A robot processing vibration control method and system

Abstract

The present invention proposes a robot processing vibration control method and system, belonging to the technical field of robot vibration control. The method includes: decoupling the acceleration of the six-axis at the end to sense the vibration at the end of the robot; establishing a model of the robot joints and the end acceleration, and obtaining the relationship between the load side and the end acceleration through the Jacobian matrix; using a Kalman filter to estimate the joint speed on the load side; designing an active damping controller, analyzing the influence of the disturbing force, establishing a feedback term and introducing a feedforward term to eliminate the steady-state error; compensating the corrected speed in the active damping controller into the speed loop controller to achieve active vibration control. The present invention can accurately sense and control the vibration of a flexible joint robot, without modifying the body, introducing the six-axis acceleration signal at the end to achieve damping adaptive adjustment and improve stability.

Description

Technical Field

[0001] The present invention relates to the technical field of vibration control, and particularly to a vibration control method and system for robot machining. Background Art

[0002] With the continuous improvement of the intelligent level of the manufacturing industry, industrial robots are increasingly widely used in various fields of the manufacturing industry, especially in the machining processes with high-precision requirements such as assembly, welding, spraying, handling, inspection, and grinding. The application of industrial robots not only improves production efficiency, reduces labor costs, but also enhances the stability of product quality through automatic control. However, due to the existence of joint flexibility, the motion states of the motor side and the load side of the robot are different, that is, the input motion of the joint does not represent the output motion of the joint. At present, most robot joints on the market only install an encoder on the motor side. Therefore, it is difficult to ensure the system stability by controlling the robot through the existing motor-side motion state information, which restricts the operation performance of the robot.

[0003] Due to the joint flexibility, the robot system has a sensing mismatch characteristic, resulting in serious vibration problems during the robot motion process. To address the sensing mismatch problem, relevant scholars perceive the motion state of the robot load side by installing a secondary encoder or a torque sensor on the robot joint side, and then feedback it to the robot joint controller to achieve full closed-loop control, so as to improve the stability of the robot system. However, the method of adding sensors at the robot joints requires modifying the robot body structure, resulting in a more complex robot system structure and higher costs. In addition, the torque sensor obtains the deformation through the internal elastic structure, further reducing the stiffness of the robot joint. Moreover, it is somewhat dangerous to perform torque loop control on the robot, and most industrial robot underlying torque loop controls are not open, making it difficult to achieve torque control.

[0004] In order not to change the structure of the robot body, the second derivative of the joint position can be used to obtain the joint vibration state. However, due to the presence of noise, the joint vibration information will be submerged. It is also possible to install an acceleration sensor on the robot body to sense the state during the robot's movement and feedback it to the robot controller, thereby improving the robot's stability. Some scholars have achieved the perception and vibration control of the vibration information on the load side of the flexible joint robot by installing multiple acceleration sensors on the robot link and combining control algorithms. However, issues such as the layout and quantity of the acceleration sensors need to be considered, and the implementation method is relatively complex, making it difficult to improve the robot control efficiency. Another part of the scholars install a three-axis acceleration sensor at the end of the robot and estimate the motion state of the joint load side through intelligent algorithms. However, an industrial robot has six degrees of freedom in the Cartesian space, namely translation along three axes and rotation around three axes in the Cartesian space. It is difficult for a three-axis acceleration sensor to decouple the angular acceleration information at the end, which will result in a low decoupling accuracy of the motion state on the load side of the robot, leading to a poor vibration control effect of the robot. Installing a gyroscope sensor at the end of the robot to obtain the angular acceleration by differentiation will have problems such as large bias drift, low shock resistance, poor durability, and high noise, resulting in a low decoupling accuracy of the information on the joint load side and a poor vibration suppression effect.

[0005] In summary, there are deficiencies in the existing robot acceleration feedback vibration control technology. Currently, methods such as three-axis acceleration and gyroscopes are mostly used to sense and decouple the information on the load side. However, the decoupling accuracy is low, resulting in a poor vibration control effect of the robot. When using intelligent control algorithms such as sliding mode control and adaptive integral sliding mode controllers to achieve vibration suppression of the robot's movement, these methods mainly rely on disturbance observers or motor-side information to achieve control. In addition, the control algorithms have high requirements for the modeling accuracy of the robot, and the control structure is complex. Therefore, it is difficult for the existing technology to achieve precise, efficient, and stable vibration control of flexible joint robots. Summary of the Invention

[0006] To solve the above problems, the present invention proposes a robot processing vibration control method and system. By establishing a six-axis acceleration decoupling strategy at the end, using a load-side speed estimation method to obtain the speed information of the robot, establishing an active damping controller and real-time feedback it to the robot controller, the flexible vibration suppression of the robot joints is achieved. While improving the vibration control effect, the implementation difficulty and complexity are reduced, making the implementation process of vibration control smoother and more feasible.

[0007] To achieve the above object, the present invention adopts the following technical solutions:

[0008] In the first aspect, the present invention provides a robot processing vibration control method, including:

[0009] Construct a six-axis acceleration measurement unit at the end of the robot to obtain the six-axis acceleration at the end of the robot and the vibration state at the end.

[0010] Based on the motor side and the load side of the robot joint, construct a double-inertia model; use the Jacobian matrix to solve the six-axis acceleration measurement unit and the double-inertia model to obtain the mapping relationship between the six-axis acceleration in the Cartesian space and the joint space acceleration; substitute the end vibration state into the mapping relationship to obtain the acceleration of the load side of the robot joint.

[0011] According to the obtained acceleration of the load side of the joint, use a self-updating Kalman filter to estimate the speed of the load side joint to obtain the estimated speed of the load side joint.

[0012] Use the feedback control law to eliminate the steady-state error of the estimated speed of the load side joint to obtain the corrected speed.

[0013] Compensate the corrected speed into the robot speed loop controller, and adjust the control gain in the active damping controller according to the vibration state of the robot end to achieve active control of robot vibration.

[0014] Preferably, the six-axis acceleration at the end of the robot includes three-axis linear acceleration and three-axis angular acceleration; the acquisition process is specifically as follows:

[0015] The specific process of obtaining the six-axis acceleration at the end of the robot is as follows:

[0016] Construct a six-axis acceleration measurement unit for the end of the robot based on the spatial rigid body coordinate system; four triaxial accelerometers are respectively set on the X, Y, Z axes and the origin of the six-axis acceleration measurement unit.

[0017] According to the relationship between the spatial rigid body coordinate system and the inertial coordinate system, obtain the acceleration of the triaxial accelerometer in the six-axis acceleration measurement unit, and correct the output value of the accelerometer based on the specific force.

[0018] Convert the four triaxial accelerometers into twelve single-axis accelerometers and construct a system measurement equation.

[0019] Simplify the system measurement equation, substitute the corrected output value of the accelerometer, and obtain the six-axis acceleration.

[0020] Preferably, constructing the double-inertia model based on the motor side and the load side of the robot joint is specifically as follows: use the Lagrangian method to establish an n-degree-of-freedom double-inertia model for the robot joint; the double-inertia model includes the dynamic equation of the motor side and the dynamic equation of the load side.

[0021] Preferably, the Jacobian matrix is used to solve the six-axis acceleration measurement unit and the double-inertia model to obtain the mapping relationship between the six-axis acceleration and the joint-space acceleration in the Cartesian space; the end vibration state is substituted into the mapping relationship to obtain the acceleration on the load side of the robot joints. Specifically, it includes:

[0022] Obtain the end six-axis acceleration in the six-axis acceleration measurement unit and calculate the end Cartesian velocity;

[0023] Obtain the dynamic equation on the load side in the double-inertia model to obtain the load-side velocity vector;

[0024] Use the Jacobian matrix to construct a velocity mapping model based on the load-side velocity vector and the end Cartesian velocity;

[0025] Input the vibration state of the robot end into the mapping relationship to obtain the acceleration on the load side of the robot joints.

[0026] Preferably, based on the obtained acceleration on the load side of the joints, a self-updating Kalman filter is used to estimate the load-side joint velocity to obtain the estimated load-side joint velocity. Specifically, it includes:

[0027] Based on the motion relationship among the robot joint position, velocity, and acceleration, establish a motion differential equation;

[0028] Input the motion state of each robot joint into the motion differential equation and use a self-updating Kalman filter to estimate and solve for the load-side joint position and velocity to obtain the estimated load-side joint velocity.

[0029] Preferably, a feedback control law is used to eliminate the steady-state error of the estimated load-side joint velocity to obtain a corrected velocity. Specifically, it includes:

[0030] Use the Laplace transform method to transform the double-inertia model with external disturbances into the frequency domain; the external disturbances include the Coriolis force and centrifugal force terms, and the gravity term;

[0031] Take the derivative of the load-side dynamic equation after frequency-domain transformation to obtain the relationship between the link-side velocity and the motor-side velocity; use a feedback control law to solve the relationship between the link-side velocity and the motor-side velocity to obtain the trajectory reference joint velocity;

[0032] Introduce the trajectory reference joint velocity into the estimated load-side joint velocity to eliminate the steady-state error.

[0033] Preferably, compensating the corrected speed into the robot speed loop controller, and adjusting the control gain in the active damping controller according to the vibration state of the robot end to achieve active control of the robot vibration, specifically: feeding back the corrected speed to the robot speed loop controller, and adjusting the damping gain and stiffness gain of the active damping controller according to the six-axis vibration state signal during the movement of the robot.

[0034] In a second aspect, the present invention provides a robot processing vibration control system, including:

[0035] An end modeling module, configured to build a six-axis acceleration measurement unit for the robot end, and obtain the six-axis acceleration of the robot end and the end vibration state;

[0036] A joint-end relationship mapping module, configured to build a double-inertia model based on the motor side and the load side of the robot joints; use the Jacobian matrix to solve the six-axis acceleration measurement unit and the double-inertia model to obtain the mapping relationship between the six-axis acceleration in the Cartesian space and the joint space acceleration; substitute the end vibration state into the mapping relationship to obtain the acceleration of the load side of the robot joints;

[0037] An estimated speed acquisition module, configured to estimate the joint speed of the load side by using a self-updating Kalman filter according to the obtained acceleration of the load side of the joints, and obtain the estimated joint speed of the load side;

[0038] A corrected speed acquisition module, configured to use a feedback control law to eliminate the steady-state error of the estimated joint speed of the load side and obtain the corrected speed;

[0039] An active control module, configured to compensate the corrected speed into the robot speed loop controller, and adjust the control gain in the active damping controller according to the vibration state of the robot end to achieve active control of the robot vibration.

[0040] In a third aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps in a robot processing vibration control method described in the first aspect are implemented.

[0041] In a fourth aspect, the present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the steps in a robot processing vibration control method described in the first aspect are implemented.

[0042] Compared with the prior art, the beneficial effects of the present invention are:

[0043] (1) By constructing a six-axis acceleration measurement unit to replace the three-axis acceleration, etc., the present invention can obtain the end vibration state more accurately, improving the information perception accuracy; based on the motor side and the load side, a double-inertia model is constructed, and through the solution of the Jacobian matrix, the acceleration mapping between the Cartesian and joint spaces is obtained, overcoming the problem of low accuracy in decoupling the motion information of the motor side and the load side; the self-updating Kalman filter is used to estimate the joint speed of the load side, which reduces the high requirements for modeling accuracy and the complexity of the control structure compared with the methods relying on the disturbance observer or the motor side information; finally, the corrected speed is compensated to the speed loop controller and the damping gain is adjusted, realizing the effective active control of the robot vibration. The method proposed by the present invention is simple, reliable, and easy to implement, accurately and efficiently solving the vibration problem generated by the flexible joints during the robot movement and improving the stability of the robot movement process.

[0044] (2) The present invention designs a six-axis acceleration measurement unit that does not require the gyroscope to differentiate the approximate angular acceleration by fusing four three-axis accelerometers, obtaining the three-axis linear acceleration and three-axis angular acceleration at the end of the robot, avoiding the problem of modifying the robot body system to realize the motion state perception of the load side of the flexible joint robot.

[0045] (3) The present invention proposes a method for perceiving the motion state of the load side of the robot. Using the six-axis acceleration at the end and combining with the Jacobian matrix, a model of the acceleration of the joint load side and the end acceleration of the robot is established, avoiding the use of the differential of the motor side position signal or the approximate decoupling of the joint vibration state by the three-axis acceleration. The self-updating Kalman filter is used to estimate the joint speed of the load side, which can avoid the deficiencies of the existing approximate decoupling methods and more accurately perceive the motion state of the load side.

[0046] (4) The active damping control method proposed by the present invention establishes an active damping controller through the joint side speed, the trajectory reference speed, and the output signal of the robot position controller, and then outputs the corrected speed and feeds it back to the robot speed control loop to realize the adaptive adjustment of the damping of the robot system.

[0047] Advantages of additional aspects of the present invention will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The specification drawings forming a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute a limitation to the present invention.

[0049] Figure 1 It is the main flowchart of a robot processing vibration control method provided by an embodiment of the present invention;

[0050] Figure 2 Schematic diagram of the flexible joint modeling of the robot provided by the embodiment of the present invention;

[0051] Figure 3 Schematic diagram of the six-axis acceleration measurement unit provided by the embodiment of the present invention;

[0052] Figure 4 Schematic diagram of the geometric configuration of the six-axis acceleration measurement unit provided by the embodiment of the present invention;

[0053] Figure 5 Schematic diagram of the control principle of the overall method of the present invention provided by the embodiment of the present invention;

[0054] Figure 6 Frequency response diagram of the first joint of the robot with and without the active damping control algorithm provided by the embodiment of the present invention;

[0055] Figure 7 Schematic diagram of the reference trajectory of the robot end provided by the embodiment of the present invention;

[0056] Figure 8 Transient error diagram of the robot joint trajectory tracking with and without the active damping control algorithm provided by the embodiment of the present invention;

[0057] Figure 9 Six-axis acceleration curve diagram of the robot end with and without the active damping control algorithm provided by the embodiment of the present invention;

[0058] Figure 10 Schematic diagram of the machining process of the robot grinding the curved surface workpiece provided by the embodiment of the present invention;

[0059] Figure 11 Six-axis acceleration curve diagram of the robot end with and without the active damping control algorithm provided by the embodiment of the present invention;

[0060] Figure 12 Fast Fourier transform diagram of the acceleration of the robot end with and without the active damping control algorithm provided by the embodiment of the present invention.

[0061] Wherein: 101 - Flexible joint motor end; 102 - Joint flexible part; 103 - Flexible joint load end; 201 - Six-axis acceleration measurement unit housing 1; 202 - Six-axis acceleration measurement unit bracket; 203 - Triaxial acceleration sensor TB; 204 - Triaxial acceleration sensor TD; 205 - Triaxial acceleration sensor TA; 206 - Triaxial acceleration sensor TC; 207 - Six-axis acceleration measurement unit housing 2; 301 - Robot; 302 - Robot end flange; 303 - Grinding tool; 304 - Six-axis acceleration measurement unit; 305 - Curved surface workpiece. Specific embodiments

[0062] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0063] Embodiment 1

[0064] As Figure 1 shown, this embodiment discloses a method for controlling robot machining vibration. By establishing a decoupling strategy for the six-axis acceleration at the end, a load-side speed estimation method is used to obtain the speed information of the robot, and an active damping controller is established and fed back to the robot controller in real time to achieve suppression of the flexible vibration of the robot joints. The specific steps are as follows:

[0065] S1: Construct a six-axis acceleration measurement unit for the robot end to obtain the six-axis acceleration at the robot end and the end vibration state;

[0066] S2: Based on the motor side and the load side of the robot joints, construct a double-inertia model; use the Jacobian matrix to solve the six-axis acceleration measurement unit and the double-inertia model to obtain the mapping relationship between the six-axis acceleration in the Cartesian space and the joint space acceleration; substitute the end vibration state into the mapping relationship to obtain the acceleration of the load side of the robot joints;

[0067] S3: According to the obtained acceleration of the load side of the joints, use a self-updating Kalman filter to estimate the load-side joint speed to obtain the estimated load-side joint speed;

[0068] S4: Use the feedback control law to eliminate the steady-state error of the estimated load-side joint speed to obtain the corrected speed;

[0069] S5: Compensate the corrected speed into the robot speed loop controller, and adjust the control gain in the active damping controller according to the robot end vibration state to achieve active control of the robot vibration.

[0070] Next, in conjunction with Figure 1 , a method for controlling robot machining vibration disclosed in this embodiment will be described in detail.

[0071] In S1, the decoupling of the six-axis acceleration at the end: A six-axis acceleration measurement unit is formed by combining 4 triaxial accelerometers, and the six-axis acceleration measurement unit is installed on the robot end flange. According to the triaxial acceleration of each measurement point in the six-axis acceleration measurement unit and based on the rigid body kinematics model, the triaxial acceleration and triaxial angular acceleration at the robot end are obtained.

[0072] Furthermore, as Figure 2 shown is a schematic diagram of robot flexible joint modeling, including a flexible joint motor end 101, a joint flexible part 102, and a flexible joint load end 103. As Figure 3 shown, the decoupling of the six-axis acceleration at the end includes the following steps:

[0073] S101: The six-axis acceleration measurement unit is configured in a tetrahedral geometry, and four triaxial accelerometers are respectively installed at the four vertices of the tetrahedron;

[0074] Install the triaxial acceleration sensor TB203, triaxial acceleration sensor TD204, triaxial acceleration sensor TA205, and triaxial acceleration sensor TC206 at Figure 4 the four vertices B, D, A, and C of the tetrahedron in

[0075] S102: The four triaxial accelerometers respectively output the acceleration values along the XYZ three axes at the four vertices. According to the spatial rigid body kinematics model, the acceleration of the measurement point on the spatial rigid body is solved

[0076]

[0077] In the formula, is the linear acceleration value of the acceleration of the measurement point P j relative to the inertial coordinate system {i}; is the linear acceleration value of the rigid body coordinate system {b} relative to the inertial coordinate system {i}; is the angular velocity value of the rigid body coordinate system {b} relative to the inertial coordinate system {i}; is the angular acceleration value of the rigid body coordinate system {b} relative to the inertial coordinate system {i}; i r j is the position vector of point P j in the inertial coordinate system {i}; is the velocity of point P j relative to the rigid body coordinate system {b}; is the acceleration of point P j relative to the rigid body coordinate system {b}.

[0078] S103: Since the acceleration value output by the accelerometer installed at point P j is not the motion acceleration relative to the inertial coordinate system {i}, but the specific force at the installation position, using the relationship between the specific force and the motion acceleration, the value output by the acceleration sensor is obtained

[0079]

[0080] In the formula, b s j represents the sensitive axis direction of the single-axis acceleration sensor of point P j in the rigid body coordinate system {b}, b f j represents the specific force detected by the accelerometer of point P j and bf j The projection in the {b} system, b f ib represents the projection of the specific force at the origin of the {b} system in the {i} system, represents the skew-symmetric matrix of, b r j represents the position vector of point P j in the rigid body coordinate system {b}, b Ω ib represents b ω ib the skew-symmetric matrix of, where the b ω ib is the angular velocity value of the rigid body coordinate system {b} relative to the inertial coordinate system {i} i ω ib The projection in {b}, is b ω ib the angular acceleration value of.

[0081] S104: Convert 4 triaxial acceleration sensors into 12 uniaxial acceleration sensors, and then obtain the system measurement equation of the 12 uniaxial acceleration sensors:

[0082]

[0083] In the formula, P represents the uniaxial acceleration sensor configuration state matrix, Q represents the uniaxial acceleration sensor configuration coupling matrix, F( b ω ib ) represents the coupling term for the rigid body rotation angular velocity, A W×1 =[a1...a j ...a 12 T is the collected values of the 12 uniaxial acceleration sensors;

[0084]

[0085] S105: Rewrite the system equation in step S104 as where, is the acceleration sensor configuration matrix, c j =[P j Q j 1×12 , is the rigid body motion parameter, and then obtain the six-axis acceleration of a certain measurement point on the rigid body:

[0086]

[0087] ​​In S2, establish the acceleration models of the robot joint load side and the end effector: Assume the robot joint flexible model as a linear torsional spring, and model the motor side and the load side as a two-inertia model. Then obtain the dynamic model of an n-degree-of-freedom robot with joint flexibility (including the dynamic equations of the load side and the motor side). Use the Jacobian matrix to obtain the mapping relationship between the Cartesian space velocity (including the three-axis linear velocity and the three-axis angular velocity) of the robot end effector and the joint velocity. By differentiating both sides of the mapping relationship with respect to time, obtain the acceleration models of the robot joint load side and the end effector.

[0088] S201, as Figure 2 shown, use the Lagrangian method to establish the dynamic model of an n-degree-of-freedom robot:

[0089] Dynamic equation of the load side:

[0090]

[0091] Dynamic equation of the motor side:

[0092]

[0093] In the equations, q l , are the position vectors of the load side and the motor side respectively, represents the load side velocity vector, represents the load side acceleration vector, represents the motor side velocity vector, represents the motor side acceleration vector. is the load side torque, is the motor side torque. is the load side inertia matrix, is the Coriolis force and centrifugal force matrix, is the gravity vector. M m , K J , D j and are diagonal matrices, and the diagonal elements are the motor inertia, the reducer stiffness, the reducer damping, and the gear reduction ratio respectively.

[0094] S202: Use the Jacobian matrix to establish the relationship between the robot joint velocity and the end effector Cartesian velocity:

[0095]

[0096] In the equations, is the translational velocity of the robot end effector and the angular velocity to form the Cartesian velocity vector.

[0097] S203: Differentiate the velocity mapping model in S202 with respect to time to obtain the mapping relationship between the six-axis acceleration and the joint space acceleration in the Cartesian space:

[0098]

[0099] where is the acceleration of the robot's end effector in the Cartesian space; is the acceleration of the robot's joint load side.

[0100] S204: Use the six-axis acceleration measurement unit to obtain the vibration state a of the robot's end effector e , and obtain the acceleration of the robot's joint load side according to the mapping relationship between the six-axis acceleration and the joint space acceleration in S203

[0101]

[0102] where is the Moore-Penrose pseudoinverse of the Jacobian matrix J(q l ). is the derivative of J(q l ) with respect to time.

[0103] In S3, the velocity of the robot's load side is obtained: According to the obtained acceleration of the joint load side, the self-updating Kalman filter is used to estimate the joint velocity of the load side. By selecting the parameters in the Kalman filter, the estimation accuracy of the joint velocity of the load side is improved, thereby avoiding directly obtaining the joint velocity of the load side by integration and solving the problems of cumulative error and signal distortion caused by integration.

[0104] S301: Establish a motion differential equation based on the motion relationship between the robot's joint position, velocity, and acceleration:

[0105]

[0106] where k is the time step, A is the system state matrix, B is the system control matrix, C is the measurement matrix, is the system process noise and follows a Gaussian distribution; is the observation noise of the system at time step k and follows a Gaussian distribution, x k is the true state of the system at time step k, y k is the observation value of the system at time step k, u k is the input of the system at time step k.

[0107] S302: Estimate the motion state of each joint of the robot using the self-updating Kalman filter to solve for the joint position and velocity of the load side.

[0108] Among them, the self-updating Kalman filter estimation method is adopted, and the specific steps are as follows:

[0109] S3021: Represent the self-updating Kalman filter as a prediction process and an update process:

[0110] Prediction process:

[0111]

[0112] In the formula, is the state predicted at time k + 1 according to the system at time k; P k+1|k is the state covariance predicted at time k + 1 according to the system at time k; P k|k is the state covariance at time k + 1 after filtering.

[0113] Update process:

[0114]

[0115] In the formula, is the state at time k + 1 after system filtering; P k+1|k+1 is the state covariance at time k + 1 after filtering; K k+1 is the Kalman gain at time k; y k+1 is the actually measured signal; is the filtered signal after being estimated by the Kalman filter.

[0116] S3022: In order to improve the estimation effect of the Kalman filter, a self-updating Kalman filtering method considering the filtering residual is adopted to improve the dynamic filtering accuracy of the Kalman filter. Dynamically adjust the observation noise covariance matrix R according to the error estimated by the Kalman filter, and dynamically adjust the process noise covariance matrix Q according to the innovation matrix. The innovation d k and the residual ε k are respectively:

[0117]

[0118] S3023: Introduce a forgetting factor α for the dynamic update of the observation noise covariance matrix R and the process noise covariance matrix Q:

[0119]

[0120] In S4, the design of the active damping controller: By establishing a robot dynamics model considering joint flexibility, based on the Laplace transform method, the time-domain equation is transformed into the s-domain to facilitate the design of the active damping controller; the dynamic equation on the load side is differentiated to obtain the relationship between the link-side speed and the motor-side speed; the perturbation function is used to analyze the influence relationship of external disturbance forces on the motion performance of the robot system, and then the feedback term of the active damping controller is obtained; the feedback term is introduced into the robot speed-loop control circuit to obtain the reference signal of the motor speed loop; by analyzing the steady-state error between the load-side joint speed and the trajectory reference joint speed, the trajectory reference speed feedforward control is introduced to eliminate the influence of the steady-state error.

[0121] S401: According to the robot dynamics model (double-inertia model) established in S201, the Coriolis force, centrifugal force term, and gravity term are unified as the external disturbance torque τ. ext . And the robot motion space is regarded as the motion of each single point. Therefore, the coupling term in the load-side inertia matrix is temporarily ignored. Using the Laplace transform method, the expression model of the robot dynamics model in the s-domain is obtained:

[0122] Load side:

[0123]

[0124] Motor side:

[0125]

[0126] Among them, q li represents the joint position of the i-th joint on the load side, s is the Laplace operator, is the inverse of the corresponding element of the load-side inertia matrix M l ; k i and d i are the parameters related to stiffness and damping respectively; θ mi is the motor-side position; is the inverse of the corresponding element of the motor-side inertia matrix M m ; τ m,i is the motor-side torque, τ f,i is the friction torque, N -1 represents the inverse matrix of the diagonal matrix of the gear reduction ratio.

[0127] S402: Differentiate the dynamic equation on the load side to obtain the relationship between the link-side speed and the motor-side speed: Use the perturbation transfer function to analyze the perturbation during the operation of the robot:

[0128]

[0129] S403: Design the feedback control law G(s) to achieve the suppression of system disturbances, and introduce a correction speed into the system As a feedback signal, let To achieve the active compensation of damping and stiffness, introduce the control gains: damping gain K di and stiffness gain K ki , and set the feedback control law G(s) as:

[0130]

[0131] In the formula, represents the joint speed on the load side of the i-th joint. The numerator K d,i d i s is for active damping compensation to enhance the damping effect of the system and reduce vibrations; K k,i k i is for active stiffness compensation to increase the system stiffness and help the system recover to the stable state faster; the denominator d i s + k i is to introduce the dynamic characteristics of the controlled system through the estimation of the system damping and stiffness, so as to achieve the purpose of suppressing disturbances.

[0132] S404: Attach the feedback control law H(s) to the speed loop of the robot controller. The reference input signal of the motor speed control loop is:

[0133]

[0134] Among them, represents the speed output by the robot position loop controller, represents the joint speed of the reference trajectory.

[0135] S405: Substitute the motor-side reference signal into the robot dynamics model established in S401 to obtain the disturbance function between the input and output signals on the load side:

[0136]

[0137] S406: Use the Laplace transform to calculate the steady-state error of the system:

[0138]

[0139] Adopt the method of desired trajectory feedforward to eliminate the steady-state error of the system. When controlling the robot motor, the performance of the motor-side speed controller is sufficient to meet the control requirements. The complete control law of the active damping controller is:

[0140]

[0141] In S5, the robot speed loop feedback control: The corrected speed in the active damping controller is compensated into the robot speed loop controller. According to the vibration suppression effect during the robot's movement, the control gain in the active damping controller is adjusted, thereby achieving the active control of the robot's vibration.

[0142] S501: Feed back the corrected speed obtained by the active damping controller in S4 to the robot speed loop controller;

[0143] S502: Adjust the damping gain K d,i and stiffness gain K k,i of the active damping controller according to the vibration signal of the end six-axis during the robot's movement.

[0144] Embodiment 1

[0145] In this embodiment, the parameters of the robot flexible joint double-inertia model are set to the parameters in Table 1. According to Figure 5 the control schematic diagram in Figure 6 obtain the frequency response diagram of the first joint of the robot with and without the active damping control algorithm. It can be found from Figure 6 that when the active damping control algorithm is adopted, the peak of the system amplitude response is reduced. This is because the introduction of damping makes the response of the robot system smoother and improves the stability of the system.

[0146] Table 1 Parameter settings of the double-inertia model

[0147]

[0148] Adopt Figure 7 the reference trajectory of the robot end in Figure 5 Use the control principle of the overall method in d,i to control the robot's movement. Select the damping gain K k,i and stiffness gain K l,r in the active damping controller. According to the acceleration measurement unit of the end six-axis of the robot, the corrected speed in the active damping controller is fed back to the robot speed loop controller in real time to achieve the vibration suppression of the robot flexible joint. The transient response ||e|| = ||q l - q Figure 8The transient error diagram of the robot joint trajectory tracking with and without the active damping control algorithm is shown. The transient response fluctuation is reduced from the range of 0.004 rad to 0.015 rad without the active damping control algorithm to the range of 0 rad to 0.004 rad with the active damping control algorithm. When the active damping control algorithm is adopted, due to the feedback of the six-axis acceleration at the end of the robot, the vibration of the robot joint trajectory is significantly controlled, especially in the initial response stage. In terms of the later trajectory tracking, the robot trajectory tracking error is smaller than that without the active damping control algorithm.

[0149] And the six-axis acceleration measurement unit installed on the end flange of the robot, the decoupled linear acceleration and angular acceleration are as Figure 9 shown. When the active damping control algorithm is adopted, the six-axis vibration amplitude at the end of the robot is the smallest compared with that without the active damping control algorithm. Therefore, this method not only effectively suppresses the trajectory vibration but also improves the robot trajectory tracking accuracy.

[0150] Embodiment 2

[0151] This embodiment provides a method for controlling the vibration of a robot during machining. Since the robot needs to meet different task requirements, the method proposed in the present invention is adopted to use the robot to Figure 10 grind the curved surface workpiece in it, so as to realize the suppression of the flexible vibration of the robot during the grinding process.

[0152] As Figure 10 shown, the six-axis acceleration measurement unit 304 is installed on the robot end flange 302 at the end of the robot 301, and the grinding tool 303 is installed at the bottom of the robot end flange 302. By setting the robot grinding processing trajectory, the grinding of the curved surface workpiece 305 is realized.

[0153] Adopt Figure 5 the six-axis acceleration feedback robot vibration active control method in it to suppress the vibration during the robot machining process, and obtain the three-axis linear acceleration and three-axis angular acceleration decoupled by the six-axis acceleration measurement unit 304, as Figure 11 shown. When the active damping control algorithm is adopted, the six-axis vibration amplitude at the end of the robot is the smallest compared with that without the active damping control algorithm, and by performing a fast Fourier transform on the Figure 11 y-axis, z-axis linear acceleration and angular acceleration about the x-axis in the appendix, as shown in the fast Fourier transform diagram of the robot end acceleration with and without the active damping control algorithm in the appendix Figure 12 , the active damping control algorithm effectively suppresses the vibration at the end of the robot.

[0154] In this specific embodiment, four triaxial accelerometers are combined to form a six-axis acceleration measurement unit, and the six-axis acceleration measurement unit is installed on the end flange of the robot. The triaxial linear acceleration and triaxial angular acceleration at the end of the robot are obtained through the system measurement equation of the acceleration sensor. The mapping relationship between the Cartesian space acceleration and the joint space acceleration at the end of the robot is obtained by using the Jacobian matrix. The self-updating Kalman filter is used to estimate the joint speed on the load side, improving the estimation accuracy of the joint speed on the load side. Furthermore, the method avoids directly obtaining the joint speed on the load side by integration, solving the problems of cumulative error and signal distortion caused by integration. By establishing a robot dynamics model considering joint flexibility, the influence relationship of external disturbing forces on the motion performance of the robot system is analyzed using a disturbance function, obtaining the feedback term of the active damping controller, and introducing trajectory reference speed feedforward control to eliminate the influence of steady-state error. Then, the reference signal of the motor speed loop is obtained and fed back into the robot control loop to achieve active control of robot vibration.

[0155] Aiming at the existing problems of the robot acceleration feedback vibration control technology, the present invention proposes a robot machining vibration control method. By establishing a six-axis acceleration decoupling strategy at the end, the end vibration state is fed back in real time, and an active damping controller is established and fed back to the robot controller in real time to achieve suppression of joint flexibility vibration of the robot and improve the stability of the robot system.

[0156] Embodiment 2

[0157] This embodiment provides a robot machining vibration control system, including:

[0158] An end modeling module for constructing a six-axis acceleration measurement unit for the end of the robot to obtain the six-axis acceleration and the end vibration state at the end of the robot;

[0159] A joint-end relationship mapping module for constructing a dual-inertia model based on the motor side and the load side of the robot joints; using the Jacobian matrix to solve the six-axis acceleration measurement unit and the dual-inertia model to obtain the mapping relationship between the six-axis acceleration in the Cartesian space and the joint space acceleration; substituting the end vibration state into the mapping relationship to obtain the acceleration on the load side of the robot joints;

[0160] An estimated speed acquisition module for estimating the joint speed on the load side using a self-updating Kalman filter according to the obtained acceleration on the load side of the joints to obtain the estimated joint speed on the load side;

[0161] A corrected speed acquisition module for eliminating the steady-state error of the estimated joint speed on the load side using a feedback control law to obtain a corrected speed;

[0162] The active control module is used to compensate the correction speed into the robot speed loop controller, adjust the control gain in the active damping controller according to the vibration state of the robot end, and achieve the active control of robot vibration.

[0163] Embodiment III

[0164] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the steps in a robot machining vibration control method as described in Embodiment I above.

[0165] Embodiment IV

[0166] This embodiment provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the steps in a robot machining vibration control method as described in Embodiment I above.

[0167] The steps or modules involved in Embodiments II to IV above correspond to those in Embodiment I. For specific implementation manners, reference may be made to the relevant description part of Embodiment I. The term "computer-readable storage medium" should be understood to include a single medium or multiple media containing one or more instruction sets; it should also be understood to include any medium that can store, encode, or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0168] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for controlling the vibration of robot processing, characterized in that, Including: Construct a six-axis acceleration measurement unit for the robot end to obtain the six-axis acceleration of the robot end and the end vibration state; Based on the motor side and the load side of the robot joint, construct a double-inertia model; Use the Jacobian matrix to solve the six-axis acceleration measurement unit and the double-inertia model to obtain the mapping relationship between the six-axis acceleration in the Cartesian space and the joint space acceleration; substitute the end vibration state into the mapping relationship to obtain the acceleration of the load side of the robot joint; According to the obtained acceleration of the load side of the joint, use a self-updating Kalman filter to estimate the speed of the load side joint to obtain the estimated speed of the load side joint; Use the feedback control law to eliminate the steady-state error of the estimated speed of the load side joint to obtain the corrected speed, specifically including: Use the Laplace transform method to transform the double-inertia model with external disturbances into the frequency domain; the external disturbances include Coriolis force and centrifugal force terms, and gravity terms; Derive the relationship between the link side speed and the motor side speed for the load side dynamic equation after frequency domain conversion; use the feedback control law to solve the relationship between the link side speed and the motor side speed to obtain the trajectory reference joint speed; Introduce the trajectory reference joint speed to the estimated speed of the load side joint to eliminate the steady-state error; Compensate the corrected speed into the robot speed loop controller, and adjust the control gain in the active damping controller according to the robot end vibration state to achieve active control of robot vibration.

2. The robot processing vibration control method according to claim 1, characterized in that, The six-axis acceleration of the robot end includes three-axis linear acceleration and three-axis angular acceleration; the acquisition process is specifically: The specific process of obtaining the six-axis acceleration of the robot end is: Construct a six-axis acceleration measurement unit for the robot end based on the spatial rigid body coordinate system; 4 triaxial accelerometers are respectively set on the X, Y, Z axes and the origin of the six-axis acceleration measurement unit; According to the relationship between the spatial rigid body coordinate system and the inertial coordinate system, obtain the acceleration of the triaxial accelerometers in the six-axis acceleration measurement unit, and correct the output value of the accelerometer based on the specific force; Convert 4 triaxial accelerometers into 12 single-axis accelerometers and construct a system measurement equation; Simplify the system measurement equation, substitute the corrected output value of the accelerometer, and obtain the six-axis acceleration.

3. A robot processing vibration control method according to claim 1, characterized in that, Based on the motor side and the load side of the robot joint, constructing a double-inertia model is specifically: use the Lagrangian method to establish an n-degree-of-freedom double-inertia model for the robot joint; the double-inertia model includes a motor side dynamic equation and a load side dynamic equation.

4. The method for controlling the vibration of a robot during machining according to claim 3, wherein, Using the Jacobian matrix to solve the six-axis acceleration measurement unit and the double-inertia model to obtain the mapping relationship between the six-axis acceleration in the Cartesian space and the joint space acceleration; Substitute the end vibration state into the mapping relationship to obtain the acceleration of the load side of the robot joint; specifically including: Obtain the end six-axis acceleration in the six-axis acceleration measurement unit and calculate the end Cartesian speed; Obtain the load side dynamic equation in the double-inertia model to obtain the load side speed vector; Use the Jacobian matrix to construct a speed mapping model based on the load side speed vector and the end Cartesian speed; Input the vibration state at the end of the robot into the mapping relationship to obtain the acceleration on the load side of the robot joints.

5. A robot processing vibration control method according to claim 1, characterized in that, Based on the obtained acceleration on the load side of the joints, use a self-updating Kalman filter to estimate the joint speed on the load side to obtain the estimated joint speed on the load side; specifically including: Based on the motion relationship among the position, speed, and acceleration of the robot joints, establish a motion differential equation. Input the motion state of each joint of the robot into the motion differential equation, and use a self-updating Kalman filter for estimation to solve for the joint position and speed on the load side to obtain the estimated joint speed on the load side.

6. The robotic machining vibration control method according to claim 1, wherein Compensate the corrected speed into the robot speed loop controller, and adjust the control gain in the active damping controller according to the vibration state at the end of the robot to achieve active control of the robot vibration. Specifically: Feed back the corrected speed to the robot speed loop controller, and adjust the damping gain and stiffness gain of the active damping controller according to the six-axis vibration state signal at the end during the robot motion.

7. A robot processing vibration control system, characterized in that, Including: An end modeling module for constructing a six-axis acceleration measurement unit at the end of the robot to obtain the six-axis acceleration and the end vibration state at the end of the robot. A joint-end relationship mapping module for constructing a dual-inertia model based on the motor side and the load side of the robot joints. Use the Jacobian matrix to solve the six-axis acceleration measurement unit and the dual-inertia model to obtain the mapping relationship between the six-axis acceleration in the Cartesian space and the joint space acceleration. Substitute the end vibration state into the mapping relationship to obtain the acceleration on the load side of the robot joints. An estimated speed acquisition module for estimating the joint speed on the load side using a self-updating Kalman filter based on the obtained acceleration on the load side of the joints to obtain the estimated joint speed on the load side. A corrected speed acquisition module for eliminating the steady-state error of the estimated joint speed on the load side using a feedback control law to obtain the corrected speed, specifically including: Use the Laplace transform method to convert the dual-inertia model with external disturbances into the frequency domain; the external disturbances include the Coriolis force and centrifugal force terms, and the gravity term. Derive the relationship between the link side speed and the motor side speed from the load side dynamic equation after frequency domain conversion; use a feedback control law to solve the relationship between the link side speed and the motor side speed to obtain the trajectory reference joint speed. Introduce the trajectory reference joint speed to the estimated joint speed on the load side to eliminate the steady-state error. An active control module for compensating the corrected speed into the robot speed loop controller and adjusting the control gain in the active damping controller according to the vibration state at the end of the robot to achieve active control of the robot vibration.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in a robot machining vibration control method as described in any one of claims 1-6.

9. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in a robot machining vibration control method as described in any one of claims 1-6.

Images