Intelligent toothbrush method and system based on method pool and self-adaptive coordination impedance control

Abstract

The invention discloses an intelligent toothbrush method and system based on a method pool and self-adaptive coordination impedance control. The method comprises the steps that the method pool containing multiple cleaning strategies is constructed, multi-mode sensor data in the tooth brushing process are collected in real time, preprocessing and feature extraction are conducted on the multi-mode sensor data, and a state vector is constructed; dynamically matching and scheduling an optimal cleaning strategy from a method pool according to the state vector; based on a target impedance parameter corresponding to the optimal cleaning strategy, through a self-adaptive law based on Lyapunov stability, calculating a target torque so as to adjust a dynamic interaction force between a brush head and teeth in real time; and controlling a motor to execute tooth brushing action according to the calculated target torque. According to the invention, the defect that cleaning mode selection and brush head force control are mutually separated in the prior art is overcome, seamless connection from what to how to make is realized, and the real-time response precision and execution consistency of a cleaning strategy are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent oral care, specifically relating to an intelligent toothbrush method and system based on method pool and adaptive coordinated impedance control. Background Technology

[0002] The intelligent development of electric toothbrushes has now entered the sensing and control stage. In existing technologies, some products attempt to improve adaptability by integrating basic sensors (such as pressure sensors), for example, automatically reducing vibration frequency or issuing a warning when excessive brushing pressure is detected. Other solutions propose automatically switching cleaning modes based on the handle angle or preset time. However, these existing technologies have fundamental limitations: First, their control logic largely relies on preset fixed rules or simple threshold judgments, such as "if the pressure exceeds threshold X, switch to mode Y." This "if-then" control cannot make fine-grained and continuous adjustments based on the complexity and dynamic changes of the oral environment (such as differences in plaque adhesion and gum sensitivity in different tooth areas). Second, existing technologies generally separate the selection of cleaning modes (macro-strategy) from the execution of brush head force (micro-control), lacking a collaborative control framework that can seamlessly connect the two. This means that the system cannot simulate the human-like ability of a dentist to "use different strategies for different areas and fine-tune the force based on tactile feedback at every moment" when using hand tools. As a result, it is difficult to ensure gentle care and absolute safety for sensitive areas while efficiently cleaning all teeth. Summary of the Invention

[0003] A brief overview of embodiments of the invention is provided below to provide a basic understanding of certain aspects of the invention. It should be understood that this overview is not an exhaustive summary of the invention. It is not intended to identify key or essential parts of the invention, nor is it intended to limit the scope of the invention. Its purpose is merely to present certain concepts in a simplified form as a prelude to the more detailed description that follows.

[0004] This invention provides a smart toothbrush method and system based on a method pool and adaptive coordinated impedance control, belonging to the field of intelligent oral care. The core innovation lies in constructing a dynamic cleaning strategy pool, which pre-stores multiple cleaning strategies for different tooth locations and oral conditions. Simultaneously, an adaptive coordinated impedance controller is introduced, whose reference target is generated by the method pool scheduling. By adjusting the dynamic interaction force between the brush head and teeth in real time, the brush head exhibits different "feelings" (e.g., gentle, strong). The system senses brushing posture, contact force, and other states through multimodal sensors (e.g., IMU, pressure sensor). The method pool calls the optimal cleaning strategy based on this state and outputs the target impedance parameter. The impedance controller accurately tracks this target using an adaptive law based on Lyapunov stability, achieving synergy between macro-strategy and micro-force control. This invention solves the problems of rigid control strategies in existing electric toothbrushes, which cannot simultaneously achieve overall cleaning efficiency and localized gentle touch, realizing dynamic, personalized adaptive cleaning during a single brushing session.

[0005] According to a first aspect of this application, a smart toothbrush control method based on method pooling and adaptive coordinated impedance control is provided, comprising the following steps: Step 1: Construct a method pool containing multiple cleaning strategies, each with m... i Defined basic cleaning parameters (vibration frequency w) i , amplitude A i ), target impedance parameters (desired quality) Desired stiffness and expected damping ) and suggested duration parameters; Step 2: Collect multimodal sensor data in real time during the brushing process, preprocess and extract features to construct a state vector. The multimodal sensor data includes brush head posture data, contact pressure data and optical reflection data; the state vector includes brush head position, average contact pressure, plaque index, area cleaning time and user history preferences. Step 3: Based on the state vector, dynamically match and schedule the optimal cleaning strategy from the method pool. Specifically, this includes calculating the matching score of each cleaning strategy and selecting the cleaning strategy with the highest score as the current cleaning strategy. Step 4: Based on the target impedance parameters (desired quality) corresponding to the optimal cleaning strategy. Desired stiffness and expected damping By using an adaptive law based on Lyapunov stability, the target torque is calculated to try to adjust the dynamic interaction force between the brush head and the teeth; Step 5: Control the motor to perform the brushing action according to the target torque.

[0006] Furthermore, step 3, which involves dynamically matching and scheduling the optimal cleaning strategy from the method pool, comprehensively evaluates regional matching degree, cleaning demand degree, and user preference degree, specifically including: Step 31: Calculate the region fit score Determine whether the current brush head position falls within the cleaning strategy. Preset applicable area; if the current brush head position is within the cleaning strategy m i The preset applicable area, =1, otherwise 0; Step 32: Calculate the cleaning requirement score The cleaning requirement score is based on the current average contact pressure. With maximum safety pressure The ratio of plaque index I p The result is calculated by weighted summation of the ratio of the current cleaning time to the recommended duration. ; These are the weighting coefficients for pressure, plaque index, and cleaning time, which were calibrated experimentally. It is the greatest safety pressure. This represents the current average contact pressure. Plaque index, This is the cleaning time for the current area. This is the suggested duration.

[0007] Step 33: Calculate user preference scores ; Query user historical data to retrieve cleaning strategies from historical records The number of times N was successfully used in this area i And the total number of times N, all cleaning strategies were successfully used in the area. total Then the user preference score =N i / N total ; Step 34: Calculate the overall score for each cleaning strategy: ; Where w z w c w u Each cleaning strategy m i The corresponding weights for regional suitability score, cleaning need score, and user preference score. Select the largest The scoring method is used to determine the optimal cleaning strategy. Among them, the area matching degree assessment is based on the geometric fit between the current brush head position and the preset area of ​​the strategy; the cleaning demand degree comprehensively considers the current contact pressure, plaque index, and cleaning time; and the user preference degree refers to the user's historical preference data.

[0008] Furthermore, in step 4, the target torque is calculated using an adaptive law, which dynamically adjusts the impedance parameters (damping and stiffness parameters) based on the force tracking error. When excessive pressure is detected, the damping parameter is automatically increased and the stiffness parameter is decreased, making the brush head have a smooth feel to buffer the pressure; specifically, this includes: Obtain the desired quality based on the optimal cleaning strategy. Desired stiffness and expected damping And the amplitude A and frequency f in the basic cleaning parameters; Establish target torque Calculation formula: ; in, These are the desired position, desired velocity, and desired acceleration, respectively. , It is the amplitude. The frequencies are all provided by the method pool; It is the first derivative of the desired position. ; It is the second derivative of the expected position. ; and These represent the actual position and velocity of the brush head, indicating the true motion state of the brush head in the oral cavity. They are obtained from the brush head posture data in the multimodal sensor data. The multimodal sensor data includes brush head posture data, contact pressure data, and optical reflection data.

[0009] This represents the effective lever arm from the motor output shaft (or equivalent drive point) to the point of contact between the brush head and the teeth. It is a fixed value set before the smart toothbrush leaves the factory. Its function is to adjust the linear force calculated by the impedance controller and applied to the brush head. This is converted into the rotational torque required to drive the motor. .

[0010] The actual contact force, i.e. the contact force between the brush head and the teeth, is obtained from the contact pressure data of the multimodal sensor.

[0011] Furthermore, the adaptive law, based on the force tracking error, adjusts the damping and stiffness parameters of the impedance controller in real time. The formal definition of this law is as follows: ; in, This represents the rate of change of damping and stiffness parameters; It is the force tracking error, Fd is determined by the expected stiffness in the method pool. and expected damping Decide, ; For the damping coefficient and stiffness coefficient, i.e. The desired contact force set for the method pool; It is an adaptive gain, which determines the speed at which the damping parameter and stiffness parameter are adjusted, respectively; It is a saturation function used to ensure the boundedness of parameters. The calculation formula is as follows: .

[0012] Indicates the error parameter. This indicates the maximum acceptable error. This represents the minimum achievable error. As stated in the formula above, This can be expressed as a speed difference ( ) or location difference ( Their maximum and minimum errors were calibrated through dental experiments.

[0013] Furthermore, the control method also includes step 6: during the execution of the cleaning strategy, evaluating the cleaning effect based on sensor data and generating a reward signal to update the weights (w) of the scheduling strategy online. z w c w u This involves optimizing the strategy parameters (mass parameters, stiffness parameters, damping parameters, amplitude, frequency, cleaning duration, etc.) in the method pool. The cleaning effect is evaluated, and the weight coefficients of each cleaning strategy in the method pool are dynamically updated based on the cleaning effect data, enabling online learning and continuous optimization of the method pool.

[0014] Furthermore, the cleaning strategy parameters in the method pool are pre-calibrated through at least one of the following methods: oral medicine experiments, cleaning efficiency testing, professional dental operation data analysis, and impedance model simulation.

[0015] Furthermore, the brush head position mapping is achieved through a preset lookup table, which maps the posture sensor data to specific tooth regions, including the incisor region, canine region, premolar region, and molar region.

[0016] Furthermore, the plaque index I pBased on the optical reflection data (the intensity of reflected light detected by the optical reflection sensor) in the multimodal sensor data, the stronger the reflected light, the cleaner the tooth surface and the lower the plaque index. In other words, the higher the calculated plaque index value, the cleaner the tooth surface.

[0017] Furthermore, the scheduling strategy of the method pool is finely and continuously adjusted according to the complexity and dynamic changes of the oral cavity region, so as to achieve seamless connection of cleaning strategies for different tooth regions.

[0018] According to a second aspect of this application, a smart toothbrush system based on method pooling and adaptive coordinated impedance control is provided, comprising: The multimodal perception module is used to collect brush head posture, contact pressure and tooth surface cleanliness data (optical reflection data) in real time, and construct a state vector containing brush head position, average contact pressure, plaque index, area cleaning time and user historical preferences through feature extraction; The method pool management module has pre-stored a variety of cleaning strategies, including basic cleaning parameters, target impedance parameters and suggested durations, and can dynamically schedule and output the optimal cleaning strategy and its target impedance parameters according to the state vector. An adaptive impedance control module is used to receive target impedance parameters (desired mass, desired stiffness, and desired damping) provided by the optimal cleaning strategy, and, in conjunction with the real-time contact pressure obtained by the multimodal sensing module, calculate the target torque through an adaptive law based on Lyapunov stability, so as to adjust the dynamic interaction force between the brush head and the teeth in real time and track the target impedance parameters. And an execution module, used to control the motor to perform the brushing action according to the target torque.

[0019] Furthermore, the multimodal sensing module includes an attitude sensor, a pressure sensor, and an optical reflection sensor, wherein: The posture sensor is used to acquire brush head posture data and map the current brush head position to a specific tooth area using a preset lookup table. The pressure sensor is used to collect contact pressure data between the brush head and the teeth; The optical reflection sensor is used to detect the cleanliness of the tooth surface and to assess the degree of plaque adhesion by the intensity of reflected light.

[0020] Furthermore, the method pool management module evaluates and selects the optimal cleaning strategy by calculating a weighted sum of regional matching degree, cleaning demand degree, and user preference degree, specifically including: The area adaptation score is full when the current brush head position is within the applicable area preset by the cleaning strategy, otherwise it is zero. The cleaning demand score is calculated by weighting the average contact pressure, plaque index, and cleaning time in the state vector. User preference score is calculated based on the user's historical preference data in the state vector; The matching score is a weighted sum of the regional adaptability score, the cleaning requirement score, and the user preference score.

[0021] Furthermore, the adaptive impedance control module includes an adaptive law for impedance parameters. This law dynamically adjusts the actual stiffness and damping parameters based on the deviation between the desired force and the actual contact force, the adaptive gain, and the brush head speed. The adaptive law is used to update the damping parameters and stiffness in real time, and its update rule is related to the force tracking error and the state error of the brush head movement to ensure system stability and achieve adaptive adjustment of microforces.

[0022] Furthermore, the method pool management module also includes an execution evaluation and online learning mechanism, which dynamically updates the weight coefficients of each cleaning strategy based on the cleaning effect data of this brushing session, thereby achieving continuous optimization of the method pool.

[0023] Furthermore, the cleaning strategies in the method pool are divided into three modes based on cleaning intensity: a powerful cleaning mode, a standard cleaning mode, and a sensitive care mode, wherein: The powerful cleaning mode corresponds to high stiffness and low damping parameters, providing a firm tactile feel; The sensitive care mode corresponds to low stiffness and high damping parameters, providing a smooth touch.

[0024] Furthermore, the desired stiffness is described by an impedance model, which represents the relationship between the brush head movement speed and the desired contact force as a second-order system including mass, damping, and stiffness terms.

[0025] According to a third aspect of this application, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the above-described intelligent toothbrush control method based on method pool and adaptive coordinated impedance control.

[0026] Compared with the prior art, the present invention adopts the above solution and has the following advantages: 1. Achieved synergistic optimization of macro-level cleaning strategies and micro-level force control: In the design of the macro-level strategy library (method pool), existing technologies (such as cloud-based model scheduling or robot task planning) typically use abstract "capability-model" mapping tables or predefined rule bases, with the strategies themselves being relatively macro-level and abstract. For application in electric toothbrushes, this invention designs the method pool as a structured, interpretable, and directly hardware-driving parameterized instruction set. By constructing a method pool containing multiple cleaning strategies, the optimal cleaning strategy is dynamically scheduled based on real-time collected oral state vectors (including brush head position, average contact pressure, plaque index, area cleaning time, and user historical preferences). The strategy parameters (desired mass, desired stiffness, and desired damping) are input into the adaptive impedance control module to generate a continuously adjustable target torque. This technical solution overcomes the shortcomings of existing technologies where cleaning mode selection and brush head force control are disconnected, achieving seamless integration from what to do to how to do it well, significantly improving the real-time response accuracy and execution consistency of the cleaning strategy. This design enables cleaning strategies in oral medicine (such as the "Bass brushing technique" and its specific action on the gingival sulcus) to be precisely encoded into machine-executable dynamic instructions, achieving a direct and unambiguous conversion from medical knowledge to machine control.

[0027] 2. Significantly improved safety and smoothness during brushing: Utilizing an adaptive law based on Lyapunov stability theory, the damping and stiffness parameters of the impedance controller are adjusted in real time. When contact pressure exceeds the desired level, the system automatically increases damping and decreases stiffness, resulting in a smooth brush head. This continuous adaptive adjustment of impedance parameters buffers excessive pressure, rather than relying on simple start-stop control or threshold protection mechanisms. This technology effectively avoids sudden force changes and cleaning interruptions caused by traditional threshold control, providing continuous force buffering protection while ensuring cleaning power, thus reducing the risk of gum damage.

[0028] 3. Achieves highly personalized and dynamically adaptive oral cleaning: Through a dynamic scheduling mechanism of the method pool, it comprehensively evaluates regional matching degree, cleaning demand degree, and user preference, dynamically matching the optimal cleaning strategy for different oral regions, different cleaning stages, and different user habits. Simultaneously, through an adaptive law, it adjusts impedance parameters online based on instantaneous force tracking errors, enabling the dynamic interaction force between the brush head and teeth to continuously adapt to the complex changes in the oral environment. This technical solution overcomes the limitations of fixed parameters or preset modes in existing technologies, achieving dual adaptation at both the strategy and parameter levels, significantly improving the consistency of cleaning effects and user adaptability.

[0029] 4. Achieved high-performance control on a low-cost hardware platform: Both the method pool scheduling algorithm and the adaptive impedance control algorithm are implemented based on conventional embedded processors (such as the ARM Cortex-M4 core) and low-cost sensors (IMU, thin-film pressure sensor, optical reflection sensor). Precise force control and compliant interaction can be achieved without introducing additional high-precision force sensors or complex mechanical structures. Furthermore, the introduction of a saturation function in the adaptive impedance control algorithm is a crucial safety valve, ensuring the compliantness and predictability of control behavior, minimizing potential risks in close physical human-machine interaction, and achieving a balance between high real-time control and high safety requirements. This technical solution improves control performance on existing smart toothbrush hardware platforms through algorithmic innovation, exhibiting good technical compatibility and engineering feasibility. Attached Figure Description

[0030] The present invention can be better understood by referring to the description given below in conjunction with the accompanying drawings, in which the same or similar reference numerals are used throughout the drawings to denote the same or similar parts. These drawings, together with the following detailed description, are incorporated in and form part of this specification, and are used to further illustrate preferred embodiments of the invention and explain the principles and advantages of the invention. In the drawings: Figure 1 This is a flowchart of the intelligent toothbrush method of the present invention. Detailed Implementation

[0031] Embodiments of the present invention will now be described with reference to the accompanying drawings. Elements and features described in one drawing or embodiment of the invention may be combined with elements and features shown in one or more other drawings or embodiments. It should be noted that, for clarity, representations and descriptions of components and processes unrelated to the present invention and known to those skilled in the art have been omitted from the drawings and description.

[0032] Example 1 This embodiment provides a smart toothbrush method based on method pooling and adaptive coordinated impedance control. See [link to documentation]. Figure 1 The specific implementation steps are as follows: Step 1: Generating the macro strategy library for the method pool Method pool This is the macro-strategy library of the present invention. Each method This is a complete description of a specific cleaning task (such as "high-efficiency cleaning of the molar area" or "gingival sulcus massage").

[0033] These are the basic cleaning parameters. It is the vibration frequency (Hz). This is the amplitude (mm). These parameters were determined through optimization using oral medicine experiments and cleaning performance tests.

[0034] It is the target impedance parameter, which is the key bridge connecting macroscopic strategies and microscopic control. For desired quality, For desired damping, The desired stiffness collectively determines the "feel" of the brush head under this method. For example, the powerful cleaning mode corresponds to high stiffness and low damping, feeling "firm"; while the sensitive care mode corresponds to low stiffness and high damping, feeling "smooth". These parameters were initially calibrated through professional dental operation data analysis and impedance model simulation.

[0035] The recommended duration (s) is the appropriate cleaning time for each area based on clinical recommendations.

[0036] Step 2: Intelligent Dynamic Scheduling of the Method Pool The intelligent dynamic scheduling of the method pool is to dynamically select the optimal cleaning strategy based on the real-time oral environment. It can be divided into four parts: multi-sensor data acquisition and preprocessing, feature extraction and state vector construction, method pool matching and dynamic scheduling strategy, execution evaluation and online learning.

[0037] ① Multi-sensor data acquisition and preprocessing The system relies on a variety of low-cost embedded sensors to sense the brushing environment and user behavior, including posture sensors, pressure sensors, and optical reflection sensors, all integrated into the smart toothbrush.

[0038] Attitude sensing (IMU): A six-axis inertial measurement unit (IMU, combined with a three-axis accelerometer and a three-axis gyroscope) is used to track the three-dimensional spatial attitude of the brush head. The pitch angle of the brush head relative to its initial calibration position is calculated in real time using traditional sensor fusion algorithms (such as complementary filtering or Kalman filtering). and roll angle And estimate the linear acceleration of the brush head. and angular velocity .

[0039] Pressure sensing: A thin-film pressure sensor is integrated at the connection between the brush head and the handle or into the bristle base to measure the contact pressure between the brush head and the tooth surface. The raw voltage signal is acquired by an ADC (analog-to-digital converter) and then passed through a low-pass filter to eliminate high-frequency noise, thus smoothing the data.

[0040] Optical sensing: An optical reflection sensor (such as a photodiode combined with a specific wavelength LED) is deployed in the brush head. By measuring the intensity of light reflection from the tooth surface, the adhesion of dental plaque can be indirectly assessed, forming a plaque index. The estimated value. The stronger the reflected light, the cleaner the tooth surface usually is.

[0041] ②Feature extraction and state vector construction The raw sensor data is further processed to extract features with clear physical meaning that can characterize the current brushing state, and these features are then constructed into a state vector. :

[0042] Identifying the tooth region: This is the primary basis for scheduling decisions. The system uses the attitude angles calculated by the IMU. It maps the current brush head position to specific tooth areas using a pre-defined lookup table. For example: when and At that time, it was identified as the lower incisor area, i.e. .when and At that time, it was identified as the upper molar area. This mapping relationship can be obtained by calibration using a large amount of experimental data. The average contact pressure is calculated by averaging the filtered pressure values ​​of the thin-film pressure sensor over a short time window (e.g., 1 second) to reflect the continuous pressure level in the area and avoid interference from instantaneous fluctuations.

[0043] The plaque index is calculated from the reflectance data of an optical sensor, for example... ,in These are calibration coefficients.

[0044] The cleaning time for the current area, from the brush head to the current... The zone begins accumulating time, a parameter output by the processor in the smart toothbrush.

[0045] To support a user's historical preferences, historical data for that user is loaded from local or cloud storage and converted into a vector representing the strategy ID that the user most frequently uses in each dental region and that is evaluated as "effective" by the system. This provides a basis for personalized recommendations.

[0046] ③Method pool matching and dynamic scheduling strategy After feature extraction and state vector construction, method pool matching and dynamic scheduling strategy design are performed. The definition of the method pool is obtained through step one. The core task of the scheduler is to calculate the current state. With each method in the method pool Match score The matching score is calculated by weighting three dimensions: (1) Regional matching degree This is a decisive factor. If currently... That's exactly the method. If the target area is specified, this item will receive a perfect score (e.g., 100 points); otherwise, it will receive 0 points. This ensures that the strategy is applied to the correct dental location.

[0047]

[0048] (2) Cleaning requirements The urgency of cleaning the current area is assessed comprehensively. It is determined by both stress and plaque index.

[0049]

[0050] This is the greatest safety pressure. These are weighting coefficients, determined experimentally. High pressure, high plaque index, and insufficient cleaning time all increase the cleaning demand, thus leading to a preference for strategies with stronger cleaning power.

[0051] (3) User preferences The key to achieving personalization is querying user historical data. .

[0052]

[0053] Finally, the combined score for each method is:

[0054] in The scheduler selects the largest The scoring method is used as the currently executed strategy. It will be dynamically updated based on common methods used in performance evaluation and online learning.

[0055] ④ Implementation Assessment and Online Learning The parameters of the selected optimal method are immediately sent to the adaptive impedance controller and motor drive circuit for execution. Simultaneously, the system initiates an evaluation and optimization loop: first, performance evaluation; during strategy execution (e.g., lasting 2 seconds), the system continues to monitor sensor data and calculate an immediate reward (…). ).For example, =+1 if the pressure remains within a safe range and the plaque index decreases; =-1 if the pressure exceeds the limit. Secondly, online learning: this reward signal is used to fine-tune the weights of the scheduling strategy. Or, the parameters of the method itself. For example, a simple gradient descent rule can be used: if a policy receives a negative reward under pressure, then the system will reduce the cleaning requirement. In the future, under similar high-pressure environments, a more lenient strategy will be preferred. (User's historical preference data) It will also be updated based on the results of this brushing session. Step 3: Adaptive Coordinated Impedance Control and Microforce Adjustment The target torque is established using an impedance model. Calculation formula:

[0056] The target values ​​are directly provided by the current policy in the method pool. They are key to controlling the brush head's "tactile feel". For desired damping, For the desired stiffness, It is also provided by the current strategy in the method pool, for the expected quality.

[0057] This represents the effective lever arm from the motor output shaft (or equivalent drive point) to the point of contact between the brush head and the teeth. It is a fixed value set before the smart toothbrush leaves the factory. Its function is to adjust the linear force calculated by the impedance controller and applied to the brush head. This is converted into the rotational torque required to drive the motor. .

[0058] The desired position, velocity, and acceleration represent the ideal motion trajectory of the brush head, with the desired position typically defined by a simple waveform function, such as... ,in It is the amplitude. The frequency is provided by the method pool. Let be the desired velocity, and be the first derivative of the desired position. . It is the desired acceleration, and it is the second derivative of the desired position. The desired position, velocity, and acceleration are determined by the microcontroller (MCU) based on the current time. and those obtained from the method pool It is calculated in real time.

[0059] , The actual position and velocity of the brush head represent its true motion state within the oral cavity. This is calculated in real-time using data from a three-axis accelerometer and a three-axis gyroscope provided by an IMU (Inertial Measurement Unit). Specifically, a common Kalman filter is used to fuse the accelerometer and gyroscope data to estimate the actual position of the bristle tips. and speed .

[0060] The external force, namely the contact force between the brush head and the teeth, is obtained through a thin-film pressure sensor.

[0061] To achieve precise force tracking and overcome the uncertain dynamics introduced by the user's manual input, this invention designs a parameter adaptive law:

[0062] It is force tracking error. It is also determined by the damping and stiffness coefficients in the method pool. For the damping coefficient and stiffness coefficient, i.e. The ideal safety force set for the method pool. This represents the rate of change of damping and stiffness parameters. The controller integrates these values ​​in real time to update... . The adaptive gain determines the speed of parameter adjustment, which in turn determines the speed of impedance parameter adjustment. Its range is determined through offline simulation and stability analysis to ensure that the system response is both fast and stable.

[0063] It is a saturation function. In order to prevent the parameters from changing drastically under instantaneous disturbances, the signals of velocity difference and position difference are limited to a reasonable range, thereby enhancing the robustness of the system. The calculation formula is as follows:

[0064] Indicates the error parameter. This indicates the maximum acceptable error. This represents the minimum achievable error.

[0065] Working mechanism: When the user applies excessive force ( Force error It is positive. The adaptive law will automatically increase the damping. And reduce stiffness This makes the brush head more "soft" and "stickier" in terms of kinetics, as if it has a "pressure-relieving" buffering effect, thus automatically buffering excessive pressure and protecting the gums. Conversely, when the cleaning power is insufficient ( When this happens, the adaptive law will automatically reduce damping. And increase stiffness This enables continuous adaptive adjustment of microscopic force perception within a macroscopic strategy framework.

[0066] Step 4: Perform physical torque drive motor The microcontroller (MCU) will calculate the target torque To convert the torque into a specific motor control signal, the first step is signal conversion. The MCU will then convert the torque signal into a specific motor control signal. Convert to the corresponding target current (Because motor torque is directly proportional to current). Next, PWM modulation is performed. The motor drive circuit inside the MCU (such as an H-bridge circuit) uses PWM (Pulse Width Modulation) technology to precisely control the average current flowing into the motor by adjusting the voltage duty cycle, ensuring it reaches the required level. .

[0067] Example 2 This embodiment provides an intelligent toothbrush system based on a method pool and adaptive coordinated impedance control, consisting of a multimodal sensing module, a method pool management module, an adaptive impedance control module, and an execution module. Its core workflow is a closed-loop process.

[0068] Among them, the multimodal perception module is used to collect brush head posture, contact pressure and tooth surface cleanliness data (optical reflection data) in real time, and construct a state vector containing brush head position, average contact pressure, plaque index, area cleaning time and user historical preferences through feature extraction; The method pool management module has pre-stored a variety of cleaning strategies, including basic cleaning parameters, target impedance parameters and suggested durations, and can dynamically schedule and output the optimal cleaning strategy and its target impedance parameters according to the state vector. An adaptive impedance control module is used to receive target impedance parameters (desired mass, desired stiffness, and desired damping) provided by the optimal cleaning strategy, and, in conjunction with the real-time contact pressure obtained by the multimodal sensing module, calculate the target torque through an adaptive law based on Lyapunov stability, so as to adjust the dynamic interaction force between the brush head and the teeth in real time and track the target impedance parameters. And an execution module, used to control the motor to perform the brushing action according to the target torque.

[0069] The multimodal sensing module includes an attitude sensor, a pressure sensor, and an optical reflection sensor. The attitude sensor is used to acquire brush head attitude data and map the current brush head position to a specific tooth area using a preset lookup table. The pressure sensor is used to collect contact pressure data between the brush head and the teeth. The optical reflection sensor is used to detect tooth surface cleanliness and assess the degree of plaque adhesion by measuring the intensity of reflected light.

[0070] The method pool management module evaluates and selects the optimal cleaning strategy by calculating a weighted sum of regional matching degree, cleaning demand degree, and user preference degree, specifically including: The area adaptation score is full when the current brush head position is within the applicable area preset by the cleaning strategy, otherwise it is zero. The cleaning demand score is calculated by weighting the average contact pressure, plaque index, and cleaning time in the state vector. User preference score is calculated based on the user's historical preference data in the state vector; The matching score is a weighted sum of the regional adaptability score, the cleaning requirement score, and the user preference score.

[0071] The adaptive impedance control module includes an adaptive law for impedance parameters. This law dynamically adjusts the actual stiffness and damping parameters based on the deviation between the desired force and the actual contact force, the adaptive gain, and the brush head speed. The adaptive law is used to update the damping parameters and stiffness in real time, and its update rule is correlated with the force tracking error and the state error of the brush head movement to ensure system stability and achieve adaptive adjustment of microforces.

[0072] The method pool management module also includes an execution evaluation and online learning mechanism, which dynamically updates the weight coefficients of each cleaning strategy based on the cleaning effect data of this brushing session, thereby achieving continuous optimization of the method pool.

[0073] The cleaning strategies in the method pool are divided into three modes based on cleaning intensity: intensive cleaning mode, standard cleaning mode, and sensitive care mode. The powerful cleaning mode corresponds to high stiffness and low damping parameters, providing a firm tactile feel; The sensitive care mode corresponds to low stiffness and high damping parameters, providing a smooth touch.

[0074] The desired stiffness is described by an impedance model, which expresses the relationship between the brush head movement speed and the desired contact force as a second-order system including mass, damping, and stiffness terms.

[0075] As a specific implementation scheme, its hardware configuration is as follows: (1) Main control MCU: High-performance ARM Cortex-M4 core processor.

[0076] (2) Sensors: 6-axis IMU (for attitude calculation), pressure sensor film integrated into the brush head substrate (for measuring contact pressure).

[0077] (3) Actuator: High-performance linear resonant motor.

[0078] Example 3 This embodiment provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the above-described intelligent toothbrush control method based on method pool and adaptive coordinated impedance control.

[0079] Here is an example of a work scenario: When the user starts brushing their teeth, the IMU identifies that the brush head has entered the "lower jaw molar area" ( =4). The method pool scheduler selects the "High-efficiency cleaning of the molar area" method ( Its target parameter is high frequency ( =300Hz), high stiffness ( =1.5N / mm). The impedance controller starts working, and the brush head vibrates strongly. At this time, if the user accidentally applies too much force, the pressure sensor detects it. Exceeding the safety threshold The adaptive law acts instantaneously, reducing [the effect] within milliseconds. and increase The brush head immediately becomes "smooth," with reduced vibration and automatic pressure buffering. When the brush head moves to the "gingival line" under the app's guidance, the method pool automatically switches to the "gingival massage" method. Using low frequency and high damping parameters, the brush head becomes very gentle to the touch.

[0080] Through the above-described scheme, this invention generates macroscopic cleaning strategies via a "method pool" and transforms them into precise microscopic force interactions through "adaptive impedance control," achieving a seamless transition from "what to do" to "how to do it well," representing a fundamental innovation in control logic. Simultaneously, the compliant characteristics provided by impedance control ensure that the brush head does not simply stop working (damaging the user experience) when encountering excessive pressure, but intelligently "yields," providing a buffer and greatly improving safety and comfort—something traditional threshold protection mechanisms cannot match. Furthermore, the method pool can dynamically schedule strategies based on user, time, and region, while the impedance controller can adaptively adjust according to the instantaneous mechanical environment, allowing the toothbrush to continuously and dynamically adapt to the user's unique oral cavity characteristics and brushing habits. Additionally, this architecture primarily relies on algorithmic innovation to improve performance, and its sensor requirements (IMU, pressure sensor) are compatible with existing high-end toothbrush hardware platforms, eliminating the need to introduce expensive new sensors, making it highly promising for industrialization.

[0081] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0082] Furthermore, the method of the present invention is not limited to being executed in the chronological order described in the specification, but may also be executed in other chronological orders, in parallel, or independently. Therefore, the execution order of the method described in this specification does not constitute a limitation on the technical scope of the present invention.

[0083] Although the invention has been disclosed above through the description of specific embodiments, it should be understood that all the embodiments and examples described above are exemplary and not restrictive. Those skilled in the art can design various modifications, improvements, or equivalents to the invention within the spirit and scope of the appended claims. These modifications, improvements, or equivalents should also be considered to be included within the protection scope of the invention.

Claims

1. A smart toothbrush control method based on method pooling and adaptive coordinated impedance control, characterized in that, Includes the following steps: Step 1: Construct a method pool containing multiple cleaning strategies, each of which defines basic cleaning parameters, target impedance parameters, and suggested duration parameters; Step 2: Collect multimodal sensor data in real time during the brushing process, preprocess and extract features to construct a state vector. The multimodal sensor data includes brush head posture data, contact pressure data and optical reflection data; the state vector includes brush head position, average contact pressure, plaque index, area cleaning time and user history preferences. Step 3: Based on the state vector, dynamically match and schedule the optimal cleaning strategy from the method pool; Step 4: Based on the target impedance parameters corresponding to the optimal cleaning strategy, calculate the target torque using an adaptive law based on Lyapunov stability to adjust the dynamic interaction force between the brush head and teeth in real time. Step 5: Control the motor to perform the brushing action according to the calculated target torque.

2. The intelligent toothbrush control method according to claim 1, characterized in that: Step 3, which involves dynamically matching and scheduling the optimal cleaning strategy from the method pool, is a comprehensive evaluation of the sum of regional matching degree, cleaning demand degree, and user preference score. Specifically, it includes: Step 31: Calculate the area fit score to determine whether the current brush head position belongs to cleaning strategy m. i Preset applicable area; if the current brush head position is within the cleaning strategy m i If the preset applicable area is selected, the score is 1; otherwise, it is 0. Step 32: Calculate the cleaning demand score, which is a weighted sum of the ratio of the current average contact pressure to the maximum safe pressure, the plaque index, and the current cleaning time of the area. Step 33: Calculate user preference score; query user historical data to obtain cleaning strategies m in historical records. i The user preference score for cleaning strategies in history is calculated by the number of times a strategy was successfully used in that area, and the total number of times all cleaning strategies were successfully used in that area. i The ratio of the number of times a cleaning strategy is successfully used in the area to the total number of times all cleaning strategies are successfully used in the area; Step 34: For each cleaning strategy m i The corresponding regional suitability score, cleaning need score, and user preference score are each assigned a weight w. z w c w u The scores of each cleaning strategy are weighted and summed to obtain a comprehensive score; the cleaning strategy with the highest comprehensive score is selected as the optimal cleaning strategy.

3. The intelligent toothbrush control method according to claim 1, characterized in that: In step 4, the adaptive law dynamically adjusts the damping and stiffness parameters based on the force tracking error. When excessive contact force is detected, the adaptive law automatically increases the damping parameter and decreases the stiffness parameter to make the brush head have a smooth feel to buffer the pressure. Specifically, it includes: Obtain the desired quality based on the optimal cleaning strategy. Desired stiffness and expected damping ; Desired quality based on the target impedance parameter Desired stiffness Expected damping The target torque is calculated based on the desired motion trajectory generated from the basic cleaning parameters, and the actual motion trajectory and contact force from the collected multimodal sensor data. The calculation formula is as follows: ； in, These are the desired position, desired velocity, and desired acceleration, respectively. , It is the amplitude. The frequencies are all provided by the method pool; It is the first derivative of the desired position. ; It is the second derivative of the expected position. ; and These are the actual position and velocity of the brush head, obtained from the brush head attitude data of the multimodal sensor. The actual contact force is obtained from the contact pressure data in the multimodal sensor data.

4. The intelligent toothbrush control method according to claim 3, characterized in that: The adaptive law, based on force tracking error, adjusts the damping and stiffness parameters in real time. The formal definition of this law is as follows: ； in, These represent the rates of change of the damping parameter and the stiffness parameter, respectively; It is the force tracking error, F d Desired stiffness from the method pool Expected damping Decide, ; These are the damping coefficient and stiffness coefficient; These are adaptive gains, used to control the speed at which damping and stiffness parameters are adjusted, respectively. It is a saturation function. The calculation formula is as follows: ； Indicates the error parameter. This indicates the maximum permissible error. This represents the minimum error achieved.

5. The intelligent toothbrush control method according to claim 1, characterized in that: It also includes step 6: During the execution of the cleaning strategy, the cleaning effect is evaluated based on sensor data and a reward signal is generated to update the weight of the scheduling strategy or the strategy parameters in the optimization method pool online.

6. The intelligent toothbrush control method according to claim 1, characterized in that: The cleaning strategy parameters in the method pool are pre-calibrated through at least one of the following methods: oral medicine experiments, cleaning efficiency testing, professional dental operation data analysis, and impedance model simulation.

7. The intelligent toothbrush control method according to claim 1, characterized in that: The brush head position mapping is achieved through a preset lookup table, which maps the posture sensor data to specific tooth regions, including the incisor region, canine region, premolar region, and molar region.

8. The intelligent toothbrush control method according to claim 1, characterized in that: The plaque index is calculated based on optical reflection data from multimodal sensor data. The stronger the reflected light, the cleaner the tooth surface and the lower the plaque index. In other words, the higher the calculated plaque index value, the cleaner the tooth surface.

9. A smart toothbrush system based on method pooling and adaptive coordinated impedance control, characterized in that, include: The multimodal perception module is used to collect posture, contact force and optical reflection data during the brushing process, and extract brush head position, average contact pressure, plaque index, area cleaning time and user historical preference features to construct a state vector; The method pool management module has pre-stored a variety of cleaning strategies, including basic cleaning parameters, target impedance parameters and suggested durations, and can dynamically schedule and output the optimal cleaning strategy and its target impedance parameters according to the state vector. An adaptive impedance control module is used to receive the target impedance parameters corresponding to the optimal cleaning strategy, and calculate the target torque through an adaptive law based on Lyapunov stability to adjust the dynamic interaction force between the brush head and the teeth in real time. And an execution module, used to control the motor to perform the brushing action according to the target torque; The intelligent toothbrush system implements the steps of the intelligent toothbrush control method based on method pool and adaptive coordinated impedance control as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the intelligent toothbrush control method based on method pool and adaptive coordinated impedance control as described in any one of claims 1-8.

Images