A robot joint absolute accuracy compensation control method, system and robot

Abstract

The application relates to the field of robots, in particular to a robot joint absolute accuracy compensation control method, a system and a robot. The method comprises the following steps: obtaining an expected torque of a robot joint at a current posture; based on the expected torque, querying corresponding joint stiffness from a calibrated stiffness characteristic relationship, the stiffness characteristic relationship comprising corresponding joint stiffnesses under a plurality of different expected torques; calculating a joint compensation coefficient according to the joint stiffness corresponding to the expected torque; calculating a stiffness compensation value through the joint compensation coefficient and the expected torque; and outputting a compensation instruction for controlling the output angle of the robot joint according to the stiffness compensation value. The compensation mode provided is based on the physical properties of the transmission chain of the robot joint, and can be applied within the load range of the robot joint after calibration and is not affected by load changes. The entire compensation process only depends on the existing expected torque information in the processor, and does not require additional sensors on the joint output side.

Description

Technical Field

[0001] This invention relates to the field of robotics, and more particularly to a method, system, and robot for absolute precision compensation control of robot joints. Background Technology

[0002] Robotic or robotic arm joints typically integrate speed reducers. When a joint is subjected to external loads, the stiffness of the speed reducer and other transmission components can cause angular deviations. For robotic arms composed of multiple joints, these angular deviations accumulate over time, leading to a significant decrease in the absolute positioning accuracy of the robotic arm's end effector.

[0003] To address the aforementioned issues, one existing solution involves adding a high-precision sensor to the output side of the joint to directly measure the actual angle at the joint's output end, and then using feedback data from the encoder on the motor side for closed-loop compensation. While this solution can improve positioning accuracy to some extent, its compensation effect is highly dependent on the accuracy of the added device on the output side, typically requiring a high-precision sensor, which significantly increases the system's hardware cost.

[0004] Therefore, how to effectively compensate for the angular deviation caused by joint stiffness without increasing hardware costs, and thus improve the absolute positioning accuracy of robot joints throughout their entire life cycle, is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] This application first proposes a robot joint absolute precision compensation control method to solve the above-mentioned problems. Secondly, this application proposes a robot joint absolute precision compensation control system. Finally, it proposes a robot.

[0006] As the first aspect of this application, a method for absolute precision compensation control of robot joints is proposed, comprising the following steps: Obtain the desired torque of the robot joints in the current posture; Based on the desired torque, the corresponding joint stiffness is queried from the calibrated stiffness characteristic relationship, which includes joint stiffness corresponding to multiple different desired torques; the joint compensation coefficient is calculated based on the joint stiffness corresponding to the desired torque. The stiffness compensation value is calculated using the joint compensation coefficient and the desired torque; Based on the stiffness compensation value, a compensation command is output to control the output angle of the robot joint.

[0007] Optionally, the stiffness characteristic relationship is pre-calibrated in the following way: within a preset range of the rated torque of the robot joint, multiple different torque percentages are selected as measurement points, and the joint stiffness at each measurement point is measured respectively to establish the stiffness characteristic relationship indexed by the torque percentage.

[0008] Optionally, the preset range includes a first interval and a second interval, wherein the torque percentage of any measurement point in the first interval is less than the torque percentage of any measurement point in the second interval, and the interval between adjacent measurement points in the first interval is less than the interval between adjacent measurement points in the second interval.

[0009] Optionally, the number of measurement points in the first interval is greater than the number of measurement points in the second interval.

[0010] Furthermore, the torque percentage in the first range is less than or equal to 10% of the rated torque; the torque percentage in the second range is greater than 10% and less than or equal to 30%.

[0011] Furthermore, the joint compensation coefficient The calculation formula is: ;in, The rated torque of the joint; This represents the percentage of the desired torque. Torque percentage of the desired torque The corresponding joint stiffness; The joint stiffness compensation value The calculation formula is: .

[0012] Optionally, it further includes: calculating the hysteresis compensation value of the robot joint, and using the stiffness compensation value and the hysteresis compensation value together to compensate the output angle of the joint.

[0013] Optionally, the calculation steps for the hysteresis compensation value are as follows: The cumulative running time of the joint is determined based on the joint operating torque, joint operating speed, joint rated torque, joint rated speed, and sampling time. Obtain the initial hysteresis value of the joint and the hysteresis value when the joint reaches its design life; The hysteresis compensation value is calculated based on the initial hysteresis value of the joint, the hysteresis value when the joint reaches its design life, and the cumulative operating time of the joint.

[0014] As a second aspect of this application, a robot joint absolute precision compensation control system is proposed, including a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0015] As a third aspect of this application, a robot is proposed, comprising at least one robot joint and the aforementioned robot joint absolute precision compensation control system.

[0016] The beneficial effects of this application are as follows: 1. This application retrieves the joint stiffness from the calibrated stiffness characteristic relationship based on the desired torque, and then calculates the stiffness compensation value to correct the output angle. The compensation method provided in this application is based on the physical properties of the robot joint transmission chain. Once calibrated, it can be applied within the load range of the robot joint and is not affected by load changes. Moreover, the entire compensation process only relies on the desired torque information already existing in the processor, without the need to add additional sensors on the joint output side. This reduces system cost and complexity, and avoids the drawback of impairing repeatability accuracy due to the introduction of real-time measurement noise. This reduces the cost and complexity of robots using this method and improves economic value.

[0017] 2. Based on the characteristic that joint stiffness changes drastically in the low torque region and tends to saturate in the high torque region, this application divides the measurement points into multiple intervals and increases the interval between measurement points in each interval in turn. This concentrates limited calibration resources in the low torque region where stiffness changes most drastically, achieving the optimal balance between calibration efficiency and compensation accuracy with relatively low experimental cost. It overcomes the contradiction of insufficient accuracy in the low torque region and waste of resources in the high torque region that exists in the traditional uniform segmentation strategy, and has high practical and economic value.

[0018] 3. This application introduces a hysteresis aging model based on cumulative running time. The hysteresis compensation value is calculated based on the proportion of the current equivalent running time in the design life, and it is superimposed with the stiffness compensation value to jointly correct the joint output angle. This allows the compensation amount to adaptively increase with the actual wear level, ensuring the absolute positioning accuracy of the robot throughout its entire life cycle. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the robot joint absolute precision compensation control method in this embodiment.

[0020] Figure 2 This is a schematic diagram of the process of introducing hysteresis compensation in this embodiment.

[0021] Figure 3 This is a line graph showing the stiffness characteristic relationship in this embodiment.

[0022] Figure 4 This is a line graph showing the results of a robotic arm's compensation-closing and compensation-opening experiments under a 0kg load in this embodiment.

[0023] Figure 5 This is a line graph showing the results of a robotic arm's compensation-closing and compensation-opening experiments under a 12kg load in this embodiment. Detailed Implementation

[0024] The present application will be further described in detail below with reference to the accompanying drawings.

[0025] As the first aspect of this specific embodiment, such as Figure 1 As shown, a method for absolute precision compensation control of robot joints is proposed, including the following steps: Obtain the desired torque of the robot joints in the current posture; Based on the desired torque, the corresponding joint stiffness is queried from the calibrated stiffness characteristic relationship, which includes the joint stiffness corresponding to multiple different desired torques; the joint compensation coefficient is calculated based on the joint stiffness corresponding to the desired torque. The stiffness compensation value is calculated using the joint compensation coefficient and the desired torque. The output compensation command controls the output angle of the robot joint based on the stiffness compensation value.

[0026] Regarding the desired torque, in the implementation of this application, the processor receives the target running command under the current joint posture and generates the desired torque according to the dynamic model or motion planning algorithm. The desired torque is the command value that the joint needs to output under the current posture. The specific expression of the desired torque can be flexibly selected according to the needs of the actual control system. In some embodiments, the desired torque can be obtained by the processor of the robotic arm / robot through real-time calculation using the dynamic model. The dynamic model integrates the joint angles, angular velocities, angular accelerations, and inertia parameters such as link lengths and joint weights. Those skilled in the art can understand that its specific derivation can be found in the relevant chapters of standard textbooks such as "Robotics". In other embodiments, the desired torque is generated by the motion planning algorithm. The processor pre-calculates the torque command value that each joint needs to output at the corresponding path point based on the position, velocity, and acceleration sequence of each path point in the planned path, combined with the robot's dynamic parameters, and uses this as the desired torque. The method of obtaining the desired torque is not limited to any specific method. Any method known to those skilled in the art is within the protection scope of the technical solutions disclosed in this application.

[0027] It should be noted that the obtained expected torque is independent of the type of drive motor; it can be applied to servo motors or stepper motors with encoders. Since the stiffness characteristic relationship is calibrated at different torque percentages, in some implementations, the processor directly converts it to a torque percentage form during the calculation of the expected torque, facilitating subsequent lookup and calculation using the stiffness characteristic relationship.

[0028] Regarding stiffness characteristic relationships, these relationships refer to pre-established mappings describing the stiffness values ​​of a joint at different torque levels. Methods for calibrating stiffness characteristic relationships include, but are not limited to: measuring joint stiffness values ​​at multiple discrete torque measurement points to create a stiffness characteristic mapping table; performing curve fitting on discrete measurement points to create a continuous stiffness characteristic curve; or describing the correspondence between torque and stiffness through a functional expression. The representation of stiffness characteristic relationships can be stored and retrieved in a processor as data tables, arrays, functions, or mathematical formulas. Regardless of the specific form used, as long as it provides a correlation between different torques and corresponding joint stiffness, it constitutes a stiffness characteristic relationship.

[0029] It should be understood that the calibration of stiffness characteristics can be completed offline, and the calibration data can be stored in the non-volatile memory of the joint controller for online compensation. The calibration process can be completed once before the joint leaves the factory, or it can be re-executed after joint maintenance or gearbox replacement to ensure the timeliness and accuracy of the stiffness characteristics.

[0030] In some implementations, the stiffness characteristic relationship is pre-calibrated in the following way: within a preset range of the rated torque of the robot joint, multiple different torque percentages are selected as measurement points, and the joint stiffness at each measurement point is measured to establish a stiffness characteristic relationship indexed by the torque percentage.

[0031] It should be noted that when the stiffness characteristic relationship is a stiffness characteristic mapping table, if the percentage of torque corresponding to the desired torque of the robot joint is exactly equal to the percentage of torque corresponding to the rated torque at a certain measurement point, the stiffness value of that point can be directly used. If the percentage of torque of the desired torque falls between the percentages of torque of the rated torque at two measurement points, the stiffness value is obtained by interpolation. The interpolation method is well known to those skilled in the art and will not be elaborated here.

[0032] Furthermore, in the description of this application, torque percentage refers to the percentage value of the current torque relative to the rated torque of the joint. It is expressed as a percentage or decimal, such as 10% or 0.1, and the two can be converted to each other. Essentially, both represent the proportional relationship between the current torque and the rated torque. It should be noted that establishing stiffness characteristic relationships using torque percentage as an index is a preferred embodiment of this application. Its advantage is that torque percentage is a dimensionless quantity, which allows the same set of stiffness characteristic relationships to be applied to joints with different rated torques. Only proportional conversion is required, without the need for separate calibration for each joint, thus improving the universality of calibration data between joints of different specifications.

[0033] It should be understood that, besides torque percentage, the index in the stiffness characteristic relationship can also take other forms. For example, in some embodiments, the absolute value of torque can be directly used as the index to establish the stiffness characteristic relationship, in which case different absolute values ​​of torque correspond to different joint stiffness values. In other embodiments, the dimensionless ratio of torque, i.e., the ratio of the current torque to the rated torque, can be used as the index. The essence of all the above-mentioned index forms is to characterize the current stress level of the joint and provide the corresponding joint stiffness value accordingly. They are all reasonable variations of the technical solution of this application and do not exceed the protection scope of this application.

[0034] Regarding joint stiffness, joint stiffness refers to the ability of a robot joint to resist torsional deformation when subjected to torque. It is usually expressed as the ratio of the change in torque to the change in torsional angle, and the unit can be Nm / rad. Due to the actual characteristics of components such as reducers used in robot joints, joint stiffness is usually not a constant value, but exhibits nonlinear characteristics as it changes with the torque level. The static load method is commonly used to measure joint stiffness in stiffness characteristic relationships. In some implementations, a rigid link of known length is installed at the joint output end, and weights of different masses are successively suspended from the end of the link to apply graded torques. The displacement of the link end before and after loading is measured using laser rangefinders or dial indicators, and the displacement is converted into joint torsional angle using geometric relationships. The joint stiffness at each torque value is obtained by dividing the torque increment by the corresponding torsional angle increment, where the torque increment is the difference between the current torque value and the zero-load torque value, and the torsional angle increment is the difference between the current torsional angle and the zero-load torsional angle. This calculation principle is based on fundamental formulas in mechanics of materials, which can be found in the torsional stiffness testing chapter of mechanics of materials textbooks, and will not be elaborated further here.

[0035] It should be noted that the joint stiffness described in this application is not limited to a specific type of reducer. Whether it is a harmonic reducer, RV reducer, planetary reducer, or cycloidal pinwheel reducer, all exhibit varying degrees of stiffness nonlinearity, which can be compensated for using the method provided in this application. Therefore, the technical solution of this application has good applicability and is not limited by specific reducer types.

[0036] The method provided in this application, based on the desired torque of the robot joint in its current posture, queries the corresponding joint stiffness from a pre-calibrated stiffness characteristic relationship, then calculates the stiffness compensation value from the joint stiffness, and finally corrects the output angle. Firstly, the stiffness compensation value is derived from the desired torque through the stiffness characteristic relationship. Furthermore, stiffness characteristics are inherent physical properties of the robot joint transmission chain; once calibrated, they are applicable across the entire load range, and the compensation effect is unaffected by load changes, eliminating the need for recalibration due to load variations.

[0037] Secondly, the compensation path used in this application is desired torque - joint stiffness - compensation value, which does not rely on any real-time external measurements. Under a given posture, the desired torque output by the dynamic model is a uniquely determined value; the joint stiffness queried based on this value and the stiffness compensation value calculated from it are also uniquely determined values. Therefore, the compensation amount of the robot joint will not introduce any random fluctuations. This ensures that while improving the absolute positioning accuracy of the joint, the repeatability accuracy remains unaffected, overcoming the drawback of existing technologies where the introduction of real-time measurement noise leads to a decrease in repeatability accuracy.

[0038] Finally, the entire compensation process can be completed using only the desired torque information already available in the processor, combined with the pre-calibrated stiffness characteristic relationship. There is no need to add any high-precision encoders or torque sensors on the joint output side, which can reduce the cost and complexity of robots using this method and improve their economic value.

[0039] In some implementations, the preset range of the robot joint's rated torque in the stiffness characteristic relationship includes a first interval and a second interval. The torque percentage at any measurement point in the first interval is less than the torque percentage at any measurement point in the second interval, and the interval between adjacent measurement points in the first interval is less than the interval between adjacent measurement points in the second interval. Furthermore, the number of measurement points in the first interval is greater than the number of measurement points in the second interval. Experiments have shown that joint stiffness changes drastically in the low torque region, and the rate of change of stiffness gradually decreases as the torque increases. For example, in the range of 2% to 10% of the rated torque, the stiffness value can change by up to 18%; while in the range of 70% to 100%, the stiffness value changes by only about 8%. If equally spaced point selection is used, abrupt changes in stiffness in the low torque region will be missed, resulting in insufficient compensation accuracy. Therefore, the interval between adjacent measurement points in the first interval is less than the interval in the second interval in order to capture the rapid changes in stiffness in the low torque region. Meanwhile, since the stiffness curve in the low torque region has a high degree of nonlinearity, more dense data points are needed to accurately depict its curve shape. Therefore, the number of measurement points in the first interval is also greater than that in the second interval, thus ensuring a more accurate relationship between the stiffness characteristics in the low torque region.

[0040] In some implementations, the torque percentage in the first interval is less than or equal to 10% of the rated torque; the torque percentage in the second interval is greater than 10% and less than or equal to 30%. In this implementation, if the stiffness characteristic relationship is a stiffness characteristic mapping table, it can satisfy the scenario where the torque percentage of the desired torque during robot joint movement is less than 30%.

[0041] In some embodiments, the preset range also includes a third and fourth interval, where the preset range of the robot joint's rated torque in the stiffness characteristic relationship further includes a third and fourth interval; the density of measurement points in each of the first, second, third, and fourth intervals decreases sequentially. The first interval corresponds to the ultra-low torque region where stiffness changes most drastically, the second interval corresponds to the medium-low torque region where stiffness changes relatively significantly, the third interval corresponds to the medium-high torque region where stiffness changes more gradually, and the fourth interval corresponds to the high torque region where stiffness tends to saturate. The sequentially decreasing density of measurement points in the four intervals further improves the compensation accuracy while maintaining the same total number of measurement points.

[0042] In some embodiments where the preset range also includes a third and a fourth interval, the torque percentage in the first interval is less than or equal to 10% of the rated torque; the torque percentage in the second interval is greater than 10% and less than or equal to 20%; the torque percentage in the third interval is greater than 20% and less than or equal to 60%; and the torque percentage in the fourth interval is greater than 60% and less than or equal to 100%. Further, a specific selection table for the corresponding points in each interval is as follows:

[0043] In other embodiments where the preset range also includes a third and a fourth range, the torque percentage of the first range is less than or equal to 10% of the rated torque; the torque percentage of the second range is greater than 10% and less than or equal to 30%; the torque percentage of the third range is greater than 30% and less than or equal to 70%; and the torque percentage of the fourth range is greater than 70% and less than or equal to 100%.

[0044] Furthermore, the intervals between the measurement points in the first, second, third, and fourth intervals increase sequentially. In some embodiments, the intervals between adjacent measurement points are: 2% for the first interval (5 points in total), 4% for the second interval (5 points in total), 8% for the third interval (5 points in total), and 10% for the fourth interval (3 points in total), totaling 18 measurement points. The specific selection table for the corresponding points in each interval is as follows:

[0045] At this point, the stiffness characteristic relationship exists in the form of a stiffness characteristic mapping table, specifically as follows:

[0046] In some implementations, the discrete points in the above stiffness characteristic mapping table can be connected sequentially in ascending order of torque percentage to form a structure like... Figure 3 The stiffness characteristic curve shown indicates that if the desired torque falls between the torque percentages corresponding to two adjacent measurement points, then linear interpolation is used to obtain the joint stiffness.

[0047] Through repeated experiments, it was verified that in the low-torque region, joint stiffness increases sharply with increasing torque; however, as torque rises further, this upward trend gradually slows down and tends to saturate. Therefore, this application divides the measurement points into multiple intervals, and increases the interval between measurement points in each interval sequentially, i.e., the measurement points are dense in the low-torque region and sparse in the high-torque region. With the total number of measurement points remaining constant, limited calibration resources are concentrated in the low-torque region where stiffness changes most drastically, allowing for accurate capture of stiffness nonlinear details in this region and improving compensation accuracy under low-load conditions. In the high-torque region where stiffness changes gradually, sufficient query accuracy can be ensured with fewer measurement points, avoiding unnecessary increases in calibration workload.

[0048] In other words, the solution provided in this application maximizes resource utilization with relatively low experimental costs, achieves the optimal balance between calibration efficiency and compensation accuracy, overcomes the contradiction of insufficient accuracy in the low torque region and waste of resources in the high torque region in the traditional uniform segmentation strategy, and is highly practical and economically valuable.

[0049] In some implementations, the joint compensation coefficient The calculation formula is: ;in, The rated torque of the joint is determined by the relevant specifications of the motor used in the joint, and the unit is [unit missing]. ; The torque percentage of the desired torque is obtained by the processor based on the desired torque calculated in real time according to the inverse dynamics model; The joint stiffness corresponding to the current desired torque percentage is obtained by querying the stiffness characteristic relationship, and the unit is... ; Joint stiffness compensation value The calculation formula is: .

[0050] In such Figure 2 In some of the embodiments shown, the cumulative running time of the robot joint is obtained; based on the cumulative running time, the initial hysteresis value of the robot joint, and the hysteresis value when the robot joint reaches its design life, the joint hysteresis compensation value is calculated; and the output angle of the joint is compensated by the stiffness compensation value and the hysteresis compensation value.

[0051] Besides the angular error caused by stiffness, robot joints also experience increased hysteresis due to wear and fatigue of transmission components (such as harmonic reducers and bearings) during long-term operation. Hysteresis also affects the absolute positioning accuracy of the joint. To address this, the method provided in this application further introduces a hysteresis compensation mechanism: obtaining the cumulative operating time of the joint since it was put into use; calculating the hysteresis compensation value of the joint in the current state based on the cumulative operating time, the initial hysteresis value of the joint at the time of manufacture, and the expected hysteresis value when the joint reaches its design life; finally, the stiffness compensation value and the hysteresis compensation value are superimposed on the joint's output angle command to form a comprehensive compensation amount, thereby simultaneously correcting the angular error caused by stiffness deformation and hysteresis.

[0052] In some implementations, the hysteresis compensation value is calculated based on the cumulative running time. The calculation formula is: ;in, The joint operating torque is obtained or estimated in real time by the processor based on the current load. The joint's operating speed is measured in real time by an encoder at the motor end. The rated torque of the joint is determined by the relevant specifications of the motor. The rated speed of the robot joints is determined by the relevant specifications of the motor. The sampling time is the period during which the processor performs the compensation operation. This is the cumulative operator; The unit is hours, used to characterize the equivalent wear of a joint under current load and speed conditions.

[0053] In some implementations, joint hysteresis compensation value The calculation formula is: ;in, This is the initial hysteresis value, which is the hysteresis value measured when the robot joint leaves the factory in a wear-free state; The hysteresis value at which the design life is reached can be determined based on the technical parameters or accelerated aging test data provided by the reducer manufacturer. The design life of the joint is determined by the specifications of the reducer or joint module.

[0054] The aforementioned hysteresis compensation values ​​and stiffness compensation values ​​are all fixed values ​​and will not introduce additional random errors. Therefore, the repeatability of the joint after comprehensive compensation is not affected.

[0055] This application further considers the changes in hysteresis characteristics of the reducer due to wear during long-term operation. As the cumulative operating time of the joint increases, the internal transmission clearance of the reducer gradually increases, causing an irreversible shift in the angular deviation at the same torque level. If compensation is based solely on the stiffness characteristic relationship calibrated at the factory, this shift will accumulate over time and eventually exceed the compensation tolerance. To address this, this application introduces a hysteresis aging model based on cumulative operating time, which can more accurately reflect the true aging state of the reducer. Based on this, this application calculates the current hysteresis compensation value according to the proportion of the current equivalent operating time in the design life, and superimposes it with the stiffness compensation value to jointly correct the joint's output angle. The compensation amount can adaptively increase with the actual wear level, ensuring that the robot maintains high absolute positioning accuracy throughout its entire lifespan.

[0056] In some embodiments, this application was experimentally verified on a six-axis robotic arm. Taking the second axis as an example, the joint was controlled to continuously perform 18 nominal angle movements at 10-degree intervals under two working conditions: 0 kg load and 12 kg load. The actual angle reached each time was measured using a laser tracker, and the average deviation between the nominal angle and the actual angle was calculated. The test results are shown in the table below:

[0057] Combining the above table and Figure 4 and Figure 5 The test results, shown in the line graph, indicate that under a 0kg load, after enabling the stiffness compensation of this application, the average angular deviation decreased from 0.0136 degrees to 0.0079 degrees, improving the absolute positioning accuracy by approximately 42%. Under a 12kg load, the average angular deviation decreased from 0.0235 degrees to 0.0111 degrees, improving the absolute positioning accuracy by approximately 53%. The accuracy improvement under a 12kg load is greater than that under a 0kg unloaded load because the joint experiences greater torque under heavier loads, resulting in greater elastic deformation due to joint flexibility. The stiffness characteristic relationship used in this application can accurately match the corresponding joint stiffness under different loads, and its compensation effect for large deviations is more significant. Therefore, the improvement in absolute positioning accuracy is relatively higher under loaded conditions.

[0058] Experimental results show that this application can effectively improve the absolute positioning accuracy of the joint under different load conditions, and the compensation effect is not weakened or failed due to load changes, verifying the characteristic that the compensation method provided by this application is unaffected by load changes. In addition, the above compensation process only relies on the desired torque information already inside the processor and the current feedback signal built into the driver, without adding any additional sensors on the joint output side, thus achieving a significant improvement in absolute positioning accuracy while reducing costs.

[0059] It should be noted that this experiment used a laser tracker as the measuring instrument, with a resolution of 0.0001 degrees and an actual accuracy of approximately 0.001 degrees. Due to the limitations of the measuring instrument's accuracy, the percentage improvement in absolute positioning accuracy mentioned above fluctuates within a certain range, but the overall trend of the experimental data is stable and can reflect the technical effect of the technical solution in this application.

[0060] It should be understood that although the above embodiments are illustrated using a single joint as an example, the technical solutions of this application can be extended to multiple joints or even all joints of a robot. For multi-joint robots, each joint can be independently calibrated and compensated for its stiffness characteristics, and the compensation values ​​are superimposed on the position commands of each joint, thereby improving the overall absolute positioning accuracy of the robot / robotic arm.

[0061] As a second aspect of this specific embodiment, a robot joint absolute precision compensation control system is proposed, including a processor and a memory. The memory stores a computer program, and when the processor executes the computer program, it implements the robot joint absolute precision compensation control method described above. This system can be embedded in a robot processor and operate as a specific module for joint precision compensation.

[0062] In some implementations, the robot joint absolute precision compensation control system can be embedded as a separate control module in the robot's main controller, operating in conjunction with the robot's motion control program. In other implementations, the robot joint absolute precision compensation control system can be deployed inside the joint actuators, directly correcting the position commands of individual joints. In still other implementations, the robot joint absolute precision compensation control system can be deployed on a host computer or in the cloud, sending compensation values ​​to each joint actuator via a communication bus.

[0063] The processor can be a central processing unit (CPU), digital signal processor (DSP), microcontroller unit (MCU), field-programmable gate array (FPGA), or other chip with program execution capability. The memory can include volatile memory (such as DRAM, SRAM) and non-volatile memory (such as flash memory, EEPROM), wherein the non-volatile memory can be used to store the calibrated segmented stiffness characteristic relationship table and hysteresis compensation related parameters to ensure that the calibration data is not lost after the system is powered off.

[0064] As a third aspect of this specific embodiment, a robot is proposed, including at least one robot joint and a robot joint absolute accuracy compensation control system as described above. Each joint can be independently compensated for by the comprehensive compensation system for stiffness and hysteresis characteristics, thereby improving the absolute positioning accuracy of each joint in a multi-joint robot and ultimately improving the overall pose accuracy of the robot's end effector.

[0065] It should be noted that the robots described in this application include, but are not limited to, industrial robots, collaborative robots, service robots, medical surgical robots, humanoid robots, and other types of automated equipment containing rotary joints. For robots with multiple joints, each joint can be calibrated with its own segmented stiffness characteristics and hysteresis aging parameters. The compensation calculations for each joint are independent and can be executed in parallel. The stiffness compensation value and hysteresis compensation value of each joint are respectively superimposed on the position control command of that joint. Through the independent improvement of the accuracy of each joint, the overall positioning accuracy of the robotic arm end effector is collaboratively optimized.

[0066] Those skilled in the art will understand that the above-mentioned robot joint absolute precision compensation system and robot are based on the same inventive concept, and their specific implementation methods and technical effects can be referred to each other, and will not be described in detail here.

[0067] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof, but do not exclude the presence or addition of one or more other features, steps, operations, devices, components and / or combinations thereof.

[0068] It should be noted that the terms "first interval," "second interval," "third interval," and "fourth interval" used in this application are for ease of description and do not imply any primary, secondary, or tertiary order of importance among the intervals. Each interval is distinguished and defined only by the torque percentage range it covers and the interval between measurement points. Specifically, the first interval corresponds to the low torque region, the second interval to the low-to-medium torque region, the third interval to the medium-to-high torque region, and the fourth interval to the high torque region. The specific torque percentage range of each interval can be adjusted according to the actual stiffness variation characteristics of the joint. The interval boundary values ​​given in the embodiments of this application are merely illustrative and do not constitute a limitation on the scope of protection of this application.

[0069] In the description of this application, the term "connection" should be interpreted broadly, for example, it can be a mechanical connection, an electrical connection, or a communication connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above term in this application based on the specific circumstances. The term "communication" refers to the process of transmitting data or instructions between two or more devices through a network or bus, and its communication methods include, but are not limited to, industrial fieldbus protocols such as EtherCAT, CAN bus, RS-485, SPI, and I2C.

[0070] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for absolute precision compensation control of robot joints, characterized in that, Includes the following steps: Obtain the desired torque of the robot joints in the current posture; Based on the desired torque, the corresponding joint stiffness is queried from the calibrated stiffness characteristic relationship, which includes multiple joint stiffnesses corresponding to different desired torques; Calculate the joint compensation coefficient based on the joint stiffness corresponding to the desired torque; The stiffness compensation value is calculated using the joint compensation coefficient and the desired torque; Based on the stiffness compensation value, a compensation command is output to control the output angle of the robot joint; The stiffness characteristic relationship is pre-calibrated in the following way: within a preset range of the rated torque of the robot joint, multiple different torque percentages are selected as measurement points, and the joint stiffness at each measurement point is measured respectively to establish the stiffness characteristic relationship indexed by the torque percentage. The preset range includes a first interval and a second interval. The torque percentage at any measurement point in the first interval is less than the torque percentage at any measurement point in the second interval, and the interval between adjacent measurement points in the first interval is less than the interval between adjacent measurement points in the second interval. Calculate the hysteresis compensation value of the robot joint, and use the stiffness compensation value and the hysteresis compensation value together to compensate the output angle of the joint; The calculation steps for the hysteresis compensation value are as follows: The cumulative running time of the joint is determined based on the joint operating torque, joint operating speed, joint rated torque, joint rated speed, and sampling time. Obtain the initial hysteresis value of the joint and the hysteresis value when the joint reaches its design life; The hysteresis compensation value is calculated based on the initial hysteresis value of the joint, the hysteresis value when the joint reaches its design life, and the cumulative operating time of the joint.

2. The robot joint absolute precision compensation control method as described in claim 1, characterized in that, The number of measurement points in the first interval is greater than the number of measurement points in the second interval.

3. The robot joint absolute precision compensation control method as described in claim 2, characterized in that, The torque percentage in the first range is less than or equal to 10% of the rated torque; the torque percentage in the second range is greater than 10% and less than or equal to 30%.

4. The robot joint absolute precision compensation control method as described in any one of claims 1-3, characterized in that, The joint compensation coefficient The calculation formula is: ;in, The rated torque of the joint; The torque percentage of the desired torque; Torque percentage of the desired torque The corresponding joint stiffness; The joint stiffness compensation value The calculation formula is: .

5. A robot joint absolute precision compensation control system, characterized in that, It includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 4.

6. A robot, characterized in that, It includes at least one robot joint and the robot joint absolute precision compensation control system as described in claim 5.

Images