Robot control method, device and equipment based on digital twinning and medium

Abstract

The embodiment of the invention provides a robot control method, device and equipment based on digital twinning and a medium, and relates to the technical field of robots. The method comprises the steps that firstly, the robot is controlled to execute a target task, and the pose of the robot after the first step is executed is obtained through measurement of all sensors on the robot and serves as the measurement pose; simulating the process of executing the first step by the digital twin model simulation robot to obtain a pose of the digital twin model after executing the first step, and taking the pose as a theoretical pose; if the error of the measurement pose comprises a static error, compensating the measurement pose according to a first error compensation method to obtain a real pose of the robot; and the robot is controlled to move from the real pose to the theoretical pose, and after the robot moves to the theoretical pose, the robot is controlled to continue executing the follow-up steps of the target task. Therefore, the high-performance control requirement of the robot in a dynamic environment can be met.

Description

Technical Field

[0001] This invention relates to the field of robotics, and in particular to robot control methods, devices, equipment and media based on digital twins. Background Technology

[0002] With the continuous advancement of technology and the expansion of the market, the application scenarios of robots are also increasing. Robots have been widely used in fields such as automobiles, electronics, food processing, photovoltaics, and metal processing, becoming an indispensable highly automated device in modern production and life. The dynamic coordination ability and result stability of robots are also receiving increasing attention.

[0003] In related technologies, linear PID (Proportion-Integral-Derivative) control systems are commonly used to control robots. A PID control system linearly controls the robot by combining the proportional (P), integral (I), and derivative (D) values ​​of the deviation between a given value r(t) and the actual output value y(t) to form the control quantity. This method is simple in structure, easy to implement, and performs well in many linear or near-linear systems. However, robot systems are inherently complex systems with strong nonlinearity and strong coupling. Their dynamic characteristics change drastically with variations in pose, load, and velocity. The fixed parameters of linear PID systems are insufficient to handle these globally changing dynamic characteristics, making it difficult to meet the high-performance control requirements of robots in dynamic environments. Summary of the Invention

[0004] The purpose of this invention is to provide a robot control method, apparatus, device, and medium based on digital twins to meet the high-performance control requirements of robots in dynamic environments. The specific technical solution is as follows:

[0005] A first aspect of this application provides a robot control method based on digital twins, the method comprising:

[0006] The robot is controlled to perform the first step of the target task, and the robot's pose after performing the first step is measured by various sensors on the robot, which is used as the measured pose.

[0007] The process of the robot performing the first step is simulated by a digital twin model to obtain the pose of the digital twin model after performing the first step, which is used as the theoretical pose.

[0008] In response to the fact that the error in the measured pose includes a static error, the measured pose is compensated according to the first error compensation method to obtain the true pose of the robot;

[0009] Control the robot to move from the actual pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0010] In one possible implementation,

[0011] The first error compensation method is obtained in advance through the following means, including:

[0012] The robot is controlled to perform multiple second steps, and the pose of the robot after each second step is measured by the sensors on the robot, which serves as the target measurement pose for each second step.

[0013] The process of the robot executing each of the second steps is simulated by a digital twin model to obtain the pose of the digital twin model after executing each of the second steps, which serves as the target theoretical pose for each of the second steps.

[0014] An error compensation method is determined that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each of the second steps, and is used as the first error compensation method.

[0015] In one possible implementation,

[0016] The first step of controlling the robot to perform the target task includes:

[0017] The robot is controlled by the first controller to perform the first step of the target task;

[0018] The method further includes:

[0019] In response to errors in the measured pose, including dynamic errors, the controller parameters of the first controller are adjusted in the direction of eliminating the dynamic errors;

[0020] The robot is controlled by the adjusted first controller to continue performing subsequent steps of the target task.

[0021] In one possible implementation,

[0022] The process of simulating the robot's execution of the first step using a digital twin model to obtain the pose of the digital twin model after executing the first step, as the theoretical pose, includes:

[0023] The process of controlling the digital twin model to execute the first step by the second controller is used to obtain the pose of the digital twin model after executing the first step, which is taken as the theoretical pose; wherein, the control strategy of the second controller is the same as the control strategy of the first controller.

[0024] In one possible implementation, the method further includes:

[0025] The first time taken for the robot to complete the first step is obtained, and the second time taken for the digital twin model to complete the first step during the simulation is obtained;

[0026] If the difference between the first duration and the second duration is greater than a preset difference threshold, then based on the relationship between the first duration and the second duration, the proportional gain parameter in the preset parameters of the first controller is adjusted to obtain the adjusted first controller;

[0027] Specifically, if the first duration is greater than the second duration, the proportional gain term parameter is decreased; or if the first duration is less than the second duration, the proportional gain term parameter is increased.

[0028] In one possible implementation, the first step of the target task is used to instruct the robot to move a first amplitude, and the first controller has a preset second amplitude.

[0029] The first step of controlling the robot to perform the target task through the first controller includes:

[0030] The first controller controls the robot to move to a third amplitude with a first driving force and controls the robot to move to a second amplitude with a second driving force, wherein the third amplitude is the difference between the first amplitude and the second amplitude, and the second driving force is less than the first driving force;

[0031] Adjusting the controller parameters of the first controller in the direction of eliminating the dynamic error includes:

[0032] If the robot shakes, the preset second amplitude in the first controller is increased;

[0033] If the robot does not vibrate and the preset adjustment end condition is not met, then the preset second amplitude in the first controller is reduced.

[0034] A second aspect of this application provides a robot control device based on digital twins, the device comprising:

[0035] The first control module is used to control the robot to perform the first step of the target task, and to measure the pose of the robot after performing the first step through various sensors on the robot, as the measured pose.

[0036] The operation simulation module is used to simulate the process of the robot performing the first step through a digital twin model, and obtain the pose of the digital twin model after performing the first step as the theoretical pose;

[0037] The pose compensation module is used to compensate the measured pose according to a first error compensation method in response to errors in the measured pose, including static errors, so as to obtain the true pose of the robot.

[0038] The second control module is used to control the robot to move from the actual pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0039] In one possible implementation,

[0040] The first error compensation method is obtained in advance through the following means, including:

[0041] The robot is controlled to perform multiple second steps, and the pose of the robot after each second step is measured by the sensors on the robot, which serves as the target measurement pose for each second step.

[0042] The process of the robot executing each of the second steps is simulated by a digital twin model to obtain the pose of the digital twin model after executing each of the second steps, which serves as the target theoretical pose for each of the second steps.

[0043] An error compensation method is determined that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each of the second steps, and is used as the first error compensation method.

[0044] In one possible implementation,

[0045] The first control module is specifically used for:

[0046] The robot is controlled by the first controller to perform the first step of the target task;

[0047] The device further includes:

[0048] A parameter adjustment module is used to adjust the controller parameters of the first controller in the direction of eliminating the dynamic error in response to the error of the measured pose, including the dynamic error.

[0049] The step execution module is used to control the robot to continue performing subsequent steps of the target task through the adjusted first controller.

[0050] In one possible implementation,

[0051] The operation simulation module is specifically used for:

[0052] The process of controlling the digital twin model to execute the first step by the second controller is used to obtain the pose of the digital twin model after executing the first step, which is taken as the theoretical pose; wherein, the control strategy of the second controller is the same as the control strategy of the first controller.

[0053] In one possible implementation,

[0054] The device further includes:

[0055] The duration acquisition module is used to acquire the first duration of the robot completing the first step and the second duration of the digital twin model completing the first step during the simulation process;

[0056] The parameter setting module is used to adjust the proportional gain parameter in the preset parameters of the first controller based on the relationship between the first duration and the second duration when the difference between the first duration and the second duration is greater than a preset difference threshold, so as to obtain the adjusted first controller.

[0057] Specifically, if the first duration is greater than the second duration, the proportional gain term parameter is decreased; or if the first duration is less than the second duration, the proportional gain term parameter is increased.

[0058] In one possible implementation, the first step of the target task is used to instruct the robot to move a first amplitude, and the first controller has a preset second amplitude; the first control module controls the robot to execute the first step of the target task through the first controller, including:

[0059] The first controller controls the robot to move to a third amplitude with a first driving force and controls the robot to move to a second amplitude with a second driving force, wherein the third amplitude is the difference between the first amplitude and the second amplitude, and the second driving force is less than the first driving force;

[0060] The parameter adjustment module adjusts the controller parameters of the first controller in order to eliminate the dynamic error, including:

[0061] If the robot shakes, the preset second amplitude in the first controller is increased;

[0062] If the robot does not vibrate and the preset adjustment end condition is not met, then the preset second amplitude in the first controller is reduced.

[0063] A third aspect of the embodiments of this application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0064] Memory, used to store computer programs;

[0065] The processor, when executing a program stored in memory, implements any of the above-described digital twin-based robot control methods.

[0066] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements any of the above-described digital twin-based robot control methods.

[0067] This invention also provides a computer program product containing instructions that, when run on a computer, causes the computer to execute any of the above-described digital twin-based robot control methods.

[0068] Beneficial effects of the embodiments of the present invention:

[0069] The robot control method, apparatus, device, and medium based on digital twin provided in this invention control the robot to execute the first step of a target task. The process of executing the first step is simulated using a digital twin model. Then, the robot's measured pose after executing the first step is obtained through various sensors on the robot, and the theoretical pose of the digital twin model after executing the first step is acquired. Since the robot's measurement results may be affected by errors caused by the sensors themselves, the robot's measured pose is inaccurate. To overcome the above technical problem, this application compensates for the measured pose according to a first error compensation method, thereby obtaining the robot's true pose. After controlling the robot to move from the true pose to the theoretical pose, the robot continues to execute subsequent steps. This can reduce (or even eliminate) the measurement pose error, improve the robot's control accuracy, and thus meet the high-performance control requirements of the robot in dynamic environments.

[0070] Of course, implementing any product or method of the present invention does not necessarily require achieving all of the advantages described above at the same time. Attached Figure Description

[0071] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings.

[0072] Figure 1 A first schematic diagram of a robot control method based on digital twin provided in an embodiment of this application;

[0073] Figure 2 A schematic diagram illustrating the determination process of the first compensation method provided in this application embodiment;

[0074] Figure 3 A second schematic diagram of a robot control method based on digital twin provided in an embodiment of this application;

[0075] Figure 4 A third schematic diagram of a robot control method based on digital twin provided in an embodiment of this application;

[0076] Figure 5 A fourth schematic diagram of a robot control method based on digital twin provided in an embodiment of this application;

[0077] Figure 6 A fifth schematic diagram of a robot control method based on digital twin provided in an embodiment of this application;

[0078] Figure 7 A schematic diagram of the structure of a robot control device based on digital twin provided in an embodiment of this application;

[0079] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0080] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art based on this application are within the scope of protection of the present invention.

[0081] With the rapid development of robotics technology, robots are rapidly becoming widespread in industrial, medical, rehabilitation, and service scenarios, especially humanoid robots, whose dynamic coordination capabilities and posture stability are receiving increasing attention. Current technologies primarily rely on linear PID control systems for robot posture control. However, this control strategy suffers from slow response, susceptibility to interference, and poor adaptability in complex tasks, making it difficult to meet the high-performance control requirements of humanoid robots in dynamic environments. Furthermore, the waist of a humanoid robot plays a crucial role in transmitting upper and lower limb movements, maintaining overall balance, and participating in task execution; however, linear control strategies struggle to cope with the highly nonlinear dynamic characteristics of the humanoid robot's waist.

[0082] Although related technologies have attempted to introduce nonlinear control algorithms and information fusion techniques to control the waist posture of humanoid robots, such as using extended Kalman filter algorithm for posture estimation and combining it with sliding mode controller to improve system stability, this method cannot achieve real-time mapping and predictive updates of robot joint states, that is, it cannot achieve visualization deduction and adaptive iteration of control strategies, resulting in low robot control performance and difficulty in meeting the high-performance control requirements of robots in dynamic environments.

[0083] For example, assuming the robot's waist is currently at a 40° angle and its current angular velocity is 1° / s, if the robot's waist is undergoing an angular acceleration of 0.8° / s... 2 If the robot rotates with uniform acceleration, its attitude angle should be 45° after 5 seconds. However, the attitude angle, angular velocity, and acceleration data collected by the robot's sensors calculate that the robot's attitude angle is 42° after 5 seconds. If there is an error in the robot's measurement results, and assuming that the robot's actual attitude angle after 5 seconds is 44° under the influence of this error, directly compensating the robot's attitude angle by 3° will lead to increasingly larger errors and lower robot control accuracy.

[0084] To improve the control performance of robots in dynamic environments, this application provides a robot control method, apparatus, device, and medium based on digital twins. The following detailed description of the robot control method based on digital twins provided in this application will be provided through specific embodiments.

[0085] In a first aspect, this application provides a robot control method based on digital twins, which can be applied to electronic devices. In a specific application, the electronic device can be a server, a terminal, or a robot. In another possible implementation, the electronic device may include a first electronic device and a second electronic device, where the first electronic device is a server or terminal, and the second electronic device is a robot. The first and second electronic devices are communicatively connected to jointly implement the robot control method based on digital twins provided in this application.

[0086] The robot control method based on digital twin provided in the embodiments of this application will be described below with reference to the accompanying drawings. Figure 1 As shown, Figure 1 A first schematic diagram of a robot control method based on digital twin provided in this application embodiment includes the following steps:

[0087] Step S10: Control the robot to execute the first step of the target task, and measure the robot's pose after executing the first step through the various sensors on the robot, and use it as the measured pose.

[0088] The target task is any task that the robot can perform in scenarios such as industry, medicine, rehabilitation, and service, such as material handling, delivering items to medical personnel, assisting people with mobility difficulties, and demonstration teaching. The target task is usually broken down into multiple steps. The first step can be any of these steps; for example, it can be the first step in the execution sequence of the target task or a step in another execution sequence. Taking the target task of delivering an item as an example, the steps could be: bending over, picking up the item from position 1, standing up, rotating, moving to position 2, and delivering the item. Therefore, the first step can be any of the steps mentioned above.

[0089] The pose of the robot after the first step is obtained by measuring the data of the robot after the first step by the various sensors on the robot. This means that the pose of the robot is obtained by fusing and calculating the data of the robot after the first step by the various sensors on the robot. The fusion calculation method can be any algorithm with fusion function, such as Kalman filtering, complementary filtering, etc. This application embodiment does not limit this.

[0090] The robot is equipped with multiple sensors to perceive its external environment and understand its own state, enabling it to maintain balance, interact, and complete complex tasks. Sensors for perceiving the external environment include one or more of the following: visual sensors, such as monocular and binocular cameras, depth cameras, and wide-angle cameras, which can identify people, objects, scenes, locate objects, recognize text, and understand complex environments; distance sensors, such as lidar, ultrasonic sensors, and infrared ranging sensors, which can detect the position and distance of obstacles; auditory sensors, such as array microphones, which can receive sound commands, perform speech recognition, sound source localization, and environmental sound analysis; and tactile sensors, such as pressure and temperature sensors installed on the robot's hands, chassis (or feet), and skin, which can sense touch, pressure distribution, control gripping force, and sense temperature and material properties. Sensors for understanding its own state include, but are not limited to, inertial measurement units, joint encoders, and torque sensors, which can detect the robot's pose.

[0091] Step S20: Simulate the process of the robot executing the first step using a digital twin model to obtain the pose of the digital twin model after executing the first step, which is used as the theoretical pose.

[0092] The digital twin model is built based on the robot's key mechanical properties, which are those that affect the robot's movement. Here, "movement" can refer to all possible movements of the robot, or it can refer to the movements that the user is concerned with. For example, in an embodiment where the user is only concerned with the robot's waist movement, "movement" here refers to the robot's waist movement and does not include the movement of the robot's chassis.

[0093] Key mechanical properties include, but are not limited to, the robot's mass distribution, inertial parameters, stiffness, damping, and joint constraints. Furthermore, because users are concerned with different movements in different scenarios, and these movements are influenced by different factors, the key mechanical properties differ across scenarios. For example, in a medical rehabilitation scenario, when assisting patients with rehabilitation movements such as standing and squatting, a humanoid robot can provide real-time feedback and predict the patient's state. For movements like standing and squatting, the posture of the waist significantly affects the stability of performing these movements. In this scenario, users are only concerned with the robot's waist movement, and the roughness of the ground does not affect this movement; therefore, ground roughness is not a key mechanical property in this scenario. However, in industrial humanoid robot scenarios, users are concerned not only with the robot's waist movement but also with the robot's chassis movement. Since ground roughness affects chassis movement, ground roughness is a key mechanical property in these scenarios.

[0094] Because digital twin models fully replicate the key mechanical properties of robots, and the virtual scene in which the digital twin model exists does not contain the errors faced by the robot, the first step of the digital twin model performing the target task can be considered the ideal benchmark for the first step of the robot performing the target task under undisturbed conditions. Therefore, the pose obtained through simulation using a digital twin model can be called the theoretical pose. Theoretical poses can be directly obtained from the simulation environment.

[0095] Step S30: In response to the error in the measured pose, including the static error, the measured pose is compensated according to the first error compensation method to obtain the robot's true pose.

[0096] It is understandable that the robot's measured pose may be affected by errors caused by the sensor itself or by the modeling deviation of the digital twin. Errors caused by the sensor itself or the modeling deviation of the digital twin have the characteristic of long-term stability. For example, the digital twin model and the robot continuously exhibit a constant pose offset under the same input conditions, i.e., static error.

[0097] Static error refers to the inherent, systematic pose deviation of a robot when it is stationary. It is discovered and determined by comparing the actual measured values ​​with the theoretical values ​​at multiple stationary points. Static error exhibits high consistency and repeatability across different points. Specifically, static error is determined based on the measured poses of the robot during the execution of other steps before the first step. For example, if the measured poses of the third, fourth, and fifth steps performed before the first step all show similar errors to the theoretical poses, then these errors are considered static errors.

[0098] In another possible embodiment, the type of error in the measured pose can be determined by pre-establishing a correspondence between the types of each step and the types of errors, based on the type of the first step. For example, assuming the robot can perform eight different types of steps, denoted as steps 1 to 8, the measurement pose errors of steps 1 to 4 are pre-defined as static errors, and the measurement pose errors of steps 5 to 8 are defined as dynamic errors. If the first step is step 4, then the measurement pose error can be considered a static error.

[0099] In step S30 above, the first compensation method is determined in advance based on the output difference between the robot and the digital twin model under undisturbed or static conditions. Undisturbed conditions refer to an idealized working state where the robot's mechanical structure is only subject to the force of its internal servo system, without any external forces or torques. Static conditions refer to a state where the robot is completely stopped, not working, and without energy flow. Under undisturbed or static conditions, the interference of dynamic errors can be avoided, and the static error can be accurately determined, thereby accurately determining the first compensation method for the static error. The specific determination process of the first compensation method is detailed below and will not be repeated here.

[0100] After compensating the entity pose according to the first compensation method, the robot's true pose can be obtained. Then, by controlling the robot to move from the true pose to the theoretical pose, the error in the measured pose can be eliminated.

[0101] Step S40: Control the robot to move from the real pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0102] Understandably, once the true pose and theoretical pose are determined, a path can be planned for the robot to move from the true pose to the theoretical pose. Controlling the robot to move from the true pose to the theoretical pose means controlling the robot to move along the planned path.

[0103] Understandable Figure 1 This is merely one possible example and does not limit the execution order of steps S20 and S30. Steps S20 and S30 can be executed simultaneously or sequentially, such as executing step S20 first and then step S30, or executing step S30 first and then step S20. All of these are possible.

[0104] The technical solution provided in this application embodiment controls the robot to execute the first step of the target task, and simulates the process of the robot executing the first step through a digital twin model. Then, the robot's measured pose after executing the first step is obtained by measuring the various sensors on the robot, and the theoretical pose of the digital twin model after executing the first step is obtained. Since the robot's measured pose may be affected by errors caused by the sensors themselves, the robot's measured pose is inaccurate. In order to overcome the above technical problems, this application compensates for the measured pose according to the first error compensation method, so as to obtain the robot's true pose. After controlling the robot to move from the true pose to the theoretical pose, the robot continues to execute subsequent steps. This can reduce or even eliminate the error of the measured pose, improve the robot's control accuracy, and thus meet the high-performance control requirements of the robot in dynamic environments.

[0105] Taking the waist rotation of the robot mentioned above as an example, in the embodiment of this application, the measured pose is compensated according to the first error compensation method to determine the true pose of the robot. For example, if the true pose angle of the robot is calculated to be 42°, and the waist of the robot is controlled to rotate to 45°, the measurement error of the robot can be eliminated and the control accuracy of the robot can be improved.

[0106] The process of determining the first compensation method is illustrated below:

[0107] See Figure 2 , Figure 2 A schematic diagram illustrating the determination process of the first compensation method provided in this application embodiment includes the following steps:

[0108] Step S21: Control the robot to execute multiple second steps respectively, and measure the pose of the robot after each second step through the sensors on the robot, which is used as the target measurement pose corresponding to each second step.

[0109] Step S22: Simulate the process of the robot executing each second step using a digital twin model to obtain the pose of the digital twin model after executing each second step, which serves as the target theoretical pose for each second step.

[0110] Step S23: Determine an error compensation method that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each of the second steps, and use it as the first error compensation method.

[0111] Each second step represents all types of operations that may be involved in the task the robot might perform. For example, assuming the robot is used for material handling, it might perform steps such as lifting the material, putting it down, and moving it while handling the material. Multiple steps can be selected as second steps from these steps. In one possible implementation, to include as many steps as possible, the selected second steps include all steps in the robot's material handling process, i.e., the aforementioned steps of lifting, putting down, and moving the material. In another possible implementation, some steps can be selected as second steps according to actual needs; for example, only the steps of lifting and putting down the material can be selected as second steps, without selecting the step of moving the material.

[0112] Step S21 is similar to step S10 above, and step S22 is similar to step S20 above. The only difference is that the steps in steps S21 and S22 that control the robot and digital twin model to perform are different from the steps in steps S10 and S20 above.

[0113] In step S23 above, determining an error compensation method that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each second step refers to, for each second step, calculating the pose difference between the target measured pose and the target theoretical pose corresponding to that second step, and then directly calculating the sum of the pose difference and the target measured pose as the compensated pose to achieve error compensation; alternatively, the pose difference and the target measured pose can be weighted and summed to obtain the result as the compensated pose to achieve error compensation; or the pose error can be corrected first using the target theoretical pose to obtain the corrected pose error, and then the sum of the corrected pose error and the target measured pose can be calculated as the compensated pose to achieve error compensation.

[0114] In another possible implementation, the error compensation method that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each of the second steps can also be: based on experience, a mapping relationship between the measurement pose error and the measurement pose is established in advance; the corresponding error is determined based on the measurement pose; then the error is corrected using the theoretical pose to obtain the corrected error; and then the sum between the corrected error and the measurement pose is calculated as the compensated pose to achieve error compensation.

[0115] For example, suppose the second step is bending over, the target theoretical pose is bending over 30°, the target measured pose is bending over 35°, and the error of the measured pose is determined to be -4° according to the mapping relationship. The error is corrected using the target theoretical pose to obtain the corrected error. Assuming the corrected error is -3°, the true pose after compensation is obtained by calculating the sum of the corrected error and the target measured error.

[0116] By selecting the embodiments of this application, and controlling the robot and digital twin model to execute each second step, an error compensation method is determined that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each second step. In the actual control process, the error of the measurement result can be compensated according to the determined error compensation method to improve the control accuracy of the robot.

[0117] It is understandable that the robot's measured pose may be affected not only by errors caused by the sensor or the digital twin modeling deviation itself, but also by errors caused by external environmental disturbances. However, external environmental disturbances do not cause the digital twin model and the robot to continuously exhibit a constant pose offset under the same input conditions. The error caused by the dynamic environment is called dynamic error.

[0118] See Figure 3 , Figure 3 A second schematic diagram of a robot control method based on digital twins provided in this application embodiment, the method including:

[0119] Step S101: The robot is controlled by the first controller to execute the first step of the target task, and the pose of the robot after executing the first step is measured by the sensors on the robot and used as the measured pose.

[0120] Step S20: Simulate the process of the robot performing the first step using a digital twin model to obtain the pose of the digital twin model after performing the first step, which is used as the theoretical pose.

[0121] Step S30: In response to the error in the measured pose, including the static error, the measured pose is compensated according to the first error compensation method to obtain the robot's true pose;

[0122] Step S40: Control the robot to move from the real pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0123] Step S50: In response to the error in the measured pose, including dynamic error, adjust the controller parameters of the first controller in the direction of eliminating dynamic error;

[0124] In step S60, the robot is controlled by the adjusted first controller to continue performing subsequent steps of the target task.

[0125] Steps S20 to S40 are described above and will not be repeated here. Step S101 is a detailed refinement of step S10. Steps S101, S50, and S60 will be explained in detail below.

[0126] In step S101 above, the first controller is a controller deployed for the robot. In one possible implementation, the first controller adopts a sliding mode control strategy, which constructs a sliding surface to make the robot's state move along the sliding surface, thereby achieving strong robust control against uncertainties, modeling errors and external disturbances.

[0127] It is understandable that if the error in the measured pose is a dynamic error, it can be assumed that the error is caused by an environmental disturbance that is not reproducible (or has an extremely low reproducibility frequency) and there is a deviation between the measured pose and the theoretical pose. However, such environmental disturbances cannot be fully modeled in a virtual simulation environment and can only be addressed by enhancing the robustness of the robot's controller. The controller's response mainly depends on the controller's control parameters. Therefore, when the error in the measured pose is a dynamic error, the controller parameters of the robot's controller can be adjusted.

[0128] In step S50 above, the controller parameters of the first controller are adjusted in the direction of eliminating dynamic error. It is understood that dynamic error is caused by the controller parameters of the first controller; therefore, dynamic error can be expressed as a function with the control parameters as independent variables. The gradient direction obtained by differentiating this function with respect to the control parameters is the direction for eliminating dynamic error. The adjustment magnitude can be adaptively adjusted according to the magnitude of the error between the measured pose and the theoretical pose, so that adjusting the controller parameters of the first controller can eliminate (or gradually eliminate) the dynamic error. Alternatively, the adjustment magnitude can be a preset fixed adjustment magnitude. When dynamic error is detected, the controller parameters of the first controller are adjusted according to the preset adjustment magnitude, so that the dynamic error decreases each time the controller parameters of the first controller are adjusted according to the preset adjustment magnitude, until the dynamic error is eliminated.

[0129] By using the embodiments of this application, when the error in measuring pose is a dynamic error, the error in measuring pose can be quickly eliminated by adjusting the robot's controller parameters, thereby enhancing the robustness of the robot's controller.

[0130] In order to ensure that the control results of the first step of the simulation of the digital twin model can serve as the theoretical benchmark for the control results of the robot executing the first step, the digital twin model and the robot should be controlled by completely identical controllers. In this paper, "completely identical" means that the controllers have the same control method, that is, the controllers will control the robot to execute the same steps when facing the same environment. However, it does not mean that the implementation methods of the controllers are the same. For example, the robot's controller can be implemented based on the first code, while the controller of the digital twin model can be implemented based on a second code that is different from the first code.

[0131] Based on this, see Figure 4 , Figure 4 A third schematic diagram of a robot control method based on digital twins provided in this application embodiment, the method including the following steps:

[0132] Step S101: The robot is controlled by the first controller to execute the first step of the target task, and the pose of the robot after executing the first step is measured by the sensors on the robot and used as the measured pose.

[0133] Step S201: The digital twin model is controlled by the second controller to execute the first step, and the pose of the digital twin model after executing the first step is obtained as the theoretical pose.

[0134] Step S30: In response to the error in the measured pose, including the static error, the measured pose is compensated according to the first error compensation method to obtain the robot's true pose;

[0135] Step S40: Control the robot to move from the real pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0136] Step S50: In response to the error in the measured pose, including dynamic error, adjust the controller parameters of the first controller in the direction of eliminating dynamic error;

[0137] In step S60, the robot is controlled by the adjusted first controller to continue performing subsequent steps of the target task.

[0138] Steps S101 and S30 to S60 are described above and will not be repeated here. Step S201 is the specific implementation of step S20 above.

[0139] In step S201, the second controller is a controller deployed for the digital twin model. The first controller and the second controller can be the same controller or different controllers, but the control strategies of the first controller and the second controller are the same. The control strategy includes, but is not limited to, control algorithms and parameter settings.

[0140] In one possible implementation, both the first controller and the second controller are sliding mode controllers with the same parameters.

[0141] By using the embodiments of this application, the first controller and the second controller adopt the same control strategy, which enables the control result of the first step of the control digital twin model simulation to serve as the theoretical benchmark for controlling the robot to execute the first step.

[0142] Understandably, the speed at which a robot performs a certain step should be within a suitable range. If the robot performs a certain step too quickly, it will affect the robot's stability, while if the speed is too slow, it will affect the robot's efficiency in performing the target task.

[0143] In order to improve the robot's stability and efficiency, one possible implementation is to compare the time taken for the digital twin model and the robot to perform the same steps. If the robot takes longer to perform the first step than the digital twin model, the robot is considered to be slow; otherwise, the robot is considered to be overshooting, i.e., too fast.

[0144] Based on this, see Figure 5 , Figure 5 A fourth schematic diagram of a robot control method based on digital twin provided in this application embodiment, the method including the following steps:

[0145] Step S101: The robot is controlled by the first controller to execute the first step of the target task, and the pose of the robot after executing the first step is measured by the sensors on the robot and used as the measured pose.

[0146] Step S20: Simulate the process of the robot performing the first step using a digital twin model to obtain the pose of the digital twin model after performing the first step, which is used as the theoretical pose.

[0147] Step S30: In response to the error in the measured pose, including the static error, the measured pose is compensated according to the first error compensation method to obtain the robot's true pose;

[0148] Step S40: Control the robot to move from the real pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0149] Step S50: In response to the error in the measured pose, including dynamic error, adjust the controller parameters of the first controller in the direction of eliminating dynamic error;

[0150] Step S60: Control the robot to continue executing subsequent steps of the target task through the adjusted first controller;

[0151] Step S70: Obtain the first time taken for the robot to complete the first step, and the second time taken for the digital twin model to complete the first step during the simulation, and calculate the difference between the first time and the second time.

[0152] Step S500: Determine whether the first duration is greater than the second duration and whether the difference is greater than a preset difference threshold.

[0153] If the first duration is greater than the second duration and the difference is greater than the preset difference threshold, then proceed to step S80; otherwise, proceed to step S90.

[0154] Step S80: Decrease the proportional gain parameter in the preset parameters of the first controller;

[0155] Step S90: Increase the proportional gain parameter in the preset parameters of the first controller.

[0156] Steps S101 to S60 are described above and will not be repeated here. The preset difference threshold is set based on experience and can be any natural number; this embodiment does not limit this. It is understood that... Figure 5 This is only one possible embodiment. In other possible embodiments, steps S70 to S90 may be executed after step S20 and before steps S30 and S50, or they may be executed simultaneously with steps S30 and S50, or they may be executed after steps S30 and S50. This application embodiment does not limit this.

[0157] When the first duration is greater than the second duration and the difference is greater than the preset difference threshold, the robot's reaction speed is slow. By increasing the proportional gain parameter in the preset parameters of the first controller, the robot's dynamic response speed can be improved. If the first duration is less than the second duration and the difference is greater than the preset difference threshold, the robot will overshoot. By automatically decreasing the proportional gain parameter in the preset parameters of the first controller, the robot's instability risk can be reduced.

[0158] Understandably, when the first duration equals the second duration or the difference between the first duration and the second duration is not greater than a preset difference threshold, it can be assumed that the robot does not have a slow reaction speed or overshoot phenomenon, and the parameters of the first controller do not need to be adjusted.

[0159] By selecting the embodiments of this application, the time taken for the robot and the digital twin model to perform the same step is used to determine whether the robot's speed in performing the first step is too fast or too slow. When the robot's speed in performing the first step is unreasonable, the proportional gain term of the preset parameters of the first controller can be adjusted to improve the robot's stability and operating speed.

[0160] Furthermore, suppose the robot needs to move to a certain pose during the first step, but the controller's control method is unreasonable, causing the robot to move beyond that pose after reaching it. In this case, the controller will control the robot to move back to that pose, which may again exceed the pose. This cycle repeats, causing the robot to experience high-frequency jitter near that pose during the first step. This high-frequency jitter may lead to task failure. For example, if the first step involves grasping material, high-frequency jitter in the robot may prevent it from successfully grasping the material.

[0161] In one possible implementation, the first step of the target task is used to instruct the robot to move a first amplitude, and a second amplitude is preset in the first controller. For example... Figure 6 As shown, Figure 6 A fifth schematic diagram of a robot control method based on digital twins provided in this application embodiment includes the following steps:

[0162] Step S1011: The robot is driven to move to a third amplitude by the first driving force through the first controller, and driven to move to a second amplitude by the second driving force. The robot's pose after performing the first step is measured by the sensors on the robot and used as the measured pose.

[0163] Among them, the third amplitude is the difference between the first amplitude and the second amplitude, and the second driving force is less than the first driving force;

[0164] Step S20: Simulate the process of the robot performing the first step using a digital twin model to obtain the pose of the digital twin model after performing the first step, which is used as the theoretical pose.

[0165] Step S30: In response to the error in the measured pose, including the static error, the measured pose is compensated according to the first error compensation method to obtain the robot's true pose;

[0166] Step S40: Control the robot to move from the real pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0167] Step S501: In response to the error in the measured pose, including dynamic error, determine whether the robot is shaking;

[0168] Step S502: If jitter occurs, increase the preset second amplitude in the first controller to obtain the adjusted first controller;

[0169] Step S503: If no jitter occurs and the preset adjustment end condition is not met, then reduce the preset second amplitude in the first controller to obtain the adjusted first controller.

[0170] Step S60: Control the robot to continue executing subsequent steps of the target task through the adjusted first controller;

[0171] Steps S20 to S60 are described above and will not be repeated here. Step S1011 is a specific implementation of step S101, and steps S501 to S503 are specific implementations of step S50. Steps S1011 and S501 to S503 will be explained in detail below.

[0172] In step S1011 above, there can be one or more second amplitudes. When there are multiple second amplitudes, each second amplitude corresponds to a second driving force. For example, suppose the first step of controlling the robot to perform the target task through the first controller is: controlling the robot to move at amplitude 1 with driving force 1, controlling the robot to move at amplitude 2 with driving force 2, and controlling the robot to move at amplitude 3 with driving force 3. Then driving force 1 is the first driving force, amplitude 1 is the third amplitude, driving force 2 and driving force 3 are the second driving forces, and amplitude 2 and amplitude 3 are the second amplitudes.

[0173] It is understandable that when the robot moves two amplitudes in succession, the second driving force corresponding to the amplitude of the later movement is less than the second driving force corresponding to the amplitude of the earlier movement. For example, in the above embodiment, the driving force 3 is less than the driving force 2.

[0174] It is understandable that when a robot shakes, the shaking will produce specific frequency characteristics in the motion signal. Sensors on the robot can capture these shaking signals, so whether shaking has occurred can be determined by whether the sensors have captured the shaking signals. In step S01 above, other methods can also be used to determine whether shaking has occurred; this application does not limit the method for detecting robot shaking.

[0175] Understandably, in sliding mode control scenarios, the robot's pose after executing the first step can be considered as the sliding surface of the sliding mode control. Since the first step is executed with a larger first driving force followed by a smaller second driving force, it can be considered that there is a dead zone near the sliding surface. When the robot enters the dead zone, the driving force is reduced. Since the robot moves a second amplitude from the start of reducing the driving force to the completion of the first step, the second amplitude is the size of the dead zone. In other words, increasing the size of the dead zone is equivalent to increasing the second amplitude. That is, in sliding mode control scenarios, the aforementioned step S502 can be achieved by increasing the size of the dead zone, thereby reducing jitter.

[0176] Since the second driving force is less than the first driving force, the speed of the robot's third amplitude movement is greater than the speed of its second amplitude movement. If the robot does not vibrate, the speed at which the robot performs the first step can be increased by reducing the second amplitude.

[0177] Understandably, as the second amplitude decreases, the robot may tremble due to its small size, requiring an increase in the second amplitude to mitigate the tremor. To prevent the robot from adjusting the second amplitude indefinitely, a pre-set adjustment termination condition can be established. For example, the tremor amplitude can be detected; if the tremor amplitude decreases below a preset threshold after the nth adjustment, the preset adjustment termination condition is considered met. Alternatively, the second amplitude can be adjusted adaptively each time, and the preset adjustment termination condition is considered met when the second amplitude shrinks to a preset value. Another example is recording the occurrence of high-frequency tremors during the adjustment process; the adjustment after the high-frequency tremors occurs is considered the final adjustment amplitude. If the robot does not tremble and the previous adjustment of the second amplitude was the final adjustment amplitude, the preset adjustment termination condition is considered met.

[0178] By using the embodiments of this application, when the robot shakes, adjusting the second amplitude can reduce the robot's shaking while ensuring the speed of the robot performing the first step, thereby improving the success rate of the robot in performing the target task.

[0179] Through this application, in the field of medical rehabilitation, real-time feedback and predictive modeling can be used to dynamically adjust the angle and speed of rehabilitation movements, thereby achieving stable support and intervention control for patients during rehabilitation processes such as assisted standing, slow squatting, and lateral swaying, significantly improving safety. In the field of intelligent service robots, robots can maintain natural human dynamics during turning, bending, and tilting, while accurately tracking and coordinating feedback to user actions, enhancing the robot's "human-like" behavioral performance. In industrial humanoid robot scenarios, such as collaborative assembly, cargo handling, and loading / unloading operations, humanoid robots can maintain high-precision control in confined spaces, ensuring balance and coordination when performing tasks such as bending over to pick up objects and rotating for assembly. In the field of educational companion robots, through digital twin models driving motion prediction and feedforward correction, companion robots can exhibit agile posture changes during activities such as games, storytelling, and demonstration teaching with children, improving the user experience.

[0180] Corresponding to the first aspect mentioned above, a second aspect of the embodiments of this application provides a robot control device based on digital twins, such as... Figure 7 The diagram shown is a structural schematic of a robot control device based on a digital twin model provided in an embodiment of this application. The device includes:

[0181] The first control module 701 is used to control the robot to perform the first step of the target task, and to measure the pose of the robot after performing the first step through various sensors on the robot, as the measured pose.

[0182] The operation simulation module 702 is used to simulate the process of the robot performing the first step through a digital twin model, and obtain the pose of the digital twin model after performing the first step as the theoretical pose;

[0183] The pose compensation module 703 is used to compensate the measured pose according to the first error compensation method in response to the error of the measured pose including static error, so as to obtain the true pose of the robot.

[0184] The second control module 704 is used to control the robot to move from the real pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0185] In this embodiment, the robot is controlled to perform the first step of the target task. The process of the robot performing the first step is simulated by a digital twin model. Then, the robot's measured pose after performing the first step is obtained by measuring the various sensors on the robot, and the theoretical pose of the digital twin model after performing the first step is obtained. Since the robot's measurement results may be affected by errors caused by the sensors themselves, the robot's measured pose is inaccurate. In order to overcome the above technical problems, this application compensates for the measured pose according to the first error compensation method, so as to obtain the robot's true pose. After controlling the robot to move from the true pose to the theoretical pose, the robot continues to perform subsequent steps. This can reduce (or even eliminate) the error of the measured pose, improve the robot's control accuracy, and thus meet the high-performance control requirements of the robot in dynamic environments.

[0186] In one possible implementation, the robot control device based on digital twins provided in this application embodiment may include four subsystems: a digital modeling module, a state perception and filtering module, a virtual-real collaborative control module, and an error diagnosis and compensation module. The first control module 701 and the operation simulation module 702 constitute the virtual-real collaborative control module subsystem, while the pose compensation module 703 and the second control module 704 constitute the error diagnosis and compensation module subsystem. The digital modeling module subsystem is used to construct a digital twin model in a virtual simulation platform, and the state perception and filtering module subsystem is used to obtain the measured pose by fusing data measured by the robot's sensors. These four subsystems are efficiently linked through asynchronous threads and shared memory, supporting real-time processing.

[0187] In one possible implementation, the first error compensation method is obtained in advance by means of:

[0188] The robot is controlled to perform multiple second steps, and the pose of the robot after each second step is measured by the sensors on the robot, which serves as the target measurement pose for each second step.

[0189] The process of the robot executing each of the second steps is simulated by a digital twin model to obtain the pose of the digital twin model after executing each of the second steps, which serves as the target theoretical pose for each of the second steps.

[0190] An error compensation method is determined that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each of the second steps, and is used as the first error compensation method.

[0191] In one possible implementation,

[0192] The first control module is specifically used for:

[0193] The robot is controlled by the first controller to perform the first step of the target task;

[0194] The device further includes:

[0195] A parameter adjustment module is used to adjust the controller parameters of the first controller in the direction of eliminating the dynamic error in response to the error of the measured pose, including the dynamic error.

[0196] The step execution module is used to control the robot to continue performing subsequent steps of the target task through the adjusted first controller.

[0197] In one possible implementation,

[0198] The operation simulation module is specifically used for:

[0199] The process of controlling the digital twin model to execute the first step by the second controller is used to obtain the pose of the digital twin model after executing the first step, which is taken as the theoretical pose; wherein, the control strategy of the second controller is the same as the control strategy of the first controller.

[0200] In one possible implementation,

[0201] The device further includes:

[0202] The duration acquisition module is used to acquire the first duration of the robot completing the first step and the second duration of the digital twin model completing the first step during the simulation process;

[0203] The parameter setting module is used to adjust the proportional gain parameter in the preset parameters of the first controller based on the relationship between the first duration and the second duration when the difference between the first duration and the second duration is greater than a preset difference threshold, so as to obtain the adjusted first controller.

[0204] Specifically, if the first duration is greater than the second duration, the proportional gain term parameter is decreased; or if the first duration is less than the second duration, the proportional gain term parameter is increased.

[0205] In one possible implementation, the first step of the target task is used to instruct the robot to move a first amplitude, and the first controller has a preset second amplitude; the first control module controls the robot to execute the first step of the target task through the first controller, including:

[0206] The first controller controls the robot to move to a third amplitude with a first driving force and controls the robot to move to a second amplitude with a second driving force, wherein the third amplitude is the difference between the first amplitude and the second amplitude, and the second driving force is less than the first driving force;

[0207] The parameter adjustment module adjusts the controller parameters of the first controller in the direction of eliminating the dynamic error, resulting in an adjusted first controller, including:

[0208] If the robot shakes, the preset second amplitude in the first controller is increased;

[0209] If the robot does not vibrate and the preset adjustment end condition is not met, then the preset second amplitude in the first controller is reduced.

[0210] This invention also provides an electronic device, such as... Figure 8 As shown, it includes a processor 801, a communication interface 802, a memory 803, and a communication bus 804, wherein the processor 801, the communication interface 802, and the memory 803 communicate with each other through the communication bus 804.

[0211] Memory 803 is used to store computer programs;

[0212] When processor 801 executes a program stored in memory 803, it performs the following steps:

[0213] The robot is controlled to perform the first step of the target task, and the robot's pose after performing the first step is measured by various sensors on the robot, which is used as the measured pose.

[0214] The process of simulating the robot's first step using a digital twin model is used to obtain the pose of the digital twin model after the first step is executed, which is then used as the theoretical pose.

[0215] If the error in the measured pose is a static error, then the measured pose is compensated according to the first error compensation method to obtain the robot's true pose;

[0216] Control the robot to move from its actual pose to its theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

[0217] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0218] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0219] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0220] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0221] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, implements the steps of any of the above-described digital twin-based robot control methods.

[0222] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the digital twin-based robot control methods described in the above embodiments.

[0223] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0224] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0225] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the apparatus embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0226] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A robot control method based on digital twins, characterized in that, The method includes: The robot is controlled to perform the first step of the target task, and the robot's pose after performing the first step is measured by various sensors on the robot, which is used as the measured pose. The process of the robot performing the first step is simulated by a digital twin model to obtain the pose of the digital twin model after performing the first step, which is used as the theoretical pose. In response to the error in the measured pose, including static error, the measured pose is compensated according to the first error compensation method to obtain the robot's true pose; Control the robot to move from the actual pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

2. The method according to claim 1, characterized in that, The first error compensation method is obtained in advance through the following means, including: The robot is controlled to perform multiple second steps, and the pose of the robot after each second step is measured by the sensors on the robot, which serves as the target measurement pose for each second step. The process of the robot executing each of the second steps is simulated by a digital twin model to obtain the pose of the digital twin model after executing each of the second steps, which serves as the target theoretical pose for each of the second steps. An error compensation method is determined that can compensate for the difference between the target measured pose and the target theoretical pose corresponding to each of the second steps, and is used as the first error compensation method.

3. The method according to claim 2, characterized in that, The first step of controlling the robot to perform the target task includes: The robot is controlled by the first controller to perform the first step of the target task; The method further includes: In response to errors in the measured pose, including dynamic errors, the controller parameters of the first controller are adjusted in the direction of eliminating the dynamic errors; The robot is controlled by the adjusted first controller to continue performing subsequent steps of the target task.

4. The method according to claim 3, characterized in that, The process of simulating the robot's execution of the first step using a digital twin model to obtain the pose of the digital twin model after executing the first step, as the theoretical pose, includes: The process of controlling the digital twin model to execute the first step by the second controller is used to obtain the pose of the digital twin model after executing the first step, which is taken as the theoretical pose; wherein, the control strategy of the second controller is the same as the control strategy of the first controller.

5. The method according to claim 3, characterized in that, The method further includes: The first time taken for the robot to complete the first step is obtained, and the second time taken for the digital twin model to complete the first step during the simulation is obtained. If the difference between the first duration and the second duration is greater than a preset difference threshold, then based on the relationship between the first duration and the second duration, the proportional gain parameter in the preset parameters of the first controller is adjusted to obtain the adjusted first controller; Specifically, if the first duration is greater than the second duration, the proportional gain term parameter is decreased; or if the first duration is less than the second duration, the proportional gain term parameter is increased.

6. The method according to claim 3, characterized in that, The first step of the target task is used to instruct the robot to move a first amplitude, and the first controller has a preset second amplitude. The first step of controlling the robot to perform the target task through the first controller includes: The first controller controls the robot to move to a third amplitude with a first driving force and controls the robot to move to a second amplitude with a second driving force, wherein the third amplitude is the difference between the first amplitude and the second amplitude, and the second driving force is less than the first driving force; Adjusting the controller parameters of the first controller in the direction of eliminating the dynamic error includes: If the robot shakes, the preset second amplitude in the first controller is increased; If the robot does not vibrate and the preset adjustment end condition is not met, then the preset second amplitude in the first controller is reduced.

7. A robot control device based on digital twin, characterized in that, The device includes: The first control module is used to control the robot to perform the first step of the target task, and to measure the pose of the robot after performing the first step through various sensors on the robot, as the measured pose. The operation simulation module is used to simulate the process of the robot performing the first step through a digital twin model, and obtain the pose of the digital twin model after performing the first step as the theoretical pose; The pose compensation module is used to compensate the measured pose according to the first error compensation method if the error of the measured pose is a static error, so as to obtain the true pose of the robot. The second control module is used to control the robot to move from the actual pose to the theoretical pose, and after the robot moves to the theoretical pose, control the robot to continue to perform the subsequent steps of the target task.

8. A robot, characterized in that, The robot is equipped with a main control chip, which is used to control the waist of the robot according to any one of claims 1 to 6.

9. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1 to 6.

Images