A method for anti-entanglement path planning of an oil storage tank operation robot

Abstract

A method for anti-entanglement path planning for an oil tank operation robot belongs to the field of oil tank operation technology. Its features include the following steps: Step a, establishing an electronic map of the environment; Step b, dividing the work area; Step c, planning the work path; Step d, identifying obstacles during movement; Step e, constructing an obstacle envelope circle; Step f, the robot performing obstacle avoidance maneuvers around the obstacle envelope circle; Step g, replanning the work path; Step h, the robot performing the operation. This method for anti-entanglement path planning for an oil tank operation robot adopts a zoned and timed operation approach to avoid cable entanglement between the detection and cleaning robots or entanglement between the robot and drainage pipes during operation. Simultaneously, while implementing local path planning, the robot's obstacle departure position is redefined, reducing redundant paths generated during obstacle avoidance and improving robot detection and cleaning efficiency.

Description

Technical Field

[0001] An anti-entanglement path planning method for an oil storage tank operation robot belongs to the field of oil storage tank operation technology. Background Technology

[0002] Oil storage tanks are extremely common storage devices in the petroleum and petrochemical industry. Due to the complexity of the internal medium composition, the tank bottom plate is highly susceptible to corrosion. Its condition is a key factor determining the tank's opening cycle, making its inspection and evaluation crucial. However, in practical applications, the tank bottom plate is often covered by the stored medium, making inspection difficult using traditional techniques. This necessitates shutdown and tank opening for maintenance, which is not only time-consuming and labor-intensive but also costly.

[0003] Traditional methods for inspecting and cleaning the bottom plates of oil storage tanks mainly rely on manual labor. This method suffers from problems such as large human errors, low efficiency, and high safety risks. To address these issues, robotics technology has been introduced into oil storage tank bottom plate inspection. However, due to the complex environment at the bottom of oil storage tanks, in addition to the known and unknown anode blocks located there, obstacles in unpredictable locations often appear. Furthermore, obstacle avoidance is far more difficult than normal movement, resulting in low overall robot efficiency in actual operations. In existing technologies:

[0004] (1) The technical solution disclosed in the literature "Ding Peng. (2021). Research on Mobile Robot Path Planning Based on Iterative A* Algorithm and Bug Algorithm (Master's Thesis, Northwest Normal University)" adopts a scheme combining the Bug algorithm and the A* algorithm. That is, when the robot moves towards the target position, it first determines whether the robot has encountered an obstacle. When the robot encounters an obstacle, the improved A* algorithm is used to search for the path, and when the robot does not encounter an obstacle, the Bug2 algorithm is used to search for the path. Although this method has better obstacle avoidance performance, this path planning method is based on global planning, that is, all environmental information in the working environment must be known in advance, which has great limitations and is difficult to implement in many uncertain working environments.

[0005] (2) In the technical solution disclosed in the literature "Yang Yang, Wen Xing, Ma Qianglong, Zhang Gang, Cheng Shangkun, Qi Jian, Chen Zhizhen, Chen Liqing. Real-time planning of obstacle avoidance path for agricultural machinery in dynamic identification area based on Bézier curve [J]. Transactions of the Chinese Society of Agricultural Engineering, 2022, 38(06):34-43", the agricultural machinery operation and walking dynamic identification area is first constructed, and obstacles are perceived in the dynamic identification area using lidar. Then, the obstacle avoidance path control point selection range is calculated using obstacle information, and then Bézier curve fitting is performed using control points to generate a path cluster that meets multiple constraints such as the minimum turning radius of the agricultural machinery. At the same time, the optimal obstacle avoidance path is selected from the path cluster with the goal of minimizing curvature.

[0006] Although the dynamic constraints are satisfied and the generated path has high smoothness, redundant paths are often created in local obstacles in most cases. This prevents timely adjustments to direction or planning of new routes, thus hindering path planning efficiency. In contrast, this invention redefines the obstacle departure position based on the original Bug obstacle avoidance algorithm, thereby avoiding redundant paths.

[0007] (3) In the technical solution disclosed in the literature “Dai Fengyan, Gao Qingshan, Chen Jiaqing, Guo Lifeng. Research on full-coverage path planning of oil tank cleaning robot [J]. Mechanical Design and Manufacturing, 2020(02):263-266”, an improved Bug path planning algorithm suitable for internal floating roof oil tanks is proposed based on a full-coverage tank bottom method to address the problem that the oil tank cleaning robot has a high overlap rate due to obstacles inside the tank during cleaning operations and cannot guarantee the cleaning coverage area. First, the cleaning point coordinates are obtained by using the cleaning principle of full-coverage tank bottom with a regular hexagonal topology structure. Second, the cleaning environment is modeled using the coordinate method. The positional relationship information between the robot cleaning target point and the tank bottom obstacle is added to the straight-line walking condition of the traditional Bug algorithm to optimize the cleaning point coordinate position and plan an obstacle-free straight-line walking path from the robot starting point to each cleaning point. However, the purpose of this improved method is to make the obstacle choose an obstacle-free path from the beginning. It can be said that the obstacle avoidance process is not reflected. The algorithm may be more effective in environments with few obstacles, but if the working scenario is more complex, the algorithm is prone to getting stuck in an infinite loop.

[0008] (4) In the technical solution disclosed in the literature “Zha Rongrui, Ma Yunhua, Yan Xiang, Zheng Shuang. Mobile robot obstacle avoidance based on scene understanding and improved BUG algorithm [J]. Computer Measurement & Control, 2023, 31(03):228-234”, in view of the limitations of existing mobile robots in visual obstacle avoidance, a robot obstacle avoidance method based on deep convolutional neural network and improved Bug algorithm is proposed by combining deep learning algorithm and path planning technology. The method uses multi-task deep convolutional neural network to extract road image features to realize image classification and semantic segmentation tasks. Secondly, a grid map is constructed based on the semantic segmentation results, and the image classification results are combined with the improved Bug algorithm to search for the optimal obstacle avoidance path. At the same time, in order to reduce redundant calculations, a feature comparison structure is designed to avoid redundant calculations of feature information and ensure the real-time performance of the robot in practical applications.

[0009] While this method effectively balances accuracy and efficiency in visual tasks, accurately planning safe obstacle avoidance paths and assisting robots in navigation and obstacle avoidance, its algorithm employs a multi-task deep convolutional neural network, requiring substantial computational resources. This results in poor real-time performance, making it difficult to apply to real-time robot obstacle avoidance scenarios. Furthermore, due to the high sensitivity of artificial intelligence algorithms to data, changes in lighting, weather, or other environmental conditions necessitate network retraining, further limiting the algorithm's practicality. Since deep learning algorithms require vast amounts of data to identify and classify objects, this method may struggle with recognizing small obstacles, making it difficult to avoid them.

[0010] (5) A technical solution is disclosed in Chinese invention patent application number 201810492962.1, application date May 4, 2021, entitled “A Dynamic Path Planning Method for Robot Combining A* Algorithm and VFH Obstacle Avoidance Algorithm”. In this technical solution, the environment in which the robot is located is represented as a grid map, and an algorithm is used to search for a global initial path in the grid map. The robot moves along the initial path towards the target point and determines whether it encounters an obstacle. If it does, the VFH algorithm is used to avoid the obstacle; otherwise, it continues to move towards the target point.

[0011] In this technical solution, the combination of A* algorithm and VFH algorithm is used for dynamic obstacle avoidance, which can be applied to most scenarios. However, the VFH algorithm is prone to dead zone phenomenon in narrow areas. This is because the VFH algorithm is an artificial potential field method. When there are obstacles on both sides, if the resistance force is the same, a dead zone phenomenon will occur. Summary of the Invention

[0012] The technical problem this invention aims to solve is to overcome the shortcomings of existing technologies and provide an anti-entanglement path planning method for oil tank operation robots. This method employs a zoned and time-based operation approach to avoid cable entanglement or robot entanglement with drainage pipes during operation. Simultaneously, while implementing local path planning, it redefines the robot's obstacle departure position, reducing redundant paths generated during obstacle avoidance and improving the robot's detection and cleaning efficiency.

[0013] The technical solution adopted by this invention to solve its technical problem is: an anti-entanglement path planning method for an oil storage tank operation robot, characterized by the following steps:

[0014] Step a, create an electronic map of the environment;

[0015] Step b: Based on the number of manholes on the top of the oil storage tank, divide the bottom of the oil storage tank into working areas and determine the order in which the robot will work in the working areas.

[0016] Step c: For each work area, determine its initial start point and end point, and use the line connecting the initial start point and end point as the work path according to the Bug algorithm.

[0017] Step d: When the robot is moving along the work path, it moves along the path at a set speed and checks whether it encounters any obstacles during the movement. If there are obstacles, proceed to step e; otherwise, continue moving along the path.

[0018] Step e: After encountering an obstacle, the robot records the speed at which it detected the obstacle the moment before as its initial speed, and records the value of the initial speed and the direction of its movement vector, while constructing the obstacle's envelope circle.

[0019] Step f: After constructing the obstacle envelope circle, the robot performs obstacle avoidance maneuvers around the obstacle envelope circle.

[0020] Step g: Determine the end point of the obstacle avoidance action based on the speed during the obstacle avoidance process, replan the operation path, and the robot moves along the replanned operation path;

[0021] Step h: After the robot reaches the end of the work path, it performs the work.

[0022] Preferably, in step b, the number of manholes is N and N≥2, and the rule for dividing the work area is as follows:

[0023] When N is an even number: the bottom of the oil storage tank is evenly divided into N or 2N sector-shaped working areas. When the number of sector-shaped areas is N, each manhole is located directly above its corresponding working area; when the number of sector-shaped areas is 2N, each manhole is located directly above the middle of two adjacent working areas.

[0024] When N is an odd number, the bottom of the oil storage tank is evenly divided into 2N sector-shaped working areas, and at least one manhole is located directly above any one of the working areas.

[0025] Preferably, in step b, the robot includes a cleaning robot and an inspection robot that are successively lowered to the bottom of the tank.

[0026] For the cleaning robot: the lowered work area is the last work area to be worked on, and the work area opposite to the lowered work area is the first work area to be worked on. Starting from the lower area, the robot cycles through all work areas at the bottom of the tank in the order of relative work area - adjacent work area, and then transfers the oil storage tank out from the manhole where it was initially lowered.

[0027] For the inspection robot: When the cleaning robot starts cleaning the second work area, the inspection operation is carried out in the same order as the cleaning robot in each work area that has been cleaned. Before the inspection operation of the next work area, the oil tank is moved out and lowered from the manhole closest to the next work area to be inspected.

[0028] Preferably, in step e, constructing the obstacle envelope circle includes the following steps:

[0029] Step e-1: The robot uses its onboard ultrasonic sensors to perceive obstacle feature points and feature edge information.

[0030] Step e-2: Construct the circumcircle of the obstacle using the points and edges, i.e., the obstacle envelope circle;

[0031] Step e-3: After completing the outer contour fitting of the envelope circle, the coordinate system of the envelope circle is transformed, and the obstacle envelope circle is transformed from the tank bottom coordinate system to the robot motion coordinate system.

[0032] Step e-4: Determine the coordinates of the center of the obstacle envelope circle in the robot's motion coordinate system.

[0033] Preferably, in step g, during the robot's obstacle avoidance action, when the real-time motion vector direction is the same as the initial velocity vector direction, the obstacle avoidance action is stopped, and the endpoint of the stopped obstacle avoidance action is taken as the end point.

[0034] Preferably, step h includes the following steps:

[0035] Step h-1: Rectify each work point that needs to be worked within the work area, so that the entire work point is located within its circumscribed rectangle, and then rasterize the circumscribed rectangle.

[0036] Step h-2: Set the detection priority for the grid cells within the detection area;

[0037] Step h-3: Assign a value to the current grid cell. After completing the task in the current grid cell, the robot moves to the grid cell with the highest priority according to the priority order, sets the next grid cell as the current position, and assigns a value.

[0038] In step h-4, the robot determines whether all grids have been traversed. If all grids have been traversed, it moves to the next work point in the work area to perform the work, or moves to another work area to perform the work; otherwise, it returns to step h-3.

[0039] Preferably, in step h-1, a set S = {(x, y) | x, y are points within the area to be detected} is defined, and the value of each grid cell is defined.

[0040]

[0041] And set the grid where the robot is currently located as the starting grid;

[0042] In step h-3, the current grid is assigned a value of 0.

[0043] Preferably, in step h-2, the operation is specified to start from the leftmost edge of the raster map, with priorities P in sequence. x-1，y Left grid, P x，y+1 Top grid, P x+1，y Right grid, P x，y-1 Bottom grid.

[0044] Compared with the prior art, the beneficial effects of this invention are:

[0045] In this anti-entanglement path planning method for the oil storage tank operation robot, a zoned and timed operation approach is adopted to avoid cable entanglement between the detection and cleaning robots or entanglement between the robot and the drainage pipe during operation. Simultaneously, while implementing local path planning, the robot's obstacle departure position is redefined, reducing redundant paths generated during obstacle avoidance and improving the robot's detection and cleaning efficiency.

[0046] In this anti-entanglement path planning method for the oil storage tank operation robot, an improvement is made to the traditional Bug obstacle avoidance algorithm to generate a path planning method that can cope with the special working scenario of oil storage tanks in an oil environment. Because the working environment contains oil sludge, which can easily cause the robot to slip and get stuck, this solution uses an obstacle generation envelope circle method to redesign obstacles based on the small size of obstacles at the bottom of the oil tank, improving the smoothness and continuity of the generated path.

[0047] This path planning method possesses the flexible obstacle avoidance capabilities of the traditional Bug algorithm, and the planned path conforms to dynamic constraints. Furthermore, it redefines the robot's departure point from obstacles based on the original Bug obstacle avoidance algorithm, reducing local redundant paths generated during obstacle avoidance, improving the efficiency of path planning, and optimizing the operation.

[0048] In the anti-entanglement path planning method of this oil storage tank operation robot, the scheme of adding envelope circle constraints to obstacles and redefining the end position of obstacle avoidance behavior using the vector method is adopted. This eliminates sharp inflection points in the path generated by the mobile robot, makes the path more in line with the robot's dynamic constraints, reduces the generation of a large number of redundant paths, and greatly improves the path planning efficiency. At the same time, the effect of saving paths will increase with the increase in the number and shape complexity of obstacles.

[0049] Furthermore, obstacles are represented as circles on the electronic map, resulting in a smooth curve that conforms to the robot's dynamic constraints, preventing slippage in the mud and ensuring the robot's stable movement.

[0050] In this anti-entanglement path planning method for the oil storage tank operation robot, the robot's tasks include cleaning and inspecting the tank bottom plate. This requires particularly precise division of areas and time. This solution combines area division with the robot's "clean first, then inspect" approach, addressing both time and space aspects. This effectively prevents the robot from cable entanglement while ensuring it completes its tasks without omissions, greatly improving work efficiency and reducing unnecessary losses. Attached Figure Description

[0051] Figure 1 A flowchart of the anti-entanglement path planning method for oil storage tank operation robots.

[0052] Figure 2 This is a coordinate transformation diagram for the anti-entanglement path planning method of an oil storage tank operation robot.

[0053] Figure 3 A schematic diagram of an obstacle avoidance robot.

[0054] Figure 4 A schematic diagram of a work robot continuously avoiding obstacles.

[0055] Figure 5 A schematic diagram of a traditional bug avoidance algorithm.

[0056] Figure 6 Flowchart of the task traversal algorithm for the robot.

[0057] Figure 7 This is a schematic diagram of the area to be detected. Detailed Implementation

[0058] Figures 1-7 This is the preferred embodiment of the present invention, which is described below in conjunction with the accompanying drawings. Figures 1-7 The present invention will be further described below.

[0059] like Figure 1 As shown, an anti-entanglement path planning method for an oil storage tank operation robot includes the following steps:

[0060] Step 1001: Create an electronic map of the tank bottom environment;

[0061] Using the tank drawings, a tank model is created, ultimately generating an electronic map. This map includes markings of tank bottom components, internal pipes, and sacrificial anode blocks. Key markings include the location of obstacles encountered by the robot during operation, weld seams, and areas where sludge accumulates.

[0062] Step 1002: Divide the work area;

[0063] In traditional oil storage tanks, there are generally 1 to 4 manholes at the top of the tank. In the anti-entanglement path planning method of this oil storage tank operation robot, for oil storage tanks with 2 or more manholes, the number of manholes is denoted as N.

[0064] When N ≥ 2 and N is even, the bottom of the oil storage tank is evenly divided into N or 2N sector-shaped working areas. When the number of sector-shaped areas is N, each manhole is located directly above its corresponding working area. When the number of sector-shaped areas is 2N, each manhole is located directly above the middle of two adjacent working areas. When N ≥ 2 and N is odd, the bottom of the oil storage tank is evenly divided into 2N sector-shaped working areas. At least one manhole is located directly above any one of the working areas.

[0065] The order in which the robots operate within their respective work areas is determined: In this anti-entanglement path planning method for the oil storage tank operation robots, the cleaning robot and the inspection robot operate sequentially within their respective work areas, adopting a "clean first, then inspect" approach. Furthermore, to further ensure the cleaning and inspection robots achieve an anti-entanglement effect when operating at the bottom of the tank, the strategy for their operation is as follows:

[0066] The cleaning robot is first lowered to the bottom of the tank through one of the manholes. It then performs cleaning operations in the designated area within the current work zone. After completing its cleaning task, the robot moves to an adjacent work zone, and so on, in a sequential order of relative-adjacent-relative-adjacent. Finally, it cleans the initial work zone below the tank. Once all work zones are cleaned, the robot is removed from the tank through the manhole below.

[0067] For the inspection robot, when the cleaning robot starts cleaning the second work area, the inspection work is carried out in the work areas that have been cleaned in sequence according to the work order of the cleaning robot. The difference is that before the inspection robot starts the inspection work in the next work area, it is lifted out of the oil storage tank and lowered from the manhole closest to the next work area to be inspected.

[0068] In this anti-entanglement path planning method for the oil storage tank operation robot, the robot's tasks include cleaning and inspecting the tank bottom plate. This requires particularly precise division of areas and time. This solution combines area division with the robot's "clean first, then inspect" approach, addressing both time and space aspects. This effectively prevents the robot from cable entanglement while ensuring it completes its tasks without omissions, greatly improving work efficiency and reducing unnecessary losses.

[0069] Step 1003: Determine the initial work path for the work area;

[0070] For each work area, the initial start and end points of the operation are determined. The robot's work environment is represented as an electronic grid map, using the area shape and known information as the main elements. A coordinate system XOY is established with the starting point as the center, and a coordinate system X'-O'-Y' is established with the robot's center O' and the horizontal and vertical directions to follow the robot's movement. The unit length of the coordinate axes is defined. The line connecting the initial start and end points is used as the initial work path according to the Bug algorithm. The center of the bottom of the tank is preferably used as the initial start point for all work areas.

[0071] Step 1004: The robot moves along the work path;

[0072] The robots include cleaning robots and inspection robots. After the cleaning robot or inspection robot is lowered to the bottom of the tank through the manhole, the cleaning robot or inspection robot travels along the planned work path to the destination within the determined work area.

[0073] In the anti-entanglement path planning method of the oil storage tank operation robot, since the cleaning robot and the inspection robot are lowered into the oil storage tank one after another to carry out the operation, the cleaning robot and the inspection robot perform corresponding actions according to steps 1004 to 1014 respectively during the movement.

[0074] Step 1005: Determine if an obstacle has been encountered;

[0075] When the robot (including the cleaning robot and the inspection robot) moves along the work path, it moves along the initial path at a set speed. Because there is no light in the oil work environment, ultrasonic sensors are used to determine whether obstacles are encountered during the movement. If the robot detects an obstacle in the work path, it proceeds to step 1008; otherwise, it proceeds to step 1006.

[0076] Step 1006, record the initial velocity;

[0077] The robot records its initial velocity as the speed at which it traveled before detecting the obstacle, and also records the value of the initial velocity and the direction of its travel vector.

[0078] Step 1007: Construct the obstacle envelope circle;

[0079] Considering the oil sludge at the bottom of the oil storage tank and the robot's dynamic constraints, and to avoid sharp inflection points in the robot's trajectory—both adhering to the dynamic constraints and preventing slippage and sludge accumulation—and taking advantage of the small size of obstacles at the bottom of the oil tank, ultrasonic sensors are used to detect the feature points and edges of obstacles when they encounter them. The process of constructing the obstacle's envelope circle and determining its position in the coordinate system includes the following steps:

[0080] Step 1007-1: The robot uses its onboard ultrasonic sensors to perceive obstacle feature points and feature edge information.

[0081] Step 1007-2: Construct the circumcircle of the obstacle using points and edges, i.e., the obstacle envelope circle;

[0082] Step 1007-3: After completing the fitting of the outer contour of the envelope circle, obtain the center coordinates and radius of the outer contour envelope circle of the obstacle in the X'-O'-Y' coordinate system;

[0083] Step 1007-4: Perform a coordinate transformation from the X'-O'-Y' coordinate system to the XOY coordinate system on the grid map for the center coordinates of the obstacle's envelope circle. The coordinate transformation formula is as follows:

[0084]

[0085] x'=Rcos(lat p )cos(lon p )+lsinθ1+x s sinθ1+y s sinθ1

[0086] y'=Rcos(lat p sin(lon) p )+lcosθ1-x s sinθ1+y s cosθ1

[0087] Transformation principle as follows Figure 2 As shown, point P is the robot ( Figure 2 The center point is indicated by the arrow c. L is the distance between the LiDAR position and the robot's center position. α is the angle between the robot's velocity direction and the Y-axis of the XOY coordinate system. β is the angle between this direction and the positive X' direction of the X'O'Y' coordinate system. S is the point to be converted in the conversion model, i.e., the identified obstacle. Figure 2 The location is indicated by the middle arrow (a). The distance from S to O' is defined as d. The latitude and longitude coordinates of point S and the included angle α are obtained through positioning equipment.s The longitude of the center point of the obstacle, lat s The latitude of the center point of the obstacle, lon p The longitude of the robot's center point, lat p The coordinates of the robot's center point are represented by x' and y', respectively, which represent the coordinates of the obstacle envelope circle in the X'-O'-Y' coordinate system. s y s Let X and Y represent the coordinates of the obstacle in the XOY coordinate system, respectively. R represents the Earth's radius, and θ1 represents the angle between the due north direction and the robot's velocity direction.

[0088] Step 1007-5: Obtain the obstacle envelope circle using the conversion formula from step 1007-4. Figure 2 The latitude and longitude coordinates of the point (point b in the middle arrow) are used to determine its rectangular coordinates in the raster electronic map, and then combined with the latitude and longitude coordinates of the XOY origin.

[0089] Step 1008: The robot performs obstacle avoidance maneuvers around the envelope of the obstacle.

[0090] After determining the location and size of the obstacle envelope, the robot moves around the circumference of the envelope to perform obstacle avoidance maneuvers when it reaches the obstacle envelope.

[0091] Let q be the start and end points of the initial job path. start0 and q goal In robot q start0 →q goal During this movement, the robot's starting coordinates are defined as (X... r Y r The endpoint coordinates are (X) g Y g At this point, the equations of motion for the robot can be determined using the two-point equation:

[0092]

[0093] Simultaneously, according to the Lagrange method, the center of the robot operating in the oil is treated as a fluid particle, and its motion law is expressed in the following vector form:

[0094] in, It is the radius vector of the robot as it moves within the electronic grid map, and the variable X is... r Y r t is a Lagrange variable, and the starting coordinate is (X... r Y r When the robot encounters an obstacle and performs obstacle avoidance behavior, the position of the obstacle detected by the robot detector is defined as q. H1. Record the velocity vector at this moment, and use the vector method to find the partial derivatives of the equations of motion:

[0095]

[0096] Define its velocity in this coordinate system at this moment.

[0097] Step 1009: Record the travel speed during obstacle avoidance;

[0098] During the robot's obstacle avoidance maneuver around the envelope of obstacles, by

[0099]

[0100] Update robot movement real-time speed

[0101] Step 1010: The robot determines whether the vector direction of the current travel speed is the same as that of the initial speed;

[0102] As can be seen from the above, the robot updates its movement speed during the movement process. When the outer product of the two velocities is zero, that is... If the vector direction of the velocity is the same as the vector direction of the initial velocity recorded in step 1006, stop the obstacle avoidance movement and execute step 1011; otherwise, return to step 1009.

[0103] Step 1011: Redetermine the job path;

[0104] When the robot performs obstacle avoidance maneuvers around the envelope circle, its position when it stops the obstacle avoidance movement is denoted as q. start1 Its coordinates are (X r1 Y r1 The location is then used as the new starting point for the operation, and the operation path is replanned. Then, the process returns to step 1004 and proceeds along the replanned path.

[0105] like Figure 5 As shown, the obstacle avoidance behavior of the traditional BUG algorithm is to first plan the path from the initial point to the destination, and the route is a two-point system. To determine this, when encountering obstacles in the image, traditional algorithms will perform a circular motion during obstacle avoidance, and are prone to getting stuck in local errors when encountering some irregular obstacles.

[0106] like Figures 3-4As shown, in the anti-entanglement path planning method of this oil tank operation robot, an envelope circle constraint is added to the obstacle. The position to stop the obstacle avoidance action is determined by comparing the vector direction of the action velocity during the obstacle avoidance action with the vector direction of the initial velocity. Compared with the traditional BUG algorithm, where the obstacle avoidance behavior only stops when the mobile robot reaches the intersection of the initially two-point determined path and the obstacle edge, the saved path planning distance is:

[0107]

[0108] Where r is the radius of the obstacle envelope circle.

[0109] Therefore, in the anti-entanglement path planning method of this oil storage tank operation robot, the scheme of adding envelope circle constraints to obstacles and redefining the end position of obstacle avoidance behavior using the vector method is adopted. This eliminates sharp inflection points in the path generated by the mobile robot, makes the path more in line with the robot's dynamic constraints, reduces the generation of a large number of redundant paths, greatly improves the path planning efficiency, and the effect of saving paths will increase with the increase in the number and shape complexity of obstacles.

[0110] Furthermore, obstacles are represented as circles on the electronic map, resulting in a smooth curve that conforms to the robot's dynamic constraints, preventing slippage in the mud and ensuring the robot's stable movement.

[0111] Step 1012: Determine if the destination has been reached;

[0112] The robot determines whether it has reached the end of the work path. If it has reached the end of the work path, it proceeds to step 1013. If it has not yet reached the end of the work path, it returns to step 1004 and continues to travel along the initial work path or the path replanned after encountering a fault.

[0113] Step 1013: Begin executing the job;

[0114] After reaching the end of the work path, the robot begins to perform the corresponding cleaning or inspection tasks, such as... Figure 6 As shown, it includes the following steps:

[0115] Step 1013-1: Grid the work area;

[0116] Combination Figure 7 The process involves rectangularizing each work point within the work area, ensuring that all work points are contained within their circumscribed rectangles. Then, these circumscribed rectangles are rasterized. First, a set S = {(x, y) | x, y are points within the area to be detected} is defined, along with the value of each raster cell.

[0117]

[0118] Then set the grid where the robot is currently located as the starting grid.

[0119] Step 1013-2: Plan the priority order of the grid;

[0120] For each grid cell within the area to be inspected, a detection priority is set, and the inspection operation is specified to start from the leftmost edge of the grid map, with a priority of P. x-1，y (Left grid) Maximum, P x，y+1 (Top grid) Next, then P x+1，y (Right grid), P x，y-1 (Bottom grid).

[0121] Step 1013-3: Assign values ​​to the grid;

[0122] The robot sets the current grid cell to 0. After completing the task in the current grid cell, it moves to the grid cell with the highest priority according to the priority order, sets the next grid cell as the current position, and assigns a value.

[0123] Step 1013-4: Process all grids;

[0124] The robot determines whether all grids have been traversed. If all grids have been traversed, it proceeds to step 1013-5; otherwise, it returns to step 1013-3.

[0125] Step 1013-5, End.

[0126] The robot has completed the task at the current work point within the work area and moves to the next work point within the work area to perform the task, or moves to another work area to perform the task.

[0127] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An anti-entanglement path planning method for an oil tank operating robot, characterized by: Includes the following steps: Step a, create an electronic map of the environment; Step b: Based on the number of manholes on the top of the oil storage tank, divide the bottom of the oil storage tank into working areas and determine the order in which the robot will work in the working areas. Step c: For each work area, determine its initial start point and end point, and use the line connecting the initial start point and end point as the work path according to the Bug algorithm. Step d: When the robot is moving along the work path, it moves along the path at a set speed and checks whether it encounters any obstacles during the movement. If there are obstacles, proceed to step e; otherwise, continue moving along the path. Step e: After encountering an obstacle, the robot records the speed at which it detected the obstacle the moment before as its initial speed, and records the value of the initial speed and the direction of its movement vector, while constructing the obstacle's envelope circle. Step f: After constructing the obstacle envelope circle, the robot performs obstacle avoidance maneuvers around the obstacle envelope circle. Step g: Determine the end point of the obstacle avoidance action based on the speed during the obstacle avoidance process, replan the operation path, and the robot moves along the replanned operation path; Step h: After the robot reaches the end of the work path, it performs the work. In step b, the number of manholes is N and N≥2, and the rules for dividing the work area are as follows: When N is an even number: the bottom of the oil storage tank is evenly divided into N or 2N sector-shaped working areas. When the number of sector-shaped areas is N, each manhole is located directly above its corresponding working area. When the number of sector-shaped areas is 2N, each manhole is located directly above the middle of two adjacent working areas. When N is an odd number, the bottom of the oil storage tank is evenly divided into 2N sector-shaped working areas, and at least one manhole is located directly above any one of the working areas. In step b, the robots include a cleaning robot and an inspection robot that are lowered sequentially to the bottom of the tank. For the cleaning robot: the lowered work area is the last work area to be worked on, and the work area opposite to the lowered work area is the first work area to be worked on. Starting from the lower area, the robot cycles through all work areas at the bottom of the tank in the order of relative work area - adjacent work area, and then transfers the oil storage tank out from the manhole where it was initially lowered. For the inspection robot: When the cleaning robot starts cleaning the second work area, the inspection operation is carried out in the same order as the cleaning robot in each work area that has been cleaned. Before the inspection operation of the next work area, the oil tank is moved out and lowered from the manhole closest to the next work area to be inspected.

2. The anti-entanglement path planning method for an oil tank servicing robot according to claim 1, characterized by: In step e, constructing the obstacle envelope circle includes the following steps: Step e-1: The robot uses its onboard ultrasonic sensors to perceive obstacle feature points and feature edge information. Step e-2: Construct the circumcircle of the obstacle using the points and edges, i.e., the obstacle envelope circle; Step e-3: After completing the outer contour fitting of the envelope circle, the coordinate system of the envelope circle is transformed, and the obstacle envelope circle is transformed from the tank bottom coordinate system to the robot motion coordinate system. Step e-4: Determine the coordinates of the center of the obstacle envelope circle in the robot's motion coordinate system.

3. The anti-entanglement path planning method for an oil tank servicing robot according to claim 1, characterized by: In step g, during the robot's obstacle avoidance action, when the real-time motion vector direction is the same as the initial velocity vector direction, the obstacle avoidance action stops, and the endpoint of the stopped obstacle avoidance action is taken as the end point.

4. The anti-entanglement path planning method of an oil tank servicing robot according to claim 1, characterized by: Step h includes the following steps: Step h-1: Rectify each work point that needs to be worked within the work area, so that the entire work point is located within its circumscribed rectangle, and then rasterize the circumscribed rectangle. Step h-2: Set the detection priority for the grid cells within the detection area; Step h-3: Assign a value to the current grid cell. After completing the task in the current grid cell, the robot moves to the grid cell with the highest priority according to the priority order, sets the next grid cell as the current position, and assigns a value. In step h-4, the robot determines whether all grids have been traversed. If all grids have been traversed, it moves to the next work point in the work area to perform the work, or moves to another work area to perform the work; otherwise, it returns to step h-3.

5. The anti-entanglement path planning method of an oil tank servicing robot according to claim 1, characterized by: In step h-1, the set S is defined as {( x, y )| x, y The points within the area to be detected are defined, and the value of each grid cell is also defined. And set the grid where the robot is currently located as the starting grid; In step h-3, the current grid is assigned a value of 0.

6. The anti-entanglement path planning method of an oil tank servicing robot according to claim 1, characterized by: In step h-2, the prescribed work starts from the left edge of the grid map, and the priority is P x-1，y left grid, P x，y+1 upper grid, P x+1，y right grid, P x，y-1 lower grid.

Images