Dynamic Adjustment System of a Multi-Degree-of-Freedom Manipulator

Abstract

The present invention relates to the technical field of robotic arm adjustment, and discloses a dynamic adjustment system for a multi-degree-of-freedom robotic arm, including a collision detection module, a processing module, and a path correction module; wherein: the collision detection module detects early collisions of the robotic arm through stress monitoring; the processing module is configured with a reference path of the robotic arm; when the collision detection module detects the early collision, the processing module divides the reference path into sub-paths and identifies the sub-paths with collision risks; the path correction module generates an obstacle avoidance path based on the motion redundancy of different joints of the robotic arm, and corrects the sub-paths with collision risks based on the obstacle avoidance path; the present application improves the motion performance and environmental adaptability of the robotic arm by identifying early collisions and optimizing the obstacle avoidance path.

Description

Technical Field

[0001] The present invention relates to the technical field of robotic arm adjustment, and particularly to a dynamic adjustment system for a multi-degree-of-freedom robotic arm. Background Art

[0002] When the robotic arm of an existing shield tunneling machine for cutter changing executes an action, it completes the action according to the control of the system. However, many robotic arms are not equipped with a feedback mechanism. In a limited space, if the robotic arm touches an obstacle and is not adjusted, the overall structure of the robotic arm will be damaged. The sensitivity of the existing technology for collision detection of robotic arms is insufficient. Most of the existing collision detection methods rely on external sensors such as lidar and vision cameras. These methods have problems such as large influence of detection accuracy by environmental factors and detection delay, and it is difficult to make an effective response in the early stage of a collision.

[0003] A multi-degree-of-freedom robotic arm is composed of multiple joints, and the movements between the joints need to be highly coordinated to achieve precise actions. However, currently, in terms of joint motion control, there is a lack of an effective coordination mechanism, and it is difficult to dynamically adjust the motion parameters of the joints according to the real-time load and working state of the robotic arm. For some robotic arms with redundant degrees of freedom, the existing technology does not make full use of these redundant degrees of freedom to optimize the motion path, improve the flexibility and obstacle avoidance ability of the robotic arm.

[0004] For example, the Chinese patent with the authorization announcement number CN113232021B discloses a method for detecting collisions in a robotic arm grasping path. First, set the starting point of the robotic arm path, obtain scene data, judge and obtain the position of the object to be grasped as the end point of the robotic arm path, and then calculate the poses of multiple equally divided points of the robotic arm in combination with the interpolation step length to judge whether there is a collision in the path. If there is no collision, send the pose to the controller and grasp the object in combination with the joint angle state of the robotic arm to complete the collision detection. This invention can quickly obtain the effectiveness of the path by controlling the step length to obtain the poses of discrete positions of the robot on the grasping path and performing collision detection between the robotic arm and the environment according to the poses; by using the method of the center point and the axial length to represent the axisymmetric bounding box of the object to be grasped, the calculation amount of the bounding box is reduced, and the calculation and detection efficiency are improved. However, this invention still has the problems raised in the background art of this application: there are problems such as large influence of detection accuracy by environmental factors and detection delay.

[0005] The information disclosed in this background art section is only intended to increase the understanding of the overall background of the present invention, and should not be regarded as an admission or any form of suggestion that this information constitutes prior art already known to those of ordinary skill in the art. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the defects of the prior art and provide a dynamic adjustment system for a multi-degree-of-freedom robotic arm, which improves the motion performance and environmental adaptability of the robotic arm by identifying early collisions and optimizing obstacle avoidance paths.

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] A dynamic adjustment system for a multi-degree-of-freedom robotic arm, including a collision detection module, a processing module, and a path correction module; where:

[0009] The collision detection module detects early collisions of the robotic arm through stress monitoring;

[0010] The processing module is configured with a reference path of the robotic arm; when the collision detection module detects the early collision, the processing module divides the reference path into sub-paths and identifies the sub-paths with collision risks;

[0011] The path correction module generates an obstacle avoidance path based on the motion redundancy of different joints of the robotic arm and corrects the path of the sub-path with collision risks based on the obstacle avoidance path.

[0012] As a preferred solution of the dynamic adjustment system for the multi-degree-of-freedom robotic arm of the present invention, where: the collision detection module includes a stress monitoring unit and a calculation unit; the stress monitoring unit is used to monitor the stress distribution of each stressed component on the robotic arm in real time; the stress monitoring unit includes stress sensors installed at different positions of any stressed component; for any stressed component, during the operation of the robotic arm, the stress values at different positions are continuously collected and recorded through the stress sensors;

[0013] The calculation unit identifies early collisions of the robotic arm based on the stress distribution of each stressed component; specifically including:

[0014] Real-time collect the motion parameters of each joint of the robotic arm;

[0015] Based on the motion parameters of each joint of the robotic arm, obtain the reference stress value corresponding to each position of each stressed component in different poses;

[0016] Based on the real-time stress value and the corresponding reference stress value of each position of each stressed component in different poses, identify early collisions of the robotic arm.

[0017] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm described in the present invention, wherein: the calculation unit is configured with a stress mapping model; obtaining the reference stress value corresponding to each position of each stressed component under different poses specifically includes: inputting the motion parameters of each joint of the robotic arm into the stress mapping model, and the stress mapping model calculates and outputs the reference stress values of different positions of each stressed component; the stress mapping model is any one of a linear regression model, a polynomial regression model, and a support vector regression model.

[0018] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm described in the present invention, wherein: the calculation unit is further configured with a collision recognition strategy; the collision recognition strategy is used to recognize the early collision of the robotic arm, specifically as follows:

[0019] Read the real-time stress value and the corresponding reference stress value of each position of the stressed component;

[0020] Calculate the stress difference between the real-time stress value and the corresponding reference stress value of each position of each stressed component;

[0021] The calculation unit is further configured with a stress deviation threshold range; if the stress difference at any position of the stressed component at the current moment exceeds the stress deviation threshold range, the corresponding stressed component has an early collision.

[0022] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm described in the present invention, wherein: the processing module includes a path segmentation unit; the path segmentation unit is used to divide the reference path into sub-paths, specifically as follows:

[0023] Starting from the current coordinate of the end effector, trace the reference path along the preset moving direction of the end effector; during the tracing process, whenever the path length passed reaches the preset segmentation threshold or the direction change of the reference path reaches m times, add a segmentation marker point; m is a positive integer; continue tracing with the latest segmentation marker point as the starting point until reaching the end of the reference path; divide the reference path at each segmentation marker point to obtain + 1 sub-paths; wherein, is the number of segmentation marker points.

[0024] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm described in the present invention, wherein: the processing module further includes an obstacle detection unit; the obstacle detection unit is used to detect the coordinate range of the obstacle;

[0025] The processing module further includes a risk identification unit; the risk identification unit identifies the sub-paths with collision risks based on the coordinate range of the obstacle; specifically includes:

[0026] Based on the inverse kinematics algorithm, calculate the motion parameters of each joint during the end effector passing through each sub-path.

[0027] Calculate the motion trajectory of each joint based on the motion parameters; if there is an intersection between the motion trajectory of at least one joint and the coordinate range of the obstacle in any sub-path, there is a collision risk for the corresponding sub-path.

[0028] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm described in the present invention, wherein: the path correction module includes a joint evaluation unit; the joint evaluation unit is used to calculate the motion redundancy of different joints; specifically including:

[0029] Determine the obstacle-avoidance related joints based on the force-bearing components that have early collisions.

[0030] Obtain the maximum motion amount, minimum motion amount, and current motion amount of each obstacle-avoidance related joint.

[0031] The joint evaluation unit is configured with an optimization strategy, and calculates the optimal motion amount of each obstacle-avoidance related joint based on the optimization strategy.

[0032] Calculate the maximum adjustable amount and obstacle-avoidance adjustment amount of each obstacle-avoidance related joint; wherein, the maximum adjustable amount of any obstacle-avoidance related joint is the difference between the maximum motion amount and the minimum motion amount; the obstacle-avoidance adjustment amount of any obstacle-avoidance related joint is the absolute value of the difference between the current motion amount and the optimal motion amount.

[0033] Calculate the motion redundancy of each obstacle-avoidance related joint based on the maximum adjustable amount and obstacle-avoidance adjustment amount; wherein, the motion redundancy of any obstacle-avoidance related joint is the ratio of the maximum adjustable amount to the obstacle-avoidance adjustment amount.

[0034] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm described in the present invention, wherein: the optimization strategy includes solving the optimal motion amount of each obstacle-avoidance related joint based on the objective function and constraint conditions.

[0035] The constraint conditions include: during the obstacle-avoidance process, the distance index is greater than a preset distance index threshold; wherein, the distance index is denoted as , indicating the distance between the force-bearing component that has an early collision and the corresponding obstacle when the core motion vector is T; the core motion vector T is a vector composed of the core motion amounts of all obstacle-avoidance related joints.

[0036] The objective function is the change amount index; the change amount index is denoted as , indicating the sum of the squares of the change amounts of the core motion amounts of all obstacle-avoidance related joints.

[0037] Adopt an optimization algorithm to solve the core motion vector that minimizes the objective function while satisfying the constraint conditions, denoted as , The i-th element in corresponds to the optimal motion amount of the i-th obstacle-avoiding associated joint.

[0038] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm according to the present invention, wherein: the path correction module further includes a path correction unit; the path correction unit is configured with an obstacle-avoiding strategy; the obstacle-avoiding strategy is used to generate an obstacle-avoiding path, specifically as follows:

[0039] Select p obstacle-avoiding associated joints with the highest redundancy as adjustable joints; p is a positive integer less than or equal to n;

[0040] Taking the current coordinates of the end effector as the obstacle-avoiding starting point, select an obstacle-avoiding end point on the sub-path without collision risk;

[0041] Search for alternative obstacle-avoiding paths between the obstacle-avoiding starting point and the obstacle-avoiding end point, and perform collision detection on the alternative obstacle-avoiding paths; the collision detection specifically includes: only adjusting the motion parameters of the p adjustable joints, obtaining the motion trajectory of each joint; if the motion trajectories of all joints have no intersection with the coordinate range of the obstacle, the alternative obstacle-avoiding path passes the collision detection;

[0042] Record the alternative obstacle-avoiding path that passes the collision detection as the obstacle-avoiding path.

[0043] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm according to the present invention, wherein: the path correction unit is further configured with a correction strategy; the correction strategy is used to correct the path of the sub-path with collision risk, specifically as follows:

[0044] Generate at least q obstacle-avoiding paths based on the path correction unit; q is a positive integer;

[0045] Calculate the regression complexity of each obstacle-avoiding path; replace all sub-paths with collision risk with the obstacle-avoiding path with the minimum regression complexity to obtain the corrected reference path;

[0046] The calculation method of the regression complexity is as follows:

[0047] Calculate the path length of each joint moving in the obstacle-avoiding path; the path correction unit is configured with an angle change threshold; based on the angle change threshold, determine the number of significant corners of each joint in the obstacle-avoiding path, specifically including: counting the number of corners where the angle of each joint in the obstacle-avoiding path is greater than the angle change threshold as the number of significant corners;

[0048] Normalize the path length of each joint moving in the obstacle-avoiding path and the number of significant corners respectively, and perform weighted summation to obtain the regression complexity of the obstacle-avoiding path.

[0049] As a preferred solution of the dynamic adjustment system of the multi-degree-of-freedom robotic arm according to the present invention, wherein: the system further includes an artificial intervention module; the artificial intervention module includes a detection and monitoring unit and an early warning unit;

[0050] The obstacle avoidance strategy further includes: if each alternative obstacle avoidance path obtained by at least N searches fails to pass the collision detection, the path correction unit sends a search failure notice to the detection and monitoring unit; the detection and monitoring unit is configured with a failure times threshold, and if the number of search failure notices received by the detection and monitoring unit is greater than the failure times threshold, the early warning unit sends a warning message for artificial intervention in obstacle avoidance to the management personnel.

[0051] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0052] In this application, by monitoring the stress distribution of the key stressed components of the robotic arm and comparing the real-time stress value with the reference stress value, early collisions are accurately identified. This detection method starts directly from the structure of the robotic arm itself, reduces the real-time stress calculation amount, avoids the problem of reduced detection accuracy caused by environmental factor interference, and improves the accuracy and timeliness of collision detection.

[0053] By calculating the joint motion redundancy, the joints with high motion redundancy are preferentially adjusted, making full use of the redundant degrees of freedom of the robotic arm, enhancing the adaptability of the robotic arm in complex environments, ensuring the stable operation of the robotic arm, and improving the working efficiency and reliability of the robotic arm. By calculating the regression complexity to screen the best obstacle avoidance path to replace the risky sub-path, the path length and tortuosity are comprehensively considered, enabling the robotic arm to save time and energy consumption while avoiding obstacles and maintaining smooth motion, effectively avoiding the problems of unsmooth motion and collision risks caused by unreasonable path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings. Among them:

[0055] Figure 1 is a schematic structural diagram of the dynamic adjustment system of the multi-degree-of-freedom robotic arm provided by the present invention;

[0056] Figure 2 is a flowchart of the calculation unit for identifying early collisions of the robotic arm provided by the present invention;

[0057] Figure 3 is a flowchart of the collision recognition strategy provided by the present invention;

[0058] Figure 4 Flow chart for the path correction unit provided by the present invention to generate an obstacle avoidance path. Detailed implementation manners

[0059] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. Without conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0060] This embodiment introduces a dynamic adjustment system for a multi-degree-of-freedom robotic arm. Referring to Figure 1 , the dynamic adjustment system for a multi-degree-of-freedom robotic arm includes a collision detection module, a processing module, a path correction module, and an artificial intervention module; where:

[0061] The collision detection module detects early collisions of the robotic arm through stress monitoring.

[0062] The collision detection module includes a stress monitoring unit and a calculation unit; where:

[0063] The stress monitoring unit is used to monitor the stress distribution of each stressed component on the robotic arm in real time; the stress monitoring unit includes stress sensors installed at different positions of any stressed component; for any stressed component, during the operation of the robotic arm, real-time stress values at different positions are continuously collected and recorded through the stress sensors.

[0064] A stressed component is a component on the robotic arm that is prone to stress changes due to the movement of the connecting joints or the change of the load state during the operation of the robotic arm, such as an end effector, a connecting rod, etc.; preferably, in the embodiments of the present application, stress sensors are arranged based on the key stress points of each stressed component. For example, stress sensors are installed on the surface of the connecting rod along the length direction to monitor the stress change of the connecting rod during movement and determine whether abnormal stress is generated due to collisions or other reasons.

[0065] Referring to Figure 2 , the calculation unit identifies early collisions of the robotic arm based on the stress distribution of each stressed component; specifically including:

[0066] Real-time collect the motion parameters of each joint of the robotic arm.

[0067] For any rotating joint, the motion parameters at any moment include angular displacement, angular velocity, and angular acceleration; for any linear joint, the motion parameters at any moment include linear displacement and linear acceleration.

[0068] Based on the motion parameters of each joint of the robotic arm, obtain the reference stress values at different positions of each stressed component.

[0069] The calculation unit is configured with a stress mapping model; obtaining the reference stress value corresponding to each position of each stressed component in different poses specifically includes: inputting the motion parameters of each joint of the robotic arm into the stress mapping model, and the stress mapping model calculates and outputs the reference stress values of different positions of each stressed component.

[0070] The calculation unit is also configured with multi-body dynamics software for providing a data basis for the training of the stress mapping model in the following manner:

[0071] Model the robotic arm in the multi-body dynamics software;

[0072] Based on the reference path of the robotic arm, solve the motion parameters of each joint of the robotic arm at each moment after starting the operation through the inverse kinematics algorithm, and generate a motion change time series table; any column in the motion change time series table represents the motion parameters of different joints at the same moment; any row in the motion change time series table represents the motion parameters of the same joint at different moments.

[0073] The method for solving the motion parameters of each joint of the robotic arm at each moment after starting the operation through the inverse kinematics algorithm is as follows: establish a Cartesian space coordinate system to quantitatively record the trajectory of the reference path of the end effector of the robotic arm; use the inverse kinematics equation (such as the geometric relationship established by the D-H parameter method) to inversely solve the motion parameters of each joint (including the angular displacement of the rotary joint, the linear displacement of the prismatic joint, etc.).

[0074] The multi-body dynamics software generates a motion change time series table based on these motion parameters, records the states of each joint at each moment, and provides input for subsequent stress simulation.

[0075] Input the physical properties of the robotic arm and the motion change time series table into the multi-body dynamics software; the multi-body dynamics software simulates the operation of the robotic arm and outputs the theoretical stress values of different positions of each stressed component at each moment as reference stress values;

[0076] Record the simulation data at each moment; the simulation data at any moment includes an input part and an output part. The input part includes the motion parameters of each joint at this moment, and the output part is the reference stress value of different positions of each stressed component at this moment; organize the simulation data at any moment into a piece of training data to form a training set; construct a stress mapping model and perform model training based on the training set; the stress mapping model is any one of a linear regression model, a polynomial regression model, and a support vector regression model.

[0077] The multi-body dynamics software generates a large amount of data through simulation. Each piece of data contains joint motion parameters (input) and the theoretical stress values of the corresponding stressed components (output). By training the stress mapping model with the simulation data, the trained stress mapping model (such as the support vector regression model) can establish a non-linear mapping relationship from joint motion parameters to stress. By inputting joint parameters in real time, the reference stress value can be output, avoiding complex real-time finite element calculations.

[0078] The physical properties of the robotic arm input into the multi-body dynamics software include mass properties, such as the mass, centroid position, and moment of inertia of each stressed component; material properties, such as the elastic modulus, Poisson's ratio, density, yield strength, etc. of each stressed component; geometric properties, such as the geometric dimensions, shapes, connection methods, etc. of each component; and contact properties, such as the contact stiffness, friction coefficient, damping coefficient, etc. of the contact surfaces of each stressed component. Based on the virtual model of the robotic arm, as well as the input motion change time sequence table and the physical properties of the robotic arm, the multi-body dynamics software performs simulation calculations through multi-body dynamics theory and directly outputs the stress change data of each stressed component during the motion process, which can be organized into a stress change curve of the stress value at each position over time. In the embodiments of the present application, ADAMS or RecurDyn is preferably used as the multi-body dynamics software configured with the calculation unit, which can not only provide a large number of modeling elements to facilitate the rapid construction of the virtual model of the robotic arm, but also ensure the simulation calculation time.

[0079] Based on the reference stress values and the corresponding real-time stress values at different positions of each stressed component, identify the early collision of the robotic arm.

[0080] The calculation unit is also configured with a collision identification strategy; referring to Figure 3 , the collision identification strategy is used to identify the early collision of the robotic arm, specifically as follows:

[0081] Read the real-time stress value and the corresponding reference stress value at each position of the stressed component;

[0082] Calculate the stress difference between the real-time stress value and the corresponding reference stress value at each position of each stressed component;

[0083] The calculation unit is also configured with a stress deviation threshold range; if the stress difference at any position of the stressed component at the current moment exceeds the stress deviation threshold range, the corresponding stressed component has an early collision.

[0084] In this application, stress sensors are built into the key force-bearing components of the robotic arm to monitor the stress changes at different positions of each force-bearing component, and to infer whether the force-bearing components of the robotic arm are subjected to abnormal forces from obstacles, so as to directly identify early collisions from the structural state of the robotic arm itself. Further, this application predicts the stress reference values at different positions of the robotic arm through a trained stress mapping model, which can reduce the real-time stress calculation amount, achieve rapid identification of early collisions, and avoid the situation where the real-time calculation resource requirements are too high through methods such as numerical calculation or finite element analysis, resulting in untimely collision detection.

[0085] The processing module is configured with a reference path of the robotic arm; when the collision detection module detects the early collision, the processing module divides the reference path into sub-paths and identifies the sub-paths with collision risks;

[0086] The processing module includes a path segmentation unit, an obstacle detection unit, and a risk identification unit; where:

[0087] The path segmentation unit is used to divide the reference path into sub-paths; the path segmentation unit is configured with a path segmentation strategy, specifically as follows:

[0088] Starting from the current coordinates of the end effector, trace the reference path along the preset moving direction of the end effector; during the tracing process, whenever the path length passed reaches the preset segmentation threshold or the direction change of the reference path reaches m times, add a segmentation marker point; m is a positive integer; continue tracing with the latest segmentation marker point as the starting point until reaching the end of the reference path; divide the reference path at each segmentation marker point to obtain + 1 sub-paths; where, is the number of segmentation marker points.

[0089] The obstacle detection unit is used to detect the coordinate range of the obstacle;

[0090] The obstacle detection unit is configured with a lidar, a multi-camera, and a data processing sub-unit; where, the lidar is used to scan the obstacle at the location of the early collision to obtain point cloud data; the data processing sub-unit obtains the first coordinate range of the obstacle based on the point cloud data; the multi-camera is used to collect obstacle images at the location of the early collision from different angles; the data processing sub-unit obtains the second coordinate range of the obstacle based on the obstacle images from different angles; the data processing sub-unit is configured with a data fusion algorithm, and based on the data fusion algorithm, fuses the first coordinate range and the second coordinate range to obtain the coordinate range of the obstacle.

[0091] The data processing subunit aggregates the point cloud data into different obstacle clusters based on a clustering algorithm, identifies the positions and shapes of each obstacle, and thus obtains the first coordinate range; extracts the edges of the obstacles from the obstacle images through an edge detection algorithm, and then locates the obstacles based on a multi-view recognition algorithm, thereby obtaining the second coordinate range. Finally, data fusion algorithms such as Kalman filtering are used to more accurately estimate the actual coordinate range of the obstacles.

[0092] The lidar point cloud data has high precision but is sparse, and the pixel data of the multi-view camera images is dense but vulnerable to light interference. Kalman filtering iteratively corrects the estimated value through the state equation and the observation equation to reduce the influence of noise. For example, the ranging error of the lidar and the pixel error of the camera are weighted and fused through the covariance matrix to output a more robust obstacle coordinate range.

[0093] The risk identification unit identifies the sub-paths with collision risks based on the coordinate range of the obstacles; specifically includes:

[0094] Based on the inverse kinematics algorithm, calculate the motion parameters of each joint during the end effector passing through each sub-path;

[0095] Based on the motion parameters, calculate the motion trajectories of each joint; read the motion parameters of each joint at each moment obtained by inverse kinematics solution, including the angular displacement of the rotary joint and the linear displacement of the prismatic joint, and calculate the coordinates of each joint according to the displacement amount; the coordinates of each joint at different moments form the motion trajectory of the corresponding joint.

[0096] Discretize the motion trajectory of each joint into a point set in the coordinate system and perform a geometric intersection judgment with the obstacle coordinate range. If there is an intersection between the motion trajectory of at least one joint and the obstacle coordinate range in any sub-path, the corresponding sub-path has a collision risk.

[0097] The path correction module generates an obstacle avoidance path based on the motion redundancy of different joints of the robotic arm and corrects the sub-paths with collision risks based on the obstacle avoidance path.

[0098] The path correction module includes a joint evaluation unit and a path correction unit; among them:

[0099] The joint evaluation unit is used to calculate the motion redundancy of different joints; specifically includes:

[0100] Determine the obstacle avoidance related joints based on the force-bearing components that have early collisions; the obstacle avoidance related joints are the joints that can directly or indirectly adjust the force-bearing components that have early collisions; by changing the motion parameters of one or more obstacle avoidance related joints, an obstacle avoidance path can be generated to bypass the obstacles for the corresponding force-bearing components.

[0101] Obtain the maximum motion amount, minimum motion amount, and current motion amount of each obstacle - bypassing related joint;

[0102] The joint evaluation unit is configured with an optimization strategy, and calculates the optimal motion amount of each obstacle - bypassing related joint based on the optimization strategy;

[0103] Calculate the maximum adjustable amount and obstacle - bypassing adjustment amount of each obstacle - bypassing related joint; wherein, the maximum adjustable amount of any obstacle - bypassing related joint is the difference between the maximum motion amount and the minimum motion amount; the obstacle - bypassing adjustment amount of any obstacle - bypassing related joint is the absolute value of the difference between the current motion amount and the optimal motion amount;

[0104] Calculate the motion redundancy of each obstacle - bypassing related joint based on the maximum adjustable amount and the obstacle - bypassing adjustment amount; wherein, the motion redundancy of any obstacle - bypassing related joint is the ratio of the maximum adjustable amount to the obstacle - bypassing adjustment amount. The obstacle - bypassing related joint with a larger motion redundancy has a larger adjustment space in the current state, and thus is more suitable for obstacle - bypassing path planning. Prioritize adjusting the joint with a high motion redundancy for obstacle - bypassing path planning, and try to keep the state of the joint with a low motion redundancy unchanged, which can enable the robotic arm to avoid obstacles while maintaining the smoothness and efficiency of the overall motion. The obstacle - bypassing redundancy quantifies the adjustment potential of the joint. For example, joints with high redundancy, such as the middle joints of the robotic arm, can be adjusted significantly without affecting the end - effector accuracy, while joints with low redundancy, such as the end joints, need to be kept as stable as possible. When optimizing, prioritize adjusting the joints with high redundancy to minimize the overall motion disturbance.

[0105] The maximum motion amount, minimum motion amount, current motion amount, and optimal motion amount of any joint are different values of the core motion amount of the joint. The link coordinate system of the robotic arm describes the relationship between each joint through the D - H parameter method; for a revolute joint, the core motion amount is its angle in the link coordinate system; for a prismatic joint, the core motion amount is its displacement in the link coordinate system.

[0106] The optimization strategy includes solving the optimal motion amount of each obstacle - bypassing related joint based on the objective function and constraint conditions;

[0107] The constraint conditions include: during the obstacle - bypassing process, the distance index is greater than a preset distance index threshold; wherein, the distance index is denoted as , indicating the distance between the stressed component that has an early collision and the corresponding obstacle when the core motion vector is T; the core motion vector T is a vector composed of the core motion amounts of all obstacle - bypassing related joints; let T = , where represents the value of the core motion amount of the i - th obstacle - bypassing related joint in T; the value range of i is 1, 2, ……, n, and n is the number of obstacle - bypassing related joints;

[0108] The objective function is a variation index; the variation index is a , represents the sum of the squares of the changes in the core motion of all obstacle-related joints, and is calculated as follows:

[0109] ;

[0110] in, Indicates the current motion of the i-th associated joint.

[0111] The optimization algorithm is used to solve the core motion vector that minimizes the objective function while satisfying the constraints, which is denoted as , The i-th element in corresponds to the optimal motion of the i-th obstacle-circumventing joint. The optimization algorithm is a genetic algorithm or a particle swarm algorithm.

[0112] As the formula clearly indicates, the objective function is the sum of the squares of joint adjustments (i.e., changes in motion). The objective function is calculated using a typical least squares objective function, quantifying the magnitude of these adjustments by summing the squares of all joint adjustments. This sum-of-squares form has the following characteristics: The squared function is convex, ensuring that the optimization problem has a unique minimum, making it easier to solve. Penalizing large deviations: The squared term amplifies the impact of large deviations, thereby prioritizing reducing drastic adjustments to individual joints and avoiding sudden changes in the robot's motion. Minimizing the objective function forces the optimization algorithm to select the solution with the smallest joint adjustments. This ensures that the robot maintains the continuity of its original trajectory as much as possible when navigating obstacles. The objective function and constraints work together to find the adjustment solution that best approximates the current state while avoiding obstacles. This combination achieves obstacle avoidance while minimizing changes to the original path, minimizing the disruption of obstacle avoidance to the operational process. Because the objective function is a convex optimization problem, algorithms such as the particle swarm optimization algorithm can be used for rapid solution, meeting the real-time requirements of dynamic adjustments. The constraints set a safe distance for mandatory obstacle avoidance. Optimization algorithms such as genetic algorithms randomly generate multiple sets of motion parameters and screen solutions that meet constraints and minimize the objective function; particle swarm algorithms approximate the optimal solution through group iteration.

[0113] The path correction unit is configured with an obstacle avoidance strategy; Figure 4, the obstacle avoidance strategy is used to generate an obstacle avoidance path, which is specifically as follows: Select p obstacle avoidance related joints with the highest redundancy as adjustable joints, where p is a positive integer less than or equal to n; Use the current coordinates of the end effector as the obstacle avoidance starting point, select the obstacle avoidance end point on the sub path without collision risk, search for alternative obstacle avoidance paths between the obstacle avoidance starting point and the obstacle avoidance end point, and perform collision detection on the alternative obstacle avoidance paths; The collision detection specifically includes: Only adjust the motion parameters of the p adjustable joints to obtain the motion trajectory of each joint; If the motion trajectories of all joints have no intersection with the coordinate range of the obstacle, the alternative obstacle avoidance path passes the collision detection; Record the alternative obstacle avoidance path that passes the collision detection as the obstacle avoidance path.

[0114] The manual intervention module includes a detection and monitoring unit and an early warning unit;

[0115] The obstacle avoidance strategy further includes: If each alternative obstacle avoidance path obtained by at least N searches fails to pass the collision detection, the path correction unit sends a search failure notice to the detection and monitoring unit; N is a positive integer; The detection and monitoring unit is configured with a failure times threshold. If the number of search failure notices received by the detection and monitoring unit is greater than the failure times threshold, the early warning unit sends a warning message for manual intervention in obstacle avoidance to the management personnel.

[0116] Furthermore, each time a search failure notice is sent to the detection and monitoring unit, the path correction unit re selects the obstacle avoidance end point, or selects more than p obstacle avoidance joints with the highest redundancy as adjustable joints, and re explores the alternative obstacle avoidance paths.

[0117] The path correction unit is also configured with a correction strategy; The correction strategy is used to correct the path of the sub path with collision risk, which is specifically as follows:

[0118] Based on the path correction unit, at least q obstacle avoidance paths are generated, where q is a positive integer;

[0119] Calculate the regression complexity of each obstacle avoidance path; Replace all sub paths with collision risk with the obstacle avoidance path with the minimum regression complexity to obtain the corrected reference path;

[0120] The calculation method of the regression complexity is as follows:

[0121] Calculate the path length of each joint movement in the obstacle avoidance path; Divide the continuous obstacle avoidance path into multiple segments at adjacent time points. For example, record the spatial position of the joint at a fixed time step (such as every 0.1 second). Sum the path lengths between all adjacent time points to obtain the path length of the joint in the entire obstacle avoidance path.

[0122] The path correction unit is configured with an angular change threshold; determining the number of significant corners of each joint in the obstacle avoidance path based on the angular change threshold, specifically including: counting the number of corners where the angle of each joint in the obstacle avoidance path is greater than the angular change threshold as the number of significant corners;

[0123] Normalize the path length and the number of significant corners of each joint movement in the obstacle avoidance path respectively, and perform weighted summation to obtain the regression complexity of the obstacle avoidance path. The formula is as follows:

[0124] ;

[0125] where C represents the regression complexity of any obstacle avoidance path; is the path length of the j-th joint movement; is the number of significant corners of the k-th joint; is a preset path length reference value; is a preset reference value for the number of significant corners; the value ranges of j and k are both 1, 2,..., M, and M is the number of joints. 、 are weight coefficients, and are both non-negative and their sum is 1.

[0126] The path length and the number of significant corners are two important indicators for measuring the quality of the obstacle avoidance path. The path length is directly related to the time and energy consumed by the manipulator movement. A shorter path usually means higher obstacle avoidance efficiency and lower energy consumption. The number of significant corners reflects the tortuosity of the path. Too many significant corners will make the movement of the manipulator more complex, increasing the instability and collision risk during the movement. By comprehensively considering these two indicators in this application, the regression complexity can more comprehensively and accurately reflect the overall rationality of the obstacle avoidance path.

[0127] This application realizes the dynamic obstacle avoidance adjustment of the manipulator through the cooperation between modules. Detection first identifies the collision area through the stress difference and triggers an early warning; then path segmentation divides the reference path by direction or length to lock the risky sub-path; then obstacle positioning fuses multi-sensor data to accurately calibrate the coordinates of the obstacle; then redundancy evaluation selects high-degree-of-freedom joints as the main adjustment objects to ensure movement flexibility; finally path correction uses an optimization algorithm to generate candidate paths and selects the optimal solution through regression complexity evaluation to complete the real-time update of the global path. Each module is closely linked to form a closed-loop control of "detection - positioning - decision - execution", maximizing the movement efficiency of the manipulator while ensuring safety.

[0128] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0129] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the purpose and scope of the present invention. All of these are within the protection scope of the present invention.

Claims

1. A dynamic adjustment system for a multi-degree-of-freedom robotic arm, characterized in that, It includes a collision detection module, a processing module, and a path correction module; among which: The collision detection module detects early collisions of the robotic arm through stress monitoring; The processing module is configured with a reference path of the robotic arm; when the collision detection module detects the early collision, the processing module divides the reference path into sub-paths and identifies the sub-paths with collision risks; The path correction module generates an obstacle avoidance path based on the motion redundancy of different joints of the robotic arm and corrects the sub-paths with collision risks based on the obstacle avoidance path; The path correction module includes a joint evaluation unit; the joint evaluation unit is used to calculate the motion redundancy of different joints; specifically including: Determine the obstacle avoidance related joints based on the stressed components where early collisions occur; Obtain the maximum motion amount, minimum motion amount, and current motion amount of each obstacle avoidance related joint; The joint evaluation unit is configured with an optimization strategy and calculates the optimal motion amount of each obstacle avoidance related joint based on the optimization strategy; Calculate the maximum adjustable amount and the obstacle avoidance adjustment amount of each obstacle avoidance related joint; where, the maximum adjustable amount of any obstacle avoidance related joint is the difference between the maximum motion amount and the minimum motion amount; the obstacle avoidance adjustment amount of any obstacle avoidance related joint is the absolute value of the difference between the current motion amount and the optimal motion amount; Calculate the motion redundancy of each obstacle avoidance related joint based on the maximum adjustable amount and the obstacle avoidance adjustment amount; where, the motion redundancy of any obstacle avoidance related joint is the ratio of the maximum adjustable amount to the obstacle avoidance adjustment amount.

2. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 1, wherein: The collision detection module includes a stress monitoring unit and a calculation unit; the stress monitoring unit is used to monitor the stress distribution of each stressed component on the robotic arm in real time; the stress monitoring unit includes stress sensors installed at different positions of any stressed component; for any stressed component, during the operation of the robotic arm, real-time stress values at different positions are continuously collected and recorded through the stress sensors; The calculation unit identifies the early collision of the robotic arm based on the stress distribution of each stressed component; specifically including: Collect the motion parameters of each joint of the robotic arm in real time; Based on the motion parameters of each joint of the robotic arm, obtain the reference stress value corresponding to each position of each stressed component in different poses; Identify the early collision of the robotic arm based on the real-time stress value and the corresponding reference stress value of each position of each stressed component in different poses.

3. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 2, wherein: The calculation unit is configured with a stress mapping model; the obtaining of the reference stress value corresponding to each position of each stressed component in different poses specifically includes: inputting the motion parameters of each joint of the robotic arm into the stress mapping model, and the stress mapping model calculates and outputs the reference stress values of different positions of each stressed component; the stress mapping model is any one of a linear regression model, a polynomial regression model, and a support vector regression model.

4. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 3, characterized in that: The calculation unit is also configured with a collision identification strategy; the collision identification strategy is used to identify the early collision of the robotic arm, specifically as follows: Read the real-time stress value and the corresponding reference stress value of each position of the stressed component; Calculate the stress difference between the real-time stress value and the corresponding reference stress value of each position of each stressed component; The calculation unit is also configured with a stress deviation threshold range; if the stress difference at any position of the force-bearing component at the current moment exceeds the stress deviation threshold range, there is an early collision in the corresponding force-bearing component.

5. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 4, characterized in that: The processing module includes a path segmentation unit; the path segmentation unit is used to divide the reference path into sub-paths, specifically as follows: Starting from the current coordinates of the end effector, trace the reference path along the preset moving direction of the end effector; during the tracing process, whenever the path length passed reaches the preset segmentation threshold or the direction change of the reference path reaches m times, add a segmentation marker point; m is a positive integer; continue tracing with the latest segmentation marker point as the starting point until the end of the reference path is reached; at each segmentation marker point, segment the reference path to obtain +1 sub-paths; where is the number of segmentation marker points.

6. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 5, characterized in that: The processing module further includes an obstacle detection unit; the obstacle detection unit is used to detect the coordinate range of the obstacle; The processing module further includes a risk identification unit; the risk identification unit identifies the sub-paths with collision risks based on the coordinate range of the obstacle; specifically includes: Based on the inverse kinematics algorithm, calculate the motion parameters of each joint during the end effector passing through each sub-path; Based on the motion parameters, calculate the motion trajectory of each joint; if there is an intersection between the motion trajectory of at least one joint and the coordinate range of the obstacle in any sub-path, the corresponding sub-path has a collision risk.

7. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 6, wherein: The optimization strategy includes solving the optimal motion amount of each obstacle avoidance related joint based on the objective function and constraint conditions; The constraint conditions include: during the obstacle avoidance process, the distance index is greater than a preset distance index threshold; where the distance index is denoted as , representing the distance between the force-bearing component that experiences an early collision and the corresponding obstacle when the core motion vector is T; the core motion vector T is a vector composed of the core motion amounts of all obstacle avoidance-related joints; The objective function is a variation index; the variation index is denoted as , representing the sum of squares of the variations of the core motion amounts of all obstacle-avoiding related joints; Using an optimization algorithm, while satisfying the constraint conditions, solve for the core motion vector that minimizes the objective function, denoted as ; The i-th element in corresponds to the optimal motion amount of the i-th obstacle avoidance associated joint.

8. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 7, characterized in that: The path correction module further includes a path correction unit; the path correction unit is configured with an obstacle avoidance strategy; the obstacle avoidance strategy is used to generate an obstacle avoidance path, specifically as follows: Select the p obstacle avoidance related joints with the highest redundancy as adjustable joints; p is a positive integer less than or equal to n, and n is the number of obstacle avoidance related joints; Take the current coordinate of the end effector as the obstacle avoidance starting point, and select the obstacle avoidance end point on the sub-path without collision risk; Search for alternative obstacle avoidance paths between the obstacle avoidance starting point and the obstacle avoidance end point, and perform collision detection on the alternative obstacle avoidance paths; the collision detection specifically includes: only adjusting the motion parameters of the p adjustable joints, obtaining the motion trajectory of each joint; if the motion trajectories of all joints have no intersection with the coordinate range of the obstacle, the alternative obstacle avoidance path passes the collision detection; Record the alternative obstacle avoidance path that passes the collision detection as the obstacle avoidance path.

9. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 8, wherein: The path correction unit is also configured with a correction strategy; the correction strategy is used to correct the path of the sub-path with collision risk, specifically as follows: Generate at least q obstacle avoidance paths based on the path correction unit; q is a positive integer; Calculate the regression complexity of each obstacle avoidance path; replace all sub-paths with collision risks with the obstacle avoidance path with the minimum regression complexity to obtain the corrected reference path; The calculation method of the regression complexity is as follows: Calculate the path length moved by each joint in the obstacle avoidance path; The path correction unit is configured with an angle change threshold; Based on the angle change threshold, determine the number of significant corners of each joint in the obstacle avoidance path, specifically including: counting the number of corners where the angle of each joint in the obstacle avoidance path is greater than the angle change threshold as the number of significant corners; Normalize the path length moved by each joint and the number of significant corners in the obstacle avoidance path respectively, and perform weighted summation to obtain the regression complexity of the obstacle avoidance path.

10. The dynamic adjustment system of the multi-degree-of-freedom robotic arm according to claim 9, characterized in that: The system further includes an artificial intervention module; The artificial intervention module includes a detection and monitoring unit, a warning unit; The obstacle avoidance strategy further includes: if each alternative obstacle avoidance path obtained by at least N searches fails to pass the collision detection, the path correction unit sends a search failure notification to the detection and monitoring unit; the detection and monitoring unit is configured with a failure times threshold, and if the number of search failure notifications received by the detection and monitoring unit is greater than the failure times threshold, the warning unit sends a warning message for manual intervention in obstacle avoidance to the management personnel.

Images