Virtual sampling-based six-axis palletizing robot adaptation method, storage medium and system

Abstract

The invention relates to the technical field of robot control, in particular to a six-axis palletizing robot adaptation method based on virtual sampling, a storage medium and a system. The method specifically comprises the steps of obtaining parameter information and actual working environment information of a robot, and determining a target working space of the robot through kinematics analysis; based on the target working space, a pickup point set and a placement point set of the materials are generated through virtual sampling, and path samples of the robot in the stacking process are generated; performing trajectory planning on the path sample, and calculating and optimizing an optimal motion parameter; performing robot action sequence scheduling optimization on the pickup point set and the placement point set to obtain an optimal scheduling scheme; and finally, integrating the optimal motion parameters and the optimal scheduling scheme to generate a control program, and completing adaptation through verification and fine tuning in an actual environment. According to the method, through an automatic program generation mode, the flexibility and adaptability of the six-axis palletizing robot to cope with various working scenes are improved.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, specifically to a method, storage medium, and system for adapting a virtual sampling six-axis palletizing robot. Background Technology

[0002] In modern industry, automation technology, with industrial robots at its core, has become a key means of improving production efficiency and ensuring operational quality. Industrial robots, in particular, have demonstrated unique advantages in areas such as material handling and stacking.

[0003] Traditionally, the programming and adaptation of such robots has relied on manual processes. Engineers need to write complex sequences of motion instructions to customize the robot's robotic arm, gripping tools, and control system to ensure that the robotic arm can match specific production processes. However, as production scenarios become increasingly complex, new requirements are being placed on existing robot program control technologies. Existing technologies mainly address how to enable the robotic arm to accurately execute a predefined control program, but they have some problems dealing with situations where there are multiple dynamically changing initial and target positions in the work environment. In such multi-tasking scenarios, traditional offline programming methods are inadequate because the preset program cannot autonomously decide and optimize the task execution sequence. Existing technologies also have some problems in automatically transforming high-level operational requirements into complete control programs that include complex path planning and optimized task scheduling.

[0004] To address this, a method, storage medium, and system for adapting a six-axis palletizing robot based on virtual sampling are proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a method, storage medium, and system for adapting a six-axis palletizing robot based on virtual sampling. First, the parameter information and actual working environment information of the robot to be adapted are acquired. Based on the parameter information, kinematic analysis is performed on the robot to obtain the workspace boundary of the robot's end effector. Then, based on the actual working environment information and workspace boundary, the target workspace is determined. Based on the target workspace, a set of material pick-up points and a set of material placement points are generated through virtual sampling, and a set of path samples for the robot during the palletizing process is generated. Next, trajectory planning is performed on the path samples in the path sample set, and the optimal motion parameters are calculated and optimized to ensure smooth robot operation and reduce losses. Objective functions and violation functions are constructed to optimize the robot's action sequence scheduling for multiple palletizing points, improving optimization efficiency while ensuring the feasibility of the scheduling scheme, resulting in the optimal scheduling scheme. Finally, the optimal motion parameters and the optimal scheduling scheme are integrated to generate a control program. The performance of the control program is verified to generate verification results, and fine-tuning is performed based on the verification results to complete the final adaptation. This invention improves the flexibility and adaptability of a six-axis palletizing robot to various working scenarios through an automated program generation method.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A virtual sampling-based adaptation method for a six-axis palletizing robot includes:

[0008] Obtain the parameter information and actual working environment information of the robot to be adapted;

[0009] Based on the parameter information, kinematic analysis of the robot is performed to obtain the workspace boundary of the robot's end effector.

[0010] Based on the actual working environment information and workspace boundaries, the target workspace is determined; based on the target workspace, a set of material pick-up points and a set of material placement points are generated through virtual sampling, and a set of path samples for the robot during the palletizing process is generated.

[0011] Trajectory planning is performed on path samples in the path sample set, and the optimal motion parameters are calculated and optimized. The path samples include point-to-point paths, straight paths, and circular paths. The optimal motion parameters constitute the core motion commands of the robot control program.

[0012] Objective function and violation function are constructed to optimize robot action sequence scheduling for the pick-up point set and the placement point set, and obtain the optimal scheduling scheme. The objective function is used to maximize the total number of palletized materials, minimize the energy consumed by the palletized materials, and maximize the uniformity of container filling at each palletizing point. Violation function is used to ensure the feasibility of the scheduling scheme.

[0013] The optimal motion parameters and optimal scheduling scheme are integrated to generate a control program. The performance of the control program is verified to generate verification results. Based on the verification results, fine-tuning is performed to complete the final adaptation.

[0014] Preferably, the kinematic analysis includes: obtaining the geometric parameters and joint limit ranges of each link of the robot based on parameter information; establishing a robot base coordinate system, and defining local coordinate systems for the robot links using the base coordinate system; obtaining DH parameters based on the local coordinate systems, wherein the DH parameters are used to describe the relative position and orientation between two adjacent links; establishing the kinematic equations of two adjacent local coordinate systems based on the DH parameters to obtain the pose matrix; and determining the workspace boundary using the limit step angle method based on the pose matrix and the limit ranges of the robot joints.

[0015] Preferably, the point-to-point path specifically includes: obtaining a set of key points in the point-to-point path; using a genetic algorithm to optimize and determine the optimal time interval sequence between key points; designing a first preset pose matrix for the robot at the key points; performing inverse kinematics on the first preset pose matrix to obtain the first preset joint angles of each joint of the robot at the key points; and performing cubic spline curve interpolation based on the first preset joint angles to generate the joint angles, angular velocities, and angular acceleration motion trajectories that constitute the core motion commands.

[0016] Preferably, the straight path specifically includes: obtaining the coordinates of the starting point and ending point of the straight path; inserting straight interpolation points at equal time intervals on the straight path using a straight interpolation algorithm with acceleration and deceleration control; designing a second preset pose matrix for the robot at the straight interpolation points; performing inverse kinematics on the second preset pose matrix to obtain a second preset joint angle of the robot joint; and performing cubic spline curve interpolation based on the second preset joint angle to generate the joint angle, angular velocity, and angular acceleration motion trajectory that constitutes the core motion command.

[0017] Preferably, the circular path specifically includes: obtaining the coordinates of the starting point, ending point, and intermediate point of the circular path; determining and calculating the center and radius of the circular path; establishing a circular coordinate system based on the center and the plane where the circular path is located; calculating the circular interpolation point in the circular coordinate system based on the center and radius using a circular interpolation algorithm; transforming the circular interpolation point from the circular coordinate system to the base coordinate system; designing a third preset pose matrix for the robot at the circular interpolation point; performing inverse kinematics on the third preset pose matrix to obtain the third preset joint angles of each joint of the robot; and performing cubic spline curve interpolation based on the third preset joint angles to generate the joint angles, angular velocities, and angular acceleration motion trajectories that constitute the core motion commands.

[0018] Preferably, the optimization process of the genetic algorithm includes constraints, including velocity constraints, acceleration constraints, and impact constraints; constructing a cubic spline function based on the time interval sequence to obtain a cubic spline curve; analyzing the cubic spline curve to obtain the angular velocity, angular acceleration, and angular jerk at any time for each joint; and retaining the time interval sequence when the maximum angular velocity, maximum angular acceleration, and maximum angular jerk satisfy all constraints.

[0019] Preferably, the robot action sequence scheduling optimization specifically includes: randomly generating a first scheduling scheme set; randomly selecting an intermediate scheduling scheme based on the first scheduling scheme set; generating a first candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; traversing the first scheduling scheme set to obtain a second scheduling scheme set; calculating the selection probability of the scheduling schemes in the second scheduling scheme set, and taking the scheduling scheme with the highest selection probability as the intermediate scheduling scheme; generating a second candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; performing a merit-based evaluation of the first candidate scheduling scheme and the second candidate scheduling scheme according to the objective function and the violation function, and deciding whether to update the scheduling scheme set based on the evaluation results; if the scheduling scheme set is updated, monitoring and replacing the scheduling schemes whose optimization effect has stalled with newly generated random scheduling schemes and re-executing the above process until the preset algorithm termination condition is reached.

[0020] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the virtual sampling six-axis palletizing robot adaptation method described above.

[0021] A virtual sampling six-axis palletizing robot adaptation system includes:

[0022] The data acquisition module is used to acquire parameter information and actual working environment information of the robot to be adapted;

[0023] The kinematics analysis module is used to perform kinematic analysis on the robot based on parameter information to obtain the workspace boundary of the robot's end effector;

[0024] The virtual sampling module is used to determine the target workspace based on the actual working environment information and workspace boundaries; based on the target workspace, it generates a set of material pick-up points and a set of material placement points through virtual sampling, and generates a set of path samples for the robot during the palletizing process.

[0025] The trajectory planning module is used to plan trajectories for path samples in the path sample set and calculate the optimal motion parameters.

[0026] The scheduling optimization module is used to construct objective functions and violation functions to optimize the robot action sequence scheduling of the pick-up point set and the placement point set, and obtain the optimal scheduling scheme.

[0027] The robot adaptation module integrates the optimal motion parameters and the optimal scheduling scheme to generate a control program, verifies the performance of the control program to generate verification results, and makes fine adjustments based on the verification results to complete the final adaptation.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0029] 1. This invention uses virtual sampling technology to adapt a six-axis palletizing robot, taking into account the situation where there are multiple palletizing points. First, the target workspace is determined through kinematic analysis. Then, multiple material pick-up points and palletizing points are determined within the target workspace. Path samples are generated and trajectory planning is performed. Finally, objective function and violation function are constructed to optimize the palletizing task scheduling for multiple palletizing points, improving optimization efficiency while ensuring the feasibility of the scheduling scheme. Finally, it is verified and fine-tuned in a real environment to ensure the practicality of the adaptation results and significantly improve the ability of the six-axis palletizing robot to adapt to different working scenarios.

[0030] 2. This invention employs a comprehensive trajectory planning method, encompassing point-to-point paths, straight-line paths, and circular paths. For point-to-point paths, a genetic algorithm is used to optimize the time interval sequence, and a smooth motion trajectory is constructed using cubic spline functions. For straight-line and circular paths, intermediate point interpolation is performed, along with cubic spline curve interpolation, ensuring the continuity of position, velocity, and acceleration, thus improving the smoothness and stability of the robot's motion. By comprehensively considering different types of paths, the robot's spatial motion capability and adaptability are significantly enhanced. This comprehensive trajectory optimization method can adapt to various complex palletizing task scenarios.

[0031] 3. This invention employs an iterative palletizing task scheduling optimization method. Through iterative updates of multiple scheduling scheme sets, the scheduling results are continuously optimized. It combines random search and probability-based selection strategies, ensuring global search capabilities while also considering local optimization. During the optimization process, both the objective function and the violation function are considered to ensure the feasibility of the selected scheduling scheme. It can effectively handle complex palletizing tasks, significantly improving the scheduling efficiency and flexibility of palletizing tasks, and enhancing the ability of the six-axis palletizing robot system to adapt to changing work requirements. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the adaptation method for a six-axis palletizing robot based on virtual sampling provided in an embodiment of the present invention;

[0033] Figure 2 A schematic diagram of the structure of a virtual sampling six-axis palletizing robot adaptation system provided in an embodiment of the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] Please see Figure 1 , Figure 2 This invention provides a method, storage medium, and system for adapting a six-axis palletizing robot based on virtual sampling. The technical solution is as follows:

[0036] A virtual sampling-based adaptation method for a six-axis palletizing robot includes:

[0037] Obtain the parameter information and actual working environment information of the robot to be adapted;

[0038] Based on the parameter information, kinematic analysis of the robot is performed to obtain the workspace boundary of the robot's end effector.

[0039] Based on the actual working environment information and workspace boundaries, the target workspace is determined; based on the target workspace, a set of material pick-up points and a set of material placement points are generated through virtual sampling, and a set of path samples for the robot during the palletizing process is generated.

[0040] Trajectory planning is performed on path samples in the path sample set, and the optimal motion parameters are calculated and optimized. The path samples include point-to-point paths, straight paths, and circular paths. The optimal motion parameters constitute the core motion commands of the robot control program.

[0041] Objective function and violation function are constructed to optimize robot action sequence scheduling for the pick-up point set and the placement point set, and obtain the optimal scheduling scheme. The objective function is used to maximize the total number of palletized materials, minimize the energy consumed by the palletized materials, and maximize the uniformity of container filling at each palletizing point. Violation function is used to ensure the feasibility of the scheduling scheme.

[0042] The optimal motion parameters and optimal scheduling scheme are integrated to generate a control program. The performance of the control program is verified to generate verification results. Based on the verification results, fine-tuning is performed to complete the final adaptation.

[0043] Example 1

[0044] Company A needs to stack three types of materials into different containers for loading and transportation. They applied a virtual sampling-based six-axis palletizing robot adaptation method. (See [link / reference]). Figure 1 ,include:

[0045] A virtual sampling-based adaptation method for a six-axis palletizing robot includes:

[0046] Obtain the parameter information and actual working environment information of the robot to be adapted;

[0047] Based on the parameter information, kinematic analysis of the robot is performed to obtain the workspace boundary of the robot's end effector.

[0048] Based on the actual working environment information and workspace boundaries, the target workspace is determined; based on the target workspace, a set of material pick-up points and a set of material placement points are generated through virtual sampling, and a set of path samples for the robot during the palletizing process is generated.

[0049] Trajectory planning is performed on path samples in the path sample set, and the optimal motion parameters are calculated and optimized. The path samples include point-to-point paths, straight paths, and circular paths. The optimal motion parameters constitute the core motion commands of the robot control program.

[0050] Objective function and violation function are constructed to optimize robot action sequence scheduling for the pick-up point set and the placement point set, and obtain the optimal scheduling scheme. The objective function is used to maximize the total number of palletized materials, minimize the energy consumed by the palletized materials, and maximize the uniformity of container filling at each palletizing point. Violation function is used to ensure the feasibility of the scheduling scheme.

[0051] The optimal motion parameters and optimal scheduling scheme are integrated to generate a control program. The performance of the control program is verified to generate verification results. Based on the verification results, fine-tuning is performed to complete the final adaptation.

[0052] The parameter information includes the linkage geometry, joint limit range, maximum allowable speed, maximum acceleration, and maximum jerk of each joint motor; the actual working environment information includes: the three-dimensional geometric boundaries of fixed obstacles in the working area, the precise position coordinate range of the material pick-up station, and the entry position, size, and expected material stacking method of each palletizing container.

[0053] Further, the kinematic analysis includes: obtaining the geometric parameters and joint limit ranges of each link of the robot based on parameter information; establishing a robot base coordinate system, and defining local coordinate systems for the robot links using the base coordinate system; obtaining DH parameters based on the local coordinate systems, the DH parameters being used to describe the relative position and orientation between two adjacent links; establishing the kinematic equations of two adjacent local coordinate systems based on the DH parameters to obtain the pose matrix; and determining the workspace boundary using the limit step angle method based on the pose matrix and the limit ranges of the robot joints; the kinematic equations are:

[0054]

[0055] Among them, A i+1θ represents the transformation matrix from the i-th local coordinate system to the (i+1)-th local coordinate system. i α represents the joint angle of the i-th link; i The l represents the torsion angle of the i-th link; i d represents the length of the i-th link; i R represents the offset of the i-th link; R represents the pose matrix of the robot's end effector; Q represents the position of the i-th link. 3*3 P represents a 3x3 rotation matrix used to describe the attitude of the end effector relative to the base coordinate system; 3*1 This represents a 3x1 coordinate vector used to describe the position of the end effector relative to the base coordinate system.

[0056] In this embodiment of the application, an example process for establishing a local coordinate system is as follows: The i-th local coordinate system first revolves around x i Axis rotation angle α i Move the origin along x i Axis travel distance l i Then around z i Axis rotation angle θ i Move the origin along z i The axis moves a distance d, resulting in the (i+1)th local coordinate system; where counterclockwise rotation is positive; the origin of the base coordinate system is located at the base of the six-axis palletizing robot, and the DH parameter of the first link describes the positional relationship between the first link and the base; the DH parameters of the six-axis palletizing robot are shown in Table 1.

[0057] Table 1 Parameter Table for the Six-Axis Palletizing Robot DH

[0058] Linkage number Twist angle α Rod length l / m Offset d / m Joint angle θ Joint angle range 1 90° 0.17 0.5 <![CDATA[θ1]]> [-180°,180°] 2 0 0.75 0 <![CDATA[θ2]]> [-150°,60°] 3 90° 1.02 0 <![CDATA[θ3]]> [-100°,90°] 4 -90° 0 0.60 <![CDATA[θ4]]> [-150°,150°] 5 90° 0.11 0 <![CDATA[θ5]]> [-120°,120°] 6 0 0 0.16 <![CDATA[θ6]]> [-180°,180°]

[0059] Through kinematic analysis of the robot, the robot's pose matrix and workspace boundaries were accurately determined, providing a reliable kinematic basis for subsequent trajectory planning and palletizing task scheduling optimization.

[0060] The extreme step angle method is as follows: First, fix one joint of the robot at the limit position of its range of motion. Then, traverse the entire range of motion of all other joints with a preset, fine angle step size (e.g., 1 degree). During this process, record the farthest spatial point that the robot end can reach. Repeat this process until every combination of extreme positions of all joints has been tested. Finally, connect all the recorded farthest points to form a complete and accurate outer boundary of the workspace.

[0061] Furthermore, the point-to-point path specifically includes: obtaining a set of key points in the point-to-point path, using a genetic algorithm to optimize and determine the optimal time interval sequence between key points; designing a first preset pose matrix for the robot at the key points; performing inverse kinematics on the first preset pose matrix to obtain the first preset joint angles of each joint of the robot at the key points; and performing cubic spline curve interpolation based on the first preset joint angles to generate the joint angles, angular velocities, and angular acceleration motion trajectories that constitute the core motion commands.

[0062] Specifically, obtain the set of key points {q0,q1,…,q} in the point-to-point path. n}, where q0 represents the starting point of the point-to-point path, q n This represents the endpoint of the point-to-point path; the time sequence corresponding to the robot's end effector reaching the key point is {t0, t1, ..., t...}. n The time interval sequence is {T1,T2,…,T}. n}, T n =t n -t n-1 The optimal time interval sequence is obtained using a genetic algorithm. The specific steps include:

[0063] S1. Randomly generate N time interval sequences as the initial population;

[0064] S2. Calculate the fitness of each time interval sequence in the initial population; the fitness is the reciprocal of the point-to-point path execution time;

[0065] S3. Sort the fitness of each time interval sequence in the initial population, and select the top M time interval sequences with the highest fitness to form the first temporary population;

[0066] S4. Perform a crossover operation on the remaining NM time interval sequences in the initial population to obtain a second temporary population; the size of the second temporary population is NM;

[0067] S5. Perform a mutation operation on each time interval sequence in the second temporary population to obtain a third temporary population; the size of the third temporary population is NM;

[0068] S6. Merge the first temporary population and the third temporary population to obtain a offspring population; the size of the offspring population is N;

[0069] S7. Determine whether the termination condition has been met. If the termination condition has not been met, replace the initial population with the offspring population and return to step S2. If the termination condition has been met, select the time interval sequence with the highest fitness in the offspring population as the optimal time interval sequence.

[0070] Design a first preset pose matrix for the robot at each key point. The position part of the first preset pose matrix is ​​determined by the spatial coordinates of the key point, and the posture part follows preset constraint rules. In this embodiment, to ensure the smoothness of the handling process, the rule is to keep the orientation of the robot's end effector consistent with the initial orientation at the starting point of the path throughout the entire path. Perform inverse kinematics on the first preset pose matrix to obtain the first preset joint angles of each joint of the robot at the key points. Based on the first preset joint angles, perform cubic spline curve interpolation to obtain the angle, angular velocity, and angular acceleration of each joint at any time. The cubic spline curve interpolation includes: obtaining n time intervals based on the optimal time interval sequence; constructing a cubic spline function for each joint in each time interval; solving the coefficients of the cubic spline function using MATLAB based on the first preset joint angles, utilizing the conditions of angle continuity, velocity continuity, and acceleration continuity to obtain the cubic spline curves of each joint of the robot; and analyzing the cubic spline curves to obtain the angle, angular velocity, and angular acceleration of each joint at any time.

[0071] In this embodiment, the point-to-point path description has two scenarios. The first is the process of the six-axis palletizing robot moving its end effector from the initial position to the location of the material to be picked up. The second is the process of the robot placing the material into the palletizing container and then moving to the next material pick-up point. In the point-to-point path, the robot's end effector does not pick up the material. Therefore, reducing the path execution time is the primary goal. The use of a genetic algorithm to optimize the time interval improves the efficiency of robot operation. At the same time, cubic spline curve interpolation ensures smooth operation of the robot joints and reduces wear and tear.

[0072] Furthermore, the straight path specifically includes: obtaining the coordinates of the start and end points of the straight path; inserting straight interpolation points at equal time intervals on the straight path using a straight interpolation algorithm with acceleration and deceleration control; designing a second preset pose matrix for the robot at the straight interpolation points, wherein the second preset pose matrix follows the same posture constraint rules as the first preset pose matrix, i.e., keeping the posture of the end effector constant; performing inverse kinematics on the second preset pose matrix to obtain the second preset joint angles of the robot joints; and performing cubic spline curve interpolation based on the second preset joint angles to generate the joint angles, angular velocities, and angular acceleration motion trajectories that constitute the core motion commands.

[0073] In this embodiment, the straight path describes the process of a six-axis palletizing robot lifting or placing materials. It is necessary to ensure the smooth operation of the trajectory. The straight path is optimized by a straight interpolation algorithm based on trapezoidal velocity curve features, and then cubic spline curve interpolation is performed to improve the smoothness of the path execution over the entire trajectory and avoid accidental damage to materials when the robot lifts or places them.

[0074] Further, the circular path specifically includes: obtaining the coordinates of the starting point, ending point, and intermediate point of the circular path; determining and calculating the center and radius of the circular path; establishing a circular coordinate system based on the center and the plane where the circular path is located; using a circular interpolation algorithm to calculate the circular interpolation point in the circular coordinate system based on the center and radius; transforming the circular interpolation point from the circular coordinate system to the base coordinate system; designing a third preset pose matrix for the robot at the circular interpolation point, wherein the third preset pose matrix follows the same attitude constraint rules as the first preset pose matrix, i.e., keeping the attitude of the end effector constant; performing inverse kinematics on the third preset pose matrix to obtain the third preset joint angles of each joint of the robot; and performing cubic spline curve interpolation based on the third preset joint angles to generate the joint angles, angular velocities, and angular acceleration motion trajectories that constitute the core motion commands.

[0075] In this embodiment, the circular arc path describes the process of the six-axis palletizing robot lifting the material and moving it to the palletizing point. It is necessary to ensure the smooth operation of the trajectory. The circular arc interpolation algorithm can generate an accurate circular arc motion trajectory, and then cubic spline curve interpolation is performed to improve the smoothness of the path execution on the entire trajectory and avoid the robot accidentally damaging the material when moving the material.

[0076] Furthermore, the optimization process of the genetic algorithm includes constraints, including velocity constraints, acceleration constraints, and impact constraints; a cubic spline function is constructed based on the time interval sequence to obtain a cubic spline curve; the cubic spline curve is analyzed to obtain the angular velocity, angular acceleration, and angular jerk at any time for each joint; the time interval sequence is retained when the maximum angular velocity, maximum angular acceleration, and maximum angular jerk satisfy all constraints.

[0077] By incorporating velocity, acceleration, and impact constraints into each iteration of the genetic algorithm, the stability and reliability of the method are improved while ensuring that the trajectory planning time is optimal and that the obtained time interval sequence meets the actual physical constraints and safety requirements.

[0078] Furthermore, the robot action sequence scheduling optimization specifically includes: randomly generating a first scheduling scheme set; randomly selecting an intermediate scheduling scheme based on the first scheduling scheme set; generating a first candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; traversing the first scheduling scheme set to obtain a second scheduling scheme set; calculating the selection probability of the scheduling schemes in the second scheduling scheme set, and taking the scheduling scheme with the highest selection probability as the intermediate scheduling scheme; generating a second candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; performing a merit-based evaluation of the first candidate scheduling scheme and the second candidate scheduling scheme according to the objective function and the violation function, and deciding whether to update the scheduling scheme set based on the evaluation results; if the scheduling scheme set is updated, monitoring and replacing the scheduling schemes whose optimization effect has stalled with newly generated random scheduling schemes and re-executing the above process until the preset algorithm termination condition is reached.

[0079] Specifically, the first set of scheduling schemes is generated randomly. Each scheduling scheme is a D-dimensional array, where each element in the array has a value... Indicates the next action of the palletizing robot; sets an evaluation counter `count`, initially set to 0; for the current scheduling scheme `x`... m ∈F, set an optimization counter t m The initial value is 0; a scheduling scheme is randomly selected from the first set of scheduling schemes as the intermediate scheduling scheme r. m Based on the intermediate scheduling scheme and the current scheduling scheme, a candidate scheduling scheme v is generated. m The calculation formula is as follows:

[0080]

[0081] Where i represents any dimension in the D-dimensional scheduling scheme;

[0082] The current scheduling scheme and the candidate scheduling scheme are compared based on the objective function and the violation function;

[0083] The formulas for the objective function f(x) and the violation function v(x) are as follows:

[0084]

[0085] Where x represents the scheduling scheme; N pallet Indicates the total quantity of palletized materials; C energy Indicates the energy consumed; ΔN pallet α represents the maximum value of the pairwise differences in quantity from each pick-up point in the palletized material; α and β represent the weighting factors for multi-objective optimization; h i(x) represents the material waiting time violation at the i-th pick-up point; c represents the number of pick-up points; g(x) represents the action violation, i.e., the number of times the material fails to arrive when the robot's end effector moves to the next material pick-up point; max t i This represents the maximum waiting time for the material at the i-th pick-up point during the palletizing process. This represents the maximum allowable waiting time for the material at the i-th pick-up point.

[0086] In this embodiment of the application, the consumed energy C energy C is a quantitative index calculated from the optimal motion parameters generated by the trajectory planning step according to the present invention. This index aims to characterize the smoothness and economy of the robot's motion trajectory during task completion. The specific calculation method is as follows: First, based on the small time steps of the motion trajectory of each joint obtained by trajectory planning, the instantaneous total driving torque required for each joint is estimated. The instantaneous total driving torque is the sum of the torques required to overcome the inertia of the robot's links and loads, resist the effect of gravity, and overcome joint friction. By multiplying the instantaneous total driving torque of each joint by its instantaneous angular velocity at that moment, the instantaneous output mechanical power of that joint is obtained. Then, the instantaneous output mechanical power of all joints is added together to obtain the instantaneous mechanical power of the robot as a whole. Then, the overall power is integrated over the time of the entire motion process to obtain the total mechanical energy output by the robot. The calculated total mechanical energy is divided by a preset comprehensive efficiency coefficient related to the motor and drive system, where the comprehensive efficiency coefficient is less than 1, thereby obtaining C, which can truly reflect the total power consumption. energy The overall efficiency coefficient and C energy The relationship is negative, meaning that the lower the efficiency, the higher the energy consumption.

[0087] The weighting factors α and β in the multi-objective optimization are configurable parameters preset before the adaptation method is executed. The parameter values ​​reflect the business priorities in specific application scenarios. To ensure the correctness of the optimization direction, the values ​​of the total number of palletized materials, energy consumption, and filling uniformity in the objective function need to be normalized before applying these weighting factors. Specifically, a set of random scheduling schemes is initially evaluated to estimate the maximum and minimum values ​​of energy consumption and filling uniformity differences. Then, the evaluation value of each scheme is scaled proportionally to a uniform range, for example, 0 to 1. After normalization, weights can be set within the range of 0 to 1 according to business priorities. For example, the weight of α can be increased in scenarios that prioritize production efficiency, while the weight of β can be adjusted accordingly in scenarios that prioritize low-cost operation.

[0088] The six-axis palletizing robot adaptation method based on virtual sampling reduces the trial-and-error costs in real-world environments. It comprehensively considers robot kinematics, trajectory planning, and multi-objective palletizing task scheduling optimization to achieve comprehensive robot adaptation. The palletizing task scheduling optimization objective function maximizes the total number of materials palletized, ensures balanced container filling at different palletizing points, and minimizes energy consumption during palletizing. Balanced container filling ensures that containers fill at similar times, enabling one-time replacement of all containers, improving efficiency and reducing repetitive operations. A violation function ensures the feasibility of the scheduling scheme. Finally, fine-tuning in a real-world environment ensures the applicability of the adaptation results.

[0089] When the candidate scheduling scheme is superior to the current scheduling scheme, the candidate scheduling scheme replaces the current scheduling scheme, and t is set to... m =0; when the current scheduling scheme is better than the candidate scheduling scheme, t m =t m +1; Update the first scheduling scheme set and the evaluation counter count = count + 1;

[0090] By traversing all scheduling schemes in the first scheduling scheme set, we obtain the second scheduling scheme set F′={x′ 1 ,…,x′ m ,…,x′ FN}; For the current scheduling scheme x′ m ∈F′, continue using the optimized counter t from the second stage. m Calculate the selection probability p of all scheduling schemes in the second scheduling scheme set. m The scheduling scheme with the highest probability will be selected as the intermediate scheduling scheme r. m The probability calculation formula is as follows:

[0091]

[0092] Where FN represents the total number of scheduling schemes in the set of scheduling schemes;

[0093] Based on the intermediate scheduling scheme and the current scheduling scheme, a candidate scheduling scheme v is generated. m The current scheduling scheme and the candidate scheduling scheme are compared based on the objective function and the violation function.

[0094] When the candidate scheduling scheme is superior to the current scheduling scheme, the candidate scheduling scheme replaces the current scheduling scheme, and t m =0; when the current scheduling scheme is better than the candidate scheduling scheme, t m =t m+1; Update the second scheduling scheme set and the evaluation counter count = count + 1;

[0095] By traversing all scheduling schemes in the second scheduling scheme set, we obtain the third scheduling scheme set F″={x″. 1 ,…,x″ m ,…,x″ FN};

[0096] Before proceeding to the next iteration, a stagnation check is performed on the third set of scheduling schemes. The scheduling schemes are monitored, and if a scheme meets a preset stagnation condition, it is replaced with a new random scheduling scheme, and the above process is repeated until a preset algorithm termination condition is reached. Specifically, the preset stagnation condition is as follows: Each scheduling scheme in the set is traversed, and its corresponding optimization counter is checked. When the optimization counter value of a scheduling scheme exceeds a preset stagnation threshold (e.g., 50 times), the scheme is considered to have stagnated and needs to be replaced by a newly generated random scheduling scheme. Simultaneously, its optimization counter is reset to 0.

[0097] After completing the stall check, the value of the evaluation counter is obtained. When the value of the evaluation counter is less than the preset value, the algorithm returns to the second stage, replaces the first scheduling scheme set with the third scheduling scheme set, and starts the next iteration. When the value of the evaluation counter reaches the maximum number of iterations, the algorithm terminates and outputs the current optimal scheduling scheme.

[0098] In the process of optimizing robot action sequence scheduling, the rules for comparing the current scheduling scheme and the candidate scheduling schemes include: if the violation function values ​​of both the current scheduling scheme and the candidate scheduling scheme are 0, the scheduling scheme with the better objective function is selected; if one violation function value is 0 while the other violation function value is not 0, the scheduling scheme with the violation function value of 0 is selected; if the violation function values ​​of the two scheduling schemes are not 0 and are not equal, the solution with the smaller violation function value is selected; if the violation function values ​​of the two scheduling schemes are not 0 and are equal, the scheduling scheme with the better objective function is selected.

[0099] In this embodiment, the set of pickup points is O = {o1, o2, o3}, and the corresponding set of palletizing points is P = {p1, p2, p3}. There is a one-to-one correspondence between pickup points and palletizing points. Therefore, the palletizing robot has four possible next actions: 1. Place the material selected from pickup point o1 to palletizing point p1; 2. Place the material selected from pickup point o2 to palletizing point p2; 3. Place the material selected from pickup point o3 to palletizing point p3; 4. Wait and do not perform any operation. The value of each element in the scheduling scheme...

[0100] Multi-stage iterative optimization of the scheduling scheme improves the comprehensiveness of the search; using objective function and violation function to evaluate and select the scheduling scheme can improve efficiency and resource utilization while ensuring feasibility; combining random search and probabilistic selection balances global exploration and local optimization; evaluation counter and optimization counter determine whether a new scheduling scheme needs to be introduced, avoiding getting trapped in local optima and increasing the possibility of finding the global optimum.

[0101] Example 2

[0102] See Figure 2 A virtual sampling six-axis palletizing robot adaptation system includes:

[0103] The data acquisition module is used to acquire parameter information and actual working environment information of the robot to be adapted;

[0104] The kinematics analysis module is used to perform kinematic analysis on the robot based on parameter information to obtain the workspace boundary of the robot's end effector;

[0105] The virtual sampling module is used to determine the target workspace based on the actual working environment information and workspace boundaries; based on the target workspace, it generates a set of material pick-up points and a set of material placement points through virtual sampling, and generates a set of path samples for the robot during the palletizing process.

[0106] The trajectory planning module is used to plan trajectories for path samples in the path sample set and calculate the optimal motion parameters.

[0107] The scheduling optimization module is used to construct objective functions and violation functions to optimize the robot action sequence scheduling of the pick-up point set and the placement point set, and obtain the optimal scheduling scheme.

[0108] The robot adaptation module integrates the optimal motion parameters and the optimal scheduling scheme to generate a control program, verifies the performance of the control program to generate verification results, and makes fine adjustments based on the verification results to complete the final adaptation.

[0109] Furthermore, the kinematic analysis module obtains the geometric parameters of each link of the robot and the limit range of each joint angle based on the parameter information; it describes the spatial position of the robot using a Cartesian coordinate system and establishes a base coordinate system; under the base coordinate system, it defines a local coordinate system for each link of the robot; based on the local coordinate system, it obtains DH parameters, which are used to describe the relative position and attitude between two adjacent links; based on the DH parameters, it establishes the kinematic equations between two adjacent local coordinate systems to obtain the pose matrix; based on the pose matrix and the limit range of each joint angle of the robot, it determines the boundary of the workspace using the extreme limit step angle method.

[0110] Furthermore, the trajectory planning module includes a first trajectory planning submodule and a second trajectory planning submodule; the second trajectory planning submodule includes a first trajectory planning unit and a second trajectory planning unit; the first trajectory planning submodule is used to plan the trajectory of a point-to-point path; the second trajectory planning submodule is used to plan the trajectory of a continuous motion path; the first trajectory planning unit is used to plan the trajectory of a straight path; and the second trajectory planning unit is used to plan the trajectory of a circular path.

[0111] Furthermore, the point-to-point path specifically includes: obtaining a set of key points in the point-to-point path, using a genetic algorithm to optimize and determine the optimal time interval sequence between key points; designing a first preset pose matrix for the robot at the key points; performing inverse kinematics on the first preset pose matrix to obtain the first preset joint angles of each joint of the robot at the key points; and performing cubic spline curve interpolation based on the first preset joint angles to generate the joint angles, angular velocities, and angular acceleration motion trajectories that constitute the core motion commands.

[0112] Furthermore, the straight path specifically includes: obtaining the coordinates of the starting point and ending point of the straight path; using a straight interpolation algorithm with acceleration and deceleration control to insert straight interpolation points at equal time intervals on the straight path; designing a second preset pose matrix for the robot at the straight interpolation points; performing inverse kinematics on the second preset pose matrix to obtain a second preset joint angle of the robot joint; and performing cubic spline curve interpolation based on the second preset joint angle to generate the joint angle, angular velocity, and angular acceleration motion trajectory that constitutes the core motion command.

[0113] Furthermore, the circular path specifically includes: obtaining the coordinates of the starting point, ending point, and intermediate point of the circular path; determining and calculating the center and radius of the circular path; establishing a circular coordinate system based on the center and the plane where the circular path is located; using a circular interpolation algorithm to calculate the circular interpolation point in the circular coordinate system based on the center and radius; transforming the circular interpolation point from the circular coordinate system to the base coordinate system; designing the third preset pose matrix of the robot at the circular interpolation point; performing inverse kinematics on the third preset pose matrix to obtain the third preset joint angle of each joint of the robot; and performing cubic spline curve interpolation based on the third preset joint angle to generate the joint angle, angular velocity, and angular acceleration motion trajectory that constitutes the core motion command.

[0114] Furthermore, the scheduling optimization module includes an initialization submodule and an optimization submodule; the initialization submodule is used to construct an objective function and a violation function for optimizing the robot action sequence scheduling; the objective function is used to maximize the total number of palletized materials, minimize the energy consumed by the palletized materials, and maximize the uniformity of container filling at each palletizing point; the violation function is used to ensure the feasibility of the scheduling scheme; the optimization submodule is used to optimize the robot action sequence scheduling.

[0115] Furthermore, the optimization submodule specifically includes: randomly generating a first set of scheduling schemes; randomly selecting an intermediate scheduling scheme based on the first set of scheduling schemes; generating a first candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; traversing the first set of scheduling schemes to obtain a second set of scheduling schemes; calculating the selection probability of the scheduling schemes in the second set of scheduling schemes, and selecting the scheduling scheme with the highest selection probability as the intermediate scheduling scheme; generating a second candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; performing a merit-based evaluation on the first candidate scheduling scheme and the second candidate scheduling scheme according to the objective function and the violation function, and deciding whether to update the set of scheduling schemes based on the evaluation results; if the set of scheduling schemes is updated, monitoring and replacing the scheduling schemes whose optimization effect has stalled with newly generated random scheduling schemes and re-executing the above process until the preset algorithm termination condition is reached.

[0116] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for adapting a six-axis palletizing robot based on virtual sampling, characterized in that, include: Obtain the parameter information and actual working environment information of the robot to be adapted; Based on the parameter information, kinematic analysis of the robot is performed to obtain the workspace boundary of the robot's end effector. Determine the target workspace based on actual work environment information and workspace boundaries; Based on the target workspace, a set of material pick-up points and a set of material placement points are generated through virtual sampling, and a set of path samples for the robot during the palletizing process is generated. Trajectory planning is performed on path samples in the path sample set, and the optimal motion parameters are calculated and optimized. The path samples include point-to-point paths, straight paths, and circular paths. The optimal motion parameters constitute the core motion commands of the robot control program. By constructing objective functions and violation functions, the robot action sequence scheduling is optimized using the pick-up point set and the placement point set to obtain the optimal scheduling scheme; The objective function is used to maximize the total number of palletized materials, minimize the energy consumed by the palletized materials, and maximize the uniformity of container filling at each palletizing point; the violation function is used to ensure the feasibility of the scheduling scheme. The optimal motion parameters and optimal scheduling scheme are integrated to generate a control program. The performance of the control program is verified to generate verification results. Based on the verification results, fine-tuning is performed to complete the final adaptation.

2. The adaptation method for a six-axis palletizing robot based on virtual sampling according to claim 1, characterized in that, The kinematic analysis includes: obtaining the geometric parameters and joint limit ranges of each link of the robot based on parameter information; establishing a robot base coordinate system and defining local coordinate systems for the robot links using the base coordinate system; obtaining DH parameters based on the local coordinate systems, wherein the DH parameters are used to describe the relative position and orientation between two adjacent links; establishing the kinematic equations of two adjacent local coordinate systems based on the DH parameters to obtain the pose matrix; and determining the workspace boundary using the limit step angle method based on the pose matrix and the limit ranges of the robot joints.

3. The adaptation method for a six-axis palletizing robot based on virtual sampling according to claim 1, characterized in that, The point-to-point path specifically includes: obtaining a set of key points in the point-to-point path; using a genetic algorithm to optimize and determine the optimal time interval sequence between key points; designing a first preset pose matrix for the robot at the key points; performing inverse kinematics on the first preset pose matrix to obtain the first preset joint angles of each joint of the robot at the key points; and performing cubic spline curve interpolation based on the first preset joint angles to generate the joint angles, angular velocities, and angular acceleration motion trajectories that constitute the core motion commands.

4. The adaptation method for a six-axis palletizing robot based on virtual sampling according to claim 1, characterized in that, The straight path specifically includes: obtaining the coordinates of the starting point and ending point of the straight path; using a straight interpolation algorithm with acceleration and deceleration control to insert straight interpolation points at equal time intervals on the straight path; designing a second preset pose matrix for the robot at the straight interpolation points; performing inverse kinematics on the second preset pose matrix to obtain a second preset joint angle of the robot joint; and performing cubic spline curve interpolation based on the second preset joint angle to generate the joint angle, angular velocity, and angular acceleration motion trajectory that constitutes the core motion command.

5. The adaptation method for a six-axis palletizing robot based on virtual sampling according to claim 1, characterized in that, The circular path specifically includes: obtaining the coordinates of the starting point, ending point, and intermediate point of the circular path; determining and calculating the center and radius of the circular path; establishing a circular coordinate system based on the center and the plane where the circular path is located; using a circular interpolation algorithm to calculate the circular interpolation point in the circular coordinate system based on the center and radius; transforming the circular interpolation point from the circular coordinate system to the base coordinate system; designing the third preset pose matrix of the robot at the circular interpolation point; performing inverse kinematics on the third preset pose matrix to obtain the third preset joint angle of each joint of the robot; and performing cubic spline curve interpolation based on the third preset joint angle to generate the joint angle, angular velocity, and angular acceleration motion trajectory that constitutes the core motion command.

6. The adaptation method for a six-axis palletizing robot based on virtual sampling according to claim 3, characterized in that, The optimization process of the genetic algorithm includes constraints, including velocity constraints, acceleration constraints, and impact constraints; a cubic spline function is constructed based on the time interval sequence to obtain a cubic spline curve; the cubic spline curve is analyzed to obtain the angular velocity, angular acceleration, and angular jerk at any time for each joint; the time interval sequence is retained when the maximum angular velocity, maximum angular acceleration, and maximum angular jerk satisfy all constraints.

7. The adaptation method for a six-axis palletizing robot based on virtual sampling according to claim 1, characterized in that, The robot action sequence scheduling optimization specifically includes: randomly generating a first scheduling scheme set; randomly selecting an intermediate scheduling scheme based on the first scheduling scheme set; generating a first candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; traversing the first scheduling scheme set to obtain a second scheduling scheme set; calculating the selection probability of the scheduling schemes in the second scheduling scheme set, and selecting the scheduling scheme with the highest selection probability as the intermediate scheduling scheme; generating a second candidate scheduling scheme by combining the intermediate scheduling scheme and the current scheduling scheme; evaluating the first candidate scheduling scheme and the second candidate scheduling scheme according to the objective function and the violation function, and deciding whether to update the scheduling scheme set based on the evaluation results; monitoring the scheduling schemes, and when a scheduling scheme meets a preset stagnation condition, replacing the scheduling scheme with a new random scheduling scheme and re-executing the above process until the preset algorithm termination condition is reached.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the virtual sampling six-axis palletizing robot adaptation method as described in any one of claims 1-7.

9. A virtual sampling six-axis palletizing robot adaptation system, characterized in that, include: The data acquisition module is used to acquire parameter information and actual working environment information of the robot to be adapted; The kinematics analysis module is used to perform kinematic analysis on the robot based on parameter information to obtain the workspace boundary of the robot's end effector; The virtual sampling module is used to determine the target workspace based on actual working environment information and workspace boundaries; Based on the target workspace, a set of material pick-up points and a set of material placement points are generated through virtual sampling, and a set of path samples for the robot during the palletizing process is generated. The trajectory planning module is used to plan trajectories for path samples in the path sample set and calculate the optimal motion parameters. The scheduling optimization module is used to construct objective functions and violation functions to optimize the robot action sequence scheduling of the pick-up point set and the placement point set, and obtain the optimal scheduling scheme. The robot adaptation module integrates the optimal motion parameters and the optimal scheduling scheme to generate a control program, verifies the performance of the control program to generate verification results, and makes fine adjustments based on the verification results to complete the final adaptation.

Images