Robot joint motor cooperative driving method based on double-encoder dynamic decoupling

Abstract

The invention relates to the technical field of robot motor control, in particular to a robot joint motor cooperative driving method based on double-encoder dynamic decoupling, which effectively overcomes the interference of elastic deformation of a transmission chain on control precision through real-time comparison of double-angle signals of a motor end and a load end of a robot and a lag compensation mechanism. The robot joint trajectory tracking stability under the variable load working condition is remarkably improved, transmission chain deformation interference can be isolated in real time under the robot joint main motor load working condition, a robot joint main motor torque instruction synchronously adapts to load disturbance, and the tail end trajectory deviation is restrained outside the unmeasurable external force influence.

Description

Technical Field

[0001] The present invention relates to the technical field of robot motor control, and in particular to a robot joint motor collaborative driving method based on dual encoder dynamic decoupling. Background Art

[0002] At present, robot joint motors are the core execution units of industrial robot power systems, and their performance directly determines the positioning accuracy, dynamic response and reliability of the entire machine.

[0003] The robot joint motors in the existing technology currently have the following technical problems: 1. Currently, traditional robot joint motors rely solely on a single encoder to measure the position of their output shafts, making it impossible to perceive the torsional deformation of the reducer and the actual position of the load end in real time. This results in the complete immeasurability of the elastic deformation of the transmission chain. Trajectory deviations caused by external force disturbances (such as workpiece collisions or sudden changes in gravity) are passively fed back to the motor control loop, leading to technical issues such as delayed system response and deteriorated positioning accuracy. This prevents the torque commands of the robot joint motors from synchronously adapting to load disturbances, resulting in trajectory deviations exceeding ±0.5° at the robot end effector.

[0004] 2. In industrial robots currently using hollow integrated joints, the traditional hysteresis compensator embedded in the joint controller uses fixed delay parameters, which can easily lead to the following problems under variable load / temperature conditions: Case 1: Low load + low temperature conditions: Overcompensation → Excitation of high-frequency resonance (>10Hz) → System oscillation.

[0005] Case 2: High load + high temperature conditions: Insufficient compensation → elastic deformation energy of the reducer accumulates → induces low-frequency resonance (2-8Hz) → fatigue damage of the reducer.

[0006] Case 3: Resonant energy cannot be actively offset → additional 15%-30% increase in servo energy consumption.

[0007] Therefore, a robot joint motor collaborative driving method based on dual encoder dynamic decoupling is needed to solve the above problems. Summary of the Invention

[0008] The present invention provides a robot joint motor collaborative driving method based on dual encoder dynamic decoupling. The present invention effectively overcomes the interference of transmission chain elastic deformation on control accuracy through real-time comparison of dual-angle signals at the robot motor end and the load end and a hysteresis compensation mechanism, significantly improving the robot joint trajectory tracking stability under variable load conditions. Under the load condition of the robot joint main motor, the transmission chain deformation interference can be isolated in real time, so that the torque command of the robot joint main motor can synchronously adapt to the load disturbance, and the terminal trajectory deviation can be suppressed outside the influence of unmeasurable external forces.

[0009] The technical solution adopted by the present invention to solve the above technical problems is: a robot joint motor collaborative driving method based on dual encoder dynamic decoupling, comprising the following steps: Step 1: collecting a first angle signal of the output shaft of the robot motor in real time through a first encoder installed on the output shaft of the robot motor; Step 2: Collect the second angle signal in real time through a second encoder installed at the load output end of the robot joint; Step 3: Generate a reference torque control instruction for the robot joint main motor according to the first angle signal and transmit it to the robot joint main motor controller; Step 4: performing a hysteresis compensation operation on the second angle signal to generate a hysteresis compensation signal; Step 5: Generate a compensation torque control instruction for the robot joint auxiliary motor based on the hysteresis compensation signal and transmit it to the robot joint auxiliary motor controller; Step 6: Calculate the difference between the second angle signal and the first angle signal in real time. When the absolute value of the difference exceeds a preset offset threshold: If the second angle signal is greater than the first angle signal, the output polarity of the compensation torque control instruction for the auxiliary motor of the robot joint is negative; If the second angle signal is smaller than the first angle signal, the output polarity of the compensation torque control instruction for the auxiliary motor of the robot joint is positive.

[0010] Furthermore, the first angle signal is the real-time rotation angle of the robot motor output shaft, and the second angle signal is the actual position angle of the robot joint end effector.

[0011] Furthermore, the hysteresis compensation operation in step 4 includes: Step 4-1: Set a dynamically adjustable delay time window for the robot joint auxiliary motor compensation torque command. The reference value of the time window is 0.8 to 1.5 times the closed-loop control period of the robot joint main motor position. Step 4-2: Query the preset mapping table to obtain the delay coefficient according to the real-time load rate of the robot joint and adjust the delay coefficient based on the temperature of the main motor of the robot joint; Step 4-3: Multiply the delay time window by the delay coefficient and apply the resultant multiplication to the second angle signal.

[0012] Furthermore, the setting logic of the delay coefficient in step 4-2 is: When the robot joint load rate is between 0% and 30%, the delay factor is set to 0.9; When the robot joint load rate is between 30% and 70%, the delay factor is set to 1.0; When the robot joint load rate is between 70% and 100%, the delay factor is set to 1.2; The time delay coefficient is updated at a frequency of 10 to 50 times per second.

[0013] Further, the adjustment of the time delay coefficient based on the temperature of the robot joint main motor in step 4-2 includes: Step 4-2-1: Real-time monitoring of the temperature of the robot joint main motor, when the temperature exceeds 65℃ but does not exceed 85℃: If the current robot joint load rate is greater than or equal to 50%, the time delay coefficient is additionally increased by 0.1; If the current robot joint load rate is less than 50%, the time delay coefficient is additionally increased by 0.05; When the temperature of the robot joint main motor exceeds 85℃, the time delay coefficient under all robot joint load rates is forced to be set to 1.3 times the original value.

[0014] Further, the time delay coefficient is a proportional factor for adjusting the length of the delay time applied to the robot joint second angle signal in the lag compensation operation.

[0015] Further, the preset offset threshold in step 6 is the actual torsion angle deviation of the robot joint load output end relative to the robot motor output shaft, and the preset offset threshold range is 0.3° to 0.8°.

[0016] Further, the setting of the preset offset threshold in step 6 includes: dividing the robot joint load rate interval and associating the threshold range, when the real-time load rate of the robot joint is less than or equal to 30%, the offset threshold range is set to 0.7° to 0.8°; when the real-time load rate of the robot joint is > 30% and ≤ 70%, the offset threshold range is set to 0.5° to 0.6°; when the real-time load rate of the robot joint is greater than 70%, the offset threshold range is set to 0.3° to 0.4°.

[0017] Further, in step 3, when generating the robot joint main motor reference torque control instruction according to the first angle signal, the robot joint load disturbance model is used to dynamically decouple the double encoder signal mismatch amount, and the robot joint load disturbance model satisfies the following formula: ; Where τ main is the robot joint main motor reference torque instruction; Where θ error is the deviation of the first angle signal from the target angle, and the target angle is the angle position that the robot joint motor output shaft should theoretically reach at the current time; Where M load is the real-time joint load rate of the robot joint; Where ω is the stiffness attenuation factor mapped by the temperature of the robot joint main motor.

[0018] The advantages of the present invention are: 1. The present invention effectively overcomes the interference of transmission chain elastic deformation on control accuracy through real-time comparison of dual-angle signals at the robot motor end and load end and hysteresis compensation mechanism, significantly improving the robot joint trajectory tracking stability under variable load conditions. Under the load condition of the robot joint main motor, the transmission chain deformation interference can be isolated in real time, so that the torque command of the robot joint main motor can synchronously adapt to the load disturbance, and the terminal trajectory deviation can be suppressed outside the influence of unmeasurable external forces.

[0019] 2. The present invention has a highly adaptive hysteresis compensation mechanism. To address the problem of fixed delay compensators easily inducing resonance under variable loads (variable operating conditions), the present invention dynamically sets the delay coefficient based on the real-time joint load rate and performs online adjustments in combination with the temperature of the robot joint main motor, thereby constructing a compensation strength firewall that is adaptive to the working conditions. Under low-load and low-temperature working conditions, the compensation intensity is reduced to avoid overcompensation and eliminate induced high-frequency resonance; under high-load and high-temperature working conditions, the compensation is significantly enhanced to fully offset the accumulation of elastic deformation energy, thereby effectively suppressing the low-frequency resonance that induces fatigue damage to the reducer. The compensation torque generated in real time by the auxiliary motor of the robot joint actively offsets the deformation energy, converting the 15%-30% additional servo energy consumption caused by the accumulation of resonance in traditional solutions into effective mechanical power output, significantly reducing the total energy consumption of the system.

[0020] 3. The present invention can achieve closed-loop protection of the transmission chain status and extend the life of key components. The present invention not only controls accuracy and stability, but also actively builds a health protection mechanism for the transmission chain (especially the reducer) at the control level. By using temperature monitoring as a key input for compensation strength, when the temperature of the robot joint main motor rises abnormally, the compensation strength is dynamically enhanced, especially at high temperatures, forcing a significant increase, effectively blocking the risk of plastic deformation caused by thermal effects. The present invention forms a closed-loop protection against transmission chain deformation through the combined effects of temperature-related compensation enhancement and intelligent switching of torque polarity, which can significantly improve the fatigue life of key components such as harmonic reducers. DETAILED DESCRIPTION

[0021] The following is a clear and complete description of the technical solution of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention. Example

[0022] The present invention proposes a robot joint motor collaborative driving method based on dual-encoder dynamic decoupling, comprising the following steps: collecting a first angle signal of the robot motor output shaft in real time by a first encoder installed on the robot motor output shaft; collecting a second angle signal in real time by a second encoder installed on the robot joint load output end; generating a robot joint main motor reference torque control instruction according to the first angle signal, and transmitting it to the robot joint main motor controller; performing a lag compensation operation on the second angle signal to generate a lag compensation signal; generating a robot joint auxiliary motor compensation torque control instruction based on the lag compensation signal, and transmitting it to the robot joint auxiliary motor controller; calculating the difference between the second angle signal and the first angle signal in real time, and when the absolute value of the difference exceeds a preset offset threshold: if the second angle signal is greater than the first angle signal, then controlling the output polarity of the robot joint auxiliary motor compensation torque control instruction to be negative; if the second angle signal is less than the first angle signal, then controlling the output polarity of the robot joint auxiliary motor compensation torque control instruction to be positive.

[0023] Among them, the first encoder refers to a sensor installed on the output shaft of the robot motor for measuring the rotation angle. Specifically, it can be implemented by an incremental photoelectric encoder or an absolute magnetic encoder, and is used to obtain the theoretical output angle of the motor shaft end in real time. Among them, the second encoder refers to a position detection device installed at the load output end of the robot joint. Specifically, it can be implemented by a hollow absolute value encoder, and is used to directly measure the actual position angle of the end effector. Among them, the lag compensation operation refers to the process of applying an adjustable time delay to the load end angle signal. Specifically, a variable delay window can be implemented through a digital signal processing chip to dynamically eliminate the phase lag caused by the elastic deformation of the transmission chain. Among them, the polarity adjustment of the compensation torque control instruction refers to the control logic that changes the torque output sign according to the direction of the angle difference. Specifically, it can be implemented through a comparator circuit or software conditional judgment, and is used to actively offset the mechanical energy accumulated by the elastic deformation of the reducer.

[0024] The present invention dynamically compares the real-time angle signals of the dual encoders at the motor end and the load end, combines a hysteresis compensation mechanism with adjustable parameters and active torque polarity control, and forms a closed-loop decoupling control architecture for the collaborative drive of the main and auxiliary motors. While maintaining the main motor's reference torque output, the auxiliary motor compensates for the trajectory deviation caused by the deformation of the transmission chain in real time, thereby achieving precise position synchronization under load disturbances.

[0025] The working process and principle of the present invention are as follows: by installing a first encoder and a second encoder on the output shaft of the robot motor and the output end of the joint load respectively, the angle signals of the two positions are collected in real time. The main motor reference torque control instruction is generated according to the first angle signal, and the auxiliary motor compensation torque control instruction is generated after the second angle signal is hysteresis compensated. The difference between the two angle signals is calculated in real time. When the difference exceeds the preset threshold, the output polarity of the auxiliary motor compensation torque is controlled according to the positive or negative value of the difference. This dual-encoder dynamic decoupling method can monitor the elastic deformation of the transmission chain in real time, and offset the trajectory deviation caused by deformation through the coordinated drive of the main and auxiliary motors. The hysteresis compensation operation can make the compensation instruction match the deformation process of the transmission chain, avoiding compensation misalignment caused by signal transmission delay. By comparing the dual angle differences in real time and intelligently switching the compensation torque polarity, it is ensured that the elastic deformation can be actively offset at the early stage of accumulation, blocking the resonant energy transfer path.

[0026] As a preferred embodiment, the solution of the present invention is specifically implemented as follows: A high-precision incremental photoelectric encoder is installed on the motor output shaft of the robot joint as the first encoder, and a magnetic encoder is installed on the joint end effector as the second encoder. The resolution of the first encoder is 10,000 lines per revolution, and the resolution of the second encoder is 4,096 lines per revolution.

[0027] The control system collects angle signals from two encoders at a frequency of 1kHz. Based on the deviation between the first angle signal and the target angle, combined with preset PID parameters, it generates a reference torque control command for the main motor.

[0028] The second angle signal is hysteresis-compensated, with the compensation time window set to 1.2 times the main motor position closed-loop control period. The compensated signal is processed by a bandpass filter to remove high-frequency noise and low-frequency drift.

[0029] Based on the lag-compensated second angle signal, a feedforward-feedback composite control algorithm is used to generate the auxiliary motor compensation torque command. The feedforward portion estimates the inertia moment based on angular velocity and angular acceleration, while the feedback portion uses fuzzy PID control.

[0030] The difference between the two angle signals is calculated in real time, with an offset threshold set to 0.5°. When the absolute value of the difference exceeds the threshold, if the second angle signal is greater than the first, the output polarity of the auxiliary motor compensation torque command is set to negative; otherwise, it is set to positive.

[0031] Control commands for the main and auxiliary motors are transmitted to their respective drivers via the CAN bus. The drivers use vector control to achieve high-precision torque output.

[0032] Through the above scheme, the present invention can monitor and compensate for the elastic deformation of the transmission chain in real time, significantly improving the trajectory tracking accuracy of the robot joints. Through the coordinated drive of the main and auxiliary motors, it can respond quickly to sudden load changes and suppress system oscillations. The dynamically adjustable hysteresis compensation mechanism matches the compensation instructions with the actual deformation process, avoiding the problem of insufficient or excessive compensation. The intelligent switching of the compensation torque polarity can actively offset the elastic deformation energy and effectively prevent the occurrence of resonance. This method not only improves the positioning accuracy and dynamic response capability of the robot, but also extends the service life of key components such as the reducer.

[0033] The present invention further proposes that the first angle signal is the real-time rotation angle of the robot motor output shaft, and the second angle signal is the actual position angle of the robot joint end effector.

[0034] The first angle signal is directly acquired by an encoder mounted on the motor output shaft, reflecting the theoretical motion state of the driver. The second angle signal, acquired by an encoder at the load output, represents the actual position of the end effector. These two signals, corresponding to the position information of the driver and load, respectively, form the basic data source for closed-loop feedback. The accuracy of the end effector's actual position angle measurement directly affects the effectiveness of the compensation torque command.

[0035] Specifically, the motor output shaft encoder captures the rotation angle in real time as input for the main motor's reference torque control command. The end-effector encoder captures the actual position angle, which is lag-compensated and then generates the auxiliary motor's compensation torque command. By comparing the dual-angle signals in real time, the mismatch caused by elastic deformation of the transmission chain is dynamically decoupled. For example, when the deviation between the end-effector's actual position angle and the motor output shaft's angle exceeds a threshold, the auxiliary motor actively offsets the deformation through polarity switching. This solution suppresses the end-effector trajectory deviation to within ±0.3°, improving positioning accuracy by 40% compared to traditional single-encoder solutions.

[0036] As a preferred embodiment, the solution of the present invention is specifically implemented as follows: The first angle signal is the real-time rotation angle of the robot's motor output shaft, and the second angle signal is the actual position angle of the robot's joint end effector. Specifically, the first encoder is installed on the output shaft of the robot's joint motor and is used to collect the rotation angle of the motor output shaft in real time. The second encoder is installed on the end effector of the robot's joint and is used to collect the actual position angle of the end effector in real time. Therefore, by using these two encoders together, it is possible to simultaneously obtain angle information from the output end and the load end of the robot's joint motor, providing basic data for subsequent control.

[0037] Through the above technical solution, the present invention can simultaneously obtain the angle information of the output end and the load end of the robot joint motor, providing the necessary data basis for achieving precise control of the robot joint motor. Furthermore, by comparing the first angle signal and the second angle signal, the elastic deformation in the transmission chain can be monitored in real time, thereby effectively overcoming the interference of the elastic deformation of the transmission chain on the control accuracy. As a result, the present invention significantly improves the trajectory tracking stability of the robot joint under variable load conditions, realizes the synchronous adaptation of the torque command of the robot joint main motor to the load disturbance, and effectively suppresses the end trajectory deviation.

[0038] The present invention further proposes a method for performing lag compensation operation on the second angle signal, including setting a dynamically adjustable delay time window for the robot joint auxiliary motor compensation torque instruction, and the time window reference value is 0.8 to 1.5 times the robot joint main motor position closed-loop control period; querying a preset mapping table according to the real-time load rate of the robot joint to obtain a delay coefficient and adjusting the delay coefficient based on the temperature of the robot joint main motor; multiplying the delay time window by the delay coefficient and applying it to the second angle signal.

[0039] The baseline value of the delay time window is set at 0.8 to 1.5 times the main motor control period, allowing the time window length to be adaptively adjusted based on the dynamic characteristics of the main motor control loop. The delay coefficient is obtained by querying a preset load factor mapping table. For example, the coefficient is set to 0.9 for a load factor between 0% and 30%, 1.0 for a load factor between 30% and 70%, and 1.2 for a load factor between 70% and 100%. A temperature adjustment mechanism further optimizes the delay coefficient. When the main motor temperature exceeds 65°C, the coefficient is increased by an additional 0.05 to 0.1 depending on the load factor. When the temperature exceeds 85°C, the coefficient is forcibly increased to 1.3 times the original value. The delay coefficient is updated 10 to 50 times per second to ensure dynamic response speed.

[0040] Specifically, the reference value of the delay time window is linked to the main motor control cycle to synchronize the compensation delay with the system control frequency to avoid phase mismatch. The setting of the load rate mapping table associates different load intervals with corresponding delay attenuation or enhancement coefficients. For example, the compensation strength is reduced at low load to prevent high-frequency resonance, and the compensation strength is increased at high load to offset elastic deformation. The temperature adjustment mechanism introduces thermal state parameters. When the motor temperature rises and the mechanical stiffness decreases, the compensation action time is extended by increasing the delay coefficient to compensate for the transmission chain deformation lag caused by thermal expansion. The high-frequency update of the delay coefficient enables the compensation parameters to track load fluctuations and temperature changes in real time, forming a dynamic closed-loop adjustment. By establishing a multi-dimensional association between load rate, temperature and delay coefficient, the compensation strength can adapt to different working conditions, effectively suppress the accumulation of resonant energy, reduce servo energy consumption and extend the life of the reducer.

[0041] As a preferred embodiment, the scheme of the present application is implemented as follows: A dynamically adjustable time delay window is set for the auxiliary motor of the robot joint to compensate the torque command, and the reference value of the time delay window is 1.2 times of the position closed-loop control period of the main motor of the robot joint. A delay coefficient is obtained by querying a preset mapping table according to the real-time load rate of the robot joint. When the load rate of the robot joint is 40%, the delay coefficient is set to 1.0. Further, the delay coefficient is adjusted based on the temperature of the main motor of the robot joint. For example, when the temperature of the main motor is 75℃, the delay coefficient is additionally increased by 0.1. Finally, the time delay window is multiplied by the adjusted delay coefficient, and 1.32 times of the position closed-loop control period of the main motor is obtained as the final time delay time, which acts on the second angle signal.

[0042] Through the above technical scheme, the present application realizes dynamic adjustment of the auxiliary motor of the robot joint to compensate the torque command. Thus, the lag compensation operation can be adaptively adjusted according to real-time load and temperature changes, avoiding the problems of insufficient or excessive compensation caused by fixed delay parameters. Further, by dynamically adjusting the delay coefficient, the generation of high-frequency and low-frequency resonance is effectively suppressed, improving the stability and precision of the robot joint. Specifically, in the low-load low-temperature working condition, the compensation strength is reduced to avoid excessive compensation; in the high-load high-temperature working condition, the compensation is enhanced to offset the accumulated elastic deformation energy. This adaptive mechanism significantly improves the performance of the robot in variable load and temperature change conditions.

[0043] The present application further proposes a setting logic for the delay coefficient: when the load rate of the robot joint is between 0% and 30%, the delay coefficient is set to 0.9; when the load rate of the robot joint is between 30% and 70%, the delay coefficient is set to 1.0; when the load rate of the robot joint is between 70% and 100%, the delay coefficient is set to 1.2; and the delay coefficient is updated at a frequency of 10 to 50 times per second.

[0044] Among them, the delay coefficient is divided into three load rate intervals corresponding to the values, the low load interval uses a coefficient lower than the reference value, and the high load interval uses a coefficient higher than the reference value. The delay coefficient update frequency is limited to the range of 10-50Hz, and the high-frequency refresh mechanism ensures the real-time matching of the compensation strength and the load state. The division of the load rate interval uses an asymmetric boundary, with 30% and 70% as key division points, forming three compensation strength gradients of low, medium and high.

[0045] Specifically, in the load rate range of 0%-30%, the delay coefficient is reduced to 0.9 to reduce the amount of compensation and avoid high-frequency resonance caused by excessive compensation under low load; in the load rate range of 30%-70%, the baseline coefficient is maintained at 1.0 to keep the compensation strength matching the standard operating conditions; when the load rate exceeds 70%, the coefficient is increased to 1.2 to enhance compensation to offset the accumulation of elastic deformation energy caused by high load. The delay coefficient is updated at a frequency of 10-50 times per second, which ensures a rapid response to changes in load status and avoids overloading the controller's computing resources due to excessive frequency. For example, under conditions where the load rate fluctuates rapidly, using a 50Hz update frequency can ensure that the delay coefficient is adjusted within 20ms, so that the compensation strength accurately follows the load changes.

[0046] As a preferred embodiment, the solution of the present invention is specifically implemented as follows: The setting logic of the delay coefficient is: when the robot joint load rate is between 0% and 30%, the delay coefficient is set to 0.9; when the robot joint load rate is between 30% and 70%, the delay coefficient is set to 1.0; when the robot joint load rate is between 70% and 100%, the delay coefficient is set to 1.2; the delay coefficient is updated at a frequency of 10 to 50 times per second.

[0047] Specifically, a table lookup method can be used to dynamically set the delay coefficient. For example, a mapping table containing load rate ranges and corresponding delay coefficients can be pre-established. During operation, the system queries this table to obtain the corresponding delay coefficient based on the real-time load rate of the robot joints. To ensure the real-time accuracy of the delay coefficient, the system periodically updates the delay coefficient at a frequency of 10 to 50 times per second.

[0048] As a possible implementation, a timer can be used to trigger the update of the delay factor. The timer period can be set to a value between 20 milliseconds and 100 milliseconds, corresponding to an update frequency of 10 to 50 times per second. Each time the timer is triggered, the system will perform the following steps: Get the real-time load rate of the current robot joint; Query the mapping table according to the load rate and determine the corresponding delay coefficient; The new delay factor is applied to the lag compensation operation.

[0049] Through the above technical solution, the present invention can dynamically adjust the delay coefficient based on the real-time load conditions of the robot joint, thereby optimizing the lag compensation effect. Under low-load conditions, a smaller delay coefficient can avoid overcompensation; under high-load conditions, a larger delay coefficient can provide sufficient compensation strength. Frequent updates of the delay coefficient ensure that the compensation operation can respond to load changes in a timely manner, improving the dynamic adaptability of the system. This adaptive compensation mechanism helps to improve the control accuracy and stability of the robot joint while reducing the risk of resonance.

[0050] The present invention further proposes to adjust the delay coefficient based on the temperature of the robot joint main motor, including: real-time monitoring of the temperature of the robot joint main motor, when the temperature exceeds 65°C but does not exceed 85°C, if the current robot joint load rate is greater than or equal to 50%, the delay coefficient is additionally increased by 0.1; if the current robot joint load rate is less than 50%, the delay coefficient is additionally increased by 0.05; when the temperature of the robot joint main motor exceeds 85°C, the delay coefficient under all robot joint load rates is forcibly set to 1.3 times the original value.

[0051] Temperature monitoring uses a PT100 temperature sensor mounted on the main motor's stator winding, sampling temperature data at a frequency of 20 times per second. The delay coefficient adjustment logic is implemented using a preset temperature-load rate two-dimensional correction table, which is stored in the joint controller's non-volatile memory. When the temperature is between 65°C and 85°C, the controller determines the compensation increment based on the real-time load rate: when the load rate is ≥50%, the compensation increment is set to 0.1; when the load rate is ≤50%, the compensation increment is set to 0.05. When the temperature exceeds 85°C, the controller triggers temperature protection mode, forcibly multiplying the current delay coefficient by a correction factor of 1.3.

[0052] Specifically, when the main motor temperature reaches 70°C and the load factor is 60%, the original delay coefficient is 1.0. At this time, an additional 0.1 is added to make the final coefficient 1.1. When the temperature rises to 90°C, regardless of the load factor range, the delay coefficient is forced to increase to 1.3 times the original value. The temperature monitoring data and the load factor data are synchronized through timestamps to ensure the timing consistency of the compensation parameter adjustment. In temperature protection mode, the corrected delay coefficient acts on the time window calculation module of the lag compensation algorithm, increasing the compensation signal delay time by 30%, thereby enhancing the ability to offset thermally induced deformation. When the reducer stiffness decreases due to high temperature, this mechanism actively suppresses the accumulation of elastic deformation by increasing the compensation strength and blocks the chain reaction of thermoplastic deformation.

[0053] As a preferred embodiment, the solution of the present invention is specifically implemented as follows: Monitor the temperature of the robot joint's main motor in real time. When the temperature exceeds 65°C but does not exceed 85°C, perform the following operations: If the current robot joint load factor is greater than or equal to 50%, increase the delay factor by an additional 0.1; if the current robot joint load factor is less than 50%, increase the delay factor by an additional 0.05. When the temperature of the robot joint's main motor exceeds 85°C, the delay factor for all robot joints at all load factors is forcibly set to 1.3 times the original value.

[0054] For example, in one specific embodiment, the temperature of the robot joint's main motor is monitored at 75°C. At this point, the system detects that the robot joint's load factor is currently 60%. Because the temperature exceeds 65°C but does not exceed 85°C, and the load factor is greater than or equal to 50%, the delay factor is increased by an additional 0.1. Assuming the original delay factor is 1.0, the adjusted delay factor is 1.1.

[0055] Furthermore, if the temperature of the robot joint's main motor continues to rise to 90°C, exceeding the 85°C threshold, the system will forcibly set the delay factor to 1.3 times the original value, regardless of the current robot joint load rate. Assuming the original delay factor is 1.0, the adjusted delay factor is 1.3.

[0056] Through the above technical solution, the present invention realizes the dynamic delay coefficient adjustment based on the temperature of the main motor of the robot joint. Thus, under high temperature conditions, the compensation strength is enhanced by increasing the delay coefficient, which effectively offsets the accumulation of elastic deformation energy caused by the temperature increase. Specifically, when the temperature is between 65°C and 85°C, different degrees of delay coefficient increase are adopted according to different load rates to achieve more refined compensation adjustments. When the temperature exceeds 85°C, a more radical delay coefficient increase strategy is adopted to cope with more serious thermal effects that may occur. This temperature-related compensation strengthening mechanism effectively blocks the risk of plastic deformation caused by thermal effects and improves the stability and reliability of the robot joints.

[0057] The present invention further proposes that the delay coefficient is a proportional factor for adjusting the length of the delay time applied to the second angle signal of the robot joint in the lag compensation operation.

[0058] The delay coefficient, acting as a proportional factor, directly determines the final calculated value of the delay time window. Multiplying this value by the time window generates the actual compensation delay applied to the second angle signal. This coefficient is mapped to the load factor using a preset mapping table, and incremental adjustments are added when temperature anomalies occur, creating a dynamic proportional adjustment mechanism. For example, when the load factor is 50% and the temperature is within the specified range, the delay coefficient is 1.0. If the temperature exceeds 65°C and the load factor is greater than 50%, the coefficient increases to 1.1.

[0059] Specifically, in the hysteresis compensation operation, the base value of the delay time window is 0.8 to 1.5 times the main motor control period, and the delay coefficient is dynamically adjusted through load rate interval division and temperature monitoring. For example, when the load rate is in the range of 30%-70%, the basic coefficient is set to 1.0; if the main motor temperature exceeds 65°C, an increment of 0.05 or 0.1 is superimposed according to the load rate. After the delay coefficient is multiplied by the time window, the final delay time is generated and acts on the second angle signal. This proportional factor is defined quantitatively so that the compensation delay time can accurately match the load and temperature changes, avoiding excessive or insufficient compensation caused by fixed parameters. For example, under high temperature and high load conditions, the delay coefficient is forced to increase to 1.3 times the original value, significantly increasing the delay time to offset the impact of thermal deformation, blocking the accumulation of elastic deformation energy, thereby suppressing low-frequency resonance and reducing the risk of fatigue damage to the reducer.

[0060] As a preferred embodiment, the solution of the present invention is specifically implemented as follows: The delay coefficient is a proportional factor used to adjust the delay time applied to the second angle signal of the robot joint during the lag compensation operation. Specifically, the delay coefficient is a dimensionless value that is multiplied by the preset baseline delay time to obtain the delay time for actual application. For example, when the baseline delay time is set to 10 milliseconds, if the delay coefficient is 0.9, the actual delay time is 9 milliseconds; if the delay coefficient is 1.2, the actual delay time is 12 milliseconds. The delay coefficient can be dynamically adjusted according to the real-time load rate of the robot joint and the temperature of the main motor to meet the compensation requirements under different working conditions.

[0061] Through the above-mentioned technical solution, the present invention achieves flexible adjustment of hysteresis compensation during the coordinated drive of robot joint motors. This allows precise control of the delay time of the second angle signal based on actual operating conditions, avoiding the problems of under- or over-compensation caused by fixed delay parameters. Furthermore, this dynamic adjustment mechanism effectively suppresses system resonance, improving the control accuracy and stability of the robot joints, while also reducing servo energy consumption and extending the life of key components.

[0062] The present invention further proposes that the setting of a preset offset threshold includes dividing the robot joint load rate interval and associating the threshold range. When the real-time load rate of the robot joint is less than or equal to 30%, the offset threshold range is set to 0.7° to 0.8°; when the real-time load rate of the robot joint is greater than 30% and ≤70%, the offset threshold range is set to 0.5° to 0.6°; when the real-time load rate of the robot joint is greater than 70%, the offset threshold range is set to 0.3° to 0.4°.

[0063] The offset threshold range is divided into three numerical intervals corresponding to the load rate intervals, specifically, the low load rate interval corresponds to 0.7° to 0.8°, the medium load rate interval corresponds to 0.5° to 0.6°, and the high load rate interval corresponds to 0.3° to 0.4°. The division boundaries of the load rate interval are set to 30% and 70%, and the joint load rate data is obtained in real time by the pre-set load rate detection module. The threshold range is negatively correlated with the load rate, that is, the higher the load rate, the smaller the threshold range.

[0064] Specifically, when the robot joint is in the low load rate interval, the end effector is less affected by external disturbances, allowing a larger angular deviation threshold to avoid false triggering of the compensation polarity switching. At this time, the threshold range is set to 0.7° to 0.8°, and the corresponding threshold parameter is called by the threshold selection module built-in the main motor controller. When the load rate enters the medium load rate interval, the end effector is affected by the increased elastic deformation of the transmission chain, and the threshold range is adjusted to 0.5° to 0.6° to balance the trajectory accuracy and system stability. When the load rate exceeds 70%, the reducer elastic deformation can quickly accumulate, and the threshold range is compressed to 0.3° to 0.4°, triggering the auxiliary motor compensation torque polarity switching in time through more sensitive deviation detection, and blocking the deformation energy accumulation. The dynamic adjustment of the threshold parameter is realized through the pre-set load rate-threshold mapping table, which is stored in the non-volatile memory of the joint controller.

[0065] As a preferred embodiment, the scheme of the present application is implemented as follows: The preset offset threshold is the actual torsional angle deviation of the robot joint load output end relative to the robot motor output shaft, and the preset offset threshold range is 0.3° to 0.8°. Specifically, during the cooperative driving process of the robot joint motor, the first angle signal of the robot motor output shaft and the second angle signal of the robot joint load output end are collected by the first encoder and the second encoder respectively. Further, the difference between the second angle signal and the first angle signal is calculated in real time, and the difference is compared with the preset offset threshold. Thus, when the absolute value of the difference exceeds the preset offset threshold, the output polarity of the robot joint auxiliary motor compensation torque control instruction is controlled according to the size relationship between the second angle signal and the first angle signal. For example, the offset threshold can be set to 0.5°. When the calculated absolute value of the difference is greater than 0.5°, if the second angle signal is greater than the first angle signal, the output polarity of the robot joint auxiliary motor compensation torque control instruction is negative; if the second angle signal is less than the first angle signal, the output polarity of the robot joint auxiliary motor compensation torque control instruction is positive.

[0066] By the technical solution, the angle deviation between the robot joint load output end and the motor output shaft can be accurately detected and compensated. By setting a reasonable offset threshold range, the problem of overcompensation or insufficient compensation can be effectively avoided. Further, the polarity of the compensation torque is controlled according to the direction of the angle deviation, so that the position of the robot joint can be accurately adjusted. Therefore, the positioning accuracy and motion stability of the robot joint are improved, and the trajectory deviation caused by the elastic deformation of the transmission chain is reduced.

[0067] The application further divides the robot joint load rate interval and associates the threshold range. When the real-time load rate of the robot joint is less than or equal to 30%, the offset threshold range is set to 0.7° to 0.8°; when the real-time load rate of the robot joint is greater than 30% and less than or equal to 70%, the offset threshold range is set to 0.5° to 0.6°; and when the real-time load rate of the robot joint is greater than 70%, the offset threshold range is set to 0.3° to 0.4°.

[0068] The offset threshold range is divided into three consecutive intervals, each corresponding to a different load condition. The load rate interval is divided at 30% and 70%, forming three states: light load, medium load, and heavy load. The threshold range decreases in steps as the load rate increases, and is dynamically adjusted within the 0.3° to 0.8° interval. The mapping relationship between the load rate and the threshold range is realized through a pre-set parameter table.

[0069] Specifically, by monitoring the real-time load rate of the robot joint, the corresponding threshold range is dynamically selected. When the load rate is less than 30%, a larger threshold of 0.7° to 0.8° is used, allowing a larger angle deviation between the primary and secondary motors without triggering polarity switching, thereby avoiding excessive compensation under low load that can cause high-frequency oscillation. When the load rate rises to the 30%-70% interval, the threshold is reduced to 0.5° to 0.6°, at which point the transmission chain stiffness decreases but has not reached its limit, and a moderate advance in compensation can balance energy consumption and stability. When the load rate exceeds 70%, the smallest threshold of 0.3° to 0.4° is used to intervene in compensation in advance under high stress conditions of the reducer, effectively blocking the accumulation of elastic deformation energy. This dynamic threshold mechanism matches the load rate with the degree of stiffness degradation of the transmission chain, so that the timing of the compensation torque polarity switching is synchronized with the deformation characteristics of the mechanical system in real time, and the balance point between resonance suppression and energy optimization is controlled in the best interval.

[0070] In a preferred embodiment, the method for setting the preset offset threshold includes: establishing a robot joint load rate monitoring module, collecting the current signal of the torque sensor at the output end of the reducer in real time and converting it into a load rate percentage value; creating a load rate interval division unit, dividing the load rate into three level intervals: low, medium and high; configuring a threshold mapping module, calling the threshold database of 0.7° to 0.8° when the load rate is monitored to be in the low interval, calling the threshold database of 0.5° to 0.6° when it is in the medium interval, and calling the threshold database of 0.3° to 0.4° when it is in the high interval; setting the threshold dynamic switching logic, when the load rate crosses the boundary of the adjacent interval, the threshold is smoothly interpolated with a transition period of 20 milliseconds to avoid control oscillation caused by step changes.

[0071] Through the above technical solution, the present invention effectively solves the problem of sensitivity mismatch of traditional fixed offset thresholds under variable load conditions. Using a larger threshold in the low-load range can prevent the joint auxiliary motor from frequently switching torque polarity due to slight deformation, thereby avoiding high-frequency oscillations; using a smaller threshold in the high-load range can timely capture the elastic deformation of the reducer and prevent deformation energy accumulation from causing low-frequency resonance. This dynamic threshold mechanism enables the system to automatically match the optimal deviation detection sensitivity under different load conditions, which not only ensures the end trajectory tracking accuracy, but also avoids energy loss caused by over-compensation, significantly improving the robustness and energy efficiency of the robot joint control.

[0072] The present invention further proposes a technical solution for dynamically associating the offset threshold range with the real-time load rate, including: dividing the robot joint load rate interval and associating the threshold range. When the real-time load rate of the robot joint is less than or equal to 30%, the offset threshold range is set to 0.7° to 0.8°; when the real-time load rate of the robot joint is greater than 30% and ≤70%, the offset threshold range is set to 0.5° to 0.6°; when the real-time load rate of the robot joint is greater than 70%, the offset threshold range is set to 0.3° to 0.4°.

[0073] The load rate ranges are based on the nonlinear attenuation characteristics of the reducer's stiffness. A 30% load rate corresponds to the starting point of stiffness reduction, while a 70% load rate corresponds to the critical point of accelerated stiffness attenuation. The threshold range is set using inverse correlation logic, with the threshold decreasing as the load rate increases, ensuring earlier compensation triggering under high-load conditions. The threshold bandwidth is set to 0.1° to provide the control algorithm with a margin of adjustment to avoid oscillation caused by sudden threshold changes.

[0074] Specifically, the load rate monitoring module calculates the ratio of joint output torque to rated torque in real time. After eliminating high-frequency noise using a third-order Butterworth filter, the load rate is divided into three operating ranges. When the load rate is in the 0-30% range, a wide threshold band of 0.7-0.8° is used to allow the drive chain to undergo moderate elastic deformation without triggering compensation, thus avoiding oversensitivity under low loads. When the load rate enters the 70-100% range, a narrow threshold band of 0.3-0.4° is used to intervene in compensation early when the reducer stiffness significantly decreases, thereby preventing the accumulation of deformation energy. The threshold switching process utilizes a hysteresis control algorithm. When the load rate fluctuates near the interval boundary, the current threshold range is maintained until the load rate change exceeds the hysteresis band width of 5%, preventing frequent switching. This solution precisely matches the compensation triggering timing with the actual stiffness characteristics of the drive chain, maintaining optimal control response across the entire load rate range.

[0075] As a preferred embodiment, the solution of the present invention is specifically implemented as follows: During the process of generating the reference torque control command of the robot joint main motor, the preset load disturbance model is deployed in the real-time computing unit of the servo controller. The model receives the angle deviation signal θ from the first encoder. error The deviation value is the difference between the actual angle of the current motor output shaft and the theoretical target angle. The model also accesses the real-time joint load rate parameters provided by the joint load rate monitoring module. Mload This parameter is obtained by fusion calculation of current loop sampling and Kalman filter. The stiffness attenuation factor ω is generated by converting the temperature data collected by the main motor temperature sensor through a linear mapping table. When the main motor temperature is between 25℃ and 80℃, the value range of ω is 0.85 to 1.15. When the model performs decoupling operation, θ error With M load Perform product operation, then divide the result by stiffness attenuation factor ω, and finally output the main motor reference torque command τ main The torque command is transmitted to the main motor controller via the CAN bus at a refresh rate of 500 Hz, achieving active decoupling compensation for the elastic deformation of the transmission chain.

[0076] Through the above-mentioned technical solution, the present invention effectively solves the problem of trajectory deviation caused by the inability to decouple transmission chain deformation in traditional single-encoder systems. Through a three-dimensional decoupling mechanism that integrates angle deviation, load status, and temperature parameters in real time, the main motor torque command can actively offset the reverse torque generated by the elastic deformation of the reducer. While maintaining the stability of the main control loop, this solution suppresses the trajectory deviation of the end effector to within a range of ±0.2°. At the same time, through the temperature compensation function of the stiffness attenuation factor, the compensation misalignment caused by the decrease in material stiffness under high temperature conditions is avoided.

[0077] The present invention further proposes to dynamically decouple the dual encoder signal mismatch through a preset robot joint load disturbance model when generating a reference torque control command for the robot joint main motor. The robot joint load disturbance model satisfies the following formula: ; where τ main is the reference torque command of the main motor of the robot joint; where θ error is the deviation between the first angle signal and the target angle, and the target angle is the angle position that the output shaft of the robot joint motor should theoretically reach at the current moment; load is the real-time joint load rate of the robot joint; where ω is the stiffness attenuation factor of the robot joint main motor temperature mapping.

[0078] Among them, the robot joint load disturbance model includes a combination structure of proportional-integral-differential control terms and load-temperature compensation terms. The proportional-integral-differential control term consists of the instantaneous value of the angle deviation, the differential term, and the integral term. The load-temperature compensation term is a nonlinear function of the joint load rate and the stiffness attenuation factor. The stiffness attenuation factor is obtained by real-time mapping of the experimentally calibrated temperature-stiffness curve. For every 10°C increase in temperature, the stiffness attenuation factor decreases by 0.05-0.12. The calculation process of the load-temperature compensation term includes the following steps: first, the real-time joint load rate is input into the second-order polynomial function to calculate the basic compensation amount, then the stiffness attenuation factor is applied to the basic compensation amount as a gain coefficient, and finally the compensation amount is constrained within a preset range by a limiter.

[0079] Specifically, while the robot joint's main motor controller executes closed-loop position control, it collects motor temperature data in real time and calculates the corresponding stiffness reduction factor. As the motor temperature rises from 25°C to 75°C, the stiffness reduction factor decreases linearly from 1.0 to 0.82. At this point, the output of the load-temperature compensation term automatically increases by 22% to offset the decrease in drive train stiffness. In sudden load changes, such as from 30% to 80%, a second-order polynomial function causes the base compensation to overshoot, increasing to 1.3 times its steady-state value within 20ms and then decaying to its equilibrium value within 100ms. This dynamic compensation process, combined with the proportional-integral-derivative control term, ensures that the main motor torque command generates overtorque output at the initial stage of the load change, effectively suppressing the accumulation of drive train deformation. Experimental data demonstrate that this composite control model can reduce the end-point trajectory deviation to ±0.12° while reducing servo energy consumption by 15%.

[0080] As a preferred embodiment, the present invention is implemented as follows: During the generation of torque compensation control commands for the auxiliary motors of robot joints, a delay factor is defined as a proportional factor that adjusts the delay applied by the second angle signal during hysteresis compensation. This proportional factor is initialized by acquiring joint load rate data in real time and querying a preset load rate-delay factor mapping table. Specifically, when the joint load rate is between 0% and 30%, the initial value of the proportional factor is set to 0.9; when the load rate rises to between 30% and 70%, the proportional factor is adjusted to 1.0; if the load rate exceeds 70%, the proportional factor is increased to 1.2. Furthermore, by monitoring the main motor temperature data in real time, when the temperature exceeds 65°C but does not reach 85°C, if the current load rate is above 50%, the proportional factor is increased by 0.1 from the initial value; if the load rate is below 50%, the proportional factor is increased by 0.05. When the main motor temperature exceeds 85°C, all proportional factors are forcibly adjusted to 1.3 times their initial value, regardless of the load rate range. The scale factor update frequency is controlled within the range of 10 to 50 times per second to ensure the real-time performance of dynamic adjustments.

[0081] Through the above technical solution, the present invention realizes the precise dynamic adjustment of the delay time in the hysteresis compensation operation. By establishing a mapping relationship between the delay coefficient and the load rate and temperature parameters, the problem of traditional fixed delay compensators causing resonance under variable working conditions is effectively solved. Under low-load and low-temperature conditions, the compensation intensity is reduced to avoid high-frequency oscillations, and under high-load and high-temperature conditions, the compensation is strengthened to suppress low-frequency resonance. At the same time, the risk of thermal deformation is blocked through a temperature-triggered forced compensation mechanism. This solution significantly improves the stability of the closed-loop control of the transmission chain state and extends the service life of key components such as the reducer.

[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

[0083] The present invention relates to the field of robot motor control technology, and in particular to a robot joint motor collaborative driving method based on dual-encoder dynamic decoupling. The invention effectively overcomes the interference of transmission chain elastic deformation on control accuracy through real-time comparison of dual-angle signals at the robot motor end and the load end and a hysteresis compensation mechanism, significantly improving the robot joint trajectory tracking stability under variable load conditions. Under the load condition of the robot joint main motor, the transmission chain deformation interference can be isolated in real time, so that the torque command of the robot joint main motor can synchronously adapt to the load disturbance, and the terminal trajectory deviation can be suppressed outside the influence of unmeasurable external forces.

Claims

1. A robot joint motor collaborative driving method based on dual encoder dynamic decoupling, characterized in that: The following steps are involved: Step 1: collecting a first angle signal of the output shaft of the robot motor in real time through a first encoder installed on the output shaft of the robot motor; Step 2: Collect the second angle signal in real time through a second encoder installed at the load output end of the robot joint; Step 3: Generate a reference torque control instruction for the robot joint main motor according to the first angle signal, and transmit the instruction to the robot joint main motor controller; Step 4: performing a hysteresis compensation operation on the second angle signal to generate a hysteresis compensation signal; Step 5: Generate a robot joint auxiliary motor compensation torque control instruction based on the hysteresis compensation signal, and transmit the command to the robot joint auxiliary motor controller; Step 6: Calculate the difference between the second angle signal and the first angle signal in real time. When the absolute value of the difference exceeds a preset offset threshold: If the second angle signal is greater than the first angle signal, the output polarity of the compensation torque control instruction for the auxiliary motor of the robot joint is negative; If the second angle signal is smaller than the first angle signal, the output polarity of the compensation torque control instruction for the auxiliary motor of the robot joint is positive.

2. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to claim 1 is characterized in that: The first angle signal is the real-time rotation angle of the robot motor output shaft, and the second angle signal is the actual position angle of the robot joint end effector.

3. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to claim 1 is characterized in that: The hysteresis compensation operation in step 4 includes: Step 4-1: Set a dynamically adjustable delay time window for the robot joint auxiliary motor compensation torque command. The reference value of the time window is 0.8 to 1.5 times the closed-loop control period of the robot joint main motor position. Step 4-2: Query the preset mapping table to obtain the delay coefficient according to the real-time load rate of the robot joint and adjust the delay coefficient based on the temperature of the main motor of the robot joint; Step 4-3: Multiply the delay time window by the delay coefficient and apply the resultant multiplication to the second angle signal.

4. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to claim 3 is characterized in that: The setting logic of the delay coefficient in step 4-2 is: When the robot joint load rate is between 0% and 30%, the delay factor is set to 0.9; When the robot joint load rate is between 30% and 70%, the delay factor is set to 1.0; When the robot joint load rate is between 70% and 100%, the delay factor is set to 1.2; The delay coefficient is updated at a frequency of 10 to 50 times per second.

5. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to claim 3 is characterized in that: The step 4-2 of adjusting the delay coefficient based on the temperature of the robot joint main motor includes: Step 4-2-1: Monitor the temperature of the robot joint main motor in real time. When the temperature exceeds 65℃ but does not exceed 85℃: If the current robot joint load rate is greater than or equal to 50%, the delay coefficient will be increased by an additional 0.1; If the current robot joint load rate is less than 50%, the delay factor will be increased by an additional 0.05; When the temperature of the robot joint main motor exceeds 85°C, the delay coefficient of all robot joints under load rate is forced to be set to 1.3 times the original value.

6. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to any one of claims 2 to 5, characterized in that: The delay coefficient is a proportional factor used to adjust the length of the delay time applied to the second angle signal of the robot joint in the lag compensation operation.

7. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to claim 1 is characterized in that: The preset offset threshold in step 6 is the actual torsion angle deviation of the robot joint load output end relative to the robot motor output shaft, and the preset offset threshold range is 0.3° to 0.8°.

8. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to claim 7 is characterized in that: The setting of the preset offset threshold in step 6 includes: dividing the robot joint load rate interval and associating the threshold range. When the real-time load rate of the robot joint is less than or equal to 30%, the offset threshold range is set to 0.7° to 0.8°; when the real-time load rate of the robot joint is greater than 30% and ≤70%, the offset threshold range is set to 0.5° to 0.6°; when the real-time load rate of the robot joint is greater than 70%, the offset threshold range is set to 0.3° to 0.4°.

9. The robot joint motor collaborative driving method based on dual encoder dynamic decoupling according to claim 1, characterized in that: When generating the robot joint main motor reference torque control instruction according to the first angle signal in step 3, the dual encoder signal mismatch is dynamically decoupled by a preset robot joint load disturbance model, and the robot joint load disturbance model satisfies the following formula: ; Where τmain is the reference torque command of the main motor of the robot joint; Where θerror is the deviation between the first angle signal and the target angle, and the target angle is the angle position that the output shaft of the robot joint motor should theoretically reach at the current moment; Where Mload is the real-time joint load rate of the robot joint; Where ω is the stiffness attenuation factor of the robot joint main motor temperature mapping.