here, is the height strength coefficient, and is a constant. Indicates the height of the grid cells to be selected. This represents the updated grid height. Indicates the height of the target point; Step 3-2: Introduce the visibility heuristic function: ； ； in, It is a path point The straight-line distance to the target point; Step 3-3: Introduce a security heuristic function: ； ； Where M represents the number of obstacles at the current location; Steps 3-4: Determine the path security evaluation index function based on the security heuristic function: ； in, This represents the safety factor of the i-th grid cell on the current path; Steps 3-5: Determine the transition rules for the improved ant colony algorithm.

4. The high-precision safety path planning method for inspection robots in chemical industry scenarios according to claim 3, characterized in that, The search in step 4 yields the optimal safe path, specifically as follows: Step 4-1: Based on the safe driving corridor determined in Step 2, determine the robot's starting point and inspection target point, initialize the relevant parameters of the ant colony algorithm, and set the threshold of the path safety evaluation index V. Step 4-2: Place all ants at the starting point and build a table to store the path nodes that the ants have already visited. If the next node to be visited by the ant already exists in this table, then the node is marked as a node that cannot be visited again. Step 4-3: Based on the improved transfer rules determined in Step 3, determine the next path node that the ant should go to, and put the visited nodes into the taboo list. When the ant reaches the target node, a search is completed. Step 4-4: For each path in the ant search, only update the pheromone on the path where the path safety evaluation index V is greater than the set threshold to ensure that the optimal path meets the evaluation criteria. Compare the optimal paths in each iteration to find the current optimal path and determine whether the set termination condition is met. If it is met, output the optimal safe path. Otherwise, return to step 4-2 and continue the next iteration until the termination condition is met; Steps 4-5: After arriving at the inspection point, calculate the distance between the robot and the inspection target. Once the inspection threshold is reached, perform the inspection work. After the inspection is completed, execute the path planning from the current point to the next inspection point.

5. A high-precision safety path planning system for inspection robots in chemical industry scenarios, characterized in that, Includes the following modules: Map building module: Used to construct a 2D raster map based on laser point cloud data, and to perform obstacle expansion processing on the map, including: Acquire laser point cloud data and determine the degree of laser influence on the grid according to the two-dimensional normal distribution formula. Obtain the influence degree of all laser points in each grid, remove point cloud data with influence degree less than the set threshold, and store the remaining point cloud data into the grid. Based on the distribution of laser point clouds in the vertical direction within the grid, the average height of each grid is determined, and the laser point cloud is matched with the grid to obtain a two-dimensional grid map containing the height information of each grid. Construct a hazard function to evaluate the collision risk level of the robot's path: ； ； in, The number of obstacles within the affected area. It is a path point To the obstacle The straight-line distance It refers to the range of influence of the obstacle, which is related to the size of the robot. These are coefficients that are proportional to the robot's velocity vector and the material of the obstacle, respectively, and determine the degree of danger posed to the robot by the collision; Extract obstacle information from the 2D raster map and puff the obstacle edges: Expand the obstacle outward along its outermost contour by L: ； ； In the formula, It is the maximum radius of the robot's dimensions. It refers to the degree of collision risk along the robot's path. It's the robot's speed. This indicates the robot's current task completion status; Representing path points The shortest distance to the target point Indicates the inspection range of the target point; Safe Driving Corridor Generation Module: Used to generate safe driving corridors for robots based on two-dimensional grid maps and expanded obstacles; Transfer rule determination module: Used to improve the ant colony algorithm by introducing height heuristic function, visibility heuristic function and spatial safety coefficient to improve the transfer rules of the ant colony algorithm; Optimal safe path generation module: used to search for routes based on the improved ant colony algorithm and the generated safe driving corridors to obtain the optimal safe path.

6. A computer apparatus comprising a memory, a processor, and a computer program stored in the memory and operating on the processor, characterized in that, When the processor executes the computer program, it performs the method steps as specified in any one of claims 1 to 4 above.

7. A computer-storeable medium storing a computer program, characterized in that, The process steps as defined in any one of claims 1 to 4 are performed by a processor on the computer program.

Images