Motor and joint angle conversion correction and kinematic calibration method for tethered robots

Abstract

The present invention discloses a method for correction and kinematic calibration of the motor and joint angle conversion of a rope-pulled flexible robot, aiming to obtain an accurate kinematic model of the rope-pulled flexible robot. The method comprises: obtaining the accurate motor and joint angle conversion relationship based on a numerical fitting method and a time-series neural network for error compensation; and proposing a kinematic calibration method based on a minimum kinematic parameter error model and combined with an artificial neural network, which takes into account the influence of non-geometric factors such as kinematic parameter errors and rope elasticity. This method addresses the model uncertainty of the rope-pulled flexible robot caused by the rope, and while being based on the actual physical model of the robot, it combines a data-driven neural network method. Compared with the robot nominal kinematic model and the conventional rigid robot kinematic calibration method, the present invention can obtain a more accurate and practical robot kinematic model.

Description

Technical Field

[0001] The present invention relates to the control field of a rope-pulled flexible robot, and in particular to a motor and joint angle conversion correction and kinematic calibration method of a rope-pulled flexible robot. Background Art

[0002] Unlike traditional industrial robots that directly drive rigid links through motors and reducers, rope-traction flexible robots use motors to pull ropes and utilize mechanical transmission structures such as pulleys to transmit force and motion to drive the robot. The introduction of rope-traction drive allows the motor and other heavier mechanical components to be placed near the base, significantly reducing the mass and inertia of the robot itself, achieving a higher load-to-weight ratio and a more compact structure. The introduction of ropes also makes the robot itself flexible, making interaction with humans safer. Due to these advantages, rope-traction drive is widely used in today's medical robots, collaborative robots and manipulators, and integrated robot arm-hand systems.

[0003] Due to factors such as rope material, load, and wear, ropes often exhibit non-repeatable nonlinear elastic characteristics. Furthermore, ropes often span multiple joints of a robot, introducing complex friction and, in some cases, delays and hysteresis, making accurate modeling of rope-pulled robots difficult. Conventional calibration methods, however, primarily target rigid robots and ignore the effects of the rope on the motor-joint angle conversion relationship and kinematic model. Therefore, for rope-pulled flexible robots, a kinematic calibration method that accounts for the effects of the rope is urgently needed to obtain an accurate robot kinematic model.

[0004] In view of this, the present invention is proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for correcting the motor and joint angle conversion and kinematic calibration of a rope robot, which can obtain an accurate rope robot kinematic model while considering the relationship between the rope and the angle conversion and the influence of the robot kinematics, thereby solving the above-mentioned technical problems existing in the prior art.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] A method for motor and joint angle conversion correction and kinematic calibration of a rope-tethered robot, comprising:

[0008] Step S1, deriving a theoretical conversion formula between the motor and joint angles of a rope robot based on a rope transmission structure model of each joint of the rope robot, wherein the rope robot refers to a rope-pulled flexible robot;

[0009] Step S2, using a sequential neural network including a gated recurrent unit to perform angle conversion error compensation on the theoretical conversion formula of the rope robot motor and joint angles obtained in step S1 to obtain an accurate angle conversion relationship;

[0010] Step S3, establishing a kinematic model of the rope robot using an exponential product formula according to the acquired accurate angle conversion relationship, and deriving a minimum kinematic parameter error model of the rope robot using the kinematic model;

[0011] Step S4: training the minimum kinematic parameter error model of the rope robot through a corresponding four-layer fully connected artificial neural network, and using the trained artificial neural network to compensate for the terminal posture predicted by the kinematic model of the rope-pulled flexible robot.

[0012] Compared with the prior art, the motor and joint angle conversion correction and kinematic calibration method of the rope robot provided by the present invention has the following beneficial effects:

[0013] Before utilizing a neural network, a theoretical conversion formula for the robot's motor and joint angles was derived based on the cable transmission structure model of each joint. A kinematic model of the robot was then established using an exponential product formula. This kinematic model was then used to derive the minimum kinematic parameter error model for the robot. This fully accounted for the impact of the cable on the robot's motor and joint angle conversion relationship and kinematic model. Based on the actual physical model, the accuracy of the conversion and kinematic model of the robot's motor and joint angles was further improved by combining a data-driven neural network. Furthermore, the proposed method is easy to use, requiring only experimental acquisition of the neural network and the required calibration data, such as motor angles, speeds, and torques. Furthermore, the method is highly versatile and can be readily extended to other flexible rope-pulled robots with similar structures, significantly alleviating the current problem of inaccurate models for flexible rope-pulled robots. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] Figure 1 This is a flow chart of a method for motor and joint angle conversion correction and kinematic calibration of a rope robot provided in an embodiment of the present invention.

[0016] Figure 2Schematic diagram of the overall kinematic structure of the rope-traction flexible robot provided in an embodiment of the present invention.

[0017] Figure 3 Schematic diagram of the shoulder joint degree of freedom configuration and joint configuration of the rope-traction flexible robot provided in an embodiment of the present invention.

[0018] Figure 4 Schematic diagram of the elbow joint degree of freedom configuration and joint configuration of the rope-traction flexible robot provided in an embodiment of the present invention.

[0019] Figure 5 Schematic diagram of the wrist joint degree of freedom configuration and joint configuration of the rope-traction flexible robot provided in an embodiment of the present invention.

[0020] Figure 6 Schematic diagram of the overall framework for correcting the motor angle-joint angle conversion relationship provided by an embodiment of the present invention.

[0021] Figure 7 A schematic diagram of kinematic calibration provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the specific content of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments, and do not constitute a limitation of the present invention. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0023] First, the following terms may be used in this article:

[0024] The term “and / or” means that either or both of them can be realized at the same time. For example, X and / or Y includes both “X” or “Y” and “X and Y”.

[0025] The terms "include," "comprises," "contains," "has," or other similar expressions should be interpreted as non-exclusive. For example, "including certain technical features (such as raw materials, components, ingredients, carriers, dosage forms, materials, dimensions, parts, components, mechanisms, devices, steps, procedures, methods, reaction conditions, processing conditions, parameters, algorithms, signals, data, products, or manufactured articles, etc.) should be interpreted as including not only the technical features explicitly listed, but also other technical features known in the art that are not explicitly listed.

[0026] The term "consisting of" excludes any technical features not explicitly listed. If used in a claim, this term renders the claim closed, excluding any technical features other than those explicitly listed, except for conventional impurities associated with them. If this term appears only in a clause of a claim, it limits only the elements explicitly listed in that clause; elements listed in other clauses are not excluded from the claim as a whole.

[0027] Unless otherwise specified or limited, the terms "mounted," "connected," "connect," and "fixed" should be interpreted broadly. For example, they can refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediary; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in this document based on specific circumstances.

[0028] When concentration, temperature, pressure, size or other parameters are expressed in the form of a numerical range, the numerical range should be understood to specifically disclose all ranges formed by the pairing of any upper limit, lower limit, or preferred value within the numerical range, regardless of whether the range is explicitly stated. For example, if a numerical range of "2 to 8" is stated, the numerical range should be interpreted as including ranges of "2 to 7," "2 to 6," "5 to 7," "3 to 4 and 6 to 7," "3 to 5 and 7," "2 and 5 to 7," etc. Unless otherwise specified, the numerical ranges stated herein include both their endpoints and all integers and fractions within the numerical range.

[0029] The terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings and are only for the convenience and simplification of description, and do not explicitly or implicitly indicate that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore should not be understood as a limitation to this document.

[0030] The following describes in detail the motor and joint angle conversion correction and kinematic calibration methods for the rope robot provided by the present invention. Any content not described in detail in the embodiments of the present invention belongs to the prior art known to professionals in this field. Where specific conditions are not specified in the embodiments of the present invention, the process is carried out in accordance with conventional conditions in the art or the conditions recommended by the manufacturer. Reagents or instruments used in the embodiments of the present invention, where the manufacturer is not specified, are all commercially available conventional products.

[0031] like Figure 1 As shown, an embodiment of the present invention provides a method for motor and joint angle conversion correction and kinematic calibration of a rope robot, comprising:

[0032] Step S1, deriving a theoretical conversion formula between the motor and joint angles of a rope robot based on a rope transmission structure model of each joint of the rope robot, wherein the rope robot refers to a rope-pulled flexible robot;

[0033] Step S2, using a sequential neural network including a gated recurrent unit to perform angle conversion error compensation on the theoretical conversion formula of the rope robot motor and joint angles obtained in step S1 to obtain an accurate angle conversion relationship;

[0034] Step S3, establishing a kinematic model of the rope robot using an exponential product formula according to the acquired accurate angle conversion relationship, and deriving a minimum kinematic parameter error model of the rope robot using the kinematic model;

[0035] Step S4: training the minimum kinematic parameter error model of the rope robot through a corresponding four-layer fully connected artificial neural network, and using the trained artificial neural network to compensate for the terminal posture predicted by the kinematic model of the rope-pulled flexible robot.

[0036] Preferably, in step S1 of the above method, the theoretical conversion formula of the rope robot motor and joint angle is derived based on the rope transmission structure model of each joint of the rope robot in the following manner, including:

[0037] According to the configuration characteristics of the shoulder, elbow and wrist joints of the rope-pulled flexible robot, the theoretical conversion formulas between the motor and joint angle of each joint are derived.

[0038] Preferably, in step S2 of the above method, the temporal neural network including a gated recurrent unit includes:

[0039] Gated recurrent unit GRU and fully connected multi-layer perceptron MLP containing temporal neural network; among them,

[0040] The gated loop unit is used to take the motor position, speed, torque and state at the previous moment as input to obtain the current moment posture;

[0041] The timing neural network including the gated cyclic unit is communicatively connected to the output end of the gated cyclic unit, and is used to obtain the angle conversion error by taking the output of the gated cyclic unit as input, perform error compensation on the theoretical conversion formula of the motor and joint angle, and correct the theoretical conversion formula of the motor and joint angle.

[0042] Preferably, in step S3 of the above method, the kinematic model of the rope-pulled flexible robot is established using the exponential product formula based on the shoulder, elbow and wrist joint configuration characteristics of the rope-pulled flexible robot, and the minimum kinematic parameter error model of the rope-pulled flexible robot is derived based on the kinematic model.

[0043] Preferably, in step S4 of the above method, the corresponding four-layer fully connected artificial neural network is used, with the motor angle of the rope-pulled flexible robot as input and the end position error of the rope-pulled flexible robot as output, to compensate for the end position predicted by the kinematic model of the rope-pulled flexible robot.

[0044] Preferably, the above-mentioned four-layer fully connected artificial neural network consists of an input layer connected in sequence to two hidden layers and an output layer.

[0045] In summary, it can be seen that the method of the embodiment of the present invention, because it fully considers the influence of the rope on the relationship between the robot motor and joint angle conversion and the kinematic model, further improves the accuracy of the rope robot motor and joint angle conversion and kinematic model based on the actual physical model and combines it with the data-driven neural network. In addition, the proposed method is easy to use. It only needs to collect the neural network and the series of data required for calibration through experiments, such as motor angle, speed and torque, to achieve calibration. At the same time, the method has high versatility and can be well extended to rope-pulled flexible robots with similar structures, greatly alleviating the problem of inaccurate models of rope-pulled flexible robots at present.

[0046] In order to more clearly demonstrate the technical solution and technical effects provided by the present invention, the motor and joint angle conversion correction and kinematic calibration method of the rope robot provided by the embodiment of the present invention is described in detail below with specific embodiments.

[0047] Example 1

[0048] The formulas in this article are for the example robot implemented in this paper;

[0049] Figure 2 The kinematic structure of the rope-pulled flexible robot of this embodiment is shown, which consists of a 3-DOF shoulder joint, a 1-DOF elbow joint, and a 3-DOF wrist joint.

[0050] Motor-driven ropes drive the movements of each joint: The robot's shoulder joint utilizes high-stiffness ropes and a large-radius active pulley to enhance the stiffness of the winch drive mechanism. The elbow joint utilizes a movable pulley system, with a single motor driving a pair of coupled ropes. Extending one rope while simultaneously contracting the other ensures a pure rolling motion. The wrist joint utilizes a rope tension-amplifying mechanical structure, and its antiparallelogram design equates the entire wrist joint to spherical rolling motion. Overall, it is essentially equivalent to a 7-DOF, 10-link tandem robotic arm driven by seven motors and subject to three independent sets of kinematic constraints.

[0051] According to relevant research, the motor angle of the rope-pulled flexible robot in this embodiment can be obtained as and joint angle The conversion relationship f s :

[0052] θ=f s (q) (1);

[0053] In formula (1), the three shoulder joints θ1-θ3 are driven by the first three motors q1-q3 respectively, and their conversion relationship is also linear: θ i =k i *q i ,i=1,2,3, where k i is the deceleration coefficient of the motor reducer and the rope pulley transmission (similar to the subsequent description); the two elbow joints θ4-θ5 are driven solely by motor q4, subject to the mechanism constraints of the rope traction transformation, and the final nonlinear conversion relationship is:

[0054]

[0055] In formula (2), ω1 is the rolling semicircle diameter of the elbow joint;

[0056] The first four wrist joints θ6-θ9 are also subject to the constraints of the rope traction mechanism and are driven by two motors q5 and q6 respectively, showing a nonlinear relationship. 10 It is driven by the seventh motor q7, and the conversion relationship after rope transmission is linear:

[0057]

[0058]

[0059] θ 10 =k7·q7 (3);

[0060] In formula (3), ω2 and ω3 are the rolling circle diameters of the wrist joint in two different rotation directions.

[0061] In order to obtain accurate conversion coefficients k1-k7, ω1-ω3, etc., after collecting data such as motor angle and joint angle through experiments, the numerical fitting method is used to fit the theoretical motor and joint angle conversion formula. On this basis, the influence of factors such as rope elasticity in actual scenarios is further considered. By designing a temporal neural network containing gated recurrent units GRU and fully connected multi-layer perceptrons MLP, the current motor angle q, speed Torque τ m As well as the state at the previous moment, further error compensation is performed on the joint angle calculated by the numerical fitting formula. A separate network is designed for error compensation of the conversion relationship between each motor angle and joint angle. The overall framework for correcting the conversion relationship between motor angle and joint angle can be found in Figure 5 , so the final accurate relationship between the motor angle and joint conversion is:

[0062] θ=f s (q)+e com (4);

[0063] In formula (4), e com It is the error compensation value predicted by the network.

[0064] Based on the accurate relationship between motor angle and joint conversion, the kinematic calibration of the rope-pulled flexible robot is performed, including:

[0065] Combining the accurate motor angle and joint conversion relationship based on the exponential product formula, the forward kinematic model of the rope-pulled flexible robot of this embodiment is obtained as follows:

[0066]

[0067] In formula (5), T∈SE(3) represents the pose of the robot end coordinate system {b} relative to the robot base coordinate system {s}; is the screw axis of joint i relative to the base coordinate system {s}; M∈SE(3) is the initial position of {b} relative to {s} when each joint of the robot is at the initial zero position; [S i ]∈se(3) is the screw axis S i 4x4 matrix representation of .

[0068] If the second-order error or above is ignored, the screw axis S of each joint can be defined as i The minimum kinematic parameter η i At the same time, the initial posture M in the forward kinematics model formula (5) is rewritten into a constant spin S st The exponential matrix form of The formula can be rewritten as:

[0069]

[0070] Take the partial derivative of the above formula (6) and multiply it by T on the right -1 , the minimum kinematic parameter error model for calibration can be obtained:

[0071]

[0072] From the above formula (7), we can know that the robot end position error δT is composed of the kinematic parameter deviation δη, joint angle deviation δθ and initial position deviation δS st The joint angle deviation has been compensated in the previous conversion formula correction, so here we choose to calibrate δη and δS st ;

[0073] Define the symbol ∨ to represent the matrix logarithmic mapping: The minimum kinematic parameter error model δTT can be derived -1 About δη and δS st The formula is:

[0074]

[0075] In formula (8),

[0076]

[0077] The same is true for the initial pose error related part, which satisfies:

[0078]

[0079] The relevant kinematic parameters in the above minimum kinematic parameter error model formula are still the screw axis S. Changing it to the minimum kinematic parameter η still requires taking

[0080]

[0081] Therefore, for the minimum kinematic parameter error model, Equation (8) can be rewritten as:

[0082]

[0083] definition:

[0084]

[0085]

[0086] Then the minimum kinematic parameter error model can be changed to the following form:

[0087] y = Γ·δx (14);

[0088] in,

[0089]

[0090] In formula (15), y is the end pose error of the robot; Γ is the calibration information matrix; δx is the minimum kinematic parameter error.

[0091] In order to ensure the adequacy of the calibration, it is necessary to collect enough data of different robot configurations. The number of groups of collected data is recorded as k, then:

[0092] Y=Γ k ·δx(16);

[0093] In formula (16), Y and Γ k To collect the terminal pose error and calibration information matrix of k sets of data, the minimum kinematic parameter deviation δx can be estimated according to the least squares method:

[0094] δx=Γ k + Y=(Γ k T Γ k ) -1 Γ k T ·Y (17);

[0095] In view of the particularity of the driving mode of the rope-pulled flexible robot, the influence of non-geometric factors such as rope elasticity on the model still needs to be considered. An artificial neural network is used to further compensate for the end-position error of the kinematic model calibrated based on the exponential product method to improve the accuracy of the model. The specific design structure can be found in Figure 6 .

[0096] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0097] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims. The information disclosed in the background technology section of this article is only intended to deepen the understanding of the overall background technology of the present invention, and should not be regarded as an admission or any form of implication that the information constitutes prior art already known to those skilled in the art.

Claims

1. A method for motor and joint angle conversion correction and kinematic calibration of a rope robot, characterized in that: include: Step S1, deriving a theoretical conversion formula between the motor and joint angles of a rope robot based on a rope transmission structure model of each joint of the rope robot, wherein the rope robot refers to a rope-pulled flexible robot; Step S2, using a sequential neural network including a gated recurrent unit to perform angle conversion error compensation on the theoretical conversion formula of the rope robot motor and joint angles obtained in step S1 to obtain an accurate angle conversion relationship; Step S3, establishing a kinematic model of the rope robot using an exponential product formula according to the acquired accurate angle conversion relationship, and deriving a minimum kinematic parameter error model of the rope robot using the kinematic model; The kinematic model of the rope robot is established as: （1）； in, Represents the robot end coordinate system Relative robot base coordinate system 's posture; It's a joint Relative to the base coordinate system The spiral axis; For the spiral shaft 4x4 matrix representation of; Indicates that the initial pose is rewritten as a constant spinor The exponential matrix form of ; The obtained minimum kinematic parameter error model of the rope robot is as follows: （2）； in, is the end pose error of the robot; is the calibration information matrix; is the minimum kinematic parameter error; （3）； In formula (3), the symbol Represents a matrix logarithmic mapping: ; （4）； Step S4: training the minimum kinematic parameter error model of the rope robot through a corresponding four-layer fully connected artificial neural network, and using the trained artificial neural network to compensate for the terminal posture predicted by the kinematic model of the rope-pulled flexible robot.

2. The method for motor and joint angle conversion correction and kinematic calibration of a rope robot according to claim 1, characterized in that: In step S1, the theoretical conversion formula between the motor and joint angle of the rope robot is derived based on the rope transmission structure model of each joint of the rope robot in the following manner, including: According to the configuration characteristics of the shoulder, elbow and wrist joints of the rope-pulled flexible robot, the theoretical conversion formulas between the motor and joint angle of each joint are derived.

3. The method for motor and joint angle conversion correction and kinematic calibration of a rope robot according to claim 1, characterized in that: In step S2, the temporal neural network including the gated recurrent unit includes: Gated recurrent unit and fully connected multilayer perceptron containing temporal neural network; among them, The gated loop unit is used to take the motor position, speed, torque and state at the previous moment as input to obtain the current moment posture; The timing neural network including the gated cyclic unit is communicatively connected to the output end of the gated cyclic unit, and is used to obtain the angle conversion error by taking the output of the gated cyclic unit as input, perform error compensation on the theoretical conversion formula of the motor and joint angle, and correct the theoretical conversion formula of the motor and joint angle.

4. The method for motor and joint angle conversion correction and kinematic calibration of a rope-robot according to any one of claims 1 to 3, characterized in that: In step S3, based on the configuration characteristics of the shoulder, elbow and wrist joints of the rope-pulled flexible robot, an exponential product formula is used to establish a kinematic model of the rope-pulled flexible robot, and a minimum kinematic parameter error model of the rope-pulled flexible robot is derived based on the kinematic model.

5. The method for motor and joint angle conversion correction and kinematic calibration of a rope-robot according to any one of claims 1 to 3, characterized in that: In step S4, a corresponding four-layer fully connected artificial neural network is used, with the motor angle of the rope-pulled flexible robot as input and the end position error of the rope-pulled flexible robot as output, to compensate for the end position predicted by the kinematic model of the rope-pulled flexible robot.

6. The method for motor and joint angle conversion correction and kinematic calibration of a rope-robot according to claim 5, characterized in that: The four-layer fully connected artificial neural network consists of an input layer connected in sequence to two hidden layers and an output layer.

Images