A robot dynamics parameter optimization method, device and equipment

Abstract

This invention discloses a method, apparatus, device, and computer-readable storage medium for optimizing robot dynamic parameters. The method includes: establishing a robot kinematic and dynamic model using robot link parameters; linearizing the established dynamic model and deriving a set of minimum parameters to be identified; using the condition number of the robot's minimum inertial parameter observation matrix as the objective function to design an excitation trajectory in Fourier series form, and optimizing the Fourier series coefficients using an improved differential evolution algorithm; driving the robot to move along the optimized excitation trajectory, and collecting data to identify the dynamic parameters using the least squares method. This invention utilizes an improved differential evolution algorithm to optimize the excitation trajectory parameters of the Fourier series, ensuring the robot can operate stably within its maximum safe space and improving the accuracy of subsequent dynamic parameter identification.

Description

Technical Field

[0001] This invention relates to the field of dynamic control optimization, and in particular to a method, apparatus, device, and computer-readable storage medium for optimizing robot dynamic parameters. Background Technology

[0002] With the continuous development of intelligent manufacturing, the application scenarios of industrial robots are becoming increasingly diverse. Consequently, the performance requirements for domestically produced robots are also rising, demanding not only faster operating speeds and higher trajectory accuracy, but also better control performance. An accurate robot dynamics model is the foundation for robot planning and control; the robot's motion accuracy, control performance, operational stability, and robustness are also determined by the robot dynamics model. While robot kinematic parameters can be obtained through calibration, robot dynamic parameters must be identified through specific identification experiments.

[0003] There are three main methods for obtaining robot dynamic parameters: disassembly measurement, computer-aided design (CAD), and overall identification. Disassembly measurement involves breaking down the robot into its individual links, measuring the geometric parameters of each link separately, and calculating the inertial parameters of the links according to the definition of rigid body inertial parameters based on the link material. However, this method is very complex for complex components, and accurate measurement of parameters such as the center of mass and link length is difficult to achieve. CAD methods use a 3D design model of the robot, with parameters automatically calculated by software or derived from theoretical formulas. This method does not consider joint factors, including the influence of joint friction and elasticity on dynamics, and the inertial parameter values ​​measured by CAD are ideal, which may introduce errors during actual manufacturing and assembly. Overall identification involves conducting identification experiments on an actual robot. Given an optimized trajectory for each joint, the torque and rotation parameters of each joint are measured during the robot's movement. The measured data are then fed into an identification model, and the constructed identification algorithm calculates the values ​​of the dynamic parameters. Holistic identification has significant advantages over the previous two identification methods. Because holistic identification can take into account the effects of various dynamic factors, it is currently the most common identification scheme for dynamic parameters. However, existing holistic identification methods rarely establish excitation trajectories, and even when they do, the excitation parameters are merely random parameters, leading to low accuracy in dynamic parameter identification and inaccurate torque prediction.

[0004] In summary, it can be seen that improving the accuracy of the overall identification dynamics identification parameters is a problem that needs to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide a method, apparatus, device, and computer-readable storage medium for optimizing robot dynamic parameters, which solves the problem that the existing technology does not consider various dynamic influencing factors, resulting in low accuracy of dynamic parameter identification.

[0006] To address the aforementioned technical problems, this invention provides a method for optimizing robot dynamic parameters, comprising:

[0007] A kinematic model of the robot is established using a robot link parameter table, and a dynamic model is constructed based on the kinematic model.

[0008] The dynamic model is linearized to obtain the minimum set of inertial parameters to be identified;

[0009] Parameters are extracted based on the minimum set of inertial parameters to be identified, and an observation matrix is ​​established.

[0010] The condition number of the observation matrix is ​​used as the objective function, and the excitation trajectory and constraints in the form of a Fourier series are constructed.

[0011] The optimal parameters of the excitation trajectory in Fourier series form are solved using an improved differential evolution algorithm.

[0012] Substituting the optimal parameters into the Fourier series form of the excitation trajectory yields the optimal excitation trajectory;

[0013] Based on the excitation trajectory, robot motion data is collected, and the robot's dynamic parameters are identified using the least squares algorithm.

[0014] Preferably, the step of collecting robot motion data based on the excitation trajectory and identifying robot dynamic parameters using a least squares algorithm includes:

[0015] Based on the identified robot dynamics parameters, a new optimal excitation trajectory is reconstructed.

[0016] Calculate the torque information of the new optimal excitation trajectory and the root mean square difference of the torque information of the optimal excitation trajectory;

[0017] If the root mean square error does not meet the preset value, the excitation trajectory and dynamic parameters are recalculated.

[0018] Preferably, establishing the robot dynamics model using the robot link parameter table includes:

[0019] Based on the robot's geometry, establish a link coordinate system diagram for each link;

[0020] The DH parameter table of the robot is obtained based on the link coordinate system diagram of each link;

[0021] Based on the robot's DH parameter table, a robot dynamics model is established using the Newton-Euler iterative method.

[0022] Where M(q)∈R n×n The inertial tensor matrix, Let G(q) ∈ R be the vector of Coriolis force and centrifugal force. n Let q be the gravity vector, q∈R n From the robot's perspective, For the robot's speed, For the robot's acceleration, τ f ∈R n This is the term for joint friction.

[0023] Preferably, the linearization of the dynamic model to obtain the minimum set of inertial parameters to be identified includes:

[0024] The state variables and inertial parameters in the dynamic model are separated to obtain the minimum set of inertial parameters P to be identified;

[0025] Its expression is:

[0026]

[0027] Where, m i Let x be the mass of the i-th joint. i y i , z i Let I be the position of the centroid of the i-th joint. xx，i I yy，i I zz，i I xy，i I xz，i I yz，i f is the value of the inertia tensor matrix of the i-th joint in the joint coordinate system. c，i f v，i These are the Coulomb friction coefficient and the viscosity model coefficient, respectively.

[0028] Preferably, the step of extracting parameters and establishing an observation matrix based on the minimum set of inertial parameters to be identified includes:

[0029] The minimum inertial parameter set to be identified is extracted to construct the observation matrix Y, which is expressed as follows:

[0030]

[0031] Where Y∈R n×10n It is a function that contains only joint angles, joint angular velocities, and joint angular accelerations, where m is the number of identifiable parameters and n is the number of joints.

[0032] Preferably, the step of using the condition number of the observation matrix as the objective function and constructing the excitation trajectory and constraints in the form of a Fourier series includes:

[0033] The condition number of the observation matrix is ​​taken as the objective function, and the expression of the objective function is:

[0034] F(Y) = cond(Y) = ||(Y) -1 ||·||Y||

[0035] min(F(Y));

[0036] The expression for the excitation trajectory in Fourier series form is:

[0037]

[0038] Among them, w f q is the fundamental frequency of the Fourier series, N is the harmonic number of the Fourier series, and q is the frequency of the Fourier series. i，0 a is a constant term. k，i b k，i These are the amplitude coefficients of the sine and cosine functions.

[0039] The constraints are as follows:

[0040] Where {s(q(t))} is the set of end-effector positions of the robot during movement, S represents the maximum workspace allowed for the robot to operate, and q max q min These represent the upper and lower limits of the angles of each joint. These represent the upper and lower limits of the velocity of each joint. These represent the upper and lower limits of the acceleration for each joint.

[0041] Preferably, the optimal parameters for solving the Fourier series form of the excitation trajectory using the improved differential evolution algorithm include:

[0042] S1: Initialize the population parameter N, and set the maximum number of iterations g. max ;

[0043] S2: Calculate the objective function value of parameter i in the parameter population, where i∈(1,2,3,…,N);

[0044] S3: Determine whether the objective function value of parameter i satisfies the constraint conditions;

[0045] S4: If the objective function value of parameter i does not satisfy the constraint condition, then set the objective function value of parameter i to 10. 10 If satisfied, proceed to S5;

[0046] S5: Determine the magnitude of the objective function value and the initial function value of parameter i;

[0047] S6: If the objective function value is greater than the initial function value, then the objective function value is used as the function value of parameter i;

[0048] S7: Mutate and crossover the parameters in the population to generate a new i-th parameter;

[0049] S8: Calculate the objective function value of the new parameter i, and determine whether the objective function value of the new parameter i is greater than the function value of parameter i;

[0050] S9: If the objective function value of the new parameter i is greater than the function value of the parameter i, then the parameter i is taken as the optimal parameter; if the objective function value of the new parameter i is less than the objective function value of the parameter, then the new parameter i is taken as the optimal parameter.

[0051] S10: Let i = i + 1, return to step S2, until i = g max Output the optimal parameters.

[0052] The present invention also provides a device for identifying robot dynamic parameters, comprising:

[0053] Model building module: used to build a robot kinematic model using a robot link parameter table, and to build a dynamic model based on the kinematic model;

[0054] The processing model module is used to linearize the dynamic model to obtain the minimum set of inertial parameters to be identified.

[0055] The parameter extraction module is used to extract parameters based on the minimum set of inertial parameters to be identified and to establish an observation matrix.

[0056] A target function module is constructed to use the condition number of the observation matrix as the target function, and to construct the excitation trajectory and constraints in the form of a Fourier series.

[0057] The optimization calculation module is used to solve for the optimal parameters of the excitation trajectory in the form of the Fourier series using an improved differential evolution algorithm;

[0058] The excitation trajectory calculation module is used to input the optimal parameters into the excitation trajectory in the form of the Fourier series to obtain the optimal excitation trajectory;

[0059] The parameter calculation module collects robot motion data based on the excitation trajectory and identifies the robot's dynamic parameters using the least squares algorithm.

[0060] The present invention also provides a device for identifying robot dynamic parameters, comprising:

[0061] A memory for storing computer programs; a processor for executing the computer programs to implement the steps of the above-described method for optimizing robot dynamic parameters.

[0062] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for optimizing robot dynamic parameters.

[0063] This invention provides a method for optimizing robot dynamic parameters. First, a robot dynamic model is established using the robot's link parameters. Then, the dynamic model is linearized to establish a minimum inertia parameter set. The parameters of the minimum inertia set are then extracted to obtain an observation matrix. The condition number of the observation matrix is ​​used as the objective function, and a Fourier series excitation trajectory is designed. An improved differential evolution algorithm is used to optimize the parameters of the Fourier series excitation trajectory to obtain optimal parameters. This ensures that the excitation trajectory allows the robot to operate stably within its maximum safe space, laying the groundwork for subsequent accurate dynamic parameter identification. Finally, the optimal parameters are substituted into the excitation trajectory to obtain the optimal excitation trajectory. Based on the optimal excitation trajectory, the robot's dynamic parameters are identified using a least squares algorithm. This invention utilizes an improved differential evolution algorithm to optimize the Fourier series excitation trajectory parameters, obtaining optimal excitation trajectory parameters that ensure the robot can operate stably within its maximum safe space, thus improving the accuracy of subsequent dynamic parameter identification. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 A flowchart of a first specific embodiment of the robot dynamics parameter optimization method provided by the present invention;

[0066] Figure 2 A flowchart of a second specific embodiment of the robot dynamics parameter optimization method provided by the present invention;

[0067] Figure 3 This is a link coordinate system diagram of a six-axis robot according to an embodiment of the present invention;

[0068] Figure 4 This is a comparison diagram of the optimization process of the improved differential evolution algorithm in an embodiment of the present invention;

[0069] Figure 5 This is an optimized excitation trajectory diagram according to an embodiment of the present invention;

[0070] Figure 6 This is an optimized excitation trajectory velocity diagram according to an embodiment of the present invention;

[0071] Figure 7 This is an optimized excitation trajectory acceleration diagram according to an embodiment of the present invention;

[0072] Figure 8 This is a comparison diagram of the predicted torque and the actual torque of robot joint 1 in an embodiment of the present invention;

[0073] Figure 9 This is a comparison diagram of the predicted torque and the actual torque of robot joint 2 in an embodiment of the present invention;

[0074] Figure 10 This is a comparison diagram of the predicted torque and the actual torque of robot joint 3 in an embodiment of the present invention;

[0075] Figure 11 This is a comparison diagram of the predicted torque and the actual torque of robot joint 4 in an embodiment of the present invention;

[0076] Figure 12 This is a comparison diagram of the predicted torque and the actual torque of robot joint 5 in an embodiment of the present invention;

[0077] Figure 13 This is a comparison diagram of the predicted torque and the actual torque of robot joint 6 in an embodiment of the present invention;

[0078] Figure 14 This is a structural block diagram of a device for optimizing robot dynamic parameters provided in an embodiment of the present invention. Detailed Implementation

[0079] The core of this invention is to provide a method, apparatus, device, and computer-readable storage medium for optimizing robot dynamic parameters. It utilizes an improved differential evolution algorithm to optimize the excitation trajectory parameters of the Fourier series, ensuring that the robot can operate stably within its maximum safe space and improving the accuracy of subsequent dynamic parameter identification.

[0080] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely 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.

[0081] Please refer to Figure 1 , Figure 1A flowchart of a first specific embodiment of the robot dynamics parameter optimization method provided by the present invention; the specific operation steps are as follows:

[0082] Step S101: Establish a robot kinematic model using the robot link parameter table, and construct a dynamic model based on the kinematic model;

[0083] Step S102: Linearize the dynamic model to obtain the minimum set of inertial parameters to be identified;

[0084] Step S103: Extract parameters based on the minimum inertial parameter set to be identified and establish an observation matrix;

[0085] Step S104: Use the condition number of the observation matrix as the objective function, and construct the excitation trajectory and constraint conditions in the form of a Fourier series;

[0086] Step S105: Solve for the optimal parameters of the excitation trajectory in Fourier series form using the improved differential evolution algorithm;

[0087] Step S106: Substitute the optimal parameters into the excitation trajectory in the Fourier series form to obtain the optimal excitation trajectory;

[0088] Step S107: Based on the excitation trajectory, collect robot motion data and use the least squares algorithm to identify the robot's dynamic parameters.

[0089] In this embodiment, a robot dynamics model is first established using the robot's link parameters. Then, the dynamics model is linearized to establish a minimum inertia parameter set. The parameters of the minimum inertia set are then extracted to obtain the observation matrix. The condition number of the observation matrix is ​​used as the objective function, and a Fourier series excitation trajectory is designed. An improved differential evolution algorithm is used to optimize the Fourier series excitation trajectory parameters to obtain optimal parameters. This ensures that the excitation trajectory allows the robot to operate stably within its maximum safe space, laying the groundwork for subsequent accurate dynamics parameter identification. Finally, the optimal parameters are substituted into the excitation trajectory to obtain the optimal excitation trajectory. Based on the optimal excitation trajectory, the robot's dynamics parameters are identified using a least squares algorithm. This invention utilizes an improved differential evolution algorithm to optimize the Fourier series excitation trajectory parameters, ensuring the robot can operate stably within its maximum safe space and improving the accuracy of subsequent dynamics parameter identification.

[0090] Based on the above embodiments, this embodiment describes in detail the specific process of the six-axis robot dynamics parameter optimization method. Please refer to [link / reference]. Figure 2 , Figure 2 A flowchart of a second specific embodiment of the robot dynamics parameter optimization method provided by the present invention; the specific operation steps are as follows:

[0091] Step S201: Construct a kinematic model of the robot based on its geometry;

[0092] Based on the geometry of the six-axis robot, establish a link coordinate system diagram for each link, such as... Figure 3 As shown;

[0093] The DH parameter table (kinematic model) of the six-axis robot is obtained from the link coordinate system diagram of each link, as shown in Table 1.

[0094] i <![CDATA[α i-1 / (°)]]> <![CDATA[a i-1 / mm]]> <![CDATA[d i / mm]]> <![CDATA[θ i / (°)]]> 1 0 0 0 <![CDATA[θ1]]> 2 -90 <![CDATA[a1=20.121]]> 0 <![CDATA[θ2]]> 3 0 <![CDATA[a2=260.037]]> 0 <![CDATA[θ3]]> 4 -90 <![CDATA[a3=19.663]]> <![CDATA[d4=279.484]]> <![CDATA[θ4]]> 5 90 0 0 <![CDATA[θ5]]> 6 -90 0 <![CDATA[d6=77.806]]> <![CDATA[θ6]]>

[0095] The parameters of each link are defined as follows:

[0096] a i Along X i Axis, from Z i Move to Z i+1 The distance;

[0097] α i Along X i Axis, from Z i Rotate to Z i+1 The distance;

[0098] d i Along Z i Axis, from X i-1 Move to X i The distance;

[0099] θ i Along Z i Axis, from X i-1 Rotate to X i The distance.

[0100] Step S202: Based on the robot's kinematic model, establish a six-axis robot dynamic model using the Newton-Euler iterative method;

[0101] Based on the DH parameter table of a six-axis robot, a dynamic model of the robot is established using the Newton-Euler iterative method.

[0102] Where M(q)∈R n×n The inertial tensor matrix, Let G(q) ∈ R be the vector of Coriolis force and centrifugal force. n Let q be the gravity vector, q∈R n From the robot's perspective, For the robot's speed, For the robot's acceleration, τ f ∈R n This is the term for joint friction.

[0103] Regarding the term τ of joint friction f Calculation method:

[0104] The joint friction model of the robot is established using the Coulomb-viscous friction model:

[0105]

[0106] In the formula, f c f is the Coulomb friction coefficient. v Let be the coefficient of viscous friction, and sign(·) be the sign function.

[0107] Step S203: Separate the state variables and inertial parameters of the six-axis robot dynamic model to obtain the minimum set of inertial parameters of the six-axis robot;

[0108] Separating the state variables and inertial parameters in the dynamic model, the inertial parameters of each link contain twelve constants related to the physical properties of the link, which is the minimum set of inertial parameters for a six-axis robot, specifically represented by P:

[0109]

[0110] Where, m i Let x be the mass of the i-th joint. i y i , z i Let I be the position of the centroid of the i-th joint. xx，i I yy，i I zz，i I xy，i I xz，i I yz,i f is the value of the inertia tensor matrix of the i-th joint in the joint coordinate system. c，i f v，i These are the Coulomb friction coefficient and the viscosity model coefficient, respectively.

[0111] Step S204: Extract the parameters of the minimum inertial parameter set of the six-axis robot to obtain the observation matrix;

[0112] Since the dynamic model consists of implicit nonlinear equations containing a set of dynamic inertial parameters, it needs to be linearized before calculation. The specific expression is as follows:

[0113]

[0114] In the formula, Y∈R n×mn It is a function that contains only joint angles, joint angular velocities, and joint angular accelerations, where m is the number of identifiable parameters and n is the number of joints.

[0115] The relationship between the inertial parameter set, the observation matrix, and the joint torques can be expressed as:

[0116] Y n×mn P mn×1 =τ n×1

[0117]

[0118] The observation matrix Y in the above equation is not full rank, so parameter identification cannot be performed directly from the observation matrix. Not every robot inertial parameter affects the torque; the column corresponding to this parameter in the observation matrix is ​​zero, and therefore can be eliminated. Ultimately, the inertial parameters that affect the robot's links can be retained to form a minimal identifiable set of inertial parameters P. k The corresponding full-rank state observation matrix is ​​Y. k :

[0119] Y k P k =τ

[0120] In this embodiment, the 72 dynamic parameters of the six joints of the six-axis robot are simplified into a minimum set of inertial parameters. The minimum parameter of a joint-link of the six-axis robot is I. zz f c f v The minimum parameter for each link from joint two to joint six is ​​I. xx I xy I xz I yz I zz ,mx,my,f c f v .

[0121] Step S205: Use the number of adjustments in the observation matrix as the objective function and construct the excitation trajectory in the form of a Fourier series;

[0122] The condition number of the observation matrix reflects the noise resistance of the excitation trajectory. The larger the condition number, the greater the ill-conditioned nature of the matrix. Therefore, the condition number of the observation matrix can be used as the optimization criterion for the excitation trajectory.

[0123] The condition number of the observation matrix is ​​taken as the objective function, and the expression of the objective function is:

[0124] F(Y) = cond(Y) = ||(Y) -1 ||·||Y||

[0125] min(F(Y));

[0126] The expression for the excitation trajectory in Fourier series form is:

[0127]

[0128] Among them, w f q is the fundamental frequency of the Fourier series, N is the harmonic number of the Fourier series, and q is the frequency of the Fourier series. i，0 a is a constant term. k，i b k，i These are the amplitude coefficients of the sine and cosine functions.

[0129] The constraints are as follows:

[0130] Where {s(q(t))} is the set of end-effector positions of the robot during movement, S represents the maximum workspace allowed for the robot to operate, and q max q min These represent the upper and lower limits of the angles of each joint. These represent the upper and lower limits of the velocity of each joint. Table 2 shows the upper and lower limits of acceleration for each joint, and the constraints for each joint are shown in Table 2.

[0131] Table 2. Constraints of each joint of the six-axis robot

[0132]

[0133] The quality of the excitation trajectory design directly affects the accuracy of robot body parameter identification. In designing the excitation trajectory for the robot body, it is necessary to consider its stability, smoothness, and noise resistance; on the other hand, to fully utilize the robot's dynamic characteristics, the robot's movement should occupy as much of the workspace as possible. A periodic Fourier series is chosen as the excitation trajectory, and an improvement is made by replacing the constant terms of the Fourier series with a fifth-order polynomial to ensure the continuity of initial angle, velocity, and acceleration.

[0134] After determining the form and constraints of the excitation trajectory, the parameters of the excitation trajectory can be optimized using an optimization algorithm. Equation min(F(Y)) is an optimization problem, which is solved using an improved differential evolution algorithm. Because of the robot's joint angles, angular velocities, angular accelerations, and end-effector position motion space limitations, the excitation trajectory is subject to certain constraints. The improved differential evolution algorithm can address these constraints by designing a penalty function that adds constraints.

[0135] Step S206: Optimize the excitation trajectory in Fourier series form using an improved difference algorithm to obtain the optimal excitation trajectory;

[0136] S61: Initialize the population parameter N, and set the maximum number of iterations g. max ;

[0137] S62: Calculate the objective function value of parameter i in the parameter population, where i∈(1,2,3,…,N);

[0138] S63: Determine whether the objective function value of parameter i satisfies the constraint conditions;

[0139] S64: If the objective function value of parameter i does not satisfy the constraint condition, then set the objective function value of parameter i to 10. 10 If satisfied, proceed to S65;

[0140] S65: Determine the magnitude of the objective function value and the initial function value of parameter i;

[0141] S66: If the objective function value is greater than the initial function value, then the objective function value is used as the function value of parameter i;

[0142] S67: Mutate and crossover the parameters in the population to generate a new i-th parameter;

[0143] S68: Calculate the objective function value of the new parameter i, and determine whether the objective function value of the new parameter i is greater than the function value of parameter i;

[0144] S69: If the objective function value of the new parameter i is greater than the function value of the parameter i, then the parameter i is taken as the optimal parameter; if the objective function value of the new parameter i is less than the objective function value of the parameter i, then the new parameter i is taken as the optimal parameter.

[0145] S610: Let i = i + 1, return to step S62, until i = g max Output the optimal parameters.

[0146] In this embodiment, the Differential Evolution (DE) algorithm and the improved Differential Evolution (IDE) algorithm of this invention are compared. The initial population size is set to 20 for both algorithms, and the maximum number of generations is 200. As shown in the figure, the initial population quality of the IDE algorithm is significantly better than that of the DE algorithm. The DE algorithm converges to 69.44 after 105 generations, while the improved algorithm converges to 54.35 after 84 generations, demonstrating improvements in both convergence speed and accuracy. Figure 4 As shown.

[0147] The optimized excitation trajectory parameters are shown in Table 3, and the optimized excitation trajectory is as follows: Figure 5 As shown, the optimized excitation trajectory velocity is as follows: Figure 6 As shown, the optimized excitation trajectory acceleration is as follows: Figure 7As shown in the figure, the comparison reveals that the joint angle curves of the optimized trajectory strictly meet the angle constraints, and the joint angular velocity and joint angular acceleration also meet the robot's constraints. Furthermore, the trajectory is continuous and smooth, satisfying the robot's constraint requirements.

[0148] Table 3. Parameters after Fourier series optimization of the excitation trajectory.

[0149]

[0150] Step S207: Input the optimal excitation trajectory into the six-axis robot, collect the robot's motion data, and identify the dynamic parameters of the six-axis robot using the least squares algorithm;

[0151] Collect motion data from a six-axis robot;

[0152] The dynamic parameter model of the six-axis robot is identified using the least squares algorithm, and its calculation formula is as follows:

[0153]

[0154] The identified dynamic parameters of the six-axis robot are shown in Table 4.

[0155] Table 4 Identification of Dynamic Parameters of a Six-Axis Robot

[0156]

[0157]

[0158] Step S208: Repeat steps S201 to S207 above to obtain new dynamic parameters. Compare the new dynamic parameters with the original dynamic parameters to verify the accuracy of the identification results.

[0159] In this embodiment, the predicted torque of each node is compared with the actual collected torque, and the root mean square error (RMS) of the torque is also compared. For example, joint 1... Figure 8 As shown, joint 2 is as follows Figure 9 As shown, joint 3 is as follows Figure 10 As shown, joint 4 is as follows Figure 11 As shown, joint 5 is as follows Figure 12 As shown, joint 6 is as follows Figure 13 As shown; from Figures 8 to 13 It can be seen that the predicted torque values ​​of each joint have a high degree of overlap with the measured torque values, and the root mean square of each joint tends to zero. Therefore, the identification accuracy of this invention is high.

[0160] In this embodiment, a DH parameter table for the six-axis robot is constructed based on its ensemble structure. A dynamic model of the six-axis robot is then established based on this DH parameter table. The dynamic parameter model is linearized to obtain the minimum set of parameters to be identified. An observation matrix is ​​then constructed, and the number of conditions in the observation matrix is ​​used as the objective function. An excitation trajectory in Fourier series form is designed, and an improved differential evolution algorithm is used to optimize the Fourier series coefficients. The robot is driven to move along the optimized excitation trajectory. Data is collected and the dynamic parameters are identified using the least squares method. The torque of the robot's dynamic model is calculated based on the identified dynamic parameters. Finally, a new excitation trajectory is designed to verify the accuracy of the identified robot dynamic parameters. This invention uses an improved differential algorithm to optimize the Fourier coefficients, ensuring the robot can operate stably within its maximum safe space, improving the accuracy of dynamic parameter identification. Furthermore, multiple new excitation trajectories are designed to verify the accuracy of the six-axis robot's dynamic parameters, further enhancing the accuracy of the identified dynamic parameters. This fully utilizes the robot's dynamic characteristics and improves its noise resistance.

[0161] Please refer to Figure 14 , Figure 14 This invention provides a structural block diagram of a device for identifying robot dynamic parameters; the specific device may include:

[0162] Model building module 100: used to build a robot kinematic model through a robot link parameter table, and to build a dynamic model based on the kinematic model;

[0163] The processing model module 200 is used to linearize the dynamic model to obtain the minimum set of inertial parameters to be identified.

[0164] The parameter extraction module 300 is used to extract parameters based on the minimum inertial parameter set to be identified and to establish an observation matrix;

[0165] A target function module 400 is constructed to use the condition number of the observation matrix as the target function and to construct the excitation trajectory and constraint conditions in the form of a Fourier series.

[0166] The optimization calculation module 500 is used to solve for the optimal parameters of the excitation trajectory in the form of the Fourier series using an improved differential evolution algorithm;

[0167] The excitation trajectory calculation module 600 is used to input the optimal parameters into the excitation trajectory in the form of the Fourier series to obtain the optimal excitation trajectory;

[0168] The parameter calculation module 700 collects robot motion data based on the excitation trajectory and identifies robot dynamic parameters using the least squares algorithm.

[0169] This embodiment of a robot dynamics parameter identification device is used to implement the aforementioned robot dynamics parameter optimization method. Therefore, the specific implementation of the robot dynamics parameter identification device can be found in the embodiment section of the robot dynamics parameter optimization method above. For example, the model construction module 100, the model processing module 200, the parameter extraction module 300, the objective function construction module 400, the optimization calculation module 500, the excitation trajectory calculation module 600, and the parameter calculation module 700 are respectively used to implement steps S101, S102, S103, S104, S105, S106, and S107 in the aforementioned robot dynamics parameter optimization method. Therefore, its specific implementation can be referred to the description of the corresponding embodiments, and will not be repeated here.

[0170] A specific embodiment of the present invention also provides a device for identifying robot dynamic parameters, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the above-described robot dynamic parameter optimization method.

[0171] A specific embodiment of the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for optimizing robot dynamic parameters.

[0172] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0173] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0174] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0175] The foregoing has provided a detailed description of a robot dynamics parameter optimization method, apparatus, device, and computer-readable storage medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and core ideas of the present invention. It should be noted that those skilled in the art can make various improvements and modifications to the present invention without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims

1. A method of robot dynamics parameter optimization, characterized by, The application relates to a robot dynamic parameter identification method and device. The robot kinematics model is established through a robot link parameter table, and a dynamics model is constructed according to the kinematics model; The dynamics model is linearized to obtain a minimum inertia parameter set to be identified; Parameter extraction is performed according to the minimum inertia parameter set to be identified, and an observation matrix is established; The condition number of the observation matrix is taken as a target function, and a Fourier series form of an excitation trajectory and a constraint condition are constructed; An improved differential evolution algorithm is used to solve the optimal parameters of the Fourier series form of the excitation trajectory; The optimal parameters are brought into the Fourier series form of the excitation trajectory to obtain an optimal excitation trajectory; Robot motion data are collected based on the excitation trajectory, and robot dynamics parameters are identified by using a least square algorithm; The improved differential evolution algorithm is used to solve the optimal parameters of the Fourier series form of the excitation trajectory, and the solving method comprises the following steps: S1: initialize parameter population N, set the maximum iteration number g max ; S2: the target function value of parameter i in the parameter population is calculated, wherein i is an integer in (1, 2, 3, …, N); S3: whether the target function value of the parameter i meets the constraint condition is judged; S4: If the objective function value of the parameter i does not satisfy the constraint condition, the objective function value of the parameter i is set to 10 10 , and if it does, S5 is performed; S5: whether the target function value of the parameter i is greater than the initial function value is judged; S6: if the target function value is greater than the initial function value, the target function value is taken as the function value of the parameter i; S7: mutate the parameters in the population, crossover operation, generate the i * new parameters; S8: Calculate the objective function value of the new parameter i * and determine whether the objective function value of the new parameter i * is greater than the objective function value of the parameter i; S9: if the objective function value of the new parameter i * is greater than the function value of the parameter i, then the parameter i is taken as the best parameter, if the objective function value of the new parameter i * is smaller than the function value of the parameter i, then the new parameter i * is taken as the best parameter; S10: Let i = i * + 1, go back to step S2 until i = g max , output the best parameter; The condition number of the observation matrix is taken as the target function, and the expression of the target function is: min (F (Y)); F(Y) = cond(Y) = ||(Y) -1 ||·||Y|| The expression of the Fourier series form of the excitation trajectory is: After the robot motion data are collected based on the excitation trajectory and the robot dynamics parameters are identified by using the least square algorithm, the following steps are further included: wherein w f is the fundamental frequency of the Fourier series, N is the harmonic number of the Fourier series, q i,0 is the constant term, a k,i , b k,i is the amplitude coefficient of the cosine function; The constraint condition is: where {s(q(t))} is the set of end position of the robot when moving, S represents the maximum workspace allowed for the robot to run, q max , q min are the upper and lower limits of each joint angle respectively, are the upper and lower limits of each joint speed respectively, are the upper and lower limits of each joint acceleration respectively.

2. The parameter optimization method of claim 1, wherein, A new optimal excitation trajectory is reconstructed based on the identified robot dynamics parameters; The torque information of the new optimal excitation trajectory and the root mean square difference of the torque information of the optimal excitation trajectory are calculated; If the root mean square difference does not meet a preset value, the excitation trajectory and the dynamics parameters are recalculated. The robot kinematics model is established through a robot link parameter table, and a dynamics model is constructed according to the kinematics model; 3. The parameter optimization method of claim 1, wherein, According to the geometric structure of the robot, a link coordinate system diagram of each link is established; The kinematics model of the robot is obtained according to the link coordinate system diagram of each link; The dynamics model is linearized to obtain a minimum inertia parameter set to be identified; Based on the kinematics model of the robot, a robot dynamics model is established using Newton-Euler iteration method where M(q) e R n×n is the inertia tensor matrix, is the Coriolis and centrifugal force vector, G(q) e R n is the gravity vector, q e R n is the angle of the robot, is the velocity of the robot, is the acceleration of the robot, τ f e R n is the joint friction term, n is the number of joints.

4. The parameter optimization method of claim 1, wherein, The state variables and inertia parameters in the dynamics model are separated to obtain the minimum inertia parameter set P to be identified, and the expression is: The minimum inertia parameter set to be identified is extracted to construct the observation matrix Y, and the expression is: The application further discloses a robot dynamic parameter identification device. where m i is the mass of the ith joint, x i , y i , z i are the center of mass position of the ith joint, I xx,i , I yy,i , I zz,i , I xy,i , I xz,i , I yz，i are the values of the inertia tensor matrix of the ith joint in the joint coordinate system, f c，i , f v,i are the Coulomb friction coefficient and the viscous model coefficient, respectively.

5. The parameter optimization method of claim 1, wherein, The model construction module is used for establishing a robot kinematics model through a robot link parameter table and constructing a dynamics model according to the kinematics model; The processing model module is used for linearizing the dynamics model to obtain a minimum inertia parameter set to be identified; where Y ∈ R n×mn is a function only containing joint angles, joint angular velocities, joint angular accelerations, m is the number of parameters to be identified, and n is the number of joints.

6. An apparatus for robot dynamics parameter optimization, characterized by, ​ ​ ​ A parameter extraction module is configured to extract parameters according to the minimum inertia parameter set to be identified, and to establish an observation matrix. A target function module is configured to take the condition number of the observation matrix as a target function, and to construct an excitation trajectory in the form of Fourier series and a constraint condition. An optimization calculation module is configured to solve the optimal parameters of the excitation trajectory in the form of Fourier series by using an improved differential evolution algorithm. A calculation excitation trajectory module is configured to bring the optimal parameters into the excitation trajectory in the form of Fourier series to obtain an optimal excitation trajectory. A calculation parameter module is configured to collect robot motion data based on the excitation trajectory, and to identify robot dynamics parameters by using a least square algorithm. The improved differential evolution algorithm includes the following steps: Initialize the population of parameters N, set the maximum number of iterations g max ; Calculating the target function value of parameter i in the parameter population, where i∈(1, 2, 3, …, N); Judging whether the target function value of the parameter i satisfies the constraint condition; if the objective function value of the parameter i does not satisfy the constraint condition, the objective function value of the parameter i is set to 10 10 , and if it does, S5 is performed; Judging the size of the target function value of the parameter i and the initial function value; If the target function value is greater than the initial function value, taking the target function value as the function value of the parameter i; The parameters in the population are mutated, crossed to generate the i * new parameters; calculating the objective function value for the new parameter i * and determining whether the objective function value for the new parameter i * is greater than the objective function value for the parameter i; If the objective function value of the new parameter i * is greater than the function value of the parameter i, the parameter i is taken as the best parameter, and if the objective function value of the new parameter i * is less than the function value of the parameter i, the new parameter i * is taken as the best parameter. Let i = i * + 1, go back to step S2 until i = g max , output the best parameters; The target function of the condition number of the observation matrix includes the following steps: Taking the condition number of the observation matrix as the target function, and the expression of the target function is: F(Y) = cond(Y) = ||(Y) -1 ||·||Y|| min(F(Y)); The expression of the excitation trajectory in the form of Fourier series is: wherein w f is the fundamental frequency of the Fourier series, N is the harmonic number of the Fourier series, q i，0 is the constant term, a k，i , b k，i is the amplitude coefficient of the cosine function; The constraint condition is: where {s(q(t))} is the set of end position of the robot when moving, S represents the maximum workspace allowed for the robot to run, q max , q min are the upper and lower limits of each joint angle respectively, are the upper and lower limits of each joint speed respectively, are the upper and lower limits of each joint acceleration respectively.

7. An apparatus for robot dynamics parameter optimization, characterized by including: A memory is configured to store a computer program. A processor is configured to implement the steps of the robot dynamics parameter optimization method according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium, characterized in that, The computer program is stored on the computer readable storage medium, and the computer program is executed by the processor to implement the steps of the robot dynamics parameter optimization method according to any one of claims 1 to 5.

Images