A method for high dynamic torque control of a series elastic actuator joint system

Abstract

This invention discloses a high-dynamic torque control method for a series elastic actuator joint system, comprising: establishing a mechatronic coupling model to clarify the dynamic transmission relationship between the electromagnetic torque of the motor and the output torque of the torsion spring, analyzing the voltage constraint mechanism and the field weakening release principle; designing a control framework containing a linearly extended state observer to estimate the output torque of the torsion spring, the rate of torque change, and the lumped disturbance in real time, and preprocessing the torque demand to avoid high-frequency oscillations; dividing the operating conditions into four categories, and combining linearized current, voltage constraints, and torque demands to solve for the optimal d-axis and q-axis current increments. The control method of this invention can smoothly switch the field weakening trajectory, achieve accurate disturbance compensation and high-dynamic torque response under voltage constraints, and significantly improve torque tracking accuracy under complex operating conditions.

Description

Technical Field

[0001] This invention belongs to the field of robot system motor control technology, specifically relating to a high dynamic torque control method for a series elastic actuator joint system. Background Technology

[0002] Series Elastic Actuators (SEAs) significantly improve shock absorption, human-machine interaction compliance, and torque transmission fidelity by introducing elastic torsion springs between the motor and the load. They have become the core execution unit in physical human-machine interaction (PHRI) scenarios such as rehabilitation robots, industrial assembly robots, and mechanical exoskeletons. Their torque control performance directly determines the robot's operating accuracy, interaction safety, and dynamic response speed. Especially under conditions of rapid load changes and complex disturbances, the requirements for dynamic torque response, tracking accuracy, and robustness are extremely high.

[0003] In the field of disturbance compensation, the Disturbance Observer (DOB), Noise Reducing Disturbance Observer (NRDOB), and Linear Extended State Observer (LESO) are the mainstream solutions. Among them, LESO is widely used because of its simple structure and easy parameter tuning. However, when dealing with the coupled disturbances of time-varying load-side dynamics and motor electrical parameters, it is difficult to balance estimation accuracy and dynamic response, which leads to an increase in torque tracking error.

[0004] More critically, voltage saturation during high-dynamic operation of the SEA (Self-Electronic Actuator) restricts torque response. As the core drive unit, the stator voltage of the permanent magnet synchronous motor (PMSM) is limited by the linear modulation region of SVPWM (Space Vector Pulse Width Modulation). Increased speed leads to increased back electromotive force, preventing the torque current from increasing and limiting electromagnetic torque, thus reducing the rate of change of the torsion spring output torque. Existing field weakening control methods, such as those described in the literature [Roozing W, Kashiri N, Tsagarakis N G. Enhancing the Explosive Motion Capability of Torque-Controlled Drivers through Magnetic Field Weakening Control, RSJ International Conference on Intelligent Robots and Systems. IEEE, 2018], while releasing voltage margin by injecting negative d-axis current, have significant drawbacks: First, a direct control correlation between the q-axis and d-axis currents is not established, the d-axis current cannot be adaptively adjusted, and the current loop decoupling effect degrades when the voltage saturates; second, the field weakening trajectory does not match torque requirements with current and voltage constraints, easily causing torque fluctuations during switching and insufficient bus voltage utilization.

[0005] Furthermore, the low-pass filtering characteristics of the SEA elastic torsion spring limit the rate of torque change, and the time-varying loads and external disturbances in human-machine interaction scenarios make it difficult to be equivalent to an ideal dual-inertia model. Existing control strategies, such as those in the literature [Zhang Tao, Du Qiang, Yang Gang. A review of compliant control research for collaborative robots, IEEE Transactions on Robotics, Vol. 37, No. 4, pp. 1234-1250, 2021], do not fully couple the dynamic characteristics of the motor electrical system and the joint mechanical system, and cannot achieve optimal balance among multiple objectives of "torque response-disturbance compensation-voltage constraint", making it difficult to meet the control requirements of high dynamics, high precision, and strong robustness. Therefore, developing a high-dynamic torque control strategy that adapts to voltage states and torque requirements, accurately compensates for disturbances, and achieves optimal current distribution has become an urgent problem to be solved in this field. Summary of the Invention

[0006] In view of the above, the present invention provides a high dynamic torque control method for a series elastic actuator joint system. By designing a control framework containing LESO, combining MTPA and field weakening control, different operating conditions are divided, and the optimal d-axis and q-axis current increments are solved to achieve accurate disturbance compensation and high dynamic torque response under voltage constraints.

[0007] A method for high dynamic torque control of a series elastic actuator joint system includes the following steps:

[0008] (1) Real-time acquisition of synchronous speed, stator current, electromagnetic torque and torsion spring output torque of motor in SEA joint;

[0009] (2) Construct LESO to estimate the output torque of the SEA joint torsion spring and the total system disturbance in real time;

[0010] (3) Calculate the torque increment and electromagnetic torque reference value at the next moment based on the observed values ​​output by LESO;

[0011] (4) Based on the different working conditions of voltage state and torque demand change, use the geometric relationship of the current increment plane to solve the optimal current increment at the next moment, and add it to the stator current vector at the current moment to obtain the expected current at the next moment.

[0012] (5) The stator reference voltage at the next moment is obtained by tracking the desired current using the deadbeat algorithm of the current loop, and then the SVPWM technology is used to generate a PWM signal to drive the power switching device in the motor inverter at the next moment.

[0013] Furthermore, the discrete expression for LESO in step (2) is as follows:

[0014]

[0015] in: For the first k The observation error at time, For the first k time The actual value, For the first k time The actual value, and The first k Time and the k +1 moment The observed values, For the torsion spring to output torque, For the electromagnetic torque of the motor, To control the cycle, and The first k Time and the k +1 moment The observed values, for The first derivative, h 1~ h 3 is the observer gain and , , , For the observer bandwidth, and The first k Time and the k +1 moment d The observed values, d For system lumped disturbances, To control the gain and , For the moment of inertia of the motor, N The reduction ratio of the harmonic reducer. This is the stiffness coefficient of the torsion spring. k It is a natural number.

[0016] Furthermore, the system lumped disturbance d The expression is as follows:

[0017]

[0018] in: This is the motor damping coefficient. This refers to the internal mechanical friction torque of the motor. For the load-side angular velocity, for The first derivative.

[0019] Furthermore, in step (3), the torque increment and electromagnetic torque reference value at the next moment are calculated using the following expression:

[0020]

[0021]

[0022] in: and The first k The smoothed expected torque and its increment at time +1 For the first k The torque increment at time +1 To control bandwidth, For the first k Reference value of electromagnetic torque of motor at +1 moment.

[0023] Furthermore, in step (4), for operating conditions where the voltage is not limited and the torque demand changes little, i.e. and The optimal current increment at the next moment can be solved using the following expression:

[0024]

[0025]

[0026] in: For the first k The vector of the optimal current increment at time +1 This represents the maximum value of the motor stator current. For the first k Stator reference voltage amplitude at time [time]. This is the DC voltage of the motor inverter. For the first k The distance from the linear increase in torque at any given moment to the current operating point. This is the motor torque coefficient. Let be a unit vector perpendicular to the line of moment and , For the first k The vector from the origin of the current increment plane to the center of the current constraint circle at any given time. , and The first k The d-axis stator current and q-axis stator current of the motor at any given time, superscript T This indicates transpose.

[0027] Furthermore, in step (4), for operating conditions where the voltage is not limited and the torque demand varies greatly, i.e. and The optimal current increment at the next moment can be solved using the following expression:

[0028]

[0029]

[0030] in: For the first k The vector of the optimal current increment at time +1 This represents the maximum value of the motor stator current. For the first k Stator reference voltage amplitude at time [time]. This is the DC voltage of the motor inverter. For the first k The distance from the linear increase in torque at any given moment to the current operating point. This is the motor torque coefficient. Let be a unit vector perpendicular to the line of moment and , For the first k The vector from the origin of the current increment plane to the center of the current constraint circle at any given time. , and The first k The d-axis stator current and q-axis stator current of the motor at any given time. Symbolic function, superscript T This indicates transpose.

[0031] Furthermore, in step (4), for operating conditions where voltage is limited and torque demand varies significantly, i.e. and The optimal current increment at the next moment can be solved using the following expression:

[0032]

[0033]

[0034]

[0035]

[0036]

[0037] in: For the first k The vector of the optimal current increment at time +1 This represents the maximum value of the motor stator current. For the first k Stator reference voltage amplitude at time [time]. This is the DC voltage of the motor inverter. For the first kThe distance from the linear increase in torque at any given moment to the current operating point. D U ( k ) is the first k The distance from the voltage constraint line to the current current operating point at any given moment. d U ( k +1) is the first k Half the length of the chord formed by the intersection of the voltage constraint line and the current constraint circle at time +1. v and u These are unit vectors parallel and perpendicular to the voltage constraint line, respectively. This is the motor torque coefficient. For the stator resistance of the motor, For the first k The vector from the origin of the current increment plane to the center of the current constraint circle at any given time. , and The first k The d-axis stator current and q-axis stator current of the motor at any given time. and The first k The d-axis stator reference voltage and q-axis stator reference voltage at time t. For the first k The synchronous speed of the motor at any given time. and These are the direct-axis inductance and quadrature-axis inductance of the motor, respectively. Symbolic function, superscript T Indicates transpose. and The first k The d-axis voltage constraint linearization coefficients and q-axis voltage constraint linearization coefficients at time t. For the first k The magnitude of the voltage constraint comprehensive coefficient at time t. For the first k The maximum allowable increment of the stator voltage amplitude at time +1 and , This is a voltage change rate constraint parameter.

[0038] Furthermore, in step (4), for operating conditions where voltage is limited and torque demand changes relatively little, i.e. and The optimal current increment at the next moment can be solved using the following expression:

[0039]

[0040]

[0041]

[0042] in: For the first k The vector of the optimal current increment at time +1 This represents the maximum value of the motor stator current. For the first k Stator reference voltage amplitude at time [time]. This is the DC voltage of the motor inverter. For the first k The distance from the linear increase in torque at any given moment to the current operating point. This is the motor torque coefficient. For the stator resistance of the motor, and The first k The d-axis stator reference voltage and q-axis stator reference voltage at time t. For the first k The synchronous speed of the motor at any given time. and These are the direct-axis inductance and quadrature-axis inductance of the motor, respectively, indicated by the superscript. T Indicates transpose. and The first k The d-axis voltage constraint linearization coefficients and q-axis voltage constraint linearization coefficients at time t. For the first k The maximum allowable increment of the stator voltage amplitude at time +1 and , This is a voltage change rate constraint parameter.

[0043] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described high dynamic torque control method for a series elastic actuator joint system.

[0044] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described high dynamic torque control method for a series elastic actuator joint system.

[0045] Based on the above technical solution, the present invention has the following beneficial technical effects:

[0046] 1. This invention utilizes a dynamic field weakening strategy to inject negative d-axis current to release voltage margin when voltage is limited, thereby increasing torque current, enhancing motor electromagnetic torque, and accelerating the rate of change of torsion spring output torque.

[0047] 2. This invention estimates the total disturbance (including mechanical friction and load disturbance) in real time based on a linear extended state observer, achieving accurate compensation for unmodeled errors and external disturbances, avoiding high-frequency oscillations, and significantly improving torque tracking accuracy under complex working conditions.

[0048] 3. This invention linearizes different constraints on motor operation, divides four dynamic operating conditions to solve for the optimal current increment, smoothly switches the field weakening trajectory, avoids torque fluctuations, and makes full use of the bus voltage. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the high dynamic torque control algorithm for the SEA joint system of the present invention.

[0050] Figure 2 This is a schematic diagram of the optimal current trajectory in this invention.

[0051] Figure 3 This is a schematic diagram of the optimal current increment solution in this invention, where (a), (b), (c), and (d) correspond to four operating conditions, respectively.

[0052] Figure 4 The diagram shows the comparative experimental results of output torque and dq-axis current under the step torque demand condition. (a) corresponds to the strategy without field weakening, and (b) corresponds to the strategy of this invention. The horizontal axis t represents time.

[0053] Figure 5 This invention addresses the issue of step-increase torque demand in operating conditions. i d - i q Schematic diagram of plane current control trajectory. Detailed Implementation

[0054] To describe the present invention in more detail, the technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] This embodiment takes an SEA joint with a rated power of 0.2kW, a rated torque of 34N·m, and a bus voltage of 36V as an example. The drive motor is a small motor with a rated torque of 0.94N·m and a rated speed of 2000r / min. The transmission ratio of the harmonic reducer is 120, and the stiffness coefficient of the torsion spring is 104N·m / rad. On the load side, different weight plates are suspended to change the inertia and torque of the load, and closed-loop control is performed on the transmission torque from the SEA to the output to the load.

[0056] like Figure 1 As shown, after the system starts, the torque requirement of the system is preprocessed according to the tracking differentiator (TD), and a smooth desired torque is output. τ sref and its increment Δ τ sref Simultaneously, system operating status data is collected and noise is suppressed to provide reliable input for subsequent control decisions. Input signals include the rotor mechanical angular velocity collected by the motor encoder. ω m The stator three-phase current is converted into dq-axis current through coordinate transformation. i d and i q The transmission torque is obtained by multiplying the difference between the position sensors at both ends of the elastic element by the stiffness coefficient. τ s The collected variables are input into the designed discrete linear extended state observer, and the output is the transmission torque. τ s Rate of change of transmission torque and aggregate disturbance d The observed values ​​are then subtracted from the expected value output by the TD to obtain the deviation of the torque requirement. Along with the collected synchronous rotation speed ω e Current i d and i q The combined input is fed into the optimal current calculation strategy based on dynamic field weakening, and the output is the desired current of a given current loop. and The current loop uses a deadbeat-free algorithm to track the given desired current and outputs the desired d-axis and q-axis stator voltages. and The inverter output voltage is adjusted using SVPWM technology to drive the motor. The motor rotation is transmitted through the harmonic reducer to compress the torsion spring and output joint torque. A torque sensor is used to measure the output torque of the SEA joint to verify this.

[0057] To achieve closed-loop control of the output torque of the SEA joint, this embodiment provides a high dynamic torque control strategy for a series elastic actuator joint system, the specific process of which is as follows:

[0058] (1) Establish the torque transmission model of the SEA joint system.

[0059]

[0060] in: J m For the moment of inertia of the motor, N For the reduction ratio of the harmonic reducer, K c This is the stiffness coefficient of the torsion spring. B m This is the motor damping coefficient. τ m For the electromagnetic torque of the motor, τ s For the torsion spring to output torque, ω L For the load-side angular velocity, τ f This refers to the internal mechanical friction torque of the motor.

[0061] (2) Design the Linear Extended State Observer (LESO).

[0062] The system friction damping and disturbances (including) τ f Load disturbance torque τ L Unified as lumped disturbance d Therefore, the third-order LESO is designed in discrete form as follows:

[0063]

[0064] in: z 1. z 2. They are respectively τ s , , d The observed values, T s To control the cycle, K τ = J m N / K c , , , ( ω o The observation bandwidth is set to 200 in this embodiment; this LESO can estimate in real time. τ s , as well as d This provides accurate data support for torque demand deviation compensation.

[0065] (3) Convert the torque requirement into the desired electromagnetic torque of the motor.

[0066] Using a tracking differentiator τ sref Preprocessing is performed to output a smoothed desired torque and its increment Δ τ sref This avoids high-frequency oscillations caused by direct differentiation. τ sref ( k+1), Δ τ sref ( k +1) respectively compared with the observations of LESO z 1( k +1) z 2( k +1) Subtracting the values ​​yields the torque requirement deviation Δ. T ( k +1), based on demand deviation Δ T ( k +1) and design control bandwidth ω c Feedforward compensation for system disturbances transforms the torque requirement into a desired reference value for the motor's electromagnetic torque. τ mref ( k +1) is:

[0067]

[0068]

[0069] (4) Design the optimal weak magnetic trajectory.

[0070] like Figure 2 As shown, the optimal field weakening trajectory design can be divided into two segments: the voltage unsaturated segment—the current point moves along the maximum torque-to-current ratio (MTPA) trajectory (OA segment), with no negative direction. i d Injection; Voltage saturation segment—the current point moves along the intersection of the voltage limit circle and the current limit circle (segment ABCD), through negative injection. i d Reducing the equivalent flux linkage lowers the d-axis induced electromotive force, releasing the voltage margin to maintain i q The torque contribution. The trajectory switching condition is: the switching point is the intersection of the voltage limit circle and the MTPA trajectory, ensuring that during the switching process... i d , i q The rate of change is continuous, avoiding torque fluctuations.

[0071] (5) Construct the current increment equation for torque demand and operating constraints.

[0072] Torque demand increment: Based on the previously established relationship between torque demand and electromagnetic torque, the motor electromagnetic torque formula is linearized with the current increment as the variable, as follows: Figure 1 As shown, the demand for the current increment is represented by a horizontal straight line.

[0073] Voltage limit constraints and incremental boundaries: Voltage limit constraints To linearly modulate the maximum voltage of the inverter, linearization is performed using the current increment as a variable, such as... Figure 1 As shown, the boundary corresponding to the current increment is a sloping line.

[0074] Current limit constraints and incremental boundaries: Current limit constraints The maximum allowable current of the motor is defined by the current increment, with the boundary of the current increment being a circle centered at the current operating point. Figure 1 As shown, the center of the circle is represented as .

[0075] (6) Design the optimal selection rules for d-axis and q-axis current increments.

[0076] The specific rules for selecting the optimal current increment include: a. The voltage change must not exceed the maximum allowable voltage of the modulation, otherwise the voltage will be over-modulated, which may cause current runaway; b. The target current point must be limited within the current limit circle to prevent the winding from overheating; c. If the current is limited by voltage or current and cannot meet the joint output torque requirement in one step, then select the target current point as close as possible to the load line to gradually meet it; d. The selected current should be selected according to the maximum torque-to-current ratio (MTPA) control, provided that it can meet the joint output torque requirement.

[0077] (7) Solve for the optimal d-axis and q-axis current increments.

[0078] Operating Condition 1: Voltage is not limited and torque demand changes little.

[0079] When the motor speed and back EMF are low, all points inside the current constraint circle satisfy the voltage constraint (rules a and b); if the joint torque demand changes little, and the distance from the torque increment line to the center of the circle is less than... I max Choose its intersection point with the vertical axis according to rule d, such as... Figure 3 For point A in (a), the optimal current increment is:

[0080]

[0081]

[0082] in: Torque increment straight to i 0( k The distance, and ; t It is a unit vector perpendicular to the line of moment. ω c The control bandwidth required for joint torque control (set to 0.5 in this embodiment).

[0083] Operating Condition 2: Voltage is not limited and torque demand varies significantly.

[0084] The motor speed and back EMF are relatively low (same as in operating condition 1, satisfying rules a and b), but the joint torque demand varies greatly, and the distance from the linear torque increment to the center of the circle is greater than... I max According to rule c, select the current increment point within the current constraint that is closest to the straight line to maximize the torque requirement, such as... Figure 3 For point B in (b), the optimal current increment is:

[0085]

[0086] Where: sgn is the sign function, and .

[0087] Operating Condition 3: Voltage is limited and torque requirements vary significantly.

[0088] When the motor speed and back EMF are high, only the voltage constraint region within the current constraint circle satisfies rules a and b, and the distance from these points to the center of the circle is less than the linear distance of the torque increment, making it impossible to meet the requirements in a single step; therefore, according to rule c, the current increment point closest to the straight line within this region is selected, such as... Figure 3 For point C in (c), the optimal current increment is:

[0089]

[0090]

[0091]

[0092]

[0093]

[0094] in: D U ( k () represents the distance from the voltage constraint line to the center of the circle. d U ( k +1) represents half the length of the chord formed by the intersection of the voltage constraint line and the current constraint circle. v , u These are unit vectors parallel and perpendicular to the voltage constraint line, respectively. ; , This is used to limit its dynamic rate of change and prevent voltage overshoot (set to 1 in this embodiment). D U ( k The voltage amplitude, which symbolizes the demand, is the key to achieving dynamic adjustment of the depth of magnetic weakening.

[0095] Operating Condition 4: Voltage is limited and torque demand changes little.

[0096] The motor speed and back electromotive force are relatively high, but the torque demand changes little; the distance from the point satisfying rules a and b to the center of the circle is equal to the linear distance of the torque increment, which can directly meet the demand; therefore, according to rule d, the intersection of the two is selected, such as... Figure 3 For point D in (d), the optimal current increment can be obtained by solving:

[0097]

[0098] at this time: The motor exits the weak magnetic field operating region.

[0099] Figure 4 The experiment showcases the comparative results of two control strategies under a step torque demand condition (a 10kg weight is hung on the load side, and the output torque jumps from -10N·m to +10N·m). Each waveform, from top to bottom, displays the expected / actual output torque, q-axis current setpoint / actual value, d-axis current setpoint / actual value, and motor speed. The control strategy without a field weakening path suffers from a large torque drop due to voltage limitation, resulting in a settling time of 56ms. The strategy of this invention, through optimal current setpoint calculation, achieves effective current tracking, reducing the settling time to only 30ms (46.4% shorter than the strategy without field weakening), demonstrating superior dynamic performance.

[0100] Figure 5 The method of the present invention is demonstrated under the above working conditions. i d - i q The in-plane current control trajectory shows that, while meeting the electromagnetic torque requirement, the method of this invention fully utilizes the available current points within the current limit circle, achieving... Figure 2 The optimal weak magnetic trajectory designed in the middle.

[0101] The above description of the embodiments is provided to enable those skilled in the art to understand and apply the present invention. Those skilled in the art can readily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without creative effort. Therefore, the present invention is not limited to the above embodiments, and any improvements and modifications made to the present invention by those skilled in the art based on the disclosure thereof should be within the scope of protection of the present invention.

Claims

1. A high dynamic torque control method for a series elastic actuator joint system, characterized in that, Includes the following steps: (1) Real-time acquisition of synchronous speed, stator current, electromagnetic torque and torsion spring output torque of motor in SEA joint; (2) Construct LESO to estimate the output torque of the SEA joint torsion spring and the total system disturbance in real time; (3) Calculate the torque increment and electromagnetic torque reference value at the next moment based on the observed values ​​output by LESO; (4) Based on the different working conditions of voltage state and torque demand change, use the geometric relationship of the current increment plane to solve the optimal current increment at the next moment, and add it to the stator current vector at the current moment to obtain the expected current at the next moment. For operating conditions where voltage is limited and torque demand varies significantly, i.e. and The optimal current increment at the next moment can be solved using the following expression: in: For the first k The vector of the optimal current increment at time +1 This represents the maximum value of the motor stator current. For the first k Stator reference voltage amplitude at time [time]. This is the DC voltage of the motor inverter. For the first k The distance from the linear increase in torque at any given moment to the current operating point. D U ( k ) is the first k The distance from the voltage constraint line to the current current operating point at any given moment. d U ( k +1) is the first k Half the length of the chord formed by the intersection of the voltage constraint line and the current constraint circle at time +1. v and u These are unit vectors parallel and perpendicular to the voltage constraint line, respectively. This is the motor torque coefficient. For the stator resistance of the motor, For the first k The vector from the origin of the current increment plane to the center of the current constraint circle at any given time. , and The first k The d-axis stator current and q-axis stator current of the motor at any given time. and The first k The d-axis stator reference voltage and q-axis stator reference voltage at time t. For the first k The synchronous speed of the motor at any given time. and These are the direct-axis inductance and quadrature-axis inductance of the motor, respectively. Indicates a sign function, superscript T Indicates transpose. and The first k The d-axis voltage constraint linearization coefficients and q-axis voltage constraint linearization coefficients at time t. For the first k The magnitude of the voltage constraint comprehensive coefficient at time t. For the first k The maximum allowable increment of the stator voltage amplitude at time +1 and , The voltage change rate constraint parameter, For the first k Reference value of electromagnetic torque of motor at +1 time; (5) The stator reference voltage at the next moment is obtained by tracking the desired current using the deadbeat algorithm of the current loop, and then the SVPWM technology is used to generate a PWM signal to drive the power switching device in the motor inverter at the next moment.

2. The high dynamic torque control method for a series elastic actuator joint system according to claim 1, characterized in that, The discrete expression for LESO in step (2) is as follows: in: For the first k The observation error at time, For the first k time The actual value, For the first k time The actual value, and The first k Time and the k +1 moment The observed values, For the torsion spring to output torque, For the electromagnetic torque of the motor, To control the cycle, and The first k Time and the k +1 moment The observed values, for The first derivative, h 1~ h 3 is the observer gain and , , , For the observer bandwidth, and The first k Time and the k +1 moment d The observed values, d For system lumped disturbances, To control the gain and , The moment of inertia of the motor. N The reduction ratio of the harmonic reducer. This is the stiffness coefficient of the torsion spring. k It is a natural number.

3. The high dynamic torque control method for a series elastic actuator joint system according to claim 2, characterized in that, The system lumped disturbance d The expression is as follows: in: This is the motor damping coefficient. This refers to the internal mechanical friction torque of the motor. For the load-side angular velocity, for The first derivative.

4. The high dynamic torque control method for a series elastic actuator joint system according to claim 2, characterized in that, In step (3), the torque increment and electromagnetic torque reference value at the next moment are calculated using the following expression: in: and The first k The smoothed expected torque and its increment at time +1 For the first k The torque increment at time +1 To control bandwidth.

5. The high dynamic torque control method for a series elastic actuator joint system according to claim 1, characterized in that, In step (4), for the working condition where the voltage is not limited and the torque demand changes little, that is... and The optimal current increment at the next moment can be solved using the following expression: in: It is a unit vector perpendicular to the line of moment and .

6. The high dynamic torque control method for a series elastic actuator joint system according to claim 1, characterized in that, In step (4), for the working condition where the voltage is not limited and the torque demand varies greatly, i.e. and The optimal current increment at the next moment can be solved using the following expression: in: It is a unit vector perpendicular to the line of moment and .

7. The high dynamic torque control method for a series elastic actuator joint system according to claim 1, characterized in that, In step (4), for the working condition where the voltage is limited and the torque demand changes little, that is... and The optimal current increment at the next moment can be solved using the following expression: 。 8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: The processor is used to execute the computer program to implement the high dynamic torque control method for the series elastic actuator joint system as described in any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, it implements the high dynamic torque control method for the series elastic actuator joint system as described in any one of claims 1 to 7.

Images