Industrial robot joint external torque estimation method and experimental system

Abstract

The invention discloses an industrial robot joint external torque estimation method and an experiment system, and the method comprises the following steps: (1) considering the influence of joint friction force and external force, and building a 3T2R configuration five-degree-of-freedom industrial robot dynamic model; (2) introducing a Stribeck friction-speed model to describe nonlinear characteristics of joint friction force of the industrial robot; (3) estimating the joint external torque of the industrial robot by using a generalized momentum observer of a self-adaptive super-spiral sliding mode algorithm, and optimizing self-adaptive law gain parameters by using a whale optimization algorithm to improve the precision of joint external torque estimation; according to the industrial robot joint external torque estimation method and the experimental system provided by the invention, the influence of friction force and external force of the robot can be fully considered, and the precision and robustness of industrial robot joint external torque estimation are effectively improved.

Description

[0001] The invention relates to the technical field of industrial robots, and in particular to an industrial robot joint external torque estimation method and an experimental system. Background Art

[0002] In recent years, intelligent robot processing has gradually become a new trend in high-quality manufacturing. However, the commonly used position and speed control modes cannot meet industrial requirements. In complex industrial environments, the interaction between industrial robots and external objects not only affects the operating accuracy, but also directly affects the stability and safety of the industrial robot system. Accurately estimating the external torque and force of industrial robots is crucial to improving the control performance of industrial robots and executing processing tasks.

[0003] Industrial robot operations, especially in tasks such as precision assembly, grinding, and welding, require industrial robots to perceive and adjust the interaction forces with the environment in real time, which further increases the complexity of estimating the external torques of industrial robot joints.

[0004] At present, the commonly used methods for estimating the external torque and external force of industrial robots include those based on filtering and observers and those based on machine learning and models. The filtering and observer methods focus on using the system state information for estimation. In the case of external disturbances or dynamic load changes, the observer's estimation may have large deviations; the machine learning and model-based methods rely on a large amount of high-quality data for training, usually require a lot of computing resources, and lack the explainability of their internal processes.

[0005] Therefore, it is necessary to propose an industrial robot joint external torque estimation method and experimental system according to the structural characteristics and operation mode of the robot, to study the robot joint external torque estimation, to realize the industrial robot joint external torque estimation, and to improve the robot's processing performance. Summary of the invention

[0006] In order to solve the problem of estimation accuracy of external torque of joints of industrial robots caused by external forces and friction during the processing of industrial robots, the present invention provides a method and experimental system for estimating external torque of joints of industrial robots, which can fully consider the influence of friction and external forces of robot joints, realize the estimation of external torque of robot joints, and improve the estimation accuracy and anti-interference of joint external torque through adaptive super-helical sliding mode algorithm, thus providing a way to realize the estimation of external torque of robot joints.

[0007] In order to achieve the above object, the present invention adopts the following technical scheme: In a first aspect, the present invention provides an industrial robot joint external torque estimation method, comprising the following steps:

[0008] Step 1: Considering the influence of joint friction and external force, establish the dynamic model of 3T2R configuration five-DOF industrial robot:

[0009] (1) According to the robot structural parameters, the Jacobian matrix of the robot system is calculated by using the mapping relationship between the velocity at the robot link and the end velocity using the vector product method;

[0010] (2) Establish the Lagrangian equation of the industrial robot in the generalized coordinate system, ignore the kinetic energy of the involved motion, only consider the kinetic energy of the connecting rod translation and rotation, obtain the inertia matrix of the robot system, and obtain the Coriolis matrix of the robot based on the symmetric positive definite characteristics of the inertia matrix of the industrial robot;

[0011] (3) Considering the influence of joint friction and external force during the movement of the industrial robot, a dynamic model of a 3T2R configuration five-degree-of-freedom industrial robot is established.

[0012] Step 2: Introduce the Stribeck friction-velocity model to describe the nonlinear characteristics of the friction force of the industrial robot joints:

[0013] (1) The Stribeck effect is introduced into the friction model of industrial robots to more accurately describe the nonlinear behavior of friction force during low-speed motion;

[0014] (2) Design a joint friction force identification experiment, ignore the Coriolis force and centrifugal force, simplify the industrial robot dynamic equation, and calculate the joint friction torque of the industrial robot;

[0015] (3) The Stribeck friction-velocity model of the friction torque of the industrial robot joints is established by combining the particle swarm optimization algorithm with the least squares method.

[0016] Step 3: Use the generalized momentum observer of the adaptive super-helical sliding mode algorithm to estimate the external torque of the industrial robot joint:

[0017] (1) Introducing the definition of generalized momentum of industrial robots and rewriting the dynamic equations of industrial robots;

[0018] (2) Design a finite-time momentum estimation observer, using a super-helical sliding mode generalized momentum observer to estimate the difference between the momentum dynamics and the actual momentum dynamics in a finite time. By using the integral form to obtain the actual control variable, high-frequency switching is avoided, thereby effectively reducing the system chattering phenomenon and enhancing the smoothness and stability of the system.

[0019] (3) The whale optimization algorithm is used to optimize the adaptive law gain parameters of the super-helical sliding mode algorithm to improve the accuracy of joint external torque estimation.

[0020] In the second aspect, the present invention provides an industrial robot joint external torque estimation method and experimental system, which are used to realize the above-mentioned industrial robot joint external torque estimation. The experimental system includes an industrial robot body, a workpiece to be processed, servo drivers for each joint, and a host industrial control computer and data processing software.

[0021] In the robot joint friction force identification experiment, the upper industrial control machine drives a joint motor to move forward at a certain speed for a certain distance, and the motor's forward output torque is measured. Then, the joint moves in the reverse direction at the same speed for a certain distance, and the motor's reverse output torque is measured. When the joint of the industrial robot returns to the initial position, different speeds are applied to the joint to continue to move in the forward and reverse directions at a constant speed, and the motor torque value during the forward and reverse constant speed motion is recorded. Finally, the Stribeck friction-speed curve model is fitted by combining the particle swarm optimization algorithm with the least squares method.

[0022] The robot joint external torque estimation experiment, during the industrial robot processing process, the upper industrial control computer TwinCAT control software samples the industrial robot joint angle, joint speed and torque information through the industrial robot body encoder with a sampling period of 2ms, and finally uses the adaptive super-helical sliding mode method based on the whale optimization algorithm to estimate the joint external torque;

[0023] Compared with the prior art, the advantages and positive effects of the present invention are:

[0024] (1) The present invention describes the dynamic equation of the robot system based on the Lagrangian dynamic equation, joint friction identification experiment and generalized momentum, and is applicable to all serial, parallel and hybrid robots, with strong versatility;

[0025] (2) The present invention models the friction force of robot joints, fully considering the influence of joint friction force, and can achieve the integrity and accuracy of industrial robot dynamics modeling;

[0026] (3) The present invention estimates the external torque of an industrial robot joint by using an adaptive super-helical sliding mode. When facing a complex dynamic environment, the method has stronger anti-interference and higher estimation accuracy, and does not require force sensors and acceleration information, further simplifying the system hardware configuration and reducing costs.

[0027] (4) By comparing the present invention with the super-helical sliding mode and the first-order generalized momentum observer method, it is verified that the present invention has higher estimation accuracy and stronger anti-interference ability.

[0028] The method and experimental system for estimating external torque of industrial robot joints using adaptive super-helical sliding mode of the present invention can achieve:

[0029] (1) Establish a friction model for each joint of an industrial robot to improve the accuracy of the robot system dynamics modeling;

[0030] (2) It can flexibly adjust the optimization strategy to cope with system changes, provide stronger robustness, and meet the real-time requirements of joint external torque estimation through the fast convergence characteristics of the superhelical algorithm. It does not need to rely on a large amount of training data, consumes less computing resources, and has good interpretability and optimization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to explain the technical solution of the present invention in detail, the drawings used in the present invention are briefly introduced.

[0032] Figure 1 Control block diagram for estimating joint external torques of the adaptive super-helical generalized momentum observer for the whale optimization algorithm;

[0033] Figure 2 This is a schematic diagram of the DH reference coordinate system of the kinematic model of an industrial robot according to the present invention;

[0034] Figure 3 This is a schematic diagram of an industrial robot joint friction torque identification experimental platform according to the present invention;

[0035] Figure 4 This is a schematic diagram of an experimental platform for estimating external torques of industrial robot joints according to an example of the present invention;

[0036] Figure 5 This is a schematic diagram comparing the external torque estimation of three super-helical sliding modes of an example industrial robot of the present invention;

[0037] Figure 6 It is a schematic diagram comparing the estimation of joint external torque of the adaptive super-helical generalized momentum observer and the estimation of joint external torque of the first-order generalized momentum observer of the whale optimization algorithm for industrial robots in the present invention;

[0038] Figure 7 This is a schematic diagram comparing the estimation of joint external torques by the adaptive super-helical generalized momentum observer and the estimation of joint external torques by the first-order generalized momentum observer of the whale optimization algorithm after a disturbance is applied to an example industrial robot of the present invention. DETAILED DESCRIPTION

[0039] In order to introduce the present invention in detail, based on the following relevant drawings, the industrial robot joint external torque estimation method and experimental system of adaptive super-helical sliding mode are explained in more detail; it is necessary to point out that the following description is only used for the detailed description of the present invention, and is not equivalent to the limitation of the protection scope of the present invention. At the same time, the industrial robot joint external torque estimation method and experimental system of adaptive super-helical sliding mode based on whale optimization algorithm can be implemented in other forms.

[0040] Figure 1 This is a control block diagram of the external torque estimation of the adaptive super-helical generalized momentum observer of the whale optimization algorithm of the present invention, which is described in detail below. The reader can estimate the external torque of the industrial robot joint according to it;

[0041] The present invention takes a 3T2R configuration high dexterity robot as an example, and its DH reference coordinate system is as follows: Figure 2 According to the present invention, the steps of performing the external torque of the joint of the industrial robot of the adaptive super-helical sliding mode are as follows:

[0042] Step 1: Considering the influence of joint friction and external force, establish the dynamic model of 3T2R configuration five-DOF industrial robot:

[0043] (1) According to the robot structural parameters, the Jacobian matrix of the robot system is calculated by using the mapping relationship between the velocity at the robot link and the end velocity using the vector product method.

[0044]

[0045] In formula (1), J Pi and J Oi Represents the i-th column of the Jacobian matrix of the moving joint and the rotating joint respectively; p e is the position vector of the working point of the end effector of the industrial robot, through the transformation matrix T e 0 The first three elements of the fourth column of are given; p i-1 By transformation matrix The first 3 elements of the fourth column are given; z i-1 is a function of the joint variable

[0046]

[0047] In formula (2), Represents the coordinate system from O0 to O i-1 The coordinate system transformation posture matrix.

[0048] Depend on Figure 2 From the DH reference coordinate system shown in the figure, it can be seen that the coupling relationship between movement and rotation is equivalently processed, and joint 1 is equivalent to a moving joint and joint 2 is equivalent to a rotating joint. According to formula (1), solve the J1~J5 columns to obtain the Jacobian matrix of the robot

[0049] J=[J1 J2 J3 J4 J5] (3)

[0050] (2) Establish the Lagrangian equation of the industrial robot in the generalized coordinate system, ignore the kinetic energy of the involved motion, only consider the kinetic energy of the connecting rod translation and rotation, obtain the inertia matrix of the robot system, and obtain the robot Coriolis matrix based on the symmetric positive definite characteristics of the inertia matrix of the industrial robot

[0051] Establishing the Lagrangian equations for industrial robots in a generalized coordinate system

[0052]

[0053] Where L is the Lagrangian function, which is the total kinetic energy of the system minus the potential energy of the system, ξ i is the generalized coordinate q i The relevant general force.

[0054] The kinetic energy of each joint link is expressed as a matrix

[0055]

[0056] Where B(q) is a symmetric positive definite n×n inertia matrix, where According to the symmetric positive definite characteristics of the inertia matrix of the industrial robot, we can get

[0057] Since potential energy is independent of angular velocity, the Lagrange equation is organized into a matrix form to obtain the dynamic equation of the industrial robot system

[0058]

[0059] In the formula, τ n represents the actual driving torque of the joint, represents the Coriolis torque and centrifugal torque, The (i, j)th element in can be decomposed into three small elements to solve

[0060]

[0061] Where b ij represents the centrifugal effect of joint i caused by the velocity of joint j

[0062] (3) Considering the influence of joint friction and external force during the movement of industrial robots, a dynamic model of a 3T2R five-degree-of-freedom industrial robot is established

[0063] When an industrial robot performs processing operations, the external forces acting on the Cartesian space at the end of the industrial robot (such as processing force, disturbance force, etc.) will be transmitted to the joints of the industrial robot through the end-operating tool of the industrial robot, thereby generating additional torque on the joints, and the robot joints will be affected by friction during movement. Therefore, the dynamic equations of the industrial robot system need to consider the influence of joint friction and external forces on the joints. At this time, the dynamic equations of the industrial robot system can be rewritten as

[0064]

[0065] In the formula, τ ext is the external torque of the industrial robot joint, τ f is the friction torque of the industrial robot joint

[0066] Step 2: Introduce the Stribeck friction-velocity model to describe the nonlinear characteristics of the friction force of the industrial robot joints:

[0067] Since the friction force of industrial robot joints shows obvious nonlinear characteristics, the Stribeck effect is introduced into the friction model of industrial robots to more accurately describe the nonlinear behavior of friction force at low speed.

[0068]

[0069] In the formula, f v 、f c and f s They are the viscous friction coefficient, the Coulomb friction coefficient and the static friction coefficient. represents an n×1 vector whose elements are given by the sign functions of the individual joint velocities, and μ are Stribeck convergence speed and curve shape coefficient respectively, μ is an empirical coefficient

[0070] (2) Design a joint friction force identification experiment, ignore the Coriolis force and centrifugal force, simplify the industrial robot dynamics equation, and calculate the joint friction torque of the industrial robot

[0071] When the joints of an industrial robot move at a constant speed and are not acted upon by external forces, is zero, remains unchanged, the Coriolis force and centrifugal force can be ignored, and the dynamic equation of the industrial robot can be simplified to

[0072] τ f +G(q)=τ n (11)

[0073] The motor is driven to move forward at a certain speed for a certain distance, and the forward output torque of the motor is measured. Then move in the opposite direction at the same speed for a certain distance, and measure the reverse output torque of the motor When the industrial robot's reciprocating positions are the same, the gravity torque is equal. Since the industrial robot joints are driven by high-precision reducers, the difference in friction between forward and reverse motion is very small. It can be considered that the friction forces are equal in magnitude but opposite in direction. The friction torque of the industrial robot joints can be calculated.

[0074]

[0075] (3) The Stribeck friction-velocity model of the joint friction torque of the industrial robot is established by combining the particle swarm optimization algorithm with the least squares method.

[0076] The actual output torque of the motor in the reciprocating motion of the industrial robot can be obtained to calculate the friction torque of the industrial robot joint at the current motion speed. By changing the motion speed of the industrial robot, the friction torque data at different joint motion speeds can be obtained. Then, the viscous friction coefficient f is calculated by combining the particle swarm optimization algorithm with the least squares method. v , Coulomb friction coefficient f c , static friction coefficient f s and Stribeck convergence rate and curve shape coefficient The Stribeck friction-velocity model of the friction torque of the industrial robot joints is established by fitting.

[0077] Step 3: Use the generalized momentum observer of the adaptive super-helical sliding mode algorithm to estimate the external torque of the industrial robot joint:

[0078] (1) Introducing the definition of generalized momentum of industrial robots and rewriting the dynamic equations of industrial robots

[0079] The generalized momentum p of an industrial robot can be defined as Taking the derivative of the generalized momentum p of the industrial robot, we can get

[0080]

[0081] Substituting equation (13) into the industrial robot dynamics equation (8) yields

[0082]

[0083] (2) Design a finite-time momentum estimation observer, using a super-helical sliding mode generalized momentum observer to estimate the difference between the momentum dynamics and the actual momentum dynamics in a finite time. By using the integral form to obtain the actual control variable, high-frequency switching is avoided, thereby effectively reducing the system chattering phenomenon and enhancing the smoothness and stability of the system.

[0084] Taking the external torque of the industrial robot joint as disturbance, a super-helical sliding mode momentum observer is established

[0085]

[0086] Where x1 is the difference between the estimated momentum and the actual momentum of the industrial robot, x2 is the auxiliary variable, sgn is the sign function, and k1 and k2 are the control gains.

[0087] From equation (15), we can see that under steady-state conditions, when x1 = 0 and hour, At this time x2=τ ext , that is, when the estimated generalized momentum of the industrial robot joint As x2 approaches the actual generalized momentum p, it approaches τ ext , so the external torque τ of the industrial robot joint can be ext Make an estimate.

[0088] Substituting the industrial robot dynamics equation into equation (15) we can get the new super-helical sliding mode observer:

[0089]

[0090] The calculated τ n , τ f , Substituting G(q) into equation (16), when x1 approaches 0, the required x2 is the estimated external torque of the industrial robot joint

[0091] (3) The whale optimization algorithm is used to optimize the adaptive law gain parameters of the super-helical sliding mode algorithm to improve the accuracy of joint external torque estimation.

[0092] In practical applications, many system parameters (such as friction, elasticity, etc.) and processing conditions are often unknown or change over time. If you rely on fixed control gains k1 and k2, these changes may cause system performance to degrade or even cause instability. By introducing an adaptive mechanism, the control gains k1 and k2 can be automatically adjusted according to the real-time status of the system to cope with these uncertainties.

[0093]

[0094] In the formula and are the estimated values ​​of k1 and k2. The whale algorithm is used to optimize the adaptive law gain parameters γ1 and γ2 to ensure the overall performance of the system and the negative definiteness of the derivative of the Lyapunov function, thereby improving the stability and control accuracy of the system.

[0095] The optimization target of the whale algorithm is the estimated value of the external torque in the joint space of the industrial robot The external moment τ in the joint space calculated by the dynamic equation of the industrial robot ext The minimum mean absolute error of

[0096]

[0097] Where N is the number of estimates.

[0098] In order to verify the effectiveness of the adaptive super-helical sliding mode industrial robot joint external torque estimation method based on the whale optimization algorithm of the present invention, the method of the present invention (WOA-ASTSM-GMO) is compared with the adaptive super-helical sliding mode (ASTSM-GMO), super-helical sliding mode (STSM-GMO) and first-order generalized momentum observer (FMGMO) methods using different gain parameters α for experimental analysis. Taking joint 1 as an example, the results are shown in Figure 5 , 6 As shown, it is verified that the proposed method has higher estimation accuracy; the method of the present invention is compared with the first-order generalized momentum observer method after applying the same random disturbance for the second time, and the results are as follows Figure 7 As shown in the figure, it is verified that the proposed method has stronger anti-interference ability.

[0099] Although the exemplary embodiments of the present invention have been described in detail, this does not limit the scope of the present invention. Those skilled in the art may modify the present invention in details and form without departing from the scope and spirit disclosed in the claims, and all changes made should belong to the protection scope of the technical solution of the present invention, and the scope of the present invention is not limited by the above technical solution, but is defined by the claims.

Claims

1. A method and experimental system for estimating external torque of an industrial robot joint, the method comprising the following steps: Step 1: Considering the influence of joint friction and external force, establish the dynamic model of 3T2R configuration five-DOF industrial robot: (1) According to the robot structural parameters, the Jacobian matrix of the robot system is calculated by using the mapping relationship between the velocity at the robot link and the end velocity using the vector product method; (2) Establish the Lagrangian equation of the industrial robot in the generalized coordinate system, ignore the kinetic energy of the involved motion, only consider the kinetic energy of the connecting rod translation and rotation, obtain the inertia matrix of the robot system, and obtain the Coriolis matrix of the robot based on the symmetric positive definite characteristics of the inertia matrix of the industrial robot; (3) Considering the influence of joint friction and external force during the movement of the industrial robot, a dynamic model of a 3T2R configuration five-degree-of-freedom industrial robot is established. Step 2: Introduce the Stribeck friction-velocity model to describe the nonlinear characteristics of the friction force of the industrial robot joints: (1) The Stribeck effect is introduced into the friction model of industrial robots to more accurately describe the nonlinear behavior of friction force during low-speed motion; (2) Design a joint friction force identification experiment, ignore the Coriolis force and centrifugal force, simplify the industrial robot dynamic equation, and calculate the joint friction torque of the industrial robot; (3) The Stribeck friction-velocity model of the friction torque of the industrial robot joints is established by combining the particle swarm optimization algorithm with the least squares method. Step 3: Use the generalized momentum observer of the adaptive super-helical sliding mode algorithm to estimate the external torque of the industrial robot joint: (1) Introducing the definition of generalized momentum of industrial robots and rewriting the dynamic equations of industrial robots; (2) Design a finite-time momentum estimation observer, using a super-helical sliding mode generalized momentum observer to estimate the difference between the momentum dynamics and the actual momentum dynamics in a finite time. By using the integral form to obtain the actual control variable, high-frequency switching is avoided, thereby effectively reducing the system chattering phenomenon and enhancing the smoothness and stability of the system. (3) The whale optimization algorithm is used to optimize the adaptive law gain parameters of the super-helical sliding mode algorithm to improve the accuracy of joint external torque estimation.

2. The method for estimating external torque of an industrial robot joint according to claim 1, characterized in that Considering the influence of joint friction and external force, the robot Coriolis matrix is ​​obtained according to the symmetric positive definite characteristics of the inertia matrix of the industrial robot, and the dynamic model of the 3T2R configuration five-degree-of-freedom industrial robot is established, including: (1) Based on the mapping matrix between the linear velocity and angular velocity of the industrial robot joints and the linear velocity and angular velocity of the corresponding end effector, the Jacobian matrix of the robot system is solved by the vector product method. In formulas (1) and (2), J Pi and J Oi Represents the i-th column of the Jacobian matrix of the moving joint and the rotating joint respectively; pe is transformed by the matrix T e 0 The first three elements of the fourth column of are given; p i-1 By transformation matrix The first three elements of the fourth column; z i-1 is a function of the joint variables; Represents the coordinate system from O0 to O i-1 The rotation matrix of (2) The Lagrangian equation of the industrial robot is established in the generalized coordinate system to obtain the inertia matrix of the robot system. The Coriolis matrix of the robot is obtained based on the symmetric positive definite property of the inertia matrix of the industrial robot. Establishing the Lagrangian equations for industrial robots in a generalized coordinate system Where L is the Lagrangian function, which is the total kinetic energy of the system minus the potential energy of the system, ξ i is the generalized coordinate q i The relevant general force. In order to calculate the kinetic energy of each link of the industrial robot, the kinetic energy is expressed as a function of the system joint variables. The kinetic energy of each link includes two parts: translational kinetic energy and rotational kinetic energy. Where, T li is the kinetic energy of joint i, and are the Jacobian matrix columns related to the translation joint velocity and the rotation joint velocity, respectively. The kinetic energy of each joint link is expressed as a matrix Where B(q) is a symmetric positive definite n×n inertia matrix, where It is obtained based on the symmetric positive definite property of the inertia matrix of the industrial robot. Since potential energy is independent of angular velocity, the Lagrange equation is organized into a matrix form to obtain the dynamic equation of the industrial robot system In the formula, τ n represents the actual driving torque of the joint, represents the Coriolis torque and centrifugal torque, The (i, j)th element in can be decomposed into three small elements to solve Where b ij represents the eccentric effect of joint i caused by the velocity of joint j; (3) Considering the influence of joint friction and external force during the movement of the industrial robot, a dynamic model of a 3T2R five-degree-of-freedom industrial robot is established The dynamic equation of the industrial robot system needs to consider the influence of joint friction and the influence of external forces on the joints. The dynamic equation of the industrial robot system can be rewritten as In the formula, τ f represents the joint friction torque, τ ext represents the external moment of the joint.

3. The method for estimating external torque of an industrial robot joint according to claim 1, characterized in that: Taking full account of the influence of robot joint friction, the Stribeck friction-velocity model is introduced to describe the nonlinear characteristics of industrial robot joint friction, including (1) The Stribeck effect is introduced to more accurately describe the nonlinear behavior of friction at low speeds. In the formula, f v 、f c and f s They are the viscous friction coefficient, the Coulomb friction coefficient and the static friction coefficient. represents an n×1 vector whose elements are given by the sign functions of the individual joint velocities, and μ are the Stribeck convergence speed and curve shape coefficient respectively. And μ is an empirical coefficient, generally 0.5 to 2. (2) Design a joint friction force identification experiment, ignore the Coriolis force and centrifugal force, simplify the industrial robot dynamics equation, and calculate the joint friction torque of the industrial robot When the joints of an industrial robot move at a constant speed and are not acted upon by external forces, is zero, remains unchanged, the Coriolis force and centrifugal force terms can be ignored, and the dynamic equation of the industrial robot can be simplified to t f +G(q)=τ n (10) By obtaining the actual output torque of the motor during the reciprocating motion of the industrial robot, the friction torque of the industrial robot joint at the current motion speed can be calculated. (3) The Stribeck friction-velocity model of the joint friction torque of the industrial robot is established by combining the particle swarm optimization algorithm with the least squares method. The particle swarm optimization algorithm is combined with the least squares method to calculate the viscous friction coefficient f. v , Coulomb friction coefficient f c , static friction coefficient f s and Stribeck convergence rate and curve shape coefficient The Stribeck friction-velocity model of the friction torque of the industrial robot joints is established by fitting.

4. The method for estimating external torque of an industrial robot joint according to claim 1, characterized in that: The generalized momentum observer of the adaptive super-helical sliding mode algorithm is used to estimate the external torque of the industrial robot joints, and the whale optimization algorithm is used to optimize the adaptive law gain parameters to improve the accuracy of the external torque estimation of the joints, including (1) Introducing the definition of generalized momentum of industrial robots and rewriting the dynamic equations of industrial robots; Introduce the generalized momentum p of the industrial robot and derive it The generalized momentum derivative Substituting into the dynamic equation of the industrial robot disclosed in the present invention, we can get (2) Design an observer for finite-time momentum estimation, using a super-helical sliding mode generalized momentum observer to estimate the difference between the momentum dynamics and the actual momentum dynamics in finite time. Taking the external torque of the industrial robot joint as disturbance, a super-helical sliding mode momentum observer is established Where x1 is the difference between the estimated momentum and the actual momentum of the industrial robot, x2 is the auxiliary variable, sgn is the sign function, k1 and k2 are the control gains, and when x1 approaches 0, the required x2 is the estimated external torque of the industrial robot joint. Based on the Lyapunov stability theory, the adaptive control law of control gains k1 and k2 is designed to strictly satisfy the derivative of the Lyapunov function. Negative conditions, design and The adaptive law is as follows (3) Using the whale optimization algorithm to optimize the adaptive law gain parameters of the super spiral sliding mode algorithm The whale algorithm is used to optimize the adaptive law gain parameters γ1 and γ2 to ensure the overall performance of the system and the negative definiteness of the derivative of the Lyapunov function, thereby improving the stability and control accuracy of the system. The optimization target of the whale algorithm is the estimated value of the external torque in the joint space of the industrial robot. The external moment τ in the joint space calculated by the industrial robot dynamics equation after dynamic parameter identification ext The minimum mean absolute error of Where N is the number of estimates.

5. The method for estimating external torque of an industrial robot joint according to claim 1, characterized in that: The external torque of each joint of the robot during the machining process is collected through the servo drivers of each joint of the robot and the upper industrial control computer, and compared with the external torque of the joint estimated by the adaptive super-helical sliding mode method. Step 1: Write the desired inertia matrix of the robot and the friction model equations of each joint into the control program of the robot's upper industrial control computer; Step 2: Fix the required workpiece; Step 3: Control the robot to perform processing operations; Step 4: Collect the torque and speed information of the servo motors of each joint in the upper industrial control computer, and input them into the external torque estimation method of the industrial robot joint of the adaptive super-helical sliding mode; Step 5: post-processing the collected torque signals in the servo motors of each joint and the estimated torque of the industrial robot joints of the adaptive super-helical sliding mode; Step 6: Calculate the mean absolute error between the estimated external torque of the industrial robot joint of the adaptive super-helical sliding mode and the external torque signal of the servo motor after friction torque compensation; Step 7: Determine the effectiveness of the method of the present invention based on the calculated mean absolute error.

Images