Multi-mode intelligent training equipment motor control method and system based on exhaustion protection

Abstract

The invention provides a multi-mode intelligent training equipment motor control method and system based on exhaustion protection, and the method comprises the steps: building a motion current model based on motor operation data and motor control current, and the training modes at least comprise a fluid resistance mode; obtaining a motion current model of each training mode based on the training modes in combination with the motion current models; acquiring the angular position of the motor based on the time sequence, and accumulating energy above a preset frequency threshold in the angular position sequence to obtain an energy density accumulated value; when the energy density accumulated value is greater than a preset energy density threshold value, switching each training mode into a fluid resistance mode; constructing a reinforcement learning model based on the motor operation data and the motion current model of each training mode; and training the reinforcement learning model to obtain the optimal motor control current of each training mode. The exhaustion condition in the training process is precisely processed, so that the intelligent fitness equipment is safer and more reliable.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of motor control, in particular to a motor control method and system for multi-mode intelligent training equipment based on muscle failure protection. BACKGROUND

[0002] Traditional counterweight training equipment is gradually being replaced by motor-driven intelligent resistance systems due to the inconvenience of adjustment and lack of dynamic adaptability. Such systems control the motor to operate in generator mode, use reverse electromagnetic torque (reverse drag force) to provide adjustable resistance, allowing trainees to perform precise load training according to their own needs, while supporting data feedback and intelligent adjustment.

[0003] However, during high-intensity or endurance training, trainees may enter a muscle failure state due to muscle fatigue, which can cause the following problems: decreased motor control ability: muscle nerve recruitment efficiency decreases when in muscle failure, leading to motion deformation (such as shoulder joint compensation during bench press, knee joint buckling during deep squat), significantly increasing the risk of tendon / ligament strain and joint injury; sudden resistance imbalance: if the motor resistance is not dynamically adjusted, the trainee's loss of strength may cause the equipment to rebound (such as uncontrolled high pull-up horizontal bar, rapid descent of barbell), causing impact or sprain; long-term health risks: abnormal stress accumulation during repeated muscle failure training may lead to chronic injuries (such as lumbar disc wear and tear, rotator cuff tears).

[0004] Currently, motor control systems for fitness equipment generally lack muscle failure state recognition and protection mechanisms, relying only on pre-set resistance curves or simple speed feedback, which cannot respond to the physiological limit state of the trainee in real time, making intelligent fitness equipment have significant defects in safety and adaptability. SUMMARY

[0005] The present application provides a motor control method and system for multi-mode intelligent training equipment based on muscle failure protection to overcome the shortcomings of the prior art.

[0006] To solve the above technical problems, the present application solves the problem by the following technical scheme: A motor control method for multi-mode intelligent training equipment based on muscle failure protection, the intelligent training equipment provides multiple training modes, including the following steps: Obtain training mode and motor operation data; based on motor operation data and motor control current, construct a motion current model; wherein the motor operation data at least includes motor angular position, motor angular velocity, motor acceleration, motor stiffness, motor damping and motor moment of inertia, the training mode at least includes fluid resistance mode; Based on the training mode, combine the motion current model to obtain the motion current model of each training mode; Collect motor angular position data for a preset monitoring period to obtain an angular position sequence; accumulate the energy above a preset frequency threshold in the angular position sequence to obtain the cumulative energy density value; When the cumulative energy density value exceeds the preset energy density threshold, the training mode will be switched to fluid resistance mode. A reinforcement learning model is constructed based on motor operation data and motion current models for each training mode. The reinforcement learning model is trained to obtain the optimal motor control current for each training mode, thereby achieving optimal control of the motor of the intelligent training equipment.

[0007] As one possible implementation method, the moving current model is represented as follows:

[0008] in, Indicates the motor control current. Indicates motor stiffness. Indicates motor damping, Indicates the moment of inertia of the motor. Indicates the angular position of the motor. Indicates the angular velocity of the motor. Indicates the angular acceleration of the motor. These represent the set angular position, set angular velocity, and set angular acceleration, respectively. This represents the motor torque constant term. This represents the motor torque constant.

[0009] As one possible implementation method, the training mode includes at least a fluid resistance mode, a centripetal isotensile mode, an eccentric isotensile mode, and an elastic mode; In the fluid resistance mode, the motor stiffness and the motor moment of inertia in the motion current model are zero, and the motor damping is negative. In the centripetal isotensile mode, the motor stiffness, motor damping, and motor moment of inertia in the motion current model are zero, and the motor angular velocity is positive. In the centrifugal isotensile mode, the motor stiffness, motor damping, and motor moment of inertia in the motion current model are zero, and the motor angular velocity is negative. In the elastic mode, the motor damping and the motor moment of inertia in the motion current model are zero, and the motor stiffness is negative.

[0010] As one possible implementation, the step of collecting motor angular position data for a preset monitoring period to obtain an angular position sequence, and accumulating the energy above a preset frequency threshold in the angular position sequence to obtain an energy density accumulation value, includes the following steps: The motor angular position in a preset monitoring period is collected based on a time sequence to obtain an initial angular position sequence; The initial angular position sequence is subjected to low-pass filtering to obtain a pretreated angular position sequence; The pretreated angular position sequence is subjected to energy and frequency domain conversion processing to obtain an angular position spectrum sequence; The energy density of a frequency band greater than or equal to a preset frequency threshold in the angular position spectrum sequence is extracted and accumulated to obtain an energy density accumulation value.

[0011] As an implementable manner, the motor control method of the multi-mode intelligent training equipment based on exhaustive protection further comprises: when the motor angular velocity is greater than a first preset angular velocity threshold, adjusting the motion current model of each training mode according to the difference between the motor angular velocity and the first preset angular velocity threshold; the adjusted motion current model of each training mode is represented as follows:

[0012] wherein, represents the adjusted motor control current, represents a proportional gain, represents the motor angular velocity, represents the first preset angular velocity threshold, represents a motor torque constant.

[0013] As an implementable manner, the reinforcement learning model comprises a state space model, an action space model and a reward function, specifically: The state space model is constructed based on the motor angular position, the motor angular velocity, the motor angular acceleration, the motor stiffness, the motor damping and the motor moment of inertia; The action space model is constructed based on the motor control current; The force position reward value is obtained based on the motor torque predicted by the reinforcement learning model and an ideal motor torque; wherein the ideal motor torque is obtained through motor operation data; The exhaustive reward value is obtained based on the motor energy density accumulation value and a preset energy density threshold; wherein the motor energy density accumulation value is obtained through the motor angular position and a preset frequency threshold; The overspeed reward value is obtained based on whether the motor angular velocity exceeds a maximum motor angular velocity; The force position reward value, the exhaustive reward value and the overspeed reward value are subjected to weighted summation to construct the reward function.

[0014] As an implementable manner, when the energy density accumulation value is greater than the preset energy density threshold, the training mode is switched to a fluid resistance mode, specifically: the motor stiffness and the motor moment of inertia are set to zero, the motor damping is a negative value, and the absolute value of the motor damping is increased.

[0015] A multi-mode intelligent training equipment motor control system based on exhaustion protection is characterized in that it can realize the multi-mode intelligent training equipment motor control method based on exhaustion protection described in any one of the above-mentioned methods. The control system includes a basic model module, a training mode model module, an energy density module, a model adjustment module, an inference model construction module, and an inference model training module. The basic model module is used to acquire training modes and motor operation data; and to construct a motion current model based on the motor operation data and motor control current; wherein the motor operation data includes at least motor angular position, motor angular velocity, motor acceleration, motor stiffness, motor damping and motor moment of inertia, and the training modes include at least fluid resistance modes. The training mode model module is used to obtain the motion current model for each training mode based on the training mode and combined with the motion current model. The energy density module is used to collect motor angular position data for a preset monitoring period to obtain an angular position sequence; and to accumulate the energy above a preset frequency threshold in the angular position sequence to obtain an energy density cumulative value. The model adjustment module is used to switch the training mode to fluid resistance mode when the cumulative energy density value is greater than the preset energy density threshold. The inference model building module is used to build a reinforcement learning model based on motor operation data and the motion current model of each training mode. The inference model training module is used to train the reinforcement learning model to obtain the optimal motor control current for each training mode in order to achieve optimal control of the motor of the intelligent training equipment.

[0016] A computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the method described in any one of the preceding descriptions.

[0017] An apparatus includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor, when executing the computer program, implements the method described in any one of the preceding methods.

[0018] The application has remarkable technical effects due to the above technical scheme: based on motor operation data and training modes, a motion current model of each training mode is obtained; whether an exhausted condition of a trainer appears is judged by detecting a jitter frequency of a motor angle position, if the exhausted condition appears, the motion current model of each training mode is adjusted, specifically, motor stiffness and motor rotational inertia in the motion current model of each training mode are set to zero, and motor damping is increased; finally, based on motor operation data and the motion current model of each training mode, a reinforcement learning model is constructed, and the reinforcement learning model is trained to obtain an optimal motor control current of each training mode to realize optimal control on the motor of the intelligent training equipment. The exhausted condition in the training process is accurately processed, so that the intelligent fitness equipment is safer and more reliable. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0020] Figure 1 A flowchart of an embodiment of a motor control method of a multi-mode intelligent training equipment based on exhaustion protection of the present application; Figure 2 A whole schematic diagram of an embodiment of a motor control system of a multi-mode intelligent training equipment based on exhaustion protection of the present application. DETAILED DESCRIPTION

[0021] The present application will be further described in detail below in combination with embodiments. The following embodiments are an explanation of the present application and the present application is not limited to the following embodiments. The features in the following embodiments can be combined with each other without conflict.

[0022] Embodiment 1: A motor control method of a multi-mode intelligent training equipment based on exhaustion protection, as shown in Figure 1 The intelligent training equipment can provide multiple training modes, including the following steps: S100: obtaining training modes and motor operation data; based on motor operation data and motor control current, constructing a motion current model; wherein the motor operation data at least includes motor angle position, motor angular velocity, motor acceleration, motor stiffness, motor damping and motor rotational inertia, and the training mode at least includes fluid resistance mode; S200: based on training modes, combining the motion current model, obtaining a motion current model of each training mode; S300: Collect motor angular position data of a preset monitoring period to obtain an angular position sequence; accumulate energy above a preset frequency threshold in the angular position sequence to obtain an energy density cumulative value; S400: When the energy density cumulative value is greater than a preset energy density threshold, the training mode is switched to a fluid resistance mode; S500: Based on the motor operation data and the motion current model of each training mode, a reinforcement learning model is constructed; S600: The reinforcement learning model is trained to obtain an optimal motor control current for each training mode to achieve optimal control of the motor of the intelligent training equipment.

[0023] The intelligent training equipment can provide multiple training modes and is an intelligent personal training equipment with a motor as the core element, generally having two motors. The motor provides resistance for training, and the motor operates in a generator mode, so that more than 80% of kinetic energy can be recovered through a super capacitor during training, and even in the fluid resistance mode, the device not only does not consume power, but also can supply power to external devices. The motor is preferably a high-torque axial flux motor, which has a fast current regulation feature with a millisecond-level response to ensure that there is no perceptual delay in mode switching.

[0024] The intelligent training equipment sets the training to be done through hardware, such as training biceps, legs, waist, etc. Through control of the motor, multiple training modes are provided for different training sites to meet different training needs of different groups of people. The intelligent training equipment has multiple training modes: fluid resistance mode, concentric isotonic mode, eccentric isotonic mode, elastic mode, etc.

[0025] The motor control target of the intelligent training equipment is to provide resistance that meets the training needs according to different training modes. In order to achieve this, the motion parameters (displacement, speed, acceleration) of the trainer need to be monitored in real time, which are reflected by the angular position, angular velocity and angular acceleration of the motor. When the speed or direction of the trainer's action changes, the system will quickly respond by adjusting the motor control current to adjust the motor output torque to provide the resistance required by different training modes.

[0026] In S100, the training mode and the motor operation data are obtained; based on the motor operation data and the motor control current, a motion current model is constructed, and the derivation process of the motion current model is as follows: (1) Based on the torque control technology, the motor torque is obtained, which is represented as follows: (1) Formula (1) represents the motor torque generation model based on impedance control. This control method simulates the mechanical behavior of a spring-damper-mass system. The essence of the formula is an extension of Newton's second law to a rotating system, which describes the torque that the motor needs to overcome the elastic, damping, and inertial effects of the system. Therefore, the motor output torque = elastic force + damping force + inertial force.

[0027] Therefore, the angular position, acceleration, and angular acceleration are all related to the error value, which is essentially a closed-loop control expression of the motor torque. Among them, represents the motor torque, represents the motor angular position, which is obtained through the motor encoder. represents the motor angular velocity, which is directly obtained through the sensor. represents the motor angular acceleration, which can be derived from the rate of change of angular velocity or directly obtained through the sensor, represents the motor torque constant term, respectively represent the set angular position, set angular velocity, and set angular acceleration. The set angular position is set by the system's built-in standard motion trajectory database based on the trainer's selected training plan (such as squats, rowing), and is directly given by controlling the motor current. The set angular velocity and set angular acceleration can be directly specified or generated through differentiation. represents the angular position error, represents the angular velocity error, represents the angular acceleration error, represents the motor stiffness (unit: Nm / rad), which is similar to the spring coefficient, indicating the strength of the motor's response to angular position deviation. represents the motor damping (unit: Nm·s / rad), indicating the strength of the motor's response to angular velocity deviation. represents the motor's own rotational inertia (unit: kg·m²), simulating the inertial force, which is used to compensate for acceleration deviation. When the trainer pulls the intelligent training equipment, the motor will generate a resistance torque based on the deviation between the set position, velocity, and acceleration and the actual value. For example, when the trainer suddenly accelerates the pull (actual angular acceleration > set angular acceleration), the inertia term will generate a reverse torque to provide resistance.

[0028] Motor stiffness reflects the ability of the motor to resist external disturbances. By adjusting the control parameters in real time (such as increasing the proportional gain Kp or increasing the current loop bandwidth), the equivalent stiffness can be dynamically changed. For example, by modifying the controller parameters through the host computer instructions, stiffness adjustment can be achieved. Motor damping represents the ability of the system to suppress oscillations and is related to the speed feedback gain (such as the derivative gain Kd) or the damping design in the current loop. The damping parameter in the control algorithm (such as the derivative gain Kd in PID or the virtual damping in the observer) can be adjusted. Motor moment of inertia is an inherent physical property of the motor and load, depending on the mass distribution. The physical meaning of motor inertia cannot be changed in real time, but in control, the effect of inertia change can be simulated through inertia identification algorithm and feedforward compensation to achieve equivalent inertia transformation. For example, the inertia parameter of the speed loop can be adjusted to make the motor response similar to a system with different inertia (but this does not change the actual physical inertia). It should be noted that the adjustment of equivalent moment of inertia is essentially an optimization of the control strategy, rather than changing the physical properties.

[0029] The motor can adopt an axial flux motor, which has high torque density and low moment of inertia. The motor rotor responds quickly to current changes, and the adjusted torque is output within milliseconds, making the motor stiffness K, motor damping B, and motor moment of inertia M effective within milliseconds, significantly improving the force control smoothness and mode switching speed of the fitness equipment. Even in high-damping mode, it can still output large torque stably, avoiding mechanical impact caused by sudden resistance changes.

[0030] (2) According to impedance control technology, the motor torque is calculated from the motor current torque constant and the motor control current, and then the motor control current is assigned to the motor to directly output the motor torque. The motor torque is represented as follows: (2) where, is the motor control current, mainly referring to the motor shaft current, is the motor current torque constant, also known as the motor torque constant, which represents the torque generated by the motor per unit current. It is an important design parameter of the motor and reflects the electromagnetic conversion efficiency of the motor. Formula (2) expresses the linear relationship between motor torque and current. For permanent magnet motors (such as axial flux motors), the motor torque is proportional to the motor control current, and the proportional coefficient is the torque constant.

[0031] (3) Combining formula (1) and formula (2), the motion current model is obtained, represented as follows:

[0032] The control device inside the intelligent fitness equipment calculates the required resistance / motor torque according to formula (1), and then converts the required resistance into motor control current through formula (2), and then drives the motor to output the target torque / resistance, which is the bottom link of motor control, and is usually realized by a current loop (inner loop).

[0033] The overall working process of the intelligent fitness equipment is as follows: Setting target: The intelligent fitness equipment converts the related settings (such as the size of constant resistance, speed-related resistance, etc.) of the training mode and the trainer into the set angular position, angular velocity, and angular acceleration.

[0034] Sensor feedback: After the trainer starts training, the sensor measures the actual angular position of the motor. The actual angular velocity of the motor can be obtained by differentiating the angular position or through the sensor. The angular acceleration of the motor can be obtained by twice differentiating the angular position or through the sensor.

[0035] Torque calculation: According to formula (1), the required motor torque, that is, the ideal motor torque in the following, is calculated.

[0036] Current control: According to formula (2), the motor torque is converted into motor control current, and the motor is driven through the inverter.

[0037] In S200, based on the training mode, a motion current model of each training mode is obtained in combination with the motion current model, and the training mode at least includes a fluid resistance mode, a centrifugal isotonic mode, a centripetal isotonic mode, and an elastic force mode.

[0038] (1) Fluid resistance mode The fluid resistance mode simulates the characteristics of wind resistance and water resistance rowing machine. The faster the speed used by the trainer, the greater the resistance. In this mode, the muscles of the legs, waist, upper limbs, front side of the upper torso, and back can be enhanced, and the whole body muscles can achieve good training effect. In the fluid resistance mode, the motor stiffness and the motor moment of inertia are both zero, the motor damping is a negative number with adjustable size, indicating that the faster the speed of the trainer, the greater the resistance. The motor torque in the fluid resistance mode is represented as follows:

[0039] Therefore, the motion current model of the fluid resistance mode is represented as follows:

[0040] wherein, is a constant compensation term, and is set to be non-adjustable.

[0041] (2) Centrifugal isotonic mode In the eccentric-isometric mode, the trainer provides resistance when performing the outward stretching action, and does not provide resistance when performing the inward contraction action. For example, in weight lifting, the eccentric-isometric mode can be used to enhance the outward stretching strength of the muscles. In the eccentric-isometric mode, the motor stiffness, the motor damping and the motor moment of inertia are zero, the motor angular velocity is positive, the normal force is positive, and the motor torque is positive and adjustable, indicating that the longer the trainer contracts, the greater the force output, and the greater the motor torque output. The motion current model of the eccentric-isometric mode is shown as follows: .

[0042] (3) Concentric-isometric mode In the concentric-isometric mode, the trainer provides resistance when performing the inward contraction action, and does not provide resistance when performing the outward stretching action. For example, in swimming, the concentric-isometric mode can be used to enhance the inward contraction strength of the muscles. In the concentric-isometric mode, the motor stiffness, the motor damping and the motor moment of inertia are zero, the motor angular velocity is negative, the normal force is negative, and the motor torque , is positive and adjustable, indicating that the longer the trainer stretches, the greater the force output, and the greater the motor torque output. The motion current model of the concentric-isometric mode is shown as follows: .

[0043] (4) Elastic mode In the elastic mode, the longer the trainer pulls the traction rope, the greater the resistance provided by the intelligent training device. This mode can effectively improve muscle strength, physical activity and flexibility, and become an aerobic training that can strengthen the heart and lung functions and improve the body shape. In the elastic mode, the motor damping and the motor moment of inertia are zero, the motor stiffness is negative, and the motor torque in the fluid resistance mode is shown as follows:

[0044] The motion current model of the elastic mode is shown as follows:

[0045] wherein, is a constant compensation term and cannot be adjusted. is negative, indicating that the longer the trainer pulls the distance / displacement, the greater the force, is adjustable in several grades.

[0046] In S300, the motor angular position data of a preset monitoring period is collected to obtain an angular position sequence; and the energy above a preset frequency threshold in the angular position sequence is accumulated to obtain an energy density cumulative value. When the trainer is in a state of exhaustion during the training process, causing the trainer to be unable to resist the existing resistance provided by the motor, the intelligent training equipment can automatically monitor and timely adjust to avoid causing harm to the trainer due to a large stroke change. The essence of the exhaustion state detection is to detect the jitter, and after it is determined that the trainer enters the exhaustion state, the system enters the fluid resistance mode, and is automatically set to a larger motor damping, so that when the trainer is exhausted, the dangerous condition of the traditional fitness equipment will not occur. The exhaustion state detection includes the following steps: S310: Collect the motor angular position of a preset monitoring period based on a time sequence to obtain an initial angular position sequence. A high-precision angle encoder can be integrated in the motor, and the encoder captures the instantaneous fluctuation of the angle by sampling the time domain signal of the angle (such as 0.1 ms sampling interval) to realize real-time measurement of the angular velocity and angular position of the motor rotor. An inertial measurement unit (IMU) can also be used at the moving part (such as the handle or weight block) of the fitness equipment, and the IMU is used to detect acceleration and angular velocity, and performs redundant checking in combination with the encoder signal to assist in identifying the jitter characteristics and improve the detection reliability. According to the Nyquist sampling theorem, it is necessary to sample at least twice the target frequency, and in practice, it can be higher to capture high-frequency jitter.

[0047] S320: Perform low-pass filtering processing on the initial angular position sequence to obtain a preprocessed angular position sequence, wherein the preprocessing includes one or more of low-pass filtering and amplification, linearization, and data conversion. The low-pass filtering processing (cutoff frequency 10 Hz) can eliminate high-frequency noise interference.

[0048] S330: Time domain to frequency domain conversion. Perform energy and frequency domain conversion processing on the preprocessed angular position sequence to obtain an angular position spectrum sequence. The time domain jitter signal is converted into the frequency domain through Laplace transform or Fourier transform to analyze its frequency components. Laplace transform is more suitable for analyzing transient signals, but in practice, fast Fourier transform (FFT) is more commonly used to realize spectrum analysis. The frequency spectrum graph of the angle change is obtained, with the horizontal coordinate being the frequency (Hz) and the vertical coordinate being the energy density (amplitude square).

[0049] S340: Based on the angular position spectrum sequence, obtain the energy density cumulative value of the frequency band above the preset frequency threshold. When the trainer is close to exhaustion, the muscle control ability of the trainer decreases, causing the motion stability to decrease, which is usually manifested as high-frequency, low-amplitude involuntary muscle tremors (similar to physiological tremor, frequency range 5-12 Hz, and accordingly the frequency threshold is set), and then the energy density cumulative value of the frequency band above the preset frequency threshold in the spectrum is calculated. The energy density cumulative value is represented as follows:

[0050] in, This represents the cumulative energy density. This indicates the preset frequency threshold, which can be set to 5Hz. This represents the energy density in the angular position spectral sequence. This represents the frequency in the angular position spectrum sequence.

[0051] In S400, when the cumulative energy density value exceeds the preset energy density threshold, the training mode is switched to fluid resistance mode.

[0052] When the cumulative energy density value exceeds the preset energy density threshold (i.e., the trainee has entered a state of exhaustion), the motor stiffness and motor moment of inertia are adjusted to zero, and the motor damping is increased to the preset damping value (e.g., by 50%). Since the motor damping is negative at this time, it is actually an increase in the absolute value of the motor damping. The increase in motor damping should be such that the increased motor damping does not exceed the maximum motor damping. When the trainee is detected to have entered a state of exhaustion, the motor immediately switches to a high (absolute value) damping fluid resistance mode (such as a viscous fluid controlled by a solenoid valve). This increases the motor torque / resistance, and the speed of the training equipment drops sharply, preventing rapid fall or loss of control. The preset energy density threshold can be set according to experimental values. Based on the adjusted motor stiffness, motor moment of inertia, and motor damping, the required torque is recalculated using formula (1), and the q-axis current of the motor is adjusted according to the new motor torque.

[0053] Furthermore, to avoid false triggering of exhaustion mode, IMU data can be used to verify the correlation between jitter direction and motion trajectory, thus preventing false triggering (e.g., external vibration interference). In calculating the cumulative energy density value, besides distinguishing between exhaustion jitter and environmental noise by setting a preset frequency threshold, different weights can be assigned to different frequency bands (e.g., 5-10Hz has a higher weight, and above 10Hz attenuates). The cumulative energy density value is then expressed as follows:

[0054] in, Both represent the weight of the frequency band, and , Both represent preset frequency thresholds. This indicates the total number of preset frequency thresholds.

[0055] In another embodiment, the multi-mode intelligent training equipment motor control method based on exhaustion protection may further include hand-drop detection and protection. Specifically, when the motor angular velocity is greater than a first preset angular velocity threshold, the motion current model of each training mode is adjusted according to the difference between the motor angular velocity and the first preset angular velocity threshold. The training equipment provides adjustable resistance through the motor. When the trainee holds the handle, the motor outputs torque to resist human force. If the trainee suddenly drops their hand, the rope and load will rebound rapidly due to inertia or elasticity after the motor resistance disappears, which may cause equipment damage or personal injury. There are many factors that can cause hand-drop, such as exhaustion, insufficient physical fitness of the trainee, improper training movements, and problems with the training equipment. Therefore, after a hand-drop occurs, in order to avoid the rope rebounding too quickly and causing injury to the intelligent training equipment itself or the trainees around, it is necessary to limit the rebound speed of the rope and automatically activate hand-drop protection. The essence of hand-drop protection is a speed limiting mode, and the motor torque output is judged at the motor end. By monitoring the motor angular velocity (i.e., the linear velocity corresponding to the rope rebound) in real time, when the detected angular velocity exceeds the first preset angular velocity threshold (e.g., 10 rad / s), it is determined to be a release state, and then the speed loop proportional control (P control) is activated. The motor outputs reverse torque to suppress the rebound. The adjusted motor torque is expressed as follows:

[0056] in, This indicates the adjusted motor torque. This indicates the first preset angular velocity threshold. It can be based on the maximum angular velocity Settings, for example . Represents the proportional gain, which indicates the increase in torque output corresponding to a unit speed deviation. Its physical meaning is the motor's "resistance strength" to speed overshoot.

[0057] The adjusted motor control current is obtained based on the adjusted motor torque. The adjusted motion current model for each training mode is expressed as follows:

[0058] in, This indicates the adjusted motor control current.

[0059] Release detection is implemented throughout the training process to protect trainees from release during normal training. Release detection generally does not occur simultaneously with exhaustion. However, if both exhaustion and release are detected under special circumstances, release detection will be prioritized. Specifically: when the motor angular velocity exceeds a first preset angular velocity threshold and the cumulative energy density exceeds a preset energy density threshold, the motion current model for each training mode is adjusted based on the difference between the motor angular velocity and the first preset angular velocity threshold. The adjusted motion current model for each training mode is expressed as follows: .

[0060] In the S500, a reinforcement learning model is constructed based on motor operation data and the motion current model for each training mode. The reinforcement learning environment simulates the dynamic behavior of the motor in the intelligent training equipment. The reinforcement learning model includes a state space model, an action space model, and a reward function, specifically including: (1) Construct a state-space model. The state space S includes the motor's angular position and angular velocity. Angular acceleration of motor Motor stiffness Motor damping and the moment of inertia of the motor The state-space model is then represented as follows:

[0061] (2) Construct the motion space model. The motion space A includes the input signals of the motor, mainly the motor control current, etc. The motion space model is represented as follows:

[0062] in, This represents a state in the state space. This represents an action in the action space. This represents a six-dimensional real vector.

[0063] (3) Construct the reward function. The reward function is constructed based on the performance of the motor (motor torque, cumulative energy density, and motor angular velocity). The reward function needs to reflect the requirements of force position memory, exhaustion protection, and overspeed protection. The reward function is expressed as follows:

[0064] in, Indicates the reward value. This indicates the bonus value for the position where the force is applied. Indicates the exhaustion reward value. This represents the speeding bonus value. Both represent weighting coefficients, and .

[0065] Power Position Bonus Value This is used to reward the motor for producing the correct torque at a specific location, and is represented as follows:

[0066] The power position reward formula represents the amount of torque applied. (That is, the motor torque predicted by the reinforcement learning model is the motor torque predicted based on the current state) equals the ideal motor torque. When this happens, the system will receive a positive reward. Otherwise, the system will be subject to a negative penalty. Ideal motor torque The final motor torque after the above steps (S100~S400 and after the hand-off detection) is the motor torque after the exhaustion protection adjustment and the hand-off protection adjustment. It is the motor torque obtained by adjusting the formula (1). For ease of understanding, the ideal motor torque can be calculated by formula (2) based on the adjusted motor control current.

[0067] Exhaustion protection means that when the motor reaches exhaustion, the system is given a negative reward to prevent further damage. The exhaustion reward value is... , means as follows:

[0068] The exhaustion reward formula indicates that if the state of exhaustion is reached, the system will receive a negative penalty. If the state of exhaustion is not reached, there is no additional reward or penalty. The determination of the state of exhaustion is consistent with that above: when the cumulative energy density value is greater than the preset energy density threshold, the state of exhaustion is reached.

[0069] Hand-off protection, also known as overspeed protection, means that when the motor speed exceeds the safety limit, the system receives a negative reward, the overspeed reward value. , means as follows:

[0070] The overspeed bonus formula represents the percentage of angular velocity that is increased. Exceeded the maximum motor angular velocity The system will be subject to a negative penalty. If the maximum allowed value is not exceeded, there is no additional reward or penalty.

[0071] (4) This application also imposes constraints on the motor's operating data, stipulating that the motor's operating data (motor angular position, motor angular velocity, motor angular acceleration, motor stiffness, motor damping and motor moment of inertia) and motor torque must operate within the safe range formed by the minimum and maximum values.

[0072] In S600, the reinforcement learning model is trained to obtain the optimal motor control current for each training mode, thereby achieving optimal control of the motor of the intelligent training equipment. This includes the following steps: S610: Construct a policy model, which includes a policy function, a policy gradient model, a cumulative reward model, and a policy update model; The policy function represents the state. Select action The probability is determined by the policy parameter θ, and the policy function is expressed as follows:

[0073] in, Indicates the state given Choose action The probability of. Indicates the state given Time action The expected value, or the mean of the action, is usually output by a neural network. These represent policy parameters used to determine the probability of choosing an action, including parameters of mean and variance networks, such as the weights of a neural network. The covariance matrix representing the actions is usually a diagonal matrix (independent action dimension).

[0074] The policy gradient model is represented as follows:

[0075] in, Describe the policy objective function Regarding strategy parameters The gradient. Representing state The steady-state distribution represents the policy. Next state The probability of occurrence. The action-value function represents the state. Take action below And follow the strategy Expected cumulative rewards The policy function represents the policy parameters. The policy gradient model is used to calculate the expected value of the policy gradient, which guides the direction of policy parameter updates, thereby maximizing the cumulative reward. The core idea of ​​the policy gradient model is to adjust... This increases the probability of high-reward actions and decreases the probability of low-reward actions.

[0076] The cumulative reward model is represented as follows:

[0077] in, This represents the cumulative reward value. Indicates from time step Initial cumulative reward value, This indicates the total number of time steps in the current round. This represents the discount factor.

[0078] The policy update model is represented as follows:

[0079] in, This indicates the updated policy parameters. Indicates the strategy parameters, This represents the learning rate (step size), used to control the magnitude of policy parameter updates.

[0080] S620: Train the reinforcement learning model for each training mode separately using the REINFORCE algorithm, including the following steps: (1) Initialize hyperparameters (learning rate, discount factor, maximum number of rounds and maximum number of steps per round), policy parameters and trajectory buffer.

[0081] (2) Obtain the initialization state As the current state, the action to be taken is obtained based on the current state and the policy function. The initial state can be obtained by the reset function, or the initial state can be randomly sampled to increase the diversity of training, or it can start from a fixed initial state. This application does not limit this.

[0082] (3) Based on the current state and the action taken, and combined with the reward function, the current reward value is obtained, and the current state, the action taken, and the current reward value are added to the trajectory buffer.

[0083] (4) Based on the action taken, obtain the next state and update the step count.

[0084] (5) Repeat steps (1) to (4) until the round termination condition is met. The round termination condition includes the environment returning a termination signal (such as the motor running data exceeding the safe range or the task being completed) or reaching the maximum number of steps in the current round.

[0085] (6) Based on the trajectory buffer and combined with the cumulative reward model, the cumulative reward value of each step in the current round is obtained, and then the cumulative reward value of the current round is obtained; based on the trajectory buffer and combined with the policy gradient model, the policy gradient of the current round is obtained, and based on the policy gradient of the current round and combined with the policy update model, the policy parameters are updated.

[0086] (7) Repeat steps (1) to (6) until the training termination condition is met, and obtain the optimal motor control current, i.e. the optimal motion current model, in each state space. The training termination condition includes the most recent average cumulative reward value of the rounds being greater than the preset reward threshold or the training reaching the maximum number of rounds.

[0087] (8) Obtain real-time motor operation data. Based on the training mode and the real-time motor operation data, and combined with the optimal motion current model of each training mode, obtain the real-time optimal motor control current for each training mode. The trained reinforcement learning model is also the optimal motion current model. According to the training mode and the real-time motor operation data, output the probability distribution of the action (motor control current). The probability distribution of the action is usually assumed to be Gaussian distribution. Then, the mean value of the motor control current is selected as the optimal motor control current.

[0088] This invention simulates scenarios of exhaustion and release conditions using a reinforcement learning model. The REINFORCE algorithm is a strictly online algorithm; its training data must be generated in real-time through interaction between the current policy and the environment. Training the reinforcement learning model using the REINFORCE algorithm eliminates the need for historical data (i.e., past interaction experience). After training, the resulting model can intelligently select the optimal action (motor control current) based on the current state (motor angular position, motor angular velocity, motor angular acceleration, motor stiffness, motor damping, and motor moment of inertia) to maximize long-term cumulative rewards (enabling exhaustion and overspeed protection), and corresponding protective measures. It corrects the uneven torque problem under low motor speed and reverse drag force, enabling the intelligent training equipment to provide force position memory, exhaustion protection, and overspeed protection in different training modes.

[0089] Example 2: A multi-mode intelligent training equipment motor control system based on exhaustion protection, such as Figure 2 As shown, the multi-mode intelligent training equipment motor control method based on exhaustion protection described in any of the above embodiments can be implemented. The control system includes a basic model module 100, a training mode model module 200, an energy density module 300, a model adjustment module 400, an inference model construction module 500, and an inference model training module 600. The basic model module 100 is used to acquire training modes and motor operation data; and to construct a motion current model based on the motor operation data and motor control current; wherein, the motor operation data includes at least motor angular position, motor angular velocity, motor acceleration, motor stiffness, motor damping and motor moment of inertia, and the training mode includes at least a fluid resistance mode; The training mode model module 200 is used to obtain the motion current model for each training mode based on the training mode and combined with the motion current model. The energy density module 300 is used to collect motor angular position data for a preset monitoring period to obtain an angular position sequence; and to accumulate the energy above a preset frequency threshold in the angular position sequence to obtain an energy density cumulative value. The model adjustment module 400 is used to switch the training mode to fluid resistance mode when the cumulative energy density value is greater than the preset energy density threshold. The inference model building module 500 is used to build a reinforcement learning model based on motor operation data and the motion current model of each training mode. The inference model training module 600 is used to train the reinforcement learning model to obtain the optimal motor control current for each training mode in order to achieve optimal control of the motor of the intelligent training equipment.

[0090] Various changes and modifications made without departing from the spirit and scope of this invention, and all equivalent technical solutions, also fall within the scope of this invention.

[0091] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0092] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] This invention is described with reference to flowchart illustrations and / or block diagrams of the method, terminal device (system), and computer program product according to the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0094] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0095] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0096] It should be noted that: The phrase "an embodiment" or "an embodiment" used in this specification means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. Therefore, the phrase "an embodiment" or "an embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment.

[0097] Furthermore, it should be noted that the shapes and names of the parts and components described in the specific embodiments described in this specification may differ. All equivalent or simple variations made to the structure, features, and principles described in this patent concept are included within the protection scope of this patent. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to replace them, as long as they do not depart from the structure of this invention or exceed the scope defined in these claims, they should all fall within the protection scope of this invention.

Claims

1. A motor control method for a multi-mode intelligent training device based on exhaustion protection, wherein the intelligent training device provides multiple training modes, characterized in that, Includes the following steps: Acquire training mode and motor operation data; A motion current model is constructed based on motor operating data and motor control current; wherein the motor operating data includes at least motor angular position, motor angular velocity, motor acceleration, motor stiffness, motor damping, and motor moment of inertia, and the training mode includes at least a fluid resistance mode; Based on the training mode and combined with the motion current model, the motion current model for each training mode is obtained; Collect motor angular position data for a preset monitoring period to obtain an angular position sequence; accumulate the energy above a preset frequency threshold in the angular position sequence to obtain the cumulative energy density value; When the cumulative energy density value exceeds the preset energy density threshold, the training mode will be switched to fluid resistance mode. A reinforcement learning model is constructed based on motor operation data and motion current models for each training mode. The reinforcement learning model is trained to obtain the optimal motor control current for each training mode, thereby achieving optimal control of the motor of the intelligent training equipment.

2. The method for controlling a multi-mode intelligent training equipment motor based on exhaustion protection according to claim 1, characterized in that, The motion current model is represented as follows: in, Indicates the motor control current. Indicates motor stiffness. Indicates motor damping, Indicates the moment of inertia of the motor. Indicates the angular position of the motor. Indicates the angular velocity of the motor. Indicates the angular acceleration of the motor. These represent the set angular position, set angular velocity, and set angular acceleration, respectively. This represents the motor torque constant term. This represents the motor torque constant.

3. The method for controlling the motor of a multi-mode intelligent training equipment based on exhaustion protection according to claim 1, characterized in that, The training modes include at least the fluid resistance mode, the centripetal isotensive mode, the eccentric isotensive mode, and the elastic mode; In the fluid resistance mode, the motor stiffness and the motor moment of inertia are zero in the motion current model, and the motor damping is negative. In the centripetal isotensile mode, the motor stiffness, motor damping, and motor moment of inertia in the motion current model are zero, and the motor angular velocity is positive. In the centrifugal isotensile mode, the motor stiffness, motor damping, and motor moment of inertia in the motion current model are zero, and the motor angular velocity is negative. In the elastic mode, the motor damping and the motor moment of inertia in the motion current model are zero, and the motor stiffness is negative.

4. The multi-mode intelligent training equipment motor control method based on exhaustion protection according to claim 1, characterized in that, The process of collecting motor angular position data over a preset monitoring period to obtain an angular position sequence, and accumulating the energy above a preset frequency threshold in the angular position sequence to obtain a cumulative energy density value, includes the following steps: The initial angular position sequence is obtained by collecting the motor angular position over a preset monitoring period based on the time series data. The initial angular position sequence is low-pass filtered to obtain the preprocessed angular position sequence; The preprocessed angular position sequence is subjected to energy and frequency domain transformation to obtain the angular position spectrum sequence; Extract and accumulate the energy density of the frequency bands in the angular position spectrum sequence that are greater than or equal to a preset frequency threshold to obtain the cumulative energy density value.

5. The method for controlling a multi-mode intelligent training equipment motor based on exhaustion protection according to claim 1, characterized in that, Also includes: When the motor angular velocity is greater than a first preset angular velocity threshold, the motion current model of each training mode is adjusted according to the difference between the motor angular velocity and the first preset angular velocity threshold; the adjusted motion current model of each training mode is expressed as follows: in, This indicates the adjusted motor control current. Indicates proportional gain. Indicates the angular velocity of the motor. This indicates the first preset angular velocity threshold. This represents the motor torque constant.

6. The method for controlling a motor in a multi-mode intelligent training equipment based on exhaustion protection according to claim 1, characterized in that, The reinforcement learning model includes a state space model, an action space model, and a reward function, specifically: A state-space model is constructed based on the motor's angular position, angular velocity, angular acceleration, stiffness, damping, and moment of inertia. Based on the motor control current, construct an action space model; The reward value for the force application position is obtained based on the motor torque predicted by the reinforcement learning model and the ideal motor torque; wherein, the ideal motor torque is obtained through motor operation data; The exhaustion reward value is obtained based on the cumulative value of motor energy density and the preset energy density threshold; wherein, the cumulative value of motor energy density is obtained by the motor angular position and the preset frequency threshold; The overspeed bonus value is obtained based on whether the motor angular velocity exceeds the maximum motor angular velocity. A reward function is constructed by weighted summation of the reward values ​​for exertion position, exhaustion, and overspeed.

7. The method for controlling a multi-mode intelligent training equipment motor based on exhaustion protection according to claim 6, characterized in that, When the cumulative energy density value is greater than the preset energy density threshold, the training mode is switched to fluid resistance mode. Specifically, the motor stiffness and the motor moment of inertia are set to zero, the motor damping is set to a negative value, and the absolute value of the motor damping is increased.

8. A multi-mode intelligent training equipment motor control system based on exhaustion protection, characterized in that, The system is capable of implementing the multi-mode intelligent training equipment motor control method based on exhaustion protection as described in any one of claims 1-7. The control system includes a basic model module, a training mode model module, an energy density module, a model adjustment module, an inference model construction module, and an inference model training module. The basic model module is used to acquire training modes and motor operation data; and to construct a motion current model based on the motor operation data and motor control current; wherein the motor operation data includes at least motor angular position, motor angular velocity, motor acceleration, motor stiffness, motor damping and motor moment of inertia, and the training modes include at least fluid resistance modes. The training mode model module is used to obtain the motion current model for each training mode based on the training mode and combined with the motion current model. The energy density module is used to collect motor angular position data for a preset monitoring period to obtain an angular position sequence; and to accumulate the energy above a preset frequency threshold in the angular position sequence to obtain an energy density cumulative value. The model adjustment module is used to switch the training mode to fluid resistance mode when the cumulative energy density value is greater than the preset energy density threshold. The inference model building module is used to build a reinforcement learning model based on motor operation data and the motion current model of each training mode. The inference model training module is used to train the reinforcement learning model to obtain the optimal motor control current for each training mode in order to achieve optimal control of the motor of the intelligent training equipment.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

10. An apparatus comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 7.

Images