Self-adaptive series elastic driver control method with preset time convergence

Abstract

The invention discloses a preset time convergence self-adaptive series elastic driver control method, which comprises the following steps of: firstly, selecting a gain function under a new time scale, performing time scale change on a state space expression of a series elastic driver, mapping time to a new time domain, and finishing conversion from preset time stability to progressive time stability. And secondly, designing an unknown parameter estimation form, and defining an estimation error dynamic state, thereby obtaining a time scale-based adaptive law of the unknown parameter, and enabling the unknown parameter of the system to converge to a field near a true value within a preset time. And then designing a tracking error preset time sliding mode surface, and ensuring that system output realizes accurate trajectory tracking within preset time. And finally, strictly proving the stability of the preset time controller and the estimator based on the Lyapunov theory. Based on the method, the problem of rapid and high-precision trajectory tracking control of the series elastic driver under the conditions of unknown model parameters and external disturbance is solved.

Description

Technical Field

[0001] This invention relates to an adaptive series flexible actuator control method with preset time convergence, belonging to the field of flexible actuator control technology. Background Technology

[0002] With the increasing aging of the population, the elderly are experiencing problems such as lower back pain, weakened lower limb motor skills, and decreased balance, which not only reduce their quality of life but also impose huge care costs on society and families. Traditional rehabilitation treatment relies on one-on-one manual assistance from therapists, which is not only labor-intensive and inefficient but also lacks quantitative control over the intensity and duration of rehabilitation training. Therefore, rehabilitation exoskeleton robots are of great significance in promoting neural remodeling, improving the effectiveness of rehabilitation training, and enhancing patients' quality of life. However, most rehabilitation exoskeleton robots use rigid actuators, which have shortcomings in terms of motion safety and human-machine interaction comfort. Therefore, researching intelligent control methods for series elastic actuators with characteristics such as safety, compliance, and force sensing to improve the motion control accuracy of rehabilitation exoskeleton robots has become a research hotspot in the field of intelligent medical rehabilitation.

[0003] There are various existing intelligent control methods for series elastic actuators, mainly including robust control, adaptive control, fuzzy control, and neural network control. Among them, the superiority of adaptive control methods in adjusting the control strategy in real time based on feedback information has attracted many scholars to conduct related research. The design of traditional adaptive control methods mainly follows the deterministic equivalence principle and uses the Lyapunov synthesis method to solve the adaptive law of unknown parameters. However, there is a lack of general or systematic methods for selecting Lyapunov functions, which can be very difficult, especially when dealing with complex nonlinear systems. In addition, traditional adaptive control methods guarantee asymptotic or exponential stability, which means that the system needs an infinite amount of time to reach a stable state, failing to meet the requirements of fast system response.

[0004] To address the fast response problem, finite-time stability was introduced to shorten the system convergence time. However, a major drawback of this method is that the convergence time depends on the initial state. In practical engineering, many system parameters are unknown, making it difficult to obtain initial values. Subsequently, fixed-time stability was developed to overcome the initial state dependency problem; however, the upper bound of the convergence time is related to the system parameters, making it easy to misestimate the upper bound and leading to inaccurate system performance descriptions. Therefore, it is necessary to study an adaptive control method with pre-set convergence time, allowing the system convergence time to be preset and reducing the dependence on the complexity of Lyapunov function construction. Summary of the Invention

[0005] To overcome the limited convergence performance of traditional adaptive control and further address the shortcomings of adaptive laws relying on Lyapunov synthesis methods, this invention proposes an adaptive series elastic actuator control method with preset time convergence.

[0006] The technical solution of this invention is:

[0007] According to a first aspect of the present invention, a pre-time convergent adaptive series elastic actuator control method is provided, comprising:

[0008] Define a series elastic actuator dynamics model; rewrite the series elastic actuator dynamics model to establish a series elastic actuator dynamics model containing lumped unknown system dynamics; based on the series elastic actuator dynamics model containing lumped unknown system dynamics, select state variables and establish... The state-space expression of the series elastic actuator in the time domain; for The time-domain serial elastic actuator state-space expression is subjected to a time-scale transformation and combined with... The system state in the time domain is obtained. The state-space expression in the time domain;

[0009] Based on the introduction of auxiliary functions, define The relationship between unknown parameters and their corresponding estimates in the state-space expression of the system in the time domain; the difference between the estimated and true values ​​of the unknown parameters is defined as an invariant manifold; based on the invariant manifold, the external equations are obtained; the derivative of the external equations is used to obtain the dynamic equation of the estimation error; Substituting the state-space expression in the time domain, the defined adaptive law, and the auxiliary control law into the dynamic equation of the estimation error, we obtain the expansion of the dynamic equation of the estimation error; based on the expansion of the dynamic equation of the estimation error, we construct an adaptive law based on the time scale with a correction term.

[0010] in accordance with Time-domain reference trajectory, definition Time-domain tracking error; Time-domain tracking error about Solve for the first and second derivatives; based on Time-domain tracking error, design sliding surface; calculate the first derivative of the sliding surface, and combine... The first and second derivatives of the time-domain tracking error are used to obtain the first derivative expression for the sliding surface; based on the first derivative expression for the sliding surface, a construction is performed. Auxiliary control law in the time domain; based on the constructed time-scale-based adaptive law and auxiliary control law, the series elastic actuator is controlled within a preset time. Tracking of reference trajectories in the time domain.

[0011] Furthermore, the aforementioned The establishment of the state-space expression in the time domain is as follows:

[0012] Based on the preset time adjustment function Define a new time scale satisfy:

[0013] ;

[0014] Define a new time scale Gain function under :

[0015] ;

[0016] in, express The inverse function; express First derivative;

[0017] Based on the new time scale The gain function under the system state Transform into System state in the time domain To satisfy:

[0018] ;

[0019] in, , express The first and second derivatives; express The first derivative; , express The first and second derivatives;

[0020] To ensure the system status is within the preset time Internal convergence, the preset time adjustment function satisfies Given the conditions, the preset time adjustment function is selected as follows:

[0021] ;

[0022] Adjust the selected preset time function with respect to time. Differentiating, we get:

[0023] ;

[0024] in, Represents the natural constant;

[0025] right The time-domain serial elastic actuator state-space expression is subjected to a time-scale transformation and combined with... The system state in the time domain is obtained. The state-space expression in the time domain is:

[0026] ;

[0027] in, , express Time domain , ; express Time domain , , , , Indicates output force. express The first derivative; ; ; and These represent the damping coefficients of the motor and the ball screw, respectively. This indicates the total mass of the motor and the ball screw; Represents the elastic coefficient; Indicates the conversion factor; express Time domain , Indicates the motor input current; express Time domain , It represents the movement of a lumped unknown system.

[0028] Furthermore, at the preset time Define a preset time adjustment function. The following conditions must be met:

[0029] For any time , and when ; express The left limit;

[0030] For any time , Second-order continuous differentiable;

[0031] For any time , , and when ;in, express The first derivative.

[0032] Furthermore, the time-scale-based adaptive law , The expression is:

[0033] ;

[0034] in, Representing a new time scale The gain function under; , express Time domain , ; , , Indicates output force. express The first derivative; , Represents an auxiliary function; express Auxiliary control law in the time domain; express Time domain , ; and express and Estimated value; ; ; and These represent the damping coefficients of the motor and the ball screw, respectively. This indicates the total mass of the motor and the ball screw; Represents the elastic coefficient; Indicates the conversion factor; , Indicates the correction parameter; , This represents the adaptive gain parameter.

[0035] Furthermore, the aforementioned Auxiliary control law in the time domain :

[0036] ;

[0037] in, Representing a new time scale Gain function under The first derivative; express Time domain , , Indicates output force The first derivative; ; ; express Time domain , express Reference trajectory in the time domain; , express The first and second derivatives; Indicates the sliding surface; , ; express Estimated value; , and These represent the damping coefficients of the motor and the ball screw, respectively. This indicates the total mass of the motor and the ball screw; Represents the elastic coefficient; Represents an auxiliary function; express Time domain , It represents the movement of a lumped unknown system.

[0038] According to a second aspect of the present invention, a preset time-converged adaptive series elastic actuator control system is provided, comprising a module of the method described in any one of the above.

[0039] According to a third aspect of the invention, a processor is provided for running a program, wherein the program, when running, performs the steps of the method as described in any one of the foregoing descriptions.

[0040] The beneficial effects of this invention are as follows: This invention uses a preset time gain function to transform the state-space expression of the series elastic actuator into a time-scale transformation, converting the preset time stability problem in the original time domain into an asymptotic stability problem in the new time domain, thereby reducing the complexity of the preset time controller design. Furthermore, it employs an adaptive law for unknown parameters based on immersion and invariance theory. Compared with traditional adaptive controllers based on deterministic equivalence principles, the adaptive law introduces a nonlinear function, changing the parameter update rule from a single integral form to a proportional-integral form, thus enhancing the correction capability of parameter estimation errors. Simultaneously, the construction method based on immersion and invariance theory alleviates the complexity of Lyapunov function construction in traditional adaptive control and ensures that the unknown parameters of the system converge to near their true values ​​within a preset time. At the same time, a preset time sliding surface for tracking error is designed to accelerate the convergence process of the system state to the reference trajectory, thereby ensuring that the output of the series elastic actuator achieves fast and high-precision trajectory tracking within a preset time. Attached Figure Description

[0041] Figure 1 This is a simplified diagram of a series flexible actuator (SEA).

[0042] Figure 2 This is a block diagram of the preset time adaptive controller design of the present invention.

[0043] Figure 3 Different initial states Tracking trajectory diagram.

[0044] Figure 4 Is the initial state taken Real-time tracking trajectory diagram.

[0045] Figure 5 Is the initial state taken Time tracking error diagram.

[0046] Figure 6 Is the initial state taken Time-controlled input diagram.

[0047] Figure 7 Is the initial state taken Time parameters Estimated curves and estimation error graphs; among which, Figure 7 (a) Corresponding estimated curve; Figure 7 (b) Corresponding estimation error.

[0048] Figure 8 Is the initial state taken Time parameters Estimated curves and estimation error graphs; among which, Figure 8 (a) Corresponding estimated curve; Figure 8 (b) Corresponding estimation error.

[0049] Figure 9 Is the initial state taken Time parameters Estimated curves and estimation error graphs; among which, Figure 9 (a) Corresponding estimated curve; Figure 9 (b) Corresponding estimation error. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other.

[0051] Example 1: As Figures 1-9 As shown, according to a first aspect of the present invention, a pre-time convergent adaptive series elastic actuator control method is provided, comprising:

[0052] 1. Modeling and time-scale variation of series elastic actuator system.

[0053] A simplified diagram of a series flexible actuator (SEA) is shown below. Figure 1 As shown. The driver mainly includes a servo motor, a ball screw, and a spring disposed between the nut of the ball screw and the external load; during operation, the servo motor drives the ball screw through transmission, converting the motor output torque into linear thrust, which compresses the spring, and the spring force is then transmitted to the load side of the SEA.

[0054] According to Newton's second law, the dynamic model of a series elastic actuator is defined as follows:

[0055] (1)

[0056] in, and These represent the masses of the motor and the ball screw, respectively. and These represent the damping coefficients of the motor and the ball screw, respectively. This indicates the displacement of the rigid connection between the motor and the ball screw. , They represent The first derivative and the second derivative, Indicates output force. , Represents the elastic coefficient. This indicates the input force generated by the motor.

[0057] The input force generated by the motor With motor input current They satisfy a linear relationship:

[0058] (2)

[0059] in, Indicates the conversion factor. This represents the torque coefficient of the servo motor. This indicates the transmission efficiency of the ball screw. This indicates the pitch of the ball screw.

[0060] Based on the consideration of the dynamics of the lumped unknown system, the dynamic model of the series elastic actuator is rewritten to establish a dynamic model of the series elastic actuator containing the dynamics of the lumped unknown system, with the following expression:

[0061] (3)

[0062] in, This indicates the total mass of the motor and the ball screw; It represents the lumped unknown system dynamics, including unmodeled dynamics (such as friction) and external disturbances.

[0063] To facilitate subsequent controller design, based on the dynamic model of a series elastic actuator containing lumped unknown system dynamics, the state variables are selected. , ,Establish The state-space expression for the series elastic actuator in the time domain is as follows:

[0064] (4)

[0065] In the formula: , , ; This indicates a dimension of 2×1.

[0066] At the preset time Define a preset time adjustment function. The following conditions must be met:

[0067] (1) For any time , and when ; express The left limit;

[0068] (2) For any time , Second-order continuous differentiable;

[0069] (3) For any time , , and when ;in, express The first derivative.

[0070] Based on the preset time adjustment function Define a new time scale satisfy:

[0071] (5)

[0072] Define a new time scale Gain function under :

[0073] (6)

[0074] in, express The inverse function of .

[0075] Based on the new time scale The gain function under the system state Transform into System state in the time domain To satisfy:

[0076] (7)

[0077] in, , , It is a function of time, that is , ; , express The first and second derivatives; express The first derivative; , express The first and second derivatives.

[0078] To ensure the system status is within the preset time Internal convergence, the preset time adjustment function satisfies the condition of equation (5), and the preset time adjustment function is selected as:

[0079] (8)

[0080] The preset time adjustment function of the selected equation (8) is adjusted with respect to time. Taking the derivative, we get:

[0081] (9)

[0082] in, Represents the natural constant.

[0083] Apply a time-scale transformation to the state-space expression of the series elastic actuator in equation (4), and combine it with equation (7). The system state in the time domain is obtained. The state-space expression in the time domain is:

[0084] (10)

[0085] in, , , express Time domain , ; , It is a function of time, that is , ; express Time domain , express Time domain , express Time domain ; This indicates transpose.

[0086] 2. Time-scale-based adaptive law design

[0087] Based on the introduction of auxiliary functions, define Unknown parameters in the state-space expression of the system in the time domain and The relationship between the corresponding estimated value and the formula:

[0088] (11)

[0089] in, and express and Estimated value and This represents an auxiliary function based on system state variables. Right now (This will be designed later; for ease of writing, please refer to the original text.) , Unlike adaptive control based on the deterministic equivalence principle, the unknown parameters in this invention... and use , The form is used for estimation, which is introduced with respect to the state. The relevant auxiliary functions enhance the adaptive law from a single integral action to a proportional-integral action, achieving a faster convergence speed.

[0090] The difference between the estimated and true values ​​of the unknown parameters is defined as an invariant manifold. :

[0091] (12)

[0092] Based on the invariant manifold, the external equations of the manifold are obtained as follows:

[0093] (13)

[0094] in, , They represent and Estimation error.

[0095] Differentiating the external equation of the manifold in equation (13), we obtain the dynamic equation for the estimation error as follows:

[0096] (14)

[0097] in, , They represent , The first derivative, , They represent , The first derivative; , They represent , The first derivative.

[0098] Define the adaptive law and the auxiliary control law as follows:

[0099] (15)

[0100] in, They represent Unknown parameters in the time domain and The adaptive law; express Motor input current in the time domain (i.e., preset time tracking controller) express Auxiliary control law in the time domain.

[0101] Substituting equations (10) and (15) into the dynamic equation of the estimation error in equation (14), we obtain the expansion of the dynamic equation of the estimation error:

[0102] (16)

[0103] Considering that a preset time gain further amplifies the impact of nonparametric uncertainties such as unmodeled disturbances on parameter estimation, traditional immersion and invariant adaptive control are more prone to parameter drift and may lead to estimate divergence. To enhance the robustness of the closed-loop system, based on the expansion of the dynamic equation of the estimation error, a method is introduced... Based on the modification terms, construct a structure with Time-scale-based adaptive law of the correction term , :

[0104] (17)

[0105] in, , Indicates the adaptive gain parameter; , express Correction parameters (in embodiments of the present invention) , ).

[0106] Furthermore, based on the adaptive law design of this invention, the following theorem is given:

[0107] Theorem 1: Consider a theorem with unknown parameters and and of The state-space expression in the time domain (10) is used to adapt to changes in the time scale, and the parameter estimation error is preset over a certain time. Internal consistency eventually leads to bounded stability.

[0108] Proof: Choose the Lyapunov function

[0109] (18)

[0110] Differentiating equation (18) yields:

[0111] (19)

[0112] Definition of combination (9) :

[0113] (20)

[0114] satisfy ,when , .

[0115] Therefore, the expansion of the dynamic equation for estimating the error in equation (16) can be rewritten as:

[0116] (twenty one)

[0117] make Combining equations (21) and (19), we can obtain:

[0118] (twenty two)

[0119] Using Young's inequality, the cross terms in equation (22) are scaled:

[0120] (twenty three)

[0121] (twenty four)

[0122] (25)

[0123] in, Right now , This indicates transpose.

[0124] Substituting equations (23), (24), and (25) into equation (22), we get:

[0125] (26)

[0126] because ,and ,so:

[0127] (27)

[0128] in, express The upper boundary.

[0129] exist , so that:

[0130] (28)

[0131] therefore:

[0132] (29)

[0133] in, .

[0134] For any :

[0135] (30)

[0136] in, Represents Lyapunov functions initial value, express The upper boundary.

[0137] For any ,when ,get:

[0138] (31)

[0139] Residual set .

[0140] Therefore, parameter estimation error , Regarding the time domain It is consistent and ultimately bounded. Therefore, the original time domain The parameter estimation error is within a finite time. Enter and remain bounded Inside.

[0141] 3. Auxiliary control law design

[0142] in accordance with Time Domain Reference Trajectory ,definition Time domain tracking error :

[0143] (32)

[0144] in, express Time domain , express Reference trajectory in the time domain.

[0145] right Time-domain tracking error about Solve for the first and second derivatives, and combine them with equation (7):

[0146] (33)

[0147] To achieve convergence of tracking error within a preset time, based on Time-domain tracking error, design of sliding surface :

[0148] (34)

[0149] Differentiating equation (34) and combining it with equation (33), we get:

[0150] (35)

[0151] make ,Right now Therefore, auxiliary control law :

[0152] (36)

[0153] In the formula, , express First and second derivatives, .

[0154] For equation (17) about Partial derivative integral:

[0155] (37)

[0156] To simplify controller design, Combining equation (36), we get expression:

[0157] (38)

[0158] Furthermore, based on the auxiliary control law designed in this invention, the following theorem is given:

[0159] Theorem 2: Consider a theorem with unknown parameters and The total number of people involved is unknown. of The state-space expression in the time domain (10) is used to select the adaptive law (17) and the control law (36) by changing the time scale. The parameter estimation error and the tracking error are preset over a time. Internal consistency eventually leads to bounded stability.

[0160] Proof: Choosing the Lyapunov function:

[0161] (39)

[0162] Differentiating equation (39) and combining it with equations (35) and (26), we get:

[0163] (40)

[0164] Using Young's inequalities and ,have:

[0165] (41)

[0166] therefore

[0167] (42)

[0168] in .

[0169] Therefore, positive constants exist. , so that:

[0170] (43)

[0171] (44)

[0172] in, Represents Lyapunov functions The initial value; Therefore, the sliding surface and estimation error , Regarding the time of change It is consistent and ultimately bounded.

[0173] By changing the time scale, the time interval Mapped to , Time-domain consistent final boundedness equivalence transformation to time domain From (44), we can obtain:

[0174] (45)

[0175] Therefore, sliding mode variables With estimation error , It remains bounded throughout the preset time and at a certain finite moment. It then enters and remains within a residual set.

[0176] As can be seen from the above technical solution, on the one hand, the present invention selects a gain function under a new time scale to change the time scale of the state-space expression of the series elastic actuator, thus changing the time... Mapping to the new time domain (due to changes in time scale) Determined by its nature, Corresponding to the initial moment ,because ,so Approaching , Unbounded growth, and If it is continuous and strictly increasing, then its range is exactly , time Mapping to the new time domain This invention achieves the transformation from preset time stability to asymptotic time stability. On the other hand, based on immersion and invariance theory, an estimation form for the unknown parameters is designed, and the estimation error dynamics are defined, thereby obtaining a time-scale-based adaptive law for the unknown parameters, ensuring that the system's unknown parameters converge to the vicinity of the true value within a preset time. Subsequently, a preset time sliding surface for the tracking error is designed to ensure accurate trajectory tracking of the system output within the preset time. Finally, the stability of the preset time controller and estimator is rigorously proven based on Lyapunov theory. Based on this, the invention solves the problem of fast and high-precision trajectory tracking control of a series elastic actuator under conditions of unknown model parameters and external disturbances.

[0177] Figure 2 This paper presents the overall implementation framework of the adaptive series elastic actuator control method with preset time convergence proposed in this invention. The original time-domain state-space expression (4) of the series elastic actuator is mapped to a new time-domain state-space expression (10) through time-scale transformation, and then an adaptive law based on the time-scale is constructed based on immersion and invariance theory. Parameter estimation information generated by the adaptive law is presented. and Feedback is sent to the tracking controller to compensate for the impact of system parameter uncertainties on control performance, thereby ensuring that the closed-loop system can still accurately track the reference trajectory for a preset time when external disturbances exist.

[0178] To verify the effectiveness of the pre-converged adaptive series elastic actuator control method proposed in this invention, simulation verification was performed. The simulation was conducted within... Reference trajectory in the time domain Centralized unknown system dynamics Estimate the initial values ​​of the parameters ( express The first and second elements in the text; Right now The value at time; Right now The value of time The simulation time is... Assuming the truth value of the unknown parameter preset time Other control parameter settings are shown in Table 1.

[0179] Table 1

[0180]

[0181] Simulation results are as follows Figures 3 to 9 As shown. Among them. Figure 3 Comparing state variables under different initial values The tracking response results show that, despite different initial values ​​of the system state variables (i.e., different initial states), all response curves converge to the reference trajectory within a preset time. The proximity indicates that the convergence time of the proposed method is not affected by the initial state, thus verifying the convergence performance of the controller at the preset time.

[0182] Figures 4 to 6 Give the initial values ​​of the state variables respectively Lower state variables Tracking trajectory response, tracking error response and control input response Figure 5 middle , express Tracking error in the time domain). State variables. Approach and track the reference trajectory within a preset time. The tracking error converges rapidly to near zero, indicating that the proposed control algorithm has good tracking control performance at the preset time. Simultaneously, the control input remains bounded and changes smoothly near the preset time switching point with small abrupt changes, demonstrating that this control strategy can reduce control discontinuities near the preset time point while ensuring rapid convergence, thereby improving the stability of system operation.

[0183] Figures 7 to 9 Give initial values ​​respectively The following parameters Estimate the response process. Parameters The estimated value can converge to near the true value within a preset time, and the corresponding estimation error The parameters converge to the zero neighborhood. Although there are transient fluctuations in the parameter response during the initial adjustment phase, they all quickly stabilize, indicating that the proposed method can effectively achieve rapid identification of unknown parameters and verifying that the designed time-scale-based adaptive law has good convergence performance within the preset time.

[0184] According to a second aspect of the present invention, a pre-time convergent adaptive series elastic actuator control system is provided, comprising modules of the method described in any one of the above embodiments. Specifically, it includes: a first module, configured to define a series elastic actuator dynamic model; rewrite the series elastic actuator dynamic model to establish a series elastic actuator dynamic model containing lumped unknown system dynamics; and, based on the series elastic actuator dynamic model containing lumped unknown system dynamics, select state variables and establish... The state-space expression of the series elastic actuator in the time domain; for The time-domain serial elastic actuator state-space expression is subjected to a time-scale transformation and combined with... The system state in the time domain is obtained. The state-space expression in the time domain; the second module, used to define, based on the introduction of auxiliary functions. The relationship between unknown parameters and their corresponding estimates in the state-space expression of the system in the time domain; the difference between the estimated and true values ​​of the unknown parameters is defined as an invariant manifold; based on the invariant manifold, the external equations are obtained; the derivative of the external equations is used to obtain the dynamic equation of the estimation error; Substituting the state-space expression in the time domain, the defined adaptive law, and the auxiliary control law into the dynamic equation of the estimation error, we obtain the expansion of the dynamic equation of the estimation error. Based on the expansion of the dynamic equation of the estimation error, we construct an adaptive law based on the time scale with a correction term. The third module is used to... Time-domain reference trajectory, definition Time-domain tracking error; Time-domain tracking error about Solve for the first and second derivatives; based on Time-domain tracking error, design sliding surface; calculate the first derivative of the sliding surface, and combine... The first and second derivatives of the time-domain tracking error are used to obtain the first derivative expression for the sliding surface; based on the first derivative expression for the sliding surface, a construction is performed. Auxiliary control law in the time domain. For parts of the modules not described in detail above, please refer to the relevant descriptions in this embodiment.

[0185] According to a third aspect of the present invention, a processor is provided for running a program, wherein the program, when running, performs the steps of the method as described in any one of the foregoing descriptions.

[0186] The specific embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A pre-time convergence adaptive series elastic actuator control method, characterized in that, include: Define a series elastic actuator dynamics model; The dynamic model of the series elastic actuator is rewritten to establish a dynamic model of the series elastic actuator containing the dynamics of the lumped unknown system. Based on the dynamic model of a series elastic actuator containing lumped unknowns in the system dynamics, state variables are selected, and a model is established. The state-space expression of the series elastic actuator in the time domain; for The time-domain serial elastic actuator state-space expression is subjected to a time-scale transformation and combined with... The system state in the time domain is obtained. The state-space expression in the time domain; Based on the introduction of auxiliary functions, define The relationship between unknown parameters and their corresponding estimates in the state-space expression of the system in the time domain; the difference between the estimated and true values ​​of the unknown parameters is defined as an invariant manifold; based on the invariant manifold, the external equations are obtained; the derivative of the external equations is used to obtain the dynamic equation of the estimation error; Substituting the state-space expression in the time domain, the defined adaptive law, and the auxiliary control law into the dynamic equation of the estimation error, we obtain the expansion of the dynamic equation of the estimation error; based on the expansion of the dynamic equation of the estimation error, we construct an adaptive law based on the time scale with a correction term. in accordance with Time-domain reference trajectory, definition Time-domain tracking error; Time-domain tracking error about Solve for the first and second derivatives; based on Time-domain tracking error, design sliding surface; calculate the first derivative of the sliding surface, and combine... The first and second derivatives of the time-domain tracking error are used to obtain the first derivative expression for the sliding surface; based on the first derivative expression for the sliding surface, a construction is performed. Auxiliary control law in the time domain.

2. The adaptive series elastic actuator control method with preset time convergence according to claim 1, characterized in that, The The establishment of the state-space expression in the time domain is as follows: Based on the preset time adjustment function Define a new time scale satisfy: ； Define a new time scale Gain function under : ； in, express The inverse function; express First derivative; Based on the new time scale The gain function under the system state Transform into System state in the time domain To satisfy: ； in, , express The first and second derivatives; express The first derivative; , express The first and second derivatives; To ensure the system status is within the preset time Internal convergence, the preset time adjustment function satisfies Given the conditions, the preset time adjustment function is selected as follows: ； Adjust the selected preset time function with respect to time. Differentiating, we get: ； in, Represents the natural constant; right The time-domain serial elastic actuator state-space expression is subjected to a time-scale transformation and combined with... The system state in the time domain is obtained. The state-space expression in the time domain is: ； in, , express Time domain , ; express Time domain , , , , Indicates output force. express The first derivative; ; ; and These represent the damping coefficients of the motor and the ball screw, respectively. This indicates the total mass of the motor and the ball screw; Represents the elastic coefficient; Indicates the conversion factor; express Time domain , Indicates the motor input current; express Time domain , It represents the movement of a lumped unknown system.

3. The adaptive series elastic actuator control method with preset time convergence according to claim 2, characterized in that, At the preset time Define a preset time adjustment function. The following conditions must be met: For any time , and when ; express The left limit; For any time , Second-order continuous differentiable; For any time , , and when ;in, express The first derivative.

4. The adaptive series elastic actuator control method with preset time convergence according to claim 1, characterized in that, The time-scale-based adaptive law , The expression is: ； in, Representing a new time scale The gain function under; , express Time domain , ; , , Indicates output force. express The first derivative; , Represents an auxiliary function; express Auxiliary control law in the time domain; express Time domain , ; and express and Estimated value; ; ; and These represent the damping coefficients of the motor and the ball screw, respectively. This indicates the total mass of the motor and the ball screw; Represents the elastic coefficient; Indicates the conversion factor; , Indicates the correction parameter; , This represents the adaptive gain parameter.

5. The adaptive series elastic actuator control method with preset time convergence according to claim 1, characterized in that, The Auxiliary control law in the time domain : ； in, Representing a new time scale Gain function under The first derivative; express Time domain , , Indicates output force The first derivative; ; ; express Time domain , express Reference trajectory in the time domain; , express The first and second derivatives; Indicates the sliding surface; , ; express Estimated value; , and These represent the damping coefficients of the motor and the ball screw, respectively. This indicates the total mass of the motor and the ball screw; Represents the elastic coefficient; Represents an auxiliary function; express Time domain , It represents the movement of a lumped unknown system.

6. A pre-time convergence adaptive series elastic actuator control system, characterized in that, The module includes the method described in any one of claims 1-5.

7. A processor, characterized in that, The processor is used to run a program, wherein the program, when running, performs the steps of the method as described in any one of claims 1-5.

Images