Offshore converter station operation inspection robot obstacle avoidance method based on path planning

Abstract

The invention discloses an offshore converter station operation inspection robot obstacle avoidance method based on path planning, and belongs to the technical field of robot intelligent obstacle avoidance, and the method comprises the steps: calculating the decrease amplitude of the current wheel rail adhesive force, and obtaining the braking distance extension of a robot reaching each position point of an offshore converter station; according to the brake distance extension amount, the conversion proportion of the slip position distribution density in the high-risk area and the base plate corrosion area proportion is determined, the moving speed upper limit and the minimum steering radius of the robot are set, and an alternative route set of the high-risk area is constructed; and based on the optimal route in the alternative route set, planning the moving speed and the wheel track steering radius of the robot, and constructing an obstacle avoidance moving path. Therefore, by implementing the method, the problem that the robot is easy to deviate from the inspection target in the actual moving process due to the fact that the corrosion condition of the bottom plate is not considered when the inspection path of the robot in the offshore converter station is planned in the prior art can be solved.

Description

Technical Field

[0001] This application belongs to the field of intelligent obstacle avoidance technology for robots, specifically relating to an obstacle avoidance method for an offshore converter station operation and maintenance robot based on path planning. Background Technology

[0002] As a core facility for marine energy transmission, offshore converter stations require inspection robots to move across narrow, slippery, and severely corroded seabeds to perform equipment status checks and troubleshooting. The robots utilize wheel-rail drive for precise positioning and agile obstacle avoidance, ensuring efficient completion of inspection tasks. However, the salt spray and humidity of the marine environment accelerate seabed corrosion, resulting in a complex and variable operating environment that places extremely high demands on the robot's stability and safety, thus requiring sophisticated path planning.

[0003] Existing solutions achieve robot obstacle avoidance by pre-setting fixed paths or using simple sensors to detect environmental obstacles. However, these methods have significant limitations when facing complex problems caused by chassis corrosion. Fixed path planning cannot adapt to real-time changes in chassis surface conditions, while single obstacle detection methods are insufficient to comprehensively perceive the decrease in wheel-rail adhesion caused by corrosion. In particular, when the robot travels in severely corroded areas, the reduced surface roughness of the chassis directly weakens the friction between the wheels and rails, leading to slippage or increased braking distance. As chassis corrosion worsens, the decreased wheel-rail adhesion prevents the robot from stopping or turning in time according to the expected path. This dynamic change makes it difficult for the robot to maintain safe obstacle avoidance on the original path, increasing the risk of collisions or deviations, and may even cause the robot to stray from the inspection target, resulting in mission failure. Summary of the Invention

[0004] This application proposes an obstacle avoidance method for an offshore converter station operation and maintenance robot based on path planning. This method can solve the problem in the prior art that the robot does not take into account the corrosion of the bottom plate when planning the inspection path of the robot in the offshore converter station, which makes the robot prone to deviating from the inspection target during actual movement.

[0005] The first aspect of this application provides an obstacle avoidance method for an offshore converter station operation and maintenance robot based on path planning, the method comprising:

[0006] By scanning the bottom surface of the offshore converter station and collecting the tire slip ratio of the robot, the decrease in the current wheel-rail adhesion is calculated, and the braking distance extension of the robot to each position point of the offshore converter station is obtained.

[0007] Based on the braking distance extension, high-risk areas in the offshore converter station are identified, and the conversion ratio between the distribution density of slippage locations and the proportion of corrosion area on the bottom plate in the high-risk areas is determined.

[0008] Based on the conversion ratio, set the upper limit of the robot's moving speed and the minimum turning radius, and construct a set of alternative routes for the high-risk area;

[0009] The optimal route is constructed by selecting the set of alternative routes. Based on the optimal route, the robot's moving speed and wheel-rail turning radius are planned, and an obstacle avoidance movement path is constructed.

[0010] The above scheme predicts the surface roughness of the bottom plate based on the size of the corrosion area of ​​the offshore converter station, thereby accurately estimating the decrease in the robot's wheel-rail adhesion in rough areas, and thus determining the required braking distance extension when the robot's traction and braking capabilities are significantly reduced. The braking distance extension identifies high-risk areas prone to slippage and other loss-of-control events. The calculated conversion ratio reflects the spatial clustering characteristics of slippage risk, providing support for subsequent planning of low-slippage-risk paths. The conversion ratio resets the upper limit of movement speed and the minimum turning radius for stable movement and accurate braking in high-risk areas, thus constructing a set of safe alternative routes. This provides a basis for the robot to avoid dynamic risks caused by bottom plate corrosion during inspection tasks. Finally, path segments that are as far away from high-risk areas as possible and meet the robot's minimum turning radius are selected from the alternative route set. This constructs a low-slippage-risk obstacle-avoidance movement path that ensures safe and accurate execution of inspection tasks, enabling the inspection robot to operate safely and stably in complex corrosive environments.

[0011] In one possible implementation of the first aspect, by scanning the bottom surface of the offshore converter station and collecting the tire slip ratio of the robot, the decrease in the current wheel-rail adhesion is calculated to obtain the extension of the braking distance of the robot to each position point of the offshore converter station, specifically:

[0012] Scan the bottom plate surface of the offshore converter station to obtain the current bottom plate roughness at the specified location;

[0013] The ratio of the current base plate roughness to the tire slip ratio is used as the friction attenuation coefficient of the region to which the location point belongs;

[0014] If the friction attenuation coefficient exceeds a preset first threshold, the decrease in the current wheel-rail adhesion force relative to the preset standard adhesion coefficient is calculated, and the robot's actual braking torque is updated based on the decrease.

[0015] Based on the updated actual braking torque, the robot's braking response time, and the current moving speed, the braking distance extension relative to the standard braking distance is calculated.

[0016] The above scheme determines the friction attenuation coefficient of each region of the bottom plate of the offshore converter station by using the current bottom plate roughness and tire slip ratio. This allows for the precise calculation of the robot's adhesion in these regions and the decrease in adhesion in corroded regions. It also yields the additional braking distance extension when the robot's traction and braking capabilities are significantly reduced, providing support for subsequent path planning and speed control.

[0017] In one possible implementation of the first aspect, high-risk areas in the offshore converter station are identified based on the braking distance extension, and a conversion ratio between the distribution density of slippage locations and the proportion of bottom plate corrosion area in the high-risk areas is determined, specifically as follows:

[0018] The bottom plate area of ​​the offshore converter station is divided using a grid method, and the area where the number of risk points exceeds a second threshold is designated as the high-risk area; wherein, the risk point is the location point where the braking distance extension exceeds a preset safety threshold.

[0019] Obtain the number of historical slippage events that have occurred in the high-risk area, and determine the distribution density of the slippage locations;

[0020] The contour of the corroded area of ​​the base plate is extracted from the high-risk area by using an edge detection algorithm, and the proportion of the area of ​​the corroded area of ​​the base plate in the area of ​​the high-risk area is calculated to obtain the proportion of the corroded area of ​​the base plate.

[0021] Using the distribution density of slippage locations as the independent variable and the proportion of corrosion area on the base plate as the dependent variable, a linear regression equation is constructed, and the regression coefficient of the linear regression equation is used as the conversion ratio; wherein, the conversion ratio represents the change in the proportion of corrosion area on the base plate caused by a unit change in the distribution density of slippage locations.

[0022] In one possible implementation of the first aspect, the number of historical slippage events that have occurred in the high-risk area is obtained to determine the distribution density of the slippage locations, specifically as follows:

[0023] Identify slip points in the high-risk area, and perform spatial smoothing on the slip points using a Gaussian kernel function to obtain a kernel density value for each slip point; wherein, the kernel density value is the number of historical slip events that have occurred within a preset radius of the slip point;

[0024] The kernel density values ​​of all the slip locations are superimposed to obtain the spatial distribution density field.

[0025] In the spatial distribution density field, the product of the number of historical slip events per unit area and the kernel density value is used to obtain the slip location distribution density.

[0026] The above scheme constructs the distribution density of slip locations to characterize the spatial clustering characteristics of slip risk, thereby more accurately identifying areas prone to slip events in the base plate area.

[0027] In one possible implementation of the first aspect, the robot's upper limit for movement speed and minimum turning radius are set according to the conversion ratio, and a set of alternative routes for the high-risk area is constructed, specifically as follows:

[0028] Based on the aforementioned conversion ratio, the potential failure probability of the robot failing to brake per unit time is analyzed when the corrosion area of ​​the base plate increases.

[0029] Based on the potential failure probability value, adjust the robot's upper limit of movement speed, and set the robot's minimum turning radius using the upper limit of movement speed;

[0030] If the minimum turning radius is less than a preset turning radius threshold, then the set of alternative routes is constructed by searching for candidate paths that meet the minimum turning radius within the high-risk area.

[0031] The above scheme directly reflects the risk of braking failure due to increased corrosion area through the potential failure probability value, thereby constraining the robot's movement speed when performing inspection tasks and thus obtaining the minimum turning radius. Then, candidate paths that meet the minimum turning radius are searched in high-risk areas to determine alternative routes that ensure the robot's safe movement. Compared to the general route, the robot's tires have stronger adhesion and its braking performance is more reliable on the alternative routes.

[0032] In one possible implementation of the first aspect, the set of alternative routes is constructed by searching for candidate paths that satisfy the minimum turning radius within the high-risk area, specifically as follows:

[0033] Search for candidate paths that meet the minimum turning radius within the high-risk area, calculate a weighted score for each candidate path based on a preset cost function, and add candidate paths whose weighted scores exceed a fourth threshold to the set of alternative routes.

[0034] The cost function is used to calculate the cost of the movement path length, the number of robot turns, and the degree of avoidance of the corrosive area. It also calculates the adhesion gradient cost and the predicted slip risk cost. The adhesion gradient cost is calculated by the adhesion change amplitude at each point on the candidate path. The predicted slip risk cost is calculated by the slip position distribution density at each point on the candidate path.

[0035] The above scheme also considers the cost of adhesion gradient and the cost of predicting slippage risk. The more drastic the change in wheel-rail adhesion, the more likely the robot is to slip or lose control; the higher the slippage density, the greater the risk of brake failure. The alternative routes constructed based on these two factors can ensure that the robot's tire adhesion is stronger, the braking response is better, and it can perform inspection tasks more reliably.

[0036] In one possible implementation of the first aspect, the optimal route is constructed from the set of alternative routes, specifically as follows:

[0037] Select the route with the smallest overlap with the high-risk area from the set of alternative routes as the candidate route;

[0038] The slip ratio of the candidate route is obtained by recording the actual rotational speed of the drive wheels at each path point along the candidate route using a wheel speed detection device.

[0039] If the slip ratio is lower than the preset allowable slip limit, then calculate the robot's braking safety distance on the candidate route;

[0040] By comparing the braking safety distance with the preset obstacle avoidance safety margin, safe path segments in the candidate routes are identified, and the safe path segments are connected according to the movement order to construct the optimal route.

[0041] The above scheme is based on avoiding high-risk areas as much as possible. It selects the optimal route from the set of alternative routes that meets the requirements of wheel-rail adhesion and has better braking performance. This effectively avoids high-risk areas, ensures that the robot can move stably on the path segment, and reduces the risk of slippage.

[0042] In one possible implementation of the first aspect, an obstacle avoidance path is constructed based on the robot's moving speed and wheel-rail turning radius according to the optimal route planning, specifically as follows:

[0043] By combining the speed threshold of each segment in the optimal route and the current roughness of the bottom plate, the upper limit of the robot's moving speed and the deceleration magnitude on the optimal path are planned to obtain the speed adjustment strategy.

[0044] Based on historical inspection data, wheel-rail slippage events and braking timeout events that occurred under the speed adjustment strategy were statistically analyzed, and a feedback dataset was constructed.

[0045] Based on the feedback dataset and the high-risk area, identify the slippage locations on the optimal route and obtain the slippage location prediction results;

[0046] Based on the slip position prediction results, the optimal route is corrected by evaluating the robot's minimum turning radius on the optimal route to obtain the obstacle avoidance movement path.

[0047] In one possible implementation of the first aspect, the slippage location on the optimal route is identified based on the feedback dataset and the high-risk area to obtain a slippage location prediction result, specifically as follows:

[0048] Extract event coordinates from the feedback dataset, calculate the shortest distance between the event coordinates and the boundary of the high-risk area, and obtain the spatial deviation rate;

[0049] The slippage region identification coefficient is adjusted based on the spatial deviation rate, and the prior distribution of the slippage position is corrected by the adjusted slippage region identification coefficient.

[0050] By combining the prior distribution, the spatial distribution of slip locations, and the current distribution of corroded areas, slip locations on the optimal route are identified, and slip location prediction results are obtained; wherein, the spatial distribution of slip locations is constructed through a feedback dataset.

[0051] The above scheme identifies slippage points on the optimal route with the help of historical data, providing data support for subsequent adjustments to the optimal route to avoid slippage points.

[0052] In one possible implementation of the first aspect, based on the slip position prediction result, the optimal route is corrected by evaluating the robot's minimum turning radius on the optimal route to obtain the obstacle avoidance movement path, specifically:

[0053] Based on the slip location prediction results, the adhesion change range of the optimal route is predicted;

[0054] When the change in adhesion exceeds the preset adhesion stability range, the minimum turning radius of the robot on the optimal route is re-evaluated.

[0055] Based on the comparison between the re-evaluated minimum turning radius and the turning curvature of the optimal route, the optimal route is smoothed to construct the obstacle avoidance path.

[0056] The above solution assesses the variation in adhesion and plans a smooth and executable obstacle avoidance path that can both avoid high-risk corrosion areas and adapt to dynamic adhesion conditions while meeting the robot's turning capability limits, thus ensuring the safe and stable execution of inspection tasks by the inspection robot in complex corrosive environments. Attached Figure Description

[0057] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0058] Figure 1 This is a schematic diagram illustrating the specific process of an obstacle avoidance method for an offshore converter station operation and maintenance robot based on path planning, provided in one embodiment of this application. Detailed Implementation

[0059] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0060] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0061] First Embodiment

[0062] Adhesion, a key factor affecting the stable operation of robots, directly determines their braking distance and steering ability. However, the floor of offshore converter stations is constantly exposed to seawater, making it susceptible to corrosion, often resulting in corroded areas. When a robot travels through severely corroded areas, the reduced surface roughness of the floor directly weakens the friction between the wheels and rails, leading to a significant decrease in wheel-rail adhesion and consequently, slippage or increased braking distance. Therefore, the main research objective of this application is to adjust the robot's operating speed and path planning on the complex and variable floor of offshore converter stations to achieve precise obstacle avoidance and prevent slippage.

[0063] like Figure 1 As shown, to address the problem in existing technologies where the planning of inspection paths for robots at offshore converter stations does not consider bottom corrosion, leading to the robot easily deviating from the inspection target during actual movement, the first embodiment of this application provides a detailed flowchart of a path planning-based obstacle avoidance method for offshore converter station operation and maintenance robots. This embodiment's path planning-based obstacle avoidance method for offshore converter station operation and maintenance robots includes steps S1 to S4, detailed below:

[0064] Step S1: By scanning the offshore converter station and collecting the tire slip ratio of the robot, the decrease in the current wheel-rail adhesion is calculated, and the braking distance extension of the robot to each position point of the offshore converter station is obtained.

[0065] The bottom plate surface of the offshore converter station is continuously measured by laser profile scanning technology. The surface height change data is obtained by emitting a laser beam and receiving the reflected signal. The standard deviation of the height change is used as the current bottom plate roughness at each location.

[0066] In this embodiment, a measurement point is set up every 0.5 meters along the travel path of the inspection robot to identify the corrosion area of ​​the base plate and construct a corrosion state distribution map.

[0067] The robot's tire slip ratio is obtained by reading the actual rotational speed of the robot's drive wheels from the wheel speed monitoring device and calculating the difference between the actual speed and the theoretical speed.

[0068] Based on the corrosion state distribution map, the current roughness of each partition of the base plate is determined, and then the friction attenuation coefficient of each partition is further determined. The friction attenuation coefficient is the ratio of the current base plate roughness of that partition to the tire slip rate of the robot in that partition.

[0069] If the surface of the base plate becomes smooth due to corrosion, the surface roughness value decreases, the micro-interlocking effect of the wheel-rail contact surface weakens, and the friction decreases.

[0070] If the friction attenuation coefficient exceeds a set threshold, the decrease in the robot's current wheel-rail adhesion force in this zone relative to a preset standard adhesion coefficient is calculated, typically expressed as a percentage decrease. The robot's standard braking torque is then updated based on this percentage decrease to obtain the actual usable braking torque.

[0071] Based on the updated actual braking torque, the basic braking distance required for the robot to decelerate from its current speed to zero is calculated. The product of the robot's braking response time and the current speed is taken as the additional braking distance. The difference between the sum of the basic braking distance and the additional braking distance and the standard braking distance is taken as the braking distance extension, and this is used to obtain the braking distance extension for the robot to reach each location point of the offshore converter station.

[0072] The braking distance extension is a key parameter for subsequent path planning and speed control, and directly affects the obstacle avoidance safety of the inspection robot in corrosive environments.

[0073] In this embodiment, a wheel speed monitoring device is installed on the drive wheel axle of the inspection robot to collect the actual rotational speed in real time. The theoretical rotational speed is calculated based on the motor control signal and transmission ratio. The tire slip ratio is equal to the difference between the theoretical and actual rotational speeds divided by the theoretical rotational speed. When the tire slip ratio exceeds 0.15, it indicates that there is significant slippage between the wheel and the rail.

[0074] Optionally, for the corrosion state distribution map, this embodiment defines areas with a corrosion depth of less than 2 mm as lightly corroded areas, 2 to 5 mm as moderately corroded areas, and more than 5 mm as heavily corroded areas. For each zone, a larger friction attenuation coefficient indicates more severe slippage under the same roughness, and a more significant attenuation of friction performance. When the friction attenuation coefficient exceeds a preset threshold of 1.5, it is determined that the wheel-rail adhesion coefficient decreases by more than 30% relative to the standard adhesion coefficient of the dry and clean base plate, at which point the robot's traction and braking capabilities are significantly reduced.

[0075] Step S2: Based on the braking distance extension, identify high-risk areas in the offshore converter station and determine the conversion ratio between the distribution density of slippage positions and the proportion of corrosion area on the bottom plate in the high-risk areas.

[0076] First, the bottom plate area of ​​the offshore converter station is divided into multiple zones to be inspected using a gridding method. By comparing the braking distance extension of the robot to the position point of each zone with the preset safe braking distance range, the number of risk points within each zone is counted.

[0077] Points where the braking distance extension exceeds a preset safe braking distance range are marked as risk points. An initial risk level distribution map is constructed based on the total number of risk points in each region, and high-risk areas are identified from the number of consecutive risk levels in the map. The density of risk points represented by the initial risk level distribution map reflects the degree of danger in the region.

[0078] Optionally, in this embodiment, the base plate area is evenly divided into several detection zones using a 5m x 5m grid. The safe braking distance range is dynamically set based on the robot's moving speed and the base plate environmental conditions. Under standard operating conditions, this range is set to 1.2 times the theoretical braking distance. When the actual braking distance extension causes the total braking distance to exceed this range, it indicates a potential safety hazard at that location.

[0079] Then, the spatial distribution density value of slip locations in the high-risk area is calculated using the kernel density estimation method.

[0080] For example, slippage locations in high-risk areas are identified, and the bandwidth of the Gaussian kernel function is set to 2m, representing the impact of a single slippage event on an area within a 2-meter radius. The Gaussian kernel function is used to spatially smooth each slippage location, obtaining a kernel density value for each location. This kernel density value represents the number of historical slippage events that have occurred within the radius surrounding the slippage location. By superimposing the kernel density values ​​of all slippage locations, a continuous spatial distribution density field is obtained. In this field, the product of the number of historical slippage events per unit area and the kernel density value is calculated to obtain the slippage location distribution density, which characterizes the spatial clustering characteristics of slippage risk in the area.

[0081] Then, the corrosion image of the base plate in the high-risk area is processed by the edge detection algorithm. By calculating the gradient magnitude and direction of the image, the boundary line between the corrosion area and the normal area is identified, and the complete outline of the corrosion area of ​​the base plate is extracted.

[0082] The percentage of the corroded area of ​​the base plate is obtained by calculating the area ratio of the corroded area to the area of ​​the high-risk area.

[0083] The corrosion area percentage is calculated using a pixel statistics method. It is obtained by dividing the number of pixels within the outline of the corrosion area on the substrate by the total number of pixels in the high-risk area. This corrosion area percentage reflects different degrees of substrate corrosion.

[0084] Finally, using the slippage location distribution density as the independent variable and the percentage of corrosion area on the base plate as the dependent variable, a quadratic linear regression equation was constructed using the least squares method. The corresponding conversion ratio was obtained from the regression coefficients of this equation. Here, the conversion ratio represents the change in the percentage of corrosion area on the base plate caused by a unit change in the slippage location distribution density.

[0085] Step S3: Set the upper limit of the robot's moving speed and the minimum turning radius according to the conversion ratio, and construct a set of alternative routes for the high-risk area.

[0086] Based on the obtained conversion ratio and its corresponding linear regression equation, the potential failure probability of the robot braking within a unit time under conditions of increased corrosion area on the base plate is analyzed using a probability mapping function. This potential failure probability value ranges from 0 to 1 and is used to reflect the risk of braking failure caused by increased corrosion area.

[0087] Based on the potential failure probability value and the robot's kinetic energy data, a braking safety evaluation index is constructed. The kinetic energy data is obtained from the robot's drive motor parameters.

[0088] Based on braking safety evaluation indicators, the upper limit of the robot's moving speed is adjusted, and then the minimum turning radius of the robot is set using the adjusted upper limit of the moving speed.

[0089] Specifically, a bisection method is used to search for the speed range that meets the braking safety evaluation criteria, and the maximum allowable acceleration value corresponding to the speed range is determined. The initial search range is set from 0 to the robot's maximum design speed. In each iteration, the braking safety evaluation criterion corresponding to the midpoint value of the current speed is calculated. If the criterion exceeds a set threshold, the upper limit of the search is adjusted to the midpoint value; if the criterion does not exceed the threshold, the lower limit of the search is adjusted to the midpoint value. After multiple iterations, iteration stops when the length of the search range is less than 0.1 m / s. The upper limit at this point is the upper limit of the speed range that meets the braking safety evaluation criteria. The lower limit of the speed range is determined by considering the inspection efficiency requirements, and is usually set as the minimum speed to ensure the completion of basic inspection tasks. The maximum allowable acceleration value is calculated by the ratio of the traction force experienced by the robot during movement to the robot's mass, while also considering the limitation of the wheel-rail adhesion coefficient.

[0090] The minimum turning radius of the robot is set within a speed range that meets braking safety evaluation criteria. Generally, the curvature limit value at each speed is obtained by dividing the maximum permissible acceleration value by the square of the different movement speeds. The minimum turning radius is then calculated based on the reciprocal of the curvature limit value.

[0091] The curvature limit is calculated based on the centripetal force constraint principle. The centripetal acceleration generated by the robot when traveling on a curved path cannot exceed the maximum permissible acceleration value; therefore, the curvature limit can be derived to be approximately equal to the maximum permissible acceleration divided by the square of the moving speed. The minimum turning radius, as the reciprocal of the curvature, directly reflects the robot's turning ability at different speeds.

[0092] If the minimum turning radius is less than a preset turning radius threshold, a path replanning mechanism is initiated. This mechanism searches for paths within the passable area of ​​the platform that meet the minimum turning radius, thus constructing a set of alternative routes for the high-risk area. The turning radius threshold is typically set to the turning radius required for the current path planning.

[0093] During the route replanning process, a cost function is used to evaluate candidate paths searched within high-risk areas, and a weighted score is calculated for each candidate path. The top-scoring candidate paths are added to the alternative route set. In other embodiments, candidate paths with weighted scores exceeding a set threshold can also be selected to form the alternative route set. When searching for candidate paths, only path nodes that satisfy the minimum turning radius are considered.

[0094] The cost function, in addition to considering the costs arising from the movement path length, the number of robot turns, and the degree of corrosion area avoidance, also incorporates adhesion gradient cost and predicted slippage risk cost as key cost factors for inspection. Through this cost function, an inspection path that avoids the dynamic risks caused by base plate corrosion can be obtained.

[0095] Furthermore, the calculation of adhesion gradient cost aims to avoid areas of drastic adhesion changes in order to maintain the robot's mobility stability. An adhesion map is constructed using the current base plate roughness and the decrease in adhesion. For each node on the candidate path, the adhesion difference between it and its neighboring nodes is calculated, and the absolute value of this difference is used to obtain the adhesion gradient of the local area. Then, the adhesion gradient cost for the entire candidate path is obtained by integrating the adhesion gradients of each path segment. A higher adhesion gradient cost indicates more unstable wheel-rail contact conditions, making the robot more prone to sideslip or loss of steering control; therefore, this cost is higher.

[0096] The purpose of predicting slippage risk cost is to proactively avoid high-probability slippage areas and reduce the probability of the robot slipping. The predicted slippage risk cost for the entire candidate path is obtained by integrating the previously calculated slippage location distribution density at each location point.

[0097] In addition, the number of robot turns is obtained by counting the number of changes in the direction of the path; the degree of avoidance of corrosive areas is obtained by the length of the path crossing the corrosive area.

[0098] The cost function calculates a weighted score for each candidate path by weighting the costs incurred from the calculated path length, the number of robot turns, and the degree of avoidance of corroded areas, as well as the costs of adhesion gradient and predicted slippage risk. The weighted coefficients can be adjusted according to the actual application scenario.

[0099] The expression for the cost function is:

[0100] Total cost = w1 × movement path length cost + w2 × number of turns cost + w3 × corrosion zone avoidance cost + w4 × adhesion gradient cost + w5 × predicted slip risk cost;

[0101] In the formula, w1, w2, w3, w4 and w5 are all weighting coefficients, which can be set according to the priority of efficiency, security and stability in the actual application scenario.

[0102] For example, when the inspection robot needs to reach a detection point 100 meters away from the starting point, there are two candidate paths: Path A is 105 meters long, has 2 turns, and crosses a 20-meter area of ​​mild corrosion; Path B is 110 meters long, has 1 turn, and completely avoids the corrosion area. Considering only these three indicators, Path B receives a higher weighted score because it has fewer turns and avoids the corrosion area, therefore Path B is added to the set of candidate routes.

[0103] Step S4: Construct the optimal route using the set of alternative routes, and plan the robot's moving speed and wheel-rail turning radius based on the optimal route to construct an obstacle avoidance movement path.

[0104] The spatial overlap between each route and the high-risk area is calculated from the set of candidate routes to obtain the overlap area between each route and the high-risk area. The route with the smallest overlap area is selected as the candidate route.

[0105] Specifically, the path is treated as a polyline composed of continuous line segments, and high-risk areas are treated as closed polygons. The length of the overlapping portion is calculated by determining the intersection points of each path segment with the polygon boundary. If the path completely traverses a high-risk area, the overlapping area equals the length of the overlapping portion of the path within the area multiplied by the robot's width. If the path partially traverses a high-risk area, the coordinates of the entry and exit points are determined using a line segment trimming algorithm, and the actual overlapping portion is calculated.

[0106] Compared to other routes, the candidate route avoids high-risk areas as much as possible, thus having a lower risk of skidding compared to other paths.

[0107] Then, the actual rotational speed of the drive wheels at each path point is recorded by the wheel speed detection device along the candidate route to obtain the slip ratio of the candidate route. The calculation of the slip ratio is the same as the previous steps.

[0108] If the slip ratio of the candidate route is lower than the preset allowable slip limit, the braking safety distance of the robot for each path segment of the candidate route is calculated. If the braking safety distance is less than the preset obstacle avoidance safety margin, the corresponding path segment in the candidate route is determined to be a safe path segment.

[0109] The obstacle avoidance safety margin specifies the minimum distance that the robot must maintain between itself and obstacles. The safe path segment is the path segment that meets the obstacle avoidance requirements.

[0110] Then, the safe path segments are connected sequentially according to the movement order starting from the starting point to construct an optimal path.

[0111] As an improvement to the above scheme, spline interpolation can be used to smoothly connect the transition points between each safe path segment, forming a continuous and executable complete path. The optimal path includes a sequence of path coordinates, suggested speeds for each segment, and expected travel times.

[0112] Based on the obtained optimal path, the robot's moving speed and wheel-rail turning radius are planned to construct an obstacle avoidance movement path.

[0113] The planning process mainly includes two parts: adjusting the robot's movement strategy and correcting the path.

[0114] The suggested speed values ​​for each segment of the optimal path are extracted. By combining the suggested speed values ​​with the previously calculated braking distance extension, the upper limit of the robot's moving speed and the deceleration magnitude on the optimal path are determined, thereby obtaining the speed adjustment strategy.

[0115] The deceleration amplitude is calculated using a piecewise linear function. In this embodiment, when the braking distance extension is less than 30% of the safety threshold, the deceleration amplitude is 0; when the extension is between 30% and 70%, the deceleration amplitude increases linearly; and when the extension exceeds 70%, the deceleration amplitude increases sharply according to an exponential law.

[0116] Based on the speed adjustment strategy, historical inspection data is used to statistically analyze wheel-rail slippage events and braking timeout events occurring at various travel speeds, forming a feedback dataset containing timestamps, location coordinates, and failure indicators. The failure indicators characterize the event type, including four types: minor slippage, severe slippage, insufficient braking force, and braking timeout.

[0117] Extract the coordinates of each event from the feedback dataset, calculate the shortest distance (usually Euclidean distance) between the event coordinates and the boundary of the high-risk area, and then divide the shortest distance by the equivalent radius of the high-risk area to obtain the spatial deviation rate.

[0118] The slippage area identification coefficient is adjusted based on the spatial deviation rate. The prior distribution of slippage position is corrected by the adjusted slippage area identification coefficient to improve the accuracy of high-risk area classification.

[0119] By combining the prior distribution, the spatial distribution of slippage locations, and the current distribution of corroded areas, slippage locations on the optimal route are identified. The spatial distribution of slippage locations is constructed using historical failure events provided by the feedback dataset.

[0120] Specifically, a gridded prediction matrix is ​​constructed by combining the prior distribution, the spatial distribution of slippage locations, and the current distribution of corroded areas. The slippage locations on the optimal route are determined by bilinear interpolation of this matrix. The gridded prediction matrix divides the base plate area into 1m × 1m grid cells, with each cell storing the slippage probability value, corrosion level index, and historical failure count (i.e., the number of times slippage, braking failure, etc., have occurred historically). The time dimension of the matrix records the dynamic trend of the slippage probability value, allowing for prediction of slippage risk evolution over a future period through time series analysis. A weighted average of the slippage probability values ​​of four adjacent grid nodes in the gridded prediction matrix is ​​calculated using bilinear interpolation to obtain the slippage probability at any consecutive location. Based on the calculated slippage probability, the slippage location on the optimal route is identified, yielding the slippage location prediction result.

[0121] The interpolation weights are set inversely to the distances from the target node to each node to ensure the spatial continuity of the prediction results.

[0122] For example, when a robot needs to traverse a 50-meter-long inspection route, the prediction matrix outputs the probability of slippage every 0.5 meters along that route. Locations with a probability value exceeding 0.3 are marked as potential slippage points, considered to be prone to slippage events and thus suitable as slippage locations.

[0123] First, based on the slippage location prediction results, an adhesion force numerical sequence for each grid point in the gridded prediction matrix is ​​generated. The adhesion force difference between adjacent grid points is calculated using this sequence to obtain the adhesion force variation range on the optimal path. When the adhesion force variation range exceeds a preset adhesion force stability range, the area exceeding the range is marked as an unstable area, and the minimum turning radius of the robot in the unstable area on the optimal route is re-evaluated and updated. The adhesion force stability range is determined based on the robot's dynamic characteristics.

[0124] The magnitude of the adhesion change reflects the degree of drastic change in adhesion in space.

[0125] Optionally, in this embodiment of the application, when the absolute value of the change in adhesion between adjacent points exceeds 0.2, it indicates that there is a sharp change in adhesion in the area.

[0126] The adhesion force variation amplitudes are merged according to the principle of spatial proximity, and adjacent grid points with similar variation amplitude characteristics are merged into the same region to obtain the boundary coordinates of the region with abrupt adhesion force changes (hereinafter referred to as the unstable region). These boundary coordinates are used to subsequently determine whether the robot is about to enter or leave the unstable region. The minimum adhesion force in the unstable region is obtained by traversing all grid points in the region, and this minimum value determines the robot's limit traction capability in the region.

[0127] The minimum adhesion force in the unstable region is used to recalculate the threshold of the robot's centripetal acceleration in that region, thus obtaining the updated minimum turning radius. The updated radius is generally larger than the original data, indicating that the robot requires more turning space.

[0128] The updated minimum turning radius is compared with the turning curvature of the optimal route. Path segments in the optimal route that do not meet the new minimum turning radius are identified and smoothed to generate an obstacle avoidance path that meets the new minimum turning radius.

[0129] First, constraints are constructed: the path segment requiring smoothing is discretized into a series of control points, including the start point, end point, and key points in between, resulting in a sequence of control points for the corrected path. Based on the premise that each point on the path must be greater than or equal to the new minimum turning radius, core geometric constraints are constructed. These core geometric constraints ensure the robot can still turn stably in areas where adhesion decreases, preventing lateral slippage due to excessively sharp turns. The previously obtained high-risk areas and slippage location predictions are used as obstacle avoidance and risk mitigation zones to construct environmental risk constraints. These environmental risk constraints enable the corrected path to automatically move away from the boundaries of these obstacle avoidance and risk mitigation zones, ensuring obstacle avoidance safety margins.

[0130] Based on the aforementioned modified path control point sequence, core geometric constraints, and environmental risk constraints, cubic spline interpolation is used to smooth path segments that do not meet the conditions. Specifically, in areas with slight corrosion and stable adhesion, spline interpolation prioritizes minimizing path length and smoothing curvature changes to improve inspection efficiency and driving smoothness.

[0131] When the interpolation path approaches an unstable region where the adhesion gradient changes drastically, the density of interpolation control points is automatically increased. By inserting more control points near points of abrupt changes in adhesion, the path curve can make more precise and gradual turns in that region, thus avoiding aggressive turning maneuvers when adhesion changes rapidly.

[0132] Specifically, during the interpolation process, additional control points are inserted into the original path. The positions of these control points are determined by an optimization algorithm, with the goal of ensuring that the corrected path satisfies both the new minimum turning radius and minimizes deviation from the original path. The corrected path control point sequence includes the coordinates of the newly added control points, suggested speed values, and expected transit times. The spacing between control points is dynamically adjusted based on the path curvature; areas with greater curvature have denser control point density.

[0133] For example, when the change in adhesion causes the minimum turning radius of a certain road section to increase from 5 meters to 8 meters, the original sharp turn path is smoothed into a gentle curve, and the path length increases by about 15%.

[0134] Finally, the obstacle avoidance path obtained through interpolation is a smooth and executable path that can avoid high-risk corrosion areas, adapt to dynamic adhesion conditions, and meet the robot's turning capability limits.

[0135] Implementing the embodiments of this application has the following beneficial effects:

[0136] This application's embodiments predict the surface roughness of the bottom plate based on the size of the corrosion area of ​​the offshore converter station, thereby accurately estimating the decrease in the robot's wheel-rail adhesion in rough areas, and thus determining the required braking distance extension when the robot's traction and braking capabilities are significantly reduced. The braking distance extension identifies high-risk areas prone to slippage and other loss-of-control events. The calculated conversion ratio reflects the spatial clustering characteristics of slippage risk, providing support for subsequent low-slippage-risk path planning. The conversion ratio resets the upper limit of movement speed and the minimum turning radius for stable movement and accurate braking in high-risk areas, thus constructing a set of safe alternative routes, providing a basis for the robot to avoid dynamic risks caused by bottom plate corrosion during inspection tasks. Finally, path segments that are as far away from high-risk areas as possible and meet the robot's minimum turning radius are selected from the alternative route set, thus constructing a low-slippage-risk obstacle-avoidance movement path that ensures safe and accurate execution of inspection tasks, enabling the inspection robot to operate safely and stably in complex corrosive environments.

[0137] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for obstacle avoidance of an offshore converter station operation and maintenance robot based on path planning, characterized in that, include: By scanning the bottom surface of the offshore converter station and collecting the tire slip ratio of the robot, the decrease in the current wheel-rail adhesion is calculated, and the braking distance extension of the robot to each position point of the offshore converter station is obtained. Based on the braking distance extension, high-risk areas in the offshore converter station are identified, and the conversion ratio between the distribution density of slippage locations and the proportion of corrosion area on the bottom plate in the high-risk areas is determined. Based on the conversion ratio, set the upper limit of the robot's moving speed and the minimum turning radius, and construct a set of alternative routes for the high-risk area; The optimal route is constructed by selecting the set of alternative routes. Based on the optimal route, the robot's moving speed and wheel-rail turning radius are planned, and an obstacle avoidance movement path is constructed.

2. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 1, characterized in that, The method involves scanning the bottom surface of the offshore converter station and collecting the tire slip ratio of the robot to calculate the decrease in current wheel-rail adhesion, thereby obtaining the extension of the braking distance for the robot to reach various points on the offshore converter station. Specifically: Scan the bottom plate surface of the offshore converter station to obtain the current bottom plate roughness at the specified location; The ratio of the current base plate roughness to the tire slip ratio is used as the friction attenuation coefficient of the region to which the location point belongs; If the friction attenuation coefficient exceeds a preset first threshold, the decrease in the current wheel-rail adhesion force relative to the preset standard adhesion coefficient is calculated, and the robot's actual braking torque is updated based on the decrease. Based on the updated actual braking torque, the robot's braking response time, and the current moving speed, the braking distance extension relative to the standard braking distance is calculated.

3. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 1, characterized in that, Based on the braking distance extension, high-risk areas in the offshore converter station are identified, and the conversion ratio between the distribution density of slippage locations and the proportion of bottom plate corrosion area in the high-risk areas is determined, specifically as follows: The bottom plate area of ​​the offshore converter station is divided using a grid method, and the area where the number of risk points exceeds a second threshold is designated as the high-risk area; wherein, the risk point is the location point where the braking distance extension exceeds a preset safety threshold. Obtain the number of historical slippage events that have occurred in the high-risk area, and determine the distribution density of the slippage locations; The contour of the corroded area of ​​the base plate is extracted from the high-risk area by using an edge detection algorithm, and the proportion of the area of ​​the corroded area of ​​the base plate in the area of ​​the high-risk area is calculated to obtain the proportion of the corroded area of ​​the base plate. Using the distribution density of slippage locations as the independent variable and the proportion of corrosion area on the base plate as the dependent variable, a linear regression equation is constructed, and the regression coefficient of the linear regression equation is used as the conversion ratio; wherein, the conversion ratio represents the change in the proportion of corrosion area on the base plate caused by a unit change in the distribution density of slippage locations.

4. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 3, characterized in that, The step of obtaining the number of historical slip events that have occurred in the high-risk area and determining the distribution density of slip locations specifically involves: Identify slip points in the high-risk area, and perform spatial smoothing on the slip points using a Gaussian kernel function to obtain a kernel density value for each slip point; wherein, the kernel density value is the number of historical slip events that have occurred within a preset radius of the slip point; The kernel density values ​​of all the slip locations are superimposed to obtain the spatial distribution density field. In the spatial distribution density field, the product of the number of historical slip events per unit area and the kernel density value is used to obtain the slip location distribution density.

5. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 1, characterized in that, The step of setting the robot's upper limit for movement speed and minimum turning radius according to the conversion ratio, and constructing a set of alternative routes for the high-risk area, specifically involves: Based on the aforementioned conversion ratio, the potential failure probability value of the robot in braking failure per unit time is analyzed when the corrosion area of ​​the base plate increases. Based on the potential failure probability value, adjust the robot's upper limit of movement speed, and set the robot's minimum turning radius using the upper limit of movement speed; If the minimum turning radius is less than a preset turning radius threshold, then the set of alternative routes is constructed by searching for candidate paths that meet the minimum turning radius within the high-risk area.

6. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 5, characterized in that, The step of constructing the alternative route set by searching for candidate paths that meet the minimum turning radius within the high-risk area specifically involves: Search for candidate paths that meet the minimum turning radius within the high-risk area, calculate a weighted score for each candidate path based on a preset cost function, and add candidate paths whose weighted scores exceed a fourth threshold to the set of alternative routes. The cost function is used to calculate the cost of the movement path length, the number of robot turns, and the degree of avoidance of the corrosive area. It also calculates the adhesion gradient cost and the predicted slip risk cost. The adhesion gradient cost is calculated by the adhesion change amplitude at each point on the candidate path. The predicted slip risk cost is calculated by the slip position distribution density at each point on the candidate path.

7. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 1, characterized in that, The process of constructing the optimal route from the set of candidate routes specifically involves: Select the route with the smallest overlap area with the high-risk area from the set of alternative routes as the candidate route; The slip ratio of the candidate route is obtained by recording the actual rotational speed of the drive wheels at each path point along the candidate route using a wheel speed detection device. If the slip ratio is lower than the preset allowable slip limit, then calculate the robot's braking safety distance on the candidate route; By comparing the braking safety distance with the preset obstacle avoidance safety margin, safe path segments in the candidate routes are identified, and the safe path segments are connected according to the movement order to construct the optimal route.

8. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 1, characterized in that, The obstacle avoidance path is constructed based on the robot's moving speed and wheel-rail turning radius according to the optimal route planning, specifically as follows: By combining the speed threshold of each segment in the optimal route and the current roughness of the base plate, the upper limit of the robot's moving speed and the deceleration magnitude on the optimal path are planned to obtain the speed adjustment strategy. Based on historical inspection data, wheel-rail slippage events and braking timeout events that occurred under the speed adjustment strategy were statistically analyzed, and a feedback dataset was constructed. Based on the feedback dataset and the high-risk area, identify the slippage locations on the optimal route and obtain the slippage location prediction results; Based on the slip position prediction results, the optimal route is corrected by evaluating the robot's minimum turning radius on the optimal route to obtain the obstacle avoidance movement path.

9. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 8, characterized in that, The step of identifying skid locations on the optimal route based on the feedback dataset and the high-risk area, and obtaining skid location prediction results, specifically involves: Extract event coordinates from the feedback dataset, calculate the shortest distance between the event coordinates and the boundary of the high-risk area, and obtain the spatial deviation rate; The slippage region identification coefficient is adjusted based on the spatial deviation rate, and the prior distribution of the slippage position is corrected by the adjusted slippage region identification coefficient. By combining the prior distribution, the spatial distribution of slip locations, and the current distribution of corroded areas, slip locations on the optimal route are identified, and slip location prediction results are obtained; wherein, the spatial distribution of slip locations is constructed through a feedback dataset.

10. The obstacle avoidance method for offshore converter station operation and maintenance robots based on path planning according to claim 8, characterized in that, Based on the slippage position prediction result, the optimal route is corrected by evaluating the robot's minimum turning radius on the optimal route to obtain the obstacle avoidance movement path, specifically: Based on the slip location prediction results, the adhesion change range of the optimal route is predicted; When the change in adhesion exceeds the preset adhesion stability range, the minimum turning radius of the robot on the optimal route is re-evaluated. Based on the comparison between the re-evaluated minimum turning radius and the turning curvature of the optimal route, the optimal route is smoothed to construct the obstacle avoidance path.

Images