Railway maintenance robot path planning algorithm updating method and system

Abstract

This application relates to the field of robot algorithm technology, and provides a method and system for updating path planning algorithms for railway maintenance robots. The method includes: acquiring the joint space constraints of the railway maintenance robot in the target maintenance scenario and determining the spatial position restriction area; analyzing the range of joint motion degrees of freedom and determining the motion limit space by combining the correlation relationship; generating maintenance operation space boundary data through intersection and complement analysis; collecting historical obstacle collision path data and environmental change data; extracting abnormal trajectory fragments through spatial comparison; determining the path conflict correlation relationship by combining feature correlation analysis to generate the basis for updating the obstacle avoidance strategy; reconstructing the rules and adjusting the parameters of the obstacle avoidance strategy of the first path planning algorithm based on motion restriction rules and spatial restriction rules; and outputting a second path planning algorithm that meets the constraint requirements. This improves the adaptability of the path planning algorithm in the railway maintenance scenario and enhances the rationality and accuracy of the robot's obstacle avoidance decision.

Description

Technical Field

[0001] This application belongs to the field of robot algorithm technology, specifically relating to a path planning algorithm update method and system for railway maintenance robots. Background Technology

[0002] As core equipment ensuring the safe and stable operation of railway lines, the performance of railway maintenance robots' path planning algorithms directly determines the efficiency and safety of maintenance operations. With the continuous expansion of the railway network and the ongoing electrification upgrade, the complexity of railway maintenance scenarios is increasing. The spatial structures in different scenarios, such as tunnels, stations, and mountainous lines, vary significantly. Simultaneously, the joint structures of railway maintenance robots are becoming increasingly complex, and the motion constraints resulting from multi-joint linkages are becoming increasingly stringent. Currently, most path planning algorithms for railway maintenance robots are developed based on traditional path search algorithms. However, existing technologies suffer from the following problems: they cannot achieve a deep integration of robot joint motion constraints and scene spatial limitations, and obstacle avoidance strategy adjustments lack data-driven conflict incentives. This leads to insufficient adaptability of path planning algorithms in complex maintenance scenarios, making it difficult to guarantee the accuracy and stability of robot obstacle avoidance decisions. Summary of the Invention

[0003] This application provides a path planning algorithm update method and system for railway maintenance robots.

[0004] This application provides a path planning algorithm update method for railway maintenance robots, applied to a path planning algorithm update system. The method includes:

[0005] Obtain the joint space constraints of the railway maintenance robot in the target maintenance scenario, and determine the spatial position restriction area contained in the target maintenance scenario;

[0006] Based on the joint space constraints, the range of motion degrees of freedom of each joint of the railway maintenance robot is analyzed, and the motion limit space of the railway maintenance robot is determined by combining the correlation between the ranges of motion degrees of freedom.

[0007] By performing a spatial range intersection and complement analysis on the motion limit space and the spatial position restriction area, the maintenance operation space boundary data of the railway maintenance robot under the target maintenance scenario is generated.

[0008] Collect historical obstacle collision path data corresponding to the railway maintenance robot performing maintenance operations using the first path planning algorithm in the target maintenance scenario, and obtain the current environmental change data of the target maintenance scenario;

[0009] The trajectory segments in the historical obstacle collision path data are compared with the maintenance operation space boundary data to extract abnormal trajectory segments. The abnormal trajectory features of the abnormal trajectory segments and the scene features in the environmental change data are combined to determine the path conflict correlation. The obstacle avoidance strategy update basis of the first path planning algorithm is determined based on the path conflict correlation.

[0010] Based on the obstacle avoidance strategy update criteria, combined with the motion restriction rules corresponding to the joint space constraints and the spatial restriction rules corresponding to the maintenance operation space boundary data, the obstacle avoidance strategy of the first path planning algorithm is reconstructed and the parameters are adjusted to output a second path planning algorithm that satisfies the joint space constraints and matches the spatial position restriction area.

[0011] This application provides a path planning algorithm update system, which includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the above method.

[0012] This application provides a computer-readable storage medium including a computer program. When the computer program is run on a path planning algorithm update system, the computer program is used to cause the path planning algorithm update system to perform the steps of the above-described method.

[0013] This application provides a complete path planning algorithm update system from constraint acquisition to algorithm iteration. By deeply integrating the joint space constraints of the railway maintenance robot with the spatial location restrictions of the target maintenance scenario, it achieves precise delineation of the workspace based on the dual constraints of the robot's physical motion limits and the scene space boundaries. This overcomes the technical limitations of traditional path planning algorithms that only consider the robot's motion capabilities or scene space limitations. Through correlation analysis of historical obstacle collision path data and environmental change data, an obstacle avoidance strategy update logic based on actual operational conflict causes is constructed. This achieves a fundamental shift in obstacle avoidance strategy from "experience-based pre-setting" to "data-driven," improving the adaptability of the path planning algorithm in complex railway maintenance scenarios.

[0014] Meanwhile, during the algorithm optimization process, dual verification of joint motion restriction rules and spatial restriction rules is introduced to ensure that the output second path planning algorithm can not only meet the physical motion boundary constraints of the robot, but also accurately adapt to the spatial restriction requirements of the scene. This enhances the rationality and accuracy of the robot's obstacle avoidance decision, optimizes the motion planning capability in complex maintenance spaces, and provides full-process algorithm support for the stable operation of railway maintenance robots. Attached Figure Description

[0015] Figure 1This is a flowchart illustrating a path planning algorithm update method for railway maintenance robots provided in an embodiment of this application.

[0016] Figure 2 This is a schematic diagram of the structure of a path planning algorithm update system provided in an embodiment of this application.

[0017] Figure 3 This is a functional block diagram of a path planning algorithm update system provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this application. Obviously, the described embodiments are only some embodiments of the technical solutions of this application, and not all embodiments. Based on the embodiments recorded in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the technical solutions of this application.

[0019] See Figure 1 This is a path planning algorithm update method for railway maintenance robots provided in the embodiments of this application. This method can be applied to the path planning algorithm update system. The specific process is as follows: steps 110-160.

[0020] The updated path planning algorithm for railway maintenance robots encompasses key steps such as joint space constraint analysis, determination of motion limit space, and generation of maintenance operation space boundary data. This process introduces various parameters with different physical meanings and dimensions, including spatial coordinates in the dimension of length, curvature in the dimension of the reciprocal of length, velocity in the dimension of the length-to-time ratio, angles in the dimension of radians / measurements, as well as dimensionless or composite dimensional proportionality coefficients and restriction levels. In parameter fusion and weighted calculation operations, directly performing mathematical operations on parameters with different dimensions can easily lead to dimensional mismatch problems, resulting in calculation results lacking physical meaning and affecting the correctness and reliability of the algorithm.

[0021] Those skilled in the art can effectively solve this problem through adaptive normalization. Normalization techniques can convert parameters with different dimensions into a unified dimensionless standard form. Common methods include min-max scaling and Z-score standardization, which makes the parameters comparable in scale, ensures the mathematical rationality of parameter fusion, correlation analysis, and other operations, and enhances the algorithm's adaptability to multi-sample data. This ensures that the path planning algorithm accurately integrates multi-source information when updating, and outputs a reliable second path planning algorithm.

[0022] In fields such as robot path planning and motion control, parameter normalization is a standard preprocessing step, widely used to eliminate the influence of dimensions and improve the generalization performance of models. For example, in machine learning feature engineering, feature scaling unifies the parameter range, and dimensionless normalization simplifies equations in physical simulation systems. For railway maintenance robots, the kinematic, geometric, and environmental parameters involved in path planning have different dimensions. Technicians can use conventional methods such as linear normalization to achieve parameter co-operation. For example, coordinate distances can be normalized by dividing by the maximum size of the scene, and trajectory direction change rate and velocity change parameters can be converted into proportional coefficients by dividing by their respective maximum allowable values, thus preserving the physical nature of the parameters while ensuring computational consistency.

[0023] The above normalization process is a fundamental step in engineering practice and requires no creative effort. For example, when establishing spatial feature correlation mapping, spatial contour parameters can be converted into dimensionless relative values ​​by dividing by a reference value, and then weighted and fused with dimensionless parameters of regional attributes; when determining path conflict correlation, abnormal trajectory feature parameters can be normalized to relative ratios and then calculated in combination with environmental scene feature parameters to ensure the physical meaning of correlation coefficient analysis, while enhancing the algorithm's robustness to dynamic environments.

[0024] Existing technologies already have mature normalization frameworks and tools to support this, such as the Robot Operating System (ROS) which converts sensor data into uniform dimensional values ​​through standardized interfaces. The cost function design of path planning algorithms such as RRT involves dimensionless weighted summation of parameters. Technical personnel can draw upon these methods to design suitable normalization functions for complex operations such as spatial intersection and complement analysis and determination of path conflict relationships, ensuring that motion constraint rules and spatial constraint rules work synergistically.

[0025] In summary, the path planning algorithm update for railway maintenance robots involves mixed calculations of heterogeneous parameters. However, those skilled in the art can easily achieve dimensional unification and scale adaptation by leveraging existing normalization knowledge, thereby eliminating dimensional calculation errors.

[0026] Step 110: Obtain the joint space constraints of the railway maintenance robot in the target maintenance scenario, and determine the spatial position restriction area contained in the target maintenance scenario.

[0027] In this embodiment, the target maintenance scenario is set as the overhead contact line maintenance scenario inside an electrified railway tunnel. The path planning algorithm update system first obtains the joint space constraints, such as the maximum and minimum rotation angles and joint torque limits of each rotary and kinetic joint of the robot, by connecting to the joint control unit of the railway maintenance robot. Simultaneously, by connecting to the 3D spatial modeling system inside the tunnel, it identifies the areas containing obstacles such as cable supports protruding from the tunnel wall, overhead contact line suspension devices, and maintenance platforms fixed to the tunnel sidewalls, defining these areas as spatial location restriction zones. By accurately obtaining the joint space constraints, the path planning algorithm update system can clearly define the physical boundaries of the railway maintenance robot's motion, avoiding the planning of paths exceeding the robot's joint movement capabilities. Determining the spatial location restriction zones draws a safety red line for the robot's working space, improving the adaptability of the path planning algorithm in railway maintenance scenarios.

[0028] Step 120: Based on the joint space constraints, analyze the range of motion degrees of freedom of each joint of the railway maintenance robot, and determine the motion limit space of the railway maintenance robot by combining the correlation between the ranges of motion degrees of freedom.

[0029] In this embodiment, the path planning algorithm update system analyzes the range of motion of each joint in the railway maintenance robot, such as the shoulder rotation joint, elbow swing joint, and wrist flexion-extension joint, to obtain the range of freedom of a single joint. Then, the system analyzes the linkage relationships between the joints. For example, the rotation of the shoulder joint causes a shift in the range of motion of the elbow joint, and the range of motion of the wrist joint is limited by the current position of the elbow joint. By constructing a joint linkage constraint model, the system spatially merges the ranges of freedom of multiple single joints to obtain the set of all spatial positions that the robot's end effector can reach, i.e., the motion limit space. This process fully considers the cooperative motion characteristics of the joints of the railway maintenance robot, avoids the calculation deviation of motion space caused by relying solely on single joint constraints, enhances the rationality of the robot's obstacle avoidance decisions, and ensures that the motion limit space can truly reflect the robot's actual reachable range.

[0030] Step 130: Perform spatial range intersection and complement analysis on the motion limit space and the spatial position restriction area to generate the maintenance operation space boundary data of the railway maintenance robot in the target maintenance scenario.

[0031] In this embodiment, the path planning algorithm update system performs spatial calculations on the motion limit space and the spatial position restriction area. By analyzing the intersection complement, it eliminates the part of the motion limit space occupied by the spatial position restriction area, thereby obtaining the actual safe operating space range of the robot. The boundary coordinates, contour features, and other information of this range are then organized into maintenance operation space boundary data. This step defines a precise boundary for the railway maintenance robot's operation activities, enabling the path planning algorithm to perform path search within a safe range. This optimizes the robot's motion planning capability in complex maintenance spaces and reduces the risk of collisions between the robot and obstacles in the tunnel.

[0032] Step 131: Extract the spatial contour features of the motion limit space and the regional attribute features of the spatial position restriction area, and combine the spatial contour features and the regional attribute features to establish a spatial feature association mapping between the motion limit space and the spatial position restriction area.

[0033] In this embodiment, the path planning algorithm update system first extracts spatial contour features such as the coordinates of the outer contour vertices of the motion limit space, the tilt trend of the contour edges, and the area distribution of the contour closed region. Simultaneously, it extracts regional attribute features such as obstacle type, space occupancy pattern, and restriction level of the spatially restricted region. Subsequently, the system uses a preset spatial feature matching algorithm to associate the contour vertices of the motion limit space with the boundary points of the spatially restricted region, and to map the contour edges of the motion limit space with the obstacle edges of the spatially restricted region, establishing a spatial feature association mapping between the two. This ensures that subsequent intersection and complement analysis can be conducted based on accurate spatial correspondences.

[0034] Step 1311: Extract the contour vertex coordinates, contour edge slope parameters, and contour closed region area parameters of the motion limit space to perform contour parameterization processing to obtain a spatial contour parameter set.

[0035] In this embodiment, the path planning algorithm update system extracts the contours of the three-dimensional model of the motion limit space, obtains the three-dimensional coordinates of all contour vertices, calculates the slope parameter of each contour edge, and calculates the area parameter of each contour closed region. These parameters are then integrated into a multi-dimensional spatial contour parameter set. For example, for the motion limit space of the end effector of a railway maintenance robot, the extracted contour vertex coordinates constitute a three-dimensional coordinate sequence, the slope parameter of the contour edge reflects the direction of the contour, and the area parameter reflects the spatial size of different regions. These parameters can serve as a quantitative basis for establishing spatial feature association mappings.

[0036] Step 1312: Perform attribute parsing processing on the regional attribute features to obtain a set of regional attribute parameters, including the restriction level parameter, space occupancy ratio parameter, and smoothness parameter of the regional boundary of the spatial location restriction region.

[0037] In this embodiment, the path planning algorithm update system analyzes the attribute information of spatially restricted areas, classifies the danger level of obstacles into different restriction level parameters, calculates the spatial occupancy ratio of each spatially restricted area in the motion limit space, analyzes the smoothness of the area boundaries to obtain the smoothness parameters of the area boundaries, and integrates the above parameters into a set of area attribute parameters. For example, a high-voltage cable support in a tunnel belongs to a high restriction level area. Its spatial occupancy ratio parameter reflects the proportion of the support in the robot's motion space, while the smoothness parameter reflects the regularity of the support's edges. These parameters can be used as the matching basis for spatial feature association mapping at the attribute level.

[0038] Step 1313: Based on the range of motion degrees of freedom corresponding to the motion limit space and the positional restriction characteristics of the spatial position restriction region, determine the association weight between the spatial contour parameter set and the regional attribute parameter set. The association weight is used to characterize the degree of influence of each parameter on spatial feature matching.

[0039] In this embodiment, the path planning algorithm update system combines the motion degree of freedom range of the motion limit space to analyze the influence of each parameter in the spatial contour parameter set on the robot's motion. Simultaneously, it combines the positional restriction characteristics of the spatial positional restriction region to analyze the influence of each parameter in the region attribute parameter set on the robot's obstacle avoidance. The association weight between the spatial contour parameter set and the region attribute parameter set is determined through weighted assignment. For example, the contour vertex coordinates of the motion limit space directly affect the robot's reachable position, and its association weight is set to a high value; the restriction level parameter of the spatial positional restriction region directly relates to the robot's operational safety, and its association weight is also set to a high value, ensuring that key parameters play a dominant role when establishing spatial feature association mapping.

[0040] Step 1314: Perform weighted fusion processing on the parameters in the spatial contour parameter set and the regional attribute parameter set according to the association weight to obtain fused feature parameters.

[0041] In this embodiment, the path planning algorithm update system performs weighted concatenation of corresponding parameters in the spatial contour parameter set and the regional attribute parameter set according to determined association weights to obtain fused feature parameters. For example, the coordinates of the contour vertices of the motion limit space are concatenated with the coordinates of the corresponding boundary points of the spatial location restriction region according to association weights, and the contour slope parameter of the motion limit space is concatenated with the regional boundary smoothness parameter of the spatial location restriction region according to association weights, forming fused feature parameters containing spatial location information and attribute information, providing a unified feature basis for establishing spatial feature association mapping.

[0042] Step 1315: Input the fused feature parameters into a preset spatial feature mapping model, and establish a one-to-one correspondence between the spatial contour features of the motion limit space and the regional attribute features of the spatial position restriction area through model calculation.

[0043] In this embodiment, the preset spatial feature mapping model adopts a convolutional neural network structure, including an input layer, a feature extraction layer, an association mapping layer, and an output layer. The input layer receives fused feature parameters, the feature extraction layer extracts deep features from the fused feature parameters through convolution operations, the association mapping layer establishes the feature correspondence between the motion limit space and the spatial location restriction area through a fully connected network, and the output layer outputs the one-to-one correspondence between the two. This model is trained using a large amount of spatial data from railway maintenance scenarios. During the training process, the cross-entropy loss function is used to optimize the model parameters to ensure that the association between spatial features can be accurately established.

[0044] Step 1316: Generate an initial spatial feature association mapping based on the one-to-one correspondence. Combine the spatial coordinate information of the motion limit space and the spatial coordinate information of the spatial position restriction area to perform parameter calibration on the initial spatial feature association mapping, correct the parameter deviation that occurs during the mapping process, and obtain the spatial feature association mapping between the motion limit space and the spatial position restriction area.

[0045] In this embodiment of the application, the path planning algorithm update system generates an initial spatial feature association mapping based on the one-to-one correspondence output by the preset spatial feature mapping model. Then, it substitutes the spatial coordinate information of the motion limit space and the spatial coordinate information of the spatial position restriction area into the initial mapping relationship, checks and corrects the parameter deviations caused by model calculation errors, such as adjusting the coordinate matching accuracy of the associated points and correcting the tilt angle deviation of the associated edges, so as to obtain an accurate spatial feature association mapping.

[0046] Step 132: Based on the spatial feature association mapping, perform spatial range intersection operation on the motion limit space and the spatial position restriction area to obtain the spatial coordinate information and first contour structure information of the overlapping part of the motion limit space and the spatial position restriction area, and generate initial overlapping space data.

[0047] In this embodiment, the path planning algorithm update system finds the region with overlapping coordinates in the motion limit space and the spatial position restriction region based on spatial feature association mapping, and extracts the three-dimensional coordinate information and contour structure information of the region, such as the boundary vertex coordinates of the overlapping region and the connection relationship of the contour edges. The above information is organized into initial overlapping space data, which reflects the part of the robot's motion limit space occupied by obstacles.

[0048] Step 133: Perform spatial range complementation on the motion limit space, delete the part of the motion limit space that overlaps with the spatial position restriction area, determine the spatial coordinate information and second contour structure information of the area in the motion limit space that is not covered by the spatial position restriction area, and generate initial non-overlapping spatial data.

[0049] In this embodiment, the path planning algorithm update system subtracts the portion overlapping with the spatial location restriction area in the motion limit space to obtain the space range in which the robot can operate safely. It then extracts the spatial coordinate information and contour structure information of this range, such as the boundary direction of the non-overlapping area and the distribution pattern of the internal space, to generate initial non-overlapping space data. This data defines the initial safe operating space for the railway maintenance robot and is the core of generating the maintenance operation space boundary data.

[0050] Step 134: Correct the spatial coordinate deviation and contour structure misalignment of the initial overlapping spatial data and the initial non-overlapping spatial data to obtain corrected spatial feature information.

[0051] In this embodiment, the path planning algorithm update system uses a preset spatial calibration algorithm to correct the accuracy of initial overlapping spatial data and initial non-overlapping spatial data. For example, it adjusts spatial coordinate deviations caused by coordinate transformation and corrects contour structure misalignments caused by contour extraction errors, ensuring that the spatial data accurately reflects the actual spatial conditions inside the tunnel. This step improves the accuracy of the spatial data.

[0052] Step 135: Based on the corrected spatial feature information, define the boundary of the movable spatial range of the railway maintenance robot, extract the key coordinate points and contour segments of the spatial range boundary, and generate maintenance operation spatial boundary data containing spatial boundary coordinates, boundary contour structure and spatial region attributes.

[0053] In this embodiment, the path planning algorithm update system determines the boundary of the safe operating space of the railway maintenance robot based on the corrected spatial feature information, extracts key coordinate points on the boundary, such as the turning points and extreme points of the boundary, as well as the contour line segments connecting the above key coordinate points, and marks the attribute information of each boundary area, such as whether the area is close to obstacles or whether it belongs to a high-risk area. The above information is integrated into structured maintenance operation space boundary data. This data provides accurate spatial constraints for the path planning of the railway maintenance robot, ensuring that the planned path is always within the safe operating range.

[0054] Step 136: Associate and label the boundary data of the maintenance operation space with the motion degree of freedom correlation features of the motion limit space and the position restriction features of the spatial position restriction area.

[0055] In this embodiment, the path planning algorithm update system associates each boundary point and boundary segment in the maintenance work space boundary data with the corresponding joint motion degree of freedom features in the motion limit space, marking the joint motion limit position corresponding to the boundary point. Simultaneously, it associates the maintenance work space boundary data with the positional constraint features of the spatial position restriction area, marking information such as obstacle type and restriction level corresponding to the boundary area. Through this association and annotation, the path planning algorithm can quickly obtain the constraint information corresponding to the boundary when searching for a path, further improving the efficiency and accuracy of path planning.

[0056] Step 140: Collect historical obstacle collision path data corresponding to the railway maintenance robot performing maintenance operations using the first path planning algorithm in the target maintenance scenario, and obtain the current environmental change data of the target maintenance scenario.

[0057] In this embodiment, the path planning algorithm update system connects to the railway maintenance robot's operation log database to collect historical path data of collisions or near-collisions that occurred when the robot applied the first path planning algorithm to perform overhead contact line maintenance in the tunnel. This data includes information such as the robot's position, speed, and joint status at the time of the collision. Simultaneously, the system uses environmental monitoring sensors within the tunnel to acquire current environmental change data, such as the location of newly added temporary maintenance equipment, displacement changes of the overhead contact line suspension device, and changes in air humidity within the tunnel.

[0058] Step 150: Spatial position comparison is performed between the trajectory segments in the historical obstacle collision path data and the maintenance operation space boundary data to extract abnormal trajectory segments. The abnormal trajectory features of the abnormal trajectory segments and the scene features in the environmental change data are combined to determine the path conflict correlation. The obstacle avoidance strategy update basis of the first path planning algorithm is determined based on the path conflict correlation.

[0059] In this embodiment, the path planning algorithm update system compares each trajectory segment in the historical obstacle collision path data with the maintenance operation space boundary data to identify trajectory segments that exceed the maintenance operation space boundary or are close to the spatial location restriction area, i.e., abnormal trajectory segments. Subsequently, the system extracts abnormal trajectory features such as the degree of deviation, deviation duration, and changes in motion state of the abnormal trajectory segments, as well as scene features in the environmental change data such as the addition of obstacles and changes in the environmental medium. Through feature correlation analysis, the system determines the causal relationship between abnormal trajectories and environmental changes, i.e., the path conflict correlation relationship. Finally, based on the path conflict correlation relationship, the system analyzes the shortcomings of the first path planning algorithm in obstacle avoidance strategy, determines the direction and content that need adjustment, and forms the basis for updating the obstacle avoidance strategy. This step can accurately locate the obstacle avoidance defects of the first path planning algorithm.

[0060] Step 151: Map the trajectory segments in the historical obstacle collision path data and the maintenance operation space boundary data to a preset spatial coordinate system to obtain the three-dimensional coordinate sequence of each trajectory segment and the boundary coordinate range of the maintenance operation space boundary data.

[0061] In this embodiment, the path planning algorithm update system transforms all trajectory points in the historical obstacle collision path data and all boundary points in the maintenance operation space boundary data into the same preset three-dimensional spatial coordinate system to ensure that their spatial positions are comparable. Subsequently, the system extracts the sequence of continuous three-dimensional coordinate points of each trajectory segment, as well as the minimum and maximum three-dimensional coordinate ranges defined by the maintenance operation space boundary data.

[0062] Step 152: Calculate the shortest spatial distance between each coordinate point in the three-dimensional coordinate sequence of each trajectory segment and the boundary coordinate range of the maintenance operation space boundary data. Compare the shortest spatial distance with a preset distance threshold and mark the trajectory segments with the shortest spatial distance less than the preset distance threshold as candidate abnormal trajectory segments.

[0063] In this embodiment of the application, the path planning algorithm update system calculates the shortest straight-line distance from each coordinate point in each trajectory segment to the boundary coordinate range of the maintenance operation space boundary data, compares the distance with a preset safe distance threshold, and if the shortest spatial distance between multiple coordinate points in a trajectory segment is less than the preset distance threshold, the trajectory segment is marked as a candidate abnormal trajectory segment. This step can initially screen out trajectory segments with collision risks.

[0064] Step 153: Perform trajectory trend analysis on the candidate abnormal trajectory segments by calculating the trajectory direction change rate and trajectory curvature change value, determine whether the candidate abnormal trajectory segments have a trend of extending beyond the boundary data limit of the maintenance operation space, and identify the candidate abnormal trajectory segments with such an extension trend as abnormal trajectory segments.

[0065] In this embodiment, the path planning algorithm update system calculates the rate of change of direction of adjacent trajectory points in candidate abnormal trajectory segments, as well as the curvature change value of the trajectory segments, and analyzes the motion trend of the trajectory. If the trajectory direction of a candidate abnormal trajectory segment gradually deviates outside the boundary of the maintenance work space, and the trajectory curvature shows a trend of bending outward from the boundary, then the candidate abnormal trajectory segment is determined to have a collision risk and is identified as an abnormal trajectory segment. This step further accurately identifies trajectory segments that truly pose a collision risk, avoiding misjudgments caused by accidental proximity to the boundary.

[0066] Step 154: Extract the abnormal trajectory features of the abnormal trajectory segment; the abnormal trajectory features include the maximum distance of the trajectory deviating from the boundary of the maintenance work space, the trajectory length during which the trajectory deviation continues, the speed change parameters during the trajectory deviation process, and the trajectory turning angle parameters.

[0067] In this embodiment, the path planning algorithm update system extracts the maximum distance of the trajectory deviating from the boundary of the maintenance work space for a determined abnormal trajectory segment, reflecting the severity of the trajectory deviation; extracts the trajectory length during which the trajectory deviation continues, reflecting the range of the deviation's impact; extracts the changes in the robot's motion speed during the trajectory deviation process, reflecting the robot's motion state during the deviation; and extracts the trajectory turning angle parameter, reflecting the change in the direction of the trajectory deviation.

[0068] Step 155: Extract scene features from the environmental change data; the scene features include the three-dimensional size parameters of the newly added obstacles, the spatial coordinates of the newly added obstacles, and the refractive index change parameters of the environmental medium.

[0069] In this embodiment, the path planning algorithm update system extracts the three-dimensional size parameters of the newly added obstacles from the environmental change data to reflect the size of the obstacles; extracts the spatial coordinates of the newly added obstacles to reflect the specific location of the obstacles; and extracts the refractive index change parameters of the environmental medium to reflect the impact of the tunnel environment on the robot's perception system.

[0070] Step 156: Construct a feature association analysis model by inputting the abnormal trajectory features and the scene features into the feature association analysis model and calculating the correlation coefficient between each abnormal trajectory feature and each scene feature.

[0071] In this embodiment, the feature association analysis model adopts a deep neural network structure, including an input layer, a hidden layer, and an output layer. The input layer receives abnormal trajectory features and scene features. The hidden layer extracts the association information between features through multi-layer convolution and pooling operations. The output layer outputs the correlation coefficient between each abnormal trajectory feature and each scene feature. The magnitude of the correlation coefficient reflects the degree of association between the two. The model is trained using a large amount of historical data from railway maintenance scenarios. During the training process, the mean squared error loss function is used to optimize the model parameters to ensure that the correlation coefficient between features can be accurately calculated.

[0072] Step 157: Select feature combinations with correlation coefficients greater than a preset correlation coefficient threshold based on the correlation coefficients, analyze the scene causes of abnormal trajectory segments based on the feature combinations, obtain the causal relationship between abnormal trajectory segments and environmental changes, and define the causal relationship as a path conflict relationship.

[0073] In this embodiment, the path planning algorithm update system combines abnormal trajectory features with scene features whose correlation coefficient is greater than a preset threshold, and analyzes the cause of abnormal trajectories under this combination. For example, when the abnormal trajectory feature is that the trajectory deviation distance is large, and the corresponding scene feature is that a new obstacle is approaching the work path, it can be determined that the new obstacle is the cause of the trajectory deviation. This causal relationship is defined as a path conflict relationship. This step clarifies the intrinsic connection between abnormal trajectories and environmental changes, and can be used as the core logic for determining the basis for updating obstacle avoidance strategies.

[0074] Step 158: Quantify the path conflict associations, determine the influence weights corresponding to different path conflict associations, and combine the influence weights with the deviation of the abnormal trajectory features to generate an obstacle avoidance strategy update basis that includes conflict trigger type, conflict impact range, conflict occurrence frequency and conflict response requirements.

[0075] In this embodiment, the path planning algorithm update system assigns an influence weight to each path conflict relationship based on factors such as the severity and frequency of occurrence. For example, the influence weight of a conflict relationship caused by a newly added large obstacle is set to a higher value, while the influence weight of a conflict relationship caused by a slight change in the environmental medium is set to a lower value. Subsequently, the system combines the deviation degree of abnormal trajectory characteristics, such as the magnitude of the deviation distance and the duration of the deviation, to comprehensively analyze the conflict trigger type, the scope of conflict impact, and the frequency of conflict occurrence, determine the corresponding conflict response requirements, and integrate the above information as the basis for updating the obstacle avoidance strategy.

[0076] Step 1581: Obtain the obstacle avoidance strategy parameters currently used by the first path planning algorithm, and establish the association mapping relationship between the obstacle avoidance strategy parameters and the conflict cause type.

[0077] In this embodiment, the path planning algorithm update system obtains the currently used obstacle avoidance strategy parameters, such as obstacle detection threshold, obstacle avoidance turning angle, and path replanning trigger conditions, from the configuration file of the first path planning algorithm. For each type of conflict trigger, such as adding obstacles or changing environmental media, a corresponding obstacle avoidance strategy parameter association mapping relationship is established. For example, adding obstacles corresponds to parameters such as obstacle detection threshold and obstacle avoidance turning angle, while changing environmental media corresponds to parameters such as sensor sensitivity adjustment. This step clarifies the direction of algorithm parameter adjustment corresponding to different conflict triggers.

[0078] Step 1582: Based on the influence weights, sort the types of conflict causes by weight to determine the dominant and secondary conflict causes. Quantify the trajectory risk level corresponding to the dominant conflict cause by combining the degree of deviation of the abnormal trajectory characteristics. The trajectory risk level is calculated by comprehensively considering the ratio of the deviation distance to the preset safety distance and the ratio of the length of the deviation trajectory to the total length of a single maintenance operation trajectory.

[0079] In this embodiment, the path planning algorithm update system sorts the conflict trigger types according to their influence weight from high to low, identifying the conflict trigger with the highest weight as the dominant conflict trigger and the rest as secondary conflict triggers. Subsequently, the system combines the degree of deviation of abnormal trajectory characteristics to calculate the ratio of deviation distance to a preset safety distance, and the ratio of the length of the continuous deviation trajectory to the total length of a single maintenance operation trajectory. By weighted summing of these two ratios, the system obtains the trajectory risk level corresponding to the dominant conflict trigger. A higher risk level indicates a greater impact of the conflict trigger on operational safety.

[0080] Step 1583: For the dominant conflict trigger and the corresponding trajectory risk level, analyze the adaptability of the current obstacle avoidance strategy parameters of the first path planning algorithm, determine whether the obstacle avoidance strategy parameters can meet the conflict response requirements in the current scenario, and if there are parameters with insufficient adaptability, mark them as parameters to be adjusted and determine the adjustment range of the parameters to be adjusted in combination with the secondary conflict triggers.

[0081] In this embodiment, the path planning algorithm update system checks whether the current obstacle avoidance strategy parameters of the first path planning algorithm can effectively cope with the dominant conflict trigger and its corresponding trajectory risk level. For example, when the dominant conflict trigger is the addition of a large obstacle, if the current obstacle detection threshold is too high, causing the robot to be unable to detect the obstacle in time, then the parameter is determined to be unsuitable and marked as a parameter to be adjusted. Simultaneously, the system considers the influence of secondary conflict triggers to determine a reasonable adjustment range for the parameters to be adjusted. For example, the adjustment range of the obstacle detection threshold needs to take into account the impact of changes in the environmental medium on sensor detection.

[0082] Step 1584: Based on the marked parameters to be adjusted and the corresponding adjustment range, integrate the conflict impact range, conflict occurrence frequency and quantified trajectory risk level to construct a structured strategy for updating the obstacle avoidance strategy. The structured strategy includes a description of the dominant conflict trigger, quantitative data of trajectory risk level, a list of parameters to be adjusted, parameter adjustment direction and core objectives of conflict response.

[0083] In this embodiment, the path planning algorithm update system integrates the marked parameters to be adjusted, the corresponding adjustment range, and information such as the scope of conflict impact, frequency of conflict occurrence, and trajectory risk level to construct a structured basis for obstacle avoidance strategy updates. The structured strategy clearly describes the type and characteristics of the dominant conflict triggers, provides a quantitative value for the trajectory risk level, lists the names and current values ​​of the parameters to be adjusted, indicates the adjustment direction for each parameter (e.g., increase or decrease), and determines the core objectives of conflict response, such as improving obstacle detection sensitivity and shortening obstacle avoidance response time. This structured strategy provides accurate operational guidelines for algorithm optimization.

[0084] Step 1585: Associate and label the structured strategy with the obstacle avoidance strategy logic of the first path planning algorithm to lock the algorithm module corresponding to each update content, and generate the obstacle avoidance strategy update basis for rule reconstruction and parameter adjustment of the obstacle avoidance strategy of the first path planning algorithm.

[0085] In this embodiment, the path planning algorithm update system associates each update in the structured strategy with the obstacle avoidance strategy logic module of the first path planning algorithm. For example, it associates the obstacle detection threshold adjustment with the obstacle detection module of the algorithm, and the obstacle avoidance turning angle adjustment with the path adjustment module of the algorithm. This locks the specific algorithm module corresponding to each update and generates the final obstacle avoidance strategy update basis. This basis ensures that the algorithm optimization can accurately locate the specific module, improving optimization efficiency and accuracy.

[0086] Step 160: Based on the obstacle avoidance strategy update criteria, and combined with the motion restriction rules corresponding to the joint space constraints and the spatial restriction rules corresponding to the maintenance operation space boundary data, the obstacle avoidance strategy of the first path planning algorithm is reconstructed and the parameters are adjusted to output a second path planning algorithm that satisfies the joint space constraints and matches the spatial position restriction area.

[0087] In this embodiment, the path planning algorithm update system reconstructs the obstacle avoidance strategy rules of the first path planning algorithm based on the obstacle avoidance strategy update criteria, adjusts relevant parameters, and ensures that the adjusted algorithm meets the constraints of joint space and maintenance operation space boundary data. Finally, it generates an optimized second path planning algorithm. This step achieves targeted optimization of the first path planning algorithm, improves the obstacle avoidance capability of the algorithm in railway maintenance scenarios, and provides reliable algorithm support for the stable operation of railway maintenance robots.

[0088] Step 161: Based on the obstacle avoidance strategy update criterion, parse the obstacle avoidance strategy logic module of the first path planning algorithm, extract the parameter configuration information and rule constraints of each obstacle avoidance strategy logic module, and establish a logical relationship diagram between the obstacle avoidance strategy logic modules.

[0089] In this embodiment, the path planning algorithm update system parses the obstacle avoidance strategy logic structure of the first path planning algorithm based on the algorithm modules marked in the obstacle avoidance strategy update criteria. It extracts the parameter configuration information of each logic module, such as the module's input and output parameters, internal calculation thresholds, and rule constraints, such as the module's triggering conditions and execution order. Subsequently, the system analyzes the data flow and control relationships between the logic modules, constructing a logical association graph between the obstacle avoidance strategy logic modules, with modules as nodes and inter-module relationships as edges, thus visually displaying the algorithm's obstacle avoidance strategy logic architecture.

[0090] Step 1611: Extract the algorithm module identification information corresponding to each update content included in the obstacle avoidance strategy update basis, locate all obstacle avoidance strategy logic modules related to the obstacle avoidance strategy in the first path planning algorithm based on the algorithm module identification information, and establish a module index list.

[0091] In this embodiment of the application, the path planning algorithm update system extracts the algorithm module identifier corresponding to each update content from the obstacle avoidance strategy update basis, such as the obstacle detection module identifier, the path replanning module identifier, etc. Based on the above identifiers, the corresponding obstacle avoidance strategy logic module is located in the code structure of the first path planning algorithm, and the name, location, function and other information of all related modules are organized into a module index list.

[0092] Step 1612: Based on the module index list, the core code snippets and parameter configuration files of each obstacle avoidance strategy logic module are called sequentially. The parameter configuration information of each obstacle avoidance strategy logic module is extracted through code semantic analysis, and the rule constraints corresponding to each obstacle avoidance strategy logic module are extracted through rule clause parsing. The rule constraints include the spatial limit threshold for trajectory generation, motion parameter constraint standards, and conflict response triggering conditions.

[0093] In this embodiment, the path planning algorithm update system sequentially accesses the core code snippets and parameter configuration files of each obstacle avoidance strategy logic module according to the module index list. Through code semantic analysis technology, it extracts the module's parameter configuration information, such as variable values ​​and calculation thresholds, from the code. Through rule clause parsing technology, it extracts the module's rule constraints from the code comments and configuration files, such as spatial coordinate restrictions during trajectory generation, the constraint range of robot movement speed, and response triggering conditions when a conflict occurs.

[0094] Step 1613: Cross-compare the parameter configuration information of each obstacle avoidance strategy logic module, identify parameter groups that are related between different modules and mark the parameter association type. The parameter association type includes direct parameter transmission association, indirect parameter dependency association and parameter feedback adjustment association.

[0095] In this embodiment, the path planning algorithm update system cross-compares the parameter configuration information of each obstacle avoidance strategy logic module to identify parameter groups that are related between different modules. For example, the detection result parameters of the obstacle detection module are directly transmitted to the path replanning module, which is a direct parameter transmission association; the path smoothing parameters of the path replanning module depend on the joint motion parameters of the robot motion control module, which is an indirect parameter dependency association; and the joint torque parameters of the motion control module will feed back to adjust the detection sensitivity parameters of the obstacle detection module, which is a parameter feedback adjustment association.

[0096] Step 1614: Based on the parameter association type, sort out the data flow and control logic relationship between each obstacle avoidance strategy logic module, determine the correspondence between input and output parameters between modules, and construct an initial logical association graph with each obstacle avoidance strategy logic module as a node and the association relationship between modules as an edge.

[0097] In this embodiment, the path planning algorithm update system, based on parameter association types, organizes the data flow direction and control logic relationships between various obstacle avoidance strategy logic modules. For example, the output parameters of the obstacle detection module serve as the input parameters of the path replanning module, the output parameters of the path replanning module serve as the input parameters of the motion control module, and the state parameters of the motion control module are fed back to the obstacle detection module. The system constructs an initial logical association graph, using each obstacle avoidance strategy logic module as a node and the relationships between modules as edges, to intuitively display the obstacle avoidance strategy logic architecture of the algorithm.

[0098] Step 1615: In the initial logical association graph, add annotations to the core fields of parameter configuration information and key clauses of rule constraints corresponding to each node. Eliminate redundant association edges and isolated nodes in the initial logical association graph through graph structure optimization algorithm to generate a logical association graph between obstacle avoidance strategy logical modules that represents the interaction relationship between each obstacle avoidance strategy logical module.

[0099] In this embodiment, the path planning algorithm update system adds corresponding core fields of parameter configuration information to each node of the initial logical association graph. For example, it labels the obstacle detection module node with fields such as detection threshold and detection range, and the rule constraint conditions with key clauses such as detection trigger conditions and data update frequency. Subsequently, the system uses a graph structure optimization algorithm to remove redundant edges (relationships that have no substantial impact on the algorithm logic) and isolated nodes (modules not associated with other modules) from the initial logical association graph, generating the final logical association graph between the obstacle avoidance strategy logic modules. This ensures that the graph accurately and concisely reflects the interaction relationships between the modules.

[0100] Step 162: Perform motion adaptability analysis on each obstacle avoidance strategy logic module according to the motion restriction rules corresponding to the joint space constraints, determine whether the rule constraints of each obstacle avoidance strategy logic module are compatible with the motion restriction rules, and extract the target logic module with insufficient adaptability.

[0101] In this embodiment, the path planning algorithm update system compares the motion constraint rules corresponding to the joint space constraints, such as the maximum joint rotation angle and the joint torque limit, with the rule constraints of each obstacle avoidance strategy logic module to analyze whether the module's rules conform to the robot's motion physical boundaries. For example, if the rule constraints of the path replanning module allow planning a path exceeding the maximum joint rotation angle, then the module is determined to have insufficient fit with the motion constraint rules and is extracted as the target logic module.

[0102] Step 163: Combine the spatial constraint rules corresponding to the maintenance operation space boundary data to perform spatial adaptability analysis on the target logic module, compare the spatial matching degree between the trajectory generation parameters output by the target logic module and the maintenance operation space boundary data, and determine the conflict points and conflict types between the trajectory generation parameters and the spatial constraint rules.

[0103] In this embodiment, the path planning algorithm update system combines the spatial constraint rules corresponding to the maintenance operation space boundary data, such as the boundary coordinate range of the operation space and regional attribute constraints, to perform spatial adaptability analysis on the target logic module. The system compares the trajectory generation parameters output by the target logic module, such as the coordinates of key trajectory points and the direction of trajectory segments, with the maintenance operation space boundary data to calculate the spatial matching degree. If the matching degree is lower than a preset threshold, it determines that there is a conflict between the trajectory generation parameters and the spatial constraint rules. The system then locates the conflict points, i.e., the specific locations in the trajectory that violate spatial constraints, and analyzes the conflict type, such as boundary crossing conflicts or regional attribute violation conflicts.

[0104] Step 1631: Parse the corresponding spatial restriction rules from the maintenance operation space boundary data. The spatial restriction rules include the boundary coordinate range of the maintenance operation space, spatial area attribute constraints, and spatial topology requirements for trajectory passage. Transform the spatial restriction rules into a quantifiable set of spatial adaptability evaluation indicators. The set of spatial adaptability evaluation indicators includes coordinate matching accuracy, area attribute fit, and topology adaptability.

[0105] In this embodiment, the path planning algorithm update system extracts spatial constraint rules from the maintenance work space boundary data. These rules include the minimum and maximum three-dimensional coordinate range of the maintenance work space, attribute requirements for each area such as no passage or restricted passage, and the spatial topology that the trajectory must follow, such as not crossing obstacles or needing to follow the overhead contact line. Subsequently, the system transforms these rules into a quantifiable set of spatial adaptability evaluation indicators. Coordinate matching accuracy measures the degree of fit between the trajectory coordinates and the work space boundary, area attribute fit measures whether the trajectory conforms to area attribute constraints, and topology adaptability measures whether the trajectory conforms to spatial topology requirements.

[0106] Step 1632: Call the output interface of the target logic module to obtain the trajectory generation parameters output by the target logic module. The trajectory generation parameters include the three-dimensional coordinates of the trajectory key points, the curvature parameters of the trajectory segments, and the attribute identifier of the trajectory travel area.

[0107] In this embodiment, the path planning algorithm update system obtains trajectory generation parameters generated by the target logic module through the output interface of the target logic module. These parameters include the three-dimensional coordinates of key points on the trajectory for locating key positions, the curvature parameters of the trajectory segments for reflecting the degree of curvature of the trajectory, and attribute identifiers of the trajectory's traversed areas for labeling attribute information of the areas traversed by the trajectory. These parameters provide basic data for subsequent spatial matching degree calculation.

[0108] Step 1633: Map the three-dimensional coordinates of the trajectory key points in the trajectory generation parameters to the spatial coordinate system corresponding to the maintenance operation space boundary data. Calculate the coordinate matching deviation between each trajectory key point and the boundary coordinate range of the maintenance operation space boundary data. Simultaneously, compare the fit weight between the trajectory passage area attribute identifier and the spatial area attribute constraint of the maintenance operation space boundary data. Combine the adaptation weight between the trajectory segment curvature parameter and the spatial topology requirements of the maintenance operation space, and obtain the spatial matching degree between the trajectory generation parameters and the maintenance operation space boundary data through weighted calculation.

[0109] In this embodiment, the path planning algorithm update system transforms the three-dimensional coordinates of the trajectory key points into the spatial coordinate system corresponding to the maintenance operation space boundary data. It calculates the coordinate matching deviation between each trajectory key point and the coordinate range of the operation space boundary, i.e., the degree to which the coordinate point exceeds or approaches the boundary. Simultaneously, the system compares the degree of fit between the trajectory passage area attribute identifier and the spatial area attribute constraints, assigning a fit weight. It also assigns an adaptation weight based on the degree of fit between the trajectory segment curvature parameters and the spatial topology requirements. By weighted summing of the coordinate matching deviation, fit weight, and adaptation weight, the system obtains the spatial matching degree between the trajectory generation parameters and the maintenance operation space boundary data, quantitatively reflecting the degree of fit between the trajectory and the operation space.

[0110] Step 1634: Based on the threshold standard of the spatial adaptability evaluation index set, select the trajectory key points and trajectory segments corresponding to the trajectory generation parameters whose spatial matching degree is lower than the preset threshold, and determine the spatial positions corresponding to the trajectory key points and trajectory segments as the conflict points between the trajectory generation parameters and the spatial restriction rules.

[0111] In this embodiment, the path planning algorithm update system filters out trajectory generation parameters with spatial matching degree lower than a preset threshold based on the threshold standards of the spatial adaptability evaluation index set, such as coordinate matching accuracy threshold, regional attribute fit threshold, and topology adaptability threshold. It then locates the corresponding trajectory key points and trajectory segments, determines the actual spatial locations corresponding to the above trajectory key points and trajectory segments as conflict points, and clarifies the specific locations in the trajectory that violate spatial restriction rules.

[0112] Step 1635: Perform feature analysis on the conflict points. Based on whether the conflict points exceed the boundary coordinate range of the maintenance work space, whether they violate spatial area attribute constraints, and whether they do not meet spatial topology requirements, classify the conflict types into three categories: boundary crossing conflict, attribute mismatch conflict, and topology adaptation failure conflict. At the same time, record the spatial coordinate information of each type of conflict point and the corresponding trajectory generation parameter values ​​to form a structured analysis report of conflict points and conflict types.

[0113] In this embodiment, the path planning algorithm update system performs feature analysis on each conflict point. If the conflict point exceeds the boundary coordinate range of the maintenance work space, it is determined to be a boundary crossing conflict; if the trajectory travel area attribute identifier of the area where the conflict point is located violates the spatial area attribute constraints, it is determined to be an attribute mismatch conflict; if the curvature parameter of the trajectory segment corresponding to the conflict point does not meet the spatial topology requirements, it is determined to be a topology adaptation failure conflict. The system records the spatial coordinate information of various conflict points and the corresponding trajectory generation parameter values, and organizes the above information into a structured analysis report.

[0114] Step 164: Based on the conflict points and conflict types, and according to the list of parameters to be adjusted, the direction of parameter adjustment, and the core objectives of conflict response in the obstacle avoidance strategy update basis, the rule constraints of the target logic module are reconstructed to optimize the logic association diagram; the parameters corresponding to the reconstructed rule constraints are calibrated in layers, and multi-layer calibration thresholds are set in combination with the motion restriction rules and the spatial restriction rules. Through iterative calculation, the calibration values ​​of each parameter simultaneously meet the quantification requirements of the motion restriction rules and the boundary requirements of the spatial restriction rules.

[0115] In this embodiment, the path planning algorithm update system reconstructs the rule constraints of the target logic module based on the conflict points and conflict types, combined with the list of parameters to be adjusted, the direction of parameter adjustment, and the core objectives of conflict response in the obstacle avoidance strategy update basis. For example, for boundary crossing conflicts, the trajectory generation rules of the path replanning module are adjusted to ensure that the planned trajectory always stays within the boundary of the work space. Simultaneously, the system reconstructs the logical association graph and updates the relationships between modules. Subsequently, the system performs layered calibration on the parameters corresponding to the reconstructed rule constraints, setting joint motion parameter calibration thresholds based on motion restriction rules and spatial coordinate calibration thresholds based on spatial restriction rules. Through multiple iterative calculations, the parameter values ​​are adjusted so that the calibration values ​​of each parameter simultaneously meet the quantitative requirements of the motion restriction rules and the boundary requirements of the spatial restriction rules.

[0116] Step 165: Integrate the reconstructed rule constraints and calibrated parameters into the parameter configuration library of the first path planning algorithm to replace the parameter configuration and generate the initial second path planning algorithm.

[0117] In this embodiment of the application, the path planning algorithm update system replaces the original content in the first path planning algorithm parameter configuration library with the reconstructed rule constraints and calibrated parameters, updates the obstacle avoidance strategy logic and parameter configuration of the algorithm, and generates an initial second path planning algorithm. This step realizes the initial optimization of the first path planning algorithm.

[0118] Step 166: Perform trajectory simulation calculation on the initial second path planning algorithm, obtain simulated trajectory data, and perform dual adaptation verification on the simulated trajectory data with the joint space constraints and the spatial position restriction area. Based on the dual adaptation verification, optimize the parameters of the initial second path planning algorithm, and output a second path planning algorithm that satisfies the joint space constraints and matches the spatial position restriction area.

[0119] In this embodiment, the path planning algorithm update system calculates the initial second path planning algorithm using a trajectory simulation platform to generate simulated trajectory data. This data is then adapted and verified against joint space constraints to check whether the joint movements corresponding to the trajectory conform to the physical limitations of the joints. Simultaneously, the simulated trajectory data is adapted and verified against spatial position constraints to check whether the trajectory collides with obstacles or approaches dangerous areas. For any issues discovered during the verification process, the system fine-tunes and optimizes the parameters of the initial second path planning algorithm, ultimately generating a second path planning algorithm that satisfies the joint space constraints and matches the spatial position constraints, ensuring that the algorithm can effectively guide the railway maintenance robot to complete maintenance operations.

[0120] As an optional embodiment, the method further includes:

[0121] Step 210: Collect the trajectory time-series features corresponding to the maintenance trajectory data output by the second path planning algorithm, and extract the trajectory continuity features and trajectory smoothness features from the trajectory time-series features.

[0122] In this embodiment of the application, after generating the second path planning algorithm, the path planning algorithm update system collects the multi-segment maintenance trajectory data output by the algorithm and extracts the temporal features of the trajectory, including trajectory continuity features, which reflect the smoothness of the connection of the trajectory in the time series, such as the time interval change of adjacent trajectory points, the continuous change of trajectory speed, etc., and trajectory smoothness features, which reflect the curvature change rate of the trajectory, such as the change frequency of the turning angle, etc.

[0123] Step 220: Perform correlation analysis on trajectory continuity features and trajectory smoothness features, identify unstable feature segments in trajectory temporal features, correlate and match unstable feature segments with joint space constraints, and determine the motion constraint correlation factors generated by unstable feature segments.

[0124] In this embodiment, the path planning algorithm update system performs correlation analysis on trajectory continuity features and trajectory smoothness features to identify unstable feature segments in the trajectory that have continuity interruptions or smoothness abrupt changes. These segments are then matched with joint space constraints to analyze the joint motion state corresponding to the unstable feature segments and determine the motion constraint correlation factors that cause trajectory instability, such as trajectory abrupt changes caused by joint motion approaching the limit position or trajectory velocity abrupt changes caused by joint torque limitations.

[0125] Step 230: Based on the motion constraint correlation factors, generate trajectory robustness optimization logic; embed the trajectory robustness optimization logic into the trajectory generation module of the second path planning algorithm to obtain the initial robustness optimization algorithm; perform correlation verification between the optimized trajectory data output by the initial robustness optimization algorithm and the stable feature standard in the trajectory time series features; filter out the algorithm parameter configuration corresponding to the optimized trajectory data that meets the stable feature standard; and determine the robustness-enhanced path planning algorithm based on the parameter configuration.

[0126] In this embodiment, the path planning algorithm update system generates trajectory robustness optimization logic based on determined motion constraint correlation factors. This logic includes adjusting the trajectory direction in advance when joint movement approaches its limit position and optimizing trajectory velocity planning when joint torque is limited. The system embeds this optimization logic into the trajectory generation module of the second path planning algorithm to obtain an initial robustness optimization algorithm. Subsequently, the system correlates and verifies the optimized trajectory data output by the initial robustness optimization algorithm with preset stability feature standards. It checks whether the continuity and smoothness of the trajectory meet the standard requirements, filters out the algorithm parameter configurations corresponding to the optimized trajectory data that meet the standards, and determines a robustness-enhanced path planning algorithm based on these parameter configurations to further improve the trajectory stability of the algorithm under complex motion constraints.

[0127] As an optional embodiment, the method further includes:

[0128] Step 310: Obtain the scenario feature benchmark data of multiple different types of railway maintenance scenarios, perform feature matching analysis between the core algorithm logic of the second path planning algorithm and the feature benchmark data of each scenario, establish the adaptation relationship between the core algorithm logic and the features of different scenarios, and identify the adaptation defects of the second path planning algorithm in different scenarios based on the adaptation relationship.

[0129] In this embodiment, after generating the second path planning algorithm, the path planning algorithm update system acquires scenario feature benchmark data for multiple different types of railway maintenance scenarios, such as spatial structure features, obstacle distribution features, and operational requirements features for open-air railway line maintenance scenarios, high-speed railway station maintenance scenarios, and mountain railway tunnel maintenance scenarios. The system performs feature matching analysis between the core algorithm logic of the second path planning algorithm, such as obstacle detection logic and path replanning logic, and the benchmark data for each scenario feature, establishing an adaptation relationship between the core algorithm logic and different scenario features. For example, a certain algorithm logic may be suitable for tunnel scenarios but not for open-air scenarios. Based on this relationship, the system identifies the adaptation defects of the second path planning algorithm in different scenarios, such as insufficient obstacle detection accuracy in open-air scenarios and low path planning efficiency in station scenarios.

[0130] Step 320: To address the aforementioned adaptation defects, a generalized adjustment logic for the algorithm is generated by combining the generalized characteristics of joint space constraints and the common characteristics of spatial location restriction areas in different scenarios. This generalized adjustment logic is then integrated into the second path planning algorithm to obtain an initial generalized algorithm. The initial generalized algorithm is then subjected to adaptation verification logic execution on feature data from multiple different types of railway maintenance scenarios to optimize the generalized parameter configuration of the algorithm and output a generalized path planning algorithm that adapts to multiple types of railway maintenance scenarios.

[0131] In this embodiment, the path planning algorithm update system, addressing identified adaptation defects, combines the generalized characteristics of joint space constraints, such as the motion limits of various robot joints and the common features of spatial positional constraints in different scenarios, such as the need to avoid obstacles, to generate generalized adjustment logic for the algorithm. This includes designing switchable obstacle detection modes and constructing an adaptive path planning framework. The system integrates this adjustment logic into the second path planning algorithm to obtain an initial generalized algorithm. Subsequently, the system performs adaptation verification on feature data from multiple different types of railway maintenance scenarios, checking the algorithm's operational effectiveness in each scenario and optimizing the algorithm's generalized parameter configuration, such as adjusting obstacle detection thresholds and path replanning trigger conditions in different scenarios. Finally, it outputs a generalized path planning algorithm adapted to multiple types of railway maintenance scenarios, improving the algorithm's scenario adaptability.

[0132] As an optional embodiment, the method further includes:

[0133] Step 410: Obtain multi-task feature data corresponding to the target maintenance scenario, and perform task association analysis on the single-task maintenance trajectory data output by the second path planning algorithm and the multi-task feature data to obtain the trajectory-task association relationship.

[0134] In this embodiment, after generating the second path planning algorithm, the path planning algorithm update system obtains multi-task feature data corresponding to the target maintenance scenario, such as the task execution order, task operation location, and task execution time requirements for tasks like contact wire wear detection, insulator cleaning, and bolt tightening. The system performs task association analysis on the single-task maintenance trajectory data output by the second path planning algorithm and the multi-task feature data to establish a correspondence between the trajectory and the task, such as a certain trajectory segment corresponding to the contact wire wear detection task, a certain trajectory segment corresponding to the insulator cleaning task, etc., thus obtaining the trajectory-task association relationship.

[0135] Step 420: Based on the trajectory-task association relationship, identify the task mismatch trajectory segments in the single-task maintenance trajectory data that cannot adapt to the multi-task collaboration requirements; perform association analysis between the task mismatch trajectory segments and the spatial restriction rules corresponding to the maintenance operation space boundary data to generate multi-task trajectory collaboration adjustment logic; integrate the multi-task trajectory collaboration adjustment logic into the second path planning algorithm to obtain the initial multi-task adaptation algorithm.

[0136] In this embodiment, the path planning algorithm update system analyzes whether single-task maintenance trajectory data can meet the needs of multi-task collaboration based on the trajectory-task association relationship. This includes checking for trajectory duplication and whether the task execution order is reasonably arranged. The system identifies mismatched trajectory segments that cannot adapt to multi-task collaboration requirements, such as two trajectories repeatedly covering the same work area or trajectory order causing excessive task execution waiting time. The system performs correlation analysis between the mismatched trajectory segments and the spatial constraint rules corresponding to the maintenance work space boundary data to generate multi-task trajectory collaboration adjustment logic, such as optimizing the trajectory order to reduce task switching time and merging duplicate trajectories to improve work efficiency. The system integrates this adjustment logic into the second path planning algorithm to obtain the initial multi-task adaptation algorithm.

[0137] Step 430: Perform correlation verification analysis on the multi-task collaborative trajectory data output by the initial multi-task adaptation algorithm and the multi-task feature data to obtain the multi-task adaptation analysis results; based on the multi-task adaptation analysis results, optimize and adjust the rules of the initial multi-task adaptation algorithm to output a multi-task adaptive path planning algorithm that adapts to the multi-task collaborative maintenance requirements.

[0138] In this embodiment, the path planning algorithm update system performs correlation verification analysis on the multi-task cooperative trajectory data output by the initial multi-task adaptation algorithm and the multi-task feature data. It checks whether the multi-task cooperative trajectory conforms to the task execution order, meets the task operation location requirements, and completes the task within the specified time, thus obtaining the multi-task adaptation analysis results. Based on the analysis results, the system optimizes and adjusts the rules of the initial multi-task adaptation algorithm, such as adjusting the trajectory merging rules and optimizing the task switching trigger conditions. Finally, it outputs a multi-task adapted path planning algorithm that adapts to the needs of multi-task collaborative maintenance, thereby improving the multi-task operation capability of railway maintenance robots.

[0139] In conjunction with the above-described embodiments of this application, in an exemplary electrified railway tunnel catenary maintenance scenario, the path planning algorithm update system first interfaces with the joint control unit of a certain type of railway maintenance robot to obtain joint spatial constraints such as the maximum rotation angle and joint torque limit values ​​of its shoulder rotation joint and elbow swing joint. Simultaneously, it interfaces with the tunnel 3D modeling system to identify spatially restricted areas such as cable supports and catenary suspension devices, labeled LmtArea1 and LmtArea2. The system analyzes the range of freedom of each joint's motion, determines the motion limit space (MovLimitSpace) of the robot's end effector based on the joint linkage relationships, and performs intersection and complement analysis between MovLimitSpace and LmtArea1 and LmtArea2. It then extracts spatial contour features and regional attribute features to establish a correlation mapping, corrects spatial coordinate deviations, and generates WorkSpaceData, which includes boundary coordinates and contour structures. This generates WorkSpaceData, defining a precise working range for the robot and improving the adaptability of the path planning algorithm in the narrow space of a tunnel.

[0140] Next, the system collects data from the robot's operation log database, applying traditional methods. The system acquires historical obstacle collision path data (ColPathData) from the path planning algorithm (i.e., the first path planning algorithm) and environmental change data (EnvChangeData) from the newly added temporary maintenance platform within the tunnel, including the 3D dimensions and spatial coordinates of the new obstacle. The system maps the trajectory segments from ColPathData to WorkSpaceData in the same preset spatial coordinate system, calculates the shortest spatial distance between the boundaries of each trajectory point, and extracts abnormal trajectory segments (AbnTrajSeg) by combining the trajectory direction change rate and curvature change value. It then extracts abnormal trajectory features such as the maximum deviation distance and continuous trajectory length from AbnTrajSeg, along with scene features from EnvChangeData. Through a feature association analysis model, correlation coefficients are calculated to determine the causal relationship between the new temporary maintenance platform and trajectory deviation. After quantifying the influence weights, an obstacle avoidance strategy update basis (UpdateBasis) is generated, including the type of conflict trigger and response requirements. This provides a clear direction for optimizing the algorithm's obstacle avoidance strategy and enhances the rationality of the robot's obstacle avoidance decisions.

[0141] Finally, the system parses the traditional data based on UpdateBasis. The obstacle avoidance strategy logic module of the algorithm constructs a logical association graph. It filters out target logic modules with insufficient adaptability by combining joint motion constraint rules and spatial constraint rules. For boundary crossing conflicts and attribute mismatch conflicts, the rule constraints are reconstructed. After hierarchical calibration of algorithm parameters, an initial second path planning algorithm is generated. Simulated trajectory data (SimTrajData) is obtained through trajectory simulation. SimTrajData is then double-fitted with joint spatial constraints, LmtArea1, and LmtArea2 for verification. After iterative optimization of parameters, the optimized algorithm is output. The path planning algorithm (i.e., the second path planning algorithm) successfully planned a working path that avoids all obstacles and meets the joint movement limits in the tunnel scenario test, thus optimizing the robot's motion planning capability in complex maintenance spaces.

[0142] In summary, this application provides a complete path planning algorithm update system from constraint acquisition to algorithm iteration. By deeply integrating the joint space constraints of the railway maintenance robot with the spatial location restrictions of the target maintenance scenario, it achieves precise delineation of the workspace based on the dual constraints of the robot's physical motion limits and the scene space boundaries, breaking through the technical limitations of traditional path planning algorithms that only consider the robot's motion capabilities or scene space limitations. Through correlation analysis of historical obstacle collision path data and environmental change data, an obstacle avoidance strategy update logic based on actual operational conflict causes is constructed, realizing a fundamental shift in obstacle avoidance strategy from "experience-based pre-setting" to "data-driven," and improving the adaptability of the path planning algorithm in complex railway maintenance scenarios.

[0143] Meanwhile, during the algorithm optimization process, dual verification of joint motion restriction rules and spatial restriction rules is introduced to ensure that the output second path planning algorithm can not only meet the physical motion boundary constraints of the robot, but also accurately adapt to the spatial restriction requirements of the scene. This enhances the rationality and accuracy of the robot's obstacle avoidance decision, optimizes the motion planning capability in complex maintenance spaces, and provides full-process algorithm support for the stable operation of railway maintenance robots.

[0144] Based on the same inventive concept, embodiments of this application also provide a path planning algorithm update system. See also... Figure 2 As shown, it is a schematic diagram of the structure of a possible path planning algorithm update system provided in an embodiment of this application. Figure 2 In the path planning algorithm update system 200, there are a processor 210 and a memory 220. The processor 210 and the memory 220 are connected to each other via a communication bus. The memory 220 stores computer programs that can be executed by the processor 210. By executing the instructions stored in the memory 220, the processor 210 can perform the steps of the path planning algorithm update method for railway maintenance robots described above.

[0145] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium including a computer program. When the computer program is run on a path planning algorithm update system, it causes the path planning algorithm update system to perform the steps of the aforementioned path planning algorithm update method for railway maintenance robots. In some possible implementations, various aspects of the path planning algorithm update method for railway maintenance robots provided in this application can also be implemented as a program product, including a computer program. When the program product is run on a path planning algorithm update system, the computer program causes the path planning algorithm update system to perform the steps of the aforementioned path planning algorithm update method for railway maintenance robots. For example, the path planning algorithm update system can perform, for example,... Figure 1 The steps are shown in the diagram. The computer-readable storage medium includes volatile or non-volatile or a combination thereof, and may be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), flash memory or other memory technologies, CD-ROM, Digital Video Disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium.

[0146] like Figure 3 The diagram shown is a functional block diagram of the path planning algorithm update system provided in this embodiment of the application. The path planning algorithm update system includes a path planning algorithm update device, which includes:

[0147] The constraint determination module is used to obtain the joint space constraints of the railway maintenance robot in the target maintenance scenario and determine the spatial position restriction area contained in the target maintenance scenario.

[0148] The motion space determination module is used to analyze the range of motion degrees of freedom of each joint of the railway maintenance robot based on the joint space constraints, and determine the motion limit space of the railway maintenance robot by combining the correlation between the ranges of motion degrees of freedom.

[0149] The spatial boundary generation module is used to perform intersection and complement analysis of the spatial range of the motion limit space and the spatial position restriction area to generate the maintenance operation spatial boundary data of the railway maintenance robot in the target maintenance scenario.

[0150] The multi-dimensional data acquisition module is used to collect historical obstacle collision path data corresponding to the railway maintenance robot when performing maintenance operations using the first path planning algorithm in the target maintenance scenario, and to acquire the current environmental change data of the target maintenance scenario.

[0151] The update basis determination module is used to compare the trajectory segments in the historical obstacle collision path data with the maintenance operation space boundary data to extract abnormal trajectory segments, combine the abnormal trajectory features of the abnormal trajectory segments with the scene features in the environmental change data to determine the path conflict correlation, and determine the obstacle avoidance strategy update basis of the first path planning algorithm based on the path conflict correlation.

[0152] The algorithm strategy update module is used to reconstruct and adjust the obstacle avoidance strategy of the first path planning algorithm by combining the motion restriction rules corresponding to the joint space constraints and the spatial restriction rules corresponding to the maintenance operation space boundary data, based on the obstacle avoidance strategy update criteria, and output a second path planning algorithm that satisfies the joint space constraints and matches the spatial position restriction area.

[0153] Accordingly, this application also provides a computer program product, which includes a computer program or instructions that, when executed by a processor, cause the processor to implement the steps in the above method embodiments. It should be understood that each step or combination of steps in the above method flow can be implemented by the computer program or instructions. Furthermore, these computer programs or instructions can be applied to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device, enabling the processor of the general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to function as an apparatus for implementing the corresponding functions in the above method embodiments.

[0154] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0155] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0156] Finally, it should be noted that the above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A path planning algorithm update method for railway maintenance robots, characterized in that, The method includes: Obtain the joint space constraints of the railway maintenance robot in the target maintenance scenario, and determine the spatial position restriction area contained in the target maintenance scenario; Based on the joint space constraints, the range of motion degrees of freedom of each joint of the railway maintenance robot is analyzed, and the motion limit space of the railway maintenance robot is determined by combining the correlation between the ranges of motion degrees of freedom. By performing a spatial range intersection and complement analysis on the motion limit space and the spatial position restriction area, the maintenance operation space boundary data of the railway maintenance robot under the target maintenance scenario is generated. Collect historical obstacle collision path data corresponding to the railway maintenance robot performing maintenance operations using the first path planning algorithm in the target maintenance scenario, and obtain the current environmental change data of the target maintenance scenario; The trajectory segments in the historical obstacle collision path data are compared with the maintenance operation space boundary data to extract abnormal trajectory segments. The abnormal trajectory features of the abnormal trajectory segments and the scene features in the environmental change data are combined to determine the path conflict correlation. The obstacle avoidance strategy update basis of the first path planning algorithm is determined based on the path conflict correlation. Based on the obstacle avoidance strategy update criteria, combined with the motion restriction rules corresponding to the joint space constraints and the space restriction rules corresponding to the maintenance operation space boundary data, the obstacle avoidance strategy of the first path planning algorithm is reconstructed and the parameters are adjusted to output a second path planning algorithm that satisfies the joint space constraints and matches the spatial position restriction area. The step of performing a spatial complement analysis of the motion limit space and the spatial position restriction area to generate the maintenance operation space boundary data of the railway maintenance robot in the target maintenance scenario includes: Extract the spatial contour features of the motion limit space and the regional attribute features of the spatial position restriction area, and combine the spatial contour features and the regional attribute features to establish a spatial feature association mapping between the motion limit space and the spatial position restriction area; Based on the spatial feature association mapping, the intersection operation of the motion limit space and the spatial position restriction area is performed to obtain the spatial coordinate information and the first contour structure information of the overlapping part of the motion limit space and the spatial position restriction area, and to generate initial overlapping space data. Perform spatial range complementation on the motion limit space, delete the part of the motion limit space that overlaps with the spatial position restriction area, determine the spatial coordinate information and second contour structure information of the area in the motion limit space that is not covered by the spatial position restriction area, and generate initial non-overlapping spatial data; The initial overlapping spatial data and the initial non-overlapping spatial data are corrected for spatial coordinate deviation and contour structure misalignment to obtain corrected spatial feature information; Based on the corrected spatial feature information, the boundary of the spatial range in which the railway maintenance robot can move is defined, the key coordinate points and contour segments of the spatial range boundary are extracted, and maintenance operation spatial boundary data containing spatial boundary coordinates, boundary contour structure and spatial area attributes is generated. The maintenance operation space boundary data is associated with the motion degree of freedom of the motion limit space and the position restriction features of the spatial position restriction area, and the associated labels are marked.

2. The method as described in claim 1, characterized in that, The step of establishing a spatial feature association mapping between the motion limit space and the spatial position restriction region by combining the spatial contour features and the regional attribute features includes: The spatial contour parameter set is obtained by extracting the contour vertex coordinates, contour edge slope parameters, and contour closed region area parameters of the motion limit space to perform contour parameterization processing. The region attribute features are parsed to obtain a set of region attribute parameters, including the restriction level parameter, space occupancy ratio parameter, and smoothness parameter of the region boundary. Based on the range of degrees of freedom of motion corresponding to the motion limit space and the positional restriction characteristics of the spatial position restriction region, the association weight between the spatial contour parameter set and the region attribute parameter set is determined. The association weight is used to characterize the degree of influence of each parameter on spatial feature matching. The parameters in the spatial contour parameter set and the regional attribute parameter set are weighted and fused according to the association weights to obtain fused feature parameters; The fused feature parameters are input into a preset spatial feature mapping model, and a one-to-one correspondence is established between the spatial contour features of the motion limit space and the regional attribute features of the spatial position restriction area through model calculation; An initial spatial feature association mapping is generated based on the one-to-one correspondence. The initial spatial feature association mapping is then calibrated by combining the spatial coordinate information of the motion limit space and the spatial coordinate information of the spatial position restriction area to correct the parameter deviations that occur during the mapping process, thereby obtaining the spatial feature association mapping between the motion limit space and the spatial position restriction area.

3. The method according to any one of claims 1-2, characterized in that, The step of comparing the trajectory segments in the historical obstacle collision path data with the maintenance operation space boundary data to extract abnormal trajectory segments, combining the abnormal trajectory features of the abnormal trajectory segments with the scene features in the environmental change data to determine the path conflict correlation, and determining the basis for updating the obstacle avoidance strategy of the first path planning algorithm based on the path conflict correlation includes: The trajectory segments in the historical obstacle collision path data and the maintenance operation space boundary data are both mapped to a preset spatial coordinate system to obtain the three-dimensional coordinate sequence of each trajectory segment and the boundary coordinate range of the maintenance operation space boundary data. Calculate the shortest spatial distance between each coordinate point in the three-dimensional coordinate sequence of each trajectory segment and the boundary coordinate range of the maintenance operation space boundary data, compare the shortest spatial distance with a preset distance threshold, and mark the trajectory segments with the shortest spatial distance less than the preset distance threshold as candidate abnormal trajectory segments; By calculating the trajectory direction change rate and trajectory curvature change value of the candidate abnormal trajectory segments, the trajectory trend analysis of the candidate abnormal trajectory segments is performed to determine whether the candidate abnormal trajectory segments have a trend of extending outside the range defined by the maintenance operation space boundary data. Candidate abnormal trajectory segments with such an extension trend are identified as abnormal trajectory segments. Extract the abnormal trajectory features of the abnormal trajectory segment; the abnormal trajectory features include the maximum distance of the trajectory deviating from the boundary of the maintenance work space, the length of the trajectory deviation, the speed change parameters during the trajectory deviation process, and the trajectory turning angle parameters; Extract scene features from the environmental change data; the scene features include the three-dimensional size parameters of the newly added obstacles, the spatial coordinates of the newly added obstacles, and the refractive index change parameters of the environmental medium; A feature association analysis model is constructed, and the abnormal trajectory features and the scene features are input into the feature association analysis model to calculate the correlation coefficient between each abnormal trajectory feature and each scene feature. Based on the correlation coefficient, feature combinations with a correlation coefficient greater than a preset correlation coefficient threshold are selected. Based on the feature combinations, the scene causes of abnormal trajectory segments are analyzed to obtain the causal relationship between abnormal trajectory segments and environmental changes. The causal relationship is defined as a path conflict relationship. The path conflict correlation is quantified to determine the influence weights corresponding to different path conflict correlations. Based on the influence weights and the deviation of the abnormal trajectory features, an obstacle avoidance strategy update basis is generated, which includes the conflict cause type, conflict impact range, conflict occurrence frequency, and conflict response requirements.

4. The method as described in claim 3, characterized in that, The method combines the degree of deviation of the influence weights and the abnormal trajectory features to generate an obstacle avoidance strategy update basis that includes conflict trigger type, conflict impact range, conflict frequency, and conflict response requirements, including: Obtain the obstacle avoidance strategy parameters currently used by the first path planning algorithm, and establish the association mapping relationship between the obstacle avoidance strategy parameters and the conflict cause type; Based on the aforementioned influence weights, each conflict trigger type is ranked by weight to determine the dominant and secondary conflict triggers. The trajectory risk level corresponding to the dominant conflict trigger is quantified by combining the degree of deviation of the abnormal trajectory characteristics. The trajectory risk level is calculated by comprehensively considering the ratio of the deviation distance to the preset safety distance and the ratio of the length of the deviation trajectory to the total length of a single maintenance operation trajectory. For the dominant conflict triggers and their corresponding trajectory risk levels, analyze the adaptability of the current obstacle avoidance strategy parameters of the first path planning algorithm, determine whether the obstacle avoidance strategy parameters can meet the conflict response requirements in the current scenario, and mark the parameters with insufficient adaptability as parameters to be adjusted, and determine the adjustment range of the parameters to be adjusted in combination with the secondary conflict triggers. Based on the marked parameters to be adjusted and the corresponding adjustment range, the conflict impact range, conflict frequency and quantified trajectory risk level are integrated to construct a structured strategy for updating the obstacle avoidance strategy. The structured strategy includes a description of the dominant conflict trigger, quantitative data of trajectory risk level, a list of parameters to be adjusted, parameter adjustment direction and core objectives of conflict response. The structured strategy is associated and labeled with the obstacle avoidance strategy logic of the first path planning algorithm to lock the algorithm module corresponding to each update content, and an obstacle avoidance strategy update basis is generated for rule reconstruction and parameter adjustment of the obstacle avoidance strategy of the first path planning algorithm.

5. The method according to any one of claims 1-2, characterized in that, The first path planning algorithm's obstacle avoidance strategy is restructured and its parameters adjusted based on the obstacle avoidance strategy update criteria, combined with the motion restriction rules corresponding to the joint space constraints and the spatial restriction rules corresponding to the maintenance operation space boundary data. This process outputs a second path planning algorithm that satisfies the joint space constraints and matches the spatial location restriction region, including: Based on the obstacle avoidance strategy update criterion, the obstacle avoidance strategy logic module of the first path planning algorithm is parsed, the parameter configuration information and rule constraints of each obstacle avoidance strategy logic module are extracted, and a logical relationship diagram between the obstacle avoidance strategy logic modules is established. Based on the motion restriction rules corresponding to the joint space constraints, a motion adaptability analysis is performed on each obstacle avoidance strategy logic module to determine whether the rule constraints of each obstacle avoidance strategy logic module are compatible with the motion restriction rules, and target logic modules with insufficient adaptability are extracted. Based on the spatial constraint rules corresponding to the maintenance operation space boundary data, a spatial adaptability analysis is performed on the target logic module. The spatial matching degree between the trajectory generation parameters output by the target logic module and the maintenance operation space boundary data is compared to determine the conflict points and conflict types between the trajectory generation parameters and the spatial constraint rules. Based on the conflict points and conflict types, and according to the list of parameters to be adjusted, the direction of parameter adjustment, and the core objectives of conflict response in the obstacle avoidance strategy update basis, the rule constraints of the target logic module are reconstructed to optimize the logic association diagram; the parameters corresponding to the reconstructed rule constraints are calibrated in layers, and multi-layer calibration thresholds are set in combination with the motion restriction rules and the spatial restriction rules. Through iterative calculation, the calibration values ​​of each parameter simultaneously meet the quantification requirements of the motion restriction rules and the boundary requirements of the spatial restriction rules. The reconstructed rule constraints and calibrated parameters are integrated into the parameter configuration library of the first path planning algorithm for parameter configuration replacement, thereby generating the initial second path planning algorithm; The initial second path planning algorithm is subjected to trajectory simulation calculation to obtain simulated trajectory data. The simulated trajectory data is then subjected to dual adaptation verification with the joint space constraints and the spatial position restriction area. Based on the dual adaptation verification, the parameters of the initial second path planning algorithm are optimized to output a second path planning algorithm that satisfies the joint space constraints and matches the spatial position restriction area.

6. The method as described in claim 5, characterized in that, The obstacle avoidance strategy logic module, which updates the obstacle avoidance strategy based on the first path planning algorithm, extracts the parameter configuration information and rule constraints of each obstacle avoidance strategy logic module and establishes a logical relationship graph between the obstacle avoidance strategy logic modules, including: Extract the algorithm module identification information corresponding to each update content included in the obstacle avoidance strategy update basis, locate all obstacle avoidance strategy logic modules related to the obstacle avoidance strategy in the first path planning algorithm based on the algorithm module identification information, and establish a module index list; Based on the module index list, the core code snippets and parameter configuration files of each obstacle avoidance strategy logic module are called sequentially. The parameter configuration information of each obstacle avoidance strategy logic module is extracted through code semantic analysis, and the rule constraints corresponding to each obstacle avoidance strategy logic module are extracted through rule clause parsing. The rule constraints include the spatial limit threshold for trajectory generation, motion parameter constraint standards, and conflict response triggering conditions. Cross-compare the parameter configuration information of each obstacle avoidance strategy logic module, identify parameter groups that are related between different modules and mark the parameter association type. The parameter association type includes direct parameter transmission association, indirect parameter dependency association and parameter feedback adjustment association. Based on the parameter association type, the data flow and control logic relationship between each obstacle avoidance strategy logic module are sorted out, the correspondence between input and output parameters between modules is determined, and an initial logic association graph is constructed with each obstacle avoidance strategy logic module as a node and the association relationship between modules as an edge. The initial logical association graph is supplemented with annotations of the core fields of parameter configuration information and key clauses of rule constraints corresponding to each node. Redundant association edges and isolated nodes in the initial logical association graph are eliminated by graph structure optimization algorithm to generate a logical association graph between obstacle avoidance strategy logical modules that represents the interaction relationship between each obstacle avoidance strategy logical module.

7. The method as described in claim 5, characterized in that, The spatial adaptability analysis of the target logic module is performed by combining the spatial constraint rules corresponding to the maintenance operation space boundary data. The spatial matching degree between the trajectory generation parameters output by the target logic module and the maintenance operation space boundary data is compared to determine the conflict points and conflict types between the trajectory generation parameters and the spatial constraint rules, including: The corresponding spatial restriction rules are parsed from the boundary data of the maintenance operation space. The spatial restriction rules include the boundary coordinate range of the maintenance operation space, spatial area attribute constraints, and spatial topology requirements for trajectory passage. The spatial restriction rules are transformed into a quantifiable set of spatial adaptability evaluation indicators, which include coordinate matching accuracy, area attribute fit, and topology adaptability. Call the output interface of the target logic module to obtain the trajectory generation parameters output by the target logic module. The trajectory generation parameters include the three-dimensional coordinates of the trajectory key points, the curvature parameters of the trajectory segments, and the attribute identifier of the trajectory travel area. The three-dimensional coordinates of the trajectory key points in the trajectory generation parameters are mapped to the spatial coordinate system corresponding to the maintenance operation space boundary data. The coordinate matching deviation between each trajectory key point and the boundary coordinate range of the maintenance operation space boundary data is calculated. At the same time, the fit weight between the trajectory passage area attribute identifier and the spatial area attribute constraint of the maintenance operation space boundary data is compared. Combined with the adaptation weight between the trajectory segment curvature parameter and the spatial topology requirements of the maintenance operation space, the spatial matching degree between the trajectory generation parameters and the maintenance operation space boundary data is obtained through weighted calculation. Based on the threshold standard of the spatial adaptability evaluation index set, the trajectory key points and trajectory segments corresponding to the trajectory generation parameters with spatial matching degree lower than the preset threshold are screened out, and the spatial positions corresponding to the trajectory key points and trajectory segments are determined as the conflict points between the trajectory generation parameters and the spatial restriction rules. The conflict points are subjected to feature analysis. Based on whether the conflict points exceed the boundary coordinate range of the maintenance work space, whether they violate the spatial area attribute constraints, and whether they do not meet the spatial topology requirements, the conflict types are divided into three categories: boundary crossing conflict, attribute mismatch conflict, and topology adaptation failure conflict. At the same time, the spatial coordinate information of each type of conflict point and the corresponding trajectory generation parameter values ​​are recorded to form a structured analysis report of conflict points and conflict types.

8. A path planning algorithm update system, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform the steps of the path planning algorithm update method for railway maintenance robots according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a computer program that, when run on the path planning algorithm update system, causes the path planning algorithm update system to perform the steps of the path planning algorithm update method for railway maintenance robots as described in any one of claims 1 to 7.

Images