Action skill training method and device based on hybrid drive, equipment and storage medium

Abstract

The invention belongs to the technical field of exoskeleton robots, and discloses an action skill training method, device and equipment based on hybrid drive and a storage medium. According to the method, deviation information is generated by comparing actions of a user with a target track in real time, an active driving unit is cooperatively controlled to provide basic assistance according to the deviation information, and meanwhile a semi-active driving unit is instructed to generate guiding damping force opposite to the deviation direction; the method further has a gradual mechanism for adjusting the assistance level according to the learning progress, and dynamic posture locking can be conducted on the associated joints at the key stage of multi-joint coordination action. According to the scheme, vector type damping guidance aiming at motion deviation is adopted, so that a user moves smoothly when the motion is correct and can obtain clear physical guidance when the motion deviates, the auxiliary learning effect of the user in high-difficulty limb motion is remarkably improved, the standardization and stability of core links in complex motion are ensured, and the user experience is improved. And finally, intelligent and bionic high-level skill teaching is realized.

Description

Technical Field

[0001] This application relates to the field of exoskeleton robot technology, and in particular to a method, apparatus, device and storage medium for training motion skills based on hybrid drive. Background Technology

[0002] With the rapid development of robotics technology, wearable exoskeleton systems have achieved remarkable results in the fields of medical rehabilitation and industrial assistance. These systems are typically driven by motors to provide support for patients with limited mobility or to reduce the burden on workers, with their core objective being to enable basic mobility functions or provide strength enhancement. However, this technological wave mainly focuses on "functional replacement" or "strength assistance," and its design logic is fundamentally different from the needs of motor skill training that pursues precise, coordinated, and efficient movements.

[0003] Specifically, in high-level motor skill training fields such as sports and dance, traditional training models heavily rely on the coach's experience, observation, and immediate feedback. However, this manual guidance method has inherent limitations, including delayed feedback, difficulty in quantification, and the inability to provide real-time physical guidance. To overcome these limitations, some studies have attempted to introduce exoskeleton technology into skills training, but existing solutions still have significant shortcomings when teaching high-dynamic, high-precision skills.

[0004] While existing technologies (such as CN108309689B1) propose the concept of "progressive" training, their core is based on the macroscopic stage division of the rehabilitation cycle, such as dividing training into early, middle, and late stages of rehabilitation, and employing position control, torque control, or resistance modes at different stages. This progressive control cannot achieve millisecond-level real-time dynamic adjustments during a single movement. Secondly, the control logic and hardware design of such devices are mainly designed for the slow rehabilitation of patients with muscle weakness, lacking adaptability to high-speed explosive movements (such as kicking), and their high-inertia drive system hinders the smoothness of movement.

[0005] Other existing technologies (such as CN110303471B1) focus on providing "comfort assist" through motion intention recognition and variable impedance control, aiming to follow user intentions and reduce user energy consumption. However, this contradicts the core requirement of "error correction teaching" in skills training. This invention aims to apply precise physical constraints for "error correction" when the user's movements deviate from the ideal trajectory; that is, "the system prevents the user from going in the wrong direction." However, this existing technology does not disclose how to achieve this direction-sensitive, fast-response "directional virtual wall" type correction.

[0006] Furthermore, some solutions (such as CN116036538A1) utilize principles like electromagnetic eddy currents to construct passive dampers for strength training. These devices are purely passive, providing only speed-related resistance and offering no active assistance when initiating a movement or overcoming gravity. Their function is limited to muscle-building loads, completely neglecting guidance and instruction of movement trajectories, and failing to achieve seamless coordination and switching between "active assistance" and "passive correction."

[0007] Therefore, how to design a method that can simultaneously achieve highly transparent motion tracking and high-precision trajectory correction in highly dynamic skill training is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0008] The main objective of this application is to provide a hybrid-driven motion skill training method, device, equipment, and storage medium, aiming to solve the technical problem in the prior art of how to design a method that can simultaneously achieve high-transparency motion following and high-precision trajectory correction in high-dynamic skill training.

[0009] To achieve the above objectives, this application proposes a hybrid-driven motor skill training method, the method comprising: Acquire actual movement data of the user's limbs; The motion vector deviation is obtained based on the actual motion data and the preset target motion trajectory; If the motion vector deviation exceeds the preset vector error boundary range, a hybrid cooperative drive signal is generated based on the motion vector deviation. The hybrid cooperative drive signal includes a basic assist signal for the active drive unit and a guide damping signal for the semi-active drive unit that is opposite in direction to the motion vector deviation. The exoskeleton system is controlled according to the hybrid collaborative drive signal to assist the user in performing skill training actions.

[0010] Optionally, obtaining the motion vector deviation based on the actual motion data and the preset target motion trajectory includes: Based on the actual motion data, obtain the spatial position coordinates and joint angles of each joint of the user at the current moment; Based on the preset target motion trajectory, obtain the target spatial position coordinates and target joint angles of each joint corresponding to the current moment; The position vector deviation is obtained based on the spatial position coordinates and the target spatial position coordinates; The motion posture vector deviation is obtained based on the current joint angle and the target joint angle. The motion vector deviation is obtained based on the position vector deviation and the motion attitude vector deviation.

[0011] Optionally, before generating the hybrid cooperative drive signal based on the motion vector deviation if the motion vector deviation exceeds a preset vector error boundary range, the method further includes: Obtain the current user's historical learning duration and historical movement deviation data; An evaluation coefficient for motion stability is obtained by evaluating the historical learning duration and historical motion deviation data. The vector error boundary range is obtained based on the motion stability evaluation coefficient and the preset scaling mapping rule.

[0012] Optionally, a hybrid cooperative drive signal is generated based on the motion vector deviation, including: The degree of motion deviation is determined based on the magnitude of the projection of the motion vector deviation onto the normal direction of the preset target motion trajectory; Based on the comparison between the degree of motion deviation and the preset primary intervention threshold, a basic assist signal for the active drive unit is generated. The direction of deviation is determined based on the projection direction of the motion vector deviation onto the tangent of the target motion trajectory; Based on the deviation direction, a guide damping signal opposite to the deviation direction is generated for the semi-active drive unit.

[0013] Optionally, generating a guide damping signal for the semi-active drive unit that is opposite to the deviation direction, based on the deviation direction, includes: The magnitude of the normal deviation is determined based on the projection of the motion vector deviation onto the normal of the target motion trajectory. The magnitude of the target damping torque is determined based on the magnitude of the normal deviation. Based on the preset magnetorheological fluid model, the target damping torque corresponding to the normal deviation amplitude is obtained; The target control current value is calculated based on the relationship between the target damping torque and the torque-current characteristic of the magnetorheological damper in the semi-active drive unit. Based on the target control current value, the guiding damping signal is generated to drive the magnetorheological damper to generate a guiding damping torque opposite to the normal deviation direction.

[0014] Optionally, the motion vector deviation, before generating the hybrid cooperative drive signal, further includes: Based on the preset number of learning periods and training data records, the average historical motion vector deviation of the user within the training period is extracted; Based on the average deviation of the historical motion vector and the trend of the average deviation, the corresponding fade-out adjustment coefficient is calculated. The gain amplitude of the basic assist signal acting on the active drive unit is adjusted according to the fade-out adjustment coefficient. Based on the fade-out adjustment coefficient, the intervention threshold amplitude used to trigger the guide damping signal is adjusted synchronously, wherein the adjustment method of the gain amplitude of the basic assist signal is negatively correlated with the adjustment method of the intervention threshold amplitude of the guide damping signal.

[0015] Optionally, the hybrid-driven motion skill training method further includes: The motion stage is identified based on the acceleration change characteristics of the main drive joint and the preset motion stage threshold to obtain the motion stage at the current moment. Based on the preset joint biomechanical coupling model and the aforementioned action phase, determine the associated joints that require posture locking; Based on the angle value that the associated joint needs to maintain, an angle locking damping signal is generated for the semi-active drive unit of the joint. Based on the angle-locked damping signal, the damper of the associated joint is controlled to output a high damping torque to maintain the joint angle.

[0016] Furthermore, to achieve the above objectives, this application also proposes a hybrid-driven motion skill training device, which includes: The data acquisition module is used to acquire actual movement data of the user's limbs; The deviation calculation module is used to calculate the motion vector deviation based on the actual motion data and the preset target motion trajectory. The signal generation module is used to generate a hybrid cooperative drive signal based on the motion vector deviation if the motion vector deviation exceeds a preset vector error boundary range. The hybrid cooperative drive signal includes a basic assist signal for the active drive unit and a guide damping signal for the semi-active drive unit that is opposite to the direction of the motion vector deviation. The control execution module is used to control the exoskeleton system according to the hybrid collaborative drive signal to assist the user in performing skill training actions.

[0017] In addition, to achieve the above objectives, this application also proposes a hybrid-driven motion skill training device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the hybrid-driven motion skill training method described above.

[0018] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the hybrid-driven motion skill training method described above.

[0019] The proposed technical solutions (one or more) have at least the following technical effects: This solution generates deviation information by comparing the user's movements with the target trajectory in real time. Based on this, it coordinates the active drive unit to provide adaptive assistance, while simultaneously instructing the semi-active damping unit to generate precise guiding resistance opposite to the deviation direction. The solution also introduces a fading mechanism that dynamically adjusts the assistance level as training progresses, and can dynamically lock the posture of related joints during key stages of multi-joint coordinated movements. Because this solution employs a directional variable damping guidance strategy based on motion vector deviation, users experience nearly unrestricted, smooth movement when their movements are correct, and receive an immediate and intuitive sense of physical boundaries when their movements deviate, thus significantly improving the accuracy and feedback efficiency of skill training. Simultaneously, the synergistic effect of the fading mechanism and dynamic posture locking not only effectively avoids the user's dependence on equipment and promotes the autonomous internalization and consolidation of motor skills, but also ensures the standardization and stability of core links in complex movement chains, ultimately achieving intelligent and biomimetic efficient skill teaching. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating an embodiment of the hybrid-driven motion skill training method of this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the hybrid-driven motion skill training method of this application; Figure 3 This is a schematic diagram of the module structure of the hybrid-driven motion skill training device according to an embodiment of this application; Figure 4 This is a schematic diagram of the device structure of the hardware operating environment involved in the hybrid-driven motion skill training method in this application embodiment.

[0023] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0024] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0025] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0026] In this embodiment, for ease of description, the following description will focus on a hybrid-driven motion skill training device as the execution subject.

[0027] This application generates deviation information by comparing user movements with the target trajectory in real time. Based on this, it coordinates the control of the active drive unit to provide adaptive assistance, while instructing the semi-active damping unit to generate precise guiding resistance opposite to the direction of deviation. The scheme also introduces a fading mechanism that dynamically adjusts the assistance level as training progresses, and can dynamically lock the posture of related joints at key stages of multi-joint coordinated movements. Because this scheme employs a directional variable damping guidance strategy based on motion vector deviation, users experience nearly unrestricted, smooth movement when their movements are correct, and receive an immediate and intuitive sense of physical boundaries when their movements deviate, thus significantly improving the accuracy and feedback efficiency of skill training. Simultaneously, the synergistic effect of the fading mechanism and dynamic posture locking not only effectively avoids user dependence on equipment and promotes the autonomous internalization and consolidation of motor skills, but also ensures the standardization and stability of core links in complex movement chains, ultimately achieving intelligent and biomimetic efficient skill teaching.

[0028] This application provides a solution aimed at addressing the technical problem in the prior art of how to design a method that can simultaneously achieve highly transparent motion tracking and highly accurate trajectory correction in highly dynamic skill training.

[0029] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or a hybrid-driven motion skill training device capable of performing the above functions. The following description uses a hybrid-driven motion skill training device as the executing entity to illustrate this embodiment and the subsequent embodiments.

[0030] Based on this, embodiments of this application provide a hybrid-driven motion skill training method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the hybrid-driven motion skill training method of this application.

[0031] In this embodiment, the hybrid-driven motion skill training method includes steps S10 to S40: Step S10: Obtain the actual movement data of the user's limbs.

[0032] It should be noted that "user limb" specifically refers to the limb part corresponding to the exoskeleton system equipped with this training system, such as the upper or lower limbs of the operator using the training system to learn motor skills; "actual motion data" refers to a series of physical quantities that reflect the real-time motion state of the user's limbs in three-dimensional space, collected in real time by sensors deployed on the exoskeleton or in the surrounding environment during the training process. This data typically includes, but is not limited to, joint angles, angular velocities, angular accelerations, end effector positions, postures, and electromyographic signals.

[0033] Understandably, actual motion data involves the simultaneous acquisition, filtering, noise reduction, coordinate unification, and data fusion processing of signals from multiple sensors, including inertial measurement units, encoders, potentiometers, and even visual capture systems. For example, during arm curl training, the system may need to obtain the joint flexion and extension angles through an encoder at the elbow joint, while simultaneously using the inertial measurement unit of the upper arm to compensate for the effects of torso sway, thereby more accurately reconstructing the true motion trajectory of the arm relative to the torso.

[0034] It should be understood that in practical applications, the system needs to ensure the real-time, synchronous, and low-latency nature of data acquisition to meet the stringent requirements for immediate feedback in highly dynamic skill training scenarios. Simultaneously, considering the differences in limb size and movement habits among individual users, the system may also need to possess a certain degree of adaptive calibration capability. For example, it could guide users to complete several standard movements before training begins to establish a personalized movement benchmark model. This would allow the acquired actual movement data to more accurately reflect the user's deviation from the ideal movement, providing precise input for subsequent intelligent guidance.

[0035] Step S20: Obtain the motion vector deviation based on the actual motion data and the preset target motion trajectory.

[0036] It should be noted that the preset target motion trajectory refers to the ideal motion path or posture change sequence corresponding to the standard and standardized movement skills stored by the system before training begins. It can be represented as a curve of joint angle changes over time, a standard path of the limb end in space, or a series of key posture points. This motion trajectory is generally obtained by professional movement instructors through motion input. The motion vector deviation, on the other hand, is a comprehensive error description that includes information on magnitude, direction, and even time dimensions. It is used to accurately characterize the differences between the beginner's current actual movement and the ideal trajectory in various aspects. The core purpose of this step is to compare the beginner's actual performance with a stable and optimal reference model, thereby transforming the abstract problem of non-standard movement into a concrete and measurable difference in physical quantities.

[0037] Understandably, motion vector deviation is not merely a simple subtraction of instantaneous spatial position or angle, but involves dynamic time warping to align the speed difference between the actual and target movements, or decomposing the tracking error along the trajectory and the deviation perpendicular to the trajectory in the Frenet coordinate system. For example, during golf swing training, the system not only calculates the radial distance between the actual clubhead position and the ideal arc path, but also analyzes the differences between the torso rotation angle, arm extension, and the standard model at various time points during the swing, thus comprehensively obtaining a multi-dimensional deviation vector that fully reflects the quality of the movement. This calculation ensures that the deviation information can accurately guide subsequent assistance or resistance intervention strategies.

[0038] To help those skilled in the art better understand the calculation process of the motion vector deviation, a specific quantitative example is given below, using the typical training movement of a front kick. In this embodiment, the system calculates the user's motion vector deviation in real time, with the following settings: θa(t) is the actual hip flexion angle measured by the sensor at time t; θt(t) is the hip joint target angle extracted from the preset target trajectory and corresponding to the current action phase; φa(t) is the actual knee extension angle measured at time t; φt(t) represents the corresponding target angle of the knee joint.

[0039] The deviation related to the joint angle can then be quantified as: Hip flexion angle deviation: Δθ(t) = θa(t) - θt(t) Knee extension angle deviation: Δφ(t) = φa(t) - φt(t) In addition, to ensure proper posture, the system also monitors the lateral tilt angle of the limb (i.e., the angle by which the leg deviates outward from the midline) γ(t). When the lateral tilt angle γ(t) > 5°, it is considered a postural error (such as knee valgus), and the system will trigger the high-damping state of the hip joint damper for correction. This specific threshold (5°) is derived from biomechanical analysis of the standard front kick movement, ensuring the timeliness and accuracy of the correction.

[0040] Based on the deviations in each dimension calculated above, the system uses weighted fusion and other methods to finally form the comprehensive motion vector deviation vector δ(t). Afterward, the system enters the intervention decision-making stage.

[0041] It should be understood that the quality of information obtained from the calculated motion vector deviation directly determines the accuracy and effectiveness of the system's intelligent intervention. A well-designed deviation calculation module can not only reflect "how much deviation there is," but also indicate "in which direction" and "at which stage of the movement the deviation occurs." This enables the system to subsequently implement refined hybrid control strategies, such as "applying damping in the normal direction where the movement deviates from the correct trajectory, while providing assistance in the tangential direction." Therefore, the robustness and computational efficiency of this step are crucial for improving the overall intelligence level of the training system and the user experience.

[0042] Step S30: If the motion vector deviation exceeds the preset vector error boundary range, a hybrid cooperative drive signal is generated based on the motion vector deviation.

[0043] It should be noted that the preset vector error boundary range refers to a pre-defined threshold area used to judge whether the user's movement is qualified. When the calculated motion vector deviation falls within this boundary range, it indicates that the user's movement basically meets the requirements, and the system may only provide suggestive feedback. Once the deviation exceeds this range, it means that the movement error has reached the level requiring physical intervention. The hybrid collaborative drive signal refers to a composite control command output by the system. This command aims to provide auxiliary power and corrective resistance simultaneously or in stages. The purpose of the collaboration is to identify subtle deviations between the high-dynamic motion trajectory and the ideal model in real time, and to apply targeted, minor assistance to guide it back to the correct path, helping the learner gradually approach the standard movement and complete muscle memory learning.

[0044] In one embodiment, before generating a hybrid collaborative driving signal based on the motion vector deviation if the motion vector deviation exceeds a preset vector error boundary range, the method further includes: acquiring the current user's historical learning duration and historical motion deviation data; evaluating the historical learning duration and historical motion deviation data to obtain a motion stability evaluation coefficient; and obtaining the vector error boundary range based on the motion stability evaluation coefficient and a preset scaling mapping rule.

[0045] It should be noted that historical learning time refers to the total cumulative time a user spends training a specific motor skill on the system; historical motion deviation data refers to the time-series set of motion vector deviations recorded within previous training cycles, reflecting the consistency and improvement trend of the user's motion accuracy. The motion stability evaluation coefficient is a comprehensive quantitative indicator that, through analysis of historical data, characterizes the user's current proficiency and stability level in performing the action. The scaling mapping rule is used to convert the stability coefficient into a specific, operable scaling ratio for the vector error boundary range. The core purpose of this series of preliminary steps is to make the system's intervention threshold adaptive, dynamically adjusting it according to the user's individual learning progress.

[0046] Understandably, the motion stability evaluation coefficient not only represents the average historical deviation of the current learner, but also examines its fluctuation range, rate of decline, and recent data central tendency. For example, for a novice user, their initial historical deviation data may be large and volatile, resulting in a low stability coefficient. Based on this, the system sets a wider error boundary range for them using scaling mapping rules, avoiding excessive interference and discouraging motivation before the user's movements are fully formed. As the user's training time increases, historical data shows that their deviation gradually decreases and tends to stabilize, leading to an increase in the stability coefficient. The system then gradually narrows the error boundary, promoting improved motion accuracy with higher standards.

[0047] It should be understood that this method of dynamically setting error boundaries based on users' historical performance is key to achieving personalized adaptive training. It allows the training system to move beyond rigidly adhering to a fixed standard, enabling it to deeply perceive the user's learning status and skill development, and adjust teaching strategies accordingly. Pre-defined scaling and mapping rules ensure a smooth and reasonable adjustment process, avoiding abrupt changes in the boundary range that could cause user discomfort. This not only improves training efficiency but also aligns better with the natural laws of skill acquisition, continuously stimulating the user's learning potential by gradually increasing the challenge, thereby optimizing the final training effect through humanized interaction.

[0048] In one feasible embodiment, generating a hybrid cooperative drive signal based on the motion vector deviation includes steps A10 to A40.

[0049] Step A10: Determine the degree of motion deviation based on the magnitude of the projection of the motion vector deviation onto the normal direction of the preset target motion trajectory.

[0050] It should be noted that the normal to the target trajectory refers to the direction perpendicular to the tangent of the ideal trajectory, representing the radial distance by which the motion deviates from the correct path. The degree of motion deviation is a quantified value obtained by projecting the vector deviation between the actual trajectory and the target trajectory onto this normal direction and calculating its scalar magnitude. Essentially, this step decomposes the complex spatial deviation, focusing on the vertical distance component that best reflects the accuracy of the motion path.

[0051] Specifically, the degree of motion deviation is precisely quantified through vector calculation. The system analyzes the motion vector deviation vector δ in the Frenet coordinate system (with the tangential vector T and normal vector N along the target trajectory): Normal deviation distance: dN = |vector δ·vector N|, whose scalar value dN directly represents the vertical distance of the current position from the ideal path.

[0052] Tangential deviation component: dT = vector δ·vector T, whose sign (positive or negative) indicates whether the action is ahead or behind the ideal timing.

[0053] The triggering condition for primary intervention can be quantified as follows: when the normal deviation distance dN is greater than a dynamically adjusted primary intervention threshold C. threshold When, i.e., dN>C threshold The system determined that a path error had occurred requiring proactive intervention. Among them, C... threshold It is not a fixed value, but rather dynamically adjusted based on the user's proficiency. This quantitative criterion, based on projection decomposition and threshold comparison, ensures precise and consistent intervention decisions.

[0054] Understandably, calculating the degree of movement deviation provides a measurement of the movement path. For example, in a certain type of training, if a user is required to perform a straight-line lifting motion with their arm, and the user's actual movement trajectory deviates outward in an arc, then the projection length of this deviation onto the normal direction of the straight-line trajectory directly reflects the severity of the deviation from the correct path.

[0055] Step A20: Based on the comparison result between the degree of motion deviation and the preset primary intervention threshold, generate a basic assist signal for the active drive unit.

[0056] It should be noted that the initial intervention threshold is a preset deviation threshold. Active drive units typically refer to components in the exoskeleton that can actively output power (such as servo motors). When the system determines that the user's movement deviates from the normal direction beyond the acceptable initial threshold, it triggers the active drive unit to operate. The basic assist signal it generates drives the exoskeleton to provide a general directional push or pull force, helping the user move along the correct path.

[0057] Understandably, the basic assist signal doesn't aim to precisely cancel out all deviations. Instead, it provides a basic regressive force when the user's deviation is too large and difficult to correct on their own. The core goal of this design is to reduce the learner's psychological burden and physical exertion, especially in the early stages of training or when the user's muscle strength is weak. This assistance effectively prevents loss of control and maintains the basic effectiveness of training. Linking this intervention to a threshold avoids frequent systemic intervention, ensuring user dominance; assistance only appears when truly needed.

[0058] Step A30: Determine the direction of deviation based on the projection direction of the motion vector deviation onto the tangential direction of the target motion trajectory.

[0059] It should be noted that the tangent of the target motion trajectory represents the direction of the movement. The deviation direction specifically refers to whether the actual motion trend is ahead or behind the ideal tangent, aiming to determine the error in the user's speed and phase along the trajectory.

[0060] Understandably, determining the tangential deviation direction is an assessment of the timing and rhythm of the movement. For example, in simulating a swing, even if the path is perfectly correct, swinging too early or too late will affect the result. By analyzing whether the projection of the deviation onto the tangential direction is positive (leading) or negative (lagging), the system can determine whether the user is "rushing" or "slowing down." Understanding this decomposition helps to achieve more refined movement assessment. It should be understood that treating the path deviation in the normal direction separately from the timing deviation in the tangential direction provides a precise basis for implementing different types of coordinated interventions.

[0061] Step A40: Based on the deviation direction, generate a guide damping signal for the semi-active drive unit that is opposite to the deviation direction.

[0062] It should be noted that a semi-active drive unit refers to a component in an exoskeleton that typically cannot actively output power but can quickly adjust damping (such as magnetorheological fluid or electrorheological fluid dampers). The guide damping signal is a control command used to increase the motion resistance of the semi-active unit in the direction opposite to the deviation. The purpose of this step is to fine-tune and guide the movement by setting flexible resistance when the user's movement rhythm deviates, rather than directly pushing or pulling.

[0063] In one embodiment, generating a guide damping signal for the semi-active drive unit that is opposite to the deviation direction based on the deviation direction includes: determining the normal deviation amplitude based on the magnitude of the projection of the motion vector deviation onto the normal direction of the target motion trajectory; determining the amplitude of the target damping torque based on the normal deviation amplitude; obtaining the target damping torque corresponding to the normal deviation amplitude based on a preset magnetorheological fluid model; calculating the target control current value based on the torque-current characteristic relationship between the target damping torque and the magnetorheological damper in the semi-active drive unit; and generating the guide damping signal based on the target control current value to drive the magnetorheological damper to generate a guide damping torque opposite to the normal deviation direction.

[0064] It should be noted that the core of the guiding damping signal is to control the input current I of the magnetorheological damper, and correspondingly, the output torque τ of the magnetorheological damper... MR With current I and joint angular velocity θ dot The relationship can be modeled as follows: τ MR =τy(I)*sgn(θ dot )+c*θ dot ; Where τy(I) represents the yield stress of the magnetorheological fluid, and its value increases monotonically with the increase of the control current I. The functional relationship between the two can be determined by experimental calibration or by existing magnetorheological fluid constitutive models (such as the Bingham plasticity model); sgn() is a sign function, indicating that the direction of the damping torque is always opposite to the direction of motion (angular velocity direction); c is the fluid viscosity damping coefficient.

[0065] Understandably, the system determines the required guiding damping torque τ based on the calculations. damp (Its direction is opposite to the normal deviation velocity vector vN, and its magnitude is related to |vector vN| and dN). Combining the above model, the required control current I is solved in reverse. Since the magnetorheological effect can be completed within <10ms, this enables the system to achieve millisecond-level tactile feedback. The physical correction force has already been applied the instant the user perceives the deviation in movement.

[0066] In one embodiment, before generating the hybrid collaborative drive signal, the motion vector deviation further includes: extracting the average historical motion vector deviation of the user during the training period based on a preset number of learning periods and training data records; calculating the corresponding fading adjustment coefficient based on the average historical motion vector deviation and its changing trend; adjusting the gain amplitude of the basic assist signal acting on the active drive unit according to the fading adjustment coefficient; and synchronously adjusting... The intervention threshold amplitude used to trigger the guiding damping signal is wherein the adjustment method of the gain amplitude of the basic assist signal is negatively correlated with the adjustment method of the intervention threshold amplitude of the guiding damping signal.

[0067] Understandably, the core of the fading adjustment mechanism lies in a fading adjustment coefficient λ, whose value ranges from 0 to 1. This coefficient is dynamically updated based on the user's historical performance to achieve a smooth decay of the assistance level. The average deviation of the user's historical motion vectors within the most recent training cycle is defined as δ. avg The system is based on δ avg The magnitude of λ and its tendency to converge are updated according to preset rules, such as using a decreasing function: λ new =λ old *f(δ avg ), where f(δ) avg )<1.

[0068] This coefficient λ also directly affects two other key parameters: Active boost gain: The amplitude A of the basic boost signal acting on the active drive unit (motor) assist Multiply by λ, i.e., A assist ∝λ. As λ decreases, the assist weakens.

[0069] Intervention threshold: The primary intervention threshold C that triggers guided damping threshold It is negatively correlated with λ and can be represented as C threshold ∝(1-λ). When λ decreases (user skill level increases), C threshold The restrictions have also tightened, with the system demanding higher standards from users in their actions.

[0070] Based on the value of λ, the training can be clearly divided into three stages: Fully Assisted Stage (0.8≤λ≤1.0): Provides high assistance and a wide tolerance range, helping users develop an initial sense of movement.

[0071] Semi-assisted stage (0.3≤λ<0.8): The assistance is halved, the user needs to actively exert force, and the system only intervenes when there is a large deviation.

[0072] Autonomous stage (0≤λ<0.3): Assistance is almost zero, users have complete autonomy, and the system only provides security protection.

[0073] It should be noted that the normal deviation amplitude refers to the absolute value of the vertical deviation of the motion path calculated in the preceding steps. It is a quantitative result of the severity of the user's action deviating from the correct path. This amplitude is used as input for calculating the amplitude of the target damping torque, meaning that the magnitude of the guiding damping force that the system expects to generate is directly proportional to the degree of path deviation. That is, the greater the deviation, the greater the guiding resistance that needs to be applied to achieve a more significant corrective effect.

[0074] It is important to emphasize that this system employs a magnetorheological damper as the actuator of the semi-active drive unit. The preset magnetorheological fluid model is a mathematical model describing the relationship between the shear stress of this smart material and the applied magnetic field strength. Based on the calculated target damping torque amplitude, the system consults this model to determine the required magnetic field strength to achieve this torque. The resulting "guided damping signal" is essentially an electrical signal containing the target control current value. When this signal is applied to the electromagnetic coil of the magnetorheological damper, it instantly changes the rheological properties of the internal magnetorheological fluid, generating a damping torque opposite to the normal deviation direction. For example, if the user's arm deviates outward, this damping torque will manifest as a resistance towards the inward trajectory, gently prompting and assisting the user to correct the movement trajectory inward. It should be understood that this method of generating damping torque based on precise current control allows for rapid and smooth fine-tuning intervention, providing a more natural and human-like tactile guidance, effectively promoting the establishment of the user's proprioception and the self-correction of movement patterns.

[0075] In one embodiment, the hybrid-driven motion skill training method further includes: identifying the motion stage based on the acceleration mutation characteristics of the main drive joint and a preset motion stage threshold to obtain the motion stage at the current moment; determining the associated joint that needs posture locking based on a preset joint mechanics coupling model and the motion stage; generating an angle locking damping signal for the semi-active drive unit of the joint based on the angle value that the associated joint needs to maintain; and controlling the damper of the associated joint to output a high damping torque to maintain the joint angle based on the angle locking damping signal.

[0076] Understandably, this damping-based guidance is a fine-tuning method that is safer and aligns with human motor intuition. For example, if the system detects that the user's arm is slightly ahead when extended forward, it applies a perceptible damping force in the direction of the forward movement, prompting the user to slow down. The user needs to exert a little more force to overcome this resistance, thus autonomously adjusting the movement rhythm. This approach returns the final control of adjustment to the user, promoting neuromuscular proprioceptive learning and active control. This combination of macroscopic path correction driven by active stimulation and microscopic rhythm adjustment guided by semi-active damping constitutes the hybrid synergistic drive described above, achieving intelligent training assistance through human-computer interaction.

[0077] Step S40: Control the exoskeleton system according to the hybrid collaborative drive signal to assist the user in performing skill training actions.

[0078] It's important to note that controlling the exoskeleton system involves a sophisticated servo control loop. The received hybrid drive signals are first decoupled into specific control commands for the exoskeleton's joint actuators. These commands may involve setpoints for current, voltage, position, or torque. The actuator responses are detected in real-time by force sensors, encoders, etc., and compared with the command values. A closed-loop control algorithm ensures that the output force or motion matches the expected result. For example, when the system determines that the user's arm lifting strength is insufficient, the generated drive signal will cause the exoskeleton's shoulder and elbow joint motors to output auxiliary torque, helping the user smoothly complete the lifting movement. Simultaneously, if the system detects unnecessary internal rotation deviation in the wrist, it will control the corresponding joint to apply a slight corrective torque, thus achieving a hybrid synergistic effect of assistance and correction.

[0079] Understandably, the ideal assistive effect should be gentle and timely, allowing users to feel natural guidance rather than harsh pulling. This requires the system to have extremely low response latency, dynamic characteristics that match human movement, and to avoid generating additional burden or discomfort. Ultimately, through this intelligent physical assistance, the system aims to lower the learning threshold for users and accelerate the formation of correct movement patterns.

[0080] This embodiment generates motion vector deviation by collecting the user's actual motion data and comparing it with the target trajectory. When the deviation exceeds an adaptive threshold, the system decomposes the deviation into the normal and tangential directions of the trajectory: excessive normal deviation triggers the active drive unit to provide basic assistance to correct the path; tangential timing deviation is fine-tuned by applying reverse damping through the semi-active drive unit. Finally, the exoskeleton is controlled collaboratively through hybrid signals to achieve personalized guidance of the user's movements.

[0081] This solution employs a hybrid, synergistic intervention strategy combining normal assist and tangential damping. This provides effective support when users exhibit significant path deviations while also allowing for flexible fine-tuning of their movement rhythm, avoiding the over- or under-intervention problems associated with single control methods. Combined with an adaptive error threshold based on learning progress, the assistance level is matched to the user's ability. This ensures training safety while promoting active participation of the user's proprioception and autonomous optimization of movement patterns, ultimately improving the efficiency and quality of skill training.

[0082] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 Step S20 in the hybrid-driven motion skill training method includes steps S201 to S205: Step S201: Based on the actual motion data, obtain the spatial position coordinates and joint angles of each joint of the user at the current moment.

[0083] It should be noted that the preset target motion trajectory refers to the ideal motion model that has been implanted into the system before the training begins and has been defined in a standardized manner. It fully describes the complete sequence of the ideal spatial positions of each joint of the human body and the ideal angles of each joint as they evolve over time during the execution of standard motion skills. The concept of time synchronization is emphasized in relation to the current moment, that is, the system needs to accurately extract the frame of standard posture data that is in the same phase or at the same moment as the user's actual action from the long target trajectory data stream. The target spatial position coordinates and target joint angles respectively represent the standard three-dimensional spatial position that each joint should be in and the standard angle value that the joint should reach at that ideal time point.

[0084] Step S202: Based on the preset target motion trajectory, obtain the target spatial position coordinates and target joint angles of each joint corresponding to the current moment.

[0085] Understandably, achieving precise "correspondence" is the core technology of this step. This is usually not a simple timestamp matching process, because different users have inherent differences in their movement speed. The system may need to use intelligent algorithms such as dynamic time warping to non-linearly align the timing of the user's actual movements with the timing of the preset target trajectory, ensuring that the comparison is performed at the same movement stage rather than at the exact same physical time. For example, during squat training, even if the user squats slower than the standard template, the system can intelligently determine that their current posture corresponds to the "halfway down" phase in the target trajectory, and extract the target angles and spatial positions of the hip and knee joints at that phase, thus achieving a truly meaningful comparison at the same stage.

[0086] It should be understood that this step is crucial for providing an accurate reference frame for subsequent calculations of motion vector deviations. The acquired target spatial coordinates and target joint angles establish clear and dynamically changing benchmarks for the user's actual motion performance. Their accuracy directly affects the fairness and effectiveness of deviation calculations, while the intelligence of achieving dynamic correspondence determines whether the system can adapt to the individualized movement rhythms of different users and avoid misjudgments due to speed differences. Only by obtaining ideal standard values ​​that truly match the user's current movement state can subsequent deviation analysis and intelligent intervention be targeted, guiding the user towards corrective and learning movements in a standardized manner.

[0087] Step S203: Obtain the position vector deviation based on the spatial position coordinates and the target spatial position coordinates.

[0088] It's important to note that position vector deviation refers to the difference vector calculated by comparing the actual spatial coordinates of the user's joints with the target spatial coordinates. This vector not only includes the magnitude of the position deviation but also specifies its direction. For example, in an arm raise movement, this deviation can be quantified as the specific difference between the actual wrist position and the ideal position in three dimensions: front-back, left-right, and up-down. The subsequent motion posture vector deviation describes the difference obtained by comparing the currently measured joint angle with the target joint angle. It focuses on the correctness of the joint's bending or rotational state. For example, if the actual bending angle of the knee joint is less than the standard angle, it indicates insufficient squat depth; this deviation vector characterizes this postural irregularity.

[0089] Taking a typical and demanding skill like a golf swing as an example, the implementation process of this solution can be described numerically as follows. Key data monitored by the system includes: Position vector deviation δpos-impact: This refers to the deviation of the actual three-dimensional position (xa, ya, za) of the clubhead relative to the ideal model position (xt, yt, zt) at the critical moment of impact. Its magnitude is calculated as follows: δpos-impact=sqrt[(xa-xt)²+(ya-yt)²+(za-zt)²].

[0090] Posture vector deviation δang-separation: During the swing, multiple core joint angles are continuously monitored. For example, at the top of the backswing, the difference between the trunk rotation angle θtrunk-actual and the pelvic rotation angle θpelvis-actual, i.e., the trunk-pelvic separation angle, is a key biomechanical indicator. Its posture deviation is: δang-separation=|(θtrunk-actual-θpelvis-actual)-(θtrunk-template-θpelvis-template)|, where the template separation angle is usually 45±5°.

[0091] Furthermore, global weighting factors for positional and angular deviations, namely α and β, can be set, satisfying α + β = 1. Their initial values ​​are determined by the characteristics of the target action. The weighting factors α and β can be dynamically adjusted according to the user's training stage: during the basic learning period, α > β can be set to guide the user to follow the macroscopic spatial trajectory; during the action refinement period, α < β can be set to strictly correct the microscopic configuration of the joint posture. The adjustment strategy can be based on a preset expert knowledge base, statistical analysis of historical training data, or adaptively generated through reinforcement learning algorithms.

[0092] Understandably, based on the golf swing motion described in the aforementioned embodiments, there is a training consensus that "posture takes precedence over absolute position." The system's preset fusion weights for this motion are biased towards posture correction, for example, setting β=0.7 and α=0.3. When the calculated overall deviation δtotal exceeds the swing quality threshold calibrated based on data from professional athletes, the system determines that the motion is not standard and immediately triggers a composite guidance mode of hip, shoulder, and wrist dampers, simulating the tactile cues of a professional coach to "suppress excessive body rotation" or "maintain wrist angle."

[0093] Step S204: Obtain the motion posture vector deviation based on the current joint angle and the target joint angle.

[0094] It's important to note that motion posture vector deviation refers to a parameter that characterizes the degree to which a limb's shape deviates from a standard posture. This is achieved by quantitatively comparing a specific joint angle measured by the user at a given moment with the corresponding ideal joint angle in a preset target trajectory. Joint angles reflect the relative orientation between connected bone segments; for example, the elbow angle describes the degree of flexion of the forearm relative to the upper arm. Therefore, this deviation represents the difference between the local limb configuration and the standard model during movement execution. It focuses on the correctness of the relative positions of various body parts, rather than the absolute spatial positions of joints.

[0095] Understandably, the core of calculating motion posture vector deviation lies in the precise measurement of joint rotation or flexion-extension states. Unlike position vector deviation, which describes spatial displacement, posture deviation focuses more on the intrinsic morphological quality of the movement. For example, during squat training, even if the exerciser's hip and knee are in near-standard positions in space, their knee joint may exhibit abnormal varus or valgus angles. Such morphological errors cannot be fully reflected by position deviation but can be keenly captured by motion posture vector deviation. This deviation is usually a multi-dimensional vector that can simultaneously reflect angular abnormalities in multiple joints or multiple rotational degrees of freedom.

[0096] It should be understood that obtaining accurate motion posture vector deviation is crucial for motion quality assessment and correction. It complements position vector deviation, together depicting the gap between the user's movement and the ideal template from both macroscopic displacement and microscopic morphological perspectives. Precise calculation of this deviation provides a direct basis for generating targeted posture correction instructions in subsequent steps, enabling the assistance system not only to guide the user to the correct position but also to instruct them to complete the movement in the correct manner, thereby achieving truly standardized motor skill learning.

[0097] Step S205: Obtain the motion vector deviation based on the position vector deviation and the motion attitude vector deviation.

[0098] It should be noted that position vector deviation reveals the linear deviation between the actual position of a limb's end or joint in three-dimensional space and its desired target position, while motion posture vector deviation describes the difference between the relative angular relationships between various body parts and the standard posture. The synthesis process here aims to construct a composite index that can comprehensively and without omission characterize the overall gap between the performance of the entire movement and the ideal state.

[0099] Understandably, this synthesis process typically requires prioritizing and weighting based on the specific learning objectives of the motor skills. For example, in tasks emphasizing precise displacement, position vector deviation may be given higher weight; while in other tasks focusing on proper body posture, motion posture vector deviation is given higher weight. The system uses pre-defined or adaptive algorithms, such as weighted fusion or multi-state-space-based synthesis methods, to unify these two physically distinct but equally important deviations into a coordinated evaluation framework, thereby forming a more informative and instructive comprehensive deviation vector.

[0100] It should be understood that the final obtained motion vector deviation provides a unified quantitative standard from the perspective of overall motion performance. This composite deviation enables assistive devices to understand the full extent of user non-standard movements, and thus may be able to simultaneously apply synergistic assistive forces that promote position correction and posture adjustment, achieving comprehensive and refined guidance from spatial trajectory to body configuration, ultimately improving the effectiveness and safety of training or work.

[0101] In this embodiment, user motion data is captured in real time and simultaneously matched with standard postures in the target trajectory. The position vector deviation, representing spatial displacement errors, and the motion posture vector deviation, representing body shape errors, are calculated separately and then fused into a comprehensive motion vector deviation. By calculating and fusing position and posture deviations separately, this scheme overcomes the limitations of single-dimensional evaluation, enabling a refined assessment of both the macroscopic trajectory and microscopic configuration of the movement. This multi-angle, comprehensive deviation measurement method allows the system to generate more accurate correction instructions, effectively guiding users to correct various errors in their movements, ultimately improving the standardization of motor skill learning and the overall effectiveness of training.

[0102] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the hybrid-driven motion skill training method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0103] This application also provides a hybrid-driven motion skill training device; please refer to [reference needed]. Figure 3 The hybrid-driven motion skill training device includes: Data acquisition module 10 is used to acquire actual movement data of the user's limbs; The deviation calculation module 20 is used to calculate the motion vector deviation based on the actual motion data and the preset target motion trajectory. The signal generation module 30 is used to generate a hybrid cooperative drive signal based on the motion vector deviation if the motion vector deviation exceeds a preset vector error boundary range. The hybrid cooperative drive signal includes a basic assist signal for the active drive unit and a guide damping signal for the semi-active drive unit that is opposite to the direction of the motion vector deviation. The control execution module 40 is used to control the exoskeleton system according to the hybrid collaborative drive signal to assist the user in performing skill training actions.

[0104] The hybrid-driven motion skill training device provided in this application, employing the hybrid-driven motion skill training method described in the above embodiments, solves the technical problem in the prior art of how to design a device that can simultaneously achieve high-transparency motion following and high-precision trajectory correction in high-dynamic skill training. Compared with the prior art, the beneficial effects of the hybrid-driven motion skill training device provided in this application are the same as those of the hybrid-driven motion skill training method provided in the above embodiments, and other technical features of the hybrid-driven motion skill training device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0105] This application provides a hybrid-driven motion skill training device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the hybrid-driven motion skill training method in the above embodiment 1.

[0106] The following is for reference. Figure 4 This document illustrates a structural schematic diagram of a hybrid-driven motion skill training device suitable for implementing embodiments of this application. The hybrid-driven motion skill training device in this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 4 The hybrid-driven motion skill training device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0107] like Figure 4As shown, the hybrid-driven motion skills training device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the hybrid-driven motion skills training device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the hybrid-driven motion skills training device to communicate wirelessly or wiredly with other devices to exchange data. Although a hybrid-driven motion skills training device with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0108] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0109] The hybrid-driven motion skill training device provided in this application, employing the hybrid-driven motion skill training method described in the above embodiments, solves the technical problem in the prior art of how to design a device that can simultaneously achieve high-transparency motion following and high-precision trajectory correction in high-dynamic skill training. Compared with the prior art, the beneficial effects of the hybrid-driven motion skill training device provided in this application are the same as those of the hybrid-driven motion skill training method provided in the above embodiments, and other technical features of this hybrid-driven motion skill training device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0110] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0111] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0112] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the hybrid-driven motion skill training method in the above embodiments.

[0113] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), or any suitable combination thereof.

[0114] The aforementioned computer-readable storage medium may be included in a hybrid-driven motion skills training device; or it may exist independently and not be assembled into a hybrid-driven motion skills training device.

[0115] The aforementioned computer-readable storage medium carries one or more programs that, when executed by a hybrid-driven motion skill training device, cause the hybrid-driven motion skill training device to: acquire actual motion data of the user's limbs; obtain a motion vector deviation based on the actual motion data and a preset target motion trajectory; if the motion vector deviation exceeds a preset vector error boundary range, generate a hybrid collaborative drive signal based on the motion vector deviation, wherein the hybrid collaborative drive signal includes a basic assist signal for the active drive unit and a guide damping signal for the semi-active drive unit that is opposite in direction to the motion vector deviation; and control the exoskeleton system based on the hybrid collaborative drive signal to assist the user in performing skill training actions.

[0116] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0117] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0118] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0119] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described hybrid-driven motion skill training method. This solves the technical problem in the prior art of designing a method that simultaneously achieves high-transparency motion following and high-precision trajectory correction in high-dynamic skill training. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the hybrid-driven motion skill training method provided in the above embodiments, and will not be repeated here.

[0120] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the hybrid-driven motion skill training method described above.

[0121] The computer program product provided in this application solves the technical problem in the prior art of how to design a method that can simultaneously achieve high-transparency motion following and high-precision trajectory correction in high-dynamic skill training. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the hybrid-driven motion skill training method provided in the above embodiments, and will not be repeated here.

[0122] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for training motor skills based on hybrid driving, characterized in that, The hybrid-driven motion skill training method includes: Acquire actual movement data of the user's limbs; The motion vector deviation is obtained based on the actual motion data and the preset target motion trajectory; If the motion vector deviation exceeds the preset vector error boundary range, a hybrid cooperative drive signal is generated based on the motion vector deviation. The hybrid cooperative drive signal includes a basic assist signal for the active drive unit and a guide damping signal for the semi-active drive unit that is opposite in direction to the motion vector deviation. The exoskeleton system is controlled according to the hybrid collaborative drive signal to assist the user in performing skill training actions.

2. The hybrid-driven motion skill training method according to claim 1, characterized in that, The step of obtaining the motion vector deviation based on the actual motion data and the preset target motion trajectory includes: Based on the actual motion data, obtain the spatial position coordinates and joint angles of each joint of the user at the current moment; Based on the preset target motion trajectory, obtain the target spatial position coordinates and target joint angles of each joint corresponding to the current moment; The position vector deviation is obtained based on the spatial position coordinates and the target spatial position coordinates; The motion posture vector deviation is obtained based on the current joint angle and the target joint angle. The motion vector deviation is obtained based on the position vector deviation and the motion attitude vector deviation.

3. The hybrid-driven motion skill training method according to claim 1, characterized in that, If the motion vector deviation exceeds a preset vector error boundary range, before generating the hybrid cooperative drive signal based on the motion vector deviation, the method further includes: Obtain the current user's historical learning duration and historical movement deviation data; An evaluation coefficient for motion stability is obtained by evaluating the historical learning duration and historical motion deviation data. The vector error boundary range is obtained based on the motion stability evaluation coefficient and the preset scaling mapping rule.

4. The hybrid-driven motion skill training method according to claim 1, characterized in that, The step of generating a hybrid cooperative drive signal based on the motion vector deviation includes: The degree of motion deviation is determined based on the magnitude of the projection of the motion vector deviation onto the normal direction of the preset target motion trajectory; Based on the comparison between the degree of motion deviation and the preset primary intervention threshold, a basic assist signal for the active drive unit is generated. The direction of deviation is determined based on the projection direction of the motion vector deviation onto the tangent of the target motion trajectory; Based on the deviation direction, a guide damping signal opposite to the deviation direction is generated for the semi-active drive unit.

5. The hybrid-driven motion skill training method according to claim 4, characterized in that, The step of generating a guide damping signal for the semi-active drive unit, which is opposite to the deviation direction, based on the deviation direction includes: The magnitude of the normal deviation is determined based on the projection of the motion vector deviation onto the normal of the target motion trajectory. The magnitude of the target damping torque is determined based on the magnitude of the normal deviation. Based on the preset magnetorheological fluid model, the target damping torque corresponding to the normal deviation amplitude is obtained; The target control current value is calculated based on the relationship between the target damping torque and the torque-current characteristic of the magnetorheological damper in the semi-active drive unit. Based on the target control current value, the guiding damping signal is generated to drive the magnetorheological damper to generate a guiding damping torque opposite to the normal deviation direction.

6. The hybrid-driven motion skill training method according to claim 4, characterized in that, Before generating the hybrid collaborative drive signal, the motion vector deviation further includes: Based on the preset number of learning periods and training data records, the average historical motion vector deviation of the user within the training period is extracted; Based on the average deviation of the historical motion vector and the trend of the average deviation, the corresponding fade-out adjustment coefficient is calculated. The gain amplitude of the basic assist signal acting on the active drive unit is adjusted according to the fade-out adjustment coefficient. Based on the fade-out adjustment coefficient, the intervention threshold amplitude used to trigger the guide damping signal is adjusted synchronously, wherein the adjustment method of the gain amplitude of the basic assist signal is negatively correlated with the adjustment method of the intervention threshold amplitude of the guide damping signal.

7. The hybrid-driven motion skill training method according to any one of claims 1 to 6, characterized in that, The hybrid-driven motion skill training method further includes: The motion stage is identified based on the acceleration change characteristics of the main drive joint and the preset motion stage threshold to obtain the motion stage at the current moment. Based on the preset joint biomechanical coupling model and the aforementioned action phase, determine the associated joints that require posture locking; Based on the angle value that the associated joint needs to maintain, an angle locking damping signal is generated for the semi-active drive unit of the joint. Based on the angle-locked damping signal, the damper of the associated joint is controlled to output a high damping torque to maintain the joint angle.

8. A motion skill training device based on hybrid drive, characterized in that, The device includes: The data acquisition module is used to acquire actual movement data of the user's limbs; The deviation calculation module is used to calculate the motion vector deviation based on the actual motion data and the preset target motion trajectory. The signal generation module is used to generate a hybrid cooperative drive signal based on the motion vector deviation if the motion vector deviation exceeds a preset vector error boundary range. The hybrid cooperative drive signal includes a basic assist signal for the active drive unit and a guide damping signal for the semi-active drive unit that is opposite to the direction of the motion vector deviation. The control execution module is used to control the exoskeleton system according to the hybrid collaborative drive signal to assist the user in performing skill training actions.

9. A motion skill training device based on hybrid drive, characterized in that, The hybrid-driven motion skill training device includes: a memory, a processor, and a hybrid-driven motion skill training program stored in the memory and executable on the processor, wherein the hybrid-driven motion skill training program is configured to implement the steps of the hybrid-driven motion skill training method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a hybrid-driven motion skill training program, which, when executed by a processor, implements the steps of the hybrid-driven motion skill training method as described in any one of claims 1 to 7.

Images