Continuous body robot motion control method, system and equipment based on data driving

Abstract

The invention is applicable to the technical field of continuum robots, and provides a continuum robot motion control method based on data driving, which comprises the following steps: acquiring a target space pose and a real-time tail end pose of a continuum robot, and constructing a position error equation based on the target space pose and the real-time tail end pose; based on the position error equation, constructing a position return-to-zero dynamic equation, and solving a target driving vector of the continuum robot; in the solving process, a speed return-to-zero dynamic equation is constructed, and jacobian matrix parameters of the continuum robot under the target driving vector are obtained; according to the target driving vector, the continuum robot is driven to the target space pose from the real-time tail end pose, a complex kinematics model related to the continuum robot does not need to be built in advance for training, and universal, high-precision, rapid and stable motion control is achieved.

Description

Technical Field

[0001] This application belongs to the field of continuum robot technology, and in particular relates to a data-driven motion control method, system and device for continuum robots. Background Technology

[0002] Continuum robots are a new type of biomimetic robot whose structure and movement mimic soft-bodied creatures such as elephant trunks, octopus tentacles, or vines in nature. They have extremely high flexibility and compliance, enabling them to autonomously navigate and operate in narrow, complex, and confined environments. They are commonly used in the medical field for minimally invasive surgery, including but not limited to puncture robots, laparoscopic robots, and neurosurgical robots.

[0003] Existing technologies for driving continuum robots typically employ static analysis methods based on physical models, such as the Cosserat link theory. A robot model that conforms to the robot's own physical characteristics is pre-established. Based on this, algorithms such as deep reinforcement learning are further combined to iteratively optimize network parameters and ultimately map out the driving commands required to drive the robot.

[0004] The aforementioned driving methods for continuum robots face a trade-off between complexity and accuracy in the modeling process. Specifically, overly simplified models result in poor control accuracy, while overly complex models have slow solution speeds, making it difficult to achieve a balance. Secondly, this method has a strong parameter dependency, and the established models are usually only applicable to robots designed with specific structural parameters. Once parameters such as robot size and materials change, mechanical modeling and data collection for network training must be performed again, which is cumbersome and lacks versatility. Furthermore, control methods relying on reinforcement learning require thousands or even tens of thousands of interactions between the robot and the environment to collect training data. This process is not only time-consuming but may also cause mechanical damage to the robot body during exploratory actions. Therefore, there is an urgent need for a method that can safely and efficiently achieve motion control of continuum robots without requiring dynamic modeling. Summary of the Invention

[0005] This application provides a data-driven motion control method, system, and device for a continuum robot, which can solve one of the problems in the prior art mentioned above.

[0006] In a first aspect, embodiments of this application provide a data-driven motion control method for a continuum robot, comprising: The target spatial pose and real-time end-effector pose of the continuum robot are obtained, and a position error equation is constructed based on the target spatial pose and the real-time end-effector pose. Based on the position error equation, construct the position zeroing dynamic equation and solve for the target driving vector of the continuum robot; During the solution process, a dynamic equation with zero velocity is constructed to obtain the Jacobian matrix parameters of the continuum robot under the target driving vector; The target driving vector drives the continuum robot from the real-time end-effector pose to the target spatial pose.

[0007] Furthermore, the step of constructing a position-zeroing dynamic equation based on the position error equation and solving for the target driving vector of the continuum robot includes: By taking the time derivative of the position error equation, the second derivative equation is obtained; Based on the position-zeroing neural network, and combined with the position network design parameters, a position-zeroing dynamic equation is constructed. By combining the position-zeroing dynamic equation and the second derivative equation, the target driving vector is solved.

[0008] Furthermore, the step of combining the position-zeroing dynamic equation and the second derivative equation to solve for the target driving vector includes: Establish the velocity mapping relationship between the drive space and the task space of the continuum robot; Based on the second derivative equation and combined with the velocity mapping relationship, the third derivative equation is obtained; Based on the position-zeroing dynamic equation, combined with the third derivative equation and the position error equation, the driving vector solution equation is obtained; The target driving vector is obtained by solving the driving vector solution equation, wherein the Jacobian matrix parameter is an unknown parameter.

[0009] Furthermore, establishing the velocity mapping relationship of the continuum robot from the driving space to the task space includes: Based on the real-time end-effector pose and driving vector, the kinematic equations of the continuum robot are constructed. By taking the time derivative of the kinematic equations, a velocity mapping relationship between the drive space and the task space of the continuum robot is established, and the velocity mapping relationship is measured based on the Jacobian matrix.

[0010] Furthermore, the location network design parameters are functions related to the real-time drive vector of the continuum robot.

[0011] Furthermore, the construction of the velocity-zero dynamic equation to obtain the Jacobian matrix parameters of the continuum robot under the target driving vector includes: Based on the velocity mapping relationship, a velocity error equation for the continuum robot under the target driving vector is constructed; Based on the velocity error equation, a velocity-zero dynamic equation is constructed to estimate the Jacobian matrix parameters of the continuum robot under the target driving vector.

[0012] Furthermore, the step of constructing a velocity-zero dynamic equation based on the velocity error equation and estimating the Jacobian matrix parameters of the continuum robot under the target driving vector includes: Taking the time derivative of the velocity error equation yields the first derivative equation; Based on the velocity-zero neural network and combined with the velocity network design parameters, a velocity-zero dynamic equation is constructed. By combining the dynamic equation for zeroing velocity and the first derivative equation, the Jacobian matrix parameters are solved.

[0013] Furthermore, the speed network design parameters are functions related to the speed error.

[0014] Secondly, embodiments of this application provide a data-driven motion control system for a continuum robot, comprising: First processing module: used to acquire the target spatial pose and real-time end-effector pose of the continuum robot, and to construct a position error equation based on the target spatial pose and the real-time end-effector pose; The second processing module is used to construct a position-zeroing dynamic equation based on the position error equation and solve for the target driving vector of the continuum robot. The third processing module is used to construct the dynamic equation with zero velocity during the solution process and obtain the Jacobian matrix parameters of the continuum robot under the target driving vector. The fourth processing module is used to drive the continuum robot from the real-time end-effector pose to the target spatial pose using the target driving vector.

[0015] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described data-driven continuous robot motion control method.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium, including a computer program stored in the computer-readable storage medium, which, when executed by a processor, implements the aforementioned data-driven continuous robot motion control method.

[0017] The beneficial effects of the embodiments in this application compared with the prior art are: This application discloses a data-driven motion control method for a continuum robot. When there is an error between the real-time end-effector pose and the target spatial pose of the continuum robot, the key control variable, i.e., the driving vector q, is estimated through a position-zeroing neural network, and the Jacobian matrix coefficients of the continuum robot under the driving vector are estimated through a velocity-zeroing neural network. This invention enables the synergistic effect of two zero-return neural networks to achieve the motion control objective of a continuum robot, allowing the robot to move from its real-time end-effector pose to the target spatial pose without the need for pre-establishing complex kinematic models related to the continuum robot for training. This achieves universal, high-precision, and fast and stable motion control. Furthermore, by designing adaptive network design parameters in the two zero-return neural networks, the convergence speed in the corresponding zero-return neural networks can be adjusted in real time to adapt to nonlinear changes or disturbances in the driving vector of the continuum robot, thereby completing closed-loop control of the continuum robot. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a data-driven motion control method for a continuum robot according to an embodiment of the present invention. Figure 2 yes Figure 1 The diagram shows a flowchart of an embodiment of a data-driven motion control method for a continuum robot. Figure 3 This is a schematic diagram of the structure of a data-driven motion control system for a continuum robot according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0020] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0021] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0022] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0023] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0024] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0025] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0026] Please see Figure 1As shown, this invention is a data-driven motion control method for continuum robots. It constructs two zero-return neural networks: a position zero-return neural network (Pose_ZNN) and a velocity zero-return neural network (Speed_ZNN). By inputting real-time end-effector pose, velocity, and acceleration information of the continuum robot, and combining the corresponding network design parameters, a kinematic error function for the continuum robot is established. Utilizing the characteristics of the zero-return neural network, the corresponding position and velocity errors are made to approach zero. Finally, the estimated Jacobian matrix and driving vector of the continuum robot are output. This continuum robot motion control method eliminates the need for dynamic modeling of the continuum robot, cumbersome data acquisition, and offline training, and exhibits stronger robustness in dealing with disturbances caused by the characteristics of continuum robots.

[0027] In addition, it should be noted that since there are many types of continuum robots, this application uses concentric tube robots for explanation and illustration. Other types, such as rope-driven robots, can also achieve motion control of the robot through the data-driven continuum robot motion control method disclosed in this application.

[0028] Specifically, the data-driven motion control method for a continuum robot according to this application includes the following steps: S100: Obtain the target spatial pose and real-time end-effector pose of the continuum robot, and construct a position error equation based on the target spatial pose and the real-time end-effector pose. In this embodiment, the target spatial pose is the preset target position of the end effector of the continuum robot, while the real-time end effector pose is captured by a sensor. For example, in one embodiment, a miniature electromagnetic sensor is implanted at the end effector of the continuum robot, and its position and orientation in space are sensed by an external magnetic field generator to obtain the corresponding real-time end effector pose.

[0029] Furthermore, such as Figure 2 As shown, based on the target spatial pose preset during the tracking and control process of the continuum robot, if there is an error between the real-time end pose and the target spatial pose, it means that the continuum robot needs to generate a drive signal to drive its end to move from the current position to the target position. In the process of generating the drive signal, it is necessary to combine the current position error to generate the corresponding drive vector in order to control the continuum robot to complete the overall motion control.

[0030] Therefore, the position error equation of the continuum robot is constructed. ,in, This represents the target spatial pose of the continuum robot. To represent the actual end-effector pose of the continuum robot, then This represents the position error between the target spatial pose and the actual end-effector pose. In this application, the position error is controlled to be a minimum value close to 0 by a position zeroing neural network (Pose_ZNN), thereby completing the tracking task.

[0031] In this application, the dynamic equations of position zeroing constructed based on the position zeroing neural network are transformed to obtain the driving vector solution equations. The Jacobian matrix parameters are used as unknown parameters. By introducing the velocity zeroing neural network, the Jacobian matrix parameters are estimated, realizing the synergy between the position zeroing neural network and the velocity zeroing neural network. The corresponding target driving vectors and Jacobian matrix parameters of the continuum robot are obtained. The target driving vectors are used to drive the end effector of the continuum robot to move from the actual end effector pose to the target spatial pose.

[0032] S200. Based on the position error equation, construct the position zeroing dynamic equation and solve for the target driving vector of the continuum robot. In some embodiments, step S200 above includes: By taking the time derivative of the position error equation, the second derivative equation is obtained; Based on the position-zeroing neural network, and combined with the position network design parameters, a position-zeroing dynamic equation is constructed. By combining the position-zeroing dynamic equation and the second derivative equation, the target driving vector is solved.

[0033] In this embodiment, the position-zeroing neural network dynamically zeroes the position error equation to solve for the target driving vector, which is used to drive the movement of the end effector of the continuum robot. Therefore, the position-zeroing dynamic equation constructed based on the position-zeroing neural network is specifically as follows: ,in, Indicates positional error. for The derivative with respect to time t, This represents the location network design parameters, used to adjust the convergence speed during the dynamic zeroing process of the error.

[0034] Furthermore, regarding the position error equation Differentiating both sides with respect to time t yields the second derivative equation. Therefore, by combining the aforementioned second derivative equation and the position-zeroing dynamic equation, we obtain the initial driving vector solution equation. ,in, For target space pose The first derivative with respect to time, and For real-time end pose The first derivative with respect to time maps the initial driving vector solution equation to the position-zeroing dynamic equation into a solution equation related to the mission objective.

[0035] In some embodiments, solving for the target driving vector by combining the position-zeroing dynamic equation and the second derivative equation includes: Establish the velocity mapping relationship between the drive space and the task space of the continuum robot; Based on the second derivative equation and combined with the velocity mapping relationship, the third derivative equation is obtained; Based on the position-zeroing dynamic equation, combined with the third derivative equation and the position error equation, the driving vector solution equation is obtained; The target driving vector is obtained by solving the driving vector solution equation, wherein the Jacobian matrix parameter is an unknown parameter.

[0036] Specifically, by combining the second derivative equation and the position-zeroing dynamic equation to obtain the initial driving vector solution equation, the position-zeroing dynamic equation is initially transformed into a solution equation related to the mission objective. Furthermore, the mission objective needs to be mapped to the driving objective to obtain a solution equation related to the driving vector.

[0037] In some embodiments, establishing the velocity mapping relationship of the continuum robot from the drive space to the task space includes: Based on the real-time end-effector pose and driving vector, the kinematic equations of the continuum robot are constructed. By taking the time derivative of all kinematic equations, a velocity mapping relationship is established for the continuum robot from the drive space to the task space, and the velocity mapping relationship is measured based on the Jacobian matrix.

[0038] In this embodiment, the kinematic equations of the continuum robot are constructed, and a nonlinear kinematic mapping is performed on the continuum robot. Specifically... Through mapping function Describes the nonlinear mapping relationship from joint space to maneuver space of a continuum robot, i.e., from the driving vector q to the real-time end-effector pose. The functional relationship is given by the driving vector, which, taking a concentric tube robot with two nested tubes as an example, represents the translation and rotation of the inner and outer tubes of the concentric tube robot at time t, corresponding to the extension and rotation of the two front sleeves of the concentric tube robot, respectively. Thus, in this embodiment, the driving vector is a 4×1 column vector, represented as q=[β1,θ1,β2,θ2]^T, where β1 represents the translation of the outer tube, θ1 represents the rotation of the outer tube, β2 represents the translation of the inner tube, and θ2 represents the rotation of the inner tube.

[0039] Furthermore, by differentiating both sides of the above kinematic equations with respect to time t, we obtain... The differential kinematic equations can be used to describe the relationship between the driving velocity and the end effector velocity of a continuum robot, where The driving vector is The Jacobian matrix of the time-continuum robot is the mapping function mentioned above. For driving vectors The partial derivatives; driving vector The first derivative with respect to time is used to represent the driving speed; For real-time end pose The first derivative with respect to time represents the end-effector velocity of the continuum robot, thereby establishing a velocity mapping relationship from the driving space to the task space of the continuum robot. This velocity mapping relationship is measured based on the Jacobian matrix, which corresponds to the driving vector q.

[0040] Therefore, by combining this velocity mapping relationship with the initial driving vector solution equation, we obtain the solution equation related to the driving vector q. Specifically, we obtain the second derivative equation. The speed mapping relationship described above By combining these methods, the second derivative equation is transformed into a third derivative equation related to the driving vector q. Furthermore, by replacing the aforementioned third derivative equation with the position-zeroing dynamic equation, we obtain the solution equation related to the driving vector q. .

[0041] Furthermore, for the above-mentioned equations related to the driving vector q, both sides of the equation are multiplied by the Jacobian matrix on the left. inverse matrix ,get By extracting from the formula Simplify to obtain the derivative of the driving vector q with respect to time. , specifically Understandably, the above relates to the derivative of the driving vector q with respect to time. The relevant equation is the driving vector solution equation. By solving this driving vector solution equation, we can obtain... Finally, by integrating q over time, we can obtain the driving vector q that moves the end effector of the continuum robot from its real-time end effector pose to its target spatial pose.

[0042] In some embodiments, the location network design parameters are functions related to the real-time drive vector of the continuum robot.

[0043] In this application, the position network design parameters are set as functions related to the real-time drive vector of the continuum robot. Taking a concentric tube robot as an example, the inner and outer tubes of the concentric tube robot are mostly designed for dominant stiffness, meaning the bending stiffness of the outer tube is much greater than that of the inner tube. When there is a significant overlap in the bending portion of the outer tube, the frictional force of the tube increases nonlinearly during extension, contraction, and torsion. Therefore, to address the enhanced nonlinearity caused by the change in the overlapping portion of the tube in the concentric tube robot, and to avoid the problem of slow control convergence speed caused by the position network design parameter η under conventional constant design, the position network design parameter η is set as a function related to the real-time drive vector q of the concentric tube robot. This allows for real-time reflection of the nonlinearity of the continuum robot and adjustment of the convergence speed during the dynamic zeroing process of the position error. Specifically... ,in, The preset basic convergence constant, A positive scaling factor. Represents the driving vector The overlap length between the inner and outer tubes of the concentric tube robot is calculated using the tube length and curvature. Specifically, through the design of the aforementioned position network parameters, the overlap length is automatically increased as the overlap of the tubes increases. This enhances the response speed of the pose tracking controller and maintains a high convergence speed.

[0044] Understandably, for different continuous robots, corresponding functions related to the real-time drive vector can be set according to the robot's motion performance to adjust the speed of the position error convergence process and improve reliability. Therefore, the specific equation for solving the drive vector is as follows: .

[0045] Furthermore, in solving the above-mentioned driving vector solution equations... At that time, there are parameters ,in Through the above steps S100-S200, the corresponding values ​​can be obtained for substitution calculations, and for the parameters... The parameters of the Jacobian matrix need to be obtained by performing matrix inverse operations, and in this solution equation, the parameters of the Jacobian matrix are unknown parameters.

[0046] It is worth noting that in traditional methods, the unknown Jacobian matrix parameters are usually obtained by physically modeling the continuum robot. However, the parameter calculation process in physical modeling is complex and highly dependent on the continuum robot, lacking versatility. In addition, the data acquisition and training process is cumbersome, and environmental interaction may damage the continuum robot. Therefore, this application introduces a velocity-zeroing neural network to solve for the Jacobian matrix parameters, improving the practicality of the continuum robot motion control method and achieving universal, high-precision, and fast and stable motion control.

[0047] S300. During the solution process, a dynamic equation with zero velocity is constructed to obtain the Jacobian matrix parameters of the continuum robot under the target driving vector. In some embodiments, step S300 above includes: Based on the velocity mapping relationship, a velocity error equation for the continuum robot under the target driving vector is constructed; Based on the velocity error equation, a velocity-zero dynamic equation is constructed to estimate the Jacobian matrix parameters of the continuum robot under the target driving vector.

[0048] Specifically, the Zero-Return Neural Network (ZNN) is a dynamic differential equation solver used to solve a time-varying matrix or vector equation in real time and online. Its "parameters" are the variables to be solved, such as the Jacobian matrix parameters to be solved in the velocity zero-return neural network of this embodiment. This is achieved by driving the velocity error equation to dynamically return to zero. As can be seen from the above step S200, the Jacobian matrix parameters need to correspond to the target driving vector. Therefore, the continuum robot needs to be mapped from the driving space to the task space to obtain the velocity error equation of the continuum robot under the driving vector, which is used to predict the Jacobian matrix parameters of the corresponding driving vector.

[0049] Specifically, the velocity mapping relationship in step S200 above maps the continuum robot from the driving space to the task space. Therefore, based on the above velocity mapping relationship, the velocity error equation of the continuum robot under the driving vector q can be determined. In a velocity-zero neural network, by making this velocity error... As the value approaches 0, the estimation of the Jacobian matrix can be completed, that is, the parameters of the Jacobian matrix under the driving vector q can be obtained. .

[0050] In some embodiments, the step of constructing a velocity-zero dynamic equation based on the velocity error equation and estimating the Jacobian matrix parameters of the continuum robot under the target driving vector includes: Taking the time derivative of the velocity error equation yields the first derivative equation; Based on the velocity-zero neural network and combined with the velocity network design parameters, a velocity-zero dynamic equation is constructed. By combining the dynamic equation for zeroing velocity and the first derivative equation, the Jacobian matrix parameters are solved.

[0051] In this embodiment, the velocity-zeroing dynamic equation constructed based on the velocity-zeroing neural network is specifically as follows: ,in, Indicates speed error, for The derivative with respect to time t, μ is a velocity network design parameter used to control the convergence speed of the velocity error during the dynamic zeroing process.

[0052] Furthermore, regarding the velocity error equation Differentiating both sides with respect to time t yields the first derivative equation. ,in, for Regarding time The derivative of for Regarding time The derivative of and for For the partial derivative with respect to time, the first derivative equation and the velocity error equation are further substituted into the velocity zero-reduction dynamic equation, transforming the velocity zero-reduction dynamic equation into a solution equation related to the Jacobian matrix parameters, specifically: This equation can be used to solve for the parameters of the Jacobian matrix.

[0053] More specifically, multiply both sides of the above equation by the time derivative of the driving vector q. inverse vector ,get By extracting from the formula And we compiled and obtained the solution equations for the Jacobian matrix, specifically... Understandably, by solving this equation, we can obtain the derivative of the Jacobian matrix of the continuum robot with respect to time t under the driving vector q. Then, by time integration, the Jacobian matrix coefficients under the required driving vector q can be obtained. Finally, the obtained Jacobian matrix coefficients The inverse is performed and passed to the driving vector solving equation obtained in step S200, realizing the collaboration of the position zeroing neural network and the velocity zeroing neural network, and finally solving to obtain the target driving vector.

[0054] In some embodiments, the speed network design parameters are functions related to speed error.

[0055] In this application, the speed network design parameters are set as functions related to the speed error. For example, in one embodiment, the speed network design parameters are set as the speed error norm. The function is used to adaptively adjust the convergence speed of the Jacobian matrix. Specifically, the design parameters of this speed network are: ,in The minimum convergence constant is preset. A positive scaling factor. For speed error The Euclidean norm, specifically, through the design of the above-mentioned speed network design parameters, when the speed error norm... When it is large, it will automatically increase. This improves the convergence speed, allowing the speed error to approach zero more quickly, and ensures the accuracy of the Jacobian matrix used to drive the solution. Therefore, the final solution equation for the Jacobian matrix is ​​specifically expressed as: .

[0056] It is worth noting that the design parameters of the velocity network enable the Jacobian matrix coefficients to be quickly and accurately adjusted when the continuum robot is subjected to external disturbances. This adjustment works in conjunction with the position-zeroing neural network to complete closed-loop control and improve the reliability of the motion control method.

[0057] S400. Using the target driving vector, the continuum robot is driven from the real-time end-effector pose to the target spatial pose.

[0058] In this application, combined with Figure 2 The control flow shown describes a process where, when there is an error between the real-time end-effector pose and the target space pose of the continuum robot, the key control variable, i.e., the driving vector q, is estimated using the position-zeroing neural network in step S200, and the velocity-zeroing neural network in step S300 estimates the Jacobian matrix coefficients of the continuum robot under the driving vector. This enables the synergistic effect of two zero-return neural networks, thereby achieving the motion control objective of the continuum robot and allowing the continuum robot to move from the real-time end-effector pose to the target space pose.

[0059] Specifically, this application achieves motion control of a continuum robot through a data-driven method using two cooperating zero-return neural networks. Specifically, through numerical iteration, the Jacobian matrix coefficients of the continuum robot are estimated in real time, and the driving vector is calculated accordingly to achieve motion control of the continuum robot without the need to pre-establish a complex kinematic model related to the continuum robot. This achieves universal, high-precision, fast and stable motion control of the continuum robot.

[0060] Please see Figure 3 As shown, the present invention also provides a data-driven motion control system for a continuum robot, the system comprising: First processing module 301: used to acquire the target spatial pose and real-time end-effector pose of the continuum robot, and to construct a position error equation based on the target spatial pose and the real-time end-effector pose; The second processing module 302 is used to construct a position-zeroing dynamic equation based on the position error equation and solve for the target driving vector of the continuum robot. The third processing module 303 is used to construct the velocity-zero dynamic equation during the solution process and obtain the Jacobian matrix parameters of the continuum robot under the target driving vector. Fourth processing module 304: used to drive the continuum robot from the real-time end-effector pose to the target spatial pose through the target driving vector.

[0061] It is understandable that, such as Figure 1 The content of the data-driven continuous robot motion control method embodiments shown herein is applicable to this data-driven continuous robot motion control system embodiment. The specific functions implemented by this data-driven continuous robot motion control system embodiment are as follows: Figure 1 The illustrated data-driven motion control method for continuum robots is the same as the one shown, and achieves the same beneficial effects. Figure 1 The beneficial effects achieved by the data-driven continuum robot motion control method embodiment shown are also the same.

[0062] It should be noted that the information interaction and execution process between the above systems are based on the same concept as the method embodiments of the present invention. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0063] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0064] Please see Figure 4As shown, this embodiment of the invention also provides a computer device 4, including: a memory 402 and a processor 401, and a computer program 403 stored in the memory 402. When the computer program 403 is executed on the processor 401, it implements the data-driven continuous robot motion control method as described in any of the above methods.

[0065] The computer device 4 may be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device 4 may include, but is not limited to, a processor 401 and a memory 402. Those skilled in the art will understand that... Figure 4 The computer device 4 is merely an example and does not constitute a limitation on the computer device 4. It may include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, etc.

[0066] The processor 401 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0067] In some embodiments, the memory 402 may be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 402 may be an external storage device of the computer device 4, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 4. Furthermore, the memory 402 may include both internal and external storage units of the computer device 4. The memory 402 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 402 can also be used to temporarily store data that has been output or will be output.

[0068] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the data-driven motion control method for a continuum robot as described in any of the above methods.

[0069] In this embodiment, if the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographic device / computer device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0070] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A data-driven motion control method for a continuum robot, characterized in that, include: The target spatial pose and real-time end-effector pose of the continuum robot are obtained, and a position error equation is constructed based on the target spatial pose and the real-time end-effector pose. Based on the position error equation, construct the position zeroing dynamic equation and solve for the target driving vector of the continuum robot; During the solution process, a dynamic equation with zero velocity is constructed to obtain the Jacobian matrix parameters of the continuum robot under the target driving vector; The target driving vector drives the continuum robot from the real-time end-effector pose to the target spatial pose.

2. The method as described in claim 1, characterized in that, The process of constructing a position-zeroing dynamic equation based on the position error equation and solving for the target driving vector of the continuum robot includes: By taking the time derivative of the position error equation, the second derivative equation is obtained; Based on the position-zeroing neural network, and combined with the position network design parameters, a position-zeroing dynamic equation is constructed. By combining the position-zeroing dynamic equation and the second derivative equation, the target driving vector is solved.

3. The method as described in claim 2, characterized in that, The step of combining the position-zeroing dynamic equation and the second derivative equation to solve for the target driving vector includes: Establish the velocity mapping relationship between the drive space and the task space of the continuum robot; Based on the second derivative equation and combined with the velocity mapping relationship, the third derivative equation is obtained; Based on the position-zeroing dynamic equation, combined with the third derivative equation and the position error equation, the driving vector solution equation is obtained; The target driving vector is obtained by solving the driving vector solution equation, wherein the Jacobian matrix parameter is an unknown parameter.

4. The method as described in claim 3, characterized in that, The establishment of the velocity mapping relationship between the drive space and the task space of the continuum robot includes: Based on the real-time end-effector pose and driving vector, the kinematic equations of the continuum robot are constructed. By taking the time derivative of the kinematic equations, a velocity mapping relationship between the drive space and the task space of the continuum robot is established, and the velocity mapping relationship is measured based on the Jacobian matrix.

5. The method as described in claim 2, characterized in that, The location network design parameters are functions related to the real-time drive vector of the continuum robot.

6. The method as described in claim 4, characterized in that, The process of constructing a velocity-zero dynamic equation to obtain the Jacobian matrix parameters of the continuum robot under the target driving vector includes: Based on the velocity mapping relationship, a velocity error equation for the continuum robot under the target driving vector is constructed; Based on the velocity error equation, a velocity-zero dynamic equation is constructed to estimate the Jacobian matrix parameters of the continuum robot under the target driving vector.

7. The method as described in claim 6, characterized in that, The process of constructing a velocity-zero dynamic equation based on the velocity error equation and estimating the Jacobian matrix parameters of the continuum robot under the target driving vector includes: Taking the time derivative of the velocity error equation yields the first derivative equation; Based on the velocity-zero neural network and combined with the velocity network design parameters, a velocity-zero dynamic equation is constructed. By combining the dynamic equation for zeroing velocity and the first derivative equation, the Jacobian matrix parameters are solved.

8. The method as described in claim 7, characterized in that, The speed network design parameters are functions related to the speed error.

9. A data-driven motion control system for a continuum robot, characterized in that, include: First processing module: used to acquire the target spatial pose and real-time end-effector pose of the continuum robot, and to construct a position error equation based on the target spatial pose and the real-time end-effector pose; The second processing module is used to construct a position-zeroing dynamic equation based on the position error equation and solve for the target driving vector of the continuum robot. The third processing module is used to construct the dynamic equation with zero velocity during the solution process and obtain the Jacobian matrix parameters of the continuum robot under the target driving vector. The fourth processing module is used to drive the continuum robot from the real-time end-effector pose to the target spatial pose using the target driving vector.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 8.

Images