An obstacle avoidance path planning system for an overhead line maintenance robot

Abstract

The application relates to the technical field of intelligent sensing and control equipment, in particular to an obstacle avoidance path planning system of an overhead line maintenance robot. The system comprises the following: a data modeling module constructs an environment map and a robot model, and obtains an AABB bounding box of an obstacle and the robot; a data acquisition module acquires posture, motion variable and obstacle distance data of the robot; a data analysis module analyzes collision risk and vibration characteristics according to the acquired data, and calculates a pose swing rotation degree and a directional inflation factor adjustment coefficient; and a collision update and path generation module dynamically updates the AABB bounding box based on the adjustment coefficient, and performs path search to plan an obstacle avoidance path. The application solves the problem that a traditional fixed bounding box cannot adapt to a motion state, leading to easy failure of path planning and missed collision reporting, and improves the environmental adaptability and safety of obstacle avoidance path planning of the maintenance robot.

Description

Technical Field

[0001] This application relates to the field of intelligent sensing and control equipment technology, specifically to an obstacle avoidance path planning system for an overhead line maintenance robot. Background Technology

[0002] Overhead transmission lines form the backbone of modern power grids, characterized by numerous points, long lines, and wide coverage. They are exposed to complex natural environments for extended periods, inevitably facing threats from various factors such as lightning strikes, icing, wind loads, external damage, and natural aging of equipment, which can lead to line faults. Overhead transmission lines also contain components such as pipe fittings, insulator strings, and vibration dampers, all of which are obstacles encountered by maintenance robots during inspections. The key technology for autonomous obstacle crossing lies in the ability to plan obstacle-crossing paths under different obstacle environments.

[0003] For obstacle avoidance path planning in overhead power lines with various fittings, the AABB (Axis-Aligned Bounding Box) algorithm not only has a simple data structure suitable for hierarchical organization, but also has low collision testing complexity. However, in the obstacle avoidance path planning of overhead power line maintenance robots, the traditional AABB algorithm only considers the minimum and maximum points of the bounding box, and cannot dynamically adjust the size of the bounding box according to the robot's real-time motion state. This can lead to excessive encroachment on the robot's feasible obstacle avoidance space, causing frequent failures in path searching in narrow sections or deadlocks. At the same time, the influence of external environmental factors such as wind vibration can cause conductor vibration, which can cause the maintenance robot to sway. The uncertainty of the robot's pose may cause the actual space occupied to exceed the bounding box range, resulting in "missed" collision detection and causing safety accidents. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide an obstacle avoidance path planning system for an overhead line maintenance robot, and the specific technical solution adopted is as follows: This application proposes an obstacle avoidance path planning system for an overhead line maintenance robot. The system includes a data modeling module, a data acquisition module, a data analysis module, and a collision update and path generation module. The data modeling module is used to construct an environmental map of overhead lines and a kinematic model of the maintenance robot, and to obtain the AABB bounding box of obstacles and the maintenance robot. The data acquisition module is used to collect the posture data, joint variable data, and distances between obstacles of the maintenance robot. The data analysis module is used to analyze the collision risk, pose speed improvement status and alternating tension and compression oscillation degree of the maintenance robot based on the collected data, and then obtain the pose sway rotation degree and directional expansion factor adjustment coefficient of the maintenance robot. The collision update and path generation module is used to update the directional expansion factor of the AABB bounding box of the maintenance robot according to the directional expansion factor adjustment coefficient, and perform path search based on the updated AABB bounding box to plan the obstacle avoidance path of the maintenance robot.

[0005] Preferably, the data modeling module performs the following steps: Collect the 3D point cloud of the overhead line environment where the maintenance robot is located, extract the vertices, center points and three-axis radii of each obstacle, construct the AABB bounding box of each obstacle and inject it into the global octree environment map; Based on the URDF kinematic model of the maintenance robot, the maintenance robot is decomposed into multiple rigid body sub-components, and AABB primitives of each rigid body sub-component are established. The composite AABB bounding box of the entire maintenance robot is synthesized in real time through forward kinematic chain transformation.

[0006] Preferably, in the data acquisition module, attitude data is acquired by an IMU attitude sensor installed inside the electrical box of the maintenance robot, motion joint variable data is acquired by a rotary encoder deployed at the shaft ends of the robotic arm and walking arm of the maintenance robot, and distance data between obstacles is acquired by an ultrasonic ranging sensor deployed at the front end of the robotic arm of the maintenance robot.

[0007] Preferably, the process by which the data analysis module analyzes the collision risk of the maintenance robot is as follows: Based on the relative distance between the maintenance robot and the nearest obstacle, and the temporal fluctuation characteristics of the three-axis radius of the obstacle caused by external environmental excitation, the quantitative results of the collision risk of the maintenance robot are determined; among them, the quantitative results of the collision risk are negatively correlated with the relative distance and positively correlated with the temporal fluctuation characteristics of the three-axis radius.

[0008] Preferably, the process by which the data analysis module analyzes the improvement status of the maintenance robot's pose and speed is as follows: Perform time-series trend analysis on the three-axis acceleration data of the maintenance robot to obtain the intensity of the acceleration increase trend; Temporal feature analysis was performed on the attitude angle data of the maintenance robot to obtain the degree of instability of the attitude angle change rate; Based on the intensity of the acceleration increase trend and the instability of the attitude angle change rate, a quantitative result of the attitude velocity increase status is calculated.

[0009] Preferably, the specific process by which the data analysis module analyzes the degree of alternating tensile and compressive oscillations is as follows: Frequency domain analysis was performed on the rotational angular velocity data of the joints of the maintenance robot to extract high-frequency oscillation characteristics; Obtain the clearance distance between the moving joint and its corresponding AABB bounding box; The degree of oscillation between alternating tensile and compressive forces is positively correlated with the high-frequency oscillation characteristics and negatively correlated with the gap distance.

[0010] Preferably, the pose swing rotation degree is the product of the quantitative result of the pose speed improvement of the maintenance robot and the degree of alternating tension and compression oscillation.

[0011] Preferably, the method for obtaining the directional expansion factor adjustment coefficient is as follows: The quantification of collision risk is used as a weighting factor, and combined with the pose sway rotation degree for weighted calculation to obtain the directional expansion factor adjustment coefficient.

[0012] Preferably, the process by which the collision update and path generation module updates the directional expansion factor of the AABB bounding box of the maintenance robot is as follows: The directional expansion factor of each pre-set planning period is adjusted using a smoothing coefficient and a directional expansion factor adjustment coefficient, and the adjusted value is mapped to a pre-set expansion ratio range to obtain the directional expansion factor mapping value of the AABB bounding box of the maintenance robot in each planning period.

[0013] Preferably, the collision update and path generation module performs path search as follows: The RRT-connect algorithm is used to search for paths on a global octree environment map. When a candidate path node is generated, the configuration of the joints of the maintenance robot is obtained and the pose of each joint is calculated using forward kinematics. The AABB bounding box of the maintenance robot is updated using the directional dilation factor mapping value within each planning time period, and collision detection of the AABB hierarchical bounding box tree is performed. When no collision occurs, the candidate path node is added to the random tree. When a collision occurs, the candidate path node is removed to obtain the final obstacle avoidance path.

[0014] This application has the following beneficial effects: This application constructs a comprehensive quantitative index system that includes collision risk status, pose velocity improvement status, and alternating tension and compression oscillation degree through multi-source data fusion analysis. It then solves the pose sway rotation degree and directional expansion factor adjustment coefficient, effectively solving the shortcomings of the traditional AABB bounding box algorithm, which cannot adapt to the robot's real-time motion state and external wind vibration interference due to the use of fixed size. This method can not only accurately quantify and characterize the robot's dynamic collision risk and pose uncertainty in each directional axis, but also realize the adaptive dynamic expansion of the bounding box size according to the working conditions. While ensuring obstacle avoidance safety margin, it can effectively release narrow working space, which helps to improve the environmental adaptability and operational safety of the obstacle avoidance path planning system. This application introduces a smoothing filter and interval mapping mechanism to achieve adaptive updating of the directional expansion factor of the AABB bounding box. The updated directional expansion factor is used to dynamically adjust the size of each primitive AABB and synthesize the whole bounding box. Then, the RRT-connect algorithm and the collision detection mechanism of the AABB hierarchical bounding box tree are combined to perform path search. This effectively solves the problem of collision underreporting and path search deadlock that is easily caused by traditional fixed-size bounding boxes in complex vibration environments, and improves the environmental adaptability and safety of obstacle avoidance path planning of maintenance robots. Attached Figure Description

[0015] Figure 1 A block diagram of an obstacle avoidance path planning system for an overhead line maintenance robot provided in one embodiment of this application; Figure 2 This is a schematic diagram illustrating the process of extracting the directional expansion factor adjustment coefficient according to an embodiment of this application. Detailed Implementation

[0016] The following description, in conjunction with the accompanying drawings, details the specific scheme of the obstacle avoidance path planning system for an overhead line maintenance robot provided in this application.

[0017] Please see Figure 1 The diagram shows a block diagram of an obstacle avoidance path planning system for an overhead line maintenance robot according to an embodiment of this application. The system includes: a data modeling module 101, a data acquisition module 102, a data analysis module 103, and a collision update and path generation module 104.

[0018] The data modeling module 101 is used to construct an environmental map of the overhead line and a kinematic model of the maintenance robot, and to obtain the AABB bounding boxes of obstacles and the maintenance robot.

[0019] The maintenance robot is deployed to the work site and mounted on the conductor. A solid-state LiDAR mounted on the front of the robot scans along the conductor axis, simultaneously collecting 3D point cloud data of the overhead line environment where the robot is located. The 3D point cloud includes 3D point cloud data of conductors, insulator strings, vibration dampers, spacers, and tower crossarms. The 3D point cloud is then processed sequentially with pass-through filtering, statistical outlier removal, and voxel downsampling. Euclidean clustering algorithm is used to segment the point cloud data corresponding to each obstacle. The extreme values, center points, and radii of all vertices of each obstacle on the three axes of the world coordinate system are extracted. Based on this, AABB bounding boxes of each obstacle are constructed, and a global octree environment map is built. The world axis coordinate system is a well-known technology and will not be described in detail here.

[0020] Meanwhile, based on the URDF kinematic model of the maintenance robot body, the whole machine is decomposed into multiple rigid body sub-components such as the walking arm, the working robot arm, the end tool and the control cabinet. Reference axes, reference points and local coordinate systems are established for each rigid body sub-component. The AABB primitives of each sub-component are constructed one by one, and the composite AABB bounding box of the whole maintenance robot is synthesized in real time through forward kinematic chain transformation.

[0021] In this embodiment, the AABB bounding box update frequency for each obstacle instance is set to 10Hz, and the composite AABB bounding box update frequency for the entire robot is set to 100Hz.

[0022] The process of obtaining the AABB bounding box and constructing the global octree environment map are well-known technologies and will not be described in detail here.

[0023] The data acquisition module 102 is used to collect the posture data, joint variable data, and distances between obstacles of the maintenance robot.

[0024] An IMU attitude sensor is installed inside the electrical box of the maintenance robot to acquire the three-axis acceleration and three-axis attitude angle (pitch angle, roll angle, yaw angle) data of the maintenance robot. The sensor coordinate system needs to be aligned with the body coordinate system of the maintenance robot, that is, the X-axis points in the forward direction, the Z-axis is vertically upward, and the Y-axis is determined according to the right-hand rule. The acquisition frequency of angular velocity and attitude angle is set to 100Hz.

[0025] Incremental and absolute rotary encoders are deployed on the ends of the robotic arm and each walking arm to acquire real-time variable data of each joint on the robotic arm, including the length and rotation angle of the robotic arm push rod, with the acquisition frequency set to 100Hz.

[0026] An ultrasonic ranging module is deployed at the front end of the maintenance robot arm to obtain the relative distance between the maintenance robot and the nearest obstacle in real time during the robot's movement.

[0027] To balance the real-time performance of high-frequency wind vibration response with the stability of inspection operations, a planning period of 2 seconds is used. The entropy weight method is used to normalize the collected data to eliminate dimensional differences. The normalization range is set to include N consecutive planning periods (N=10) including the current planning period, in order to meet the needs of long-term and short-term trend analysis of bounding box collision risk.

[0028] The process of normalizing data using the entropy weight method is a well-known technique and will not be elaborated further.

[0029] The data analysis module 103 is used to analyze the collision risk, posture speed improvement status and alternating tension and compression oscillation degree of the maintenance robot based on the collected data, and then obtain the posture swing rotation degree and directional expansion factor adjustment coefficient of the maintenance robot.

[0030] In the obstacle avoidance path planning process of overhead line maintenance robots, the relative distance and orientation relationship between the maintenance robot and obstacles in space determines the degree of collision risk, which in turn determines the expansion adjustment requirements of the AABB bounding box on each world coordinate axis; at the same time, the three-axis radius of the obstacle instance can effectively characterize the torsional swaying of the slender structure around the suspension point under the coupling effect of external wind vibration and conductor vibration excitation.

[0031] Specifically, during the movement of the inspection robot, the collision risk increases as the distance between the robot and the nearest obstacle instance decreases due to the combined effects of robot motion, wire excitation, and external wind vibration load. This increases the urgency of the expansion requirement of the AABB bounding box on each axis. At the same time, the three-axis radius fluctuation of the obstacle instance intensifies and exhibits significant anisotropic peak characteristics.

[0032] Based on this, this embodiment analyzes the collision risk of the maintenance robot during its movement to characterize the urgency of obstacle avoidance and the level of collision risk increase caused by the coupling effect of multiple factors. Based on the relative distance between the maintenance robot and the nearest obstacle, and the temporal fluctuation characteristics of the three-axis radius of the obstacle caused by external environmental excitation, the quantitative result of the collision risk of the maintenance robot is determined. The calculation formula for the quantitative result of the collision risk is as follows: ; In the formula, Indicates the first Quantitative results of the collision risk of maintenance robots within a planning period; This represents the distance between the maximum vertex of the maintenance robot along the direction of travel of the guide wire and the center point of the bounding box of the nearest obstacle AABB during the i-th planning time period; This indicates the total number of axes used to measure the radius of the obstacle (fixed to 3 in this embodiment). Indicates the first The nearest obstacle to the maintenance robot within a planned time period. The mean DTW distance between the peak times of the corresponding radii of each directional axis and all other axes; Indicates the first The obstacle closest to the maintenance robot at all times within a planning period. Information entropy of the peak value in the radius corresponding to each directional axis; Let represent the normalization function, which uses the minimum-maximum normalization method to map the values ​​to the interval [0,1]. The normalization range is a continuous range including the i-th planning period. Each planning period ( (Take 10), to meet the needs of long-term and short-term trend analysis of collision risk; This represents a preset constant greater than 0, used to prevent the denominator from being 0. In this embodiment... The value is 0.01. In practical applications, as other implementation methods, implementers can also set it according to specific circumstances. This embodiment does not impose any special restrictions.

[0033] It should be noted that in the first Within each planned time period, the nearest obstacle to the maintenance robot at all given times is first counted. The probability distribution of the peak value in the radius corresponding to the bar axis is then used. The Shannon entropy calculation formula is then used to sum the product of the probability of each value in the probability distribution and its logarithm, and the negative value is taken. The result is the information entropy of the radius peak value. In this embodiment, the information entropy is calculated using Shannon entropy, which aims to quantify the degree of uncertainty in the distribution of obstacle radius peak data.

[0034] The calculation processes for DTW distance, Shannon entropy, and data normalization using the minimum-maximum normalization method are all well-known techniques and will not be elaborated further.

[0035] The quantification of collision risk for maintenance robots is influenced by the robot's movement, the excitation of wire motion, and external wind vibration loads, resulting in the spatial proximity between the robot and obstacles and the urgency of obstacle avoidance. It reflects the comprehensive risk level of collisions during the planning period. A higher collision risk indicates that the distance between the robot and the nearest obstacle is closer, and the degree of torsional swaying of the obstacle caused by environmental excitation is more severe. This means that the possibility of interference between the robot and the obstacle is extremely high, and the vigilance of obstacle avoidance path planning needs to be greatly improved. Conversely, a lower collision risk indicates that the robot's movement space is relatively open and the obstacle state is stable, resulting in a lower potential risk of collision. Path planning can focus on improving passage efficiency.

[0036] Since the coordinate axes of the AABB bounding box are strictly parallel to the world coordinate axes, its size is determined by the extreme differences of the object's vertices on each axis. During maintenance operations, the robot's center of gravity shifts due to the robot arm's movements, obstacle avoidance by the walking arm, and external wind vibrations. Furthermore, fluctuations in joint pose may cause the actual space occupied to exceed the bounding box range for a short period, leading to collisions. Therefore, further analysis of the robot's posture changes and real-time variables is necessary.

[0037] Specifically, when the maintenance robot is affected by the movement operation and external wind vibration, resulting in aggravated posture swaying and deflection, the acceleration data on the direction axis shows a clear upward trend, the attitude angle swing range increases, and the attitude angle change rate exhibits violent fluctuation characteristics.

[0038] Based on the above analysis, this embodiment quantifies the pose velocity improvement status to reflect the intensity of the acceleration improvement trend and the degree of instability of the attitude angle change rate caused by vibration affecting the maintenance robot. The calculation formula for the pose velocity improvement status is as follows: ; In the formula, Indicates the first The maintenance robot within the planned time period Quantification results of pose velocity improvement on each directional axis; This indicates the intensity of the acceleration increase trend of the maintenance robot on the k-th direction axis during the i-th planning period; It represents the product of the coefficient of variation of all elements in the first-order difference sequence of the attitude angles of the maintenance robot on the k directional axes during the i-th planning period and the range of the attitude angles.

[0039] The process of obtaining the acceleration improvement trend strength is as follows: Rank the acceleration data of the maintenance robot on any axis within each planning period. The smaller the acceleration data, the lower the corresponding ranking. At the same time, number all sampling times within each planning period in ascending order. Calculate the correlation coefficient between the ranking of the acceleration data on any axis within each planning period and the corresponding number of all sampling times. Use an exponential function with the natural constant as the base and the correlation coefficient as the exponent as the exponent, and use it as the acceleration improvement trend strength of the maintenance robot on any axis within each planning period.

[0040] In this embodiment, Spearman's correlation coefficient is used to calculate the correlation coefficient. In actual applications, as other implementation methods, implementers may choose other correlation coefficient calculation methods according to specific circumstances. This embodiment does not impose any special restrictions.

[0041] The calculation process of the Spearman correlation coefficient is a well-known technique and will not be elaborated further.

[0042] The intensity of the acceleration increase trend is used to characterize the monotonically increasing trend of acceleration data of the maintenance robot over time, caused by travel vibration and external wind vibration in each axis. It reflects the degree of deterioration of the maintenance robot's motion state. When the intensity of the acceleration increase trend is greater, it indicates that the acceleration data of the maintenance robot in the corresponding axis shows a significant and continuous increase, which means that the external excitation causes the robot's vibration amplitude to rise continuously, the vibration response to become more and more violent, and it is very easy to cause the risk of loss of control. At this time, it is necessary to increase the enclosure box expansion to ensure safety. Conversely, when the intensity of the acceleration increase trend is smaller, it indicates that the vibration of the maintenance robot is stable or the acceleration data has no obvious increasing trend, and the operating state is safe and controllable. At this time, the enclosure box expansion can be appropriately reduced to release the working space.

[0043] Furthermore, the movement and wind-induced vibrations will cause pose fluctuations in the joints of the maintenance robot. When the pose sway is severe, the vibration is transmitted to the push rod through the robotic arm, generating alternating tension and compression, resulting in drastic fluctuations in the push rod length and high-frequency oscillations in the rotation angle; at the same time, the gap between the end point of the robotic arm push rod and the planes of the AABB enclosure box in each direction decreases.

[0044] In response, this embodiment constructs an alternating tensile and compressive oscillation level to quantify the collision risk caused by real-time variable fluctuations in the joints of the maintenance robot.

[0045] The rotation angle of any joint on the maintenance robot at all sampling times within each planning period is processed by FFT Fourier transform and Hilbert transform. The sum of the differences between the energy of the main frequency component and the energy of the other frequency components is extracted and used as the high-frequency oscillation characteristic value of the rotation angle. At the same time, the minimum distance between the vertex of the push rod end of any joint on the maintenance robot and the three-directional plane of the AABB bounding box within each planning period is calculated and recorded as the gap distance.

[0046] The formula for calculating the degree of alternating tensile and compressive oscillation is as follows: In the formula, This indicates the degree of alternating tension and compression oscillation on the k-th direction axis of the maintenance robot during the i-th planning time period; It represents the mean of the product of the standard deviation of the length of the robotic arm push rod corresponding to all the motion joints of the maintenance robot and the high-frequency oscillation characteristic value of the rotation angle during the i-th planning period; Let $\frac{ ... This represents a preset constant greater than 0, used to prevent the denominator from being zero. Implementers can also set this value according to their specific needs, provided it does not affect the calculation. The value of .

[0047] The alternating tension and compression oscillation degree is used to characterize the combined severity of the fluctuations in the mechanical push rod size and the high-frequency oscillations in the rotation angle caused by the coupling effect of travel vibration and wind vibration on the motion joints of the maintenance robot. When the alternating tension and compression oscillation degree is larger, it indicates that the length fluctuation of the robot arm push rod is more severe and the high-frequency oscillation characteristics of the rotation angle are more significant. This means that the alternating load on the motion joints of the maintenance robot is large and the mechanical clearance is extremely small, and the probability of interference or collision is extremely high. The size of the enclosure box must be increased to avoid risks. Conversely, when the alternating tension and compression oscillation degree is smaller, it indicates that the mechanical joint transmission is smooth, the force is uniform, and there is sufficient clearance between the joint and the enclosure box wall. The risk of the actual space occupation exceeding the enclosure box is low. At this time, the expansion factor can be tightened to improve the flexibility of path planning.

[0048] Furthermore, the pose swing rotation is calculated using the following formula: In the formula, Indicates the first The maintenance robot within the planned time period pose yaw rotation on each directional axis; Indicates the first The maintenance robot within the planned time period Quantification results of pose velocity improvement on each directional axis; Indicates the first Within the planning period, the first The degree of alternating tension and compression oscillations along each directional axis; norm() represents the normalization function. Normalization method and Consistency: This indicator, through the fusion of multiple factors, comprehensively reflects the risk that the robot's actual space occupancy exceeds the bounding box range.

[0049] The pose yaw rotation degree is used to characterize the overall posture instability and local joint sway of the maintenance robot caused by the coupling effect of travel vibration and external wind vibration on each axis. It reflects the possibility that the actual space occupied by the maintenance robot will break through the original AABB bounding box limit. When the pose sway and rotation degree is greater, it indicates that the overall acceleration of the maintenance robot is increasing rapidly and the joints are oscillating severely. This means that the end effector or protruding parts of the maintenance robot may experience a large displacement and deflection in an instant, which can easily cause the actual occupied space to exceed the static bounding box range, resulting in collisions that are missed. Therefore, the requirement for obstacle avoidance safety margin is extremely high. Conversely, when the pose sway and rotation degree is smaller, it indicates that the overall operating posture of the maintenance robot is stable and the joint movement fluctuations are small. This means that the actual movement trajectory of the robot is highly consistent with the theoretically planned path, and the space occupied boundary is stable. At this time, the bounding box expansion can be appropriately tightened to improve the passage ability in narrow spaces while ensuring safety.

[0050] Finally, the quantification result of the collision risk of the maintenance robot in the i-th planning period is used as the weight, and its product with the pose swing rotation degree corresponding to the i-th direction axis is calculated as the directional expansion factor adjustment coefficient of the maintenance robot in the i-th planning period. Preferably, the schematic diagram of the directional expansion factor adjustment coefficient extraction process provided in this embodiment is as follows: Figure 2 As shown.

[0051] The larger the directional expansion factor adjustment coefficient, the higher the collision risk of the maintenance robot, the more violent the obstacle fluctuations, and the more severe the robot pose swaying during the corresponding planning period. The higher the expansion requirement of the AABB bounding box in that directional axis direction, the more important it is to balance safety and bounding box tightness. Based on this, the expansion status of each world coordinate axis direction needs to be adjusted accordingly to obtain the dynamic bounding box directional expansion factor.

[0052] The collision update and path generation module 104 is used to update the directional expansion factor of the AABB bounding box of the maintenance robot according to the directional expansion factor adjustment coefficient, and perform path search based on the updated AABB bounding box to plan the obstacle avoidance path of the maintenance robot.

[0053] The directional expansion factor of each axis of the AABB enclosure of the maintenance robot is updated and adjusted using the directional expansion factor adjustment coefficient. The specific adjustment method is as follows: First, a first-order lag filtering algorithm is used to smoothly update the directional expansion factor to avoid frequent jumps in the factor value. The calculation formula is as follows: ; In the formula, This represents the adjusted directional expansion factor of the maintenance robot on the k-th direction axis during the i-th planning time period; Represents the smoothing coefficient. The purpose is to avoid frequent jumps in the directional expansion factor in this embodiment. The value is 0.3; This is the directional expansion factor of the maintenance robot on the k-th direction axis during the (i-1)-th planning time period (at initialization, the bounding box expansion factor in each world coordinate axis direction is set to 1 by default, indicating that the bounding box has not expanded). It represents the directional expansion factor adjustment coefficient of the maintenance robot on the k-th directional axis during the i-th planning period.

[0054] Then, Mapped to a preset expansion range The final directional expansion factor used for adjustment is obtained, and the calculation formula is as follows: ; In the formula, This represents the mapping value of the directional expansion factor on the k-th direction axis of the maintenance robot during the i-th planning time period; This represents the directional expansion factor of the maintenance robot on the k-th direction axis during the i-th planning time period; This represents the maximum expansion factor of the AABB enclosure box of the maintenance robot, with a value range of [1.3, 1.5], and 1.3 in this embodiment; The maximum value of the directional expansion factor on the k-th direction axis of the maintenance robot within N consecutive planning periods, including the i-th planning period and the preceding historical planning periods; when The smaller the value, the more stable the maintenance robot's movement and the more open its working space on overhead lines, with minimal impact from operational vibrations and external wind. In this case, the AABB enclosure should be tightened (i.e., the expansion should be reduced) to free up workspace, reduce computational load, and decrease the frequency of path searching. Conversely, when... The larger the value, the more severe the swaying of the maintenance robot during operation, the narrower the working space, and the more severely it is affected by operational vibrations and external wind vibrations. In this case, it is necessary to expand the AABB enclosure box to avoid collisions.

[0055] Based on the mapping value of the directional expansion factor, the AABB dimensions of each primitive in the maintenance robot are updated within each planning period. Specifically, the product of the AABB reference dimension of each primitive on any directional axis and the directional expansion factor mapping value on the corresponding directional axis is used as the expansion increment. The sum of the reference dimension and the increment is used as the updated AABB dimensions of each primitive on the maintenance robot. Subsequently, a bottom-up merging operation is performed on the updated AABB of each primitive to obtain the updated AABB bounding box of the entire robot.

[0056] The RRT-connect algorithm is used to search for paths on a global octree environment map. When a candidate path node is generated, the following processing is performed: the current joint configuration is obtained, the world pose of each joint is calculated using forward kinematics, the AABB bounding box of the robot is updated using the final directional expansion factor of each axis in the current planning time period, and collision detection of the AABB hierarchical bounding box tree is performed. If the detection result is no collision, the candidate node is added to the random tree; otherwise, the candidate node is removed. In this embodiment, the spatial step size of the maintenance robot for each movement is set to 30mm. Since the RRT-connect algorithm and the AABB axis-aligned bounding box algorithm are both well-known technologies, the specific principles will not be elaborated. Thus, obstacle avoidance path planning for the overhead line maintenance robot is realized through the above method.

Claims

1. An obstacle avoidance path planning system for an overhead power line maintenance robot, characterized in that, The system includes a data modeling module, a data acquisition module, a data analysis module, and a collision update and path generation module. The data modeling module is used to construct an environmental map of overhead lines and a kinematic model of the maintenance robot, and to obtain the AABB bounding box of obstacles and the maintenance robot. The data acquisition module is used to collect the posture data, joint variable data, and distances between obstacles of the maintenance robot. The data analysis module is used to analyze the collision risk, pose speed improvement status and alternating tension and compression oscillation degree of the maintenance robot based on the collected data, and then obtain the pose sway rotation degree and directional expansion factor adjustment coefficient of the maintenance robot. The collision update and path generation module is used to update the directional expansion factor of the AABB bounding box of the maintenance robot according to the directional expansion factor adjustment coefficient, and perform path search based on the updated AABB bounding box to plan the obstacle avoidance path of the maintenance robot.

2. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 1, characterized in that, The data modeling module performs the following steps: Collect the 3D point cloud of the overhead line environment where the maintenance robot is located, extract the vertices, center points and three-axis radii of each obstacle, construct the AABB bounding box of each obstacle and inject it into the global octree environment map; Based on the URDF kinematic model of the maintenance robot, the maintenance robot is decomposed into multiple rigid body sub-components, and AABB primitives of each rigid body sub-component are established. The composite AABB bounding box of the entire maintenance robot is synthesized in real time through forward kinematic chain transformation.

3. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 1, characterized in that, In the data acquisition module, attitude data is acquired through an IMU attitude sensor installed inside the electrical box of the maintenance robot, motion joint variable data is acquired through a rotary encoder deployed at the shaft ends of the robotic arm and walking arm of the maintenance robot, and distance data between obstacles is acquired through an ultrasonic ranging sensor deployed at the front end of the robotic arm of the maintenance robot.

4. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 2, characterized in that, The data analysis module analyzes the collision risk of the maintenance robot as follows: Based on the relative distance between the maintenance robot and the nearest obstacle, and the temporal fluctuation characteristics of the three-axis radius of the obstacle caused by external environmental excitation, the quantitative results of the collision risk of the maintenance robot are determined; among them, the quantitative results of the collision risk are negatively correlated with the relative distance and positively correlated with the temporal fluctuation characteristics of the three-axis radius.

5. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 2, characterized in that, The data analysis module analyzes the improvement of the maintenance robot's pose and speed as follows: Perform time-series trend analysis on the three-axis acceleration data of the maintenance robot to obtain the intensity of the acceleration increase trend; Temporal feature analysis was performed on the attitude angle data of the maintenance robot to obtain the degree of instability of the attitude angle change rate; Based on the intensity of the acceleration increase trend and the instability of the attitude angle change rate, a quantitative result of the attitude velocity increase status is calculated.

6. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 1, characterized in that, The specific process by which the data analysis module analyzes the degree of alternating tensile and compressive oscillations is as follows: Frequency domain analysis was performed on the rotational angular velocity data of the joints of the maintenance robot to extract high-frequency oscillation characteristics; Obtain the clearance distance between the moving joint and its corresponding AABB bounding box; The degree of oscillation between alternating tensile and compressive forces is positively correlated with the high-frequency oscillation characteristics and negatively correlated with the gap distance.

7. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 5, characterized in that, The pose swing rotation degree is the product of the quantitative result of the pose speed improvement of the maintenance robot and the degree of alternating tension and compression oscillation.

8. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 4, characterized in that, The method for obtaining the directional expansion factor adjustment coefficient is as follows: The quantification of collision risk is used as a weighting factor, and combined with the pose sway rotation degree for weighted calculation to obtain the directional expansion factor adjustment coefficient.

9. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 1, characterized in that, The process by which the collision update and path generation module updates the directional expansion factor of the AABB bounding box of the maintenance robot is as follows: The directional expansion factor of each pre-set planning period is adjusted using a smoothing coefficient and a directional expansion factor adjustment coefficient, and the adjusted value is mapped to a pre-set expansion ratio range to obtain the directional expansion factor mapping value of the AABB bounding box of the maintenance robot in each planning period.

10. The obstacle avoidance path planning system for an overhead line maintenance robot according to claim 9, characterized in that, The collision update and path generation module performs path search as follows: The RRT-connect algorithm is used to perform path search on the global octree environment map. When a candidate path node is generated, the configuration of the joints of the maintenance robot is obtained and the pose of each joint is calculated using forward kinematics. The AABB bounding box of the maintenance robot is updated using the directional expansion factor mapping value in each planning time period, and AABB hierarchical bounding box tree collision detection is performed. When no collision occurs, the candidate path node is added to the random tree. When a collision occurs, the candidate path node is removed to obtain the final obstacle avoidance path.

Images