Precision self-adaptive control system of high-torque-density cycloid speed reducer joint

Abstract

The invention relates to a precision self-adaptive control system for a high-torque-density cycloid speed reducer joint, and belongs to the technical field of robot joint driving and control. The method comprises the steps that a robot joint is driven to reach a designated position through a communication device and a positioner, and collected joint posture data are transmitted to a main control device; through a joint encoder equipped on an angle driving framework, attitude data during equipment movement are analyzed based on the joint encoder, and precision self-adaptive adjustment in the joint angle direction is carried out according to the input attitude data; calculating the actual deviation between the cycloid speed reducer and the target joint through a backlash deviation model; based on the actual deviation, a motor is controlled to drive the framework to move in combination with the corresponding initial compensation parameters; and real-time precision evaluation is carried out on the collected joint postures through a precision evaluation framework, and the optimal compensation position and related posture data are recorded after the evaluation is passed. Under the same output capacity, the small size, the small weight and the large output torque are achieved.

Description

Technical Field

[0001] This invention belongs to the field of robot joint drive and control technology, specifically relating to a precision adaptive control system for a high torque density cycloidal reducer joint. Background Technology

[0002] With the widespread application of robots in industrial manufacturing, medical rehabilitation, service robots, and special operations, higher requirements are being placed on the precision, stability, and adaptability of their joint transmission components. As a commonly used high-torque-density transmission component, cycloidal reducers are widely integrated into robot joint systems due to their small size, high load-bearing capacity, and high transmission efficiency. However, during long-term operation, cycloidal reducers are often affected by load fluctuations, friction and wear, assembly errors, and environmental factors, often resulting in backlash deviations and decreased posture accuracy, thus affecting the robot's performance in complex tasks. Most existing robot joint control systems rely on single encoder feedback or fixed-parameter compensation methods, which cannot accurately reflect the error changes of the cycloidal reducer under dynamic conditions in real time, nor can they achieve adaptive control of precision under different working environments. Furthermore, some systems lack in-depth analysis and multi-dimensional evaluation of posture data during the compensation process, resulting in limited compensation effects and potentially leading to compensation lag or over-correction in highly dynamic operations. Therefore, there is an urgent need for a comprehensive control system that can combine multi-sensor data fusion, backlash deviation modeling, dynamic compensation parameter library and precision adaptive optimization processing to improve the accuracy and stability of cycloidal reducer joints under multiple working conditions, and achieve smaller size and greater output torque with the same output capacity. Summary of the Invention

[0003] To address the aforementioned problems in the prior art, this invention provides a precision adaptive control system for the joints of a high torque density cycloidal reducer. The objective of this invention can be achieved through the following technical solutions: The acquisition and control module receives remote control commands through the mobile communicator deployed on the robot joint, drives the movement of the robot joint in conjunction with the integrated positioner, and transmits the acquired joint movement data to the main control device based on the motion sensor to generate accuracy judgment commands. The attitude adjustment module integrates an angle drive and a cycloidal reducer assembly. The angle drive is equipped with a joint encoder and generates attitude parameters with adjusted accuracy by inputting attitude data collected by the cycloidal reducer assembly. The backlash measurement module acquires the deviation data between the cycloidal reducer and the target joint surface in real time through the deviation sensor installed on the robot platform. Combined with the attitude and position information of the joint gimbal, the module calculates the actual deviation between the cycloidal reducer and the target joint through the backlash deviation model. The accuracy assessment module selects the compensation parameters of the target joint from the pre-established compensation parameter information library based on the actual deviation, controls the motor drive architecture to drive the cycloidal reducer component architecture to move along the rotation direction according to the compensation parameters, and simultaneously activates the attitude sensor to collect attitude data in real time, and transmits the collected attitude data to the main control device to generate the accuracy assessment result. The compensation drive module analyzes the accuracy evaluation results through the accuracy evaluation architecture to obtain the accuracy index value of each attitude data, and performs accuracy optimization processing based on the accuracy index value to generate the compensated accuracy control command.

[0004] Specifically, the joint movement data is generated into a motion dataset based on multi-sensor covariance cross-fusion, and the motion data is transmitted to the main control device through the motion sensor to generate accuracy determination instructions.

[0005] Specifically, the operation process of the backlash deviation model includes: The attitude angle is measured by combining the main joint encoder with the attitude sensor, and the position deviation is calculated according to the installation position of the joint gimbal and the current state of the environmental adaptability adjustment mechanism of the robot platform. The deviation between the cycloidal reducer and the target joint is obtained by the sensor. Based on the system parameters of the gimbal, the corresponding parameter values ​​are retrieved from the pre-established parameter database. The actual deviation between the cycloidal reducer and the target joint is calculated using the backlash deviation formula and compared. When the deviation exceeds the set threshold, the data re-acquisition and calculation process is triggered.

[0006] Specifically, the motor drive architecture is equipped with a position sensor, which monitors the position information of the cycloidal reducer assembly architecture along the rotation direction in real time and feeds the position information back to the main control device. The main control device completes the accuracy correction of the compensation process by comparing the feedback position information with the target compensation position.

[0007] Specifically, the compensation driving module performs real-time accuracy evaluation of the collected attitude data in the following way: The collected multidimensional attitude data is converted into standardized attitude data, retaining the key information of the attitude data, and preprocessed attitude data is generated by filtering and denoising. The gradient magnitude and gradient direction of the sampling points in the preprocessed attitude data are calculated, wherein the gradient magnitude is used to represent the magnitude of attitude change and the gradient direction is used to represent the directional information of the attitude edge. The overall accuracy of the attitude data is evaluated by analyzing the concentration and peak position of the gradient histogram. An orientation histogram of the attitude data is constructed based on gradient orientation information. The orientation distribution characteristics of the attitude edges are analyzed, and the accuracy index of the attitude data is calculated by combining the gradient magnitude and orientation information.

[0008] Specifically, the generation of the orientation histogram of the attitude data involves dividing the attitude data into local regions, calculating the accuracy index value of each local region, obtaining the average value of the accuracy index values ​​of all local regions as the overall accuracy index value of the attitude data, and dividing the size and position of the local regions based on the principle of region edge detail information.

[0009] Specifically, the accuracy optimization process involves constructing a training dataset by acquiring joint sample data, labeling the data in the training dataset, and generating an accuracy level model based on the labeled sample joint data. The accuracy level model ensures that the accuracy index values ​​of the attitude data collected during actual control meet the requirements of control analysis by setting accuracy threshold ranges for each type of joint.

[0010] Specifically, during the compensation process, the accuracy evaluation architecture calculates the accuracy index value of the attitude data by collecting multiple attitude data, and predicts the position of the joint gimbal corresponding to the accuracy peak based on the change sequence of the accuracy index value. The predicted position drives the motor to drive the movement direction of the architecture.

[0011] Specifically, the method for precision optimization processing is as follows: Based on the accuracy index value of the attitude data and the optimal compensation position information of the gimbal, the deviation details in the attitude data are recovered by performing deconvolution processing on the attitude data. Based on the requirements of control analysis in terms of torque and speed, and combined with the motor parameters of the robot platform, torque correction and speed adjustment are performed on the posture data; the optimized posture data is compared with the original acquired posture data, and the optimized parameter data is recorded. Based on joint control, the main control device organizes and analyzes the collected posture data, extracts key feature information of the joints, and compares and matches them with the pre-stored joint database to complete the rapid identification and classification of joints.

[0012] Specifically, the main control device has a built-in dynamic compensation mechanism, including a compensation control subsystem based on attitude feedback and deviation feedback: the attitude feedback compensation control subsystem calculates the required compensation amount according to the direction and magnitude of the attitude change, and controls the motor drive architecture to move the cycloidal reducer assembly architecture along the rotation direction by a corresponding compensation distance; the deviation feedback compensation control subsystem recalculates the actual deviation between the cycloidal reducer and the target joint based on the deviation change, and controls the motor drive architecture to perform compensation adjustment based on the corresponding compensation parameters selected from the compensation parameter database.

[0013] Specifically, the accuracy level model establishes an energy efficiency characteristic curve by monitoring the power consumption of the motor drive architecture under different load and speed conditions, and the main control device selects the corresponding optimal control parameters based on the energy efficiency characteristic curve.

[0014] Specifically, the compensation parameter information database includes a set of compensation parameters corresponding to different operating conditions. The compensation parameter information database is generated by experimental calibration data and historical operating data, and is dynamically updated based on real-time feedback during operation.

[0015] The beneficial effects of this invention are as follows: This invention provides a precision adaptive control system for the joint of a high-torque-density cycloidal reducer, enabling high-precision and stable control of the joint under complex working environments and multiple operating conditions. By setting up a data acquisition and control module, an attitude adjustment module, a backlash measurement module, a precision evaluation module, and a compensation drive module, the system can monitor and analyze the attitude data and backlash deviation of the cycloidal reducer joint in real time. Combined with the dynamic retardation of compensation parameters from the compensation parameter information database, it achieves adaptive optimization of joint precision. Compared with existing technologies, this invention has significant advantages in the following aspects: First, the system introduces a multi-sensor fusion and backlash deviation modeling method, which can effectively eliminate accuracy errors caused by joint wear, environmental changes, and load fluctuations; second, the use of a dual-encoder redundancy structure improves the reliability and safety of joint attitude measurement; third, based on gradient and orientation histogram analysis of attitude data, quantitative evaluation of accuracy indicators is achieved, avoiding the shortcomings of relying on single position feedback in traditional compensation processes; in addition, the system has a dynamic threshold adjustment and online compensation mechanism, which can improve response speed and energy efficiency while ensuring accuracy; finally, through the accuracy optimization processing of the compensation drive module, this invention can significantly reduce the risk of compensation lag and overcorrection, thereby improving the long-term operational stability and reliability of the cycloidal reducer joint in complex environments. Attached Figure Description

[0016] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0017] Figure 1 This is a flowchart illustrating the precision adaptive control system for the joint of a high torque density cycloidal reducer according to the present invention.

[0018] Figure 2 This is a schematic diagram of the structure of a precision adaptive control system for a high torque density cycloidal reducer joint according to the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0020] Please see Figure 1 A precision adaptive control system for the joints of a high torque density cycloidal reducer: The acquisition and control module receives remote control commands through the mobile communicator deployed on the robot joint, drives the movement of the robot joint in conjunction with the integrated positioner, and transmits the acquired joint movement data to the main control device based on the motion sensor to generate accuracy judgment commands. The attitude adjustment module integrates an angle drive and a cycloidal reducer assembly. The angle drive is equipped with a joint encoder and generates attitude parameters with adjusted accuracy by inputting attitude data collected by the cycloidal reducer assembly. The backlash measurement module acquires the deviation data between the cycloidal reducer and the target joint surface in real time through the deviation sensor installed on the robot platform. Combined with the attitude and position information of the joint gimbal, the module calculates the actual deviation between the cycloidal reducer and the target joint through the backlash deviation model. The accuracy assessment module selects the compensation parameters of the target joint from the pre-established compensation parameter information library based on the actual deviation, controls the motor drive architecture to drive the cycloidal reducer component architecture to move along the rotation direction according to the compensation parameters, and simultaneously activates the attitude sensor to collect attitude data in real time, and transmits the collected attitude data to the main control device to generate the accuracy assessment result. The compensation drive module analyzes the accuracy evaluation results through the accuracy evaluation architecture to obtain the accuracy index value of each attitude data, and performs accuracy optimization processing based on the accuracy index value to generate the compensated accuracy control command.

[0021] Specifically, the joint movement data is generated into a motion dataset based on multi-sensor covariance cross-fusion, and the motion data is transmitted to the main control device through the motion sensor to generate accuracy determination instructions.

[0022] Specifically, the operation process of the backlash deviation model includes: The attitude angle is measured by combining the main joint encoder with the attitude sensor, and the position deviation is calculated according to the installation position of the joint gimbal and the current state of the environmental adaptability adjustment mechanism of the robot platform. The deviation between the cycloidal reducer and the target joint is obtained by the sensor. Based on the system parameters of the gimbal, the corresponding parameter values ​​are retrieved from the pre-established parameter database. The actual deviation between the cycloidal reducer and the target joint is calculated using the backlash deviation formula and compared. When the deviation exceeds the set threshold, the data re-acquisition and calculation process is triggered.

[0023] In this embodiment, the specific formula used in the backlash deviation model is as follows: , , , Where Δθ is the joint's angular backlash deviation, θm is the theoretical joint angle measured by the main encoder, θs is the actual angle measured by the attitude sensor, and Δd is the actual position deviation between the output end of the cycloidal reducer and the target joint surface (x m y m , z m (x) represents the system parameters of the articulated gimbal and the theoretical position calculated by the main encoder. s y s , z s ) represents the actual position coordinates acquired by the sensor, Δtotal represents the overall backlash deviation of the joint, and w1 and w2 are the weighting coefficients selected from the compensation parameter information library.

[0024] Specifically, the motor drive architecture is equipped with a position sensor, which monitors the position information of the cycloidal reducer assembly architecture along the rotation direction in real time and feeds the position information back to the main control device. The main control device completes the accuracy correction of the compensation process by comparing the feedback position information with the target compensation position.

[0025] In this embodiment, the compensation drive module includes a compensation calculation unit, a drive signal adjustment unit, and an execution drive unit. The compensation calculation unit is connected to the data processing module of the main control device and is used to calculate the compensation amount required by the current joint under the target motion trajectory in real time based on the accuracy judgment command issued by the main control device and the backlash compensation coefficient, friction compensation coefficient, etc. in the compensation parameter information library. The drive signal adjustment unit is connected to the driver circuit of the angle drive architecture and is used to convert the compensation amount output by the compensation calculation unit into control signals in the form of PWM signals, current set values, etc., to correct the original drive signal and form a corrected drive signal. The execution drive unit includes servo motors and cycloidal reducers installed on each rotary joint. After receiving the corrected drive signal, it executes the corresponding rotation action, thereby realizing the precision control of attitudes such as pitch angle, roll angle, and yaw angle.

[0026] Specifically, the compensation driving module performs real-time accuracy evaluation of the collected attitude data in the following way: The collected multidimensional attitude data is converted into standardized attitude data, retaining the key information of the attitude data, and preprocessed attitude data is generated by filtering and denoising. The gradient magnitude and gradient direction of the sampling points in the preprocessed attitude data are calculated, wherein the gradient magnitude is used to represent the magnitude of attitude change and the gradient direction is used to represent the directional information of the attitude edge. The overall accuracy of the attitude data is evaluated by analyzing the concentration and peak position of the gradient histogram. An orientation histogram of the attitude data is constructed based on gradient orientation information. The orientation distribution characteristics of the attitude edges are analyzed, and the accuracy index of the attitude data is calculated by combining the gradient magnitude and orientation information.

[0027] In this embodiment, the attitude adjustment module consists of an angle drive architecture and a cycloidal reducer component architecture, and is equipped with two encoders, a main encoder and a secondary encoder. The main encoder provides real-time position and angle data, while the secondary encoder serves as a redundancy backup, monitoring the operating status of the main encoder in real time. When a data anomaly or signal interruption is detected in the main encoder, the system can automatically switch to the secondary encoder to ensure the continuity and reliability of the attitude adjustment process. The cycloidal reducer component architecture is as follows: Figure 2 As shown, it connects to the angle drive architecture to achieve high torque density transmission output. This component transmits power through built-in input and output shafts, and, combined with high-precision bearings and rigid couplings, ensures stable output torque during attitude adjustment.

[0028] Specifically, the generation of the orientation histogram of the attitude data involves dividing the attitude data into local regions, calculating the accuracy index value of each local region, obtaining the average value of the accuracy index values ​​of all local regions as the overall accuracy index value of the attitude data, and dividing the size and position of the local regions based on the principle of region edge detail information.

[0029] Specifically, the accuracy optimization process involves constructing a training dataset by acquiring joint sample data, labeling the data in the training dataset, and generating an accuracy level model based on the labeled sample joint data. The accuracy level model ensures that the accuracy index values ​​of the attitude data collected during actual control meet the requirements of control analysis by setting accuracy threshold ranges for each type of joint.

[0030] Specifically, during the compensation process, the accuracy evaluation architecture calculates the accuracy index value of the attitude data by collecting multiple attitude data, and predicts the position of the joint gimbal corresponding to the accuracy peak based on the change sequence of the accuracy index value. The predicted position drives the motor to drive the movement direction of the architecture.

[0031] Specifically, the method for precision optimization processing is as follows: Based on the accuracy index value of the attitude data and the optimal compensation position information of the gimbal, the deviation details in the attitude data are recovered by performing deconvolution processing on the attitude data. Based on the requirements of control analysis in terms of torque and speed, and combined with the motor parameters of the robot platform, torque correction and speed adjustment are performed on the posture data; the optimized posture data is compared with the original acquired posture data, and the optimized parameter data is recorded. Based on joint control, the main control device organizes and analyzes the collected posture data, extracts key feature information of the joints, and compares and matches them with the pre-stored joint database to complete the rapid identification and classification of joints.

[0032] Specifically, the main control device has a built-in dynamic compensation mechanism, including a compensation control subsystem based on attitude feedback and deviation feedback: the attitude feedback compensation control subsystem calculates the required compensation amount according to the direction and magnitude of the attitude change, and controls the motor drive architecture to move the cycloidal reducer assembly architecture along the rotation direction by a corresponding compensation distance; the deviation feedback compensation control subsystem recalculates the actual deviation between the cycloidal reducer and the target joint based on the deviation change, and controls the motor drive architecture to perform compensation adjustment based on the corresponding compensation parameters selected from the compensation parameter database.

[0033] Specifically, the accuracy level model establishes an energy efficiency characteristic curve by monitoring the power consumption of the motor drive architecture under different load and speed conditions, and the main control device selects the corresponding optimal control parameters based on the energy efficiency characteristic curve.

[0034] Specifically, the compensation parameter information database includes a set of compensation parameters corresponding to different operating conditions. The compensation parameter information database is generated by experimental calibration data and historical operating data, and is dynamically updated based on real-time feedback during operation.

[0035] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A high-torque-density cycloid reducer joint precision adaptive control system, characterized in that, The application relates to a robot joint precision compensation system, which comprises the following modules: a collection control module, which receives remote control instructions through a mobile communicator arranged on a robot joint, drives the movement of the robot joint in combination with an integrated positioner, and transmits collected joint movement data to a master control device to generate precision judgment instructions based on a movement sensor; a posture adjustment module, which integrates an angle drive and a cycloidal reducer assembly, wherein the angle drive is equipped with a joint encoder, and generates posture parameters after precision adjustment by inputting posture data collected by the cycloidal reducer assembly; a backlash ranging module, which obtains deviation data of a cycloidal reducer and a target joint surface in real time through a deviation sensor installed on a robot platform, combines posture and position information of the joint holder, calculates the actual deviation of the cycloidal reducer and the target joint through a backlash deviation model, and generates precision evaluation results by transmitting collected posture data to the master control device; a precision evaluation module, which selects compensation parameters of a target joint from a pre-established compensation parameter information library based on the actual deviation, controls a motor drive architecture to drive the cycloidal reducer assembly architecture to move in a rotating direction according to the compensation parameters, starts a posture sensor to collect posture data in real time, and transmits the collected posture data to the master control device to generate precision evaluation results; a compensation drive module, which analyzes the precision evaluation results through a precision evaluation architecture, obtains precision index values of each posture data, and generates compensation precision control instructions after precision optimization processing according to the precision index values.

2. The system of claim 1, wherein, The joint movement data is generated into a motion data set based on multi-sensor covariance cross fusion, and the motion data is transmitted to a master control device to generate precision judgment instructions.

3. The system of claim 1, wherein, The operation process of the backlash deviation model comprises the following steps: a posture angle is measured through the master joint encoder in combination with a posture sensor, a position deviation is calculated according to the installation position of the joint holder and the current state of the environment adaptability adjustment mechanism of the robot platform, and the deviation of the cycloidal reducer and the target joint is obtained through a sensor; corresponding parameter values are called from a pre-established parameter database based on system parameters of the joint holder, the actual deviation of the cycloidal reducer and the target joint is calculated through a backlash deviation formula and compared, and a data reacquisition and calculation process is triggered when the deviation exceeds a set threshold value.

4. The system of claim 1, wherein, The motor drive architecture is equipped with a position sensor, the position sensor monitors position information of the cycloidal reducer assembly architecture in a rotating direction in real time, and feeds back the position information to the master control device, and the master control device completes precision correction of the compensation process by comparing the feedback position information with a target compensation position.

5. The system of claim 1, wherein, The real-time precision evaluation mode of the collected posture data in the compensation drive module is as follows: multi-dimensional posture data is converted into standardized posture data, key information of the posture data is reserved, and preprocessed posture data is generated through filtering and denoising; the gradient amplitude and the gradient direction of a sampling point in the preprocessed posture data are calculated, wherein the gradient amplitude is used to represent the posture change amplitude, and the gradient direction is used to represent the direction information of the posture edge, and the overall precision of the posture data is evaluated by analyzing the concentration degree and the peak position of the gradient histogram; The direction histogram of the pose data is constructed based on gradient direction information, the direction distribution characteristics of the pose edge are analyzed, and the precision index value of the pose data is calculated based on the gradient amplitude and direction information.

6. The system of claim 4, wherein, The generation of the direction histogram of the pose data is performed by dividing the pose data into local regions, calculating the precision index value of each local region, and obtaining the average value of the precision index values of all local regions as the overall precision index value of the pose data; and the size and position of the local region are divided based on the principle of regional edge detail information.

7. The system of claim 6, wherein, The precision optimization process is performed by constructing a training data set based on the joint sample data, labeling the data in the training data set, and generating a precision level model based on the labeled sample joint data; the precision level model sets the precision threshold range corresponding to each type of joint to ensure that the precision index value of the pose data collected in the actual control process meets the requirements of control analysis.

8. The system of claim 2, wherein, The cycloid reducer is connected with the driving motor, the rotating power of the motor is transmitted to the inside of the reducer, and the relative meshing motion of the input end and the output end is completed according to the eccentricity of the column pin and the rolling tooth groove; the multi-tooth meshing is formed by the needle tooth shell and the cycloid wheel which are distributed in the outer periphery of the cycloid wheel assembly, and the power after reduction is output to the robot joint through the connection of the column pin and the cycloid wheel assembly.

9. The system of claim 6, wherein, The method for the precision optimization process is: Based on the precision index value of the pose data and the best compensation position information of the joint gimbal, the deviation details in the pose data are recovered by performing deconvolution processing on the pose data; Based on the requirements of control analysis in terms of torque and speed, and combined with the motor parameters of the robot platform, the torque correction and speed adjustment are performed on the pose data; the pose data after optimization processing is compared with the original collected pose data, and the parameter data of the optimization processing is recorded; Based on joint control, the main control device arranges and analyzes the collected pose data, extracts the key feature information of the joint, and compares and matches with the pre-stored joint database to complete the rapid identification and classification of the joint.

10. The system of claim 4, wherein, The main control device is built-in dynamic compensation mechanism, including compensation control subsystem based on pose feedback and deviation feedback: the compensation control subsystem of the pose feedback calculates the required compensation amount according to the direction and amplitude of the pose change, and drives the cycloid reducer assembly architecture to move a corresponding compensation distance in the rotating direction through the control of the motor drive architecture; the compensation control subsystem of the deviation feedback recalculates the actual deviation of the cycloid reducer and the target joint through the deviation change amount, and controls the motor drive architecture to perform compensation adjustment based on the corresponding compensation parameters selected from the compensation parameter database.

11. The system of claim 7, wherein, The precision level model establishes the energy efficiency characteristic curve by monitoring the power consumption of the motor drive architecture under different load and speed conditions, and selects the corresponding optimal control parameters based on the energy efficiency characteristic curve by the main control device.

12. The system of claim 1, wherein, The compensation parameter information library includes a set of compensation parameters corresponding to different working condition parameters, and is generated based on experimental calibration data and historical operation data, and is dynamically updated based on real-time feedback during operation.

Images