CN-122008235-A - Method, device, equipment and storage medium for training action skills based on hybrid drive

CN122008235ACN 122008235 ACN122008235 ACN 122008235ACN-122008235-A

Abstract

The invention belongs to the technical field of exoskeleton robots, and discloses a hybrid-drive-based action skill training method, a hybrid-drive-based action skill training device, hybrid-drive-based action skill training equipment and a hybrid-drive-based storage medium. The method generates deviation information by comparing the action of a user with a target track in real time, cooperatively controls the active driving unit to provide basic assistance according to the deviation information, instructs the semi-active driving unit to generate guiding damping force opposite to the deviation direction, and has a fading mechanism for adjusting the assistance level according to learning progress, and can lock the dynamic posture of the associated joint in a key stage of the cooperative action of multiple joints. According to the scheme, vector damping guidance aiming at movement deviation is adopted, so that a user moves smoothly when the movement is correct, clear physical guidance can be obtained when the movement is deviated, the auxiliary learning effect of the user in the movement of the limbs with high difficulty is remarkably improved, the normalization and stability of the core link in the complex movement are ensured, and finally intelligent and bionic high-level skill teaching is realized.

Inventors

LIU ZHEN
YE PENG
WANG FENG

Assignees

深圳陨墨星人工智能有限公司

Dates

Publication Date: 20260512
Application Date: 20260331

Claims (10)

1. A method for training action skills based on hybrid driving is characterized in that, the method for training the action skills based on the hybrid drive comprises the following steps: acquiring actual motion data of limbs of a user; Obtaining motion vector deviation according to the actual motion data and a preset target motion track; If the motion vector deviation exceeds a preset vector error boundary range, generating a mixed cooperative driving signal according to the motion vector deviation, wherein the mixed cooperative driving signal comprises a basic power-assisted signal used for an active driving unit and a guiding damping signal used for a semi-active driving unit and opposite to the motion vector deviation; and controlling the exoskeleton system according to the mixed cooperative driving signal so as to assist the user to perform skill training actions.
2. The hybrid drive-based motor skill training method according to claim 1, wherein the obtaining a motion vector deviation according to the actual motion data and a preset target motion trajectory comprises: according to the actual motion data, acquiring the space position coordinates and joint angles of all joints of a user at the current moment; acquiring target space position coordinates and target joint angles of all joints corresponding to the current moment according to the preset target motion trail; Obtaining a position vector deviation according to the space position coordinate and the target space position coordinate; obtaining a motion attitude vector deviation according to the joint angle at the current moment and the target joint angle; and obtaining the motion vector deviation according to the position vector deviation and the motion gesture vector deviation.
3. The hybrid drive-based motor skills training method of claim 1, wherein if the motion vector deviation exceeds a preset vector error boundary range, before generating a hybrid cooperative drive signal according to the motion vector deviation, further comprising: Acquiring historical learning duration and historical motion deviation data of a current user; evaluating according to the historical learning duration and the historical motion deviation data to obtain an action stability evaluation coefficient; And obtaining the vector error boundary range according to the motion stability evaluation coefficient and a preset scaling mapping rule.
4. The hybrid drive-based motor skills training method of claim 1, wherein generating a hybrid co-drive signal from the motion vector bias comprises: determining the degree of motion deviation according to the upward projection size of the motion vector deviation on a preset target motion track method; Generating a basic power-assisted signal for the active driving unit according to a comparison result of the motion deviation degree and a preset primary intervention threshold value; Determining a deviation direction according to the projection direction of the motion vector deviation in the tangential direction of the target motion track; according to the deviation direction, a pilot damping signal for the semi-active drive unit is generated opposite to the deviation direction.
5. The hybrid drive-based motor skills training method of claim 4, wherein the generating a pilot damping signal for a semi-active drive unit opposite the bias direction as a function of the bias direction comprises: determining a normal deviation amplitude according to the projection size of the motion vector deviation in the normal direction of the target motion track; determining the amplitude of the target damping moment according to the normal deviation amplitude; obtaining a target damping moment corresponding to the normal deviation amplitude according to a preset magnetorheological fluid model; calculating to obtain a target control current value according to the relation between the target damping moment and the moment-current characteristic of the magneto-rheological damper in the semi-active driving unit; and generating the guide damping signal according to the target control current value so as to drive the magnetorheological damper to generate guide damping moment opposite to the normal deviation direction.
6. The hybrid drive-based motor skills training method of claim 4, wherein prior to generating the hybrid co-drive signal, the method further comprises: according to the preset learning period number and training data record, extracting and obtaining a historical motion vector deviation mean value of a user in a training period; Calculating to obtain a corresponding fading-out adjusting coefficient according to the historical motion vector deviation mean value and the change trend of the mean value; according to the fading-out adjusting coefficient, adjusting the gain amplitude of a basic power-assisted signal acting on the active driving unit; And synchronously adjusting an intervention threshold amplitude for triggering the guiding damping signal according to the fading-out adjusting coefficient, wherein the adjustment mode of the gain amplitude of the basic power-assisted signal is inversely related to the adjustment mode of the intervention threshold amplitude of the guiding damping signal.
7. A hybrid drive-based motor skills training method according to any of claims 1 to 6, further comprising: performing action phase identification according to the acceleration mutation characteristic of the main driving joint and a preset action phase threshold value to obtain an action phase at the current moment; determining an associated joint needing posture locking according to a preset joint mechanical coupling model and the action stage; Generating an angle locking damping signal for the joint semi-active driving unit according to the angle value to be maintained of the associated joint; and controlling the damper of the related joint to output high damping moment for maintaining the joint angle according to the angle locking damping signal.
8. A hybrid drive-based motor skills training apparatus, the apparatus comprising: the data acquisition module is used for acquiring actual movement data of the limbs of the user; the deviation calculation module is used for calculating and obtaining a motion vector deviation according to the actual motion data and a preset target motion track; The signal generation module is used for generating a mixed cooperative driving signal according to the motion vector deviation if the motion vector deviation exceeds a preset vector error boundary range, wherein the mixed cooperative driving signal comprises a basic power-assisted signal used for an active driving unit and a guiding damping signal used for a semi-active driving unit and opposite to the motion vector deviation direction; And the control execution module is used for controlling the exoskeleton system according to the mixed cooperative driving signal so as to assist a user to execute skill training actions.
9. A hybrid drive-based motor skills training apparatus comprising a memory, a processor and a hybrid drive-based motor skills training program stored on the memory and executable on the processor, the hybrid drive-based motor skills training program being configured to implement the steps of the hybrid drive-based motor skills training method of any of claims 1 to 7.
10. A storage medium having stored thereon a hybrid drive-based motor skill training program which, when executed by a processor, implements the steps of the hybrid drive-based motor skill training method of any of claims 1 to 7.

Description

Method, device, equipment and storage medium for training action skills based on hybrid drive Technical Field The application relates to the technical field of exoskeleton robots, in particular to a hybrid drive-based action skill training method, a hybrid drive-based action skill training device, hybrid drive-based action skill training equipment and a hybrid drive-based storage medium. Background With the rapid development of robotics, the wearable exoskeleton system has achieved significant results in the fields of medical rehabilitation and industrial assistance. Such systems are typically motor driven to provide support for patients with impaired mobility or to reduce weight on workers, with the central objective of performing basic locomotion functions or providing strength enhancement. However, this technological surge is mainly focused on "functional replacement" or "strength assistance", whose design logic is essentially different from the need to pursue motor skills training with precision, coordination, and efficiency of motion. In particular to the field of high-order action skill training such as sports competition, dance art and the like, the traditional training mode is highly dependent on experience observation and instant feedback of a coach. However, this manual guidance approach has inherent limitations such as feedback lag, difficulty in quantification, inability to provide real-time physical guidance, etc. To overcome these limitations, some research attempts to introduce exoskeleton technology into skill training, but existing solutions still suffer from significant drawbacks when directed to high-dynamic, high-precision skill teaching. The prior art (such as CN108309689B 1) has proposed the concept of "progressive" training, but its core is based on macroscopic phase division of the rehabilitation cycle, e.g. dividing the training into early, mid and late phases of rehabilitation, and employing position control, torque control or resistance modes at different phases. The progressive control cannot realize millisecond-level real-time dynamic adjustment in the single action execution process, and secondly, the control logic and hardware design of the equipment mainly aims at the slow rehabilitation of patients with impaired muscle strength, the adaptability to high-speed burst actions (such as kicking) is lacking, and the high-inertia driving system can obstruct the fluency of the actions. Another prior art (e.g., CN110303471B 1) focuses on providing "comfort assistance" through motion intent recognition and variable impedance control, with the goal of following the user's intent and reducing user energy consumption, but this is contrary to the core needs of "error correction teaching" in skill training. The present invention aims to impose accurate physical constraints for "error correction" when the user action deviates from the ideal trajectory, i.e. "the system prevents the user from going to the wrong direction". However, this prior art does not disclose how to achieve such a direction sensitive, fast responding "directional virtual wall" type correction. Furthermore, some solutions (such as CN116036538 A1) use the principles of electromagnetic eddy current to construct passive dampers for strength training. The device is purely passive, can only provide resistance related to speed, can not provide any active assistance when the action is started or gravity needs to be overcome, has the function limited to muscle-increasing load, does not relate to guidance and teaching of action tracks, and can not realize cooperation and seamless switching between 'active assistance' and 'passive correction'. Therefore, how to design a trajectory correction device capable of simultaneously realizing high-transparency motion following and high-precision trajectory correction in high dynamic skill training is a technical problem to be solved in the art. Disclosure of Invention The application mainly aims to provide a method, a device, equipment and a storage medium for training action skills based on hybrid driving, which aim to solve the technical problem of how to design a method for realizing action following with high transparency and high-precision track deviation correction in high dynamic skill training in the prior art. To achieve the above object, the present application provides a method for training motor skills based on hybrid driving, the method comprising: acquiring actual motion data of limbs of a user; Obtaining motion vector deviation according to the actual motion data and a preset target motion track; If the motion vector deviation exceeds a preset vector error boundary range, generating a mixed cooperative driving signal according to the motion vector deviation, wherein the mixed cooperative driving signal comprises a basic power-assisted signal used for an active driving unit and a guiding damping signal used for a semi-active driving unit and opposite to the motion vector devi