Variable stiffness actuator force planning and active stiffness matching anti-interference control method and system

Through stable contact front force planning and active flexible and anti-disturbance force control, the problems of instantaneous overshoot and environmental stiffness changes during robot grinding and polishing are solved, and high-precision force control and stiffness matching are achieved, which improves the surface quality and production efficiency of the workpiece .

CN119017373AActive Publication Date: 2024-11-26HUAZHONG UNIV OF SCI & TECH

Patent Information

Application Number
CN202410960743.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-26
Estimated Expiration
2044-07-17

AI Technical Summary

Technical Problem

Existing robotic grinding and polishing end effectors are prone to overshoot at the moment of contact, and the force speed planning ignores changes in environmental stiffness, resulting in inaccurate accuracy, and the force control accuracy decreases in complex surfaces, affecting the surface quality of the workpiece.

Method used

The stable contact front force planning method is used to constrain the collision force by controlling the total impulse of the input system, and optimize the force, position and stiffness control in combination with environmental stiffness, actuator stiffness, robot dynamics and kinematic constraints, and use environmental stiffness estimator and synovial membrane. The controller, RBF network and extended state observer perform active compliance and immunity control.

Benefits of technology

It effectively avoids force overshoot at the moment of contact, improves force control accuracy and stiffness estimation accuracy, improves the polishing quality and production efficiency of complex curved workpieces, and reduces labor costs and failure rates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119017373A_ABST
    Figure CN119017373A_ABST
Patent Text Reader

Abstract

The invention belongs to the field of robot control, and discloses a force planning and active stiffness matching anti-interference control method and system for a variable stiffness actuator at the tail end of a grinding and polishing robot, and the method comprises a force planning method before stable contact, a force planning method after stable contact and an active compliance anti-interference force control method. According to the force planning method before stable contact, the maximum collision force generated by collision is restrained by controlling the total impulse of an input system; according to the post-stable-contact force planning method, environment rigidity constraints, actuator rigidity constraints, robot dynamics and kinematics constraints, actuator driving error modeling and material removal constraints are considered, all the constraints are combined, optimization is carried out with the optimal efficiency T as the target, and optimal force, position and rigidity control target values are obtained; the active compliance anti-disturbance force control method comprises a noise and disturbance suppression environment rigidity estimator, a rigidity controller, a sliding mode controller, a radial basis function (RBF) network, an extended state observer (ESO) and a switching controller. The method shows excellent noise suppression and rigidity estimation precision. In the grinding and polishing process of a robot, the force control precision is excellent.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of robot control, and in particular relates to a method and system for force planning and active stiffness matching anti-disturbance control of a variable stiffness actuator at the end of a grinding and polishing robot. Background Art

[0002] With the increasing requirements for the surface quality of free-form workpieces, such as large fan blades, aircraft turbine blades and propellers, the polishing process becomes more complicated with the requirements for higher and higher precision levels. At present, most polishing still requires skilled technicians to perform manually, which has the problems of high risk, high labor intensity, high processing cost, low operating efficiency and low precision. In order to solve the limitations and shortcomings of manual polishing, robotic polishing systems and other types of automatic polishing equipment have been proposed and developed. Among them, industrial robots have the advantages of low price, large workspace and strong flexibility, but they lack force control technology for the workpiece polishing process. Therefore, force control technology is the key technology to realize the surface processing of free-form workpieces by robots, which needs to be solved urgently. Active compliant motion control method is the main research trend of high polishing quality. However, there are still the following major problems in the research field of force-controlled end effectors:

[0003] (1) The contact distance between the robot grinding and polishing force-controlled end effector tool and the workpiece is unknown, which can easily lead to overshoot at the moment of contact. The existing method uses force-position switching control, which on the one hand increases the demand for sensors, and on the other hand, it is difficult to ensure that the force caused by the collision at the moment of contact will not overshoot through the detection of force threshold.

[0004] (2) In order to ensure the material removal accuracy of robot grinding and polishing, it is often necessary to perform force-speed planning of the robot. However, the current planning ignores the changes in the force response performance of the actuator caused by changes in environmental stiffness, which leads to inaccurate force-speed planning results.

[0005] (3) Research on variable stiffness actuators at the end of grinding and polishing robots has gradually increased, but the current variable stiffness actuator stiffness control mainly focuses on the actuator's own stiffness control (i.e., controlling its own stiffness to reach the expected value), thus ignoring the estimation of the stiffness. The environmental stiffness estimation is seriously affected by noise, which also seriously affects the practical application of variable stiffness actuators.

[0006] (4) When the robot faces a complex curved surface, it is affected by multiple sources of disturbance and interference, which leads to a decrease in the robot's force control accuracy and a decrease in the surface quality of the workpiece. Summary of the invention

[0007] In view of the problems existing in the prior art, the present invention provides a method for force planning and active stiffness matching and anti-disturbance control of a variable stiffness actuator at the end of a grinding and polishing robot.

[0008] The present invention is implemented as follows: a method for force planning and active stiffness matching anti-disturbance control of a variable stiffness actuator at the end of a grinding and polishing robot, the method comprising: a stable pre-contact force planning method, a stable post-contact force planning method, and an active compliant anti-disturbance force control method;

[0009] The stable pre-contact force planning method is to control the total impulse of the input system to constrain the maximum collision force generated by the collision;

[0010] The stable post-contact force planning method considers environmental stiffness constraints, actuator stiffness constraints, robot dynamics and kinematics constraints, actuator drive error modeling, and material removal constraints, and combines all constraints to optimize with the optimal efficiency T as the goal to obtain the optimal force, position, and stiffness control target values;

[0011] The active compliance anti-disturbance force control method includes an environmental stiffness estimator for noise and disturbance suppression, a stiffness controller, a sliding film controller, a radial basis function (RBF) network, an extended state observer (ESO), and a switching controller.

[0012] Furthermore, the stable pre-contact force planning method specifically includes:

[0013] S1.1: First, establish the system's motion equation. Assuming that the total input impulse of the system before contact is p, and that the contact process is a spring model with an environmental stiffness of k, the system's motion equation can be established as Where m is the mass of the actuator, c f is the actuator damping, k is the environment stiffness, and F is the driving force of the actuator during collision;

[0014] S1.2: Obtain the general expression of the maximum collision force-impulse relationship. The initial solution of the motion equation can be obtained through the initial collision conditions:

[0015]

[0016] Where x is the displacement after collision with time t, and A, B, C, α, and β are auxiliary expression parameters. The maximum collision displacement of the system can be obtained as:

[0017]

[0018] Further, the general expression of the maximum collision force can be obtained as

[0019]

[0020] S1.3: Given an impulse control strategy, obtain a special expression for the maximum collision force-impulse relationship. The control impulse is step force control, and the driving force is controlled to be F before the total impulse of the system reaches the preset impulse. mAfter the total impulse of the system reaches the preset value, the driving force is maintained To ensure that the total impulse value of the system input remains unchanged;

[0021] When the collision occurs before the total impulse of the system reaches the preset impulse, the maximum collision expression can be further derived as:

[0022]

[0023] When the collision occurs after the total impulse of the system reaches the preset impulse, the maximum collision expression can be further derived as:

[0024]

[0025] S1.4: Design to obtain the maximum allowable total input impulse. When constraining, F max1 , F max2 Take the maximum collision force of the two for impulse constraint and calculate the maximum allowable impulse p max ,By controlling the maximum impulse allowed by the driver input system, it can ,effectively ensure that the contact collision force does not overshoot.

[0026] Furthermore, the environmental stiffness constraint considers the environmental stiffness variation range k of different workpiece positions. range ; The actuator stiffness constraint considers the variable stiffness actuator stiffness control capability limit constraint k limit The dynamic constraints include the positions of the six joints of the robot range , speed v range and acceleration a range Limit constraints.

[0027] Furthermore, the actuator error modeling is to establish the transfer function G(s) of the entire system and consider the response capability of the driver servo capability in each control cycle, thereby obtaining the actuator response capability E of each control cycle under the system position and system stiffness (including environmental stiffness and stiffness generated by the actuator) limit .

[0028] Furthermore, the material removal constraint considers the relationship between the material removal amount and the contact force to ensure the material removal accuracy constraint d limit .

[0029] Furthermore, based on the construction of the actuator control expected stiffness B-spline signal curve, the contact expected force B-spline signal curve, and the robot motion expected speed B-spline signal curve, the particle swarm optimization algorithm is used to optimize the B-spline nodes, with the optimization target as T and the optimization constraint as k. range ,k limit ,p range ,vrange ,a range ,E limit ,d limit , so as to plan the expected contact force, expected control stiffness and expected movement speed of the variable stiffness actuator at the end of the grinding and polishing robot, and realize the coordinated planning and control of force, speed and stiffness of the robot grinding and polishing.

[0030] Furthermore, the environment stiffness estimator obtains the environment stiffness by observing the environment, and inputs the environment stiffness into the stiffness controller, and the stiffness controller outputs a stiffness control signal to control the output stiffness of the actuator;

[0031] When there is an expected stiffness signal input, the stiffness controller controls the output stiffness based on the expected stiffness signal. When there is no expected stiffness signal input, the system damping ratio is 1, the expected stiffness signal is calculated by the feedback environmental stiffness estimation value, and outputted based on this.

[0032] Further, the synovial controller is responsible for tracking the magnitude of the contact force to ensure that the contact force is consistent with the desired force;

[0033] The RBF network is responsible for adaptively adjusting the parameters of the synovial controller to improve the robustness and anti-interference performance of the controller.

[0034] Furthermore, the extended state observer acts as a feedforward control to reduce the control burden of the synovial controller and further improve the anti-disturbance performance of the control system;

[0035] The switching controller is responsible for switching the force planning control strategy before and after contact.

[0036] Furthermore, the environment stiffness estimator is designed as follows.

[0037] S2.1: First construct the environmental stiffness estimate:

[0038]

[0039] In the formula is the initial estimated stiffness, the initial estimated distance between the actuator and the environment, describes the estimated stiffness, F0 represents the contact force, x e describes the estimated distance between the actuator and the environment, r1, r2 are identified constants;

[0040] S2.2: Construct normalized residual matrix

[0041] At time i, there is a sequence of force data F0 and displacement data X0 with a buffer length of n

[0042]

[0043] According to the environmental stiffness estimation rate, the estimated stiffness sequence can be calculated

[0044]

[0045] According to the three sequences of force data, displacement data and estimated stiffness sequence, the normalized residual can be calculated

[0046]

[0047] Here are the normalized residuals of force, displacement, and stiffness, respectively. are the average values ​​of force, displacement and stiffness, respectively. are 587 standard deviation values ​​of force, displacement, and stiffness, respectively; the online weighting matrix Ω is obtained by setting the retention threshold based on the normalized residual;

[0048] Based on the normalized residual, a 95% retention threshold can be set to obtain a real-time weighting matrix:

[0049]

[0050] S2.3: Calculate the stiffness estimation result of the normalized residual weighted matrix, multiply the weighted matrix by the stiffness estimation matrix, and calculate the mean after removing zero values, so as to calculate the output value of the stiffness estimation at time i:

[0051]

[0052] Among them, mean() refers to the average value after excluding zero values, K e (i) represents the stiffness estimate of the final output at time i.

[0053] Another object of the present invention is to provide a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the force planning and active stiffness matching anti-disturbance control method of the variable stiffness actuator at the end of the grinding and polishing robot.

[0054] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are: the problems existing in the technology, excellent noise suppression and rigid

[0055] First, the present invention effectively solves the estimation accuracy of the existing force-controlled end effector and shows excellent force control accuracy in robot grinding and polishing.

[0056] The invention effectively solves the problems existing in the existing force-controlled end effector technology, and exhibits excellent noise suppression and stiffness estimation accuracy. In robot grinding and polishing, it exhibits excellent force control accuracy.

[0057] Second, the expected benefits and commercial value of the technical solution of the present invention after transformation are:

[0058] Improve production efficiency: The present invention achieves efficient automated polishing operation by optimizing the force planning and active stiffness matching anti-disturbance control method of the variable stiffness actuator at the end of the robot polishing. Compared with traditional manual polishing, the present invention can significantly improve production efficiency and reduce labor costs and labor intensity.

[0059] Improve polishing quality: The force control accuracy and stiffness estimation accuracy of the present invention are extremely high, ensuring high-quality polishing of complex curved surface workpieces, reducing surface defects and errors, and improving the surface quality and accuracy of the final product.

[0060] Market competitive advantage: Since the present invention can solve key problems in the prior art, its technical solution has high market competitiveness and is expected to be widely used in high-end manufacturing industries such as aerospace, automobile manufacturing, and medical equipment, thereby bringing significant economic benefits and market share.

[0061] The technical solution of the present invention fills the technical gap in the industry at home and abroad:

[0062] Innovative force planning method: The present invention proposes a stable pre-contact and post-contact force planning method, which achieves precise control of the contact force by controlling the total impulse of the input system and combining multiple constraints for optimization, thus filling the gaps in force control accuracy and stiffness estimation of traditional methods.

[0063] Active compliant anti-disturbance control technology: By designing key components such as environmental stiffness estimator, stiffness controller, sliding membrane controller, RBF network, extended state observer (ESO) and switching controller for noise and disturbance suppression, the present invention realizes efficient anti-disturbance control and stiffness matching, filling the gap in the existing force-controlled end-effector technology in terms of anti-disturbance and compliance.

[0064] The technical solution of the present invention solves the technical problems that people have been eager to solve but have never been able to solve successfully:

[0065] The problem of overshoot of contact of force-controlled end effector: In the prior art, when the robot grinding and polishing force-controlled end effector tool contacts the workpiece, it is easy to cause overshoot, which affects the processing accuracy. The present invention effectively constrains the collision force by stabilizing the force planning method before contact, and solves the problem of overshoot at the moment of contact.

[0066] Impact of environmental stiffness changes on force-speed planning: Traditional planning methods ignore the impact of environmental stiffness changes, resulting in inaccurate force-speed planning results. This invention optimizes the force, position, and stiffness control target values ​​by considering environmental stiffness constraints and actuator stiffness constraints, ensuring the accuracy of planning results.

[0067] Problems of multi-source disturbances and interference on complex surfaces: The present invention designs an active compliant anti-disturbance force control method, which adopts technologies such as sliding film controller, RBF network and extended state observer to achieve effective suppression of multi-source disturbances and interference on complex surfaces, thereby improving the robot force control accuracy and the surface quality of workpiece grinding and polishing.

[0068] The technical solution of the present invention overcomes technical prejudice:

[0069] Fusion innovation of force control technology: The present invention combines a variety of advanced force control technologies and control methods, such as synovial membrane control, RBF network, extended state observation, etc., breaking through the limitations of traditional single control methods, overcoming the technical bias of force control accuracy and stiffness estimation in complex environments, and achieving more efficient and precise control effects.

[0070] Innovative design of environmental stiffness estimation: By constructing the environmental stiffness estimation rate and the normalized residual matrix, the present invention achieves accurate estimation of environmental stiffness, overcomes the technical bias of traditional environmental stiffness estimation methods that are greatly affected by noise, and provides a reliable foundation for achieving high-precision force control.

[0071] Force planning for precision material removal: Force overshoot at the moment of contact can damage the workpiece surface and reduce machining accuracy and quality. The present invention controls the total impulse of the input system to constrain the maximum collision force generated, thereby avoiding the problem of force overshoot at the moment of contact and improving the stability and reliability of the system.

[0072] Third, the present invention has achieved significant technical breakthroughs in the field of force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of the grinding and polishing robot, which is mainly reflected in the application of its key parameters, innovative algorithms and mathematical models. The following is a four-paragraph summary of the present invention's solution to the existing technical problems and the technological progress achieved:

[0073] First, the invention precisely controls the collision force and contact force during the grinding and polishing process through carefully designed key parameters, such as total impulse, estimated environmental stiffness, and actuator response capability within the control cycle, effectively avoiding workpiece damage or actuator failure caused by excessive impact force. The precise setting of these parameters not only improves the stability of the operation, but also ensures processing accuracy and efficiency.

[0074] Secondly, the present invention adopts advanced algorithms, such as Bayesian optimization, particle swarm optimization algorithm and adaptive adjustment strategy based on radial basis function (RBF) network, to achieve the optimization solution of force, position and stiffness control target values. The application of these algorithms enables the present invention to quickly and accurately adjust the control strategy in a complex and changeable working environment to adapt to different processing requirements and environmental changes.

[0075] Furthermore, the present invention constructs a series of innovative mathematical models, such as the system motion equation, the maximum collision force-impulse relationship expression, the environmental stiffness estimation rate, etc., which provide a solid theoretical basis for the precise control of the grinding and polishing process. These mathematical models not only reveal the physical laws in the grinding and polishing process, but also provide strong support for the design and optimization of the control algorithm.

[0076] In summary, the present invention successfully solves the technical problems of existing grinding and polishing robots in force planning and stiffness matching through precise key parameter setting, advanced algorithm application and innovative mathematical model construction. The implementation of the present invention not only significantly improves the accuracy and efficiency of grinding and polishing operations, but also reduces maintenance costs and failure rates, making important contributions to the development of industrial automation and intelligent manufacturing.

[0077] Fourth, the present invention proposes a new method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of the grinding and polishing robot. This method achieves precise control of force, position, and stiffness during the grinding and polishing process through force planning before and after stable contact, as well as active and compliant anti-disturbance force control. This innovative method not only improves the accuracy and efficiency of grinding and polishing operations, but also effectively reduces disturbances caused by factors such as collisions, and improves the robot's operating stability and reliability.

[0078] In terms of force planning before stable contact, the present invention constrains the maximum collision force during collision by controlling the total impulse of the input system, thereby avoiding damage to the workpiece and the actuator caused by excessive impact force. In terms of force planning after stable contact, the present invention comprehensively considers multiple constraints such as environmental stiffness, actuator stiffness, robot dynamics and kinematics, optimizes with the goal of optimal efficiency, obtains the optimal force, position, and stiffness control target values, and realizes fine control of the grinding and polishing process.

[0079] In terms of active compliance anti-disturbance force control, the present invention introduces multiple components such as environmental stiffness estimator, stiffness controller, sliding film controller, radial basis function (RBF) network, extended state observer (ESO) and switching controller, which together constitute a powerful anti-disturbance control system. This system can estimate the environmental stiffness in real time, adjust the output stiffness of the actuator, track the contact force, and adaptively adjust the controller parameters, thereby effectively suppressing the influence of noise and disturbance on the grinding and polishing process.

[0080] The implementation of the present invention successfully solves the technical problems of the existing grinding and polishing robot end effector in force planning and stiffness matching, and has achieved significant technical progress. Through the application of the present invention, precise control of the grinding and polishing process can be achieved, the processing quality and efficiency of the workpiece can be improved, and the maintenance cost and failure rate of the robot can be reduced. It has broad application prospects and practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 is a schematic diagram of a pre-contact force planning method provided by an embodiment of the present invention;

[0082] Figure 2 is a schematic diagram of a stable post-contact force planning method provided by an embodiment of the present invention;

[0083] Figure 3 is a schematic diagram of an active compliance anti-disturbance force control method provided by an embodiment of the present invention;

[0084] Figure 4 is a schematic diagram of the design of an environmental stiffness estimator provided by an embodiment of the present invention;

[0085] Figure 5 is a schematic diagram showing the performance of a stiffness estimator provided by an embodiment of the present invention;

[0086] Figure 6 It is a schematic diagram of force control performance of a variable stiffness actuator at the end of a grinding and polishing robot provided by an embodiment of the present invention, and a method for anti-disturbance control of force planning and active stiffness matching;

[0087] Figure 7 Schematic diagram of the grinding and polishing force control effect in a variable stiffness environment provided by an embodiment of the present invention;

[0088] Figure 8 is a schematic diagram of the instantaneous contact force response at a contact distance of 10 mm provided by an embodiment of the present invention;

[0089] Fig. 9 It is a schematic diagram of the instantaneous contact force response at a contact distance of 5 mm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0090] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0091] Example 1: Surface polishing of automobile parts

[0092] In the process of automobile manufacturing, many parts (such as engine housing, transmission shaft, etc.) need to be polished to improve their surface finish and corrosion resistance. Traditional polishing methods are difficult to meet the requirements of high precision and high efficiency, and have certain limitations in the processing of complex-shaped parts.

[0093] 1) Stable pre-contact force planning:

[0094] The system motion equation and impulse control strategy are used to control the total input impulse of the polishing robot before it contacts the surface of the component to avoid surface damage of the component due to sudden collision.

[0095] By presetting the impulse strategy, the magnitude and direction of the driving force are controlled before the polishing robot contacts the surface of the component to ensure that the maximum collision force is within the allowable range of the component material.

[0096] 2) Stable contact force planning:

[0097] The control target values ​​of polishing force, displacement and stiffness are optimized by combining the surface environment stiffness of automotive parts, the stiffness control capability of the polishing actuator, the dynamics and kinematic constraints of the robot, and the amount of material removal during the polishing process.

[0098] Ensure the force control and position control accuracy during the polishing process to ensure the polishing effect and consistency of the parts surface.

[0099] 3) Active and flexible anti-disturbance force control:

[0100] The environmental stiffness estimator is used to monitor the stiffness changes of the component surface in real time, and the estimation results are input into the stiffness controller.

[0101] The stiffness controller dynamically adjusts the output stiffness of the polishing actuator according to the real-time stiffness estimation results to ensure the force control stability between the robot and the component surface during the polishing process.

[0102] The synovial controller and RBF network work together to adaptively adjust the parameters of the synovial controller and improve the robustness and anti-disturbance of the system.

[0103] Through the above method, the polishing robot can achieve high-precision and high-efficiency polishing operations when polishing the surface of automotive parts, while effectively avoiding surface damage caused by collisions, thereby improving the surface quality and production efficiency of parts.

[0104] Example 2: Mobile phone screen glass grinding

[0105] Surface grinding of mobile phone screen glass is a key step to ensure its smoothness and transparency. Traditional grinding methods can easily cause scratches or cracks on the glass surface, affecting product quality and user experience.

[0106] 1) Stable pre-contact force planning:

[0107] The motion equation of the grinding robot is established, assuming the total input impulse of the system before contact, considering the environmental stiffness of the mobile phone screen glass, and ensuring that the collision force is constrained by controlling the total impulse before the grinding robot contacts the glass surface.

[0108] Through the expression of the maximum collision force-impulse relationship, a reasonable impulse control strategy is designed to avoid excessive collision force during the contact process, which may cause damage to the glass surface.

[0109] 2) Stable contact force planning:

[0110] The force control and position control target values ​​in the grinding process are optimized by combining the environmental stiffness variation range of the mobile phone screen glass, the stiffness control capability limit constraints of the grinding actuator, and the robot dynamics and kinematic constraints.

[0111] Consider the relationship between material removal and contact force during glass grinding to ensure the accuracy and uniformity of material removal and improve the smoothness and transparency of the glass surface.

[0112] 3) Active and flexible anti-disturbance force control:

[0113] The environmental stiffness estimator monitors the stiffness changes of the mobile phone screen glass surface in real time and inputs the stiffness controller, which then outputs the corresponding stiffness control signal to ensure the force control stability during the grinding process.

[0114] The sliding film controller and RBF network adaptively adjust the control parameters to enhance the robustness and anti-disturbance of the system and reduce the force control deviation caused by external disturbances during the grinding process.

[0115] The extended state observer is used as feedforward control to further reduce the control burden of the synovial controller and improve the anti-disturbance performance of the control system.

[0116] Through the above method, the grinding robot can achieve high-precision force control and position control during the grinding process of mobile phone screen glass, avoiding scratches or cracks on the glass surface, improving the surface quality and user experience of the product, and at the same time improving production efficiency.

[0117] In view of the problems existing in the prior art, the present invention provides a method for force planning and active stiffness matching and anti-disturbance control of a variable stiffness actuator at the end of a grinding and polishing robot. The present invention is described in detail below in conjunction with the accompanying drawings.

[0118] The method comprises a stable pre-contact force planning method, a stable post-contact force planning method, and an active compliant anti-disturbance force control method.

[0119] like Figure 1As shown, the pre-contact force planning method constrains the maximum collision force generated by the collision by controlling the total impulse of the input system. The steps are as follows:

[0120] S1.1 First, establish the system's motion equation. Assuming that the total input impulse of the system before contact is p, and the contact process is considered to be a spring model with an environmental stiffness of k, the system's motion equation can be established as Where m is the mass of the actuator, c f is the actuator damping, k is the environmental stiffness, and F is the driving force of the actuator during collision.

[0121] S1.2 Obtain the general expression of the maximum collision force-impulse relationship. The initial solution of the motion equation can be obtained through the initial collision conditions:

[0122]

[0123] Where x is the displacement after collision with time t, and A, B, C, α, and β are auxiliary expression parameters. The maximum collision displacement of the system can be obtained as:

[0124]

[0125] Further, the general expression of the maximum collision force can be obtained as

[0126]

[0127] S1.3 Given the impulse control strategy, obtain the special expression of the maximum collision force-impulse relationship. The control impulse is step force control, and the driving force is controlled to be F before the total impulse of the system reaches the preset impulse. m After the total impulse of the system reaches the preset value, the driving force is maintained To ensure that the total impulse value of the system input remains unchanged.

[0128] When the collision occurs before the total impulse of the system reaches the preset impulse, the maximum collision expression can be further derived as:

[0129]

[0130] Furthermore, when the collision occurs after the total impulse of the system reaches the preset impulse, the expression of the maximum collision can be further derived as:

[0131]

[0132] S1.4 Design to obtain the maximum allowable total input impulse. When constraining, F max1 , F max2 Take the maximum collision force of the two for impulse constraint and calculate the maximum allowable impulse p max,By controlling the maximum impulse allowed by the driver input system, it can ,effectively ensure that the contact collision force does not overshoot, e.g. Figure 7 , as shown in 8.

[0133] like Figure 2 As shown in the figure, the stable post-contact force planning method considers the environmental stiffness constraints, actuator stiffness constraints, robot dynamics and kinematics constraints, actuator drive error modeling, and material removal constraints, and combines all constraints to optimize with the optimal efficiency T as the goal to obtain the optimal force, position, and stiffness control target values.

[0134] Environmental stiffness constraint considers the environmental stiffness variation range k of different workpiece positions range .

[0135] Actuator stiffness constraint Considering variable stiffness actuator stiffness control capability limit constraint k limit .

[0136] The dynamic constraints include the positions of the six joints of the robot range , speed v range and acceleration a range Limit constraints.

[0137] Furthermore, the actuator error modeling is achieved by establishing the transfer function G(s) of the entire system and considering the response capability of the driver servo capability in each control cycle, thereby obtaining the actuator response capability E of each control cycle under the system position and system stiffness (including environmental stiffness and stiffness generated by the actuator). limit .

[0138] The material removal constraint considers the relationship between the material removal amount and the contact force to ensure the material removal accuracy constraint d limit .

[0139] Based on the B-spline signal curve of the expected stiffness of the actuator control, the B-spline signal curve of the expected contact force, and the B-spline signal curve of the expected speed of the robot motion, the particle swarm optimization algorithm is used to optimize the B-spline nodes, with the optimization target as T and the optimization constraint as k. range ,k limit ,p range ,v range ,a range ,E limit ,d limit Thus, the expected contact force, expected control stiffness, and expected robot motion speed of the variable stiffness actuator at the end of the grinding and polishing robot are planned to realize the coordinated planning control of force, speed, and stiffness of the robot grinding and polishing.

[0140] like Figure 3As shown, the active compliant anti-disturbance force control method includes an environmental stiffness estimator for noise and disturbance suppression, a stiffness controller, a sliding film controller, a radial basis function (RBF) network, an extended state observer (ESO), and a switching controller.

[0141] The environment stiffness estimator obtains the environment stiffness by observing the environment and inputs it into the stiffness controller. The stiffness controller outputs a stiffness control signal to control the output stiffness of the actuator.

[0142] When there is an expected stiffness signal input, the stiffness controller controls the output stiffness based on the expected stiffness signal. When there is no expected stiffness signal input, the system damping ratio is 1, and the expected stiffness signal is calculated through the feedback of the estimated value of the environmental stiffness, and the output is based on this.

[0143] The synovial controller is responsible for tracking the contact force to ensure that it is consistent with the desired force.

[0144] The RBF network is responsible for adaptively adjusting the parameters of the synovial controller to improve the robustness and anti-disturbance of the controller.

[0145] The extended state observer is used as feedforward control to reduce the control burden of the synovial controller and further improve the anti-disturbance performance of the control system.

[0146] The switching controller is responsible for switching the force planning control strategy before and after contact

[0147] like Figure 4 As shown, the environment stiffness estimator is designed as follows.

[0148] S2.1 First construct the environmental stiffness estimation rate

[0149]

[0150] In the formula is the initial estimated stiffness, the initial estimated distance between the actuator and the environment, describes the estimated stiffness, F0 represents the contact force, x e describes the estimated distance between the actuator and the environment, and r1 and r2 are identified constants.

[0151] S2.2 Constructing the normalized residual matrix

[0152] At time i, there is a sequence of force data F0 and displacement data X0 with a buffer length of n

[0153]

[0154] According to the environmental stiffness estimation rate, the estimated stiffness sequence can be calculated

[0155]

[0156] According to the three sequences of force data, displacement data and estimated stiffness sequence, the normalized residual can be calculated

[0157]

[0158] Here are the normalized residuals of force, displacement, and stiffness, respectively. are the average values ​​of force, displacement and stiffness, respectively. 587 standard deviation values ​​of force, displacement and stiffness respectively. The online weight matrix Ω is obtained by setting the retention threshold based on the normalized residual. Further, a retention 95% threshold can be set based on the normalized residual to obtain a real-time weight matrix.

[0159]

[0160] S2.3 calculates the stiffness estimation result of the normalized residual weighted matrix. Multiply the weighted matrix by the stiffness estimation matrix, and calculate the mean after removing zero values, so as to calculate the output value of the stiffness estimation at time i.

[0161]

[0162] Among them, mean() refers to the average value after excluding zero values, K e (i) represents the stiffness estimate of the final output at time i.

[0163] like Figure 5 The results show the performance of the stiffness estimator, which shows excellent noise suppression and stiffness estimation accuracy. Figure 6 The figure shows the performance demonstration results of the force planning and active compliant anti-disturbance control method of the variable stiffness actuator at the end of a grinding and polishing robot. In the robot grinding and polishing, it shows excellent force control accuracy.

[0164] An embodiment of the present invention provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the force planning and active stiffness matching anti-disturbance control method of the variable stiffness actuator at the end of the grinding and polishing robot.

[0165] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0166] Through simulation and experimental verification, Figure 7 , Figure 8 As shown, by comparing different control methods at different contact distances, the force planning provided by the present invention can effectively suppress the overshoot force of the collision at different distances.

[0167] Fig. 9 As shown, experiments were conducted using standard parts and the stiffness estimation method provided by the present invention can effectively reduce the stiffness estimation noise peak by more than 50%.

[0168] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for force planning and active stiffness matching anti-disturbance control of a variable stiffness actuator at the end of a grinding and polishing robot, characterized in that: The method includes: a stable pre-contact force planning method, a stable post-contact force planning method, and an active compliant anti-disturbance force control method; The stable pre-contact force planning method is to control the total impulse of the input system to constrain the maximum collision force generated by the collision; The stable post-contact force planning method considers environmental stiffness constraints, actuator stiffness constraints, robot dynamics and kinematics constraints, actuator drive error modeling, and material removal constraints, and combines all constraints to optimize with the optimal efficiency T as the goal to obtain the optimal force, position, and stiffness control target values; The active compliance anti-disturbance force control method includes an environmental stiffness estimator for noise and disturbance suppression, a stiffness controller, a sliding film controller, a radial basis function (RBF) network, an extended state observer (ESO), and a switching controller.

2. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of a grinding and polishing robot as claimed in claim 1 is characterized in that: The stable pre-contact force planning method specifically includes: S1.1: First, establish the system's motion equation. Assuming that the total input impulse of the system before contact is p, and the contact process is considered to be a spring model with an environmental stiffness of k, the system's motion equation can be established as Where m is the mass of the actuator, c f is the actuator damping, k is the environment stiffness, and F is the driving force of the actuator during collision; S1.2: Obtain the general expression of the maximum collision force-impulse relationship. The initial solution of the motion equation can be obtained through the initial collision conditions: Where x is the displacement after collision with time t, and A, B, C, α, and β are auxiliary expression parameters. The maximum collision displacement of the system can be obtained as: Further, the general expression of the maximum collision force can be obtained as S1.3: Given an impulse control strategy, obtain a special expression for the maximum collision force-impulse relationship. The control impulse is step force control, and the driving force is controlled to be F before the total impulse of the system reaches the preset impulse. m After the total impulse of the system reaches the preset value, the driving force is maintained To ensure that the total impulse value of the system input remains unchanged; When the collision occurs before the total impulse of the system reaches the preset impulse, the maximum collision expression can be further derived as: When the collision occurs after the total impulse of the system reaches the preset impulse, the maximum collision expression can be further derived as: S1.4: Design to obtain the maximum allowable total input impulse. When constraining, F max1 , F max2 Take the maximum collision force of the two for impulse constraint and calculate the maximum allowable impulse p max ,By controlling the maximum impulse allowed by the driver input system, it can ,effectively ensure that the contact collision force does not overshoot.

3. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of the polishing robot as claimed in claim 1 is characterized in that: The environmental stiffness constraint considers the environmental stiffness variation range k of different workpiece positions. range ; The actuator stiffness constraint considers the variable stiffness actuator stiffness control capability limit constraint k limit The dynamic constraints include the positions of the six joints of the robot range , speed v range and acceleration a range Limit constraints.

4. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of a grinding and polishing robot as claimed in claim 1, characterized in that: The actuator error modeling is to establish the transfer function G(s) of the whole system and consider the response capability of the driver servo capability in each control cycle, so as to obtain the actuator response capability E of each control cycle under the position and system stiffness (including the environment stiffness and the stiffness generated by the actuator) of the system. limit .

5. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of the polishing robot as claimed in claim 1, characterized in that: The material removal constraint is to consider the relationship between the material removal amount and the contact force, and ensure the material removal accuracy constraint d limit .

6. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of a grinding and polishing robot as described in claim 4 is characterized in that: Based on the construction of the actuator control expected stiffness B-spline signal curve, the contact expected force B-spline signal curve, and the robot motion expected speed B-spline signal curve, the particle swarm optimization algorithm is used to optimize the B-spline nodes, with the optimization target as T and the optimization constraint as k. range ,k limit ,p range ,v range ,a range ,E limit ,d limit , so as to plan the expected contact force, expected control stiffness and expected movement speed of the variable stiffness actuator at the end of the grinding and polishing robot, and realize the coordinated planning and control of force, speed and stiffness of the robot grinding and polishing.

7. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of a grinding and polishing robot as claimed in claim 1, characterized in that: The environment stiffness estimator obtains the environment stiffness by observing the environment, and inputs the environment stiffness into the stiffness controller, and the stiffness controller outputs a stiffness control signal to control the output stiffness of the actuator; When there is an expected stiffness signal input, the stiffness controller controls the output stiffness based on the expected stiffness signal. When there is no expected stiffness signal input, the system damping ratio is 1, the expected stiffness signal is calculated by the feedback environmental stiffness estimation value, and outputted based on this.

8. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of a grinding and polishing robot as claimed in claim 1, characterized in that: The synovial controller is responsible for tracking the contact force to ensure that the contact force is consistent with the expected force; The RBF network is responsible for adaptively adjusting the parameters of the synovial controller to improve the robustness and anti-interference performance of the controller.

9. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of a grinding and polishing robot as claimed in claim 1, characterized in that: The extended state observer acts as a feedforward control to reduce the control burden of the synovial controller and further improve the anti-disturbance performance of the control system; The switching controller is responsible for switching the force planning control strategy before and after contact.

10. The method for force planning and active stiffness matching anti-disturbance control of the variable stiffness actuator at the end of a grinding and polishing robot as claimed in claim 1, characterized in that: The environment stiffness estimator is designed as follows. S2.1: First construct the environmental stiffness estimate: In the formula is the initial estimated stiffness, The initial estimated distance between the actuator and the environment, describes the estimated stiffness, F0 represents the contact force, x e describes the estimated distance between the actuator and the environment, r1, r2 are identified constants; S2.2: Construct normalized residual matrix At time i, there is a sequence of force data F0 and displacement data X0 with a buffer length of n According to the environmental stiffness estimation rate, the estimated stiffness sequence can be calculated According to the three sequences of force data, displacement data and estimated stiffness sequence, the normalized residual can be calculated Here are the normalized residuals of force, displacement, and stiffness, respectively. are the average values ​​of force, displacement and stiffness, respectively. They are 587 standard deviation values ​​of force, displacement and stiffness; The online weight matrix Ω is obtained by setting the retention threshold based on the normalized residual; Based on the normalized residual, a 95% retention threshold can be set to obtain a real-time weighting matrix: S2.3: Calculate the stiffness estimation result of the normalized residual weighted matrix, multiply the weighted matrix by the stiffness estimation matrix, and calculate the mean after removing zero values, so as to calculate the output value of the stiffness estimation at time i: Among them, mean() refers to the average value after excluding zero values, K e (i) represents the stiffness estimate of the final output at time i.

11. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the force planning and active stiffness matching anti-disturbance control method of the variable stiffness actuator at the end of the grinding and polishing robot as described in any one of claims 1-10.

Citation Information

Patent Citations

  • Constant-force grinding device and grinding control method thereof

    CN109202688A

  • Target-grabbing-oriented space robot fixed time impedance control method

    CN116661342A

  • Robot force compliance control method based on variable speed impedance control

    CN116852356A

  • Robot constant force control grinding method based on PPO reinforcement learning

    CN117226613A

  • Robot control device

    JP1991086462A

Cited By

  • Variable stiffness actuator force-stiffness closed-loop control method and system based on force feedback

    CN119526418A