Oral care equipment positioning method, device and system based on gyroscope

Abstract

The invention provides a gyroscope-based oral care equipment positioning method, device and system. The method comprises the following steps: acquiring real-time attitude data acquired by a first gyroscope mounted on oral care equipment; the real-time attitude data comprises a triaxial angular velocity, a triaxial acceleration, a pitch angle, a roll angle, a yaw angle and a timestamp; based on the real-time posture data, identifying a motion state of the oral care equipment; correcting a yaw angle in the real-time attitude data based on the motion state to obtain a corrected yaw angle; the position of the oral care device in the oral cavity is positioned based on the corrected yaw angle and other data in the real-time attitude data. The yaw angle can be accurately measured without a magnetometer or a gyroscope with the magnetometer, the cost is reduced, meanwhile, the positioning accuracy of the oral care equipment provided with the first gyroscope is guaranteed, and the user experience is improved.

Description

Technical Field

[0001] The present application relates to the technical field of oral care, and particularly to a positioning method, device and system for an oral care device based on a gyroscope. Background Art

[0002] Currently, due to factors such as changes in lifestyle, diet structure, and oral hygiene habits, oral diseases are a common health problem. Oral diseases cover many different types of diseases and involve various tissues and structures in the oral cavity. Some of the more common problems include: tooth decay (dental caries), periodontal disease, tooth loss, etc. At present, some devices that are helpful for oral care and oral lesion detection (such as electric toothbrushes, dental irrigators, and oral endoscopes) have gradually entered the public eye.

[0003] Currently, for oral positioning, a 6-axis or 9-axis gyroscope is mostly used and assembled on devices for oral care and oral lesion detection. The relative position between the real-time device attitude angle and the oral cavity is analyzed to fit the oral cavity position information. The following problems exist in the use of the two gyroscope solutions in the oral cavity:

[0004] 9-axis gyroscope: There is an error in the magnetometer and it needs to be calibrated irregularly.

[0005] 6-axis gyroscope: There is a problem of yaw angle drift, and orientation error is likely to occur during use.

[0006] The six-axis gyroscope mainly includes a three-axis accelerometer sensor and a three-axis gyroscope sensor. These two sensors together can be used to measure the rotation and acceleration of the device. On the basis of the six-axis gyroscope, the nine-axis gyroscope additionally adds a three-axis geomagnetic sensor (magnetometer), which enables the nine-axis gyroscope to measure the direction of the device relative to the earth's magnetic field, thereby providing more accurate attitude and direction perception. In addition, as a consumer-grade sensor, the six-axis sensor has the characteristics of low cost, average performance, limited accuracy, low durability, and suitability for light use. As an industrial-grade sensor, the nine-axis sensor has the characteristics of high cost, durability, high quality, high performance, higher accuracy, and a wider measurement range. Summary of the Invention

[0007] In order to solve the above technical problems, the present invention provides a positioning method, device and system for an oral care device based on a gyroscope, which can achieve accurate measurement of the yaw angle without a magnetometer or a gyroscope with a magnetometer. While reducing costs, it ensures the positioning accuracy of the oral care device equipped with the first gyroscope and improves the user experience.

[0008] The present invention adopts the following technical solutions:

[0009] On the one hand, a positioning method for an oral care device based on a gyroscope includes:

[0010] Obtain the real-time attitude data collected by the first gyroscope installed on the oral care device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp;

[0011] Based on the real-time attitude data, identify the motion state of the oral care device;

[0012] Based on the motion state, correct the yaw angle in the real-time attitude data to obtain a corrected yaw angle;

[0013] Based on the corrected yaw angle and other data in the real-time attitude data, locate the position of the oral care device in the oral cavity.

[0014] Preferably, based on the motion state, correct the yaw angle in the real-time attitude data to obtain a corrected yaw angle, specifically including:

[0015] Based on the motion state, use the trained LSTM model to correct the yaw angle in the real-time attitude data to obtain a corrected yaw angle.

[0016] Preferably, during training, the LSTM model uses a bidirectional LSTM structure, takes the first gyroscope data collected by the first gyroscope and the motion state as inputs, labels the yaw angle error, and outputs the yaw angle correction value; the yaw angle error is the error between the yaw angle of the second gyroscope data collected by the second gyroscope and the yaw angle of the first gyroscope data.

[0017] Preferably, the installation directions of the first gyroscope and the second gyroscope are the same. When obtaining training data, the first gyroscope and the second gyroscope collect data at the same frequency, and align the respective timestamps to align each frame of data.

[0018] Preferably, during real-time positioning, the LSTM model uses a unidirectional LSTM structure with the same optimization parameters as the bidirectional LSTM structure, takes the real-time attitude data collected by the first gyroscope as the input, and outputs the yaw angle correction value.

[0019] Preferably, the first gyroscope data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp; the second gyroscope data includes pitch angle, roll angle, yaw angle, and timestamp.

[0020] Preferably, based on the real-time attitude data, identify the motion state of the oral care device, specifically including:

[0021] Input the real-time attitude data into the trained hidden Markov model HMM to identify the motion state of the oral care device.

[0022] Preferably, during training, the hidden Markov model is modeled based on the first gyroscope data collected by the first gyroscope. The first gyroscope data is used as the observation space to input into the model, the motion states are labeled, the state transition probabilities and emission probabilities are initialized, and the model is trained to generate the motion state space probabilities.

[0023] Preferably, the first gyroscope is a six-axis gyroscope; the second gyroscope is a nine-axis gyroscope.

[0024] On the other hand, a positioning device for an oral care device based on a gyroscope includes:

[0025] A real-time attitude data acquisition module for acquiring the real-time attitude data collected by the first gyroscope installed on the oral care device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp;

[0026] A motion state recognition module for identifying the motion state of the oral care device based on the real-time attitude data;

[0027] A yaw angle correction module for correcting the yaw angle in the real-time attitude data based on the motion state to obtain a corrected yaw angle;

[0028] A position positioning module for positioning the position of the oral care device in the oral cavity based on the corrected yaw angle and other data in the real-time attitude data.

[0029] On yet another hand, a positioning system for an oral care device based on a gyroscope includes: an oral care device, a first gyroscope, a first gyroscope circuit board, and a terminal device; the first gyroscope and the first gyroscope circuit board are installed on the oral care device; the first gyroscope circuit board is respectively connected to the first gyroscope and the terminal device;

[0030] When the oral care device moves, the first gyroscope circuit board acquires the real-time attitude data collected by the first gyroscope and sends it to the terminal device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp;

[0031] The terminal device receives the real-time attitude data, identifies the motion state of the oral care device based on the real-time attitude data; corrects the yaw angle in the real-time attitude data based on the motion state to obtain a corrected yaw angle; and positions the position of the oral care device in the oral cavity based on the corrected yaw angle and other data in the real-time attitude data.

[0032] Preferably, the gyroscope-based oral care device positioning system further includes: a second gyroscope and a second gyroscope circuit board; during the positioning system training, the second gyroscope and the second gyroscope circuit board are installed on the oral care device; the second gyroscope circuit board is respectively connected to the second gyroscope and the terminal device; the installation directions of the first gyroscope and the second gyroscope are the same; when the oral care device moves, the first gyroscope collects first gyroscope data, and the second gyroscope collects second gyroscope data;

[0033] The terminal device obtains a yaw angle error based on the error between the yaw angle of the second gyroscope data and the yaw angle of the first gyroscope data, and trains a network model for obtaining a corrected yaw angle based on the first gyroscope data, the marked yaw angle error, and the marked motion state.

[0034] The present invention has the following beneficial effects:

[0035] (1) The present invention identifies the motion state of the oral care device based on the real-time attitude data collected by the first gyroscope; based on the motion state, corrects the yaw angle in the real-time attitude data to obtain a corrected yaw angle; based on the corrected yaw angle and other data in the real-time attitude data, locates the position of the oral care device in the oral cavity; it is possible to accurately measure the yaw angle without a magnetometer or a gyroscope with a magnetometer, while reducing costs, ensuring the positioning accuracy of the oral care device equipped with the first gyroscope, and improving the user experience;

[0036] (2) The present invention corrects the yaw angle in the real-time attitude data by using a trained LSTM model to obtain a corrected yaw angle. In particular, a bidirectional LSTM structure is used in the model training stage to better understand and capture the dependency relationships and context information in the sequence data; a unidirectional LSTM structure is used in the usage stage, which is convenient for deployment and improves efficiency and speed;

[0037] (3) The present invention accurately identifies the motion state of the oral care device based on the hidden Markov model HMM and provides it for use by the LSTM model, ensuring the correction accuracy of the LSTM model;

[0038] (4) The first gyroscope of the present invention is a six-axis gyroscope, and the second gyroscope is a nine-axis gyroscope, that is, the yaw angle of the six-axis gyroscope is corrected by the yaw angle of the nine-axis gyroscope, and the nine-axis gyroscope is only installed during model training. After the model training is completed, during the actual use process, only a six-axis gyroscope needs to be installed, and the corrected yaw angle can be obtained based on the real-time attitude data collected by the six-axis gyroscope, while reducing costs and ensuring the positioning accuracy.

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the following-described accompanying drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings. Description of the Drawings

[0040] Figure 1 Flowchart of the positioning method of the oral care device based on the gyroscope according to the embodiment of the present invention;

[0041] Figure 2 Schematic diagram of the model according to the embodiment of the present invention;

[0042] Figure 3 Effect display diagram of dividing the oral cavity into 8 fixed regions according to the embodiment of the present invention;

[0043] Figure 4 Effect display diagram of dividing the oral cavity into 16 fixed regions according to the embodiment of the present invention;

[0044] Figure 5 Effect display diagram of dividing the oral cavity into 28 fixed regions according to the embodiment of the present invention;

[0045] Figure 6 Block diagram of the structure of the positioning device of the oral care device based on the gyroscope according to the embodiment of the present invention;

[0046] Figure 7 Schematic diagram of the structure of the positioning system of the oral care device based on the gyroscope according to the embodiment of the present invention;

[0047] Figure 8 Schematic diagram of the gyroscope part of the positioning system of the oral care device based on the gyroscope according to the embodiment of the present invention. Detailed Embodiments

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention; obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0049] In the description of the present invention, it should be noted that the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including one..." does not exclude the presence of additional identical elements in the process, method, article or device including the said element.

[0050] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper", "lower", "inner", "outer", "top / bottom end", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0051] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "install", "be provided with", "sheath / connect", "connect", etc. should be understood in a broad sense. For example, "connect" can be a fixed connection, a detachable connection, or an integral connection. It can be a mechanical connection or an electrical connection. It can be directly connected or indirectly connected through an intermediate medium. It can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0052] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the step identifiers S101, S102, S103, etc. are only used for convenient expression and do not represent the execution order, and the corresponding execution order can be adjusted.

[0053] See Figure 1 As shown, a positioning method for an oral care device based on a gyroscope according to the present invention includes:

[0054] S101, obtaining real-time attitude data collected by a first gyroscope installed on the oral care device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle and timestamp;

[0055] S102, identifying the motion state of the oral care device based on the real-time attitude data;

[0056] S103, correcting the yaw angle in the real-time attitude data based on the motion state to obtain a corrected yaw angle;

[0057] S104. Based on the corrected yaw angle and other data in the real-time attitude data, locate the position of the oral care device in the oral cavity.

[0058] Specifically, the execution subject of a gyroscope-based oral care device positioning method is a terminal device, etc. The terminal device includes but is not limited to terminal devices such as mobile phones and tablets. This embodiment is not limited as long as it can execute the above method.

[0059] The first gyroscope is a six-axis gyroscope; the oral care device includes an oral endoscope, an electric toothbrush, a dental irrigator, etc.

[0060] In this embodiment, after the terminal device obtains the real-time attitude data collected by the first gyroscope, it inputs the real-time attitude data into the trained positioning model for processing, and finally outputs the position of the oral care device in the oral cavity, that is, the above S102 - S104 are implemented through the positioning model.

[0061] See Figure 2 As shown, specifically, the positioning model includes a state generation part, a recurrent neural network part, and a feature extraction part.

[0062] The state generation part inputs the real-time attitude data into the trained hidden Markov model HMM to identify the motion state of the oral care device.

[0063] The state generation part models the data generation process. It focuses on the relationship between the state sequence and the observation sequence, and generating observation data from the hidden state. Its role is to model the state changes in the sequence, especially in the case of state transitions. The state generation part of this embodiment can be used to judge the motion state of the gyroscope, and the output sequence of the state generation part is used as the input of the recurrent neural network to help the recurrent neural network part understand its motion behavior and improve the estimation accuracy of the yaw angle.

[0064] During training, the hidden Markov model is modeled based on the first gyroscope data collected by the first gyroscope. The first gyroscope data is used as the observation space to input into the model, the motion state is labeled, the motion state space is defined, the state transition probability and the emission probability are initialized, and methods such as the EM algorithm are used for model training to generate the motion state space probability. The first gyroscope data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp.

[0065] The state space defines different motion states, including stationary, horizontal brushing, vertical brushing, inward flipping, outward flipping, left turning, right turning, etc.; the observation space defines the observation space, and the observation data includes the acceleration, angular velocity, Euler angles of the six-axis gyroscope and the motion state of the device; the state transition probability defines the state transition probability, that is, the probability of transitioning from one state to another, and these probabilities can be estimated from the training data; the observation probability distribution defines the observation probability distribution, that is, the probability distribution of observing different observation data under a given state, which can be modeled using a probability distribution function (such as a Gaussian distribution); the initial state probability defines the initial state probability distribution of the model, that is, the probability of being in different states at the beginning; the emission probability can be used to describe the probability of observing a specific action under a given hidden state. For example:

[0066] P(observing "horizontal brushing" | hidden state is "left-right jittering") = 0.8

[0067] P(observing "vertical brushing" | hidden state is "left-right jittering") = 0.1

[0068] P(observing "stationary" | hidden state is "left-right jittering") = 0.1

[0069] P(observing "horizontal brushing" | hidden state is "not moving") = 0.1

[0070] P(observing "vertical brushing" | hidden state is "not moving") = 0.1

[0071] P(observing "stationary" | hidden state is "not moving") = 0.8

[0072] P(observing "horizontal brushing" | hidden state is "up-down jittering") = 0.1

[0073] P(observing "vertical brushing" | hidden state is "up-down jittering") = 0.8

[0074] P(observing "stationary" | hidden state is "up-down jittering") = 0.1

[0075] The specific training process of the Hidden Markov Model HMM is as follows.

[0076] (1) Data collection: Collect a set of labeled training data, which includes the six-axis data of the gyroscope and the corresponding motion state labels.

[0077] (2) Model parameter initialization: Initialize the model parameters of the HMM, including the state transition probability, the observation probability distribution, and the initial state probability distribution.

[0078] (3) Expectation Maximization (EM) algorithm: The Expectation Maximization algorithm (EM algorithm) is used to estimate the parameters of the model. In the E-step of the EM algorithm, the posterior probability of the hidden state at each time step is calculated given the observed data. In the M-step, these posterior probabilities are used to update the model parameters.

[0079] (4) Repeated iteration: The E-step and M-step are repeatedly executed until the model parameters converge or a predetermined number of iterations is reached.

[0080] The recurrent neural network part includes a discriminative model, which is mainly used for modeling and predicting sequence data. The time-dependent relationships and patterns in the sequence data are emphasized. The present invention is based on LSTM for recurrent neural network modeling, and in the oral positioning problem, it can help predict and correct angle changes, thereby improving the accuracy of oral positioning. Specifically, based on the motion state, the trained LSTM model is used to correct the yaw angle in the real-time pose data to obtain the corrected yaw angle. Different motion states have different effects on the yaw angle drift, so training needs to be carried out based on different motion states.

[0081] During training, the LSTM model uses a bidirectional LSTM structure, takes the first gyroscope data collected by the first gyroscope and the motion state as inputs, labels the yaw angle error, and outputs the yaw angle correction value; the yaw angle error is the error between the yaw angle of the second gyroscope data collected by the second gyroscope and the yaw angle of the first gyroscope data.

[0082] The installation directions of the first gyroscope and the second gyroscope are the same. When obtaining training data, the first gyroscope and the second gyroscope collect data at the same frequency, and their respective timestamps are corresponding to align each frame of data.

[0083] During real-time positioning, the LSTM model uses a unidirectional LSTM structure with the same optimization parameters as the bidirectional LSTM structure. The input is the real-time pose data collected by the first gyroscope, and the output is the yaw angle correction value.

[0084] In this embodiment, the second gyroscope is a nine-axis gyroscope, and the second gyroscope data includes pitch angle, roll angle, yaw angle, and timestamp.

[0085] Specifically, when training, the bidirectional LSTM receives the input sequence and performs calculations in two directions (forward and backward) at each time step. In the forward calculation process, it processes the input data in the order of time steps, from front to back. In the backward calculation process, it processes the input data in the reverse order of time steps, from back to front. The bidirectional LSTM combines the results of the forward and backward calculations, so that each time step contains the context information before and after the current moment. This helps the model better understand and capture the dependency relationships and context information in the sequence data.

[0086] During use, a unidirectional LSTM is used to process the input sequence, and there is no longer a need to consider forward and backward information. One of the reasons is that inference usually places more emphasis on efficiency and speed, while bidirectional LSTMs are computationally more expensive. In addition, the model size of the unidirectional LSTM is smaller and it is easier to deploy on different platforms. During the inference phase, the unidirectional LSTM can be regarded as a simpler and lighter model.

[0087] Using a bidirectional LSTM during the training phase and a unidirectional LSTM with the same optimization parameters during use has the following effects.

[0088] (1) Better sequence modeling: The bidirectional LSTM can consider both past and future information of the input sequence, thus better capturing the context relationships in the sequence data. This is helpful for tasks such as sentiment analysis, named entity recognition, and language generation in natural language processing (NLP) to improve model performance.

[0089] (2) More accurate feature extraction: During training, the bidirectional LSTM can learn richer feature representations because it can see both directions of the input sequence. This can provide more accurate features for tasks such as time series prediction, text classification, and speech recognition.

[0090] (3) Better solving of context-dependency problems. The bidirectional LSTM helps to better handle these dependencies.

[0091] (4) Bidirectional LSTMs can reduce the vanishing / exploding gradient problem during model training because they calculate gradients through forward and backward propagation, which can improve the stability of the model.

[0092] (5) For generative tasks, bidirectional LSTMs can provide richer context information, which helps to generate more coherent and logical outputs.

[0093] (6) The unidirectional LSTM is more computationally efficient during the inference phase because it only needs to consider forward information. This is very useful for resource-constrained situations such as real-time applications and embedded systems.

[0094] (7) The unidirectional LSTM model usually has fewer parameters and is therefore lighter. This is an advantage in some resource-limited environments.

[0095] (8) Due to the smaller size of the unidirectional LSTM model, it is easier to deploy to different platforms and devices.

[0096] The said feature extraction part is based on the data of the 3-axis acceleration of a six-axis sensor, and the processing process is as follows.

[0097] (1) Calculate the average value [3]: The average acceleration value (for each axis);

[0098] (2) Calculate the standard deviation [3]: Standard deviation (for each axis) Mean absolute difference [3]: Among 200 readings, the mean absolute difference ED between each absolute value and the mean value of these 200 values (for each axis);

[0099] (3) Calculate the average resultant acceleration [1]: The square root of the sum of the squares of the values on each axis of ED;

[0100] (4) Data grouped distribution

[30] : Determine the value range (maximum - minimum) for each axis, divide this range into 10 equally sized bins, and then record the fraction of the 200 values that fall into each bin. First, obtain this acceleration data, calculate the eigenvalue related to acceleration, and then use the acceleration eigenvalue as the input of the LSTM model to participate in the LSTM model training and the use of the LSTM model.

[0101] Function: Capture more information. Increasing the dimension can capture the information in the original data more fully. In some cases, the original data in three dimensions may not be able to fully describe the complexity of the system, and by introducing more dimensions, the characteristics of the data can be better represented.

[0102] Improve the model's expressive ability: Increasing the feature dimension can improve the expressive ability of machine learning models. Some models are more likely to learn complex patterns in a high - dimensional space, thus improving the fitting ability to the data.

[0103] Reduce redundancy: While increasing the dimension, through reasonable feature selection and processing methods, redundant information can be reduced. This helps to remove the noise and irrelevant information in the original data and improve the robustness of the model.

[0104] Better generalization ability: Increasing the dimension helps to improve the generalization ability of the model, making it more adaptable to different data distributions. This is important for the model to perform well on unseen data.

[0105] Adapt to different models: Some machine learning models are more adaptable to high - dimensional data. For example, deep learning models (neural networks) usually perform better in a high - dimensional space.

[0106] Extract complex features: Through specific data processing methods, more complex and abstract features can be extracted from the original data, which helps to better describe the internal structure of the data.

[0107] Solve non - linear relationships: The high - dimensional feature space helps to better capture complex non - linear relationships. This may be beneficial for dealing with some complex problems, such as pattern recognition and classification tasks.

[0108] Further, in order to achieve the positioning of the oral care device in the oral cavity, before using the oral care device, it is necessary to calibrate it at a fixed position in the oral cavity to establish a coordinate system between the oral care device and the oral cavity. During the use of the oral care device, the oral position remains unchanged. By changing the Euler angles (pitch angle, roll angle, yaw angle), the relative position between the device and the oral cavity is determined. Specifically, according to the needs, when dividing different regions in the oral cavity, the position of the oral care device is output. Refer to Figure 3 as shown, it is an effect display diagram of dividing the oral cavity into 8 fixed regions; refer to Figure 4 as shown, it is an effect display diagram of dividing the oral cavity into 16 fixed regions; refer to Figure 5 as shown, it is an effect display diagram of dividing the oral cavity into 28 fixed regions. Each region has corresponding left and right sides, upper and lower sides, and inner and outer occlusal surfaces of the oral cavity.

[0109] Refer to Figure 6 as shown, a positioning device for an oral care device based on a gyroscope includes:

[0110] A real-time attitude data acquisition module 601, configured to acquire real-time attitude data collected by a first gyroscope installed on the oral care device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp;

[0111] A motion state recognition module 602, configured to recognize the motion state of the oral care device based on the real-time attitude data;

[0112] A yaw angle correction module 603, configured to correct the yaw angle in the real-time attitude data based on the motion state to obtain a corrected yaw angle;

[0113] A position positioning module 604, configured to position the oral care device in the oral cavity based on the corrected yaw angle and other data in the real-time attitude data.

[0114] The specific implementation of a positioning device for an oral care device based on a gyroscope is the same as a positioning method for an oral care device based on a gyroscope, and will not be repeated in this embodiment.

[0115] Refer to Figure 7 as shown, a positioning system for an oral care device based on a gyroscope includes: an oral care device 10, a first gyroscope 20, a first gyroscope circuit board 30, and a terminal device 40; the first gyroscope 20 and the first gyroscope circuit board 30 are installed on the oral care device 10; the first gyroscope circuit board 30 is respectively connected to the first gyroscope 20 and the terminal device 40;

[0116] When the oral care device 10 moves, the first gyroscope circuit board 30 obtains the real-time attitude data collected by the first gyroscope 20 and sends it to the terminal device 40; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle and timestamp;

[0117] The terminal device 40 receives the real-time attitude data, identifies the motion state of the oral care device 10 based on the real-time attitude data; based on the motion state, corrects the yaw angle in the real-time attitude data to obtain a corrected yaw angle; based on the corrected yaw angle and other data in the real-time attitude data, locates the position of the oral care device 10 in the oral cavity.

[0118] See Figure 8 As shown, further, the positioning system of the oral care device based on the gyroscope further includes: a second gyroscope 50 and a second gyroscope circuit board 60; during the positioning system training, the second gyroscope 50 and the second gyroscope circuit board 60 are installed on the oral care device 10; the second gyroscope circuit board 60 is respectively connected to the second gyroscope 50 and the terminal device 40; the installation directions of the first gyroscope 20 and the second gyroscope 50 are the same, specifically referring to the same directions of the X-axis, Y-axis and Z-axis of the first gyroscope 20 and the second gyroscope 50; when the oral care device 10 moves, the first gyroscope 20 collects the data of the first gyroscope 20, and the second gyroscope 50 collects the data of the second gyroscope 50;

[0119] The terminal device 40 obtains a yaw angle error based on the error between the yaw angle of the second gyroscope 50 data and the yaw angle of the first gyroscope 20 data, and trains a network model for obtaining the corrected yaw angle based on the first gyroscope 20 data, the marked yaw angle error and the marked motion state.

[0120] In this embodiment, the first gyroscope 20 is a six-axis gyroscope, and the second gyroscope 50 is a nine-axis gyroscope. The oral care device 10 is a dental irrigator, an oral endoscope, an electric toothbrush, a toothbrush peripheral, etc. The oral care device 10 is provided with a sensor 101, and the sensor 101 is a toothbrush head of an electric toothbrush, a water jet nozzle of a dental irrigator or a camera of an oral endoscope, etc. The oral care device 10 is a handheld device, including a handheld part 102 and including a housing 103. The first gyroscope 20 and the first gyroscope circuit board 30 are installed inside the housing 103. The first gyroscope circuit board 30 uploads data to the terminal device 40 through a communication module (such as a wireless communication module). The model in the terminal device 40 calculates the positioning information of the oral care device 10 in the oral cavity and displays it in the terminal device 40.

[0121] See Figure 7 and Figure 8As shown, the installation positions of the first gyroscope 20 and the second gyroscope 50 are marked in the same direction as the indication 80, where the Z-axis direction is the same as that of the sensor 101. The horizontal positions of the first gyroscope 20 and the second gyroscope 50 are not restricted, and only the X, Y, and Z angles in the figure need to be kept consistent. The Euler angles are defined as follows: rotation around the X-axis is the pitch angle; rotation around the Y-axis is the roll angle; rotation around the Z-axis is the yaw angle.

[0122] Specifically, when training the model in the terminal device 40, the oral care device positioning system based on the gyroscope needs to use the second gyroscope 50 and the second gyroscope circuit board 60. Specifically, the usage method and training process of the positioning system during the training process are as follows.

[0123] See Figure 8 As shown, the first gyroscope circuit board 30 and the second gyroscope circuit board 60 are relatively fixed through the positioning post 70 (the arrow therein indicates the direction 90 of the handheld part). Preferably, the front projection of the first gyroscope circuit board 30 overlaps with the second gyroscope circuit board 60. During the data acquisition process, by operating the handheld part 102, the motion postures of the two circuit boards are synchronously changed, and the physical change amounts of the Euler angles (pitch angle, roll angle, yaw angle) are the same. Collect the original data of the gyroscope, including 3 Euler angles, 3 accelerations, 3 angular velocities, and time stamps of the six-axis gyroscope; 3 Euler angles and time stamps of the nine-axis gyroscope; and evaluate and mark the motion state of the oral care device 10 in real time. Different motion states are set according to different application requirements. Combining with the international standard brushing posture, the states are divided into: stationary, horizontal brushing, vertical brushing, inner flip, outer flip, left turn, right turn, etc., 20 states.

[0124] The tester conducts multiple data acquisitions according to the application scenarios of the embodiments. Each data acquisition process is used as a data sample. The six-axis gyroscope and the nine-axis gyroscope need to be set to collect data at the same frequency, and their respective time stamps are corresponding to align each frame of data. The initial yaw angle of the two gyroscopes is set to 0°, and the yaw angle range is between -180° and 180°. Different data samples are divided into a training set and a test set.

[0125] Using the training set, the motion state of the oral care device 10 and the differential compensation angle between the six-axis gyroscope and the nine-axis gyroscope (the yaw angle of the six-axis gyroscope is not accurate enough, and the yaw angle of the nine-axis gyroscope is regarded as accurate, and the difference of each frame of yaw angle is used to obtain the compensation angle) are used as the labels of the state generation part and the recurrent neural network part for model training and verification. The algorithm is continuously optimized and iterated to obtain the optimized parameters, and the positioning model with the optimized parameters is maintained for subsequent use.

[0126] After the model training is completed, in subsequent applications, only through Figure 7Data collection and positioning are performed on the gyroscope-based oral care device positioning system shown as follows.

[0127] (1) Real-time data acquisition: Input the data obtained by the six-axis gyroscope into the positioning model. During the use of the product, the number of frames of gyroscope data acquisition is consistent with that in the training process, and the operation direction of the gyroscope is consistent with the data acquisition direction in the model training process.

[0128] (2) Upload the real-time data to the terminal device 40 for processing. The terminal device 40 corrects the yaw angle error in the recurrent neural network part through the data features output by the state generation part and the feature extraction part in the generation model for the six-axis gyroscope data;

[0129] (3) The terminal device 40 determines the position of the oral care device 10 in the oral cavity according to the positioning and displays it.

[0130] As described above, it is only a preferred specific embodiment of the present invention; however, the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and its improvement concept of the present invention, makes equivalent substitutions or changes, and all should be covered by the protection scope of the present invention.

Claims

1. A positioning method for an oral care device based on a gyroscope, characterized in that, comprising: Obtaining real-time attitude data collected by a first gyroscope installed on the oral care device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp; Based on the real-time attitude data, identifying the motion state of the oral care device; Based on the motion state, correcting the yaw angle in the real-time attitude data to obtain a correction; Based on the corrected yaw angle and other data in the real-time attitude data, positioning the position of the oral care device in the oral cavity.

2. The positioning method for an oral care device based on a gyroscope according to claim 1, characterized in that, Based on the motion state, correcting the yaw angle in the real-time attitude data to obtain a corrected yaw angle, specifically including: Based on the motion state, using a trained LSTM model to correct the yaw angle in the real-time attitude data to obtain a corrected yaw angle.

3. The positioning method for an oral care device based on a gyroscope according to claim 2, characterized in that, During training, the LSTM model uses a bidirectional LSTM structure, takes the first gyroscope data collected by the first gyroscope and the motion state as inputs, labels the yaw angle error, and outputs the yaw angle correction value; the yaw angle error is the error between the yaw angle of the second gyroscope data collected by the second gyroscope and the yaw angle of the first gyroscope data.

4. The positioning method for an oral care device based on a gyroscope according to claim 3, characterized in that, The installation directions of the first gyroscope and the second gyroscope are the same. When obtaining training data, the first gyroscope and the second gyroscope collect data at the same frequency, and correspond their respective timestamps to align each frame of data.

5. The positioning method for an oral care device based on a gyroscope according to claim 3, characterized in that, During real-time positioning, the LSTM model uses a unidirectional LSTM structure with the same optimization parameters as the bidirectional LSTM structure, takes the real-time attitude data collected by the first gyroscope as the input, and outputs the yaw angle correction value.

6. The positioning method for an oral care device based on a gyroscope according to claim 3, characterized in that, The first gyroscope data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle, and timestamp; the second gyroscope data includes pitch angle, roll angle, yaw angle, and timestamp.

7. The positioning method for an oral care device based on a gyroscope according to claim 1, characterized in that, Based on the real-time attitude data, identifying the motion state of the oral care device, specifically including: Inputting the real-time attitude data into a trained hidden Markov model HMM to identify the motion state of the oral care device.

8. The positioning method for an oral care device based on a gyroscope according to claim 7, characterized in that, During training, the hidden Markov model is modeled based on the first gyroscope data collected by the first gyroscope, takes the first gyroscope data as the observation space and inputs it into the model, labels the motion state, initializes the state transition probability and the emission probability, conducts model training, and generates the motion state space probability.

9. The method for positioning an oral care device based on a gyroscope according to claim 3, characterized in that, the first gyroscope is a six-axis gyroscope; the second gyroscope is a nine-axis gyroscope.

10. An apparatus for positioning an oral care device based on a gyroscope, characterized in that, comprising: a real-time attitude data acquisition module, configured to acquire real-time attitude data collected by a first gyroscope installed on the oral care device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle and timestamp; a motion state recognition module, configured to recognize the motion state of the oral care device based on the real-time attitude data; a yaw angle correction module, configured to correct the yaw angle in the real-time attitude data based on the motion state to obtain a corrected yaw angle; a position positioning module, configured to position the oral care device in the oral cavity based on the corrected yaw angle and other data in the real-time attitude data.

11. A system for positioning an oral care device based on a gyroscope, characterized in that, comprising: an oral care device, a first gyroscope, a first gyroscope circuit board and a terminal device; the first gyroscope and the first gyroscope circuit board are installed on the oral care device; the first gyroscope circuit board is respectively connected to the first gyroscope and the terminal device; when the oral care device moves, the first gyroscope circuit board acquires the real-time attitude data collected by the first gyroscope and sends it to the terminal device; the real-time attitude data includes three-axis angular velocity, three-axis acceleration, pitch angle, roll angle, yaw angle and timestamp; the terminal device receives the real-time attitude data, recognizes the motion state of the oral care device based on the real-time attitude data; corrects the yaw angle in the real-time attitude data based on the motion state to obtain a corrected yaw angle; positions the oral care device in the oral cavity based on the corrected yaw angle and other data in the real-time attitude data.

12. The system for positioning an oral care device based on a gyroscope according to claim 11, characterized in that, further comprising: a second gyroscope and a second gyroscope circuit board; during the training of the positioning system, the second gyroscope and the second gyroscope circuit board are installed on the oral care device; the second gyroscope circuit board is respectively connected to the second gyroscope and the terminal device; the installation directions of the first gyroscope and the second gyroscope are the same; when the oral care device moves, the first gyroscope collects first gyroscope data, and the second gyroscope collects second gyroscope data; the terminal device obtains a yaw angle error based on the error between the yaw angle of the second gyroscope data and the yaw angle of the first gyroscope data, and trains a network model for obtaining a corrected yaw angle based on the first gyroscope data, the marked yaw angle error and the marked motion state.

Images