Humanoid model POSE action synchronization method, system and device and storage medium

Abstract

The invention relates to the technical field of robots, and discloses a mannequin POSE action synchronization method and system, and the method comprises the following steps: S1, obtaining action instruction data: obtaining the action instruction data containing target positions and motion tracks of a plurality of joints of a mannequin through an input device; s2, the current posture is monitored in real time, wherein the posture and the environment of the mannequin are monitored in real time; s3, joint control is achieved, specifically, based on an intelligent joint controller, a closed-loop control system and a servo motor are used for controlling a plurality of joints of the mannequin; s4, action adjustment: dynamically adjusting action parameters of the joints through a self-adaptive control algorithm according to attitude data and external environment changes which are monitored and fed back in real time; and S5, data processing and feedback: processing feedback data from the multi-sensor system and the joint control module in parallel in real time. By adopting a high-performance servo motor and a closed-loop control system, each joint of the humanoid model can be accurately controlled, and the accuracy and natural fluency of actions are ensured.

Description

Technical Field

[0001] The present invention relates to the field of robotics, and particularly to a method, system, device, and storage medium for synchronizing the POSE actions of a humanoid model. Background Art

[0002] In modern clothing displays, fashion photography, and other related fields, traditional model displays rely on live models to demonstrate actions and postures. However, the use of live models has many limitations in terms of time, location, cost, and availability. To overcome these limitations, with the progress of technology, mechanically or electronically controlled humanoid models have been gradually developed and applied to simulate the actions and postures of real people. However, the existing humanoid model technology still has obvious deficiencies in terms of natural smoothness, synchronization, and the presentation of complex postures.

[0003] When the existing technology realistically simulates the actions and postures of live models, especially in dynamic scenarios such as clothing displays, it is difficult to balance high-precision action synchronization, natural and smooth posture transformation, and the stable presentation of complex POSEs (the postures or action shapes of models). In addition, the existing systems still have limitations in sensor fusion, real-time feedback, and control accuracy, especially when dealing with complex environmental changes or external interactions, the flexibility and stability of action control are poor. Therefore, developing a method and system for synchronizing the POSE actions of a humanoid model that can overcome these technical bottlenecks has become an urgent need in the current technical field. Summary of the Invention

[0004] The main objective of the present invention is to provide a method, system, device, and storage medium for synchronizing the POSE actions of a humanoid model, aiming to solve the problems of the deficiencies in action synchronization, natural smoothness, and complex POSE performance in the existing technology.

[0005] To achieve the above-mentioned invention objective, the first aspect of the present invention proposes a method for synchronizing the POSE actions of a humanoid model, including the following steps: S1. Obtain action instruction data: Obtain action instruction data containing the target positions and movement trajectories of multiple joints of the humanoid model through an input device; S2. Monitor the current posture in real time: Use a multi-sensor fusion system to monitor the posture and environment of the humanoid model in real time; S3. Implement joint control: Based on an intelligent joint controller, use a closed-loop control system and servo motors to control multiple joints of the humanoid model. The control includes driving the joint actions according to the obtained action instruction data and adjusting the posture in combination with the real-time monitored data; S4. Action adjustment: Dynamically adjust the action parameters of the joints through an adaptive control algorithm according to the posture data feedback in real time and the changes in the external environment; S5. Data Processing and Feedback: Through hardware acceleration technology, the feedback data from the multi-sensor system and the joint control module are processed in parallel in real time, and the processing results are immediately fed back to the intelligent joint controller to dynamically adjust the joint motion parameters.

[0006] Further, the action instruction data includes a preset action sequence or custom action instruction data input by the user; Further, the multi-sensor fusion system includes an accelerometer, a gyroscope, a vision sensor, and a force sensor, and can detect the environmental information around the humanoid model through sensor data fusion technology, specifically including the recognition of obstacles and target objects; Further, the intelligent joint controller is integrated with a self-diagnosis module. The self-diagnosis module detects joint overload or abnormal actions by monitoring the data of the load sensor and the position sensor of the joint in real time, and issues a control signal to adjust or stop the joint action when an abnormality is detected; Further, the adaptive control algorithm dynamically adjusts the motion parameters of the joint, including speed, angle, and acceleration, by comprehensively analyzing the input action instruction data, the sensor feedback data monitored in real time, and the environmental change information; Further, the method further includes using a safety monitoring module to monitor the working state of the joint and adjusting or stopping the joint action when an abnormality occurs.

[0007] In a second aspect of the present invention, a humanoid model POSE action synchronization system is proposed. The system includes: An action capture module: used to acquire and process action instruction data, where the action instruction data includes the target positions and motion trajectories of multiple joints; A multi-sensor fusion module: real-time monitors the posture and external environment of the humanoid model through an accelerometer, a gyroscope, a vision sensor, and a force sensor; An intelligent joint control module: based on the action instruction data, uses a servo motor and a closed-loop control system to achieve the control of multiple joints; An adaptive control module: comprehensively analyzes the input action instruction data, the sensor feedback data monitored in real time, and the environmental change information, and dynamically adjusts the motion parameters of the joint, including speed, angle, and acceleration; A data processing module: adopts a parallel data processing architecture and hardware acceleration technology for real-time parallel processing, and immediately feeds back the processing results to the intelligent joint controller.

[0008] Further, the action capture module can process different formats of action data, including preset action sequences and user-defined action instructions; The multi-sensor fusion module can identify obstacles or target objects in the environment and provide data to the humanoid model POSE motion synchronization system for processing; The intelligent joint control module has a self-diagnosis function and can detect and handle joint overload and abnormal motion conditions; The adaptive control module can adjust the execution mode of joint motions according to historical data and input information; The data processing module adopts a parallel data processing architecture. Through hardware acceleration technology and multi-core parallel computing, it synchronously processes real-time data from multiple sensors and provides dynamic data support during the joint motion adjustment process; The humanoid model POSE motion synchronization system further includes a safety monitoring module: monitoring the operating state of the humanoid model POSE motion synchronization system, and restricting motion execution or entering the safety mode in case of abnormalities.

[0009] In a third aspect, the present invention provides the following technical solution. A computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned humanoid model POSE motion synchronization method is implemented.

[0010] In a fourth aspect, the present invention provides the following technical solution. A readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned humanoid model POSE motion synchronization method is implemented. Beneficial effects

[0011] By adopting high-performance servo motors and a closed-loop control system, the present invention can achieve precise control of each joint of the humanoid model, ensuring the accuracy and natural smoothness of motions. The intelligent joint controller integrates high-precision sensors (position sensors, speed sensors), can monitor the joint state in real time and respond quickly to instructions, significantly improving the speed and stability of motion execution. In addition, the controller has a self-diagnosis function and can automatically adjust motion parameters or stop working when the joint is overloaded or abnormal, effectively preventing equipment damage and improving the safety of the system.

[0012] The present invention combines an accelerometer, a gyroscope, a vision sensor, and a force sensor through a multi-sensor fusion system, improving the accuracy of attitude perception and the reliability of motion control through multi-source data fusion. The accelerometer and gyroscope can accurately monitor the attitude changes of the model, the vision sensor can identify obstacles in the surrounding environment, and the force sensor ensures flexible motion adjustment when the model comes into contact with external objects. Through the fusion of these sensors, the system can obtain the motion state of the joints and external environment information in real time, ensuring synchronization and accuracy during complex motion execution.

[0013] The system of the present invention adopts an adaptive control algorithm based on machine learning, which can dynamically optimize the motion control parameters of joints according to real-time data. By continuously learning user feedback and environmental changes, the adaptive control algorithm can self-adjust during the execution of actions to achieve higher naturalness and smoothness of actions. As the usage time increases, the system will gradually improve its adaptability to user needs and provide more personalized action performances, especially suitable for scenarios requiring high-precision action synchronization, such as clothing displays and model performances.

[0014] The system of the present invention uses an efficient parallel data processing architecture, which can complete the acquisition, processing, and feedback of sensor data within milliseconds, thereby realizing low-latency action execution. Through hardware acceleration technology and multi-core parallel processing, this architecture ensures that multi-sensor data can be processed simultaneously, significantly improving the system response speed. During the execution of actions, the system can quickly adjust joint actions to ensure low-latency action responses even under high loads, enhancing the synchronization and natural smoothness of actions. Brief Description of the Drawings

[0015] Figure 1 is a flowchart of the humanoid model POSE action synchronization method according to the first embodiment of the present invention; Figure 2 is an architecture diagram of the humanoid model POSE action synchronization system according to the second embodiment of the present invention.

[0016] The realization, functional features, and advantages of the object of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Embodiments

[0017] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0018] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. are based on the orientation or positional relationships shown in the drawings. These are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically and clearly defined.

[0019] In the description of the present invention, it should be noted that unless otherwise clearly specified and limited, the terms "mounted", "connected", and "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium, and it can be the communication inside two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0020] In the present invention, unless otherwise clearly specified and limited, the first feature being "above" or "below" the second feature may include the first and second features being in direct contact, or may include the first and second features not being in direct contact but being in contact through additional features therebetween. Moreover, the first feature being "above", "over", and "on" the second feature includes the first feature being directly above and obliquely above the second feature, or merely indicating that the first feature has a higher horizontal height than the second feature. The first feature being "below", "beneath", and "under" the second feature includes the first feature being directly below and obliquely below the second feature, or merely indicating that the first feature has a lower horizontal height than the second feature.

[0021] Referring to Figure 1 , Embodiment 1 of the present invention provides a method for synchronizing the POSE actions of a humanoid model, including the following steps: S1. Obtain action instruction data: Obtain action instruction data including the target positions and movement trajectories of multiple joints of the humanoid model through an input device; The following are the explanations of the terms in step S1: Action instruction data It refers to the instruction information of the specific actions and postures performed by the humanoid model, including the target positions, angles, speeds, and movement trajectories of each joint. This data can be sourced from user input, preset action sequences, or real-time capture by external devices.

[0022] Specific implementation: The user selects action instructions through a graphical interface or specifies that the model execute a specific POSE through voice input. For example, the user can select a standing posture, a walking action, or a rotating action. An external motion capture device (such as a camera with sensors) can capture the motion data of a real human body and convert it into action instructions for the model.

[0023] Data structure: Action instructions usually exist in the form of a time series, and each time point defines the target positions, speeds, and posture changes of each joint of the model. The system stores these instruction data and gradually transmits them to the control module for execution.

[0024] S2. Real-time monitor the current posture: Adopt a multi-sensor fusion system to real-time monitor the posture of the humanoid model and the environment; The following are the explanations of the terms in step S2: Multi-sensor fusion system A system composed of multiple sensors that can simultaneously collect and process data from different sensors to provide information about the current state of the humanoid model and the surrounding environment. Common sensors include: Accelerometer: Measures the acceleration and linear motion of an object.

[0025] Gyroscope: Measures the angular velocity and rotation state of an object.

[0026] Vision sensor: Captures environmental images through a camera to identify obstacles and target objects.

[0027] Force sensor: Measures the force and pressure when the model comes into contact with external objects.

[0028] Multi-sensor fusion system: The system real-time monitors the current posture of the model and external environment information through multiple sensors. The sensors include accelerometers, gyroscopes, vision sensors, force sensors, etc.

[0029] Data fusion: The system uses algorithms such as Kalman filtering to integrate data from multiple sensors to form an overall posture map of the model at present. This data is real-time fed back to the control module to ensure that the system has a comprehensive understanding of the current posture of the model.

[0030] S3. Implement joint control: Based on an intelligent joint controller, use a closed-loop control system and servo motors to control multiple joints of the humanoid model. The control includes driving the joint actions according to the acquired action instruction data and adjusting the posture in combination with the data of real-time monitoring; Specifically, a closed-loop control system: Control mechanism: The closed-loop control system uses a proportional-integral-derivative (PID) control algorithm for real-time adjustment. The PID controller adjusts the proportional, integral, and derivative gains to ensure that the error between the joint movement and the target posture approaches zero in a short time. Specifically, the proportional part (P) is used to reduce the current error, the integral part (I) is used to eliminate the cumulative error, and the derivative part (D) is used to predict the change trend of the error, thus effectively preventing overshoot and oscillation phenomena.

[0031] Enhanced robustness: To cope with non-linear external disturbances, the system also incorporates a robust control algorithm, such as Sliding Mode Control (SMC), to maintain control accuracy when external perturbations are large. Sliding Mode Control combines the velocity and acceleration feedback of the joints, enhancing the system's adaptability to complex operating environments.

[0032] Servo motor control: High-precision servo control: The servo motor control communicates with the control system via the CAN bus. The real-time performance and anti-interference ability of the CAN bus keep the motor's response time at the millisecond level. In addition, the servo motor integrates a position sensor and a speed encoder to measure the real-time position and speed of the joint, and maintains the high precision of the system through closed-loop feedback control.

[0033] Current control: The servo motor drive includes current loop control, that is, the torque output of the motor is precisely controlled by adjusting the current flowing through the motor windings. Combining the position loop and the speed loop, the motor control achieves precise adjustment of multi-loop cooperation, ensuring the smoothness and reliability of joint movement.

[0034] Data acquisition and feedback mechanism: Multi-sensor fusion and Kalman filtering: The joint control system combines multiple sensors, including accelerometers, gyroscopes, etc., and fuses the data of multiple sensors through Extended Kalman Filtering (EKF) to filter out noise and improve the accuracy of joint motion state estimation. This enables the control system to adjust the actions of the servo motors based on accurate position and speed estimation values, reducing feedback delay and improving the control response speed.

[0035] The following are the explanations of the terms in step S3: Intelligent joint controller An electronic device responsible for controlling the movement of each joint of the humanoid model. It receives action instruction data and combines real-time feedback to adjust the actions of the servo motors, ensuring the accuracy and coordination of joint movement.

[0036] Servo motor A precision motor is used to control the movement of joints. The servo motor can adjust its rotation angle and speed according to the control signal, and is suitable for high-precision motion control.

[0037] Closed-loop control system A control system whose output affects the input through a feedback mechanism to achieve precise control. The closed-loop control system can monitor the error between the actual position and the target position of the joint in real time, and adjust the control signal according to the error, so as to ensure the accurate execution of the action.

[0038] Servo motor drive: The joint movement is driven by a high-precision servo motor. The movement of each joint is controlled by an independent servo motor, and the motor can precisely adjust the angle, speed and acceleration.

[0039] Closed-loop control system: The movement of each joint is monitored and corrected in real time through a closed-loop control system. The system adjusts the rotation angle of the servo motor in real time according to the difference between the target position and the actual position of the joint, ensuring that the joint movement is consistent with the instruction.

[0040] Control process: Initialization: The system sends an initial motion instruction to the servo motor according to the joint target position and motion trajectory in the action instruction.

[0041] Motion execution and feedback: The servo motor starts to move, and at the same time the position sensor monitors the actual motion state of the joint and feeds it back to the controller. The controller calculates the difference (error) between the target position and the actual position in real time, and continuously adjusts the rotation of the motor to eliminate the error.

[0042] Dynamic adjustment: If an interaction between the model and the environment is detected during the movement (such as being affected by an external force or hitting an obstacle), the system will immediately adjust the joint movement according to the sensor feedback to avoid action errors or damage to the model.

[0043] S4. Action adjustment: Dynamically adjust the joint action parameters according to the attitude data and external environment changes feedback by real-time monitoring through an adaptive control algorithm; Specifically, the implementation of the adaptive control algorithm: Deep reinforcement learning model: Adaptive control adopts deep reinforcement learning (DRL), such as deep Q-network (DQN) or proximal policy optimization (PPO) algorithm. The model continuously updates the parameters of the policy network through real-time perception of the environment (data from visual sensors and force sensors), enabling the control system to adapt to dynamic environment changes. The training process of the model uses an experience replay mechanism, storing the previous actions and environment feedback data in the replay pool for repeated use in training to reduce the problem of low sample utilization rate.

[0044] Distributed Training: To improve the real-time performance of the adaptive control algorithm, the system uses a distributed learning framework to train the reinforcement learning model through multi-GPU parallel processing, ensuring that when faced with complex action requirements, the weights of the policy network can be quickly updated, thereby dynamically adjusting the control parameters of the joints.

[0045] Model Predictive Control (MPC): Future State Prediction: In real-time control, the adaptive control module uses Model Predictive Control (MPC) to predict the future state of the system. By establishing a dynamic model, MPC can predict the future positions and velocities of the joints and optimize the current action parameters based on the prediction results. For example, when the system detects an impending collision, MPC will adjust the motion trajectory of the joints in advance to avoid obstacles.

[0046] Constraint Handling: MPC allows the system to optimize the control input while considering physical constraints such as the maximum torque and maximum speed of the joints, ensuring that the humanoid model does not exceed the mechanical limits during high-precision actions and avoiding equipment damage.

[0047] Real-time Path Planning: Combination of A* and RRT: To handle action adjustments in a dynamic environment, the adaptive control module combines the A* algorithm and the Rapidly-exploring Random Tree (RRT) algorithm for path planning. The A* algorithm is used to plan the optimal path in a known environment, while RRT is used to handle the random path generation in a dynamic or unknown environment. This combination can ensure finding the optimal path in a complex environment and quickly responding to sudden changes.

[0048] Path Smoothing: After generating the path, the path is smoothed through B-spline interpolation to ensure the coherence and naturalness of joint movements and avoid sudden stops or abrupt accelerations caused by path discontinuities.

[0049] Control Adjustment Based on Fuzzy Logic: Enhanced Environmental Adaptability: To further improve the adaptability of actions, the adaptive control module also adopts fuzzy logic control. Based on the feedback from visual sensors, the fuzzy logic system can process fuzzy quantities such as "degree of approaching an obstacle" and "current joint velocity" to generate more suitable action parameters. This enables joint movements to more flexibly adapt to environmental changes. For example, when the humanoid model walks in a narrow environment, it automatically decelerates or adjusts its posture.

[0050] The following are the explanations of the terms in step S4: Adaptive Control Algorithm A control algorithm that can dynamically adjust control parameters according to real-time monitoring feedback and environmental changes to optimize system performance. In this method, the adaptive control algorithm automatically adjusts the movement mode of the joints based on sensor data and user input to ensure smooth and natural movements.

[0051] Adaptive control algorithm: This algorithm optimizes actions based on real-time feedback and historical data to make the actions more natural and smooth. The algorithm can learn the movement characteristics of the model and external environmental changes, and automatically adjust the control parameters of the joints.

[0052] Learning process: After each action is executed, the system records the response data of the joints, including movement time, angular velocity change, user feedback, etc. Based on this data, the system will train and optimize the action strategy through a machine learning model. For example, if the execution speed of a certain action is slow, the system will automatically adjust the acceleration or speed parameters of the joints to improve the efficiency of the next execution.

[0053] Real-time adjustment: During the execution process, the algorithm dynamically adjusts the joint actions according to real-time sensor data. For example, when performing a jump or a quick turn, the system may automatically adjust the action trajectory and speed of the joints according to the posture change of the model to ensure that the actions are coherent and in line with the actual situation.

[0054] Scene adaptability: Through adaptive control, the system can automatically select the optimal action execution method according to different scenes. For example, in an outdoor environment, the system will adjust the action trajectory of the model according to the uneven ground or external interference; while in an indoor shooting environment, the system will pay more attention to the accuracy and beauty of the actions.

[0055] S5. Data processing and feedback: Through hardware acceleration technology, the feedback data from the multi-sensor system and the joint control module are processed in parallel in real time, and the processing results are immediately fed back to the intelligent joint controller to dynamically adjust the joint action parameters; The following are the explanations of the terms in step S5: Low-latency data processing architecture A data processing structure designed to achieve fast response and real-time feedback. The low-latency data processing architecture usually uses parallel computing and hardware acceleration technologies to reduce the data processing time and ensure that the system can respond to user instructions and sensor feedback within milliseconds; Feedback data Refers to the data (such as joint position, speed, etc.) monitored by the system after the action is executed and returned to the control module in real time for dynamic adjustment. Feedback data is an important basis for realizing closed-loop control and optimizing action execution.

[0056] Data Processing Architecture: The system adopts a parallel computing architecture, allowing data from multiple sensors to be processed simultaneously to ensure real-time performance during action execution. The data processing module can collect, analyze, and process all sensor data within milliseconds and feedback the processing results to the control module.

[0057] Hardware Acceleration: The system incorporates hardware acceleration technologies, such as using specialized computing units (e.g., GPUs, FPGAs) to process sensor data, further enhancing the processing speed and reducing the latency of action execution.

[0058] Feedback Mechanism: The processing results of all sensor data are real-time feedback to the control module for guiding the dynamic adjustment of joints. For example, during a quick turn action, the system adjusts the rotation speed of the joints based on the feedback from the gyroscope and accelerometer to ensure the synchronization and accuracy of the turn action.

[0059] This embodiment realizes the method for synchronizing the POSE actions of a humanoid model. By adopting high-performance servo motors and a closed-loop control system, combined with multi-sensor fusion technology, it achieves precise control of the model's joints and real-time monitoring of the posture. The multi-sensor system integrates accelerometers, gyroscopes, vision sensors, and force sensors, which can provide accurate posture perception and environmental detection data to ensure the natural smoothness and synchronization of actions. The adaptive control algorithm is based on machine learning technology, which can dynamically optimize the motion parameters of the joints according to real-time feedback and gradually improve the system's adaptability to user needs, providing personalized action performances. The system adopts a low-latency parallel data processing architecture to ensure real-time response and coordination of actions. At the same time, it has self-diagnosis and safety monitoring functions, which can automatically adjust or stop actions when abnormalities occur to ensure the safety of the equipment and the environment. Through an intuitive graphical operation interface and voice control function, users can conveniently and quickly set and control the actions of the model, enhancing the operation experience and the usability of the system.

[0060] Optionally, in practical applications, the action instruction data can be obtained in various ways, including a preset action library, custom actions input by users, or real-time data generated by external motion capture devices. These data define the motion parameters such as the target position, angle, and speed of each joint of the model.

[0061] Optionally, during the implementation process, the multi-sensor fusion system of the method consists of an accelerometer, a gyroscope, a vision sensor, and a force sensor. These sensors work together to be able to monitor the environmental information around the model in real time.

[0062] Accelerometer: Real-time monitors the linear motion and acceleration of the model, obtains the motion state of the model in different directions, and is used to correct the posture of the model.

[0063] Gyroscope: Used to detect the rotation angle and angular velocity of the model in three-dimensional space, ensuring precise movements when the model performs rotations or swings.

[0064] Vision sensor: Used to capture information about the environment around the model, identify obstacles or specific target objects, and help the model avoid collisions with objects during movements.

[0065] Force sensor: Used to detect the acting force when the model comes into contact with external objects, ensuring safety and flexibility during movement execution.

[0066] Through sensor data fusion technology, the system can accurately identify obstacles and adjust the model's movements to avoid collisions.

[0067] Optionally, the intelligent joint controller is integrated with a self-diagnosis module. This module can detect joint overload or abnormal movements by real-time monitoring the data of the joint's load sensor and position sensor. When an abnormality is detected, the system will send a control signal to adjust or stop the joint's movement to avoid equipment damage or movement failure. In actual operation, when the joint load of the model exceeds the set threshold, the self-diagnosis module will react immediately to ensure the safe operation of the system.

[0068] Optionally, the adaptive control algorithm dynamically adjusts the joint's motion parameters by real-time monitoring the system's data, including the input action instructions, sensor feedback data, and environmental changes. After each action execution, the system records the joint's motion data (such as speed, angle, acceleration, etc.) and optimizes the subsequent action performance based on this data. For example, when it detects that the model's action speed is too fast, the adaptive control algorithm will reduce the joint's acceleration to ensure the naturalness and coherence of the action.

[0069] Optionally, this method further includes a safety monitoring module responsible for monitoring the working state of the joints. The safety monitoring module can real-time detect parameters such as the temperature, load, and working state of the sensors of the joints. When an abnormal situation is detected, such as joint overload or too high temperature, the safety monitoring module will issue an alarm and automatically adjust or stop the joint's movement to ensure that the model will not be damaged due to failures.

[0070] Refer to Figure 2 , Embodiment 2 of the present invention provides a humanoid model POSE action synchronization system, including: Action capture module Function: Used to obtain action instruction data, and the action capture module is the core module of the system input. This module receives action data from multiple sources, including custom instructions input by users, preset actions in the standard action library, and motion data obtained in real-time through external action capture devices.

[0071] Implementation process: Users can generate action instruction data through various input methods provided by the system. For example, users can select preset POSE actions through a graphical user interface (GUI), or use an external motion capture device (such as a camera or sensor array) to capture human movements in real time. After being processed in real time, these data are converted into parameters such as joint positions, angles, and speeds that the system can recognize, forming instruction data.

[0072] Data format: This module supports multiple data formats, such as standardized joint coordinates, Euler angles, quaternions, etc., for defining the motion trajectories and pose changes of the humanoid model. These data can be stored and processed in the form of time series to ensure that the motion of each joint has clear instruction points on the time axis.

[0073] Multi-sensor fusion module Function: It is used to monitor the pose of the model and external environment information, and integrate the data of multiple sensors to ensure precise control of the actions of the humanoid model.

[0074] Hardware components: Accelerometer: It monitors the linear motion and acceleration of the model in real time, and obtains the motion state of the model in different directions.

[0075] Gyroscope: It is used to detect the rotation angle and angular velocity of the model in three-dimensional space to ensure the accuracy of the model's rotation actions.

[0076] Vision sensor: It captures the image data of the surrounding environment, identifies obstacles and target objects in the scene, and helps the model adjust its actions in a dynamic environment.

[0077] Force sensor: It senses the contact force between the model and external objects or people, ensures the coordination of actions during interaction, and prevents damage or accidents caused by excessive force.

[0078] Data fusion method: It adopts data fusion algorithms such as Kalman filtering to integrate the data from multiple sensors into a unified model. These fused data are used to calculate the current pose of the model in real time, including position, speed, acceleration, and rotation angle, and are provided to the intelligent joint control module for further processing.

[0079] Intelligent joint control module Function: This module is the core for controlling the actions of each joint of the humanoid model. It is responsible for receiving action instruction data and achieving precise joint actions through servo motors.

[0080] Implementation method: Each joint is driven by a high-precision servo motor, and the servo motor is connected to the control system through a closed-loop control system. The closed-loop control system can monitor the current position and speed of the joint in real time, continuously correct the control signal through the feedback data of the sensor, and ensure that the joint movement is completely synchronized with the instruction. The response speed and accuracy of the servo motor can reach the millisecond level, which is suitable for complex posture and motion transformations.

[0081] Control process: After the system receives the action instruction, the intelligent joint control module calculates the rotation angle and speed instruction of the servo motor according to the target position, speed, acceleration of the joint and external feedback.

[0082] The motor starts to move, and at the same time, the sensor monitors the movement of the motor in real time and returns the feedback data to the control system.

[0083] The closed-loop control system adjusts the motor action in real time according to the feedback data to ensure that the joint movement trajectory is highly consistent with the target instruction.

[0084] Adaptive control module Function: The adaptive control module optimizes the action performance of the model in real time through an adaptive algorithm to ensure the coherence and smoothness of the action.

[0085] Algorithm type: This module adopts an adaptive control algorithm based on machine learning, which can dynamically adjust the action parameters according to historical data and environmental feedback. For example, after each action execution, the system records data such as the response time, position error, and speed change of the joint. Through the analysis of these data, the adaptive algorithm can automatically optimize the execution strategy of the next action.

[0086] Learning process: When running for the first time, the system executes instructions through a preset algorithm and model. As the usage time increases, the system gradually accumulates data and continuously optimizes through a learning mechanism. For example, if it is detected that the execution speed of some actions is too fast or too slow, the system will automatically adjust the movement trajectory and time parameters of the joint to make the subsequent execution smoother and more natural.

[0087] Real-time adjustment: During the execution process, the adaptive control module will adjust the motion parameters in real time according to the sensor feedback, such as acceleration or angular velocity, to cope with sudden environmental changes or posture instability problems.

[0088] Data processing module Function: It is used to perform parallel processing on multi-sensor data to ensure the real-time performance and accuracy of the system response.

[0089] Data Processing Architecture: This module uses a multi-core processor combined with hardware acceleration technology for data processing, and can complete the acquisition, analysis, and processing of each sensor's data within milliseconds. The parallel processing architecture allows the system to simultaneously process data from the accelerometer, gyroscope, vision sensor, and force sensor, and quickly feedback the processing results to the control module.

[0090] Data Caching and Transmission: To improve processing efficiency, the system designs a multi-level caching mechanism. First, the sensor data is cached in the local memory, and then processed in the order of priority. High-speed data transmission between modules is achieved through the data bus to ensure that the joint control module can obtain the latest sensor data within milliseconds.

[0091] Safety Monitoring Module Function: The safety monitoring module is responsible for real-time monitoring of the working status of the entire system to ensure the safety of system operation.

[0092] Monitoring Content: It includes monitoring the working temperature of the servo motor, joint load conditions, sensor working status, etc. Through the built-in sensors and diagnostic mechanisms, the system can obtain the status information of each joint in real time.

[0093] Safety Measures: When the system detects overload of the servo motor, abnormal temperature, sensor failure, or abnormal action execution, the safety monitoring module will immediately trigger an emergency handling mechanism. At this time, the system will limit the action range of the model or directly stop all actions to prevent further damage or dangerous situations.

[0094] In the third embodiment of the present invention, based on the same inventive concept, a computer-readable storage medium is proposed by the present invention. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of the humanoid model POSE action synchronization method in the above embodiment.

[0095] In the fourth embodiment of the present invention, based on the same inventive concept, a terminal is proposed by the present invention. The terminal includes: a processor and a memory; the processor and the memory communicate with each other; the memory is used to store instructions; the processor is used to execute the instructions in the memory to execute the humanoid model POSE action synchronization method in the above embodiment.

[0096] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention. Any equivalent structural or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. Humanoid model POSE action synchronization method, characterized in that It includes the following steps: S1. Obtain action instruction data: Obtain action instruction data including the target positions and movement trajectories of multiple joints of a humanoid model through an input device; S2. Monitor the current posture in real time: Adopt a multi-sensor fusion system to monitor the posture and environment of the humanoid model in real time; S3. Implement joint control: Based on an intelligent joint controller, use a closed-loop control system and servo motors to control multiple joints of the humanoid model. The control includes driving the joints to move according to the obtained action instruction data and adjusting the posture in combination with the data monitored in real time; S4. Action adjustment: Dynamically adjust the action parameters of the joints through an adaptive control algorithm according to the posture data feedback monitored in real time and changes in the external environment; S5. Data processing and feedback: Through hardware acceleration technology, perform real-time parallel processing on the feedback data from the multi-sensor system and the joint control module, and immediately feedback the processing results to the intelligent joint controller to dynamically adjust the joint action parameters.

2. The humanoid model POSE action synchronization method according to claim 1, characterized in that, The action instruction data includes a preset action sequence or custom action instruction data input by the user.

3. The method for synchronizing the POSE actions of a mannequin according to claim 1, characterized in that, The multi-sensor fusion system includes an accelerometer, a gyroscope, a vision sensor, and a force sensor, and can detect the environmental information around the humanoid model through sensor data fusion technology, specifically including the recognition of obstacles and target objects.

4. The humanoid model POSE action synchronization method according to claim 1, wherein The intelligent joint controller is integrated with a self-diagnosis module. The self-diagnosis module detects joint overload or abnormal actions by monitoring the data of the load sensor and position sensor of the joint in real time, and issues a control signal to adjust or stop the joint action when an abnormality is detected.

5. The method for synchronizing the POSE actions of a mannequin according to claim 1, characterized in that, The adaptive control algorithm comprehensively analyzes the input action instruction data, the sensor feedback data monitored in real time, and the environmental change information, and dynamically adjusts the motion parameters of the joints, including speed, angle, and acceleration.

6. The method for synchronizing the POSE actions of a mannequin according to claim 1, wherein The method further includes using a safety monitoring module to monitor the working state of the joints and adjusting or stopping the joint action when an abnormality occurs.

7. Humanoid model POSE action synchronization system, characterized in that, Applied to the humanoid model POSE action synchronization method according to any one of claims 1-6, the system includes: An action capture module: Used to obtain and process action instruction data, and the action instruction data includes the target positions and movement trajectories of multiple joints; A multi-sensor fusion module: Monitor the posture of the humanoid model and the external environment in real time through an accelerometer, a gyroscope, a vision sensor, and a force sensor; An intelligent joint control module: Based on the action instruction data, use servo motors and a closed-loop control system to achieve control of multiple joints; An adaptive control module: Comprehensively analyzes the input action instruction data, the sensor feedback data monitored in real time, and the environmental change information, and dynamically adjusts the motion parameters of the joints, including speed, angle, and acceleration; A data processing module: Adopt a parallel data processing architecture and hardware acceleration technology to perform real-time parallel processing, and immediately feedback the processing results to the intelligent joint controller.

8. The humanoid model POSE action synchronization system according to claim 7, characterized in that, The action capture module can process action data in different formats, including preset action sequences and user-defined action instructions; The multi-sensor fusion module can identify obstacles or target objects in the environment and provide data to the humanoid model POSE motion synchronization system for processing; The intelligent joint control module has a self-diagnosis function and can detect and process joint overload and abnormal motion conditions; The adaptive control module can adjust the execution mode of joint motions according to historical data and input information; The data processing module adopts a parallel data processing architecture and, through hardware acceleration technology and multi-core parallel computing, synchronously processes real-time data from multiple sensors and provides dynamic data support during the adjustment of joint motions; The humanoid model POSE motion synchronization system further includes a safety monitoring module: monitoring the operating state of the humanoid model POSE motion synchronization system and restricting motion execution or entering a safe mode in case of an abnormality.

9. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the humanoid model POSE motion synchronization method according to any one of claims 1-6.

10. A readable storage medium, characterized in that, A computer program is stored on the readable storage medium, and when the computer program is executed by the processor, it implements the humanoid model POSE motion synchronization method according to any one of claims 1 to 6.

Images