Intelligent fixed-time super-spiral sliding mode control method for hybrid mechanism

Abstract

The invention discloses an intelligent fixed-time super-spiral sliding mode control method for a hybrid mechanism. Establishing an uncertain hybrid mechanism dynamic model by adopting a Lagrangian method; based on an uncertain hybrid mechanism dynamic model, a nonlinear saturation item in super-spiral sliding mode control is improved into a transcendental function, so that the problem that the convergence speed is low relative to first-order sliding mode control is solved; secondly, designing a power term with a fixed time performance parameter in a superspiral sliding mode control algorithm transcendental function; then, in order to break through the limited range of the fixed-time super-spiral sliding mode control to overcome the system uncertainty, introducing a radial basis function neural network which can approach unknown nonlinear functions including system inherent dynamics and uncertainty in a hybrid mechanism dynamic model; and finally, implementing through software programming. According to the invention, through improving the super-spiral sliding mode control algorithm, rapid convergence of the series-parallel mechanism system within a preset fixed time can be realized, and the convergence performance of the series-parallel mechanism system is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of hybrid mechanism control and relates to an intelligent fixed-time super-helical sliding mode control method for a hybrid mechanism. Background Art

[0002] In order to solve the problems of existing serial and parallel mechanisms, such as poor heavy load bearing capacity and low flexibility, a hybrid mechanism is developed by combining the advantages of serial and parallel mechanisms. The mechanism has the advantages of high precision, high rigidity, and strong load-bearing capacity. However, the actual operation of the hybrid mechanism system is affected by various uncertain factors, such as modeling errors, joint friction, external disturbances, etc. These uncertainties affect the motion control accuracy of the hybrid mechanism and even destroy the stability of the system. Therefore, how to achieve high-performance tracking control of the hybrid mechanism is a challenging problem.

[0003] In the article "Cao Yuanyuan, Gao Guoqin, Wu Xintong. Global fast terminal sliding mode control of a new hybrid conveying mechanism" (Cao Yuanyuan, Gao Guoqin, Wu Xintong, Information Technology, 2016, Issue 4, Pages 5-9+13), a global fast terminal sliding mode controller is designed for a new hybrid automobile electrophoretic coating conveying mechanism. This method uses the Lagrangian method to establish its dynamic model, and designs a global fast terminal sliding mode controller based on it. The theoretical convergence time of the system is given, and its stability is proved by the Lyapunov stability theorem. However, this method is designed based on the finite time stability theory. Compared with fixed time control, its convergence time theoretically depends on the initial state information of the system, and when the hybrid mechanism faces a wide range of uncertainties, the robustness of the designed global fast terminal sliding mode control will be disabled due to the limitation of the sliding mode existence condition. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the present invention adopts the Lagrangian method to establish a dynamic model of an uncertain hybrid mechanism in view of the problem that the hybrid mechanism is affected by uncertainties such as modeling errors, joint friction, and external disturbances during the actual operation of the system; based on the dynamic model of the uncertain hybrid mechanism, the nonlinear saturation term in the super-helical sliding mode control is improved into a transcendental function to solve the problem that the super-helical sliding mode control has a slow convergence speed relative to the first-order sliding mode control; secondly, a power term with a fixed-time performance parameter is designed in the transcendental function of the super-helical sliding mode control algorithm to construct a fixed-time super-helical sliding mode surface to improve its global convergence performance in the arrival stage and the sliding stage; then, in order to break through the limited range of fixed-time super-helical sliding mode control to overcome the uncertainty of the system, a radial basis neural network that can approximate the unknown nonlinear function including the inherent dynamics and uncertainty of the system in the dynamic model of the hybrid mechanism is introduced, and an intelligent fixed-time super-helical sliding mode controller is further designed, and then an intelligent fixed-time super-helical sliding mode control method for the hybrid mechanism is proposed to break through the limited range of fixed-time super-helical sliding mode control to overcome the uncertainty of the system and improve the robustness of the hybrid mechanism system to a large range of changes in uncertainty factors.

[0005] The technical solution of the present invention is: a hybrid mechanism intelligent fixed-time super-helical sliding mode control method, comprising the following steps:

[0006] 1) For the uncertain hybrid mechanism, the inverse kinematics analysis of the hybrid mechanism is carried out by using analytical method, and the Jacobian matrix is ​​further obtained;

[0007] 2) The Lagrangian method is used to establish the dynamic model of the uncertain hybrid mechanism in the joint space;

[0008] 3) Based on the uncertain hybrid mechanism dynamics model established in step 2), a new super-helical sliding mode control algorithm is designed by improving the nonlinear saturation term into a transcendental function to solve the problem of its slow convergence speed compared with the first-order sliding mode control;

[0009] 4) Based on step 3), a power term with a fixed time performance parameter is designed in the transcendental function of the superhelical sliding mode control algorithm to construct a fixed time superhelical sliding mode surface and improve its global convergence performance in the arrival stage and the sliding stage;

[0010] 5) Based on step 4), in order to break through the limited range of fixed-time super-helical sliding mode control to overcome system uncertainty, a radial basis neural network is further designed to approximate the unknown nonlinear function in the dynamic model of the hybrid mechanism, including the inherent dynamics and uncertainty of the system;

[0011] 6) Based on the uncertain hybrid mechanism dynamics model and fixed-time super-helical sliding mode surface designed in steps 2, 3, 4, and 5), and combined with the radial basis function neural network, an intelligent fixed-time super-helical sliding mode controller is designed;

[0012] 7) Through software programming, an intelligent fixed-time super-helical sliding mode control of a hybrid mechanism is realized.

[0013] Furthermore, in step 3), based on the uncertain hybrid mechanism dynamics model, the nonlinear saturation term in the super-helical sliding mode control is improved to a transcendental function

[0014]

[0015] Among them, s, w. are sliding surface, sliding surface derivative, superhelical term and its derivative respectively; k1 and k2 are adjustable parameters and satisfy k1, k2>0; sig(s)=|s|sign(s).

[0016] Furthermore, in step 4), a power term with a fixed time performance parameter is designed in the transcendental function of the superhelical sliding mode control algorithm to construct a fixed time superhelical sliding mode surface to improve its global convergence performance in the arrival stage and the sliding stage s = [s1, s2, s3, s4, s5, s6] T :

[0017]

[0018] Where α and β are adjustable parameters and 0<α<1, β>1; e=x1-x 1d is the active joint tracking error vector, x1 is the active joint pose vector, x 1d is the desired pose vector of the active joint; is the active joint velocity error vector, x2 is the active joint velocity vector, is the desired velocity vector of the active joint; sig α (e)=|e| α sign(e), sig β (e)=|e| β sign(e);

[0019] Based on the super spiral algorithm, the following reaching law is designed:

[0020]

[0021] In the formula, w. are the sliding surface derivative, superhelical term and its derivative respectively; k1, k2, k3, k4 are adjustable parameters and satisfy k1, k2, k3, k4>0;

[0022] Furthermore, in step 5), in order to break through the limited range of fixed-time super-helical sliding mode control to overcome the uncertainty of the system, a radial basis neural network is further designed to approximate the unknown nonlinear function in the dynamic model of the hybrid mechanism, including the inherent dynamics and uncertainty of the system; the neural network uses the actual posture, expected posture and tracking error of each active joint of the hybrid mechanism as the network input, the Gaussian function as the membership function, the network nodes are 36, and the network output is the approximate value of the uncertainty term

[0023] Further, the intelligent fixed-time super-helical sliding mode controller in step 6) is:

[0024]

[0025] Where u is the intelligent fixed-time super-helical sliding mode controller, is the derivative of the desired velocity vector of the active joint, e is the tracking error vector of the active joint, is the active joint velocity error vector, w is the superhelical term; α and β are adjustable parameters and 0<α<1, β>1; k1 and k3 are adjustable parameters and k1, k3>0;

[0026] And the reaching law based on the super spiral algorithm is:

[0027]

[0028] In the formula, w. They are the sliding surface derivative, superhelical term and its derivative respectively; h(x) and ε are network weight error, membership function, and approximation error, respectively; T is vector transpose; k1, k2, k3, and k4 are adjustable parameters and satisfy k1, k2, k3, and k4>0;

[0029]

[0030] Then, under the action of the designed intelligent fixed-time super-helical sliding mode control method, the system state can be controlled in a fixed time. It reaches stability and converges to a compact set Ω within Ω, and the convergence time is

[0031]

[0032] Among them, λ max is the maximum eigenvalue of the matrix, λ min is the minimum eigenvalue of the matrix, B, T1, T2 are positive definite matrices respectively; η0 are positive constants and satisfy 0<η0<∞; V1(x) is the selected Lyapunov function.

[0033] The present invention proposes for the first time an intelligent fixed-time super-helical sliding mode control method for a series-parallel mechanism to achieve high-precision trajectory tracking control of the series-parallel mechanism, and its characteristics and beneficial effects are:

[0034] 1. Considering that the actual operation of the hybrid mechanism system is affected by uncertainties such as modeling errors, joint friction, and external disturbances, based on the uncertain dynamic model of the hybrid mechanism, the nonlinear saturation term in the super-helical sliding mode control is improved into a transcendental function to solve the problem of its slow convergence speed relative to the first-order sliding mode control;

[0035] 2. Design a power term with fixed time performance parameters in the transcendental function of the superhelical sliding mode control algorithm to construct a fixed time superhelical sliding mode surface, improve its global convergence performance in the arrival stage and the sliding stage, so that the superhelical sliding mode control hybrid mechanism system can converge quickly within a preset fixed time, thereby enabling the uncertain hybrid mechanism system to obtain better convergence performance;

[0036] 3. In order to break through the limited range of fixed-time super-helical sliding mode control to overcome system uncertainty, a radial basis neural network that can approximate the unknown nonlinear functions in the dynamic model of the hybrid mechanism, including the inherent dynamics and uncertainty of the system, is introduced, and an intelligent fixed-time super-helical sliding mode control algorithm for the hybrid mechanism is further designed to improve the robustness of the hybrid mechanism system to a large range of changes in uncertainty factors. Therefore, this method can effectively improve the motion tracking control performance of the hybrid mechanism in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a structural diagram of a hybrid mechanism. In the figure: 1. Guide rail 2. Base 3. Travel drive motor 4. Reducer 5. Moving slide block 6. Lifting drive motor 7. Connecting rod 8. Driven wheel 9. Driving wheel 10. Connecting rod 11. Car body 12. Flipping drive motor 13. Electric cylinder.

[0038] Figure 2 It is the principle diagram of the intelligent fixed-time super-helical sliding mode control method for the hybrid mechanism.

[0039] Figure 3 It is a simplified diagram of the lifting and flipping mechanism structure.

[0040] Figure 4 It is the distributed structure diagram of the hybrid mechanism control system.

[0041] Figure 5 It is the trajectory tracking curve of the unilateral active joint of the hybrid mechanism.

[0042] Figure 6 It is the trajectory tracking error curve of the unilateral active joint of the hybrid mechanism. DETAILED DESCRIPTION

[0043] The specific implementation manner of the present invention is further described below with reference to the accompanying drawings.

[0044] Firstly, the hybrid mechanism is inversely kinematically analyzed by analytical method, and the Jacobian matrix J is further obtained; the Lagrangian method is used to establish the dynamic model of the uncertain hybrid mechanism in the joint space; based on the dynamic model of the uncertain hybrid mechanism, the nonlinear saturation term in the superhelical sliding mode control is improved into a transcendental function to solve the problem of its slow convergence speed relative to the first-order sliding mode control; secondly, a power term with fixed time performance parameters is designed in the transcendental function of the superhelical sliding mode control algorithm to construct a fixed time superhelical sliding mode surface to improve its global convergence performance in the arrival stage and the sliding stage; then, in order to break through the limited range of fixed time superhelical sliding mode control to overcome the uncertainty of the system, a radial basis neural network that can approximate the unknown nonlinear function including the inherent dynamics and uncertainty of the system in the dynamic model of the hybrid mechanism is introduced, and the intelligent fixed time superhelical sliding mode control sliding mode controller of the hybrid mechanism is further designed, and then a hybrid mechanism intelligent fixed time superhelical sliding mode control method is proposed; finally, the intelligent fixed time superhelical sliding mode control of the hybrid mechanism is realized through software programming. The specific method is as follows:

[0045] 1) Use analytical method to perform inverse kinematic analysis on the hybrid mechanism and obtain the Jacobian matrix

[0046] The motion coordinate system is established with the midpoint of the connecting rod as the origin. The position parameters of the end effector of the hybrid mechanism in generalized coordinates are defined as q = (z, β) T , where z represents the lifting displacement of the midpoint of the connecting rod along the z-axis (in m); β represents the counterclockwise flip angle of the midpoint of the connecting rod around the y-axis (in rad). The analytical method is used to perform inverse kinematics analysis on the hybrid mechanism to obtain the position inverse solution equation. By taking the derivative of the two ends of the position inverse solution equation, we get:

[0047]

[0048] In the formula, is the speed of the connecting rod midpoint in the z-axis direction (in m / s), is the angular velocity of the connecting rod midpoint flipping counterclockwise around the y-axis (in rad / s), is the active joint velocity vector, are the speed variables of the four sliders, are the speed variables of the two driving wheels, J is the Jacobian matrix, is the connecting rod midpoint velocity vector.

[0049] 2) The Lagrangian method is used to establish the dynamic model of the uncertain hybrid mechanism in the joint space

[0050] According to the definition of Lagrangian function, the standard dynamic equation of the hybrid mechanism is established:

[0051]

[0052] Where M(q) is the positive definite inertia matrix, are the Coriolis force and centrifugal force, G(q) is the gravity term, Q is the generalized force (unit: N·m); q, They are respectively the end position vector, velocity vector and acceleration vector of the hybrid mechanism.

[0053] Considering that the hybrid mechanism system is affected by uncertainties such as modeling errors, joint friction, and external disturbances during operation, the dynamic model of the uncertain hybrid mechanism in joint space is established:

[0054]

[0055] In the formula, is the nominal inertia matrix; are the nominal Coriolis force and centrifugal force terms; is the nominal gravity term; ΔM(x), ΔG(x) is the modeling error; N(t) is the friction term (in N·m); d is the external disturbance term; x, and are the active joint posture, velocity, and acceleration vector respectively; τ is the driving torque of the motor corresponding to the active joint (in N·m);

[0056] The dynamic model of the uncertain hybrid mechanism can be further expressed as:

[0057]

[0058] Where x1 is the active joint pose vector; is the derivative of the active joint posture vector; x2 is the active joint velocity vector; x2 is the active joint acceleration vector; is the nonlinear inherent dynamics of the system; is the inverse matrix of the inertia matrix; u = τ is the motor driving torque (in N·m); is the lumped disturbance term including modeling error, joint friction and external disturbance.

[0059] 3) Based on the dynamic model of uncertain hybrid mechanism, the nonlinear saturation term in super-helical sliding mode control is improved into a transcendental function

[0060]

[0061] Among them, s, w. are sliding surface, sliding surface derivative, superhelical term and its derivative respectively; k1 and k2 are adjustable parameters and satisfy k1, k2>0; sig(s)=|s|sign(s).

[0062] 4) Design a power term with fixed time performance parameters in the transcendental function of the super-helical sliding mode control algorithm to construct a fixed time super-helical sliding mode surface and improve its global convergence performance in the arrival stage and sliding stage s = [s1, s2, s3, s4, s5, s6] T :

[0063]

[0064] Where α and β are adjustable parameters and 0<α<1, β>1; e=x1-x 1d is the active joint tracking error vector, x1 is the active joint pose vector, x 1d is the desired pose vector of the active joint; is the active joint velocity error vector, x2 is the active joint velocity vector, is the desired velocity vector of the active joint; sig α (e)=|e| α sign(e), sig β (e)=|e| β sign(e);

[0065] Based on the super spiral algorithm, the following reaching law is designed:

[0066]

[0067] In the formula, w. are the sliding surface derivative, superhelical term and its derivative respectively; k1, k2, k3, k4 are adjustable parameters and satisfy k1, k2, k3, k4>0;

[0068] 5) In order to break through the limited range of fixed-time super-helical sliding mode control to overcome the uncertainty of the system, a radial basis neural network is further designed to approximate the unknown nonlinear function in the dynamic model of the hybrid mechanism, including the inherent dynamics and uncertainty of the system; the neural network takes the actual posture, expected posture and tracking error of each active joint of the hybrid mechanism as the network input, Gaussian function as the membership function, network nodes as 36, and network output as the approximate value of the uncertainty term

[0069] 6) Based on the designed uncertain hybrid mechanism dynamics model, fixed-time super-helical sliding mode surface, and radial basis neural network, an intelligent fixed-time super-helical sliding mode controller is designed as follows:

[0070]

[0071] Where u is the intelligent fixed-time super-helical sliding mode controller, is the derivative of the desired velocity vector of the active joint, e is the tracking error vector of the active joint, is the active joint velocity error vector, w is the superhelical term; α and β are adjustable parameters and 0<α<1, β>1; k1 and k3 are adjustable parameters and k1, k3>0;

[0072] And the reaching law based on the super spiral algorithm is:

[0073]

[0074] In the formula, w. They are the sliding surface derivative, superhelical term and its derivative respectively; h(x) and ε are network weight error, membership function, and approximation error, respectively; T is vector transpose; k1, k2, k3, and k4 are adjustable parameters and satisfy k1, k2, k3, and k4>0;

[0075]

[0076] Then, under the action of the designed intelligent fixed-time super-helical sliding mode control method, the system state can be controlled in a fixed time. It reaches stability and converges to a compact set Ω within Ω, and the convergence time is

[0077]

[0078] Among them, λ max is the maximum eigenvalue of the matrix, λ min is the minimum eigenvalue of the matrix, B, T1, T2 are positive definite matrices respectively; η0 are positive constants and satisfy 0<η0<∞; V1(x) is the selected Lyapunov function.

[0079] 7) Through software programming, the intelligent fixed-time super-helical sliding mode control of the hybrid mechanism is realized.

[0080] The motor drive control quantity of each active joint of the hybrid mechanism is calculated according to formula (15), and the analog quantity obtained by digital-to-analog conversion of the control quantity is sent to the servo driver corresponding to the motor to drive the hybrid mechanism to move according to the desired trajectory.

[0081] The following provides embodiments of the present invention:

[0082] Example 1

[0083] like Figure 1 As shown, 1-guide rail 2-base 3-travel drive motor 4-reducer 5-moving slider 6-lifting drive motor 7-connecting rod 8-driven wheel 9-driving wheel 10-connecting rod 11-car body in white 12-flipping drive motor 13-electric screw.

[0084] The present invention mainly focuses on realizing high-performance operation of a series-parallel mechanism by using an intelligent fixed-time super-helical sliding mode control technology. The principle block diagram of the intelligent fixed-time super-helical sliding mode control system of the series-parallel mechanism is shown in FIG. Figure 2 As shown, the specific implementation of the control method is as follows:

[0085] 1) Use analytical method to perform inverse kinematic analysis on the hybrid mechanism and obtain the Jacobian matrix J

[0086] according to Figure 3 The lifting and flipping mechanism diagram shown in the figure uses the analytical method to perform inverse kinematics analysis on the hybrid mechanism and obtain the position inverse solution equation. By taking the derivative of both ends of the position inverse solution equation, we get:

[0087]

[0088] Where J is the Jacobian matrix; is the active joint velocity vector; is the connecting rod midpoint velocity vector.

[0089] 2) The Lagrangian method is used to establish the dynamic model of the uncertain hybrid mechanism in the joint space

[0090] According to the definition of Lagrangian function, the standard dynamic equation of the hybrid mechanism is established:

[0091]

[0092] Where M(q) is the inertia matrix; are the Coriolis force and centrifugal force; G(q) is the gravity term; Q is the generalized force (unit: N·m); q, They are respectively the end position vector, velocity vector and acceleration vector of the hybrid mechanism.

[0093] Considering that the hybrid mechanism system is affected by uncertainties such as modeling errors, joint friction, and external disturbances during operation, the dynamic model of the uncertain hybrid mechanism in joint space is established:

[0094]

[0095] In the formula, is the nominal inertia matrix; are the nominal Coriolis force and centrifugal force terms; is the nominal gravity term; ΔM(x), ΔG(x) is the modeling error; N(t) is the friction term (in N·m); d is the external disturbance term; x, and are the active joint posture, velocity, and acceleration vector respectively; τ is the driving torque of the motor corresponding to the active joint (in N·m);

[0096] The dynamic model of the uncertain hybrid mechanism can be further expressed as:

[0097]

[0098] Where x1 is the active joint pose vector; is the derivative of the active joint pose vector; x2 is the active joint velocity vector; is the active joint acceleration vector; is the nonlinear inherent dynamics of the system; is the inverse matrix of the inertia matrix; u = τ is the motor driving torque (in N·m); is the lumped disturbance term including modeling error, joint friction and external disturbance.

[0099] 3) Based on the dynamic model of uncertain hybrid mechanism, the nonlinear saturation term in super-helical sliding mode control is improved into a transcendental function

[0100]

[0101] Among them, s, w. are sliding surface, sliding surface derivative, superhelical term and its derivative respectively; k1 and k2 are adjustable parameters and satisfy k1, k2>0; sig(s)=|s|sign(s).

[0102] 4) Design a power term with fixed time performance parameters in the transcendental function of the super-helical sliding mode control algorithm to construct a fixed time super-helical sliding mode surface and improve its global convergence performance in the arrival stage and sliding stage s = [s1, s2, s3, s4, s5, s6] T :

[0103] s=e+sig α (e)+sig β (e) (24)

[0104] Where α and β are adjustable parameters and 0<α<1, β>1; e=x1-x 1d is the active joint tracking error vector, x1 is the active joint pose vector, x 1d is the desired pose vector of the active joint; is the active joint velocity error vector, x2 is the active joint velocity vector, is the desired velocity vector of the active joint; sig α (e)=|e| α sign(e), sig β (e)=|e| β sign(e);

[0105] Based on the super spiral algorithm, the following reaching law is designed:

[0106]

[0107] In the formula, w. are the sliding surface derivative, superhelical term and its derivative respectively; k1, k2, k3, k4 are adjustable parameters and satisfy k1, k2, k3, k4>0;

[0108] 5) In order to break through the limited range of fixed-time super-helical sliding mode control to overcome the uncertainty of the system, a radial basis neural network is further designed to approximate the unknown nonlinear function in the dynamic model of the hybrid mechanism, including the inherent dynamics and uncertainty of the system; the neural network takes the actual posture, expected posture and tracking error of each active joint of the hybrid mechanism as the network input, Gaussian function as the membership function, network nodes as 36, and network output as the approximate value of the uncertainty term

[0109] 6) Based on the uncertain hybrid mechanism dynamics model, the fixed-time super-helical sliding mode surface, and the radial basis function neural network, an intelligent fixed-time super-helical sliding mode controller is designed as follows:

[0110]

[0111] Where u is the intelligent fixed-time super-helical sliding mode controller, is the derivative of the desired velocity vector of the active joint, e is the tracking error vector of the active joint, is the active joint velocity error vector, w is the superhelical term; α and β are adjustable parameters and 0<α<1, β>1; k1 and k3 are adjustable parameters and k1, k3>0;

[0112] And the control law based on the super helical algorithm is:

[0113]

[0114] In the formula, w. They are the sliding surface derivative, superhelical term and its derivative respectively; h(x) and ε are network weight error, membership function, and approximation error, respectively; T is vector transpose; k1, k2, k3, and k4 are adjustable parameters and satisfy k1, k2, k3, and k4>0;

[0115] Then, under the action of the designed intelligent fixed-time super-helical sliding mode control method, the system state can be controlled in a fixed time. It reaches stability and converges to a compact set Ω within Ω, and the convergence time is

[0116]

[0117] Among them, λ max is the maximum eigenvalue of the matrix, λ min is the minimum eigenvalue of the matrix, B, T1, T2 are positive definite matrices respectively; η0 are positive constants and satisfy 0<η0<∞; V1(x) is the selected Lyapunov function.

[0118] 7) Through software programming, the intelligent fixed-time super-helical sliding mode control of the hybrid mechanism is realized.

[0119] The intelligent fixed-time super-helical sliding mode control system of the hybrid mechanism is based on the "upper computer + lower computer" distributed control structure. It has high flexibility and reliability and is suitable for large-scale, complex control systems and scenarios with high requirements for system response speed and real-time performance. The control system structure is as follows Figure 4As shown in the figure. The control system consists of four parts: the upper computer PC, the lower computer UMAC, the servo system and the hybrid mechanism prototype. Among them, the upper computer (PC) system is 32-bit Win7-Ultimate, and the processor uses Intel Core i7-4790 (3.60GHz), which is used to complete human-computer interaction, code initialization and compilation, status monitoring and other functions. The lower computer (UMAC) mainly includes 1 TURBO PMAC2 OPT-5C0 CPU motherboard card, 2 ACC-24E2A axis boards, 1 ACC-65E I / O board, 1 ACC-E1 power board, etc., which are used to complete data calculation, digital-to-analog conversion and motion control. The servo system is divided into three parts: servo drive, servo motor and reducer. The servo drive receives the torque analog signal from UMAC and drives the servo motor through the signal. The reducer increases the torque by reducing the motor speed. The servo drivers include Mitsubishi MRJ4-70A (30W) for the lifting motor, Mitsubishi MR-J4-100A (30W) for the walking motor and the flipping motor. The servo motors include 4 HG-KR73BJ (750W) for lifting and 4 HG-SR102BJ (30W) for walking and flipping. The reducer model is EVH-115-20-K7-28HF24. In addition, in order to realize the feedback control loop of the system, the real-time position information of the servo motors of each joint needs to be known, so a 22-bit (4194304 pulses / rev) high-resolution absolute position encoder is selected.

[0120] Through MATLAB simulation and hybrid mechanism prototype experiments, the proposed intelligent fixed time super helical sliding mode control method (IFTSTSMC) is compared with the fixed time super helical sliding mode control (FTSTSMC), and the results are respectively Figure 5 The track tracking curve of the unilateral active joint of the hybrid mechanism shown in the figure, Figure 6 The unilateral active joint trajectory tracking error curve is shown.

[0121] Depend on Figure 5 and Figure 6 It can be seen that compared with the fixed-time super-helical sliding mode control method, the hybrid mechanism system under the action of the intelligent fixed-time super-helical sliding mode control method proposed in the present invention can not only make the system converge within a fixed time, but also improve the system's robustness to large-scale changes in uncertainty factors.

[0122] It should be understood that the above embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent modifications of the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

Claims

1. An intelligent fixed-time super-helical sliding mode control method for a series-parallel mechanism, characterized in that: The steps include: 1) For the uncertain hybrid mechanism, the inverse kinematics analysis of the hybrid mechanism is carried out by using analytical method, and the Jacobian matrix is ​​further obtained; 2) The Lagrangian method is used to establish the dynamic model of the uncertain hybrid mechanism in the joint space; 3) Based on the uncertain hybrid mechanism dynamics model established in step 2), a new super-helical sliding mode control algorithm is designed by improving the nonlinear saturation term into a transcendental function to solve the problem of its slow convergence speed compared with the first-order sliding mode control; 4) Based on step 3), a power term with a fixed time performance parameter is designed in the transcendental function of the superhelical sliding mode control algorithm to construct a fixed time superhelical sliding mode surface and improve its global convergence performance in the arrival stage and the sliding stage; 5) Based on step 4), in order to break through the limited range of fixed-time super-helical sliding mode control to overcome system uncertainty, a radial basis neural network is further designed to approximate the unknown nonlinear function in the dynamic model of the hybrid mechanism, including the inherent dynamics and uncertainty of the system; 6) Based on the uncertain hybrid mechanism dynamics model and fixed-time super-helical sliding mode surface designed in steps 2, 3, 4, and 5), and in combination with the radial basis function neural network, an intelligent fixed-time super-helical sliding mode controller is designed; 7) Through software programming, an intelligent fixed-time super-helical sliding mode control of a hybrid mechanism is realized.

2. The method according to claim 1, characterized in that: In the step 3), based on the uncertain hybrid mechanism dynamics model, the nonlinear saturation term in the super-helical sliding mode control is improved to a transcendental function Among them, s, w. are sliding surface, sliding surface derivative, superhelical term and its derivative respectively; k1 and k2 are adjustable parameters and satisfy k1, k2>0; sig(s)=|s|sign(s).

3. The method according to claim 1, characterized in that: In the step 4), a power term with a fixed time performance parameter is designed in the transcendental function of the superhelical sliding mode control algorithm to construct a fixed time superhelical sliding mode surface: s=[s1,s2,s3,s4,s5,s6] T : Where α and β are adjustable parameters and 0<α<1, β>1; e=x1-x 1d is the active joint tracking error vector, x1 is the active joint pose vector, x 1d is the desired pose vector of the active joint; is the active joint velocity error vector, x2 is the active joint velocity vector, is the desired velocity vector of the active joint; sigα(e)=|e| α sign(e), sig β (e)=|e| β sign(e); Based on the super spiral algorithm, the following reaching law is designed: In the formula, w. are the sliding surface derivative, superhelical term and its derivative respectively; k1, k2, k3, k4 are adjustable parameters and satisfy k1, k2, k3, k4>0; 4. The method according to claim 1, characterized in that: In the step 5), the neural network uses the actual posture, expected posture and tracking error of each active joint of the hybrid mechanism as the network input, Gaussian function as the membership function, the network nodes are 36, and the network output is the uncertainty approximation value 5. The method according to claim 1, characterized in that: In step 6), the intelligent fixed-time super-helical sliding mode controller is: Where g(x) is the inverse matrix of the inertia matrix; u is the intelligent fixed-time super-helical sliding mode controller, is the derivative of the desired velocity vector of the active joint, e is the tracking error vector of the active joint, is the active joint velocity error vector, w is the superhelical term; α and β are adjustable parameters and 0<α<1, β>1; k1 and k3 are adjustable parameters and k1, k3>0; And the reaching law based on the super spiral algorithm is: In the formula, w. They are the sliding surface derivative, superhelical term and its derivative respectively; h(x) and ε are network weight error, membership function, and approximation error, respectively; T is vector transpose; k1, k2, k3, and k4 are adjustable parameters and satisfy k1, k2, k3, and k4>0; Then, under the action of the designed intelligent fixed-time super-helical sliding mode control method, the system state can be controlled in a fixed time. It reaches stability and converges to a compact set Ω within Ω, and the convergence time is Among them, λ max is the maximum eigenvalue of the matrix, λ min is the minimum eigenvalue of the matrix, B, T1, T2 are positive definite matrices respectively; η0 are positive constants and satisfy 0<η0<∞; V1(x) is the selected Lyapunov function.

Images