A robot joint friction identification method, a robot system, and an operating method

Abstract

The application discloses a robot joint friction identification method, a robot system and an operation method, and relates to the technical field of industrial robots. The friction identification method comprises the following steps: detecting the actual driving torque of the driving motor of each joint of a robot; performing constant-speed tracking experiments on each joint of the robot, and detecting the actual friction data of each joint in combination with the actual driving torque; obtaining model friction data of each joint of the robot by using a Stribeck friction model; and comparing the actual friction data of each joint of the robot with the model friction data, and evaluating the robot joint friction identification accuracy. The use of the Stribeck friction model to obtain the model friction data of each joint of the robot can ensure the accuracy and applicability of the model friction data. Then, the actual friction data of each joint of the robot is compared with the model friction data, and the robot joint friction identification accuracy is evaluated.

Description

Technical Field

[0001] This invention relates to the field of industrial robot technology, and in particular to a method for identifying robot joint friction, a robot system, and an operating method. Background Technology

[0002] In the industrial manufacturing sector, where robots are widely used, existing industrial robot arms typically operate in structured environments, repeatedly performing specified motion tasks. To improve the efficiency and quality of production tasks, industrial robotic arms possess characteristics of high precision, high speed, and high load capacity.

[0003] However, during the use of robots, due to the large number of joints in the robotic arm of an industrial robot, friction occurs at the joints, which reduces the robot's operational accuracy. Furthermore, as the robot's operating time increases, the joint friction becomes more severe, significantly reducing the robot's accuracy in tracking command trajectories. Summary of the Invention

[0004] (I) Purpose of the Invention

[0005] The purpose of this invention is to provide a robot joint friction identification method, robot system, and operation method to solve the problem that frictional resistance in existing robot joints severely reduces the robot's tracking accuracy of command trajectories.

[0006] (II) Technical Solution

[0007] To address the above problems, this invention provides a method for identifying robot joint friction, comprising:

[0008] Detect the actual driving torque of the drive motors of each joint of the robot;

[0009] A constant speed tracking experiment was conducted on each joint of the robot, and the actual friction data of each joint was detected in combination with the actual driving torque.

[0010] The model friction data of each joint of the robot were obtained by using the Stribeck friction model;

[0011] The actual friction data of each joint of the robot is compared with the model friction data to evaluate the joint friction identification accuracy of the robot.

[0012] Optionally, detecting the actual driving torque of the drive motors for each joint of the robot includes:

[0013] Sample the current signal of the current loop of the servo driver of the drive motor of each joint of the robot.

[0014] The actual driving torque of the drive motor is obtained based on the current signal of the current loop of the servo driver of the drive motor and the torque coefficient of the drive motor.

[0015] Optionally, a constant-speed tracking experiment is performed on each joint of the robot, and the actual friction data of each joint is detected in conjunction with the actual driving torque, including:

[0016] A unidirectional constant speed tracking experiment was conducted on multiple first-type joints one by one, and the actual friction data of the first-type joints in the state where the axis is parallel to the direction of gravity was obtained according to the corresponding actual driving torque.

[0017] A forward and reverse constant speed tracking experiment was conducted on multiple second-type joints one by one, and the actual friction data of the second-type joints were obtained based on the corresponding actual driving torque.

[0018] The first type of joint has its axis that can rotate to a state parallel to the direction of gravity, while the second type of joint has its axis that cannot rotate to a state parallel to the direction of gravity.

[0019] Optionally, performing forward and reverse constant speed tracking experiments on multiple second-type joints one by one, and obtaining the actual friction data of the second-type joints based on the corresponding actual driving torque includes:

[0020] Use the first formula: Obtain the actual friction data of the second type of joint;

[0021] in, u1 is the frictional torque of the second type of joint, u2 is the actual driving torque of the second type of joint during the forward constant speed tracking experiment, and u2 is the actual driving torque of the second type of joint during the reverse constant speed tracking experiment. The actual frictional data of the second type of joint includes the frictional torque of the second type of joint.

[0022] Optionally, the model friction data of each joint of the robot calculated using the Stribeck friction model includes:

[0023] The Stribeck friction model was fitted using the least squares method to the actual friction data of each joint. The Stribeck friction model formula is as follows:

[0024]

[0025] Among them, f c For Coulomb friction, f s For the maximum static friction force, v sLet ξ be the Stribeck velocity, sgn(·) be the sign function, v be the relative velocity between the two contact surfaces of the joint, σ be the coefficient of viscous friction, and F be the relative velocity between the two contact surfaces of the joint. e For external force, f c f s v s σ and σ are the parameters that need to be identified.

[0026] Optionally, the robot's joints are RV-driven joints, and the empirical parameter ξ is set to 1;

[0027] The robot's joints are harmonic-driven joints, and the empirical parameter ξ is set to 2.

[0028] In addition, the present invention also provides a robot system, the system comprising a multi-joint robot and a computer-readable storage medium, wherein the multi-joint robot is connected to the computer-readable storage medium.

[0029] The computer-readable storage medium stores a computer program, which is read and executed by a processor to implement the robot joint friction identification method described above.

[0030] In addition, the present invention also provides an operation method for controlling the robot system, the operation method comprising:

[0031] calibrate the specified limit points for designated joints of the robot;

[0032] Move the robot to the designated limit point;

[0033] To enable the processor to read and execute computer programs;

[0034] After the computer program finishes running, it is checked whether there was a pause during the running of the computer program. If there was a pause, the robot is manually moved to the starting point of the pause line and the computer program is continued to run.

[0035] (III) Beneficial Effects

[0036] The actual driving torque of the drive motors at each joint of the robot is detected, and the actual friction data of each joint is calculated in conjunction with this actual driving torque. This fully utilizes the close correlation between the actual driving torque and the actual friction force, thus ensuring the accuracy of the detected friction data. The Stribeck friction model, as a typical representative of static friction models, can describe the changing trend of friction force at low speeds. Using the Stribeck model for parameter identification and adding compensation terms based on this model fully meets the needs of the robot system. Therefore, using the Stribeck friction model to calculate the model friction data for each joint of the robot ensures the accuracy and applicability of the model friction data. Then, the actual friction data of each joint of the robot is compared with the model friction data to evaluate the accuracy of the robot joint friction identification. This achieves feedforward compensation for robot joint friction, ensuring the accuracy of robot joint friction identification and improving the tracking accuracy of the robot system for command trajectories. Attached Figure Description

[0037] Figure 1 This is a schematic flowchart illustrating the main steps of the robot joint friction identification method according to a specific embodiment of the present invention;

[0038] Figure 2 This is a schematic diagram of the Stribeck friction model used in the robot joint friction identification method according to a specific embodiment of the present invention;

[0039] Figure 3 This is a schematic flowchart illustrating the specific execution steps of the operation method according to a specific embodiment of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0041] The accompanying drawings illustrate a layer structure according to an embodiment of the present invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0042] Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0043] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0044] The invention will now be described in more detail with reference to the accompanying drawings. In the various drawings, the same elements are indicated by similar reference numerals. For clarity, the various parts in the drawings are not drawn to scale.

[0045] For existing components that do not involve the improvements of this invention, they will be briefly described or not described at all, while the focus will be on describing the components that have been improved relative to the prior art.

[0046] See Figure 1 and Figure 2 This embodiment provides a method for identifying robot joint friction, including:

[0047] Detect the actual driving torque of the drive motors of each joint of the robot;

[0048] A constant speed tracking experiment was conducted on each joint of the robot, and the actual friction data of each joint was detected in combination with the actual driving torque.

[0049] The model friction data of each joint of the robot were obtained by using the Stribeck friction model;

[0050] The actual friction data of each joint of the robot is compared with the model friction data to evaluate the joint friction identification accuracy of the robot.

[0051] It should be noted that the actual friction data here can include frictional torque and the corresponding angular velocity, from which the corresponding actual Coulomb friction force, maximum static friction force, and Stribeck velocity can be obtained. The Coulomb friction force, maximum static friction force, and Stribeck velocity can be used as identification parameters. The Stribeck velocity here can be obtained through identification. It should also be noted that the frictional torque here can be calculated using the least squares method based on the model.

[0052] The actual driving torque of the drive motors at each joint of the robot is detected, and the actual friction data of each joint is calculated in conjunction with this actual driving torque. This fully utilizes the close correlation between the actual driving torque and the actual friction force, thus ensuring the accuracy of the detected friction data. The Stribeck friction model, as a typical representative of static friction models, can describe the changing trend of friction force at low speeds. Using the Stribeck model for parameter identification and adding compensation terms based on this model fully meets the needs of the robot system. Therefore, using the Stribeck friction model to calculate the model friction data for each joint of the robot ensures the accuracy and applicability of the model friction data. Then, the actual friction data of each joint of the robot is compared with the model friction data to evaluate the accuracy of the robot joint friction identification. This achieves feedforward compensation for robot joint friction, ensuring the accuracy of robot joint friction identification and improving the tracking accuracy of the robot system for command trajectories.

[0053] In addition, the accuracy of the robot joint friction identification can be evaluated by the maximum error, relative error, and root mean square; the smaller these values ​​are, the higher the accuracy.

[0054] Furthermore, analysis revealed that, regardless of whether it is an RV-driven joint or a harmonic-driven joint, if various complex dynamic characteristics are disregarded, such as fluctuations in load torque and angular position-related fluctuations in harmonic drives, the relationship between friction and angular velocity tends to approximate the Stribeck curve, as follows: Figure 2 As shown, this is also an important basis for choosing this model.

[0055] See Figure 1 and Figure 2 Furthermore, detecting the actual driving torque of the drive motors for each joint of the robot includes:

[0056] Sample the current signal of the current loop of the servo driver of the drive motor of each joint of the robot.

[0057] The actual driving torque of the drive motor is obtained based on the current signal of the current loop of the servo driver of the drive motor and the torque coefficient of the drive motor.

[0058] Measuring frictional torque requires measuring the torque signal output by the motor. This signal is obtained by sampling the current signal in the current loop of the motor servo driver and then multiplying it by the motor's torque coefficient. The robot platform's system software can provide a real-time signal acquisition module, allowing for pre-set sampling periods to sample motor output torque, speed, and other signals in real time. The sampled data can be used directly or filtered.

[0059] See Figure 1 and Figure 2 Furthermore, a constant-speed tracking experiment is performed on each joint of the robot, and the actual friction data of each joint is detected in conjunction with the actual driving torque, including:

[0060] A unidirectional constant speed tracking experiment was conducted on multiple first-type joints one by one, and the actual friction data of the first-type joints in the state where the axis is parallel to the direction of gravity was obtained according to the corresponding actual driving torque.

[0061] A forward and reverse constant speed tracking experiment was conducted on multiple second-type joints one by one, and the actual friction data of the second-type joints were obtained based on the corresponding actual driving torque.

[0062] The first type of joint has its axis that can rotate to a state parallel to the direction of gravity, while the second type of joint has its axis that cannot rotate to a state parallel to the direction of gravity.

[0063] See Figure 1 and Figure 2 Furthermore, a forward and reverse constant speed tracking experiment was conducted on multiple second-type joints one by one, and the actual friction data of the second-type joints were obtained based on the corresponding actual driving torque, including:

[0064] Use the first formula: Obtain the actual friction data of the second type of joint;

[0065] in, u1 is the frictional torque of the second type of joint, u2 is the actual driving torque of the second type of joint during the forward constant speed tracking experiment, and u2 is the actual driving torque of the second type of joint during the reverse constant speed tracking experiment. The actual frictional data of the second type of joint includes the frictional torque of the second type of joint.

[0066] For the first type of joint, if its axis is parallel to the direction of gravity, when the first type of joint moves at a constant speed (the other joints are locked), the inertial force is zero, and the actual driving torque, that is, the input torque, is equal to the frictional torque (the other joints are locked at the initial angle).

[0067] For the second type of joint, since it is impossible to adjust to a position unaffected by gravity, its constant-speed movement requires overcoming both frictional torque and gravitational torque. To extract frictional data, the following steps can be taken:

[0068] Experiment 1) The second type of joint is set to rotate clockwise by an angle θ;

[0069] Experiment 2) The second type of joint was set to rotate counterclockwise by an angle θ.

[0070] That is, the joint performs a reciprocating periodic rotation. If the input torque of the second type of joint in the aforementioned experiment 1) is:

[0071]

[0072] (1) In the formula, τ g (θ) is the gravitational torque. This is the frictional torque. In experiment 2), the direction of motion is opposite to that in 1), and the input torque satisfies:

[0073]

[0074] For most robots, the difference between forward and reverse friction is small, therefore the following equation holds:

[0075]

[0076] The frictional torque can be calculated as follows:

[0077]

[0078] Therefore, based on the sampling data in the experiment, a set of friction torque sequences under angular velocity can be obtained from the above equation (4).

[0079] Based on this, the specific method for conducting constant velocity tracking experiments on each joint of the robot can be:

[0080] For robots used in industry, due to the use of high-precision RV and harmonic reducer transmission, the difference between forward and reverse friction is very small. Therefore, only the friction torque when the speed is positive is measured here, and the friction torque when the speed is negative is its opposite, in order to improve the efficiency of identification work.

[0081] The following explanation uses a 6-joint robot as an example.

[0082] 1) Joints 6, 4, and 1 are the first type of joints, and constant speed tracking experiments are conducted. Starting with joint 6, the robot is initially in its initial state. Joint 6's axis is aligned with the direction of gravity by rotating joint 5. Then, a constant speed tracking experiment is performed, locking the remaining joints. On the motor side of joint 6, the rotational speed is set to [0, maximum speed * 80%] r / min. At least 30 sets of constant speed tracking experiments at different speeds are conducted, with each set of speeds representing 5-10 cycles of joint movement. Considering the significant nonlinearity of friction at low speeds, a denser sampling rate (more than half of the total number of sets) is taken within the [0, maximum speed * 10%] r / min speed range. The sampling period is set to 20ms, and the average value of the sampled torque data at each fixed speed is calculated to obtain the correspondence between speed and friction torque. The operation of joint 4 is the same as that of joint 6. Since joint 1 has a lower speed, the number of experimental sets can be adjusted appropriately, requiring at least 15 sets of tracking experiments at different speeds. After the operation is completed, the robot returns to its initial state.

[0083] 2) Joints 5, 3, and 2 cannot rotate continuously in one direction, which is the second type of joint, and the influence of gravity must be eliminated. A reciprocating constant-speed tracking experiment is conducted on this type of joint. As mentioned in the previous section, gravity can be eliminated by performing forward and reverse rotation operations, allowing for the calculation of friction. First, for joint 5, the other joints are locked. Within the rotational speed range [0, maximum speed * 80%] r / min, 30 sets of reciprocating constant-speed tracking experiments are performed at different speeds, with more frequent speed values ​​in the low-speed range. At each rotational speed, the joint reciprocates for 5-10 cycles between -120° and 120°. According to the first formula: Torque data of clockwise and counterclockwise motion at a fixed speed can be collected, and the average values ​​calculated by combining equations (1) to (4) can be subtracted to obtain the friction torque of the joint, which is the friction torque of the second type of joint, thus obtaining the correspondence between speed and friction force. Since the speeds of joints 3 and 2 are relatively small, the number of experimental groups can be adjusted appropriately, and more than 15 sets of tracking experiments at different speeds can be carried out. After the operation is completed, the robot returns to its initial state.

[0084] See Figure 1 and Figure 2 Furthermore, the model friction data of each joint of the robot calculated using the Stribeck friction model includes:

[0085] The Stribeck friction model was fitted using the least squares method to the actual friction data of each joint. The Stribeck friction model formula is as follows:

[0086]

[0087] Among them, f c For Coulomb friction, f sFor the maximum static friction force, v s Let ξ be the Stribeck velocity, sgn(·) be the sign function, v be the relative velocity between the two contact surfaces of the joint, σ be the coefficient of viscous friction, and F be the relative velocity between the two contact surfaces of the joint. e For external force, f c f s v s σ and σ are the parameters that need to be identified.

[0088] The data points show that the relationship between friction and angular velocity is reflected in the Stribeck curve, and the parameters can be obtained by fitting an exponential curve using the least squares method. Stribeck velocity v s The curvature of the friction curve is affected, and the empirical parameter ξ affects the decay rate of the exponential decay segment. The empirical parameter is a known value.

[0089] Preferably, the robot's joints are RV-driven joints with an empirical parameter ξ of 1; the robot's joints are harmonic-driven joints with an empirical parameter ξ of 2.

[0090] Furthermore, the complete expansion of the Stribeck friction model is as follows:

[0091]

[0092] In addition, this embodiment also provides a robot system, which includes a multi-joint robot and a computer-readable storage medium, wherein the multi-joint robot is connected to the computer-readable storage medium.

[0093] The computer-readable storage medium stores a computer program, which is read and executed by a processor to implement the robot joint friction identification method.

[0094] Since the technical effects achieved by this robot system are the same as those achieved by the robot joint friction identification method, the robot system will not be explained further.

[0095] See Figures 1 to 3 In addition, this embodiment also provides an operation method for controlling the robot system, the operation method including:

[0096] calibrate the specified limit points for designated joints of the robot;

[0097] Move the robot to the designated limit point;

[0098] To enable the processor to read and execute computer programs;

[0099] After the computer program finishes running, check if there was a pause during the program's execution. If there was a pause, manually move the robot to the starting point of the pause line and continue running the computer program.

[0100] To address the issue of inaccurate detection caused by robot joints pausing during the execution of the computer program, a method is proposed that detects when a pause occurs during program execution and manually moves the robot to the starting point of the paused line to resume program operation. This eliminates the impact of pauses during program execution on detection accuracy.

[0101] Specifically, taking the identification of frictional force at joint 1 as an example, such as Figure 3 As shown:

[0102] 1. Open the pre-edited motion program J1frictioncal_err. This is the program containing the robot joint friction identification method described above.

[0103] 2. LP0 is calibrated as 95% of the positive limit of joint 1, and LP1 is 95% of the negative limit of joint 1.

[0104] 3. Move the robot to LP1.

[0105] 4. On the friction identification interface of the robot system's teach pendant, click the "Data Acquisition" button to notify the robot system's controller to start data acquisition. After the data acquisition switch is turned on, the controller will acquire data during the uniform speed segment of joint movement in reproduction mode.

[0106] 5. Use the teach pendant to exit the friction identification interface, switch to reproduction mode, and run the program.

[0107] 6. Check for pauses before the program ends. If a pause occurs, switch to teach mode, manually move the robot to the starting point of the paused line, switch to playback mode, and continue running (do not switch line numbers during this process, as switching line numbers will result in incorrect data collection and inaccurate friction force identification).

[0108] 7. Return to the friction identification interface, click "Start Identification" to calculate the value of friction. If the calculation is successful, the calculated value will be displayed on the interface. If the calculation fails, a failure message will be displayed.

[0109] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. A method for identifying joint friction in a robot, characterized in that: include: Detect the actual driving torque of the drive motors of each joint of the robot; A constant speed tracking experiment was conducted on each joint of the robot, and the actual friction data of each joint was detected in combination with the actual driving torque. The model friction data of each joint of the robot were obtained by using the Stribeck friction model; The actual friction data of each joint of the robot is compared with the model friction data to evaluate the joint friction identification accuracy of the robot.

2. The robot joint friction identification method according to claim 1, characterized in that, The actual driving torque of the drive motors for each joint of the robot is detected, including: Sample the current signal of the current loop of the servo driver of the drive motor of each joint of the robot. The actual driving torque of the drive motor is obtained based on the current signal of the current loop of the servo driver of the drive motor and the torque coefficient of the drive motor.

3. The robot joint friction identification method according to claim 1, characterized in that, The constant-speed tracking experiment was performed on each joint of the robot, and the actual friction data of each joint was detected in combination with the actual driving torque, including: A unidirectional constant speed tracking experiment was conducted on multiple first-type joints one by one, and the actual friction data of the first-type joints in the state where the axis is parallel to the direction of gravity was obtained according to the corresponding actual driving torque. A forward and reverse constant speed tracking experiment was conducted on multiple second-type joints one by one, and the actual friction data of the second-type joints were obtained based on the corresponding actual driving torque. The first type of joint has its axis that can rotate to a state parallel to the direction of gravity, while the second type of joint has its axis that cannot rotate to a state parallel to the direction of gravity.

4. The robot joint friction identification method according to claim 3, characterized in that, A forward and reverse constant speed tracking experiment was conducted on multiple second-type joints one by one, and the actual friction data of the second-type joints were obtained based on the corresponding actual driving torque, including: Use the first formula: Obtain the actual friction data of the second type of joint; in, This refers to the frictional torque of the second type of joint. This refers to the actual driving torque during the second type of joint forward constant velocity tracking experiment. The actual driving torque during the reverse constant speed tracking experiment of the second type of joint is the actual friction data of the second type of joint, which includes the friction torque of the second type of joint.

5. The robot joint friction identification method according to claim 3, characterized in that, The friction data of each joint of the robot, calculated using the Stribeck friction model, includes: The Stribeck friction model was fitted using the least squares method to the actual friction data of each joint. The Stribeck friction model formula is as follows: in, Coulomb friction, For maximum static friction, For Stribeck speed, These are empirical parameters. For symbolic functions, Let be the relative velocity of the two contact surfaces of the joint. The coefficient of viscous friction is... , , and These are the parameters that need to be identified.

6. The robot joint friction identification method according to claim 5, characterized in that, The robot's joints are RV-driven joints, based on empirical parameters. The value is 1; The robot's joints are harmonic-driven joints, based on empirical parameters. The value is 2.

7. A robot system, characterized in that, The system includes a multi-jointed robot and a computer-readable storage medium, the multi-jointed robot being connected to the computer-readable storage medium. The computer-readable storage medium stores a computer program, which is read and executed by a processor to implement the robot joint friction identification method as described in any one of claims 1-6.

8. An operating method, characterized in that, The method for operating the robot system of claim 7 includes: calibrate the specified limit points for designated joints of the robot; Move the robot to the designated limit point; To enable the processor to read and execute computer programs; After the computer program finishes running, it is checked whether there was a pause during the running of the computer program. If there was a pause, the robot is manually moved to the starting point of the pause line and the computer program is continued to run.

Images