CN120269576A - Robot compliance control method for learning in data-driven interaction process - Google Patents
Robot compliance control method for learning in data-driven interaction process Download PDFInfo
- Publication number
- CN120269576A CN120269576A CN202510764146.1A CN202510764146A CN120269576A CN 120269576 A CN120269576 A CN 120269576A CN 202510764146 A CN202510764146 A CN 202510764146A CN 120269576 A CN120269576 A CN 120269576A
- Authority
- CN
- China
- Prior art keywords
- robot
- instruction
- control
- motion
- force
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1628—Programme controls characterised by the control loop
- B25J9/1651—Programme controls characterised by the control loop acceleration, rate control
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1628—Programme controls characterised by the control loop
- B25J9/163—Programme controls characterised by the control loop learning, adaptive, model based, rule based expert control
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1628—Programme controls characterised by the control loop
- B25J9/1633—Programme controls characterised by the control loop compliant, force, torque control, e.g. combined with position control
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1656—Programme controls characterised by programming, planning systems for manipulators
- B25J9/1664—Programme controls characterised by programming, planning systems for manipulators characterised by motion, path, trajectory planning
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Manipulator (AREA)
Abstract
本发明公开了一种数据驱动交互过程学习的机器人柔顺控制方法，包括：工件表面方向运动指令获取步骤，得到机器人在沿着工件表面方向的运动指令；姿态误差与姿态运动指令获取步骤，得到机器人末端的跟踪期望姿态的末端角速度指令；力控制方向上的运动控制指令获取步骤，得到机器人在力控制方向上的运动控制指令；归一化控制指令构建步骤，基于所述机器人在沿着工件表面方向的运动指令、机器人末端的跟踪期望姿态的末端角速度指令以及机器人在力控制方向上的运动控制指令来构建统一的机器人速度指令。本发通过构建统一的机器人速度指令，实现对对象曲面集合形状的高效适应，同时保证足够的接触力精度，实现高精度力控制。
The present invention discloses a robot compliance control method of data-driven interactive process learning, including: a workpiece surface direction motion instruction acquisition step, obtaining the robot's motion instruction along the workpiece surface direction; a posture error and posture motion instruction acquisition step, obtaining the end angular velocity instruction of the robot's end tracking the desired posture; a motion control instruction acquisition step in the force control direction, obtaining the robot's motion control instruction in the force control direction; a normalized control instruction construction step, based on the robot's motion instruction along the workpiece surface direction, the end angular velocity instruction of the robot's end tracking the desired posture and the robot's motion control instruction in the force control direction to construct a unified robot speed instruction. The present invention realizes efficient adaptation to the shape of the object surface set by constructing a unified robot speed instruction, while ensuring sufficient contact force accuracy, and realizing high-precision force control.
Description
Technical Field
The invention relates to a robot control technology, in particular to a robot compliance control method for learning in a data-driven interaction process.
Background
With the high emphasis of industrial robots and the increasing market demand of countries, the environmental interactive task demands of robots are increasing. In this case, the robot tip interacts with the external environment. At this time, the simple end position control cannot completely ensure the processing quality and the production efficiency. Particularly, under the condition that the information such as the position, the gesture and the material quality of the contact object are uncertain, how to realize the efficient adaptation to the curved surface set shape of the object according to limited sensing information, and meanwhile, the method ensures enough contact force precision, so that the control problem to be solved urgently at present also brings great challenges to the online efficient flexible control of the robot.
The current method for controlling the compliance of the robot is mostly based on an impedance control method, namely, the interaction process of the robot and an unknown environment is modeled as a mass-spring-damping robot system, and the motion control quantity of the robot is corrected by taking the mass-spring-damping robot system as a reference, so that the robot presents certain compliance characteristics.
However, this model assumes a low adaptability to workpieces of complex materials, physical properties that vary with shape, and the like. Meanwhile, in the process of flexible control, the surface profile of the workpiece changes (namely, the vertical direction of the workpiece changes in real time), so that the robot is required to have good adaptability in the force control direction, and real-time identification and tracking of the vertical direction of the workpiece are required to be realized.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a method for controlling the compliance of a robot for learning in a data-driven interaction process, so as to have good adaptability.
In order to achieve the above purpose, the technical scheme of the invention is as follows:
a robot compliance control method for learning in a data-driven interaction process comprises the following steps:
A step of obtaining a movement instruction of the surface direction of the workpiece, wherein the movement instruction of the robot along the surface direction of the workpiece is obtained;
Acquiring an attitude error and an attitude motion instruction to obtain an end angular velocity instruction of a robot end for tracking an expected attitude;
A step of acquiring a motion control instruction in the force control direction, wherein the motion control instruction of the robot in the force control direction is acquired;
And a normalization control instruction construction step of constructing a unified robot speed instruction based on the motion instruction of the robot in the direction along the surface of the workpiece, the end angular speed instruction of the robot end for tracking the desired gesture, and the motion control instruction of the robot in the force control direction.
Optionally, the step of acquiring the attitude error and the attitude motion instruction includes:
A force sensor arranged at the tail end of the robot is adopted to acquire the contact force between the tail end of the current robot and an uncertain workpiece, and the normal information of a contact point is acquired according to the contact force, so that a rotation matrix of the expected gesture of the tail end of the robot is obtained;
And calculating an end angular velocity instruction of the robot end for tracking the expected gesture based on the rotation matrix of the expected gesture in the gesture control space.
Optionally, the step of obtaining the contact force between the current robot end and the uncertain workpiece by using a force sensor installed at the robot end, and obtaining the normal information of the contact point based on the contact force to obtain a rotation matrix of the expected gesture of the robot end includes:
Based on the acquired contact force of the current robot tip and the uncertain workpiece under the coordinate system of the force sensor of the current robot tip Normalizing the external force data fed back by the force sensor:
(1);
Wherein the vector is The rotation angle of the end of the robot is indicated,A norm representing the robot tip contact force;
Vector-based Constructing an antisymmetric feature matrix S describing the tail end gesture of the robot:
(2);
wherein n (3) represents a vector The 3 rd element in (2) represents a vectorThe 2 nd element in (2), n (1) represents a vectorElement 1 of (a);
The rotation matrix R d of the desired pose of the robot tip is expressed as:
(3);
Wherein, the Representing a three-dimensional identity matrix.
Optionally, in the gesture control space, the calculating, based on the rotation matrix of the desired gesture, a tip angular velocity instruction of the robot tip for tracking the desired gesture includes:
In the gesture control space, an angular velocity control amount is designed based on a rotation matrix R d of a desired gesture of the robot tip, and the rotation matrix of the desired gesture is converted into a quaternion form:
;
Wherein, the Representing a desired quaternion, r 11 representsRow 1, column 1 elements, r 22 representsRow 2, column 2 elements, r 33 representsRow 3, column 3 elements, r 32 representsRow 3, column 2 elements, r 23 representsRow 2, column 3 elements, r 13 representsRow 1, column 3 elements, r 31 representsRow 3, column 1 elements, r 21 representsRow 2, column 1 elements, r 12 representsRow 1, column 2 elements; Representing the desired quaternions in sequence, respectively The 1 st-4 th element value of the column vector;
By reading the quaternion of the current gesture of the robot The described posing error is expressed as:
(4);
Based on the formula (4), an error quaternion Is expressed as the real and imaginary parts ofConverting the error quaternion into a form of a rotation matrix:
(5);
Representing error quaternions in sequence respectively The 1 st-4 th element value of the column vector;
And then converting the rotation matrix into Euler angle form:
(6);
Wherein the attitude error Represented asE roll denotes a roll attitude error vector, e pitch denotes a pitch attitude error vector, and e yaw denotes a yaw attitude error vector;
Representation of The elements of row 2 and column 1 of (c),Representation ofThe elements of row 1 and column 1 of (A) are the same as the other elements, R 31 representsElements of row 3, column 1, R 32 representsElements of row 3 and column 2, R 33 representingRow 3, column 3 elements;
Then within the pose control space, the tip angular velocity command for the robot tip to track the desired pose is expressed as:
(7-1);
Wherein, the Gain coefficients representing robot tip angle posing errors.
Optionally, the step of obtaining the movement instruction of the workpiece surface direction comprises the steps of obtaining expected track information of a robot, calculating a movement tracking error and obtaining the movement instruction along the workpiece surface direction.
Optionally, the acquiring the expected track information of the robot, calculating a motion tracking error, and obtaining a motion instruction along the surface direction of the workpiece includes:
the expected derivatives of the expected trajectories of the robots are respectively ,The motion tracking error is:
;
wherein x is the actual position of the end of the current robot;
The motion instruction of the robot along the surface direction of the workpiece is:
(7-2);
Wherein, the Gain coefficients representing position errors.
Optionally, the step of acquiring a motion control instruction in the force control direction includes:
And constructing a nonlinear discrete system model of interactive force in the force control direction, and acquiring a flexible control strategy without model self-adaptive control to obtain a motion control instruction in the force control direction.
Optionally, the nonlinear discrete system model of the interaction force is:
(8);
Wherein, the 、The contact force at the moment k of the robot system and the acceleration command of the force control direction are respectively given,AndThe orders of the output and input of the robot system are respectively, and f represents an unknown nonlinear function;
and introducing the following performance index functions into the nonlinear discrete system model of the interaction force to optimize:
(9);
Wherein, the Is a control law penalty factor; substituting the data model into the performance index function for the desired output signal, relating to And let it be zero, the design data driven force control law is:
(10);
Wherein, the Is a control law step factor; Pseudo partial derivative for robot system 。
Optionally, the pseudo bias of the robot system is estimated online by adopting the following performance index function:
(11);
Wherein, the Penalty factors for estimating pseudo-bias;
Solving the middle of the related art And let it equal to zero, the pseudo-partial derivative on-line estimation algorithm is obtained as follows:
(12);
Wherein, the Estimating a pseudo-bias step factor;
the reset mechanism for estimating the pseudo-bias value is set as follows:
(13);
Wherein, the Referred to as a pseudo-bias reset threshold; Is pseudo bias guide Is set to an initial value of (1);
based on the designed data-driven force control law, a motion control instruction in the force control direction is obtained:
(14);
representing the update time interval of the motion control command in the force control direction.
Optionally, the unified robot speed command is:
;
Wherein, the ;
R d is a rotation matrix of the desired pose of the robot tip,Is the transposed moment of R d,Is a motion instruction of the robot along the surface direction of the workpiece,Is a motion control command in the force control direction.
Compared with the prior art, the invention has the beneficial effects that:
According to the invention, the motion instruction of the robot in the surface direction of the workpiece, the terminal angular velocity instruction of the robot terminal for tracking the expected gesture and the motion control instruction of the robot in the force control direction are obtained from three aspects, so that a unified robot velocity instruction is constructed, the efficient adaptation to the shape of the object curved surface set is realized, meanwhile, enough contact force precision is ensured, and high-precision force control is realized.
Drawings
Fig. 1 is a main flowchart of a robot compliance control method for learning a data-driven interactive process according to an embodiment of the present application.
FIG. 2 is a complete process flow diagram of a robot compliance control method for data-driven interactive process learning provided by an embodiment of the present application.
Detailed Description
Examples:
The technical scheme of the invention is further described below with reference to the accompanying drawings and examples.
Referring to fig. 1, the method for controlling compliance of a robot for learning a data-driven interactive process provided in this embodiment mainly includes the following steps;
A step of obtaining a movement instruction of the surface direction of the workpiece, wherein the movement instruction of the robot along the surface direction of the workpiece is obtained;
The gesture error and gesture movement command obtaining step obtains a terminal angular velocity command of the robot terminal for tracking the expected gesture
A step of acquiring a motion control instruction in the force control direction, wherein the motion control instruction of the robot in the force control direction is acquired;
And a normalization control instruction construction step of constructing a unified robot speed instruction based on the motion instruction of the robot in the direction along the surface of the workpiece, the end angular speed instruction of the robot end for tracking the desired gesture, and the motion control instruction of the robot in the force control direction.
In summary, the method obtains the motion instruction of the robot along the surface direction of the workpiece, the terminal angular velocity instruction of the terminal of the robot for tracking the expected gesture and the motion control instruction of the robot in the force control direction from three aspects, thereby constructing a unified robot velocity instruction, realizing the efficient adaptation to the shape of the curved surface set of the object, ensuring enough contact force precision and realizing high-precision force control.
In a specific embodiment, the step of acquiring the attitude error and the attitude motion instruction comprises the steps of acquiring the contact force between the current robot tail end and an uncertain workpiece by adopting a force sensor arranged at the robot tail end, acquiring normal information of a contact point according to the contact force, and acquiring a rotation matrix of an expected attitude of the robot tail end, wherein the attitude error and the attitude motion instruction are calculated and acquired based on the rotation matrix of the expected attitude in an attitude control space.
More specifically, the contact force information is acquired under the current robot end force sensor coordinate systemNormalizing the external force data fed back by the force sensor:
(1);
Wherein the vector is The rotation angle of the end of the robot is indicated,Representing the norm of the robot tip contact force. Vector-basedAn antisymmetric feature matrix describing the robot tip pose may be constructed:
(2);
wherein n (3) represents a vector The 3 rd element in (2) represents a vectorThe 2 nd element in (2), n (1) represents a vectorElement 1 of (a);
The rotation matrix of the desired pose of the robot tip can be expressed as:
(3);
Wherein, the Representing a three-dimensional identity matrix.
Thus, by the above-described operations (i.e., formulas (1) - (3)), contact surface normal information can be obtained and described as a quaternion form.
Within the gesture control space, a rotation matrix based on the desired gestureThe angular velocity control amount is designed. On the basis, the rotation matrix of the expected gesture is converted into a quaternion form:
;
Wherein, the Representing a desired quaternion, r 11 representsRow 1, column 1 elements, r 22 representsRow 2, column 2 elements, r 33 representsRow 3, column 3 elements, r 32 representsRow 3, column 2 elements, r 23 representsRow 2, column 3 elements, r 13 representsRow 1, column 3 elements, r 31 representsRow 3, column 1 elements, r 21 representsRow 2, column 1 elements, r 12 representsRow 1, column 2 elements; Representing the desired quaternions in sequence, respectively The 1 st-4 th element value of the column vector.
The traditional rotation matrix calculation method may cause singular situations in robot gesture control, cause available control instructions to be difficult to calculate under specific conditions, by adopting the mode, the application converts the rotation matrix with the expected gesture into the quaternion form, thereby effectively avoiding the problem.
Further, as the rotation matrix of the expected gesture is converted into the quaternion form, the quaternion of the current gesture of the robot is utilizedThe described posing error is expressed as:
(4);
that is, the error between the robot tip pose and the contact surface normal information is obtained by using the quaternion operation according to the current robot tip pose (i.e., equation 4).
Based on the formula (4), an error quaternionIs expressed as the real and imaginary parts ofOn the basis, the error quaternion is converted into a form of a rotation matrix:
(5);
And then converting the rotation matrix into Euler angle form:
(6);
Wherein the attitude error Represented asE roll denotes a roll attitude error vector, e pitch denotes a pitch attitude error vector, and e yaw denotes a yaw attitude error vector;
Representation of The elements of row 2 and column 1 of (c),Representation ofThe elements of row 1 and column 1 of (A) are the same as the other elements, R 31 representsElements of row 3, column 1, R 32 representsElements of row 3 and column 2, R 33 representingRow 3, column 3 elements;
The tip angular velocity command to track the desired pose of the robot tip in space can be expressed as:
(7-1);
Wherein, the Gain coefficients representing robot tip angle posing errors.
Thus, by the above-described operation, a robot attitude angular velocity command is constructed using an attitude error in the form of a description of the euler angle, which can align the contact surface normal direction and the robot tip direction.
Therefore, in the step of acquiring the gesture error and the gesture movement instruction, the method characterizes the normal direction of the uncertain contact surface through the acquired force information, then constructs the error between the normal direction of the contact surface and the direction of the tail end of the robot according to the direction of the tail end of the current robot, and designs the control speed of the gesture to align the two. Meanwhile, the problem that the conventional rotation matrix calculation mode possibly causes singular conditions in robot gesture control is solved, and the rotation matrix of the expected gesture is converted into a quaternion form by the method, so that the problem can be effectively avoided.
In one embodiment, the workpiece surface direction movement command obtaining step includes:
the motion control strategy of the robot is designed in the direction along the surface of the workpiece. Defining the expected derivatives of the expected trajectories of the robot as respectively ,And defines the motion tracking error as:
;
the motion reference instructions of the robot in the direction along the surface of the workpiece are designed as follows:
(7-2);
Wherein, the Gain coefficients representing position errors.
In one embodiment, the step of acquiring the motion control command in the force control direction includes:
And constructing a nonlinear discrete system model of interactive force in the force control direction, and acquiring a flexible control strategy without model self-adaptive control to obtain a motion control instruction in the force control direction.
In the step, the method acquires a soft control strategy (namely a data-driven soft control strategy) without model self-adaptive control by constructing a nonlinear discrete system model of the interactive force, and the high-precision force control can be realized by the prior information of the robot interactive object and the internal information in the control instruction data and the actual interactive force data of the robot.
More specifically, the interactive force contact process of the robot with the unknown workpiece is represented as a nonlinear discrete robot system of the form:
(8);
Wherein, the 、The contact force at the moment k of the robot system and the acceleration command of the force control direction are respectively given,AndThe orders of the robot system output and input, respectively, f represents an unknown nonlinear function.
The following performance index functions are introduced for optimization:
(9);
Wherein, the Is a control law penalty factor, which is introduced to limit the controller outputThe visual appearance in the robot system is to ensure that the input signal curve of the robot system is smoother.Substituting the data model into the performance index function for the desired output signal, relating toAnd let it be zero, the design data driven force control law is:
(10);
Wherein, the Is a control law step factor, and the function of introducing the parameter is to make the control law algorithm more general and flexible. For a nonlinear robotic system with an unknown precise mathematical model, the robotic system pseudo-partial derivativeAnd is also unknown time-varying. Therefore, the estimation of the pseudo bias value of the robot system needs to be performed by using the input/output data information of the controlled robot system. The pseudo bias guide of the robot system is estimated on line by adopting the following performance index function:
(11);
Wherein, the To estimate the penalty factor of the pseudo bias, the parameter is introduced to prevent the mutation of the estimated value of the pseudo bias of the robot system caused by factors such as external interference. Solving the middle of the related artAnd let it equal to zero, the pseudo-partial derivative on-line estimation algorithm can be obtained as follows:
(12);
Wherein, the To estimate the pseudo-offset step size factor. Finally, a reset mechanism for estimating the pseudo-partial derivative value is set, and the aim is to improve the dynamic tracking performance of the time-varying robot system by the algorithm.
(13);
Wherein, the Is a relatively small number, called the pseudo-bias reset threshold,Is pseudo bias guideIs set to be a constant value.
Based on the designed data-driven force control law, motion control instructions in the force control direction can be obtained:
(14);
representing the update time interval of the motion control command in the force control direction.
Thus, through the operation, the flexible control strategies (formulas (10) and (12)) can be obtained, prior information of the robot interaction object is not needed, and high-precision force control can be realized through the control instruction data of the robot and the information in the actual interaction force data.
Finally, a unified robot speed command can be constructed based on the motion command of the robot in the direction along the surface of the workpiece, the end angular velocity command of the robot end to track the desired gesture, and the motion control command of the robot in the force control direction:
;
Wherein, the ;
AndOperators are operations of orthogonal decomposition.
Therefore, the method based on the workpiece contour recognition and orthogonal decomposition of the sensor force information does not need to introduce other high-cost measuring equipment, and has the advantages of low cost, easiness in implementation and the like;
as shown in fig. 2, a complete flowchart of the robot compliance control method for learning the data-driven interaction process is provided in this embodiment.
The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement the same, and are not intended to limit the scope of the present invention. All equivalent changes or modifications made in accordance with the essence of the present invention are intended to be included within the scope of the present invention.
Claims (10)
1. A robot compliance control method for learning in a data-driven interaction process is characterized by comprising the following steps:
A step of obtaining a movement instruction of the surface direction of the workpiece, wherein the movement instruction of the robot along the surface direction of the workpiece is obtained;
Acquiring an attitude error and an attitude motion instruction to obtain an end angular velocity instruction of a robot end for tracking an expected attitude;
A step of acquiring a motion control instruction in the force control direction, wherein the motion control instruction of the robot in the force control direction is acquired;
And a normalization control instruction construction step of constructing a unified robot speed instruction based on the motion instruction of the robot in the direction along the surface of the workpiece, the end angular speed instruction of the robot end for tracking the desired gesture, and the motion control instruction of the robot in the force control direction.
2. The method for controlling compliance of a robot for data driven interactive process learning of claim 1, wherein said step of obtaining said posing error and posing motion command comprises:
A force sensor arranged at the tail end of the robot is adopted to acquire the contact force between the tail end of the current robot and an uncertain workpiece, and the normal information of a contact point is acquired according to the contact force, so that a rotation matrix of the expected gesture of the tail end of the robot is obtained;
And calculating an end angular velocity instruction of the robot end for tracking the expected gesture based on the rotation matrix of the expected gesture in the gesture control space.
3. The method for controlling compliance of a robot for learning a data-driven interactive process according to claim 2, wherein the step of obtaining a rotational matrix of a desired posture of the robot tip by obtaining a contact force between the current robot tip and an uncertain workpiece using a force sensor installed at the robot tip and obtaining normal information of a contact point based on the contact force, comprises:
Based on the acquired contact force of the current robot tip and the uncertain workpiece under the coordinate system of the force sensor of the current robot tip Normalizing the external force data fed back by the force sensor:
(1);
Wherein the vector is The rotation angle of the end of the robot is indicated,A norm representing the robot tip contact force;
Vector-based Constructing an antisymmetric feature matrix S describing the tail end gesture of the robot:
(2);
wherein n (3) represents a vector The 3 rd element in (2) represents a vectorThe 2 nd element in (2), n (1) represents a vectorElement 1 of (a);
The rotation matrix R d of the desired pose of the robot tip is expressed as:
(3);
Wherein, the Representing a three-dimensional identity matrix.
4. A method of controlling compliance of a robot for data driven interactive process learning as claimed in claim 2 or 3, wherein said calculating, in the attitude control space, the tip angular velocity command for tracking the desired attitude of the robot tip based on the rotation matrix of the desired attitude comprises:
In the gesture control space, an angular velocity control amount is designed based on a rotation matrix R d of a desired gesture of the robot tip, and the rotation matrix of the desired gesture is converted into a quaternion form:
;
Wherein, the Representing a desired quaternion, r 11 representsRow 1, column 1 elements, r 22 representsRow 2, column 2 elements, r 33 representsRow 3, column 3 elements, r 32 representsRow 3, column 2 elements, r 23 representsRow 2, column 3 elements, r 13 representsRow 1, column 3 elements, r 31 representsRow 3, column 1 elements, r 21 representsRow 2, column 1 elements, r 12 representsRow 1, column 2 elements; Representing the desired quaternions in sequence, respectively The 1 st-4 th element value of the column vector;
By reading the quaternion of the current gesture of the robot The described posing error is expressed as:
(4);
Based on the formula (4), an error quaternion Is expressed as the real and imaginary parts ofConverting the error quaternion into a form of a rotation matrix:
(5);
Representing error quaternions in sequence respectively The 1 st-4 th element value of the column vector;
And then converting the rotation matrix into Euler angle form:
(6);
Wherein the attitude error Represented asE roll denotes a roll attitude error vector, e pitch denotes a pitch attitude error vector, and e yaw denotes a yaw attitude error vector;
Representation of The elements of row 2 and column 1 of (c),Representation ofThe elements of row 1 and column 1 of (A) are the same as the other elements, R 31 representsElements of row 3, column 1, R 32 representsElements of row 3 and column 2, R 33 representingRow 3, column 3 elements;
Then within the pose control space, the tip angular velocity command for the robot tip to track the desired pose is expressed as:
(7-1);
Wherein, the Gain coefficients representing robot tip angle posing errors.
5. The method of claim 1, wherein the step of obtaining the motion command of the surface of the workpiece comprises obtaining expected track information of the robot, and calculating a motion tracking error to obtain the motion command along the surface of the workpiece.
6. The method for controlling compliance of a robot for learning a data driven interactive process according to claim 5, wherein said obtaining the expected trajectory information of the robot, calculating a motion tracking error, and obtaining a motion command along the surface of the workpiece, comprises:
the expected derivatives of the expected trajectories of the robots are respectively ,The motion tracking error is:
;
wherein x is the actual position of the end of the current robot;
The motion instruction of the robot along the surface direction of the workpiece is:
(7-2);
Wherein, the Gain coefficients representing position errors.
7. The method for robot compliance control for data driven interactive process learning of claim 1, wherein the step of obtaining motion control commands in the force control direction comprises:
And constructing a nonlinear discrete system model of interactive force in the force control direction, and acquiring a flexible control strategy without model self-adaptive control to obtain a motion control instruction in the force control direction.
8. The method for controlling compliance of a robot for data driven interactive process learning of claim 7, wherein the model of the nonlinear discrete system of interactive forces is:
(8);
Wherein, the 、The contact force at the moment k of the robot system and the acceleration command of the force control direction are respectively given,AndThe orders of the output and input of the robot system are respectively, and f represents an unknown nonlinear function;
and introducing the following performance index functions into the nonlinear discrete system model of the interaction force to optimize:
(9);
Wherein, the Is a control law penalty factor; substituting the data model into the performance index function for the desired output signal, relating to And let it be zero, the design data driven force control law is:
(10);
Wherein, the Is a control law step factor; Pseudo partial derivative for robot system 。
9. The method for controlling compliance of a robot for data driven interactive process learning of claim 8, wherein the robot system pseudo-bias is estimated on-line using a performance index function as follows:
(11);
Wherein, the Penalty factors for estimating pseudo-bias;
Solving the middle of the related art And let it equal to zero, the pseudo-partial derivative on-line estimation algorithm is obtained as follows:
(12);
Wherein, the Estimating a pseudo-bias step factor;
the reset mechanism for estimating the pseudo-bias value is set as follows:
(13);
Wherein, the Referred to as a pseudo-bias reset threshold; Is pseudo bias guide Is set to an initial value of (1);
based on the designed data-driven force control law, a motion control instruction in the force control direction is obtained:
(14);
representing the update time interval of the motion control command in the force control direction.
10. The method for controlling compliance of a robot for data driven interactive process learning of claim 1, wherein said unified robot speed command is:
;
Wherein, the ;
R d is a rotation matrix of the desired pose of the robot tip,Is the transposed moment of R d,Is a motion instruction of the robot along the surface direction of the workpiece,Is a motion control command in the force control direction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510764146.1A CN120269576B (en) | 2025-06-10 | 2025-06-10 | A data-driven interactive process learning method for robot compliant control |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510764146.1A CN120269576B (en) | 2025-06-10 | 2025-06-10 | A data-driven interactive process learning method for robot compliant control |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN120269576A true CN120269576A (en) | 2025-07-08 |
| CN120269576B CN120269576B (en) | 2025-08-26 |
Family
ID=96236947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510764146.1A Active CN120269576B (en) | 2025-06-10 | 2025-06-10 | A data-driven interactive process learning method for robot compliant control |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120269576B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5497061A (en) * | 1993-03-31 | 1996-03-05 | Hitachi, Ltd. | Method of controlling robot's compliance |
| CN110125936A (en) * | 2019-05-15 | 2019-08-16 | 清华大学深圳研究生院 | A kind of the Shared control method and ground experiment verifying system of robot for space |
| CN110948504A (en) * | 2020-02-20 | 2020-04-03 | 中科新松有限公司 | Method and device for normal constant force tracking of robot machining operations |
| CN117656062A (en) * | 2023-11-30 | 2024-03-08 | 广东省科学院智能制造研究所 | A robot's surface normal tracking method for uncertain workpieces |
-
2025
- 2025-06-10 CN CN202510764146.1A patent/CN120269576B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5497061A (en) * | 1993-03-31 | 1996-03-05 | Hitachi, Ltd. | Method of controlling robot's compliance |
| CN110125936A (en) * | 2019-05-15 | 2019-08-16 | 清华大学深圳研究生院 | A kind of the Shared control method and ground experiment verifying system of robot for space |
| CN110948504A (en) * | 2020-02-20 | 2020-04-03 | 中科新松有限公司 | Method and device for normal constant force tracking of robot machining operations |
| CN117656062A (en) * | 2023-11-30 | 2024-03-08 | 广东省科学院智能制造研究所 | A robot's surface normal tracking method for uncertain workpieces |
Also Published As
| Publication number | Publication date |
|---|---|
| CN120269576B (en) | 2025-08-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Qi et al. | Contour moments based manipulation of composite rigid-deformable objects with finite time model estimation and shape/position control | |
| Lepora et al. | Pose-based tactile servoing: Controlled soft touch using deep learning | |
| CN111360827A (en) | A visual servo switching control method and system | |
| CN106003033B (en) | A kind of sixdegree-of-freedom simulation standard Chinese character Writing method with power control | |
| CN115582831B (en) | Automatic correction method and system for robotic arm | |
| CN115480583B (en) | Visual servo tracking and impedance control method for flying operation robot | |
| CN113733088A (en) | Mechanical arm kinematics self-calibration method based on binocular vision | |
| CN118559698B (en) | Fixed time visual servo control method based on fractional obstacle Lyapunov | |
| CN117067209A (en) | A collaborative robot multi-axis hole assembly method, system, electronic equipment and medium | |
| CN114434441B (en) | Visual servo tracking control method for mobile robot based on adaptive dynamic programming | |
| CN117331309A (en) | A robotic polishing method based on super-twisted sliding mode adaptive admittance control | |
| CN113910232B (en) | Self-adaptive attitude tracking method and device, storage medium and electronic equipment | |
| WO2026020720A1 (en) | Human-guided vision-force fused impedance iterative learning control method for robotic arm | |
| CN116038697B (en) | A method and system for automatic spraying of jeans based on manual teaching | |
| CN109542094A (en) | Mobile robot visual point stabilization without desired image | |
| CN117325182A (en) | Robot compliance control method integrating multiple self-adaptive regulation and control mechanisms | |
| CN117647952A (en) | Digital modeling position accuracy compensation method and system for industrial robots | |
| CN120269576B (en) | A data-driven interactive process learning method for robot compliant control | |
| CN114800523B (en) | Mechanical arm track correction method, system, computer and readable storage medium | |
| CN118938914B (en) | Track tracking control method, system, equipment and storage medium for improving robot bending precision and inspection efficiency | |
| CN115328116A (en) | A four-wheeled robot and its motion control method | |
| CN109032071A (en) | Numerically-controlled machine tool kinematic error based on deep learning network is traced back in real time because of method | |
| CN114734435B (en) | An encoder calibration method, device and system based on hypersphere | |
| CN117921675A (en) | Robust control method and system for overhead contact operation robot | |
| Lepora et al. | Pose-based servo control with soft tactile sensing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |