Optimization method and system for stopping and operating point pose of distribution network live working robot

Abstract

The application provides a method and system for optimizing a stop position of a distribution network live working robot and an operation point position, and belongs to the field of intelligent operation and maintenance of power systems and robots. The method comprises the following steps: constructing a three-dimensional global environment model of a working area and locking an effective working space range; generating a candidate operation point position variable set and a candidate stop position variable set in the working space range; constructing a continuous non-uniform live working safety constraint potential field model, and calculating a comprehensive risk value of any point in the working space; taking the operation point position and the stop position as coupled variables for multidimensional constraint checking to obtain a joint feasible region; constructing a multidimensional normalized operation performance evaluation function; and solving an optimal operation point position and an optimal stop position by minimizing the comprehensive risk and maximizing the operation performance. The application realizes the collaborative optimization of the stop position and the operation point position, and improves the safety of live working, the operation stability of a mechanical arm and the overall working efficiency.

Description

Technical Field

[0001] This invention belongs to the field of intelligent operation and maintenance and robotics technology of power systems, and particularly relates to a method and system for optimizing the stopping position and operation point posture of a live-line working robot in a distribution network. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Live-line working on power distribution networks is a crucial operational method for ensuring uninterrupted power supply and improving power reliability. With the development of intelligent power operation and maintenance, live-line working robots mounted on boom trucks have gradually replaced manual labor in high-risk operations such as wire disconnection, wiring, and hardware installation, becoming the mainstream operational equipment. Currently, robot operation posture planning generally adopts a one-way process of first determining the operation point and then reversing the bucket's stopping posture: the boom truck delivers the robot to the vicinity of the work area, completes environmental mapping using LiDAR and vision, the operator manually selects the work area, the system discretely generates operation points within the area, and finally, based on the operation point positions, the boom truck's stopping posture is calculated in reverse, and then the robotic arm performs the operation.

[0004] Existing pose planning methods have significant technical flaws. First, the stop pose and the operating point pose are isolated, with the stop position constrained only by the operating point in one direction. The reverse impact of the stop pose on the accessibility of the operating point and the robot arm's posture is not considered, resulting in a lack of global collaborative optimization and a tendency for the operating point to be feasible but the robot arm unable to operate after stopping. Second, the live-line safety constraint modeling is crude, often relying on fixed safety distance thresholds for simple judgments without considering voltage levels, electric field spatial distribution, and the risk of multiple conductor superposition. This fails to achieve continuous risk quantification and poses safety hazards in complex scenarios such as multiple lines and densely packed towers. Third, the robot arm's operability evaluation is lacking, only judging whether a point is "accessible" without comprehensively considering joint comfort, motion singularity, load torque, and visual visibility. This can easily lead to the robot arm being in extreme postures, singular poses, or high-load states, reducing operational stability and efficiency. Furthermore, existing methods lack a unified multi-objective optimization model, failing to globally balance safety, operability, and operational efficiency. They rely heavily on local rules or manual experience for selection, making it difficult to obtain the globally optimal operational solution. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a method and system for optimizing the stopping position and operating point posture of a live-line working robot in power distribution networks. Under the constraint of manually limited working area, it realizes joint modeling and collaborative optimization of the operating point posture and the stopping position posture, thereby improving the safety of operation, operational stability and overall work efficiency.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: The first aspect of this invention provides a method for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network; Methods for optimizing the stopping bucket and operating point pose of live-line working robots in power distribution networks include: The environmental modeling step is used to construct a three-dimensional global environmental model of the work area and to lock the effective work space range based on the model; The candidate pose generation step is used to generate a discretized set of candidate operation point pose variables within the work space, and to generate a set of candidate bucket stop pose variables based on the kinematic model of the bucket truck. The steps for constructing a safety risk field are used to build a continuous non-uniform energized safety constraint potential field model. The energized conductor is used as the risk source, and the comprehensive risk value of any point in the work space is calculated based on the voltage level and spatial distance. The joint feasible region construction step is used to construct a two-dimensional combined joint space by taking the operation point pose variable and the stop pose variable as coupling variables, and to perform multi-dimensional constraint verification on each combination in combination with the comprehensive risk value to obtain the joint feasible region. The operational performance evaluation step is used to construct a multi-dimensional normalized operational performance evaluation function within the joint feasible domain to quantitatively evaluate the operational quality of each feasible combination. The collaborative optimization decision-making steps are used to construct a multi-objective collaborative optimization model with the joint feasible domain as a constraint and the objectives of minimizing comprehensive risk and maximizing operational performance. The optimal operational point pose and the optimal stopping pose are obtained by solving the optimization model in the coupled variable space.

[0007] As a further technical solution, a three-dimensional global environment model of the work area is constructed, and the effective work space range is locked based on the model, including: By collecting 3D point cloud data of the work area using visual sensors and lidar, the contours and spatial location information of the surrounding environment are identified and extracted to construct a 3D global environment model. Based on the aforementioned three-dimensional global environment model, the effective working space range is locked in response to the operator's manual selection command in the three-dimensional visualization model or by using a semantic recognition algorithm.

[0008] As a further technical solution, a discretized set of candidate operation point pose variables is generated within the workspace, including: Perform spatial discrete sampling at equal intervals along the target cable route to form a uniformly distributed set of basic candidate points; extract the structural feature points of the line and add the feature points to the candidate point set; The set of candidate bucket parking pose variables generated based on the kinematic model of the bucket truck includes: The global reachability space of the boom is solved by using a forward kinematics model. Based on the global reachability space of the boom, the constraints of the operation coverage range and the macroscopic obstacles of the boom are combined for coarse trimming to construct the initial task feasibility space. Discrete sampling is performed within the initial feasible space of the task to generate an original set of candidate stopping poses. Each pose in the set is verified to obtain a set of candidate stopping pose variables.

[0009] As a further technical solution, a continuous non-uniform energized safety constraint potential field model is constructed, taking the energized conductor as a risk source. Based on the voltage level and spatial distance, the comprehensive risk value at any point within the work space is calculated, including: Collect parameters of all live conductors in the work area, model each conductor as a three-dimensional spatial line segment, and record the coordinates of each conductor's spatial endpoints and the corresponding nominal voltage level; For any three-dimensional coordinate point within the work space, the shortest spatial distance from that point to all live conductors is calculated comprehensively, and a multi-source superposition type continuous function for live safety risks is constructed:

[0010] in, For spatial points The comprehensive electrical safety risk value at the location; the higher the value, the higher the electric field strength and the stronger the risk of electric arc and electric shock. Let be any point in space; For the first The nominal voltage rating of the live conductor; This is the shortest Euclidean distance from a spatial point to the corresponding traverse segment; is the electric field attenuation coefficient.

[0011] As a further technical solution, the operation point pose variable and the stop pose variable are used as coupling variables to construct a two-dimensional combined joint space. The comprehensive risk value is then used to perform multi-dimensional constraint verification on each combination to obtain the joint feasible region, including: The operation point pose variables and the stop pose variables in the candidate pose variable set are combined in pairs to construct a two-dimensional combined joint space; Using the joint space as the screening object, a full-dimensional constraint verification is performed on each coupling combination. The full-dimensional constraint verification includes coupling electric safety quantification constraints based on comprehensive risk value, inverse kinematics reachability constraints of robotic arm end-effector operation, collision constraints of robotic arm end-effector operation space, joint comfort constraints, robotic arm singularity constraints, and visual occlusion visibility constraints. The compliant operating conditions that satisfy all constraints are retained to form the joint feasible domain.

[0012] As a further technical solution, within the joint feasible domain, a multi-dimensional normalized operational performance evaluation function is constructed to quantitatively evaluate the operational quality of each feasible combination, including: Within the joint feasible domain, a multi-dimensional normalized operational performance evaluation function is established, which quantifies and scores five dimensions: robotic arm posture, motion stability, electrical safety, visual observation, and joint load. The multi-dimensional normalized operational performance evaluation function is as follows:

[0013] in, This is a function for evaluating operational performance. , , , , These are the weighting coefficients for each performance indicator; For joint comfort index; As an indicator of motion stability; For electric field safety performance indicators; For visual observation quality indicators; This refers to the joint torque index.

[0014] As a further technical solution, using the joint feasible region as a constraint and aiming to minimize comprehensive risk and maximize operational performance, a multi-objective collaborative optimization model is constructed as follows:

[0015] in, , The target weight coefficient; This is a function for evaluating operational performance. For combined comprehensive risk assessment functions; It is a multi-objective collaborative optimization function.

[0016] The second aspect of the present invention provides a system for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network.

[0017] The system for optimizing the stopping position and operating point posture of a live-line working robot in power distribution networks includes: The environment modeling module is configured to: construct a three-dimensional global environment model of the work area, and lock the effective work space range based on the model; The candidate pose generation module is configured to: generate a set of discretized candidate operation point pose variables within the workspace, and generate a set of candidate bucket stop pose variables based on the kinematic model of the bucket boom vehicle; The safety risk field construction module is configured to: construct a continuous non-uniform energized safety constraint potential field model, take the energized conductor as the risk source, and calculate the comprehensive risk value of any point in the work space based on the voltage level and spatial distance; The joint feasible region construction module is configured to: use the operation point pose variable and the stop pose variable as coupling variables to construct a two-dimensional combined joint space, and combine the comprehensive risk value to perform multi-dimensional constraint verification on each combination to obtain the joint feasible region; The operational performance evaluation module is configured to: construct a multi-dimensional normalized operational performance evaluation function within the joint feasible domain to quantitatively evaluate the operational quality of each feasible combination; The collaborative optimization decision module is configured to: construct a multi-objective collaborative optimization model with the joint feasible region as a constraint and the objectives of minimizing comprehensive risk and maximizing operational performance; and obtain the optimal operational point pose and the optimal stopping pose by solving the optimization model in the coupled variable space.

[0018] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the method for optimizing the stopping position and operating point pose of a live-line working robot in a power distribution network as described in the first aspect of the present invention.

[0019] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the method for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network as described in the first aspect of the present invention.

[0020] The above one or more technical solutions have the following beneficial effects: (1) This invention uses the operation point pose and the bucket stop pose as coupled variables for unified modeling, breaking through the limitations of traditional unidirectional planning and mutual isolation, and constructing a joint decision space and collaborative optimization mechanism to fundamentally avoid problems such as the operation point being feasible but the bucket stop being unable to operate, and local optima. The method does not rely on human experience and discrete screening, but obtains the global optimal combination through continuous optimization, which greatly reduces human intervention and planning errors, and makes the bucket truck parking position highly matched with the robot arm operation point, significantly improving the scientific nature, completeness and execution success rate of the overall operation planning, and providing a reliable premise for the stable and efficient operation of the robot.

[0021] (2) This invention abandons the traditional fixed safety distance threshold mode and establishes a continuous non-uniform live-line safety risk potential field model based on voltage level, spatial distance, and the superposition of electric fields of multiple conductors. This model can accurately quantify and characterize the risks in the entire spatial domain. This model can adaptively adapt to complex scenarios such as different voltage levels, multiple circuits, and dense towers, dynamically distinguish between dangerous areas and safe areas, embed continuous safety constraints into the optimization process, and effectively avoid safety hazards such as electric arcs, electric shocks, and interference. Compared with traditional methods, safety management is more refined, risk warning is more timely, and the inherent safety level of live-line work in distribution networks is greatly improved.

[0022] (3) This invention constructs a comprehensive evaluation model from five dimensions: joint comfort, motion stability, electric field safety, visual observation quality, and joint torque, and quantifies and ranks feasible solutions. Through constraint screening and optimization decision-making, it actively avoids adverse working conditions such as extreme postures, unusual postures, and high torque loads of the robotic arm, ensuring that the robotic arm is always in a state of reasonable posture, smooth motion, and balanced force. It can effectively reduce problems such as structural stress concentration, control jitter, and decreased accuracy, extend the service life of the equipment, improve the stability and operation quality of the operation process, and ensure the reliable completion of complex live tasks.

[0023] (4) This invention integrates real-time perception and online replanning mechanisms, which can monitor abnormal working conditions such as bucket posture deviation, environmental changes, new obstacles, and visual occlusion throughout the process. When a threshold condition is triggered, the entire process iterative optimization is automatically started, and the optimal operation point and bucket posture combination are quickly updated to form a closed-loop correction. This mechanism can effectively cope with time-varying interference such as wind disturbance, temporary obstacles, and equipment position deviation in the field, solve the defects of traditional planning that is generated at one time and cannot adapt to dynamic changes, significantly improve the robustness of the system and the adaptability to on-site engineering, and ensure the continuous, safe, and efficient progress of operations.

[0024] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0025] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0026] Figure 1 This is a flowchart of the method in the first embodiment.

[0027] Figure 2 This is a system structure diagram of the second embodiment. Detailed Implementation

[0028] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0029] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0030] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0031] Example 1 This embodiment discloses a method for optimizing the stopping and operation point pose of a live-line working robot in a power distribution network. By environmental perception modeling, operation area constraints, candidate point and stopping pose generation, a live-line safety constraint field and joint feasible domain are constructed. Combined with multi-dimensional performance evaluation, collaborative optimization decision-making is achieved, and dynamic replanning is supported to improve the safety, stability and efficiency of the operation.

[0032] like Figure 1 As shown, the method for optimizing the stopping bucket and operating point pose of a live-line working robot in power distribution networks includes: Step S1, Environment Modeling. This step involves constructing a three-dimensional global environment model of the work area and, based on this model, defining the effective work space.

[0033] First, 3D point cloud data of the work area is collected by fusing visual sensors and LiDAR to complete environmental feature detection and semantic segmentation, accurately identifying and extracting the outline and spatial location information of cable structures, tower hardware, power equipment, and surrounding natural and artificial obstacles. Second, geometric contour parameters and semantic category labels are fused to construct an integrated, quantifiable, and queryable 3D global environment model, providing a unified environmental data foundation for subsequent distance calculation, occlusion judgment, collision detection, live risk solution, and multi-constraint optimization calculation.

[0034] Specifically, the above process can utilize existing technologies in the robotics field, such as multi-sensor fusion mapping and semantic segmentation. Specifically, it can rely on algorithms like LeGO-LOAM and PCL point cloud libraries to complete the registration, denoising, downsampling, and global stitching of 3D point clouds, achieving accurate reconstruction of the 3D geometric contours of the work scene. By combining a multimodal fusion scheme of 2D image semantic segmentation networks such as Mask R-CNN and 3D point cloud segmentation networks like PointNet++, and through joint calibration of cameras and LiDAR, complementary label mapping between pixels and 3D point clouds can be achieved, enabling accurate semantic classification of targets such as cables, poles, hardware, and obstacles.

[0035] Finally, based on the constructed 3D global environment model, the effective working area can be locked by responding to operators' instructions to manually select and delineate the target cable segment and corresponding tower working area in the 3D visualization model; or, based on the fusion of the above-mentioned 2D image semantic segmentation network and 3D point cloud segmentation network, key point detection, component-level point cloud segmentation, and geometric / topological reasoning can be added. For example, a perceptual semantic recognition algorithm based on Keypoint R-CNN + Part-A2 + spatial topological constraints can automatically match key structures of cables, insulators, and towers, and adaptively generate restricted working areas.

[0036] In this step, the initial pose of the robot during mapping is used as the spatial reference origin. The environmental model is clipped using a custom 3D boundary. The clipping range is adjustable, and only the local spatial model required for the task is retained to avoid invalid global calculations.

[0037] Step S2, candidate pose generation. This involves generating a discretized set of candidate operational point pose variables within the workspace, and generating a candidate bucket-stopping pose variable set based on the bucket boom vehicle's kinematic model.

[0038] Step S2.1, in this embodiment, a set of candidate operation points that combines spatial coverage integrity and task-specific structure is constructed by combining uniform sampling across the entire domain with encrypted sampling of key features. Specifically: Spacing is discretely sampled at equal intervals along the target cable route to form a uniformly distributed set of basic candidate points, ensuring full coverage of the work area. For critical weak points and locations where work is urgently needed, structural feature points such as conductor connection points, clamps, insulator ends, and hardware connection points are extracted and added to the candidate point set.

[0039] Through the above steps, within the limited operating area, a set of discretized candidate operation point pose variables is generated by fusing structural sampling and key feature enhancement strategies. for:

[0040] in, The number of operation points contained in the set of operation point pose variables.

[0041] Step S2.2: For the candidate bucket position variable set, based on the overall structural parameters of the boom truck, joint motion limits, macroscopic environmental obstacles, and operational coverage constraints, a tiered screening is performed within the reach of the boom truck to generate a compliant bucket position variable set. :

[0042] in, This represents the number of operation points contained in the set of stop bucket pose variables.

[0043] Specifically, a single stop pose t contains three-dimensional spatial coordinates and six-degree-of-freedom attitude information, and its generation and filtering process is as follows: Step S2.2.1: Retrieve the factory calibration technical parameters of the bucket truck to obtain the limit ranges of each motion joint, including the rotation angle range of the slewing joint, the pitch angle range of the boom, the extension and retraction stroke of the forearm, and the leveling angle range of the bucket body. Use the DH parameter method to establish a forward kinematic model of the bucket truck, construct the mapping relationship between the motion variables of each joint and the end-effector pose of the bucket body, and solve for the global reachable geometric space of the bucket body, denoted as the global reachable space of the bucket. .

[0044] Step S2.2.2: Based on the global reachability space of the boom, multiple pre-constraints are integrated to complete a large-scale coarse-cutting and construct the initial task feasibility space. Among these, multiple pre-existing constraints include operational coverage constraints and macroscopic obstacle constraints on the boom.

[0045] The operation coverage constraint is based on the candidate points of the operation point pose, and a spherical coverage envelope is constructed in combination with the maximum extension radius of the end-effector. Only the intersection area between the reachable space of the boom and the operation envelope is retained, and the invalid space at the far end that the boom cannot cover is eliminated in advance.

[0046] Macroscopic obstacle constraints on the boom are based on the geometric contours of poles, power equipment, and large fixed obstacles in the 3D environment model. Large-size bounding box coarse-grained collision detection is used to eliminate restricted areas where the main boom frame, telescopic boom, slewing mechanism, and working bucket body are prone to compression and interference, thus avoiding macroscopic collision risks of the large structure of the boom truck from a spatial perspective.

[0047] Step S2.2.3, in the initial feasible space after trimming Internally, an adaptive step-size three-dimensional discrete sampling method is adopted. The sampling step size is dynamically adjusted according to the complexity of the operation scenario and the planning accuracy requirements to generate a batch of candidate bucket stopping poses. :

[0048] The individual bucket stop pose t includes three-dimensional spatial coordinates XYZ and a six-degree-of-freedom pose RPY. To avoid generating a large number of invalid poses that deviate from the working area, physical structural constraints are applied to the bucket stop pose, retaining only the valid working orientation, thus compressing the number of candidate poses from the source. The specific implementation method is as follows: The bucket of the boom truck lacks active pitch and roll leveling functions; only the overall slewing angle is adjustable. Therefore, the bucket's attitude is such that only the yaw angle changes with the direction of cable operation, while the pitch and roll angles remain constant at the horizontal reference, eliminating all large tilt angles and abnormal parking postures.

[0049] First, the spatial axis of the work object in the work area, such as a cable or crossarm, is fitted to extract the target orientation vector. Then, the central axis of the work object in the work area is fitted using the least squares method to solve for the overall spatial orientation unit vector of the work object. This serves as the reference orientation for the yaw attitude of the fighting vehicle.

[0050] Secondly, based on the trajectory vector of the work object, calculate the reference yaw angle of the bucket body facing the work direction. Angle discrete sampling is performed only within a narrow range of ±15° to the left and right of the reference yaw angle to ensure that all stopped bucket attitudes face the work area.

[0051] Finally, considering the actual mechanical structure of the boom truck, the working bucket platform is a rigid structure and does not support active pitch and roll adjustments. Therefore, the pitch angle and roll angle are fixed at the horizontal reference of 0°, and no angle traversal sampling is performed.

[0052] Ultimately, only the effective stopping positions set with variable XYZ position, slightly adjustable yaw, and fixed horizontal pitch and roll are retained. .

[0053] Step S2.2.4: Process the original set of candidate stopping poses. Each pose in Each pose is verified and invalid poses that do not meet the basic operating conditions are eliminated. In this step, invalid poses are eliminated by using the reachability constraints of the boom truck body and global visual visibility constraints.

[0054] Among them, the reachability constraint of the boom truck body motion is verified by combining the multi-joint linkage interference boundary conditions of the boom truck to check whether the rotation angle and extension of each joint are within the legal motion range, and to eliminate unreachable poses with mechanical limit conflicts and self-interference. The global visual visibility constraint is calculated by solving the global pose of the camera and the optical axis orientation from the pose of the bucket body based on the inherent field of view of the visual sensor and the fixed installation extrinsic parameters. Through field of view range judgment and rapid detection of line of sight occlusion, it is ensured that the visual sensor can observe at least one candidate operation point without obstruction under the current pose of the bucket body, thus ensuring the basic conditions for operation observation.

[0055] Step S2.2.5: After multi-layer spatial clipping, sampling, and secondary constraint screening, all valid poses that satisfy mechanical mobility, macroscopic collision-free operation, and visual observability are retained, forming the final candidate stopping pose set T, which provides finite discrete search samples for subsequent coupled joint optimization.

[0056] in, This represents the number of operation points contained in the set of stop bucket pose variables.

[0057] Step S3, Safety Risk Field Construction. This step is used to construct a continuous, non-uniform energized safety constraint potential field model. The energized conductor is used as the risk source, and the comprehensive risk value at any point within the work space is calculated based on the voltage level and spatial distance.

[0058] Existing traditional live-line work planning generally adopts isolation methods with fixed safety distance thresholds, which determine risks based solely on the distance of a single conductor. This approach cannot adapt to complex scenarios such as substations, multiple parallel circuits, and multiple conductors crossing each other. It also cannot quantify the risks of continuous electric field superposition in space, resulting in rigid constraints, poor adaptability, and unreasonable safety redundancy.

[0059] To overcome the shortcomings of existing technologies, this embodiment constructs a continuous non-uniform charged safety constraint potential field model, abstracting each high-voltage charged conductor as a linear electric field risk source. Through the superposition calculation of risks from multiple conductors, a continuous quantitative characterization of charged risks across the entire spatial domain is achieved. The specific modeling method is as follows: Collect parameters of all live conductors within the work area and model each conductor as a three-dimensional line segment. Record the coordinates of each conductor's spatial endpoints and their corresponding nominal voltage levels. Such as 10kV, 110kV, 220kV, etc. For any three-dimensional coordinate point within the workspace. By comprehensively calculating the shortest spatial distance from this point to all live conductors, a multi-source superposition type continuous function for live safety risk is constructed:

[0060] in, For spatial points The comprehensive electrical safety risk value at the location; the higher the value, the higher the electric field strength and the stronger the risk of electric arc and electric shock. For any point in space, it is represented by coordinates (x, y, z) in a three-dimensional rectangular coordinate system; For the first The nominal voltage rating of a strip conductor; This is the shortest Euclidean distance from a spatial point to the corresponding traverse segment; Let be the electric field attenuation coefficient, and The near-field, near-distance range is 1.8-2.0, the far-field, far-distance range is 1.2-1.5, and the normal operating condition is 1.5. It can be calibrated and corrected by combining the measured data of the power frequency electric field on site to simulate the natural attenuation law of the high voltage electric field with spatial distance.

[0061] The risk function constructed in this step is a globally continuous and computable model, breaking through the traditional one-size-fits-all constraint mode with fixed thresholds. It can accurately distinguish the electrical hazard levels of different areas. It can be used for visualizing workspace risks and dynamically verifying safety distances. Furthermore, it can embed continuous safety constraints with voltage coupling and spatial location coupling into subsequent joint screening and optimization objectives, achieving refined and quantitative control of electrical safety constraints.

[0062] Step S4, Joint Feasibility Region Construction. This step uses the operational point pose variable and the stationary pose variable as coupling variables to construct a two-dimensional combined joint space. The combined risk value is then used to perform multi-dimensional constraint verification on each combination to obtain the joint feasibility region.

[0063] Traditional planning methods typically solve the operation point and the bucket's stationary posture separately. The two variables lack coupling relationship, which can easily lead to local optimum problems where one option is feasible but the combination of working conditions is not feasible. At the same time, they lack comprehensive constraints on the kinematics, posture stability, and fine-grained interference of the robotic arm.

[0064] In this embodiment, the operation point variable p and the bucket pose variable t are used as integrated coupling variables. A spatial filtering domain is constructed with a single operation point as the center and the maximum working radius of the vehicle-mounted robotic arm as the radius. Only the bucket pose falling within the sphere domain is coupled and paired with the current operation point to construct a two-dimensional combined joint space. The specific implementation method is as follows: Based on the candidate operation point pose variable set P and the candidate stop pose variable set T, construct a two-dimensional combined joint space for all pairwise combinations. :

[0065] With joint space To screen the objects, a full-dimensional constraint verification is carried out for each coupling combination. If any constraint is not met, the combination is directly removed, and only the working conditions that meet all constraints are retained.

[0066] The full-dimensional constraint verification includes coupled electrified safety quantification constraints based on comprehensive risk values, inverse kinematics reachability constraints for robotic arm end-effector operations, collision constraints for robotic arm end-effector operation space, joint comfort constraints, robotic arm singularity constraints, and visual occlusion visibility constraints.

[0067] (1) Coupled Quantitative Constraints for Electricity Safety Based on the multi-source superposition type live safety risk continuity function constructed in step S3, the risk value at the operating point is calculated respectively. Risk value of the center position of the buoy Construct a comprehensive risk assessment formula for the portfolio:

[0068] in, This is a weighting coefficient. Based on the actual operational characteristics of locations near high-voltage equipment and where risk weights are higher, a uniform safety risk threshold is set across the entire area. Only retain those that meet the requirements. A low-risk portfolio.

[0069] (2) Reverse kinematics reachability constraints of robotic arm end-effector The inverse kinematics reachability constraints of the robotic arm's end effector are used to determine whether the combination of candidate bucket stop pose and candidate operation point pose satisfies the kinematic reachability conditions of the robotic arm. Since the position and attitude of the robotic arm base are different under different bucket stop poses for the same operation point, the inverse kinematics solution results are also different. Therefore, in this embodiment, the bucket stop pose and operation point pose are used as coupled variables for unified analysis, and reachability constraint verification is carried out in the joint space.

[0070] For any combination of stopping pose and operational point pose in the joint space, firstly, a bucket body coordinate system is established with the bucket body platform as the local reference coordinate system. Then, the pose transformation relationship between the bucket body coordinate system and the global coordinate system is determined based on the candidate stopping pose. Let the homogeneous transformation matrix corresponding to the candidate stopping pose be... for:

[0071] in, This represents the rotation matrix of the fighting body's posture; This represents the translation vector of the bucket's position.

[0072] For the target operation point in the global coordinate system, the coordinate transformation relationship is used to transform it to the local coordinate system of the bucket body, so as to obtain the position representation of the target point in the coordinate system of the robot arm base, thereby realizing the spatial coupling mapping between the candidate bucket stopping pose and the candidate operation point pose, as shown in the following formula:

[0073] in, This represents the position of the target point in the coordinate system of the robot arm base. The target operation point in the global coordinate system.

[0074] After obtaining the local coordinates of the target point, a kinematic model is established based on a six-degree-of-freedom insulated robotic arm. The DH parameter method is used to describe the connection relationships between the joints of the robotic arm, and a forward kinematic model of the robotic arm is constructed.

[0075] in, This is the joint angle vector of the robotic arm. The pose of the robotic arm's end effector. Based on the target workpoint pose. Solve the inverse kinematic equations:

[0076] The set of all inverse solutions that satisfy the target pose requirements is obtained.

[0077] Then, the joint constraint verification was performed on each set of inverse solutions obtained. Let the first set be... The permissible range of motion for each joint is:

[0078] If there exists at least one inverse kinematics solution that satisfies all joint angles within the corresponding allowable range, then the combination of the stop pose and the operation point pose is determined to satisfy the inverse kinematics reachability condition; if there is no valid solution in the inverse kinematics solution, or if all inverse kinematics solutions exceed the joint limit range, then the combination is determined to be unreachable and is directly removed from the joint space.

[0079] To avoid the unfavorable working condition where the robotic arm, despite having an inverse solution, remains in a position close to the joint limits for an extended period, a joint margin evaluation index is introduced.

[0080] in, This represents the average margin between the current posture and the limit boundaries of each joint. When the joint margin is lower than a preset threshold, the posture is considered to have a significant risk of motion restriction, and even if the inverse kinematics reachability condition is met, it is judged as an invalid solution and discarded. Finally, only the combination of the stopping pose and the operating point pose that simultaneously satisfies the inverse kinematics reachability constraint and the joint margin constraint is retained for subsequent collision detection, singularity analysis, and comprehensive optimization stages to ensure that the robotic arm can complete the power grid live-line operation task safely, stably, and with sufficient motion margin.

[0081] (3) Collision constraints of the robotic arm end-effector's working space For joint postures with valid inverse kinematics, small-scale, high-precision collision interference detection is conducted. AABB bounding boxes, OBB oriented bounding boxes, mesh model collision detection, or point cloud distance detection methods are used to perform component-by-component collision assessments on each link of the robotic arm, insulated operating tools, and end effector. This verifies whether the robotic arm's motion trajectory and static posture exhibit spatial overlap and interference with wires, hardware, equipment, or local obstacles, eliminating corresponding combinations with interference risks.

[0082] Collision constraints in the end-effector's working space are used to determine whether there are spatial collision or interference risks during the robot's operation, provided that the candidate bucket position and candidate operating point position combinations meet the inverse kinematics reachability conditions. Since the robot needs to work closely with conductors, fittings, insulators, crossarms, and other electrical equipment in live-line working environments, simply meeting end-effector reachability requirements does not guarantee operational safety. Therefore, collision constraints in the robot's end-effector's working space are further introduced into the combined space to perform refined collision detection on the robot's static posture and motion process, eliminating bucket position and operating point position combinations that pose a risk of spatial interference.

[0083] For any combination filtered by inverse kinematics reachability constraints, based on its corresponding joint angle vector By using the forward kinematics model of the robotic arm, the spatial position and orientation of each joint and link in the global coordinate system are calculated, thus obtaining the spatial model of each key component of the robotic arm. :

[0084] in, Indicates the first The robotic arm consists of components including robotic arm links, insulated operating rods, end effectors, and auxiliary tools.

[0085] Simultaneously, the set of surrounding obstacles is extracted from the 3D environment model generated in the environmental modeling steps. ,in, A spatial model representing conductors, insulators, fittings, power equipment, tower components, and other obstacles.

[0086] To improve collision detection efficiency, axis-aligned bounding boxes (AABBs) are first used for rapid coarse screening of the robotic arm components and environmental obstacles. For any robotic arm component bounding box... and obstacle enclosure Calculate the projection intervals of the two bounding boxes along the three coordinate axes. If the following conditions are met, it is determined that the two bounding boxes may overlap spatially, and the process proceeds to the fine collision detection stage; otherwise, it is considered that there is no risk of collision:

[0087]

[0088]

[0089] For component pairs that pass the initial screening, high-precision collision analysis is further performed using oriented bounding box (OBB), triangular mesh models, or point cloud distance detection methods. Let the point set on the surface of the robotic arm components be... The point set on the obstacle surface is Calculate the minimum Euclidean distance between the two. :

[0090] Among them, the surface points of the robotic arm components Obstacle surface points .

[0091] Set minimum safety interval thresholds according to live-line working safety requirements. When satisfied When the current robotic arm posture is deemed to pose a risk of collision or insufficient safety distance; when the condition is met... At that time, it is determined that the current attitude meets the space safety requirements.

[0092] To avoid the risk of dynamic collisions during the robotic arm's movement, not only is the target working posture detected, but the discrete trajectory of the robotic arm moving from the initial posture to the target posture is also continuously sampled. Let the trajectory parameters be... The trajectory of the robotic arm is represented as follows: Multiple intermediate poses are obtained by uniformly sampling along the trajectory: The collision detection process described above is repeated for each sampled pose. If any sampled point exists... If the robotic arm is at risk of collision or interference during its movement, it is determined that the robotic arm is at risk of collision or interference. If all sampled postures meet the safety distance requirements, the combination is considered to meet the dynamic space collision constraints.

[0093] (4) Joint comfort constraints Based on the mechanical motion range of each joint, calculate the relative position ratio of the current inverse joint angle within the range. The calculation formula is as follows:

[0094] in, 0 indicates that the joint is at its lower limit, and 1 indicates that it is at its upper limit.

[0095] A preset threshold for the proportion of the robot approaching its limit is used, with exemplary values ​​of 0.1 and 0.9. That is, if the joint angle is within 10% of the extreme range of motion, it is considered to be severely off-center and close to the limit. To avoid structural stress concentration, control jitter, and lifespan reduction caused by the robot arm being in a state of extreme inflection point for a long time, this type of combination is eliminated.

[0096] (5) Singularity constraints of robotic arms To prevent the robotic arm from entering singular poses during operation, thus avoiding loss of control, jitter, and control failure, and to improve the control stability of the robotic arm, singularity constraints need to be introduced. Singularity constraints are used to prevent the robotic arm from entering singular pose regions during operation, which can lead to problems such as degraded end-effector mobility, sharp increases in joint velocity, decreased trajectory tracking accuracy, and control instability. During the joint feasible region construction process, for the combination of stopping pose and operating point pose that already satisfies inverse kinematic reachability and spatial collision constraints, singularity detection is further performed to ensure that the robotic arm always operates within a region with good kinematic performance and stable control.

[0097] The speed of the robotic arm's end effector and the speed of its joints satisfy a Jacobian mapping relationship:

[0098] in, Let be the linear and angular velocity vectors of the end effector in the task space. For each joint angular velocity vector, Let be the Jacobian matrix of the robotic arm in its current joint state. The Jacobian matrix reflects the motion transmission relationship between joint space and Cartesian space, and its numerical characteristics can directly characterize the robotic arm's motion capability in its current posture. When the robotic arm is in its normal working range, the motion capability in each direction is relatively balanced; however, when the robotic arm approaches a singular pose, some degrees of freedom gradually degenerate, causing the small motion requirements of the end effector to correspond to extremely large joint movements, thereby leading to robotic arm tremors, decreased control accuracy, or even loss of control.

[0099] To quantify the degree to which the current pose of the robotic arm is close to a singular pose, this invention uses the condition number of the Jacobian matrix as a singularity evaluation index. First, singular value decomposition is performed on the Jacobian matrix to obtain its maximum singular value. and minimum singular value And calculate the condition number:

[0100] Among them, condition number This reflects the balance of the robotic arm's motion capabilities in all directions. When the robotic arm is moving away from the singular pose, the difference between the maximum and minimum singular values ​​is small, and the condition number remains at a low level; as the robotic arm gradually approaches the singular region, the minimum singular value continuously decreases and approaches zero, i.e. At this point, the condition number rapidly increases and tends towards infinity, indicating a significant deterioration in the kinematic performance of the robotic arm. Therefore, in this embodiment, a singularity threshold is preset. The following judgment criteria are adopted: when When the condition number of the Jacobian matrix is ​​less than the preset threshold under the current posture of the robotic arm, it means that the motion capability of the robotic arm in all directions is evenly distributed, there is a good mapping relationship between the joint motion and the end effector motion, and the motion of the end effector will not cause abnormally amplified joint speed. Therefore, the robotic arm is determined to be in a non-singular state, and the corresponding combination of the stop pose and the operation point pose satisfies the singularity constraint.

[0101] Conversely, when If the robotic arm approaches or enters a singular region, it indicates that there is a risk of degraded mobility, amplified joint speed, and decreased control stability. In this case, the combination is determined not to meet the singularity constraint and is eliminated.

[0102] (6) Visual occlusion visibility constraints By relying on the fixed installation of sensors to participate in the current position and posture of the bucket, the position of the camera's optical center and the direction of the optical axis are solved. The field of view boundary is used to determine whether the target point is within the effective shooting range. A spatial detection ray is constructed with the camera's optical center as the starting point and the operation point as the ending point to detect whether there are obstacles blocking the ray path. Only the coupled combination with compliant viewing angle and unobstructed line of sight is retained to ensure visual closed-loop control throughout the operation.

[0103] After the joint screening of all the above constraints, all coupled combinations that simultaneously satisfy all constraints constitute a compact and feasible joint feasible region. By combining feasible domains Eliminating all dangerous, unreachable, unstable, and unobservable invalid combinations provides a convergent, finite solution space for subsequent multi-dimensional performance evaluation and collaborative optimization.

[0104] Step S5, Operational Performance Evaluation. This step involves constructing a multi-dimensional normalized operational performance evaluation function within the joint feasible domain to quantitatively assess the operational quality of each feasible combination.

[0105] Traditional screening methods only complete a binary judgment of feasibility, and cannot distinguish between the superior and inferior among multiple feasible solutions. They are prone to selecting suboptimal solutions with poor attitude, large torque, weak stability, and poor observation conditions.

[0106] Within the joint feasible domain, this invention establishes a multi-dimensional normalized operational performance evaluation model, which quantifies and scores five dimensions: robotic arm posture, motion stability, electrical safety, visual observation, and joint load. A weighted comprehensive evaluation function is then constructed to achieve a refined ranking of feasible solutions.

[0107] The expression for the comprehensive operational performance evaluation function is as follows:

[0108] in, This is a function for evaluating operational performance. , , , , These are the weighting coefficients for each performance indicator, which can be adaptively adjusted according to the operational requirements of prioritizing safety, stability, and efficiency. For joint comfort index; As an indicator of motion stability; For electric field safety performance indicators; For visual observation quality indicators; This refers to the joint torque index.

[0109] The specific definitions of each indicator in the above formula are as follows: (1) Joint comfort index Joint comfort index is used to evaluate the rationality of the posture of the end-effector in the corresponding joint state, and to avoid problems such as excessive structural stress, reduced control accuracy and shortened service life caused by the end-effector being in the extreme joint posture. The closer to the middle posture of the joint, the higher the comfort and the stronger the operation stability.

[0110] The specific calculation process is as follows: Based on the joint variable q obtained from the inverse kinematics solution in step S4, the physical limits of each joint are obtained, and the intermediate poses of each joint are calculated. By calculating the current joint state q and the intermediate pose... The deviation percentage is used to quantify joint comfort, and the specific expression is as follows:

[0111] In the formula, This refers to the number of joints in the end-effector robotic arm. For the first The current state of each joint. For the first The intermediate posture of each joint. , The first The maximum and minimum range of motion of each joint; The value range is [0,1]. The closer the value is to 1, the closer the joint posture is to the middle area, the higher the comfort and the better the stability.

[0112] (2) Motion stability index Based on the completion of singular pose hard rejection and the fact that all combinations are in the non-singular region, this index further quantifies the degree of redundancy of each combination away from the singular region, characterizing the motion stability and control robustness of the robotic arm.

[0113] The specific calculation process is as follows: Based on the joint state q obtained in step S4, calculate the Jacobian matrix J corresponding to the robotic arm, and solve for the condition number of the Jacobian matrix J. Preset singularity threshold The singularity index is obtained through normalization. The specific expression is as follows:

[0114] in, The value range is [0,1]. The closer the value is to 1, the further the robotic arm is from the singular pose, and the stronger its motion stability. This indicates that the robotic arm is in an unusual position and does not have stable operating conditions.

[0115] (3) Electric field safety performance indicators The electric field safety performance index is used to quantify the safety risks of live-line work in the coupled combination (p, t). Combined with the live-line safety risk continuous function constructed in step S3, it realizes the synergistic evaluation of risk and operation performance. The lower the risk value, the higher the score and the stronger the safety of the operation.

[0116] The specific calculation process is as follows: based on the security risk function in step S3. Calculate the risk value of operation point p respectively. Risk value of stopping position t Combining the comprehensive risk value defined in step S3, the electric field exposure risk index is obtained through normalization. The specific expression is as follows:

[0117] In the formula, For joint feasible domain The maximum combined risk value of all combinations is used for normalization. The value range is [0,1]. The closer the value is to 1, the lower the electrical safety risk of the combination and the better the operational safety.

[0118] (4) Visual observation quality indicators The visual observation quality index is used to evaluate the observation effect of the visual sensor on the operation point p, ensuring that the sensor can clearly and stably acquire the image information of the operation point during the operation, and provide reliable feedback for operation control. The better the observation effect, the higher the score.

[0119] The specific calculation process is as follows: Using the multi-view ratio method, with the optical center of the visual sensor as the observation starting point, the visibility and viewpoint rationality of the operating point p are judged; the proportion of unobstructed views within the effective field of view of the sensor to the total number of observed views is statistically analyzed, serving as a visual visibility index. The specific expression is as follows:

[0120] In the formula, The preset total number of observation angles, The number of observation angles where the operation point p is unobstructed and the viewing angle meets the requirements; The value range is [0,1]. The closer the value is to 1, the better the visual observation effect and the stronger the controllability of the operation.

[0121] (5) Joint torque index The joint torque index is used to evaluate the force stability of the end-effector during operation, avoiding problems such as mechanical vibration, structural damage, and decreased control accuracy caused by excessive joint torque. The smaller the joint torque, the higher the score and the stronger the operational stability.

[0122] The specific calculation process is as follows: Based on the Jacobian matrix J of the robotic arm and the end effector load F, the torque τ of each joint of the robotic arm is solved through the dynamic equation. The end effector load F is a comprehensive load, including tool weight, working contact force (such as the force of the contact wire), wind load, and reaction force of the insulating tool. Its vector expression is: The magnitude of the overall joint torque is calculated using the L2 norm. The joint torque index was obtained through normalization. The specific expression is as follows:

[0123] In the formula, This represents the maximum permissible joint torque of the end-effector robotic arm. The value range is [0,1]. The closer the value is to 1, the smaller the joint torque, the more stable the operation of the robotic arm, and the less structural damage.

[0124] By weighting and fusing the above five dimensions of indicators, each coupling combination is obtained. Operational performance rating A higher score indicates better operational quality, stronger stability, and higher safety for the combination. This evaluation model uses the operational point pose and the stop pose as coupled variables to conduct quantitative evaluation within a unified parameter space. Unlike existing technologies that evaluate single variables independently, this model can more comprehensively and accurately reflect the actual operational quality and provide core evaluation basis for collaborative optimization decisions.

[0125] Step S6, Collaborative Optimization Decision. This step involves constructing a multi-objective collaborative optimization model with the joint feasible region as a constraint and the objectives of minimizing overall risk and maximizing operational performance. The optimal operational point pose and optimal stopping pose are obtained by solving the optimization model within the coupled variable space.

[0126] Unlike the traditional separate optimization mode, this embodiment uses the joint feasible region as a fixed constraint boundary, integrates the two core objectives of live safety risk and operational performance, constructs a multi-objective collaborative optimization model, solves the globally optimal matching scheme in the coupled variable space, avoids local optima, and achieves a comprehensive balance of safe, stable and efficient operation.

[0127] The multi-objective optimization model aims to minimize both risk and maximize operational performance, and is uniformly transformed into a minimization solution form:

[0128] in, , The target weighting coefficients are normalized to balance the priority of live-line safety risks and operational performance, satisfying the following requirements. ,in Safety weights in high-voltage live-line working scenarios Priority is greater than performance weight ; This is a function for evaluating operational performance. For combined comprehensive risk assessment functions; This is a multi-objective collaborative optimization function; the smaller the value, the better the overall combined effect.

[0129] In solving the above multi-objective optimization model, considering the discrete characteristics of the candidate set, the following three solution strategies can be adopted: Enumerated traversal search: Suitable for small-scale candidate sets, it performs a full domain traversal to calculate the objective function, directly selecting the optimal solution. The results are accurate and the logic is simple. Specifically, for each feasible combination within the joint feasible domain, the corresponding comprehensive risk assessment function R(p,t), operational performance assessment function Q(p,t), and optimization objective function F(p,t) are calculated. By comparing the objective function values ​​of all combinations through a full domain traversal, the optimal solution is selected. The optimal solution is obtained by finding the global optimum. This method guarantees that the search results cover all feasible combinations and obtains the global optimum. It is suitable for application scenarios with a small number of candidate combinations and relatively low real-time requirements.

[0130] Heuristic intelligent optimization algorithms employ iterative optimization strategies such as genetic algorithms and particle swarm optimization, making them suitable for large-scale complex scenarios and enabling rapid convergence to near-optimal solutions. Specifically, the stopping pose and the operation point pose are encoded as optimized individuals, constructing an initial population or particle swarm. Intelligent optimization methods such as genetic algorithms and particle swarm optimization iteratively search the joint space. In each iteration, the objective function F(p,t) is used as the fitness evaluation index. Selection, crossover, mutation, or velocity / position updates are performed on candidate individuals to guide the search direction towards a region with a better objective function, ultimately obtaining a near-optimal solution that meets the optimization requirements. This method significantly reduces the computational cost in large-scale search spaces and improves solution efficiency in complex operation scenarios.

[0131] Local Neighborhood Search Optimization: Based on the initial optimal combination, minor neighborhood adjustments are made to further improve the adaptability of the solution. Specifically, using the current optimal combination of stopping pose and operation point pose as the search center, several minor perturbation combinations are generated within its neighborhood to locally adjust the stopping position, stopping posture, and operation point position, and the corresponding objective function value is recalculated. By continuously comparing the performance difference between the neighborhood combination and the current optimal combination, the combination with the better objective function value is selected as the new search center until the objective function no longer improves significantly or reaches the preset termination condition after several consecutive iterations, thereby obtaining an optimization result that better meets the needs of on-site operations. This method can further improve the adaptability and optimization accuracy of pose combinations while maintaining overall computational efficiency.

[0132] In practical applications, depending on the size of the candidate combinations, computational resource conditions, and real-time requirements, an enumeration method can be used alone, or a combination of heuristic intelligent optimization algorithms and local neighborhood search optimization can be employed. Finally, after completing the optimization solution, a complete, directly implementable job execution plan is output, including: the optimal operation point pose. Optimal fighting vehicle positioning The comprehensive live-line safety risk value corresponding to the optimal combination The overall operational performance score corresponding to the optimal combination. .

[0133] Furthermore, the method provided in this embodiment also includes the steps of dynamic updating and replanning mechanisms.

[0134] Throughout the entire operation, the system collects GNSS positioning data of the boom truck, environmental point cloud perception data, robotic arm joint status, and visual occlusion changes in real time. It continuously monitors multiple triggering conditions, including external dynamic disturbances, preset pose offset thresholds, environmental change thresholds, electrical risk exceeding limits, and visual occlusion change thresholds.

[0135] When abnormal conditions such as temporary changes in the working environment, the addition of obstacles, or excessive bucket body posture deviation are detected, the entire process of replanning is automatically triggered. This involves iterative execution of environment modeling, candidate posture generation, joint constraint screening, performance evaluation, and collaborative optimization, dynamically updating the optimal matching combination. Through online closed-loop correction, it adapts to complex and time-varying field environments, significantly improving the environmental robustness and on-site engineering adaptability of the entire planning method.

[0136] Example 2 This embodiment discloses a system for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network; like Figure 2 As shown, the bucket stopping and operating point pose optimization system for a live-line working robot includes: The environment modeling module is configured to: construct a three-dimensional global environment model of the work area, and lock the effective work space range based on the model; The candidate pose generation module is configured to: generate a set of discretized candidate operation point pose variables within the workspace, and generate a set of candidate bucket stop pose variables based on the kinematic model of the bucket boom vehicle; The safety risk field construction module is configured to: construct a continuous non-uniform energized safety constraint potential field model, take the energized conductor as the risk source, and calculate the comprehensive risk value of any point in the work space based on the voltage level and spatial distance; The joint feasible region construction module is configured to: use the operation point pose variable and the stop pose variable as coupling variables to construct a two-dimensional combined joint space, and combine the comprehensive risk value to perform multi-dimensional constraint verification on each combination to obtain the joint feasible region; The operational performance evaluation module is configured to: construct a multi-dimensional normalized operational performance evaluation function within the joint feasible domain to quantitatively evaluate the operational quality of each feasible combination; The collaborative optimization decision module is configured to: construct a multi-objective collaborative optimization model with the joint feasible region as a constraint and the objectives of minimizing comprehensive risk and maximizing operational performance; and obtain the optimal operational point pose and the optimal stopping pose by solving the optimization model in the coupled variable space.

[0137] Example 3 The purpose of this embodiment is to provide a computer-readable storage medium.

[0138] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the method for optimizing the stopping position and operating point pose of a live-line working robot in a power distribution network as described in Example 1.

[0139] Example 4 The purpose of this embodiment is to provide an electronic device.

[0140] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the method for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network as described in Embodiment 1.

[0141] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0142] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0143] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for optimizing the stopping bucket and operating point pose of a live-line working robot in power distribution networks, characterized in that, include: The environmental modeling step is used to construct a three-dimensional global environmental model of the work area and to lock the effective work space range based on the model; The candidate pose generation step is used to generate a discretized set of candidate operation point pose variables within the work space, and to generate a set of candidate bucket stop pose variables based on the kinematic model of the bucket truck. The steps for constructing a safety risk field are used to build a continuous non-uniform energized safety constraint potential field model. The energized conductor is used as the risk source, and the comprehensive risk value of any point in the work space is calculated based on the voltage level and spatial distance. The joint feasible region construction step is used to construct a two-dimensional combined joint space by taking the operation point pose variable and the stop pose variable as coupling variables, and to perform multi-dimensional constraint verification on each combination in combination with the comprehensive risk value to obtain the joint feasible region. The operational performance evaluation step is used to construct a multi-dimensional normalized operational performance evaluation function within the joint feasible domain to quantitatively evaluate the operational quality of each feasible combination. The collaborative optimization decision-making steps are used to construct a multi-objective collaborative optimization model with the joint feasible domain as a constraint and the objectives of minimizing comprehensive risk and maximizing operational performance. The optimal operational point pose and the optimal stopping pose are obtained by solving the optimization model in the coupled variable space.

2. The method for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network as described in claim 1, characterized in that, Construct a three-dimensional global environment model of the work area, and based on the model, determine the effective work space range, including: By collecting 3D point cloud data of the work area using visual sensors and lidar, the contours and spatial location information of the surrounding environment are identified and extracted to construct a 3D global environment model. Based on the aforementioned three-dimensional global environment model, the effective working space range is locked in response to the operator's manual selection command in the three-dimensional visualization model or by using a semantic recognition algorithm.

3. The method for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network as described in claim 1, characterized in that, Within the workspace, a discretized set of candidate operation point pose variables is generated, including: Perform spatial discrete sampling at equal intervals along the target cable route to form a uniformly distributed set of basic candidate points; extract the structural feature points of the line and add the feature points to the candidate point set; The set of candidate bucket parking pose variables generated based on the kinematic model of the bucket truck includes: The global reachability space of the boom is solved by using a forward kinematics model. Based on the global reachability space of the boom, the constraints of the operation coverage range and the macroscopic obstacles of the boom are combined for coarse trimming to construct the initial task feasibility space. Discrete sampling is performed within the initial feasible space of the task to generate an original set of candidate stopping poses. Each pose in the set is verified to obtain a set of candidate stopping pose variables.

4. The method for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network as described in claim 1, characterized in that, A continuous, non-uniform energized safety constraint potential field model is constructed, with energized conductors as risk sources. The comprehensive risk value at any point within the work space is calculated based on voltage level and spatial distance, including: Collect parameters of all live conductors in the work area, model each conductor as a three-dimensional spatial line segment, and record the coordinates of each conductor's spatial endpoints and the corresponding nominal voltage level; For any three-dimensional coordinate point within the work space, the shortest spatial distance from that point to all live conductors is calculated comprehensively, and a multi-source superposition type continuous function for live safety risks is constructed: in, For spatial points The comprehensive electrical safety risk value at the location; the higher the value, the higher the electric field strength and the stronger the risk of electric arc and electric shock. Let be any point in space; For the first The nominal voltage rating of a strip conductor; This is the shortest Euclidean distance from a spatial point to the corresponding traverse segment; is the electric field attenuation coefficient.

5. The method for optimizing the stopping bucket and operating point pose of a live-line working robot as described in claim 1, characterized in that, By using the operational point pose variable and the stationary pose variable as coupling variables, a two-dimensional combined joint space is constructed. The comprehensive risk value is then used to perform multi-dimensional constraint verification on each combination, resulting in a joint feasible region, including: The operation point pose variables and the stop pose variables in the candidate pose variable set are combined in pairs to construct a two-dimensional combined joint space; Using the joint space as the screening object, a full-dimensional constraint verification is performed on each coupling combination. The full-dimensional constraint verification includes coupling electric safety quantification constraints based on comprehensive risk value, inverse kinematics reachability constraints of robotic arm end-effector operation, collision constraints of robotic arm end-effector operation space, joint comfort constraints, robotic arm singularity constraints, and visual occlusion visibility constraints. The compliant operating conditions that satisfy all constraints are retained to form the joint feasible domain.

6. The method for optimizing the stopping position and operating point posture of a live-line working robot in a power distribution network as described in claim 1, characterized in that, Within the joint feasible domain, a multi-dimensional normalized operational performance evaluation function is constructed to quantitatively assess the operational quality of each feasible combination, including: Within the joint feasible domain, a multi-dimensional normalized operational performance evaluation function is established, which quantifies and scores five dimensions: robotic arm posture, motion stability, electrical safety, visual observation, and joint load. The multi-dimensional normalized operational performance evaluation function is as follows: in, This is a function for evaluating operational performance. , , , , These are the weighting coefficients for each performance indicator; For joint comfort index; As an indicator of motion stability; For electric field safety performance indicators; For visual observation quality indicators; This refers to the joint torque index.

7. The method for optimizing the stopping bucket and operating point pose of a live-line working robot as described in claim 1, characterized in that, Using the joint feasible region as a constraint, and aiming to minimize overall risk and maximize operational performance, the multi-objective collaborative optimization model is constructed as follows: in, , The target weight coefficient; This is a function for evaluating operational performance. For combined comprehensive risk assessment functions; This is a multi-objective collaborative optimization function.

8. A system for optimizing the stopping bucket and operating point posture of a live-line working robot in power distribution networks, characterized in that, include: The environment modeling module is configured to: construct a three-dimensional global environment model of the work area, and lock the effective work space range based on the model; The candidate pose generation module is configured to: generate a set of discretized candidate operation point pose variables within the workspace, and generate a set of candidate bucket stop pose variables based on the kinematic model of the bucket boom vehicle; The safety risk field construction module is configured to: construct a continuous non-uniform energized safety constraint potential field model, take the energized conductor as the risk source, and calculate the comprehensive risk value of any point in the work space based on the voltage level and spatial distance; The joint feasible region construction module is configured to: use the operation point pose variable and the stop pose variable as coupling variables to construct a two-dimensional combined joint space, and combine the comprehensive risk value to perform multi-dimensional constraint verification on each combination to obtain the joint feasible region; The operational performance evaluation module is configured to: construct a multi-dimensional normalized operational performance evaluation function within the joint feasible domain to quantitatively evaluate the operational quality of each feasible combination; The collaborative optimization decision module is configured to: construct a multi-objective collaborative optimization model with the joint feasible region as a constraint and the objectives of minimizing comprehensive risk and maximizing operational performance; and obtain the optimal operational point pose and the optimal stopping pose by solving the optimization model in the coupled variable space.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the method for optimizing the stopping position and operating point pose of the live-line working robot in power distribution networks as described in any one of claims 1-7.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the method for optimizing the stopping position and operating point posture of the live-line working robot in power distribution networks as described in any one of claims 1-7.

Images