A collaborative robot drag teach control method, device, equipment and medium

Abstract

The application discloses a collaborative robot drag teaching control method, device, equipment and medium. The method comprises the following steps: calculating the dynamics feedforward compensation of the robot; calculating the corresponding friction compensation for different motion trends; calculating the damping of the end of each connecting rod of the robot; adding the dynamics feedforward compensation torque, the friction compensation torque and the damping torque to obtain the expected torque, and then issuing the expected torque to the joint motor. The application can improve the flexibility of the collaborative robot drag without adding a joint torque sensor, and improve the deployment efficiency of the operator.

Description

Technical Field

[0001] This application relates to the field of robot control technology, and in particular to a collaborative robot drag teaching control method, device, equipment and medium. Background Technology

[0002] The dragging function of collaborative robots, also known as the free-drive function, allows operators to directly drag the robot to a designated position and posture, thereby quickly teaching the required pose or continuous trajectory. During the dragging process, the smoothness and coordination of the dragging are crucial.

[0003] Existing collaborative robots have the following problems:

[0004] 1. For collaborative robots without torque sensors, existing feedforward compensation techniques based on Coulomb & viscous friction models and gravity models suffer from laborious start-stop and reversal processes in dragging mode, failing to meet the requirements for rapid dragging deployment. This is partly due to high acceleration noise, which cannot compensate for inertial forces, and partly due to the lack of compensation for static friction.

[0005] 2. Operators typically drag the robot by its end effector. Because the force transmitted from the end effector to the robot's joints is uneven, joints with longer lever arms experience greater torque and begin moving first. Other joints, however, do not receive sufficient external force to initiate movement and remain stationary, making it difficult to achieve synchronized movement across all joints. This results in the robot's end effector not moving in the direction the operator expects. This asynchronous movement leads to an uncoordinated feel for the operator, making it difficult to achieve precise positioning (i.e., simultaneous arrival of posture and position) during dragging. Operators usually drag the robot to the vicinity of the designated pose and then use a teach pendant for fine adjustments, which undoubtedly reduces the deployment efficiency of collaborative robots. Although adding a six-dimensional force sensor can improve this problem, it significantly increases costs. Summary of the Invention

[0006] This application provides a collaborative robot drag teaching control method, device, equipment and medium, which has the advantage of improving the compliance of the collaborative robot dragging without installing joint torque sensors, thereby improving the operator's deployment efficiency.

[0007] The technical solution of this application is as follows:

[0008] On the one hand, this application provides a collaborative robot drag-and-drop teaching control method, including the following steps:

[0009] Computational robot dynamics feedforward compensation;

[0010] Calculate the corresponding friction compensation for different motion trends;

[0011] Calculate the damping at the ends of each link of the robot;

[0012] The dynamic feedforward compensation torque, friction compensation torque, and damping torque are added together and sent to the joint motor as the desired torque.

[0013] Furthermore, computational robot dynamics feedforward compensation includes:

[0014] Establish the robot's inverse dynamics model and output the dynamic feedforward compensation torque;

[0015] The inverse dynamics model of the robot is as follows:

[0016]

[0017] Where, τ dyn =(τ1,τ2,τ3,τ4,τ5,τ6) T Let τ be the equivalent moment vector of the robot's joint space dynamic model. i Represents the equivalent torque of the dynamic model of each joint; Let V be the velocity vector in the robot's joint space. Indicates the rotational speed of each joint; Let be the acceleration vector in the robot's joint space. Let M(q) represent the rotational acceleration of each joint; M(q) is the inertia matrix in the robot's inverse dynamics model. G(q) represents the Coriolis force and centrifugal force matrices in the robot's inverse dynamics model; G(q) represents the gravity vector in the robot's inverse dynamics model.

[0018] Where α is the scaling factor.

[0019] Through the above scheme, this application controls the motor to output high-frequency harmonic torque to assist in starting when the joint is stationary.

[0020] Furthermore, the calculation of friction compensation includes:

[0021] Configure dynamic friction compensation start speed Obtain the joint velocities of the robot

[0022] When the absolute value of the robot's joint velocity Less than the dynamic friction compensation start-up speed At this time, static friction compensation is used for friction compensation:

[0023] Define the frequency f of the static friction compensation function, and calculate the period of the static friction function. and angular frequency ω f =2π*f, define the initial value t of the static friction compensation timer. i =0;

[0024] Update static friction compensation timer t i When it is greater than T, make t i =t i -T, where i represents the i-th joint;

[0025] Calculate the static friction compensation torque, i.e.

[0026]

[0027] in It is static friction compensation torque, ω f It is the angular frequency of the static friction compensation function, t i It is the time of the static friction compensation timer, k c,i It is the Coulomb friction coefficient of the joint;

[0028] The friction feedforward compensation torque required by each joint motor is equal to the static friction compensation torque, that is...

[0029]

[0030] Furthermore, the calculation of friction compensation includes:

[0031] When the absolute value of the robot's joint velocity Greater than the dynamic friction compensation start-up speed and Less than 0.9*k c,i At that time, the output will be based on the calculated value of the current static friction compensation, i.e.

[0032]

[0033] When the absolute value of the robot's joint velocity Greater than the dynamic friction compensation start-up speed and Greater than 0.9*k c,i When the joint is determined to have entered the dynamic friction compensation stage, the acceleration is first corrected:

[0034]

[0035] in It is the corrected acceleration, a deadZone It is used to correct the dead zone of acceleration. It is the acceleration calculated from the differential encoder at the motor end, referred to the output end, and then filtered by mean value;

[0036] Then calculate the correction factor c for viscous friction. i :

[0037]

[0038] Where k a It is an adjustment coefficient determined based on actual working conditions;

[0039] Finally, calculate the dynamic friction compensation torque.

[0040]

[0041] Where, k v,i The coefficient of viscous friction of each joint of the robot;

[0042] The friction feedforward compensation torque required by each joint motor is equal to the dynamic friction compensation torque, that is...

[0043]

[0044] Furthermore, when using dynamic friction compensation, reset the static friction compensation timing:

[0045] when When greater than 0, when When less than 0

[0046] Through the above scheme, this application adopts different compensation strategies for different motion states and trends: when the joint is stationary, the motor outputs high-frequency harmonic torque to assist in starting; after the joint begins to move, viscous friction is corrected based on modified acceleration, which significantly improves the deceleration and reversing performance in drag mode. This technology improves the compliance of collaborative robot dragging without installing joint torque sensors, thereby increasing the operator's deployment efficiency.

[0047] Furthermore, the damping at the ends of each link of the robot is calculated, including:

[0048] Set the linear velocity v0 at the end of the damped opening link; set the linear velocity v of the end of the damped link to remain constant. end That is, when the joint velocity exceeds v end The damping torque remains constant.

[0049] Obtain the angular velocity of each joint of the robot at present. Where i represents the i-th link; the joint velocity is converted to the end linear velocity v in the coordinate system of each link through the link length model. act,i ,Right now Where L link,i It is the distance from the end of link i to the current joint axis;

[0050] When the end linear velocity v in each link coordinate system act,i When the absolute value is greater than the linear velocity v0 at the end of the damping opening link, the damping torque of the joint motor is calculated using the following damping & link length model:

[0051]

[0052] Where f damp It is the maximum damping force in each link coordinate system that is opposite to the direction of motion of the link end.

[0053] By employing the aforementioned method, this application calculates the end-effector velocities of each link and determines whether the velocities exceed preset values. If they do, it calculates the corresponding damping torque of the joint and generates a damping torque opposite to the dragging direction via the joint motor. This slows down the joint that moves first, allowing the force to be transmitted to other non-moving joints, thereby driving the other joints to move synchronously. This technology, without the need for an end-effector six-dimensional force sensor, significantly improves the coordination of the robotic arm in joint torque dragging mode, enhances precision positioning, and increases the accuracy and efficiency of operator deployment.

[0054] On another front, this application provides a collaborative robot control device, including...

[0055] The dynamic feedforward compensation calculation module is used to obtain the dynamic feedforward compensation torque of the robot.

[0056] Friction compensation calculation module, used to obtain the friction compensation torque of the robot;

[0057] The robot end effector damping calculation module is used to obtain the damping torque at the ends of each link of the robot.

[0058] The system also includes a data processing module, which sums the obtained dynamic feedforward compensation torque, friction compensation torque, and damping torque and sends them to the corresponding joint motors of the robot.

[0059] In another aspect, this application provides a collaborative robot control device, including a memory and a processor. The memory stores a computer program, and when the computer program is called and executed by the processor, it implements the steps in the collaborative robot drag teaching control method described above.

[0060] In another aspect, this application provides a computer-readable medium storing a computer program, which, when called and executed by a computer, implements the steps in the collaborative robot drag teaching control method described above.

[0061] In summary, the beneficial effects of this application are as follows:

[0062] 1. This invention addresses the issue of laborious starting, stopping, and reversing in collaborative robot dragging mode by developing a segmented static-dynamic friction compensation method. Different compensation strategies are employed for different motion states and trends: when the joint is stationary, the motor outputs high-frequency harmonic torque to assist in starting; after the joint begins to move, viscous friction is corrected based on modified acceleration, significantly improving deceleration and reversing performance in dragging mode. This technology enhances the compliance of collaborative robot dragging without the need for joint torque sensors, thereby increasing operator deployment efficiency.

[0063] 2. To address the issues of uncoordinated and inaccurate positioning in collaborative robot dragging modes, this invention proposes a method of adding damping to the ends of each link in the robot. Based on the calculated end-speed of each link, it is determined whether the speed exceeds a preset value. If it does, the corresponding damping torque of the joint is calculated, and a damping torque opposite to the dragging direction is generated by the joint motor. This slows down the joint that moves first, allowing force to be transmitted to other non-moving joints, thereby driving the other joints to move synchronously. This technology significantly improves the coordination of the robotic arm in joint torque dragging mode without the need for a six-dimensional force sensor at the end effector, improves the accuracy of positioning, and enhances the precision and efficiency of operator deployment.

[0064] 3. This invention does not require external force estimation and does not have high accuracy requirements for dynamic model parameters. Therefore, it has high versatility and can be extended to industrial robot application scenarios with dragging requirements. Attached Figure Description

[0065] Figure 1 This is a schematic diagram of the overall process of this application;

[0066] Figure 2 This is a flowchart illustrating the process of calculating friction compensation in this application;

[0067] Figure 3 This is a flowchart illustrating the calculation of the damping torque of each joint of the robot in this application;

[0068] Figure 4 This is a schematic diagram of the robot joint motor control described in this application. Detailed Implementation

[0069] The specific embodiments of this application are described in detail below with reference to the accompanying drawings.

[0070] Example: Reference Figure 1 This application provides a collaborative robot drag-and-teach control method in one specific embodiment, comprising the following steps:

[0071] (1) Calculate the dynamic feedforward compensation of the robot

[0072] The required dynamic feedforward compensation torque τ of the joint motor is calculated using a drag-and-drop dynamics model. dyn ,include:

[0073] Given the dynamic parameters of the robot, such as the mass, center of mass, and inertia tensor of each moving part, an inverse dynamics model of the robot is established, i.e.

[0074]

[0075] Where, τ dyn =(τ1,τ2,τ3,τ4,τ5,τ6) T Let τ be the equivalent moment vector of the robot's joint space dynamic model. i Represents the equivalent torque of the dynamic model of each joint; Let V be the velocity vector in the robot's joint space. Indicates the rotational speed of each joint; Let be the acceleration vector in the robot's joint space. Let M(q) represent the rotational acceleration of each joint; M(q) is the inertia matrix in the robot's inverse dynamics model. Let G(q) be the Coriolis force and centrifugal force matrix in the robot inverse dynamics model; G(q) is the gravity vector in the robot inverse dynamics model. Here, α is a scaling factor to reduce the negative impact of actual acceleration noise on drag compliance in engineering.

[0076] (2) Calculate the corresponding friction compensation for different motion trends;

[0077] refer to Figure 2 The friction feedforward compensation torque τ required by the joint motor is calculated using a segmented friction model. f ,include:

[0078] Define the frequency f of the static friction compensation function, and calculate the period of the static friction function. and angular frequency ω f =2π*f, define the initial value t of the static friction compensation timer. i =0, defining the dynamic friction compensation start-up speed.

[0079] Update static friction compensation timer t i When it is greater than T, make t i =t i -T, where i represents the i-th joint;

[0080] Given the Coulomb friction coefficients of each joint of the robot, calculate the static friction compensation torque, i.e.

[0081]

[0082] in It is static friction compensation torque, ω f It is the angular frequency of the static friction compensation function, t i It is the time of the static friction compensation timer, k c,i It is the Coulomb friction coefficient of the joint;

[0083] Obtain the joint speed calculated from the output encoder. when Less than the preset dynamic friction compensation start speed At this point, the joint is considered to be in a static state, entering the static friction compensation stage. The friction feedforward compensation torque required by each joint motor is equal to the output of the static friction compensation model, i.e.

[0084]

[0085] when Greater than the preset dynamic friction compensation start speed To ensure the continuity of the friction compensation torque, the following two cases are discussed:

[0086] when Less than 0.9*k c,i At that time, continue to output the calculated value according to the current static friction compensation, that is...

[0087]

[0088] when Greater than 0.9*k c,i When the joint is determined to have entered the dynamic friction compensation stage, the acceleration is first corrected:

[0089]

[0090] in It is the corrected acceleration, a deadZone It is used to correct the dead zone of acceleration. The acceleration is calculated from the differential encoder at the motor end and then referred to the output end, and then filtered by the mean. The reason for not using an encoder at the output end is that the sampling accuracy of the output is low after the motor motion is scaled by the reducer.

[0091] Then calculate the correction factor c for viscous friction. i :

[0092]

[0093] Where k a It is an adjustment coefficient determined based on actual working conditions.

[0094] Finally, given the known Coulomb friction coefficient and viscous friction coefficient k of each joint of the robot... v,iCalculate the dynamic friction compensation torque under the following conditions.

[0095]

[0096] It can be seen that when velocity and acceleration are in the same direction, c i =1, the dynamic friction compensation model is the same as the Coulomb viscous friction model. When the velocity and acceleration are in opposite directions, and the acceleration exceeds the acceleration dead zone, it is determined to be a deceleration or reversal phase. Viscous friction is corrected to reduce friction compensation in the direction of motion, even when the absolute value of the corrected acceleration is large. i <0, the motor compensates for the torque opposite to the direction of motion, thereby achieving rapid deceleration and reversal.

[0097] When using dynamic friction compensation, the static friction compensation timing needs to be reset to ensure the continuity of output torque when switching from the dynamic friction model to the static friction model; when When greater than 0, when When less than 0

[0098] Finally, output the calculated value τ of the friction compensation model. f,i .

[0099] (3) Calculate the damping at the ends of each link of the robot.

[0100] refer to Figure 3 The damping torque of each joint is calculated using a damping model, including:

[0101] Set the linear velocity v0 of the link end when damping is engaged; set the linear velocity v of the link end when damping is maintained constant. end That is, when the joint velocity exceeds v end The damping torque remains constant.

[0102] Get the current angular velocity of each joint. Where i represents the i-th link; the joint velocity is converted to the end linear velocity v in the coordinate system of each link through the link length model. act,i ,Right now Where L link,i It is the distance from the end of link i to the current joint axis;

[0103] When the end linear velocity v in each link coordinate system act,i When the absolute value of the damping torque is greater than the preset damping opening linear velocity v0, the damping torque of the joint motor is calculated using the following damping & rod length model:

[0104]

[0105] Where f dampIt is the maximum damping force opposite to the direction of motion of the link end in each link's coordinate system. That is, when the linear velocity v of each link end in the current coordinate system... act,i When v is positive and equal to v0, the damping torque is 0; act,i Greater than or equal to v end At that time, the damping torque is -f damp *L link,i When v act,i At v0 and v end When the time interval is between, the end damping is calculated proportionally, and v act,i The larger the value, the greater the damping torque. Similarly, when the linear velocity v at the end of each link in the current coordinate system... act,i When v is negative and equal to -v0, the damping torque is 0; act,i Less than or equal to -v end At that time, the damping torque is f damp *L link,i When v act,i In -v end When the value is between -v0 and -v0, the end damping is calculated proportionally, and v act,i The smaller the damping, the greater the torque.

[0106] (4) As attached Figure 4 The damping torque τ of each joint is shown. damp,i Dynamic feedforward torque τ dyn,i With friction compensation feedforward torque τ f,i The summation is used to determine the desired torque of the joint motor.

[0107] The above steps are explained in detail below using a typical six-degree-of-freedom robot. The robot used as an example is shown in the patent with patent publication number CN308343078S, and has a mass of 10kg.

[0108] (1) Calculate the required dynamic feedforward compensation torque τ of the joint motor through the drag dynamic model. dyn ,include:

[0109] The dynamic parameters of each moving part of the robot, such as mass, center of mass, and inertia tensor, are obtained in advance. These parameters can be exported using third-party numerical modeling software or obtained through dynamic parameter identification algorithms.

[0110] Calculate the dynamic feedforward compensation torque using the robot's inverse dynamics model:

[0111]

[0112] Where, τ dyn =(τ1,τ2,τ3,τ4,τ5,τ6) T Let τ be the equivalent moment vector of the robot's joint space dynamic model. iRepresents the equivalent torque of the dynamic model of each joint; Let V be the velocity vector in the robot's joint space. Indicates the rotational speed of each joint; Let be the acceleration vector in the robot's joint space. Let M(q) represent the rotational acceleration of each joint; M(q) is the inertia matrix in the robot's inverse dynamics model. G(q) represents the Coriolis force and centrifugal force matrices in the robot inverse dynamics model; G(q) represents the gravity vector in the robot inverse dynamics model, where α is a scaling factor to reduce the negative impact of actual acceleration noise on drag compliance in engineering, usually set within 20%.

[0113] (2) Calculate the friction feedforward compensation torque τ required by the joint motor using a friction model. f ,include:

[0114] Define the frequency of the static friction compensation function as f = 50Hz, and calculate the period of the static friction function. and angular frequency ω f =2π*f=314, define the initial value t of the static friction compensation timer. i =0, defining the dynamic friction compensation start-up speed.

[0115] Update static friction compensation timer t i When it is greater than T, t i =t i -T, where i represents the i-th joint;

[0116] Given the Coulomb friction coefficients of each joint of the robot, calculate the static friction compensation torque, i.e.

[0117]

[0118] in This is static friction compensation torque. To ensure that the motor does not move under this compensation, k c,i It is recommended to define ω as 80-90% of the actual value. f It is the angular frequency of the static friction compensation function, t i It is the time of the static friction compensation timer, k c,i It is the Coulomb friction coefficient of the joint;

[0119] Get the current joint velocity when When the dynamic friction compensation start-up speed is less than the preset speed of 0.02 rad / s, the friction feedforward compensation torque required by each joint motor is equal to the output of the static friction compensation model, that is...

[0120]

[0121] when When the dynamic friction compensation start speed is greater than the preset 0.02 rad / s, the following two cases are discussed to ensure the continuity of the friction compensation torque:

[0122] when Less than 0.9*k c,i At that time, continue to output the calculated value according to the current static friction compensation, that is...

[0123]

[0124] when Greater than 0.9*k c,i When the joint is determined to have entered the dynamic friction compensation stage, the acceleration is first corrected:

[0125]

[0126] in It is the corrected acceleration; It is the acceleration calculated from the differential encoder at the motor end, referred to the output end, and then filtered by mean; a deadZone It is used to correct the dead zone of acceleration. The setting depends on the quality of the filtered acceleration signal and generally needs to be greater than 1 rad / s. 2 , .

[0127] Then calculate the correction factor c for viscous friction. i :

[0128]

[0129] Where k a It is an adjustment coefficient determined based on actual working conditions.

[0130] Finally, given the known Coulomb friction coefficient and viscous friction coefficient k of each joint of the robot... v,i Calculate the dynamic friction compensation torque under the following conditions.

[0131]

[0132] It can be seen that when velocity and acceleration are in the same direction, c i =1, the dynamic friction compensation model is the same as the Coulomb viscous friction model. When the velocity and acceleration are in opposite directions, and the acceleration exceeds the acceleration dead zone, it is determined to be a deceleration or reversal phase. Viscous friction is corrected to reduce friction compensation in the direction of motion, even when the corrected acceleration is large. i <0, compensating for torque opposite to the direction of motion, thereby achieving rapid deceleration or reversal.

[0133] When using dynamic friction compensation, the static friction compensation timing needs to be reset to ensure the continuity of output torque when switching from the dynamic friction model to the static friction model; when When greater than 0, when When less than 0

[0134] Output the calculated value τ of the friction compensation model f,i ;

[0135] (3) Calculate the damping torque of each joint using a damping model, including:

[0136] Set the linear velocity of the link end when damping is engaged to v0 = 0.01 m / s; set the linear velocity of the link end when damping is maintained constant to v end =0.1m / s, meaning when the joint velocity exceeds v end The damping torque remains constant.

[0137] Get the current angular velocity of each joint. Where i represents the i-th link; and the joint velocity is converted to the end linear velocity v in the coordinate system of each link through the link length model. act,i ,Right now Where L link,i It is the distance from the end of link i to the current joint axis;

[0138] When the end linear velocity v in each link coordinate system act,i When the absolute value of the damping torque is greater than the preset damping opening linear velocity v0, the damping torque of the joint motor is calculated using the following damping & rod length model:

[0139]

[0140] Where f damp =10N is the maximum damping force opposite to the direction of motion of the link end in each link's coordinate system. For example, when the length of link 2 is 0.72m, and the velocity of its end in the current coordinate system reaches 0.05m / s, τ damp = -3.2 Nm;

[0141] The damping torque τ damp,i Unlike other feedforward compensations, τ damp As resistance, the rotational speed of joint 2 is reduced, preventing the end force from being dissipated due to the movement of joint 2, so that the force can be transmitted more evenly to other non-moving joints, driving the other joints to move simultaneously.

[0142] (4) Finally, the damping torque τ of each joint is calculated. damp,i Dynamic feedforward torque τ dyn,i With friction compensation feedforward torque τ f,iThe summation is used to determine the desired torque of the joint motor.

[0143] Another specific embodiment of this application provides a collaborative robot control device, including...

[0144] The dynamic feedforward compensation calculation module is used to obtain the dynamic feedforward compensation torque of the robot.

[0145] Friction compensation calculation module, used to obtain the friction compensation torque of the robot;

[0146] The robot end effector damping calculation module is used to obtain the damping torque at the ends of each link of the robot.

[0147] The system also includes a data processing module, which sums the obtained dynamic feedforward compensation torque, friction compensation torque, and damping torque and sends them to the corresponding joint motors of the robot.

[0148] Another specific embodiment of this application provides a collaborative robot control device, including a memory and a processor. The memory stores a computer program, and when the computer program is called and executed by the processor, it implements the steps in the collaborative robot drag teaching control method described above.

[0149] Another specific embodiment of this application provides a computer-readable medium storing a computer program, which, when called and executed by a computer, implements the steps in the collaborative robot drag teaching control method described above.

[0150] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the inventive concept of this application, and these all fall within the protection scope of this application.

Claims

1. A method for drag-and-drop teaching control of a collaborative robot, characterized in that, Includes the following steps: Computational robot dynamics feedforward compensation; Calculate the corresponding friction compensation for different motion trends; Calculate the damping at the ends of each link of the robot; The dynamic feedforward compensation torque, friction compensation torque, and damping torque are added together and sent to the joint motor as the desired torque. The calculation of friction compensation includes: Configure dynamic friction compensation start speed Obtain the joint velocities of the robot ; When the absolute value of the robot's joint velocity Less than the dynamic friction compensation start-up speed At this time, static friction compensation is used for friction compensation; When the absolute value of the robot's joint velocity Greater than the dynamic friction compensation start-up speed And the static friction compensation torque is less than hour, It is the Coulomb friction coefficient of the joint, output according to the calculated value of the current static friction compensation; When the absolute value of the robot's joint velocity Greater than the dynamic friction compensation start-up speed And the static friction compensation torque is greater than When the joint is determined to enter the dynamic friction compensation stage, the friction feedforward compensation torque required by the motors of each joint is equal to the dynamic friction compensation torque.

2. The collaborative robot drag-and-teach control method according to claim 1, characterized in that, The dynamics feedforward compensation for computational robots includes: Establish the robot's inverse dynamics model and output the dynamic feedforward compensation torque; The inverse dynamics model of the robot is as follows: ； in, This is the equivalent moment vector of the dynamic model of the robot's joint space. Represents the equivalent torque of the dynamic model of each joint; Let V be the velocity vector in the robot's joint space. Indicates the rotational speed of each joint; Let be the acceleration vector in the robot's joint space. This indicates the rotational acceleration of each joint; The inertia matrix in the robot's inverse dynamics model. The Coriolis force and centrifugal force matrices in the robot inverse dynamics model; This represents the gravity vector in the robot's inverse dynamics model. in It is the scaling factor.

3. The collaborative robot drag-and-teach control method according to claim 1, characterized in that, When the absolute value of the robot's joint velocity Less than the dynamic friction compensation start-up speed At this time, static friction compensation is used for friction compensation: Define the frequency of the static friction compensation function Calculate the period of the static friction function. Circular frequency Define the initial value of the static friction compensation timer. ; Update static friction compensation timer When it is greater than At that time, ,in Indicates the first One joint; Calculate the static friction compensation torque, i.e. ； in It is static friction compensation torque. It is the angular frequency of the static friction compensation function. It is the time of the static friction compensation timer. It is the Coulomb friction coefficient of the joint; The friction feedforward compensation torque required by each joint motor is equal to the static friction compensation torque, that is... 。 4. The collaborative robot drag-and-teach control method according to claim 3, characterized in that, When the absolute value of the robot's joint velocity Greater than the dynamic friction compensation start-up speed and Less than At that time, the output will be based on the calculated value of the current static friction compensation, i.e. ； When the absolute value of the robot's joint velocity Greater than the dynamic friction compensation start-up speed and Greater than When the joint is determined to have entered the dynamic friction compensation stage, the acceleration is first corrected: ； in It is the corrected acceleration. It is used to correct the dead zone of acceleration. It is the acceleration calculated from the differential encoder at the motor end, referred to the output end, and filtered by mean value; Then calculate the correction factor for viscous friction. : ； in It is an adjustment coefficient determined based on actual working conditions; Finally, calculate the dynamic friction compensation torque. ； in, The coefficient of viscous friction of each joint of the robot; The friction feedforward compensation torque required by each joint motor is equal to the dynamic friction compensation torque, that is... 。 5. The collaborative robot drag-and-teach control method according to claim 4, characterized in that, When using dynamic friction compensation, reset the static friction compensation timing: when When greater than 0, ,when When less than 0 .

6. The collaborative robot drag-and-teach control method according to claim 1, characterized in that, The damping at the ends of each link in the robot is calculated, including: Set the linear velocity at the end of the damping opening link. Set damping to maintain a constant linear velocity at the end of the link. That is, when the joint speed exceeds The damping torque remains constant. Obtain the angular velocity of each joint of the robot at present. ,in Indicates the first Each link; the joint velocity is converted to the end linear velocity in the coordinate system of each link through the link length model; ,Right now ,in yes The distance from the end of the link to the current joint axis; When the end linear velocity in each link coordinate system The absolute value is greater than the linear velocity at the end of the damping opening link. At that time, the damping torque of the joint motor is calculated using the following damping & rod length model: ; in It is the maximum damping force in each link coordinate system that is opposite to the direction of motion of the link end.

7. A collaborative robot control device, used to implement the steps in the collaborative robot drag teaching control method as described in any one of claims 1-6, characterized in that, include The dynamic feedforward compensation calculation module is used to obtain the dynamic feedforward compensation torque of the robot. Friction compensation calculation module, used to obtain the friction compensation torque of the robot; The robot end effector damping calculation module is used to obtain the damping torque at the ends of each link of the robot. The system also includes a data processing module, which sums the obtained dynamic feedforward compensation torque, friction compensation torque, and damping torque and sends them to the corresponding joint motors of the robot.

8. A collaborative robot control device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is called and executed by the processor, it implements the steps in the collaborative robot drag teaching control method as described in any one of claims 1-6.

9. A computer-readable medium, characterized in that, The computer-readable medium stores a computer program, which, when called and executed by a computer, implements the steps of the collaborative robot drag teaching control method as described in any one of claims 1-6.

Images