Simulation cloth intelligent toy mouth shape control system and method based on action linkage

Abstract

The invention discloses a simulation cloth intelligent toy mouth shape control system and method based on action linkage, and the method comprises the steps: capturing user actions in real time through a plurality of sensors, and carrying out the denoising, filtering and normalization processing through a signal processing module; the processed data is analyzed by the intelligent control unit, user actions are recognized, a mouth shape control instruction is generated, and cloth is driven to form a mouth shape; the system optimizes a control effect according to historical data and a feedback adaptive adjustment algorithm; power management adjusts power according to use conditions, and a power-saving mode is started when the battery is low; environmental conditions affect sensor sensitivity and a recognition algorithm; through the communication interface, a user can remotely control the toy function and carry out wireless software upgrading; according to the simulation cloth intelligent toy mouth shape control system and method based on action linkage, through a high-precision sensor, an intelligent control algorithm and a low-power-consumption design, the toy can accurately adjust the mouth shape and respond to user actions, and interactivity and entertainment are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of action linkage simulation fabric intelligent toy mouth shape control, and particularly relates to a simulation fabric intelligent toy mouth shape control system and method based on action linkage. BACKGROUND

[0002] A simulation fabric intelligent toy mouth shape control system based on action linkage, generally refers to a system that combines simulation technology, intelligent control, fabric material, and action sensing, commonly used in intelligent toys, especially in toys that simulate animals, characters, and other objects with mouth movements. The goal of this system is to simulate real mouth movements by controlling the opening and closing, expression, and other actions of the toy mouth, and to enable it to be controlled in linkage with external actions or environmental changes. Through built-in sensors, the system can sense real-time changes in the toy or external environment, such as when the toy is pressed, touched, or moved, the sensors will detect these changes and transmit the data to the control unit. Through built-in microprocessors or control chips, the received action signals are analyzed and specific control logic is executed, which may involve controlling motors or actuators to drive the movement of the toy mouth to match the action or instruction. Simulation fabric is usually a flexible material that can simulate the texture and dynamic performance of real objects. In intelligent toys, the mouth area may use this fabric to simulate the opening and closing, stretching, and other lively movements of the mouth. When the control system drives the fabric, the fabric will change according to the input signal, making the mouth performance more realistic. Action linkage refers to the interaction between different actions. In this system, the changes in the toy's mouth may be coordinated with other actions to produce a comprehensive effect. For example, when the toy's head tilts, the mouth may also change, enhancing the realism.

[0003] However, with the advancement of technology, the interactivity and intelligence of intelligent toys are constantly improving. Existing intelligent toys rely on simple mechanical actions and electric drives to complete various actions and reactions. In existing technology, the mouth control of intelligent toys mainly relies on fixed trigger conditions such as buttons or remote control operations, lacking dynamic linkage control based on user actions. This method can achieve a certain level of interactivity, but has certain limitations in user experience, especially in terms of action recognition accuracy, real-time response, energy efficiency management, and other aspects, which have not yet reached the ideal interactive effect. SUMMARY

[0004] In view of the deficiencies in the prior art, the present application aims to provide a simulation fabric intelligent toy mouth shape control system and method based on action linkage, which introduces high-precision sensors, intelligent control algorithms, and low-power design technology to enable the toy to accurately adjust the mouth shape and respond to user's real-time actions, enhancing interactivity and entertainment.

[0005] The technical scheme adopted by the present application to solve its technical problems is:

[0006] An action linkage-based simulation fabric intelligent toy mouth shape control system, comprising:

[0007] A simulation fabric mouth shape control component module is used to adopt flexible fabric and a control driving mechanism to adjust the deformation of the fabric through an intelligent control system. The material selection of the fabric adopts flexibility and adjustability, and a three-dimensional mouth shape effect is formed according to the response of the control signal;

[0008] An action sensing module is used to monitor the user's actions in real time. The sensors include but are not limited to pressure sensors, acceleration sensors, and infrared sensors. The user's touch, pressing, and shaking actions are detected through these sensors, and the sensing data is transmitted to the control unit;

[0009] An intelligent control unit module is used to receive the data of the action sensing module and generate corresponding control signals according to a preset action linkage algorithm, instructing the driving simulation fabric mouth shape control component to execute mouth shape changes;

[0010] A power management system module is used to adopt a low-power design, with a built-in rechargeable battery, automatically adjusting power consumption. The system dynamically adjusts power according to usage and sends a warning signal when the battery power is low;

[0011] A communication interface module is used to adopt wireless communication functions to connect with intelligent devices for remote control and system upgrade. Users can adjust and set through smartphones or other devices.

[0012] An action linkage-based simulation fabric intelligent toy mouth shape control method, comprising the following steps:

[0013] Through various sensors in the action sensing module, the user's action data, including pressing, shaking, and approaching actions, is captured in real time;

[0014] The collected action data is preprocessed by a signal processing module. The preprocessing step includes denoising, signal filtering, and data normalization;

[0015] The processed action data is transmitted to the intelligent control unit, which uses a preset action recognition algorithm to analyze the data, recognize the user's actions, and generate corresponding mouth shape control instructions. According to the analysis results, the control unit transmits the generated instructions to the simulation fabric mouth shape control component to drive the fabric material to form a three-dimensional mouth shape;

[0016] The system adapts to the action recognition algorithm based on historical usage data and user feedback, and continuously optimizes the mouth shape control effect by learning the user's operation habits;

[0017] In use, the power management system automatically adjusts the power of the system according to the actual use, when the battery power is low, the system sends a warning signal, and automatically starts the power saving mode;

[0018] The system adjusts the sensitivity of the sensor and the action recognition algorithm in real time according to the ambient temperature, humidity, and light environment conditions;

[0019] Through the communication interface module, the user remotely controls the function setting of the toy, adjusts the details of the mouth shape control, and updates the function wirelessly and remotely. The system regularly or according to the demand, performs remote software upgrade.

[0020] As preferred, through various sensors in the motion sensing module, real-time capture of user motion data, including pressing, shaking, and proximity action methods are:

[0021] The pressure sensor is deployed on the external surface of the toy. When the user presses the toy, the sensor senses the change in surface pressure and outputs the corresponding pressure value;

[0022] The acceleration sensor is used to capture the shaking action and detect the acceleration change of the object in three-dimensional space. When the user shakes the toy, the sensor senses the dynamic acceleration change of the object and identifies the strength, direction, and frequency of the shaking action;

[0023] The infrared sensor detects the distance between the object and the sensor through the reflection of the infrared light beam. When the user approaches the toy with his hand or other objects, the infrared sensor identifies the proximity of the object and triggers the corresponding control signal;

[0024] The sensor collects motion data in real time, records pressure values, acceleration data, and infrared distance value output signals, and transmits these data to the signal processing module.

[0025] As preferred, the collected motion data is preprocessed by the signal processing module. The preprocessing step includes the methods of denoising, signal filtering, and data normalization:

[0026] Smooth the data by sliding window averaging method. For each data point, calculate the average value within a certain range of its neighborhood. Sort the neighborhood around the data point and select the middle value to eliminate outliers;

[0027] Low-pass filtering: filter out high-frequency components and retain low-frequency signals;

[0028] High-pass filtering: filter out low-frequency components and retain high-frequency change signals;

[0029] Combine low-pass and high-pass filtering to retain signals within a specific frequency band;

[0030] Methods of data normalization include:

[0031] Normalization is performed by mapping data to a specific range, with the formula:

[0032]

[0033] where x is the original data, min(x) and max(x) are the minimum and maximum values in the data set, respectively;

[0034] Normalization is performed by the mean and standard deviation of the data, so that the mean is 0 and the standard deviation is 1, with the formula:

[0035]

[0036] where x is the original data, \mu is the mean, and \sigma is the standard deviation.

[0037] Preferably, the processed action data is transmitted to the intelligent control unit, which uses a preset action recognition algorithm to analyze the data, recognize the user's action, and generate corresponding mouth shape control instructions. According to the analysis result, the control unit transmits the generated instructions to the simulated fabric mouth shape control component to drive the fabric material to form a three-dimensional mouth shape, and the method is:

[0038] The intelligent control unit analyzes the transmitted action data and uses a preset action recognition algorithm to determine the type of user action. The action recognition algorithm includes a classification algorithm and a neural network.

[0039] The result of action recognition is represented as a certain control instruction vector, with the formula:

[0040] A = f(D)

[0041] where D represents the preprocessed action data, f(D) represents the action result identified by the algorithm, and A is the action instruction after recognition.

[0042] According to the recognition result, the corresponding mouth shape control instruction is generated. The mouth shape control simulates the deformation of the fabric through a kinematic model to form a predetermined three-dimensional mouth shape. The control model of the mouth shape is based on the physical properties of the fabric. The control instruction depends on the changes in these properties. The control instruction C is represented by the following formula:

[0043] C = g(A, P)

[0044] where A is the instruction after action recognition, P is the physical parameter of the fabric, and g(A, P) is the control instruction generation function.

[0045] The generated mouth shape control instruction C is transmitted to the simulation fabric mouth shape control component to drive the fabric material to form a three-dimensional mouth shape. Based on the finite element analysis method, the deformation process of the fabric is simulated.

[0046] The displacement vector of the fabric is set as U, which represents the displacement of the fabric. The deformation of the fabric is described by the following mechanical formula:

[0047] F = KU + F ext

[0048] Where F is the internal force of the fabric, K is the stiffness matrix of the fabric, U is the displacement vector of the fabric, F ext is the external force;

[0049] Finally, the mouth shape control instruction C will form a three-dimensional mouth shape through a driving model related to the physical properties of the fabric. Through the dynamic simulation system, the fabric will change according to the calculated control instruction, and finally form the user's desired mouth shape.

[0050] As a preferred embodiment, the system adaptively adjusts the motion recognition algorithm based on historical usage data and user feedback, and continuously optimizes the mouth shape control effect by learning the user's operation habits. The method is as follows:

[0051] The system continuously records the user's operation data and feedback information, and the motion recognition algorithm adjusts in real time according to the new data. With the passage of time, the system gradually adjusts and optimizes the algorithm parameters;

[0052] The learning process of the system optimizes the motion recognition function f by minimizing the error, and the incremental learning is represented by the following formula:

[0053]

[0054] Where f t is the motion recognition function of the current model; η is the learning rate; L(D t , θ t ) is the loss function;

[0055] The Q-learning method in reinforcement learning is represented as:

[0056]

[0057] Where Q(s t , a t ) is the value of performing action a t in state s t ;

[0058] α is the learning rate;

[0059] r t is the current feedback reward;

[0060] gamma is a discount factor;

[0061] max a Q(s t|1 , a) is the maximum Q-value in the next state;

[0062] By analyzing historical data and user feedback, the system automatically selects and adjusts the current user's action recognition algorithm. Different users have different operation habits, and the system selects from multiple algorithms to gradually optimize the accuracy and responsiveness of the algorithm. Algorithm selection includes model selection and hyperparameter adjustment.

[0063] Let the algorithm used by the system be A, and each algorithm has a set of hyperparameters theta A Optimize the hyperparameters through Bayesian optimization method:

[0064]

[0065] Where theta * is the optimal set of hyperparameters, and P(Feedback | theta A ) is the posterior probability based on feedback information.

[0066] Collaborative filtering uses a user-action matrix R, where each user's feedback on an action is an element of the matrix. The goal of the algorithm is to minimize the error between the predicted feedback and the actual feedback:

[0067]

[0068] Where U and V are the latent factor matrices for users and actions, respectively, is the predicted feedback of user i on action j.

[0069] As a preferred, during use, the power management system automatically adjusts the power of the system according to the actual use, when the battery power is low, the system sends a warning signal, and automatically starts the power saving mode. The method is:

[0070] The remaining battery power is calculated in real time by monitoring the battery voltage and current. The battery power E(t) changes over time and is estimated by the following formula:

[0071]

[0072] Where E0 is the initial battery power, I(τ) is the current at time τ, and V(τ) is the battery voltage.

[0073] During system operation, the power management system automatically adjusts power consumption according to actual usage requirements. Power management methods include dynamic voltage frequency adjustment, task scheduling optimization, and dynamic adjustment of hardware and software.

[0074] Equation:

[0075] P total (t) = P CPU (t) + P Display (t) + P Periphrals (t)

[0076] Where P total (t) is the total power consumption of the system; P CPU (t), P Display (t), and P Peripherals (t) are the power consumptions of the processor, battery display, and peripheral module, respectively.

[0077] When the battery level is below a preset threshold, the power management system detects this change through the battery management system and sends a warning signal. The threshold is set at 20% or lower of the battery level. When a low battery level is detected, the system sends a warning signal or starts a power saving mode.

[0078] Let the system battery level be E(t). When E(t) < E threshold , the system sends a warning and starts a power saving mode. The set power consumption limit is P limit . The triggering of the warning and power saving mode is represented as:

[0079] If E(t) < E threshold , P total (t) → P limit

[0080] When the power saving mode is started, the system takes measures to reduce power consumption, such as reducing screen brightness, limiting background tasks, turning off Bluetooth, Wi-Fi, GPS hardware components, and reducing processor frequency.

[0081] The power consumption model in power saving mode is represented as:

[0082]

[0083] Where α is a coefficient less than 1.

[0084] When the battery level is sufficient, the system automatically returns to normal mode and restores the original power settings. During the recovery process, the system gradually increases power consumption based on the battery level to ensure a return to normal operating state. The equation is:

[0085] If E(t) > E threshold , P total (t) → P normal

[0086] Where P normal is the power in normal operating mode.

[0087] As a preferred, the system adjusts the sensitivity of the sensor and the action recognition algorithm in real time according to the ambient temperature, humidity, light environment conditions, and the method is:

[0088] The system collects environmental data in real time through the sensor, and dynamically adjusts the sensitivity of the sensor according to the environmental data;

[0089] According to the collected environmental data, the system adjusts the sensitivity of the sensor in real time, and the adjustment mode includes temperature compensation, humidity correction, and light self-adaptive adjustment;

[0090] The sensitivity of the sensor used by the system is S(t), the temperature, humidity and light intensity are T(t), H(t) and L(t) respectively, and the sensitivity adjustment is represented as:

[0091] S(t) = S0 + αT(t) + βH(t) + γL(t)

[0092] Where S0 is the default sensitivity; α, β, γ are adjustment coefficients;

[0093] In addition to the adjustment of the sensitivity of the sensor, the system dynamically optimizes the action recognition algorithm according to the environmental conditions. The environmental factors are based on the quality of the sensor data, and the algorithm is equipped with adaptive ability to handle the interference of environmental changes on data;

[0094] The accuracy of the action recognition algorithm is based on the sensor data D(t), which is related to the environmental conditions T(t), H(t) and L(t). The adjustment of the algorithm is realized by modifying the weight coefficient w(T, H, L). The system uses a machine learning model to recognize actions, which is represented by the following model:

[0095] A(t) = f(D(t), w(T(t), H(t), L(t)))

[0096] Where A(t) is the action recognition result, f(·) is the action recognition algorithm, and w(T, H, L) is the adjustment coefficient under the environmental conditions;

[0097] The system updates the adjustment strategy according to the historical environmental data D history and real-time data D current , which is represented as:

[0098] w(T(t), H(t), L(t)) = g(D history , D currcnt )

[0099] Where g(·) is the adjustment strategy update function, which adjusts the sensitivity of the sensor and the action recognition algorithm in real time.

[0100] As preferred, the user remotely controls the function settings of the toy through the communication interface module, adjusts the details of the mouth shape control, and the system performs remote software upgrading periodically or according to the demand through the wireless remote updating function:

[0101] The system is remotely connected with the user equipment through the Wi-Fi, Bluetooth, ZigBee wireless communication interface module, the user sends instructions through the control equipment, and adjusts the function and setting of the toy;

[0102] The user remotely controls the function settings of the toy through wireless communication, including but not limited to the adjustment of the details of the mouth shape control, the function enabling and disabling;

[0103] The system performs software upgrading periodically or according to the demand, and the process includes the following steps:

[0104] The toy periodically checks whether there is a new software version through the communication interface module, if it is a periodic check, the toy automatically connects the server according to the preset time interval, and queries whether there is an available upgrade package;

[0105] The user selects to start the software upgrading function through the control interface;

[0106] The toy downloads a new software update package through the communication interface module, and verifies the integrity and safety of the downloaded file;

[0107] After verification, the toy performs the upgrading operation in a safe mode;

[0108] A rollback mechanism is set, if there is a problem in the upgrading process, the system automatically rolls back to the last stable version;

[0109] The system includes a server end and a plurality of toy equipment, the server is responsible for storing and distributing the latest software update package, and also receives the control instructions sent by the user, each toy equipment is connected with the server through wireless communication, and performs remote control and software updating according to the demand.

[0110] Another technical problem to be solved by the present application is to provide an electronic device, including a memory, a processor and a computer program stored on the memory and executable on the processor, when the processor executes the program, a simulation cloth intelligent toy mouth shape control system and method based on action linkage are realized.

[0111] Another technical problem to be solved by the present application is to provide a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize a simulation cloth intelligent toy mouth shape control system and method based on action linkage.

[0112] The beneficial effects of the present application are:

[0113] The change of the toy mouth shape is realized through precise action linkage control, making the interactive experience more natural and accurate; through the joint use of adaptive algorithms and multiple sensors, the accuracy of action recognition is improved, and the delay and error are reduced; in terms of energy efficiency management, low-power design and intelligent battery management are adopted, prolonging the use time of the toy and adding wireless charging function; environmental adaptability and personalized adjustment function improve the interactivity of the toy, enabling it to optimize itself according to different users and environmental conditions. BRIEF DESCRIPTION OF DRAWINGS

[0114] Figure 1 It is a flowchart of a simulation cloth intelligent toy mouth shape control system based on action linkage. DETAILED DESCRIPTION

[0115] The principles and features of the present application are described below, and the examples are only used to explain the present application and are not intended to limit the scope of the present application. In the following paragraphs, the present application is described in more detail by way of example. The advantages and features of the present application will be more apparent from the following description and claims.

[0116] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used in the specification of the present application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0117] EMBODIMENT

[0118] The technical solution adopted by the present application to solve its technical problems is:

[0119] A simulation cloth intelligent toy mouth shape control system based on action linkage, comprising:

[0120] The simulation cloth mouth shape control component module is used to adopt flexible cloth and control driving mechanism to form a three-dimensional mouth shape effect by adjusting the deformation of the cloth through an intelligent control system. The material selection of the cloth adopts flexibility and adjustability to respond to the control signal.

[0121] The action perception module is used to monitor the user's actions in real time. The sensors include but are not limited to pressure sensors, acceleration sensors, and infrared sensors. The user's touch, press, and shake actions are detected through these sensors, and the perception data is transmitted to the control unit.

[0122] The intelligent control unit module is used to receive data from the action perception module and generate corresponding control signals according to the preset action linkage algorithm, instructing the driving simulation cloth mouth shape control component to execute the mouth shape change.

[0123] A power management system module is used to adopt a low-power design, with a built-in rechargeable battery, automatically adjusting power consumption. The system dynamically adjusts power according to usage, and sends a warning signal when the battery is low;

[0124] A communication interface module is used to connect with smart devices through wireless communication function, for remote control and system upgrade. Users can adjust and set through smart phones or other devices.

[0125] The flexible cloth mouth control component adopts flexible and adjustable cloth, allowing the toy to present a more natural and dynamic mouth shape. Through the intelligent control system, the deformation of the cloth can be adjusted to achieve complex and flexible mouth movements with higher simulation. The motion sensing module monitors user movements in real time through various sensors. The intelligent control unit can obtain data from the motion sensing module and generate corresponding control signals according to the preset motion linkage algorithm. The power management system adopts a low-power design and has a built-in rechargeable battery. The communication interface module connects with smart devices through wireless communication function, allowing users to remotely control the toy through smart phones or other devices for convenient and efficient settings.

[0126] A simulation cloth intelligent toy mouth control method based on motion linkage includes the following steps:

[0127] The motion sensing module captures user motion data in real time through various sensors, including pressing, shaking, and approaching actions.

[0128] The collected motion data is preprocessed by the signal processing module, including denoising, signal filtering, and data normalization.

[0129] The processed motion data is transmitted to the intelligent control unit, which uses a preset motion recognition algorithm to analyze the data, recognize the user's actions, and generate corresponding mouth control instructions. According to the analysis results, the control unit transmits the generated instructions to the simulation cloth mouth control component to drive the cloth material to form a three-dimensional mouth shape.

[0130] The system adjusts the motion recognition algorithm adaptively based on historical usage data and user feedback, continuously optimizing the mouth control effect by learning user operation habits.

[0131] During use, the power management system automatically adjusts the system's power based on actual usage. When the battery is low, the system sends a warning signal and automatically starts the power saving mode.

[0132] The system adjusts the sensitivity of the sensors and the motion recognition algorithm in real time based on surrounding temperature, humidity, and light environmental conditions.

[0133] Through the communication interface module, users can remotely control the function settings of the toy and adjust the details of the mouth shape control. Through the wireless remote update function, the system can perform remote software upgrades regularly or as needed.

[0134] The motion perception module uses a variety of sensors to capture user movements in real time and can accurately reflect user interactions; the signal processing module processes sensor data through steps such as denoising, signal filtering, and data normalization to ensure that the system can work stably in different environments, reduce false movements, and improve response accuracy; the system uses historical usage data and user feedback to adaptively learn and adjust using motion recognition algorithms; the power management system has a low-power design and can dynamically adjust system power according to actual usage to maximize battery life; the system can adjust the sensor sensitivity and motion recognition algorithm in real time according to ambient environmental conditions such as temperature, humidity, and light; the communication interface module realizes wireless communication functions, and users can remotely control toys through smart devices and adjust the details of mouth control. The wireless remote update function enables the system to perform remote software upgrades as needed to maintain continuous innovation and improvement of system functions.

[0135] The various sensors in the motion sensing module capture the user's motion data in real time, including pressing, shaking, and approaching motions. The method is as follows:

[0136] The pressure sensor is deployed on the external surface of the toy. When the user presses the toy, the sensor senses the pressure change on the surface and outputs the corresponding pressure value;

[0137] The accelerometer is used to capture shaking motions and detect changes in the acceleration of an object in three-dimensional space. When the user shakes the toy, the sensor senses the dynamic acceleration changes of the object and identifies the intensity, direction, and frequency of the shaking motion.

[0138] The infrared sensor detects the distance between the object and the sensor by reflecting the infrared beam. When the user brings the hand or other object close to the toy, the infrared sensor recognizes the approach of the object and triggers the corresponding control signal;

[0139] The sensor collects motion data in real time, records pressure values, acceleration data, infrared distance value output signals, and transmits these data to the signal processing module.

[0140] By capturing the user's actions in real time, the toy can dynamically adjust its response based on the user's behavior, providing a more interactive and personalized experience; different sensors are responsible for different types of actions, which can enable the toy to make accurate responses even when faced with complex operations; real-time data collection by sensors enables the toy to accurately reflect the user's operation; after the data is collected by the sensors, it is transmitted to the signal processing module, and after processing and analysis, the system can remove noise, optimize data, and ensure more accurate action responses; the system can adjust the sensitivity of the sensors according to the user's operation habits and different environmental conditions; through real-time data collection and processing, the system can determine when to enter energy-saving mode or adjust power to avoid excessive power consumption, thereby extending the battery life.

[0141] The collected action data is preprocessed by the signal processing module, and the preprocessing step includes denoising, signal filtering, and data normalization methods:

[0142] Smooth the data by sliding window averaging method, for each data point, calculate the average value within a certain range of its neighborhood, sort the neighborhood around the data point, and select the middle value to eliminate outliers;

[0143] Low-pass filtering: filter out high-frequency components and retain low-frequency signals;

[0144] High-pass filtering: filter out low-frequency components and retain high-frequency change signals;

[0145] Combine low-pass and high-pass filtering to retain signals within a specific frequency band;

[0146] The data normalization method includes:

[0147] Normalize the data by mapping it to a specific range, the formula is:

[0148]

[0149] The above x is the original data, min(x) and max(x) are the minimum and maximum values in the data set, respectively;

[0150] Normalize the data by its mean and standard deviation, so that the mean is 0 and the standard deviation is 1, the formula is:

[0151]

[0152] The above x is the original data, \mu is the mean, and \sigma is the standard deviation.

[0153] The sliding window average method is a simple and effective smoothing method that smooths the original data by calculating the average value of the neighborhood data within a certain range before and after each data point. The low-pass filter filters out high-frequency signals and only retains low-frequency parts. High-frequency signals usually represent noise or rapid, irrelevant changes. Low-pass filtering can help remove these unnecessary high-frequency noises. The high-pass filter filters out low-frequency parts and retains high-frequency signals. High-frequency signals usually represent rapid, abrupt changes and can help capture subtle motion details such as rapid gestures or slight vibrations. By combining low-pass and high-pass filters, the system can retain signals within a specific frequency band, allowing both slower, stable signals and rapid, instantaneous changes to be retained. This combination makes signal processing more flexible and enables comprehensive capture of different characteristics of user actions.

[0154] The processed motion data is transmitted to the intelligent control unit, which analyzes the data using a preset motion recognition algorithm to identify the user's motion and generates corresponding mouth shape control instructions. According to the analysis result, the control unit transmits the generated instructions to the simulated fabric mouth shape control component to drive the fabric material to form a three-dimensional mouth shape.

[0155] The intelligent control unit analyzes the transmitted motion data and uses a preset motion recognition algorithm to determine the user's motion type. The motion recognition algorithm includes classification algorithms and neural networks.

[0156] The result of motion recognition is represented as a certain control instruction vector, with the formula being:

[0157] A = f(D)

[0158] Where D represents the pre-processed motion data, f(D) represents the motion result identified by the algorithm, and A is the motion instruction after recognition.

[0159] According to the recognition result, the corresponding mouth shape control instruction is generated. Mouth shape control simulates the deformation of fabric through a kinematic model to form a predetermined three-dimensional mouth shape. The control model for establishing the mouth shape is based on the physical properties of the fabric. The control instruction depends on the changes in these properties. The control instruction C is represented by the following formula:

[0160] C = g(A, P)

[0161] Where A is the instruction after motion recognition, P is the physical parameter of the fabric, and g(A, P) is the control instruction generation function.

[0162] The generated mouth shape control instruction C is transmitted to the simulated fabric mouth shape control component to drive the fabric material to form a three-dimensional mouth shape. Based on the method of finite element analysis, the deformation process of the fabric is simulated.

[0163] The displacement vector of the cloth is set as U, representing the displacement of the cloth, and the deformation of the cloth is described by the following mechanical formula:

[0164] F = KU + F ext

[0165] Where F is the internal force of the cloth, K is the stiffness matrix of the cloth, U is the displacement vector of the cloth, F ext is the external force;

[0166] Finally, the mouth shape control command C will form a three-dimensional mouth shape through the driving model related to the physical properties of the cloth, and through the dynamic simulation system, the cloth will change according to the calculated control command, and finally form the user's desired mouth shape.

[0167] By using classification algorithms and neural networks, real-time analysis and accurate identification of user actions are ensured, ensuring the efficiency and accuracy of data processing; combined with the physical properties of the cloth, accurate control commands are generated to ensure the natural change of the mouth shape; through the finite element analysis method, the accurate deformation of the cloth under the action of external force is simulated to ensure the dynamic reality of the mouth shape; the system can adjust the cloth form according to real-time input data, so that the mouth shape always meets the user's expectations, enhancing the reality of interactive experience.

[0168] The system adjusts the action recognition algorithm adaptively according to historical usage data and user feedback, and continuously optimizes the mouth shape control effect by learning the user's operation habits:

[0169] The system continuously records the user's operation data and feedback information, and the action recognition algorithm adjusts in real time according to the new data, and with the passage of time, the system gradually adjusts and optimizes the algorithm parameters;

[0170] The learning process of the system optimizes the action recognition function f by minimizing the error, and the incremental learning is represented by the following formula:

[0171]

[0172] Where f t is the action recognition function of the current model; η is the learning rate; L(D t , θ t ) is the loss function;

[0173] The Q-learning method in reinforcement learning is represented as:

[0174]

[0175] Where Q(s i , a t ) is the value of performing action a t in state s t .

[0176] a is a learning rate;

[0177] r t is the current feedback reward;

[0178] g is a discount factor;

[0179] max a Q(s t|1 , a) is the maximum Q-value in the next state;

[0180] By analyzing historical data and user feedback, the system automatically selects and adjusts the current user's action recognition algorithm. Different users have different operation habits, and the system selects from multiple algorithms to gradually optimize the accuracy and responsiveness of the algorithm. Algorithm selection includes model selection and hyperparameter adjustment.

[0181] Let the algorithm used by the system be A, and each algorithm has a set of hyperparameters A Optimize the hyperparameters through Bayesian optimization method:

[0182]

[0183] where * is the optimal set of hyperparameters, and P(Feedback|θ A ) is the posterior probability based on feedback information.

[0184] Collaborative filtering uses a user-action matrix R, where each user's feedback on an action is an element of the matrix. The goal of the algorithm is to minimize the error between the predicted feedback and the actual feedback:

[0185]

[0186] where U and V are the latent factor matrices for users and actions respectively, is the predicted feedback of user i on action j.

[0187] By minimizing the error, the system can adjust the action recognition function in real time, so that the algorithm is gradually optimized and the system's adaptive ability is improved. Reinforcement learning method optimizes the decision-making process by analyzing user feedback, so that the system can better understand and adapt to different users' operation behavior. Bayesian optimization method automatically adjusts the hyperparameters of the algorithm to ensure that the system converges to the best performance quickly with less data. Through the user-action matrix, collaborative filtering technology analyzes historical data and generates personalized feedback prediction, improving the system's response speed and accuracy.

[0188] In use, the power management system automatically adjusts the power of the system according to actual use, when the battery power is low, the system sends a warning signal, and automatically starts the power saving mode method is:

[0189] The remaining battery power is calculated in real time by monitoring the battery voltage and current, the battery power E(t) changes with time, which is estimated by the following formula:

[0190]

[0191] Where E0 is the initial battery power, I(τ) is the current at time τ, and V(τ) is the battery voltage;

[0192] During system operation, the power management system automatically adjusts power consumption according to actual usage requirements, and the power management method includes dynamic voltage frequency adjustment, task scheduling optimization, dynamic adjustment of hardware and software;

[0193] The formula is:

[0194] P total (t)=P CPU (t)+P Display (t)+P Peripherals (t)

[0195] Where P total (t) is the total power consumption of the system; P CPU (t), P Display (t) and P Peripherals (t) are the power consumptions of the processor, battery display and peripheral modules, respectively;

[0196] When the battery power is lower than the preset threshold, the power management system detects this change through the battery management system and sends a warning signal, the threshold is set to 20% or less of the battery power, when the low power is detected, the system sends a warning signal or starts the power saving mode;

[0197] Let the system battery power be E(t), when E(t) < E theshold , the system sends a warning and starts the power saving mode, the power consumption limit is set to P limit , the triggering of the warning and power saving mode is represented as:

[0198] If E(t) < E threshold , P total (t) → P limit

[0199] When the power saving mode is started, the system takes measures to reduce power consumption such as reducing screen brightness, limiting background tasks, turning off Bluetooth, Wi-Fi, GPS hardware components, and reducing processor frequency;

[0200] The power consumption model in the power saving mode is expressed as:

[0201]

[0202] wherein, a is a coefficient less than 1;

[0203] When the battery power is sufficient, the system automatically restores to the normal mode and restores the original power setting. During the restoration process, the system gradually increases the power consumption according to the battery power to ensure the restoration to the normal operating state, and the formula is:

[0204] If E(t)>E threshold , P total (t)→P normal

[0205] wherein, P normal is the power in the normal operating mode.

[0206] By dynamically adjusting the system power and starting the power saving mode, the device usage time can be maximally prolonged under low power condition. The system automatically adjusts the power consumption according to the actual usage and the battery state to avoid invalid energy waste. The timely warning of low power and the automatic starting of the power saving mode enable the user to make good battery management in advance, avoid sudden shutdown, and increase the reliability of device usage. By gradual restoration and timely power limitation, the pressure on the battery is reduced, thereby effectively prolonging the life cycle of the battery. The system supports dynamic voltage frequency adjustment, task scheduling optimization, and hardware component adjustment, and other methods, and performs real-time optimization according to the battery power and usage demand to ensure the balance of performance and energy efficiency.

[0207] The method for the system to adjust the sensitivity of the sensor and the motion recognition algorithm in real time according to the ambient temperature, humidity, and light environment conditions is:

[0208] The system collects environmental data in real time through the sensor, and dynamically adjusts the sensitivity of the sensor according to the environmental data;

[0209] According to the collected environmental data, the system adjusts the sensitivity of the sensor in real time, and the adjustment method includes temperature compensation, humidity correction, and light self-adaptive adjustment;

[0210] The sensitivity of the sensor used by the system is S(t), the temperature, humidity, and light intensity are T(t), H(t), and L(t) respectively, and the sensitivity adjustment is expressed as:

[0211] S(t)=S0+αT(t)+βH(t)+γL(t)

[0212] wherein, S0 is the default sensitivity; a, β, and γ are adjustment coefficients;

[0213] In addition to adjusting the sensitivity of the sensor, the system dynamically optimizes the motion recognition algorithm based on environmental conditions. The environmental factors are based on the quality of the sensor data, and the algorithm used is equipped with adaptive capabilities to handle environmental changes that interfere with the data.

[0214] The accuracy of the motion recognition algorithm is based on the sensor data D(t) related to environmental conditions T(t), H(t), L(t). The adjustment of the algorithm is achieved by modifying the weight coefficient, w(T, H, L). The system uses a machine learning model to recognize motion, represented by the following model:

[0215] A(t) = f(D(t), w(T(t), H(t), L(t)))

[0216] Where A(t) is the motion recognition result, f(·) is the motion recognition algorithm, and w(T, H, L) is the adjustment coefficient under environmental conditions.

[0217] The system updates the adjustment strategy based on historical environmental data D history and real-time data D current , represented by:

[0218] w(T(t), H(t), L(t)) = g(D history , D current )

[0219] Where g(·) is the adjustment strategy update function, which adjusts the sensor sensitivity and motion recognition algorithm in real-time.

[0220] By dynamically adjusting the sensitivity of the sensor, the system can maintain high-precision data acquisition capability under various environmental changes, effectively compensating for the impact of environmental changes on the sensor, thereby ensuring the reliability of the data. The system adjusts the motion recognition algorithm in real-time based on the quality of the environmental data, enhancing the system's adaptability to external changes and improving the accuracy of motion recognition, especially in complex environments. By combining a machine learning model, the system can continuously update the adjustment strategy based on historical and real-time data. Each adjustment and optimization makes the system more intelligent and able to adapt to a wider range of application scenarios. By adjusting the sensitivity of the sensor in real-time, the system can reduce the additional energy consumption and errors caused by environmental interference, improve energy efficiency, and thus prolong the service life of the sensor and the overall system.

[0221] Through the communication interface module, users can remotely control the function settings of the toy and adjust the details of the mouth control. The system periodically or according to demand performs remote software upgrades through wireless remote update functions:

[0222] Through the Wi-Fi, Bluetooth, ZigBee wireless communication interface module, the system is remotely connected with the user equipment, and the user sends instructions through the control equipment to adjust the functions and settings of the toy.

[0223] The user remotely controls the function settings of the toy through wireless communication, including but not limited to the adjustment of the details of the mouth shape control, the enablement and disablement of the function.

[0224] The system regularly or according to the demand performs software upgrading, and the process includes the following steps:

[0225] The toy regularly checks whether there is a new software version through the communication interface module. If it is a regular check, the toy automatically connects the server according to the preset time interval to query whether there is an available upgrade package.

[0226] Through the control interface, the user selects to start the software upgrading function.

[0227] The toy downloads a new software update package through the communication interface module and verifies the integrity and security of the downloaded file.

[0228] After verification, the toy performs the upgrading operation in a safe mode.

[0229] A rollback mechanism is set. If a problem occurs during the upgrading process, the system automatically rolls back to the previous stable version.

[0230] The system includes a server and multiple toy devices. The server is responsible for storing and distributing the latest software update package and also receives the control instructions sent by the user. Each toy device is connected with the server through wireless communication and performs remote control and software updating according to the demand.

[0231] The user can remotely control the function settings of the toy through wireless communication without personally operating the toy. Regular software updating can extend the functions and improve the performance of the toy with the progress of technology or changes in demand. Through wireless remote updating, the hardware of the toy does not need to be updated, and new functions or bug fixes can be obtained through software updating, thereby greatly prolonging the life cycle of the toy. Through the safe mode and the rollback mechanism, the system can effectively avoid the failure of the toy function caused by the failure of software upgrading or problems. Through the wireless communication interface, the toy device can automatically connect with the server and check the update, simplifying the operation process of the user. Traditional toys need to be upgraded or maintained through offline mode. Through remote wireless upgrading, the manufacturer does not need to provide update services to the user face to face, which saves time and reduces cost.

[0232] The embodiment also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the action linkage-based simulation cloth intelligent toy mouth shape control system and method when executing the program.

[0233] The embodiment also provides a computer readable storage medium, which stores a computer program, wherein the program is executed by a processor to implement the action linkage-based simulation cloth intelligent toy mouth shape control system and method.

[0234] The sensor unit is used for sensing the user action and converting it into an electrical signal, and the sensor unit comprises at least one sensor for detecting physical changes;

[0235] In the process of implementing the present application, the action sensing module comprises an infrared sensor, an acceleration sensor and a pressure sensor, the infrared sensor is used for detecting the distance change between the user and the toy and providing proximity sensing, the acceleration sensor is used for capturing the action acceleration of the user and dynamically adjusting the mouth shape according to the gesture change of the user, and the pressure sensor is used for sensing the touch pressure of the user and triggering the corresponding reaction;

[0236] The intelligent control unit adjusts the control signal through an adaptive algorithm according to the received sensor data.

[0237] In the present application, the intelligent control unit adopts a neural network-based adaptive learning algorithm, which is firstly trained according to an initial data set and is online learned and optimized through continuous interaction data of the user.

[0238] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of each method can be included. Any reference to memory, storage, database or other medium used in each embodiment provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0239] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the system is divided into different functional units or modules to complete all or part of the above-mentioned functions.

[0240] The above-mentioned embodiments of the present application are not a limitation on the protection scope of the present application, and the embodiments of the present application are not limited thereto. According to the above-mentioned content of the present application, other various forms of modification, replacement or change of the above-mentioned structure of the present application can be made according to ordinary technical knowledge and conventional means in the art without departing from the above-mentioned basic technical idea of the present application, which should fall within the protection scope of the present application.

Claims

1. A mouth shape control system for a simulated cloth intelligent toy based on motion linkage, characterized in that: Includes: The simulation fabric mouth shape control component module is composed of flexible fabric and a control drive mechanism. The deformation of the fabric is adjusted through an intelligent control system. The fabric is selected for its flexibility and adjustability, and responds to the control signal to form a three-dimensional mouth shape effect. The motion sensing module is used to monitor the user's movements in real time. The sensors include but are not limited to pressure sensors, acceleration sensors, and infrared sensors. These sensors detect the user's touch, press, and shake movements and transmit the sensing data to the control unit. The intelligent control unit module is used to receive data from the motion sensing module and generate corresponding control signals according to the preset motion linkage algorithm, instructing the driving of the simulated fabric mouth shape control component to execute mouth shape changes; A power management system module, designed to use low-power design, built-in rechargeable batteries, and automatically regulate power consumption. The system dynamically adjusts power based on usage and issues a warning signal when the battery is low; The communication interface module is used to connect to smart devices using wireless communication functions for remote control and system upgrades, and users can make adjustments and settings through smartphones or other devices.

2. A method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage, characterized in that: The following steps are involved: Through the various sensors in the motion sensing module, the user's motion data is captured in real time, including pressing, shaking, and approaching movements; The collected motion data is preprocessed by the signal processing module. The preprocessing steps include denoising, signal filtering, and data normalization. The processed motion data is transmitted to the intelligent control unit, which uses a preset motion recognition algorithm to analyze the data, identify the user's motion, and generate corresponding mouth shape control instructions. Based on the analysis results, the control unit transmits the generated instructions to the simulated fabric mouth shape control component to drive the fabric material to form a three-dimensional mouth shape; The system adaptively adjusts the motion recognition algorithm based on historical usage data and user feedback. By learning the user's operating habits, the system continuously optimizes the mouth shape control effect. During use, the power management system automatically adjusts the system power according to actual usage. When the battery power is low, the system sends a warning signal and automatically starts the power saving mode. The system adjusts the sensor sensitivity and motion recognition algorithm in real time according to the surrounding temperature, humidity, and lighting conditions; Through the communication interface module, users can remotely control the function settings of the toy and adjust the details of the mouth shape control. Through the wireless remote update function, the system can perform remote software upgrades regularly or as needed.

3. The method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage according to claim 2, characterized in that: The various sensors in the motion sensing module capture the user's motion data in real time, including pressing, shaking, and approaching motions. The method is as follows: The pressure sensor is deployed on the external surface of the toy. When the user presses the toy, the sensor senses the pressure change on the surface and outputs the corresponding pressure value; The accelerometer is used to capture shaking motions and detect changes in the acceleration of an object in three-dimensional space. When the user shakes the toy, the sensor senses the dynamic acceleration changes of the object and identifies the intensity, direction, and frequency of the shaking motion. The infrared sensor detects the distance between the object and the sensor by reflecting the infrared beam. When the user brings the hand or other object close to the toy, the infrared sensor recognizes the approach of the object and triggers the corresponding control signal; The sensor collects motion data in real time, records pressure values, acceleration data, infrared distance value output signals, and transmits these data to the signal processing module.

4. The method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage according to claim 3, characterized in that: The collected motion data is preprocessed by the signal processing module. The preprocessing steps include denoising, signal filtering, and data normalization. The method is as follows: Smooth the data using the sliding window averaging method. For each data point, calculate the average value of its surrounding neighborhood within a certain range. Then sort the neighborhood around the data point and select the middle value to eliminate outliers. Low-pass filtering: retains low-frequency signals by filtering out high-frequency components; High-pass filtering: By filtering out low-frequency components, retaining high-frequency changing signals; Combine low-pass and high-pass filtering to retain signals within a specific frequency band; Methods for data normalization include: Normalization is performed by mapping the data to a specific range using the formula: The above x is the original data, min(x) and max(x) are the minimum and maximum values ​​in the data set respectively; Normalize the data by its mean and standard deviation so that its mean is 0 and its standard deviation is 1. The formula is: The above x is the original data, \mu is the mean, and \sigma is the standard deviation.

5. The method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage according to claim 4, characterized in that: The processed motion data is transmitted to the intelligent control unit, which uses a preset motion recognition algorithm to analyze the data, identify the user's motion, and generate corresponding mouth shape control instructions. Based on the analysis results, the control unit transmits the generated instructions to the simulated fabric mouth shape control component to drive the fabric material to form a three-dimensional mouth shape. The method is as follows: The intelligent control unit analyzes the transmitted motion data and uses a preset motion recognition algorithm to determine the user's motion type. The motion recognition algorithm includes a classification algorithm and a neural network. The result of action recognition is expressed as a control instruction vector, the formula is: A=f(D) Where D represents the pre-processed action data, f(D) represents the action result identified by the algorithm, and A is the action instruction after identification; Based on the recognition results, the corresponding mouth shape control instructions are generated. The mouth shape control simulates the deformation of the cloth through the kinematic model to form a predetermined three-dimensional mouth shape. The mouth shape control model is based on the physical properties of the cloth. The control instructions depend on the changes in these properties. The control instruction C is expressed by the following formula: C=g(A,P) Among them, A is the instruction after action recognition, P is the physical parameter of the cloth, and g(A, P) is the control instruction generation function; The generated mouth shape control command C is transmitted to the simulated cloth mouth shape control component, which drives the cloth material to form a three-dimensional mouth shape. Based on the finite element analysis method, the deformation process of the cloth is simulated; The displacement vector of the cloth is set as U, which represents the displacement of the cloth. The deformation of the cloth is described by the following mechanical formula: F=KU+F ext Among them, F is the internal force of the cloth, K is the stiffness matrix of the cloth, U is the displacement vector of the cloth, F ext For external forces; Ultimately, the mouth shape control instruction C will form a three-dimensional mouth shape through a driving model related to the physical properties of the cloth. Through the dynamic simulation system, the cloth will change according to the calculated control instruction, and finally form the mouth shape expected by the user.

6. The method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage according to claim 5, characterized in that: The system adaptively adjusts the motion recognition algorithm based on historical usage data and user feedback. By learning the user's operating habits, the system continuously optimizes the mouth shape control effect in the following ways: The system continuously records user operation data and feedback information, and the action recognition algorithm is adjusted in real time based on the new data. Over time, the system gradually adjusts and optimizes the algorithm parameters; The learning process of the system is established to optimize the action recognition function f by minimizing the error. The incremental learning is expressed by the following formula: Among them, f t is the action recognition function of the current model; η is the learning rate; L(D t ,θ t ) is the loss function; The Q-learning method in reinforcement learning is expressed as: Among them, Q(s t , a t ) is in state s t Next, perform action a t the value of α is the learning rate; r t Reward for current feedback; γ is the discount factor; max a Q(s t|1 ,a) is the maximum Q value in the next state; By analyzing historical data and user feedback, the system automatically selects and adjusts the current user's action recognition algorithm. Different users have different operating habits. The system selects from multiple algorithms and gradually optimizes the accuracy and responsiveness of the algorithm. Algorithm selection includes model selection and hyperparameter adjustment. The algorithm used in the system is A, and each algorithm has a set of hyperparameters θ A , hyperparameter optimization is performed using the Bayesian optimization method: Among them, θ * is the optimal hyperparameter set, P(Feedback|θ A ) is the posterior probability based on feedback information; Collaborative filtering uses a user-action matrix R, where each user's feedback on an action is used as a matrix element. The goal of the algorithm is to minimize the error between the predicted feedback and the actual feedback: Among them, U and V are the latent factor matrices of users and actions respectively, is the predicted feedback for user i on action j.

7. The method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage according to claim 6, characterized in that: During use, the power management system automatically adjusts the system power according to actual usage. When the battery power is low, the system issues a warning signal and automatically activates the power saving mode. The method is as follows: The remaining capacity is calculated in real time by monitoring the battery voltage and current. The battery capacity E(t) changes over time and is estimated using the following formula: Where E0 is the initial charge of the battery, I(τ) is the current at time τ, and V(τ) is the battery voltage; During system operation, the power management system automatically adjusts power consumption according to actual usage needs. Power management methods include dynamic voltage and frequency adjustment, task scheduling optimization, and dynamic adjustment of hardware and software. The formula is: P total (t)=P CPU (t)+P Display (t)+P Peripherals (t) Among them, P total (t) is the total power consumption of the system; P CPU (t), P Display (t) and P Peripherals (t) are the power consumption of the processor, battery display and peripheral modules respectively; When the battery level drops below a preset threshold, the power management system detects this change through the battery management system and issues a warning signal. The threshold is set at 20% or less of the battery level. When low battery is detected, the system issues a warning signal or activates power saving mode. Set the system battery power as E(t), when E(t)<E threshold , the system issues a warning and starts the power saving mode, setting the power consumption limit to P limit , the triggering of warning and power saving mode is expressed as: If E(t)<E threshold ,P total (t)→P limit When power saving mode is activated, the system reduces power consumption by lowering screen brightness, limiting background tasks, turning off Bluetooth, Wi-Fi, and GPS hardware components, and reducing processor frequency. The power consumption model in power saving mode is expressed as: Among them, α is a coefficient less than 1; When the battery is fully charged, the system automatically returns to normal mode and restores the original power settings. During the recovery process, the system gradually increases power consumption according to the battery charge to ensure that it returns to normal operation. The formula is: If E(t)>E threshold ,P total (t)→P normal Among them, P normal is the power in normal operating mode.

8. The method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage according to claim 7, characterized in that: The system adjusts the sensor sensitivity and motion recognition algorithm in real time according to the ambient temperature, humidity, and lighting conditions as follows: The system collects environmental data in real time through sensors and dynamically adjusts the sensitivity of the sensors based on the environmental data; Based on the collected environmental data, the system adjusts the sensitivity of the sensor in real time, including temperature compensation, humidity correction, and adaptive light adjustment; The sensor sensitivity used in the system is S(t), and the temperature, humidity, and light intensity are T(t), H(t), and L(t) respectively. The sensitivity adjustment is expressed as: S(t)=S0+αT(t)+βH(t)+γL(t) Among them, S0 is the default sensitivity; α, β, γ are adjustment coefficients; In addition to adjusting the sensor sensitivity, the system dynamically optimizes the motion recognition algorithm based on environmental conditions. Environmental factors are based on the quality of sensor data, and the algorithm used is equipped with adaptive capabilities to handle the interference of environmental changes on the data. The accuracy of the action recognition algorithm is based on the influence of sensor data D(t) and is related to environmental conditions T(t), H(t), and L(t). The algorithm is adjusted by modifying the weight coefficients w(T, H, L). The system uses a machine learning model to recognize actions, which is represented by the following model: A(t)=f(D(t),w(T(t),H(t),L(t))) Where A(t) is the action recognition result, f(·) is the action recognition algorithm, and w(T, H, L) is the adjustment coefficient under environmental conditions; Establish a system based on historical environmental data history and real-time data D current Update adjustment strategy, expressed as: w(T(t),H(t),L(t))=g(D history ,D current ) Among them, g(·) is the adjustment strategy update function, which adjusts the sensor sensitivity and action recognition algorithm in real time.

9. The method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage according to claim 8, characterized in that: Through the communication interface module, users can remotely control the function settings of the toy and adjust the details of the mouth shape control. Through the wireless remote update function, the system can perform remote software upgrades regularly or as needed. The method is as follows: Through Wi-Fi, Bluetooth, and ZigBee wireless communication interface modules, the system is remotely connected to the user's device, and the user can send commands through the control device to adjust the functions and settings of the toy; The user remotely controls the function settings of the toy via wireless communication, including but not limited to detailed adjustments to the mouth shape control, and enabling and disabling functions; The system performs software upgrades regularly or as needed. The process includes the following steps: The toy regularly checks whether there is a new software version through the communication interface module. If it is a regular check, the toy automatically connects to the server according to the preset time interval to check whether there is an available upgrade package; Through the control interface, the user chooses to start the software upgrade function; The toy downloads the new software update package through the communication interface module and verifies the integrity and security of the downloaded file; After verification, the toy will be upgraded in safe mode; Set up a rollback mechanism. If a problem occurs during the upgrade process, the system will automatically roll back to the previous stable version. The system includes a server and multiple toy devices. The server is responsible for storing and distributing the latest software update packages, and also receives control instructions sent by users. Each toy device maintains a connection with the server through wireless communication and performs remote control and software updates as needed.

10. An electronic device, characterized in that: The invention comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage as claimed in any one of claims 2 to 9 is implemented; A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a method for controlling the mouth shape of a simulated cloth intelligent toy based on motion linkage as described in any one of claims 2 to 9 is implemented.

Images