Mechanical arm trajectory tracking method and system based on interference estimation compensation

Abstract

The invention discloses a mechanical arm trajectory tracking method and system based on interference estimation compensation, and belongs to the technical field of trajectory tracking. The method comprises the steps that expected trajectory information of a mechanical arm is obtained, the expected trajectory information comprises the expected position qd, the expected angular speed # imgabs0 # and the expected angular acceleration # imgabs1 #, and a mechanical arm dynamic model is established based on the given system total disturbance d; designing a neural network interference observer, estimating the total disturbance of the system, and obtaining a system interference estimated value y; designing a super-spiral sliding mode controller and a global nonsingular terminal sliding mode surface s, and obtaining an equivalent control moment ueq based on expected trajectory information of the mechanical arm and total disturbance d of a given system; and based on the system interference estimated value y and the equivalent control moment ueq, a mechanical arm control law u meeting Lyapunov stability analysis is obtained, and mechanical arm trajectory tracking is achieved. The stability and trajectory tracking performance of the mechanical arm system can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of trajectory tracking, and particularly relates to a robotic arm trajectory tracking method and system based on disturbance estimation and compensation. Background Art

[0002] In the field of robotics, trajectory tracking is one of the key tasks in the robotic arm control system, and has important engineering and practical application significance. When a robotic arm performs complex operations such as assembly, welding, and grinding, the end effector often needs to move precisely along the expected trajectory to ensure operation accuracy and efficiency. However, a robotic arm is a complex system with non-linearity and coupling, and there are often disturbances and unknown errors in the actual operation process, which greatly increases the difficulty of precise control. Therefore, in the field of robotics, a robotic arm trajectory tracking control method with strong robustness, fast response, and high precision for disturbances and errors during operation is the core technical basis for improving the intelligent level of robotic arms and expanding their application scenarios.

[0003] To solve the problem of robotic arm trajectory tracking, domestic and foreign scholars have proposed control methods such as PID control, sliding mode control, adaptive control, and intelligent control based on deep learning. The PID control has a simple structure and is easy to implement, but it is sensitive to model parameter changes and external disturbances and has poor adaptability; although the traditional sliding mode control has a good implementation effect, it is prone to chattering, which can cause the robotic arm joint motors to heat up, increase wear, and even malfunction; the adaptive control enhances the system's ability to handle uncertainties, but the design is complex and parameter adjustment is difficult; the intelligent control method based on deep learning has good prediction and estimation capabilities, but has a large computational cost, high computational complexity, and high requirements for system real-time performance. For a robotic arm system with external disturbances, dynamic errors, and modeling uncertain parameter errors, the methods proposed above cannot achieve the accuracy required in the modern industrial manufacturing field. Therefore, for this robotic arm system, it is necessary to compensate for the total disturbance composed of external disturbances, dynamic errors, and modeling uncertain parameter errors. Currently, the vast majority of control methods use disturbance observers to compensate for the total disturbance, but such a processing method cannot accurately estimate the total disturbance due to the limited performance of the disturbance observer, and finally affects the trajectory tracking performance of the robotic arm.

[0004] Based on the above analysis, the main deficiencies in the current research work in the relevant field are as follows: (1) In the robotic arm system, there are various types of disturbances and complex dynamic changes. Traditional trajectory tracking control algorithms are difficult to achieve both high precision and strong robustness, and cannot meet the dual requirements of precision and stability in modern industrial manufacturing; (2) Although some existing control algorithms perform outstandingly in certain performance indicators, such as control precision or response speed, they are often difficult to be popularized and applied in actual robotic arm systems due to problems such as high computational complexity, poor robustness, or excessive burden on the actuators; (3) In the control design for actual disturbance scenarios, there are still challenges in the effective estimation and compensation of the total disturbance in the system. The lack of reliable observation and compensation affects the stability of the system. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a robotic arm trajectory tracking method and system based on disturbance estimation and compensation, which can improve the stability and trajectory tracking performance of the robotic arm system.

[0006] The present invention provides the following technical solutions:

[0007] In a first aspect, a robotic arm trajectory tracking method based on disturbance estimation and compensation is provided, including:

[0008] S1: Obtain the desired trajectory information of the robotic arm, including the desired position q d , the desired angular velocity the desired angular acceleration

[0009] S2: Based on the given total system disturbance d, establish the dynamic model of the robotic arm;

[0010] S3: Design a neural network disturbance observer to estimate the total system disturbance and obtain the system disturbance estimation value y;

[0011] S4: Design a super-twisting sliding mode controller and a global non-singular terminal sliding mode surface s to obtain the equivalent control torque u based on the desired trajectory information of the robotic arm and the given total system disturbance d eq ;

[0012] S5: Based on the system disturbance estimation value y and the equivalent control torque u eq , obtain the robotic arm control law u that satisfies the Lyapunov stability analysis to achieve robotic arm trajectory tracking.

[0013] Optionally, the dynamic model of the robotic arm is:

[0014]

[0015] where q, and They are respectively the angular positions, angular velocities, and angular accelerations of the joints of the robotic arm, and τ is the torque applied to the joints of the robotic arm. is the estimated value of the centrifugal force and Coriolis force matrix. is the estimated value of the gravity matrix. is the estimated value of the inverse matrix of the positive definite symmetric bounded inertia matrix.

[0016] Optionally, step S3 is specifically as follows:

[0017] S31: The neural network disturbance observer includes an input layer, two hidden layers, and an output layer; the input layer receives input signals related to the system state (position and state) and control signals.

[0018] The output R k (t) of the k-th neuron in the first hidden layer is:

[0019]

[0020] where φ k (·) is the activation function of the first hidden layer. is the weight connection from the l-th neuron in the input layer to the k-th hidden neuron in the first hidden layer, and the value range of k is [1, m]. a l (t) is the output of the l-th neuron in the input layer.

[0021] The output S j (t) of the j-th neuron in the second hidden layer is:

[0022]

[0023] where the value range of j is [1, n]. represents the weight connection from the k-th neuron in the first hidden layer to the j-th neuron in the second hidden layer. ψ j (·) is the activation function of the second hidden layer.

[0024] The output y(t) of the output layer is:

[0025]

[0026] where represents the weight connection from the j-th neuron in the second hidden layer to the output layer.

[0027] S32: Use the training data to train the neural network disturbance observer, and use the trained neural network disturbance observer to estimate the total system disturbance to obtain the system disturbance estimate value y.

[0028] Optionally, in step S32, during the training process, update the weights of the output layer and the two hidden layers, specifically:

[0029]

[0030] where d(t) is the total system disturbance at time t, and η is the neural network learning rate.

[0031] Optionally, step S4 is specifically:

[0032] S41: Design the global nonsingular terminal sliding mode surface s;

[0033]

[0034] where e is the position tracking error, is the angular velocity tracking error, and c and b are two diagonal constant coefficient matrices;

[0035] S42: Design the super-twisting sliding mode controller as follows:

[0036]

[0037] where, is the derivative of the sliding mode surface, v is the integral of , λ and α are the first and second control gains respectively, and sign(·) is the sign function;

[0038] S43: Utilize the global nonsingular terminal sliding mode surface s and the manipulator dynamic model to transform the derivative of the sliding mode surface, and obtain the equivalent control torque u eq ;

[0039]

[0040] Optionally, step S5 is specifically:

[0041] S51: Based on the system disturbance estimate value y and the equivalent control torque u eq , obtain the manipulator control law u, and the expression is:

[0042]

[0043] where s is the sliding mode surface, v is the integral of , e is the position tracking error, is the angular velocity tracking error, and c and b are two diagonal constant coefficient matrices;

[0044] S52: Set the Lyapunov stability function as:

[0045]

[0046] Among them, w h is the neural network weight, is the ideal weight, γ h is the Lyapunov weight factor, h = 1, 2, 3;

[0047] S53: Differentiate the Lyapunov stability function to obtain the constraint conditions for the asymptotic stability of the manipulator system and the constraint conditions for convergence within the effective time;

[0048] S54: Based on the constraint conditions for the asymptotic stability of the manipulator system and the constraint conditions for convergence within the effective time, derive the manipulator control law u that satisfies the constraint conditions.

[0049] Optionally, step S53 is specifically:

[0050] S531: Differentiate the Lyapunov stability function;

[0051]

[0052] Among them, η is the neural network learning rate, dw h is the difference between the previous moment value and the current moment value of the gradient J, and Tr represents the matrix trace;

[0053] S532: Let The constraint condition for the asymptotic stability of the manipulator system is and negative definite;

[0054] When dw h and are in the same direction, then

[0055] When sv - α·v·sign(s) ≤ 0, or sv - α·v·sign(s) > 0 and satisfies α ≥ s max at this time, then negative definite;

[0056] S533: The constraint condition for convergence within the effective time is:

[0057] Among them, ε is the convergence rate parameter, ε > 0, and γ is the fractional power parameter, 0 < γ < 1.

[0058] On the second aspect, a manipulator trajectory tracking system based on disturbance estimation compensation is provided, including:

[0059] Desired trajectory information acquisition module: Acquire the desired trajectory information of the manipulator, including the desired position q d, Desired angular velocity Desired angular acceleration

[0060] Model construction module: Based on the given total system disturbance d, establish the dynamic model of the robotic arm;

[0061] Disturbance estimation module: Design a neural network disturbance observer to estimate the total system disturbance and obtain the system disturbance estimation value y;

[0062] Equivalent control torque acquisition module: Design a super-twisting sliding mode controller and a global non-singular terminal sliding mode surface s to obtain the equivalent control torque u based on the desired trajectory information of the robotic arm and the given total system disturbance d eq ;

[0063] Trajectory tracking module: Based on the system disturbance estimation value y and the equivalent control torque u eq , obtain the robotic arm control law u that satisfies the Lyapunov stability analysis to achieve robotic arm trajectory tracking.

[0064] In a third aspect, a computer device is provided, including a processor and a memory; wherein, when the processor executes the computer program stored in the memory, the steps of the robotic arm trajectory tracking method based on disturbance estimation compensation described in any item of the first aspect are implemented.

[0065] In a fourth aspect, a computer-readable storage medium is provided for storing a computer program; when the computer program is executed by a processor, the steps of the robotic arm trajectory tracking method based on disturbance estimation compensation described in any item of the first aspect are implemented.

[0066] Compared with the prior art, the beneficial effects of the present invention are:

[0067] The present invention compensates for the errors in the robotic arm system, including external disturbances and dynamic errors, through a neural network disturbance observer, and compensates for the unobservable errors in the robotic arm system, that is, the modeling uncertain parameter errors, using a global super-twisting sliding mode controller. Finally, the stability and trajectory tracking performance of the robotic arm system are achieved, improving the accuracy of robotic arm trajectory tracking; through the neural network, the computational cost of robotic arm trajectory tracking is reduced; an integral term is introduced to effectively reduce the chattering phenomenon in the robotic arm sliding mode control process; the stability of the robotic arm system is ensured through the disturbance observer and the sliding mode controller; the present invention takes into account both high precision and strong robustness, meeting the dual requirements of modern industrial manufacturing for precision and stability. Description of the Drawings

[0068] Figure 1 is a flowchart of the robotic arm trajectory tracking method based on disturbance estimation compensation of the present invention;

[0069] Figure 2It is the component diagram of the neural network disturbance observer of the present invention;

[0070] Figure 3 It is the comparison diagram of the disturbance observer estimation provided by the embodiment of the present invention;

[0071] Figure 4 It is the tracking error curve diagram provided by the embodiment of the present invention;

[0072] Figure 5 It is the control torque output curve diagram provided by the embodiment of the present invention. Detailed implementation manners

[0073] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0074] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the term "including" and any deformation thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0075] Embodiment 1

[0076] As Figure 1 shown, a manipulator trajectory tracking method based on disturbance estimation compensation is provided, including:

[0077] S1: Obtain the manipulator desired trajectory information, including the desired position q d , the desired angular velocity the desired angular acceleration

[0078] The acquisition of the desired trajectory information can refer to the prior art. As an alternative, set the desired positions of the joints of the manipulator to q d1 = q d2 = q d3 = 10sin(0.5πt) + 10sin(πt), q d4 = q d5 = q d6 = 5sin(0.1πt), and the desired angular velocity of each joint The desired angular acceleration of each joint can be obtained by taking the first derivative of the desired trajectory of each joint. The desired angular acceleration of each joint can be obtained by taking the second derivative of the desired trajectory of each joint.

[0079] S2: Based on the given total system disturbance d, establish the dynamic model of the robotic arm.

[0080] Specifically, S21: Establish the dynamic model of the robotic arm as follows:

[0081]

[0082] In the formula, q, respectively represent the angular position, angular velocity, and angular acceleration of each joint of the robotic arm. M(q) is a positive definite symmetric bounded inertia matrix, is the centrifugal force and Coriolis force matrix, G(q) is the gravity matrix, τ is the torque applied to each joint of the robotic arm, and d is the total system disturbance;

[0083] Exemplarily, set the initial state of the system q = [0, 0, 0, 0, 0, 0] T ,

[0084] S22: There are modeling errors in the robotic arm system, which are expressed as follows:

[0085]

[0086] Among them, and are the estimated values of M(q), and G(q) respectively;

[0087] Thus, the total disturbance of the system can be written as:

[0088]

[0089] Among them is the dynamic error;

[0090] Exemplarily, let the dynamic error Let the external disturbance τ d = [τ d1 , τ d2 , τ d3 , τ d4 , τ d5 , τ d6 T , where τ d1 = 0.04sin(πt), τ d2 = 0.04sin(0.5πt) + 0.4sin(πt), τ d3= 0.04sin(πt) + 0.4sin(0.5πt), τ d4 = 0.01cos(0.2πt) + 0.01cos(0.4πt), τ d5 = 0.01cos(0.5πt), τ d6 = 0.01cos(πt). Set the modeling uncertainty parameter errors as ΔM = 0.2M, ΔC = 0.2C, ΔG = 0.2G.

[0091] S23: Rewrite the manipulator dynamics model into the following form:

[0092]

[0093] S3: Design a neural network disturbance observer to estimate the total system disturbance and obtain the system disturbance estimate value y.

[0094] S31: The neural network disturbance observer includes an input layer, two hidden layers, and an output layer; the input layer receives input signals related to the system state (position and state) and control signals.

[0095] As Figure 2 shown, set the input layer to have i neurons, which receive input signals related to the system state and control signals, and the output of the l-th neuron is a l (t), where the value range of l is [1, i];

[0096] Exemplarily, i = 7, and the value range of l is [1, 7].

[0097] Set the first hidden layer to have m neurons, then the output R k (t) of the k-th neuron is:

[0098]

[0099] where, φ k (·) is the activation function of the first hidden layer, is the weight connection from the l-th neuron in the input layer to the k-th hidden neuron in the first hidden layer, and the value range of k is [1, m]; as an option, m = 10, the value range of k is [1, 10], and φ k (·) = tanh(·).

[0100] Set the second hidden layer to have n neurons, and the output S j (t) of the j-th neuron in the second hidden layer is:

[0101]

[0102] where, the value range of j is [1, n], represents the weight connection from the k-th neuron in the first hidden layer to the j-th neuron in the second hidden layer, ψ j (·) is the activation function of the second hidden layer; optionally, n = 5, and the value range of j is [1, 5], φ k (·) = tanh(·).

[0103] The output y(t) of the output layer is:

[0104]

[0105] where represents the weight connection from the j-th neuron in the second hidden layer to the output layer.

[0106] S32: Train the neural network disturbance observer using the training data, and estimate the total system disturbance using the trained neural network disturbance observer to obtain the system disturbance estimation value y.

[0107] During the training process, update the weights of the output layer and the two hidden layers, specifically:

[0108]

[0109] where d(t) is the total system disturbance at time t, and η is the neural network learning rate.

[0110] The specific training process of the neural network disturbance observer refers to the prior art, and the acquisition of the training data also refers to the prior art.

[0111] Figure 3 The (a)-(f) in are respectively the comparison diagrams of the disturbance observer estimations of the six joints of the robotic arm. In Figure 3 , the red solid line is the disturbance value of each joint, the blue dotted line is the estimation result of the disturbance value by the neural network disturbance observer method provided by the present invention, and the black dashed line is the estimation result of the disturbance value by the nonlinear disturbance observer design method. Based on Figure 3 it can be seen that the present application greatly improves the estimation result of the disturbance value.

[0112] S4: Design a super-twisting sliding mode controller and a global nonsingular terminal sliding mode surface s to obtain the equivalent control torque u based on the given total system disturbance d eq .

[0113] Specifically, S41: Design a global nonsingular terminal sliding mode surface s;

[0114]

[0115] where e is the position tracking error, is the angular velocity tracking error, and c and b are two diagonal constant coefficient matrices;

[0116] Optionally, c = diag[2, 2, 2, 2, 2, 2], b = diag[3, 3, 3, 3, 3, 3], where diag[·] represents that the main diagonal elements are the values within [·], and the elements outside the main diagonal are all 0.

[0117] S42: Design the super-twisting sliding mode controller as follows:

[0118]

[0119] where, is the derivative of the sliding mode surface, v is the integral of λ and α are the first and second control gains respectively, and sign(·) is the sign function, satisfying λ, α ≥ 0;

[0120] Exemplarily, λ = 1, α = 1.

[0121] S43: Utilize the global nonsingular terminal sliding mode surface s and the manipulator dynamics model to transform the derivative of the sliding mode surface and obtain the equivalent control torque u for the manipulator trajectory tracking eq .

[0122] Specifically, according to the manipulator desired trajectory information obtained in step S1, including the desired position q d , the desired angular velocity the desired angular acceleration obtain the position tracking error e, the angular velocity tracking error and the angular acceleration tracking error as follows:

[0123]

[0124] Take the derivative of the global nonsingular terminal sliding mode surface s according to the super-twisting sliding mode controller:

[0125]

[0126] Obtain the equivalent control u for the manipulator trajectory tracking eq :

[0127]

[0128] S5: Based on the system disturbance estimate value y and the equivalent control torque u eq , obtain the manipulator control law u that satisfies the Lyapunov stability analysis to achieve the manipulator trajectory tracking.

[0129] S51: Based on the system interference estimation value y and the equivalent control torque u eq , obtain the manipulator control law u, and the expression is:

[0130]

[0131] where s is the sliding surface, v is the integral of , e is the position tracking error, is the angular velocity tracking error, and c and b are two diagonal constant coefficient matrices.

[0132] S52: Set the Lyapunov stability function as:

[0133]

[0134] where w h is the neural network weight, is the ideal weight, γ h is the Lyapunov weight factor, and h = 1, 2, 3;

[0135] S53: Differentiate the Lyapunov stability function to obtain the constraint conditions for the asymptotic stability of the manipulator system and the constraint conditions for convergence within the effective time.

[0136] Step S53 is specifically:

[0137] S531: Differentiate the Lyapunov stability function;

[0138]

[0139] where η is the neural network learning rate, dw h is the difference between the previous moment value and the current moment value of the gradient J, and Tr represents taking the trace of the matrix; As an option, the gradient J is

[0140] S532: Let where is used to analyze the stability of the super-twisting sliding mode control, is used to analyze the stability of the neural network disturbance observer, and the constraint conditions for the asymptotic stability of the manipulator system are and is negative definite;

[0141] When dw h is in the same direction as , then

[0142] When sv - α·v·sign(s) ≤ 0, or sv - α·v·sign(s) > 0 and satisfies α≥s max When negative definite;

[0143] S533: The constraint condition for convergence within the effective time is:

[0144] where ε is the convergence rate parameter, ε > 0, γ is the fractional power parameter, 0 < γ < 1.

[0145] That is, the manipulator system is asymptotically stable and satisfies in the form of, where ε > 0, 0 < γ < 1, then the system converges within the effective time T f inside;

[0146] where

[0147] S54: Based on the constraint condition of the asymptotic stability of the manipulator system and the constraint condition of convergence within the effective time, the manipulator control law u that satisfies the constraint condition is derived.

[0148] Figure 4 In (a)-(f) of, they are respectively the tracking error curves of the six joints of the manipulator. In Figure 4 , the red solid line is the tracking error curve of each joint of the traditional sliding mode control method based on neural network disturbance observation, the purple dashed line is the tracking error curve of each joint of the non-singular terminal sliding mode control method based on neural network disturbance observation, the blue dotted line is the tracking error curve of each joint of the fast non-singular terminal sliding mode control method based on neural network disturbance observation, and the black solid line is the tracking error curve of each joint of the manipulator trajectory tracking method based on neural network disturbance observer and global super-twisting sliding mode control provided by the present invention. Based on Figure 4 it can be seen that the tracking error of each joint of the manipulator trajectory tracking method based on neural network disturbance observer and global super-twisting sliding mode control of the present application is the lowest.

[0149] As Figure 5 shown, the control torque output curve diagram provided by the present invention. In Figure 5 , (a) is the control torque output curve of each joint of the traditional sliding mode control method based on neural network disturbance observation, (b) is the control torque output curve of each joint of the non-singular terminal sliding mode control method based on neural network disturbance observation, (c) is the control torque output curve of each joint of the fast non-singular terminal sliding mode control method based on neural network disturbance observation, and (d) is the control torque output curve of each joint of the manipulator trajectory tracking method based on neural network disturbance observer and global super-twisting sliding mode control provided by the present invention. Based on Figure 5 it can be seen that the output of the control torque of the present application is more stable and avoids the chattering phenomenon.

[0150] Example 2

[0151] A robotic arm trajectory tracking system based on interference estimation and compensation, comprising:

[0152] Desired trajectory information acquisition module: Acquire the desired trajectory information of the robotic arm, including the desired position q d , desired angular velocity desired angular acceleration

[0153] Model construction module: Based on the given total system disturbance d, establish the dynamic model of the robotic arm;

[0154] Disturbance estimation module: Design a neural network disturbance observer to estimate the total system disturbance and obtain the system disturbance estimation value y;

[0155] Equivalent control torque acquisition module: Design a super-twisting sliding mode controller and a global non-singular terminal sliding mode surface s to obtain the equivalent control torque u based on the desired trajectory information of the robotic arm and the given total system disturbance d eq ;

[0156] Trajectory tracking module: Based on the system disturbance estimation value y and the equivalent control torque u eq , obtain the robotic arm control law u that satisfies the Lyapunov stability analysis and achieve robotic arm trajectory tracking.

[0157] For a more specific process of the above method, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be elaborated herein.

[0158] Example 3

[0159] The present invention provides a computer device, comprising a processor and a memory; wherein, when the processor executes the computer program stored in the memory, the steps of the above-mentioned robotic arm trajectory tracking method based on interference estimation and compensation are implemented.

[0160] For a more specific process of the above method, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be elaborated herein.

[0161] Example 4

[0162] The present invention provides a computer-readable storage medium for storing a computer program; when the computer program is executed by a processor, the steps of the above-mentioned robotic arm trajectory tracking method based on interference estimation and compensation are implemented.

[0163] For a more specific process of the above method, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details will not be elaborated herein.

[0164] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For the systems, devices, and storage media disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description of the method part.

[0165] Those skilled in the art can clearly understand that the technologies in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions in the embodiments of the present invention, in essence, or the parts that contribute to the prior art can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0166] The above are only the preferred embodiments of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions within the idea of the present invention belong to the protection scope of the present invention. It should be pointed out that for those of ordinary skill in the art, several improvements and refinements made without departing from the principle of the present invention should be regarded as within the protection scope of the present invention.

Claims

1. A robot arm trajectory tracking method based on interference estimation and compensation, characterized in that: include: S1: Get the expected trajectory information of the robot arm, including the expected position q d , the expected angular velocity Expected angular acceleration S2: Based on the given system total disturbance d, establish the robot arm dynamics model; S3: Design a neural network disturbance observer to estimate the total disturbance of the system and obtain the estimated value y of the system disturbance; S4: Design a super-helical sliding mode controller and a global non-singular terminal sliding mode surface s to obtain the equivalent control torque u based on the desired trajectory information of the manipulator and the given system total disturbance d eq ; S5: Based on the system disturbance estimate y and equivalent control torque u eq , obtain the robot control law u that satisfies the Lyapunov stability analysis and realize the robot trajectory tracking.

2. The robot arm trajectory tracking method based on interference estimation and compensation according to claim 1, characterized in that: The mechanical arm dynamics model is: Among them, q, and are the angular position, angular velocity and angular acceleration of each joint of the robotic arm, τ is the torque applied to each joint of the robotic arm, are estimates of the centrifugal and Coriolis force matrices, is the estimated value of the gravity matrix, is an estimate of the inverse of the inertia matrix that is positive definite, symmetric, and bounded.

3. The robot arm trajectory tracking method based on interference estimation and compensation according to claim 1, characterized in that: Step S3 is specifically: S31: The neural network disturbance observer includes an input layer, two hidden layers and an output layer; the input layer receives input signals related to the system state and the control signal; The output R of the kth neuron in the first hidden layer k (t) is: Among them, φ k (·) is the activation function of the first hidden layer, is the weighted connection from the lth neuron in the input layer to the kth hidden neuron in the first hidden layer, where the value range of k is [1, m], and a l (t) is the output of the lth neuron in the input layer; The output S of the jth neuron in the second hidden layer is j (t) is: The value range of j is [1, n]. represents the weighted connection from the kth neuron in the first hidden layer to the jth neuron in the second hidden layer, ψ j (·) is the activation function of the second hidden layer; The output y(t) of the output layer is: in, Represents the weighted connection from the jth neuron in the second hidden layer to the output layer; S32: Using the training data to train a neural network disturbance observer, and using the trained neural network disturbance observer to estimate the total disturbance of the system, and obtain a system disturbance estimation value y.

4. The robot arm trajectory tracking method based on interference estimation and compensation according to claim 3, characterized in that: Step S32, during the training process, the weights of the output layer and the two hidden layers are updated, specifically: Among them, d(t) is the total disturbance of the system at time t, and η is the learning rate of the neural network.

5. The robot arm trajectory tracking method based on interference estimation and compensation according to claim 2, characterized in that: Step S4 is specifically: S41: Design the global non-singular terminal sliding surface s; Where, e is the position tracking error, is the angular velocity tracking error, c and b are two diagonal constant coefficient matrices; S42: Design the super-helical sliding mode controller as follows: in, is the derivative of the sliding surface s, v is The integral of , λ and α are the first and second control gains respectively, sign(·) is the sign function; S43: Using the global non-singular terminal sliding surface s and the robot dynamics model, the derivative of the sliding surface Convert and obtain the equivalent control torque u of the robot trajectory tracking eq ; 6. The robot arm trajectory tracking method based on interference estimation and compensation according to claim 2, characterized in that: Step S5 is specifically: S51: Based on the system disturbance estimation value y and equivalent control torque u eq , get the robot control law u, the expression is: Where s is the sliding surface, v is The integral of , e is the position tracking error, is the angular velocity tracking error, c and b are two diagonal constant coefficient matrices; S52: Set the Lyapunov stability function to: Among them, w h is the neural network weight, is the ideal weight, γ h is the Lyapunov weight factor, h = 1, 2, 3; S53: Deriving the Lyapunov stability function, obtaining the constraints of the asymptotic stability of the manipulator system and the constraints of convergence within the effective time; S54: Based on the constraints of asymptotic stability of the robot system and the constraints of convergence within the effective time, the robot control law u that satisfies the constraints is derived.

7. The robot arm trajectory tracking method based on interference estimation and compensation according to claim 6, characterized in that: Step S53 is specifically: S531: Derivation of the Lyapunov stable function; Among them, η is the neural network learning rate, dw h is the difference between the gradient J value at the previous moment and the current moment, Tr represents the matrix trace; S532: Order The constraint condition for the asymptotic stability of the robot system is: and Negative determination; When dw h and If the directions are the same, When sv-α·v·sign(s)≤0, or sv-α·v·sign(s)>0 and satisfies α≥s max When Negative determination; S533: The constraints for convergence within the effective time are: Among them, ε is the convergence rate parameter, ε>0, γ is the fractional power parameter, 0<γ<1.

8. A robot arm trajectory tracking system based on interference estimation and compensation, characterized in that: include: Expected trajectory information acquisition module: obtains the expected trajectory information of the robot arm, including the expected position q d , the expected angular velocity Expected angular acceleration Model building module: Based on the given system total disturbance d, the robot arm dynamics model is established; Disturbance estimation module: design a neural network disturbance observer, estimate the total disturbance of the system, and obtain the estimated value y of the system disturbance; Equivalent control torque acquisition module: Design a super-helical sliding mode controller and a global non-singular terminal sliding mode surface s to obtain the equivalent control torque u based on the desired trajectory information of the manipulator and the given system total disturbance d eq ; Trajectory tracking module: based on the system disturbance estimate y and equivalent control torque u eq , obtain the robot control law u that satisfies the Lyapunov stability analysis and realize the robot trajectory tracking.

9. A computer device, characterized in that: It comprises a processor and a memory; wherein, when the processor executes the computer program stored in the memory, it implements the steps of the robot arm trajectory tracking method based on interference estimation and compensation as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: Used to store computer programs; when the computer programs are executed by the processor, the steps of the robot arm trajectory tracking method based on interference estimation and compensation as described in any one of claims 1 to 7 are implemented.

Images