Assembly line feeding and discharging robot track design method and system and robot

Abstract

The invention discloses an assembly line feeding and discharging robot trajectory design method and system and a robot, and the method comprises the steps: designing a periodic Fourier series mathematical trajectory model, building a kinetic model, carrying out the global search through employing a simulated annealing method, carrying out the local constraint optimization through mixing with an Active-Set algorithm in fmincon, and carrying out the conversion of an optimized trajectory matrix, the dynamic parameters obtained through identification are input into a dynamic algorithm in the assembly line feeding and discharging robot, the actual motion range of six axes of the assembly line feeding and discharging robot is limited, and the optimal trajectory of the assembly line feeding and discharging robot is obtained through global optimization and local optimization algorithms. A hybrid simulated annealing algorithm and a five-level Fourier series mathematical trajectory model scheme are combined, so that the trajectory has higher reliability and representativeness, and the precision of grabbing and placing actions of the robot in the operation process and the stability of the robot in the carrying process are ensured.

Description

Technical Field

[0001] The present invention relates to the field of robot technology, and in particular to a method and system for designing a trajectory of a loading and unloading robot on an assembly line, and the robot. Background Art

[0002] With the development of science and technology, robots are gradually replacing manual labor in actual industrial production, especially in loading and unloading materials at fixed positions on assembly lines, where robots need to frequently grab and carry goods of different weights and sizes. Through the optimized excitation trajectory design, the robot can accurately identify its own dynamic parameters, and can flexibly respond to light goods or heavy pallets in a larger workspace, ensuring the accuracy of the grabbing action and the stability of the handling process.

[0003] One of the keys to the accuracy of robot dynamic parameter identification lies in the design of the excitation trajectory. The existing algorithms for the design of the excitation trajectory are relatively simple and cannot effectively improve the accuracy of dynamic parameter identification, and cannot traverse all the workspaces of the robot. With the continuous development of related technologies such as robots and artificial intelligence, it is urgent to optimize the excitation trajectory and introduce the robot through a hybrid algorithm to improve the design accuracy.

[0004] Existing excitation trajectory design methods mostly rely on simple mathematical models, such as common polynomial trajectories or basic Fourier series trajectories. Although these methods can achieve basic excitation trajectory planning, they have obvious deficiencies in accuracy and generalization. Especially in complex industrial environments, robots need to traverse a wider range of workspaces. Existing trajectory design methods often cannot meet the high-precision requirements of robots for multi-tasking and long-term operation, resulting in inaccurate dynamic identification and instability in the drag-teaching grasping process. Summary of the invention

[0005] The embodiments of the present application provide a method, system and robot for designing the trajectory of an assembly line loading and unloading robot, which solves the technical problem in the prior art that when an industrial assembly line loading and unloading robot frequently grabs objects of different weights and sizes, the running trajectory has deviations due to inaccurate dynamic parameter settings, resulting in the inability to accurately complete the grabbing and placing actions of different tasks. The present application uses global optimization and local optimization algorithms, combined with a hybrid simulated annealing algorithm and a five-level Fourier series mathematical trajectory model scheme, to make the trajectory more reliable and representative, ensuring the accuracy of the robot's grabbing and placing actions during operation and the stability of the handling process.

[0006] The present application embodiment provides a method for designing a trajectory of a loading and unloading robot on an assembly line, comprising:

[0007] Step S1, design a periodic Fourier series mathematical trajectory model, establish a dynamic model, initialize the Fourier series and coefficients, and import the coefficient matrix of the loading and unloading robot on the assembly line;

[0008] Step S2, global optimization of trajectory matrix: using simulated annealing method to perform global search, optimize the objective function, generate a globally optimized trajectory matrix, regularize the trajectory matrix and normalize the trajectory matrix, calculate and minimize the condition number of the global optimization matrix;

[0009] Step S3, local optimization of trajectory matrix: the Active-Set algorithm in the hybrid fmincon is used to perform local constraint optimization, find the optimal Fourier coefficient combination, obtain the locally optimized trajectory matrix, calculate the condition number of the local optimization matrix and minimize it;

[0010] Step S4, converting the optimized trajectory matrix to obtain a stable optimal trajectory, using the trajectory model obtained in step S1 to identify the dynamic parameters, reconstructing the dynamic parameters, recalculating the torque value, and comparing it with the estimated value; if the error is large, return to step S2 and step S3 for re-optimization; if it is accurate, the dynamic parameters can be used in subsequent drag teaching;

[0011] Step S5, inputting the identified dynamic parameters into the dynamic algorithm of the loading and unloading robot on the assembly line, and limiting the actual motion range of the six axes of the loading and unloading robot on the assembly line.

[0012] Preferably, the Fourier series mathematical trajectory model in step S1 is:

[0013]

[0014] in, Represents the amplitude, and also represents the parameter to be optimized by Fourier, ω f represents the fundamental frequency of the Fourier series, q i,0 Indicates the starting point of the joint excitation trajectory, in degrees.

[0015] Preferably, in step S1, the Fourier series and coefficients are initialized, and the Fourier series is set to five levels.

[0016] Preferably, the global optimization of the trajectory matrix in step S2 is specifically as follows: setting the optimization options of the simulated annealing algorithm in MATLAB: setting the number of iterations, the convergence tolerance of the objective function, and the random number, calculating the optimized random number and the six trajectory trajectories of the six axes of the robot and the speed and acceleration of the trajectory, and generating a globally optimized trajectory matrix.

[0017] Preferably, the condition number of the global optimization matrix in step S2 is 6.00-7.00.

[0018] Preferably, the local optimization of the trajectory matrix in step S3 is specifically as follows:

[0019] When the robot starts and stops, the joint angle, angular velocity and angular acceleration of the robot are set to zero, so that the robot can run continuously when starting and stopping. The constraint conditions of the parameters to be optimized are:

[0020]

[0021] Among them, t0 represents the starting time of a single trajectory cycle, t T Indicates the end time of a single trajectory. When the robot starts and ends the movement, the joint angle, angular velocity and acceleration are zero;

[0022] Inequality constraints:

[0023] q min ≤q i (t)≤q max

[0024] Among them, q min ,q max They respectively represent the lower and upper limits of the joint angles of this model of robot.

[0025] Preferably, the condition number of the local optimization matrix in step S3 is 5.00-6.00.

[0026] This application also proposes a trajectory design system for a robot for loading and unloading materials on an assembly line, comprising:

[0027] Model building module, designs the periodic Fourier series mathematical trajectory model, establishes the dynamic model, initializes the Fourier series and coefficients, and imports the coefficient matrix of the loading and unloading robots on the assembly line;

[0028] Global optimization module, used for global optimization of trajectory matrix: using simulated annealing method to perform global search, optimize the objective function, generate the trajectory matrix after global optimization, regularize the trajectory matrix and normalize the trajectory matrix, calculate the condition number of the global optimization matrix and minimize it;

[0029] Local optimization module, used for local optimization of trajectory matrix: the Active-Set algorithm in hybrid fmincon is used for local constraint optimization to find the optimal combination of Fourier coefficients, obtain the trajectory matrix after local optimization, and calculate and minimize the condition number of the local optimization matrix;

[0030] The error calibration module is used to convert the optimized trajectory matrix to obtain a stable optimal trajectory, identify the dynamic parameters using the trajectory model, reconstruct the dynamic parameters, recalculate the torque value, and compare it with the estimated value; if the error is large, re-optimize; if it is accurate, the dynamic parameters can be used in subsequent drag teaching;

[0031] The trajectory generation module is used to input the identified dynamic parameters into the dynamic algorithm of the assembly line loading and unloading robot and limit the actual motion range of the six axes of the assembly line loading and unloading robot.

[0032] The present application also proposes an assembly line loading and unloading robot, including the above-mentioned assembly line loading and unloading robot trajectory design system.

[0033] The solution provided in the embodiments of the present application has the following technical effects or advantages:

[0034] 1. This application uses a hybrid simulated annealing algorithm to optimize the drag teaching excitation trajectory design and parameter identification system of the robot arm, which improves the accuracy and efficiency of the robot arm's grasping tasks, especially in complex industrial scenarios such as assembly lines, automated assembly lines, etc. It allows operators to perform tasks without programming.

[0035] 2. This application adopts a combination of global optimization and local optimization. Please use matlab and python to mix simulated annealing with the Active-Set algorithm in fmincon. Global optimization avoids the dilemma of local optimal solution, while local constraint optimization ensures the continuity and stability of the trajectory. Through multiple consecutive iterations, the condition number is lower, the recognition accuracy is higher, and the drag teaching is more accurate, so that the robot is more stable and has lower errors in the process of grabbing materials on the assembly line, so that the robot can complete tasks more accurately in complex industrial environments.

[0036] 3. In this application, the design of the excitation trajectory adopts the Fourier series model, which can effectively express complex motion trajectories, allowing the robot arm to move flexibly in different workspaces. The Fourier series model is widely used in excitation trajectory optimization because it can decompose the trajectory into multiple frequency components, effectively handle nonlinear motion trajectories, and improve the flexibility and accuracy of path planning. Compared with the traditional simple trajectory model, the Fourier series shows higher accuracy and efficiency in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a flow chart of the trajectory design method in Example 1 of the present application;

[0038] Figure 2This is a global optimization flow chart for trajectory design in Example 1 of the present application;

[0039] Figure 3 This is a module diagram of the trajectory design system in the second embodiment of the present application;

[0040] Figure 4 This is a structural diagram of the loading and unloading robot on the assembly line in Example 3 of the present application. DETAILED DESCRIPTION

[0041] The embodiment of the present application solves the technical problem of unstable operation of industrial robots in grasping and handling operations in the prior art by providing a method and system for designing the trajectory of a loading and unloading robot on an assembly line. A simulated annealing algorithm is applied in combination with a periodic trajectory model of Fourier series, and a method combining global optimization with local constraints is adopted to enhance the dynamic identification accuracy of the robot, thereby achieving a high reliability of the robot's operation trajectory in multi-task and long-term assembly line operations, thereby ensuring high-precision and high-stability operation of the industrial robot.

[0042] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0043] Embodiment 1

[0044] like Figure 1 As shown, a method for designing a trajectory of a loading and unloading robot on an assembly line comprises:

[0045] Step S1, design a periodic Fourier series mathematical trajectory model and establish a dynamic model.

[0046]

[0047] in, Represents the amplitude, and also represents the parameter to be optimized by Fourier; ω f represents the fundamental frequency of the Fourier series; q i,0 Indicates the starting point of the joint excitation trajectory, in degrees; N represents the Fourier series, which controls the complexity of the Fourier series, where N is 5; k represents the frequency order in the Fourier series, and k starts from 1. The initial and final positions, velocities, and accelerations are all 0.

[0048] Establish a dynamic model, use the Sympybotics tool in Python, input the dh value of the assembly line loading and unloading robot, obtain the coefficient matrix and minimum parameter set data of the assembly line loading and unloading robot and save them, and express the Fourier function mathematical trajectory model in the form of code through Matlab.

[0049] Initialize the Fourier series and coefficients, and import the coefficient matrix of the loading and unloading robot on the assembly line. Set the Fourier series to five levels, the total time T to 12 seconds, the angular frequency to 2*pi / T, the step size to 0.006 seconds, and copy the coefficient matrix to MATLAB after modifying its form.

[0050] Step S2, such as Figure 2 As shown, the simulated annealing method is used to perform global search, optimize the objective function, calculate the optimized random number x1 and the six trajectory trajectories of the six axes of the robot and the speed and acceleration of the trajectory, and generate a trajectory matrix.

[0051] Set the optimization options of the simulated annealing algorithm in MATLAB. The default starting temperature T of the simulated annealing is 100. When a new solution is generated, the difference in the objective function Δf=f is calculated. new -f old , if it is negative, then the acceptance probability is calculated using the following formula.

[0052]

[0053] if is negative, indicating that the new solution is better, then the new solution is accepted; if If it is positive, it means the new solution is worse, and the probability P is used to decide whether to accept it.

[0054] Set 5000 iterations and the convergence tolerance of the objective function. -6 Stop the iteration when , and set 66 random numbers x. Associated with the random term x, the optimized random number x1 and the six trajectory trajectories of the six axes of the robot and the velocity and acceleration of the trajectory are calculated through the sine and cosine terms of the Fourier series, and the sum of these trajectories is stored in the F matrix.

[0055] Regularize the matrix and normalize it, calculate the condition number of the matrix and minimize it. And observe whether the condition number is greater than 100. If it is greater than 100, return to step S4. The regularization term is added to the matrix by F+lambda*eye(size(F)), where lambda is a regularization parameter. The normalization operation ensures that the trajectory vector of each column has unit length, and the final condition number is around 6.00-7.00.

[0056] Step S3, the Active-Set algorithm in the hybrid fmincon is used to perform local constraint optimization, combined with the data obtained in step S2, and the optimal combination of Fourier coefficients is found to obtain the optimized trajectory matrix, and the condition number of the matrix is ​​minimized.

[0057] According to x1 in step S2, the Active-Set algorithm in fmincon is used to perform local constraint optimization, where the joint angle, angular velocity and angular acceleration of the robot are set to zero when the robot starts and stops, so that the robot can run continuously when starting and stopping. The constraint conditions converted into the parameters to be optimized are:

[0058]

[0059] Among them, t0 represents the starting time of a single trajectory cycle, t T Indicates the end moment of a single trajectory. When the robot starts and ends the movement, the joint angle, angular velocity and acceleration are zero.

[0060] Inequality constraints:

[0061] q min ≤q i (t)≤q max

[0062] Among them, q min ,q max Respectively represent the lower and upper limits of each joint angle of this model of robot

[0063] Repeat the steps in step 5 to obtain the optimized x1, and finally the condition number is in the range of 5.00-6.00.

[0064] Step S4, convert the generated trajectory matrix, intercept the rotation angle of the six axes every 0.006 seconds, program and run in CoppeliaSim simulation. After obtaining the stable optimal trajectory, use the least squares method to process the data generated in the process of generating the excitation trajectory in MATLAB, including speed, acceleration, torque, etc., use the coefficient matrix obtained in S1 to obtain the dynamic parameters, and reconstruct the torque of the identified parameters, recalculate the torque value, and compare it with the estimated value. If the error is large, return to steps S2 and S3 for re-optimization. If it is accurate, the identified dynamic parameters can be used in subsequent drag teaching.

[0065] The angle, angular velocity, angular acceleration, and torque data collected after running the excitation trajectory are used as system inputs into the robot dynamics parameter identification model of this system:

[0066] τ=YP

[0067] In the formula, τ represents the torque, Y represents the coefficient matrix, and P represents the set of dynamic parameters. The dynamic parameters are obtained by the least squares method, and the torque is reconstructed using the model. The torque value is compared with the actual value to verify whether the result is accurate.

[0068] Step S5, input the identified dynamic parameters into the dynamic algorithm of the robot for loading and unloading materials on the assembly line, and limit the actual motion range of the robot's six axes to achieve stable and accurate drag-teaching grasping.

[0069] Embodiment 2

[0070] like Figure 4 As shown, the present application also proposes a trajectory design system for a loading and unloading robot on an assembly line, comprising: a model building module 01, a global optimization module 02, a local optimization module 03, an error calibration module 04, and a trajectory generation module 05;

[0071] Model building module, designs the periodic Fourier series mathematical trajectory model, establishes the dynamic model, initializes the Fourier series and coefficients, and imports the coefficient matrix of the loading and unloading robots on the assembly line;

[0072] Global optimization module, used for global optimization of trajectory matrix: using simulated annealing method to perform global search, optimize the objective function, generate the trajectory matrix after global optimization, regularize the trajectory matrix and normalize the trajectory matrix, calculate the condition number of the global optimization matrix and minimize it;

[0073] Local optimization module, used for local optimization of trajectory matrix: the Active-Set algorithm in hybrid fmincon is used for local constraint optimization to find the optimal combination of Fourier coefficients, obtain the trajectory matrix after local optimization, and calculate and minimize the condition number of the local optimization matrix;

[0074] The error calibration module is used to convert the optimized trajectory matrix to obtain a stable optimal trajectory, identify the dynamic parameters using the trajectory model, reconstruct the dynamic parameters, recalculate the torque value, and compare it with the estimated value; if the error is large, re-optimize; if it is accurate, the dynamic parameters can be used in subsequent drag teaching;

[0075] The trajectory generation module is used to input the identified dynamic parameters into the dynamic algorithm of the assembly line loading and unloading robot and limit the actual motion range of the six axes of the assembly line loading and unloading robot.

[0076] Embodiment 3

[0077] like Figure 4 As shown, the present application also proposes an assembly line loading and unloading robot, including the above-mentioned assembly line loading and unloading robot trajectory design system.

[0078] The embodiments of this specific implementation method are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

[0079] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0080] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for designing the trajectory of a loading and unloading robot on an assembly line, characterized in that: include: Step S1, design a periodic Fourier series mathematical trajectory model, establish a dynamic model, initialize the Fourier series and coefficients, and import the coefficient matrix of the loading and unloading robot on the assembly line; Step S2, global optimization of trajectory matrix: using simulated annealing method to perform global search, optimize the objective function, generate a globally optimized trajectory matrix, regularize the trajectory matrix and normalize the trajectory matrix, calculate and minimize the condition number of the global optimization matrix; Step S3, local optimization of trajectory matrix: the Active-Set algorithm in the hybrid fmincon is used to perform local constraint optimization, find the optimal Fourier coefficient combination, obtain the locally optimized trajectory matrix, calculate the condition number of the local optimization matrix and minimize it; Step S4, converting the optimized trajectory matrix to obtain a stable optimal trajectory, using the trajectory model obtained in step S1 to identify the dynamic parameters, reconstructing the dynamic parameters, recalculating the torque value, and comparing it with the estimated value; if the error is large, return to step S2 and step S3 for re-optimization; if it is accurate, the dynamic parameters can be used in subsequent drag teaching; Step S5, inputting the identified dynamic parameters into the dynamic algorithm of the loading and unloading robot on the assembly line, and limiting the actual motion range of the six axes of the loading and unloading robot on the assembly line.

2. The robot trajectory design method according to claim 1, characterized in that: The Fourier series mathematical trajectory model in step S1 is: in, Represents the amplitude, and also represents the parameter to be optimized by Fourier, ω f represents the fundamental frequency of the Fourier series, q i,0 Indicates the starting point of the joint excitation trajectory, in degrees.

3. The robot trajectory design method according to claim 1, characterized in that: In step S1, the Fourier series and coefficients are initialized, and the Fourier series is set to five levels.

4. The robot trajectory design method according to claim 1, characterized in that: The global optimization of the trajectory matrix in step S2 is specifically as follows: setting the optimization options of the simulated annealing algorithm in MATLAB: setting the number of iterations, the convergence tolerance of the objective function, and the random number, calculating the optimized random number and the six trajectory trajectories of the six axes of the robot and the speed and acceleration of the trajectory, and generating a globally optimized trajectory matrix.

5. The robot trajectory design method according to claim 1, characterized in that: The condition number of the global optimization matrix in step S2 is 6.00-7.

00.

6. The robot trajectory design method according to claim 1, characterized in that: The local optimization of the trajectory matrix in step S3 is specifically as follows: When the robot starts and stops, the joint angle, angular velocity and angular acceleration of the robot are set to zero, so that the robot can run continuously when starting and stopping. The constraint conditions of the parameters to be optimized are: Among them, t0 represents the starting time of a single trajectory cycle, t T Indicates the end time of a single trajectory. When the robot starts and ends the movement, the joint angle, angular velocity and acceleration are zero; Inequality constraints: q min ≤q i (t)≤q max Among them, q min ,q max They respectively represent the lower and upper limits of the joint angles of this model of robot.

7. The robot trajectory design method according to claim 1, characterized in that: The condition number of the local optimization matrix in step S3 is 5.00-6.

00.

8. A robot trajectory design system for loading and unloading materials on an assembly line, characterized in that: include: Model building module, designs the periodic Fourier series mathematical trajectory model, establishes the dynamic model, initializes the Fourier series and coefficients, and imports the coefficient matrix of the loading and unloading robots on the assembly line; Global optimization module, used for global optimization of trajectory matrix: using simulated annealing method to perform global search, optimize the objective function, generate the trajectory matrix after global optimization, regularize the trajectory matrix and normalize the trajectory matrix, calculate the condition number of the global optimization matrix and minimize it; Local optimization module, used for local optimization of trajectory matrix: the Active-Set algorithm in hybrid fmincon is used for local constraint optimization to find the optimal combination of Fourier coefficients, obtain the trajectory matrix after local optimization, and calculate and minimize the condition number of the local optimization matrix; The error calibration module is used to convert the optimized trajectory matrix to obtain a stable optimal trajectory, identify the dynamic parameters using the trajectory model, reconstruct the dynamic parameters, recalculate the torque value, and compare it with the estimated value; if the error is large, re-optimize; if it is accurate, the dynamic parameters can be used in subsequent drag teaching; The trajectory generation module is used to input the identified dynamic parameters into the dynamic algorithm of the assembly line loading and unloading robot and limit the actual motion range of the six axes of the assembly line loading and unloading robot.

9. A robot for loading and unloading materials on an assembly line, characterized in that: Includes the trajectory design system described in claim 8.

Images