Binaural transducer adjustment method and system for in-ear bone conduction hearing aid

Abstract

The application relates to the technical field of hearing aid devices and discloses an in-ear bone-air conduction auxiliary hearing bilateral vibrator adjusting method and system and an earphone. The method comprises the following steps: collecting structural characteristic data such as a user's bilateral ear canal cartilage contact pressure distribution and three-dimensional geometric parameters through a micro sensor, constructing a personalized acoustic transmission model containing bone conduction and air conduction sound wave transmission path characteristics on the basis of the data, and generating initial adjusting parameters of bone conduction vibrator vibration frequency, amplitude and air conduction loudspeaker frequency response compensation values according to the data. Subsequently, dynamic change feedback signals of bilateral bone conduction and air conduction perception intensity of the user are monitored in real time, the initial parameters are dynamically adjusted according to the signals, optimized bilateral vibrator control parameters are generated, and the collaborative matching of bilateral bone-air conduction sound output is realized. The method can accurately adapt to the individual ear canal characteristics of the user, dynamically respond to hearing state changes, and improve the sound transmission adaptability and hearing experience of the hearing aid device.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hearing aid devices, in particular to an in-ear bone-air conduction auxiliary hearing bilateral vibrator adjustment method, system and earphone. BACKGROUND

[0002] With the continuous development of hearing aid technology, in-ear bone-air conduction auxiliary hearing devices gradually become an important auxiliary tool for hearing impaired people because they take into account the advantages of bone conduction and air conduction sound transmission. However, there are still many problems to be solved in the bilateral vibrator adjustment process of existing devices.

[0003] The human ear canal structure has significant individual differences. The ear canal geometry, cartilage distribution and contact pressure characteristics of different users are different. Traditional hearing aids mostly use general acoustic transmission models, without fully considering such individual differences, resulting in that the transmission paths of bone conduction and air conduction sound waves cannot be accurately matched with the user's ear canal structure, and problems such as sound distortion and low transmission efficiency often occur. At the same time, the existing adjustment method is mostly based on fixed parameter settings, and lacks dynamic response ability to the real-time hearing state of the user. The hearing needs of the user in different environments will change, and the perception intensity of the bilateral ears may also fluctuate dynamically. Static adjustment parameters are difficult to adapt to such changes, which easily causes the imbalance of bilateral hearing experience and affects the hearing aid effect.

[0004] In the prior art, the adjustment of bone conduction vibrator and air conduction speaker is often independent, and no effective coordination mechanism is formed. The transmission characteristics of bone conduction and air conduction sound waves are different, and independent adjustment easily leads to mismatching of the output sound in phase and intensity, which further affects the overall perception of the user to the sound, and even may cause auditory fatigue. At the bilateral adjustment level, the traditional method mostly uses symmetrical parameter settings, ignoring the possible asymmetry of the user's bilateral ear canal structure, resulting in unilateral adaptation deviation, further reducing the use experience of the hearing aid device. The existence of these problems limits the practical application effect of the in-ear bone-air conduction auxiliary hearing device, and it is difficult to fully meet the user's demand for accurate and comfortable hearing aid. SUMMARY

[0005] The purpose of the present application is to provide an in-ear bone-air conduction auxiliary hearing bilateral vibrator adjustment method, system and earphone to solve the problems raised in the background art.

[0006] To achieve the above purpose, the present application provides an in-ear bone-air conduction auxiliary hearing bilateral vibrator adjustment method, which comprises:

[0007] Obtaining the structure characteristic data of the user's bilateral ear canal, collecting the contact pressure distribution of the ear canal cartilage and the three-dimensional geometric parameters of the ear canal shape through a micro sensor;

[0008] constructing a personalized acoustic transmission model according to the structure characteristic data, the personalized acoustic transmission model comprising bone conduction sound wave transmission path characteristics and air conduction sound wave transmission path characteristics;

[0009] generating initial adjustment parameters of the bilateral oscillators based on the personalized acoustic transmission model, the initial adjustment parameters comprising a vibration frequency and a vibration amplitude of the bone conduction oscillator and a frequency response compensation value of the air conduction loudspeaker;

[0010] monitoring a bilateral auditory feedback signal of the user in real time, the auditory feedback signal comprising dynamic changes of bone conduction perception intensity and air conduction perception intensity;

[0011] dynamically adjusting the initial adjustment parameters according to the auditory feedback signal to generate optimized bilateral oscillator control parameters, so as to realize cooperative matching of bilateral bone and air conduction sound outputs.

[0012] Preferably, the structure characteristic data of the bilateral ear canals of the user comprises:

[0013] measuring a pressure distribution of the cartilage region of the ear canal by means of a contact pressure sensor array, and extracting a peak pressure position and a pressure decay gradient in the pressure distribution;

[0014] collecting axial curvature radius and radial cross-sectional area change data of the ear canal by means of a miniature three-dimensional scanning module;

[0015] integrating the peak pressure position, the pressure decay gradient, the axial curvature radius and the radial cross-sectional area change data into an ear canal structure characteristic data set.

[0016] Preferably, the constructing a personalized acoustic transmission model according to the structure characteristic data comprises:

[0017] calculating a mechanical impedance matching coefficient of the bone conduction oscillator based on the ear canal structure characteristic data set, the mechanical impedance matching coefficient reflecting vibration coupling efficiency of the oscillator and the ear canal cartilage;

[0018] extracting acoustic resonance frequency and attenuation characteristics of the air conduction path of the ear canal, and generating an air conduction frequency response compensation curve in combination with the axial curvature radius and the radial cross-sectional area change data;

[0019] correlating and mapping the mechanical impedance matching coefficient and the air conduction frequency response compensation curve to form the personalized acoustic transmission model.

[0020] Preferably, the generating initial adjustment parameters of the bilateral oscillators based on the personalized acoustic transmission model comprises:

[0021] setting a reference vibration frequency and a vibration amplitude of the bone conduction oscillator according to the mechanical impedance matching coefficient, so that vibration energy is concentrated at the peak pressure position of the ear canal cartilage;

[0022] adjusting high-frequency gain of the air-conduction speaker according to the air-conduction frequency response compensation curve to compensate for phase shift of sound waves caused by axial curvature of the ear canal;

[0023] combining the reference vibration frequency, the vibration amplitude, and the high-frequency gain into the initial adjustment parameter.

[0024] Preferably, the real-time monitoring of the user's bilateral auditory feedback signals comprises:

[0025] acquiring a vibration waveform of the vibrator feedback through a bone conduction microphone, and extracting a harmonic distortion component and an energy decay rate in the waveform;

[0026] monitoring a residual sound pressure level in the ear canal through an air conduction microphone, and calculating an interference strength of sound waves in the air conduction path;

[0027] quantifying the harmonic distortion component, the energy decay rate, and the interference strength of sound waves into the auditory feedback signal.

[0028] Preferably, the dynamic adjustment of the initial adjustment parameter according to the auditory feedback signal comprises:

[0029] comparing the difference in the harmonic distortion component between the two sides, and correcting the vibration frequency of the bone conduction vibrator to balance the harmonic energy on both sides;

[0030] adjusting the phase delay of the air-conduction speaker according to the interference strength of sound waves to suppress the sound wave cancellation effect in the ear canal;

[0031] optimizing the dynamic range of the vibration amplitude based on the energy decay rate to generate the optimized bilateral vibrator control parameter.

[0032] Preferably, the method further comprises:

[0033] establishing a parameter coordination constraint rule for the bilateral vibrators, which limits the frequency band overlap interval of the bone conduction vibrator and the air-conduction speaker;

[0034] when the optimized bilateral vibrator control parameter triggers the frequency band overlap interval, reallocating the frequency band weight of bone conduction and air conduction.

[0035] Preferably, the method further comprises:

[0036] updating the personalized acoustic transmission model according to the user's head rotation angle data, which is collected by an inertial measurement unit;

[0037] resetting the optimized bilateral vibrator control parameter based on the updated personalized acoustic transmission model to adapt to changes in the sound source position.

[0038] Preferably, the application further comprises an in-ear bone-air conduction hearing aid bilateral vibrator adjustment system for implementing the in-ear bone-air conduction hearing aid bilateral vibrator adjustment method as described above, the system comprising:

[0039] an ear canal feature acquisition module for acquiring structural feature data of the user's bilateral ear canals;

[0040] an acoustic modeling module for constructing a personalized acoustic transmission model based on the structural feature data;

[0041] a parameter generation module for generating initial adjustment parameters of the bilateral vibrator based on the personalized acoustic transmission model;

[0042] a feedback monitoring module for monitoring the user's bilateral auditory feedback signals in real time;

[0043] a dynamic adjustment module for dynamically adjusting the initial adjustment parameters based on the auditory feedback signals to generate optimized bilateral vibrator control parameters.

[0044] Preferably, the application further comprises an in-ear bone-air conduction hearing aid, the earphone comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the in-ear bone-air conduction hearing aid bilateral vibrator adjustment method as described above.

[0045] Compared with the prior art, the application has the following beneficial effects:

[0046] By collecting personalized structural feature data of the user's bilateral ear canals and constructing a dedicated acoustic transmission model, the limitations of traditional general models are broken, and the simulation of the acoustic transmission path is more in line with the actual situation of the user's individual ear canal. Since the model fully incorporates the specific path characteristics of bone conduction and air conduction sound transmission, the initial adjustment parameters generated based on this can accurately match the ear canal structure characteristics of the user, making the initial output state of the bone conduction vibrator and the air conduction speaker more in line with the individual auditory transmission requirements, reducing distortion and adaptation deviation in the sound transmission process from the source.

[0047] The real-time auditory feedback monitoring mechanism introduced by the method can dynamically capture the changes in the user's bilateral ear's perception of bone conduction and air conduction sound, so that the adjustment process is no longer limited to fixed parameters, but can be flexibly adjusted according to the user's real-time auditory state. This dynamic adjustment mode can respond to changes in the user's auditory needs in different environments in a timely manner, and can keep the auditory perception of the bilateral ears balanced through real-time optimization of parameters in both noisy environments and quiet scenes, avoiding experience degradation caused by environmental changes or individual auditory fluctuations.

[0048] The method realizes the cooperative linkage of the bone conduction vibrator and the air conduction speaker adjustment, and through unified parameter optimization logic, the bilateral bone air conduction sound output forms a good match in key dimensions such as phase and intensity. This cooperative effect effectively solves the problem of incoordination between the two outputs in the traditional independent adjustment mode, makes the sound received by the user more integral and natural, and reduces the auditory fatigue caused by improper coordination between the two. In addition, in view of the possible asymmetry of the bilateral ear canal structure of the user, the method can adapt to the unique characteristics of the bilateral ears through the combination of personalized models and dynamic adjustment, avoiding the problem of insufficient unilateral adaptation caused by symmetrical parameter setting, and further improving the consistency and comfort of bilateral hearing experience.

[0049] In practical application, the adjustment method can make the in-ear bone air conduction hearing aid device better adapt to the individual differences of different users, whether the user has special ear canal morphology or has bilateral auditory perception difference, and can obtain a special adjustment scheme. At the same time, the dynamic cooperative adjustment mode also widens the application scenarios of the device, whether it is daily communication, outdoor activities or indoor quiet environment, it can stably output adaptive sound signals, and improve the practicability and user acceptance of the hearing aid device. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 Figure 1 is a diagram for analyzing the influence of head rotation angle on bilateral hearing sensitivity;

[0051] Figure 2 Figure 3 is a flow chart for constructing a personalized acoustic transmission model;

[0052] Figure 3 Figure 4 is a comparison diagram of bilateral ear canal mechanical impedance matching coefficient frequency response;

[0053] Figure 4 Figure 5 is a flow chart for dynamically adjusting the initial adjustment parameters. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0055] Please refer to Figure 1The application provides an in-ear bone-air conduction hearing aid bilateral vibrator adjustment method, which comprises integrated sensor technology and acoustic modeling to realize personalized sound output adjustment for user ear canal characteristics. The in-ear bone-air conduction hearing aid bilateral vibrator adjustment method comprises a plurality of coordination steps to optimize the sound energy transmission efficiency of the bone conduction and air conduction paths. The structural characteristic data of the user's bilateral ear canals is the initial link, the contact pressure distribution and three-dimensional geometric parameters of the ear canal cartilage are collected by a micro sensor array, which provides a basis for subsequent modeling. A personalized acoustic transmission model is constructed according to the structural characteristic data, which covers the mathematical expression of the bone conduction sound wave transmission path characteristics and the air conduction sound wave transmission path characteristics. The initial adjustment parameters of the bilateral vibrator are generated based on the personalized acoustic transmission model, which include the vibration frequency and amplitude of the bone conduction vibrator and the frequency response compensation value of the air conduction loudspeaker. The bilateral auditory feedback signals of the user are monitored in real time, which involve the dynamic change amount of the bone conduction perception intensity and the air conduction perception intensity. The initial adjustment parameters are dynamically adjusted according to the auditory feedback signals to generate optimized bilateral vibrator control parameters, which promote the cooperative matching of bilateral bone-air conduction sound output. The construction of the personalized acoustic transmission model depends on the accurate input of the ear canal structural characteristic data, which is obtained by high-precision sensors to ensure that the model reflects the acoustic characteristics of the individual ear canal. The bone conduction sound wave transmission path characteristic quantifies the propagation loss of vibration energy in the cartilage, and the air conduction sound wave transmission path characteristic simulates the reflection and interference of sound waves in the ear canal cavity. The initial adjustment parameters are set based on the model output, the vibration frequency and amplitude are selected to maximize the bone conduction efficiency, and the frequency response compensation value is calibrated for the resonance peak of the air conduction path. The monitoring of the auditory feedback signal uses an embedded microphone system, the bone conduction microphone captures the vibration waveform, and the air conduction microphone records the sound pressure level change. The dynamic adjustment process is based on real-time analysis of the feedback signal, and the parameters are corrected by algorithm iteration to reduce the bilateral auditory difference. The implementation of the in-ear bone-air conduction hearing aid bilateral vibrator adjustment method relies on the close integration of hardware components, the processing unit performs model calculation and parameter optimization to realize low-delay adjustment cycles.

[0056] In the stage of obtaining the structural feature data of the user's bilateral ear canals, a contact type pressure sensor array is used to measure the pressure distribution of the ear canal cartilage region, the contact type pressure sensor array is composed of 256 micro piezoelectric elements arranged in an 8x32 matrix, the piezoelectric elements are made of lead zirconate titanate material, the size of each element is 0.5mmx0.5mm, and the center distance between the elements is 0.8mm. When the user wears the device, the contact type pressure sensor array forms mechanical contact with the surface of the ear canal cartilage, the piezoelectric elements convert mechanical stress into charge signals, the charge signals are converted into voltage signals through a charge amplifier, the data acquisition system records the voltage signal values at a sampling rate of 1kHz, the voltage signal values are converted into pressure values through a pre-calibrated pressure-voltage conversion coefficient matrix to form a two-dimensional pressure distribution map of the ear canal cartilage region. The measurement process of the contact type pressure sensor array includes an initial contact stage and a stable measurement stage, the initial contact stage lasts for 2 seconds, the device applies a pre-pressure of 0.5N to ensure that the sensor array is in full contact with the ear canal cartilage, and the stable measurement stage lasts for 5 seconds, during which a continuous pressure data sequence is collected.

[0057] The processing of the pressure distribution data adopts digital signal processing technology, the original pressure data is smoothed by a fourth-order Butterworth low-pass filter, the cutoff frequency is set to 100Hz, and high-frequency noise interference is eliminated. The Grubbs criterion is used for abnormal value detection, the pressure data sequence of each sensor element is statistically tested, and the data points deviating from the mean value by more than three times the standard deviation are removed, and the missing data is filled by the bilinear interpolation method. The extraction algorithm of peak pressure position is based on the region growing principle, the continuous high-pressure region is identified from the pressure distribution map, the centroid coordinates of each high-pressure region are calculated as the peak pressure position, and the centroid coordinates are obtained by weighted average method, and the weight is the pressure value of each point. The calculation of pressure attenuation gradient adopts the central difference method, takes the peak pressure position as the origin, calculates the pressure change rate along the radial direction with a step of 0.5mm, and the pressure change rate is obtained by linear regression of the pressure difference value and the distance ratio.

[0058] The miniature three-dimensional scanning module adopts frequency domain optical coherence tomography technology, and the light source is a super radiation light-emitting diode with a central wavelength of 850 nm and a bandwidth of 50 nm. The axial resolution is 10 μm, and the lateral resolution is 15 μm. During the scanning process, the miniature three-dimensional scanning module performs spiral scanning along the axis of the ear canal, the pitch of the scanning track is 0.2 mm, the scanning speed is 2 mm / s, the obtained interference signal is converted into depth information through Fourier transform, and ear canal three-dimensional point cloud data is formed. The point cloud data processing includes two steps of noise filtering and surface reconstruction. The statistical outlier removal algorithm is used for noise filtering, and the discrete points with a distance from the mean value of more than 1 standard deviation are removed. The surface reconstruction adopts the moving cube algorithm to generate the three-dimensional grid model of the ear canal. The calculation of the axial curvature radius is based on the ear canal centerline extraction algorithm. The centerline extraction adopts the minimum path method to find the optimal path from the entrance of the ear canal to the tympanic membrane. The path optimization target is to minimize the curvature change. The centerline is discretized into 100 equally spaced points, and the curvature radius at each point is obtained by fitting the spatial circular arc composed of the adjacent five points. The acquisition of the radial cross-sectional area change data adopts the cross-sectional analysis method, and a vertical cross section is generated every 0.5 mm along the axis of the ear canal. The cross-sectional profile is identified by the alpha-shape algorithm, and the cross-sectional area is calculated by the Green formula to calculate the area surrounded by the profile.

[0059] The integration of peak pressure position, pressure decay gradient, axial curvature radius and radial cross-sectional area change data uses a multi-modal data registration method, with the ear canal coordinate system as the registration reference, the origin of the coordinate system set as the center of the ear canal entrance, and the Z-axis along the main axis direction of the ear canal. Data registration first performs time synchronization, and the pressure data and three-dimensional scanning data are aligned through hardware timestamps. Spatial registration uses the iterative closest point algorithm to unify the pressure sensor coordinates and three-dimensional point cloud coordinates in the same coordinate system. The integrated ear canal structure feature dataset contains four data layers: the pressure distribution layer stores the pressure value matrix, the gradient distribution layer stores the pressure gradient vector, the geometric parameter layer stores the curvature radius sequence and the cross-sectional area sequence, and the spatial coordinate layer stores the three-dimensional coordinates of all data points. The calibration of the contact pressure sensor array is carried out in a constant temperature and humidity environment, using a standard pressure calibration device to apply a reference pressure of 0.1N to 5N, recording the response voltage of each sensor element, and establishing a quadratic polynomial fitting relationship between pressure and voltage. The calibration of the miniature three-dimensional scanning module uses a standard stepped calibration block with known geometric dimensions, and the correction matrix is calculated by comparing the scanning data with the true size. During the measurement process, the verification of the device wearing position is realized by real-time monitoring of the pressure distribution uniformity, and when the pressure distribution uniformity index is less than 0.3, the user is automatically prompted to adjust the wearing posture. The update mechanism of the ear canal structure feature dataset is triggered based on the number of uses, and after every 10 uses, the data is automatically reacquired, and the update process uses the sliding window average method, with a weight ratio of 7:3 between new data and historical data. Data storage uses a block storage structure, with each ear canal data stored independently as an HDF5 format file, and the file header containing metadata such as acquisition time, device serial number, and user identification. Data transmission uses the AES-256 encryption protocol to ensure user privacy and security.

[0060] The dynamic tracking of the peak pressure position is achieved by continuous monitoring with a monitoring period of 100 ms, and the pressure distribution is automatically re-measured when the peak pressure position moves more than 2 mm. The calculation of the pressure decay gradient introduces directional decomposition, which decomposes the gradient into axial and radial components to analyze the pressure change characteristics in different directions. The scanning parameters of the miniature three-dimensional scanning module are adaptively adjusted, and the scanning range is dynamically adjusted according to the ear canal size. For large ear canals, an 8mm x 8mm scanning field is used, and for small ear canals, a 6mm x 6mm scanning field is used. The calculation results of the axial curvature radius are verified by using the cross-validation method, and the curvature radius is calculated by using cubic spline interpolation and polynomial fitting respectively, and the average value of the two results is taken as the final value. The analysis of the radial cross-sectional area change data includes the calculation of the area change rate, which is obtained by the differential of the area difference and the distance ratio of adjacent cross sections. The quality evaluation of the ear canal structure feature dataset is based on data integrity check, and the check items include data point number, value range and spatial distribution uniformity. Unqualified data automatically triggers the re-collection process. The power management of the contact pressure sensor array adopts a dynamic power supply strategy, and only the sensors in the active area are powered during the measurement stage, and the non-active area sensors are in a dormant state. The optical protection of the miniature three-dimensional scanning module adopts a dirt-repellent coating to prevent earwax from contaminating the optical window, and integrates a self-cleaning mechanism that automatically starts a cleaning cycle after each use. User feedback during the data collection process is achieved through vibration prompts, and a specific pattern of vibration signal is emitted when the measurement is complete, prompting the user to remove the device. The compressed storage of the ear canal structure feature dataset uses a lossy compression algorithm that retains key features while compressing the data size to 30% of the original size. The compression algorithm is based on principal component analysis, which retains 95% of the original information. The data access interface provides standard API functions to support upper-level module calls to the dataset content, including data query, update, deletion, and other operations. The abnormal handling mechanism of the measurement process includes timeout detection, which automatically terminates the process when the single ear canal measurement time exceeds 15 seconds, and generates an error log for technical personnel to analyze.

[0061] Example 2: see Figure 2In the construction of the individualized acoustic transmission model stage, the mechanical impedance matching coefficient of the bone conduction vibrator is calculated based on the ear canal structure feature dataset. The calculation of the mechanical impedance matching coefficient adopts the method of combining the lumped parameter model and the distributed parameter model. The lumped parameter model simplifies the ear canal cartilage as a mass-spring-damper system, and the mass parameter is calculated by the cartilage density and volume. The density adopts the standard value of 1.1 g / cm³, and the volume is extracted from the three-dimensional scanning data. The distributed parameter model divides the cartilage into small units to independently calculate the impedance, and the overall matching coefficient is obtained by superposition. The matching coefficient calculation uses the circuit analogy method to compare the vibrator output impedance with the cartilage input impedance, with a numerical range of 0 to 1, and the closer to 1, the higher the matching degree. The calculation of the acoustic resonance frequency adopts the transfer matrix method, which models the ear canal as a cylindrical waveguide to solve the wave equation to obtain the fundamental frequency and harmonic frequency. The attenuation characteristic is quantified by calculating the energy loss in the sound wave propagation, considering the viscous loss and heat conduction loss. The air conduction frequency response compensation curve is generated by combining the axial curvature radius and the radial cross-sectional area data, and the ear canal is divided into small segments to calculate the acoustic characteristics based on the acoustic transmission line theory. The mechanical impedance matching coefficient and the air conduction frequency response compensation curve are associated and mapped through the state space modeling to form the individualized acoustic transmission model. The state space model includes four state variables of vibration displacement, vibration velocity, sound pressure and particle vibration velocity, the input is the driving voltage and the input electric signal, and the output is the eardrum sound pressure and the bone vibration acceleration.

[0062] Based on the individualized acoustic transmission model, the initial adjustment parameters of the bilateral vibrator are generated, and the reference vibration frequency and vibration amplitude of the bone conduction vibrator are set according to the mechanical impedance matching coefficient. The reference vibration frequency is selected by scanning the range of 200 Hz to 4000 Hz to select the peak point of the matching coefficient, and the vibration amplitude is calculated based on the impedance matching principle to correspond to the maximum power transmission amplitude. The air conduction frequency response compensation curve is used to adjust the high-frequency gain of the air conduction speaker, and the phase shift is compensated for the frequency band above 3000 Hz. The initial adjustment parameters are stored in the form of a parameter matrix, including frequency value, amplitude value and high-frequency gain value. The parameter optimization adopts the gradient descent algorithm to iteratively adjust the target function of the eardrum sound pressure uniformity. The calculation of the mechanical impedance matching coefficient introduces temperature compensation, which corrects the material parameters by monitoring the temperature change in the ear canal through the temperature sensor. The calculation of the acoustic resonance frequency considers the influence of humidity, and adjusts the sound speed of air according to the relative humidity. The air conduction frequency response compensation curve adjusts the hearing sensitivity parameters according to the user's age and gender. The individualized acoustic transmission model is updated regularly, and recalibrated after 100 hours of use. The fine tuning of the reference vibration frequency adopts the sweep excitation method, and the best frequency point is selected by scanning near the reference frequency. The vibration amplitude setting considers the hearing comfort, and sets the maximum amplitude limit. The high-frequency gain adjustment adopts gradual change, and uses a gentle gain slope near the resonance frequency. The bilateral parameter coordination is realized through mutual constraint, limiting the frequency difference, amplitude difference and gain difference range.

[0063] The real-time running of the personalized acoustic transmission model adopts an embedded system, the processor uses an ARM Cortex-M7 core, and the task scheduling adopts a priority preemptive algorithm. The initial adjustment parameters are stored in a non-volatile memory, and the transmission adopts a digital serial interface. The adaptive learning of the mechanical impedance matching coefficient introduces a neural network algorithm, and the acoustic resonance frequency tracking adopts an adaptive filtering technology. The smoothing processing of the air conduction frequency response compensation curve uses a minimum phase filter, and the model robustness is enhanced through Monte Carlo analysis. The temperature compensation of the reference vibration frequency is realized through a lookup table, and the vibration amplitude safety monitoring includes an overload protection mechanism. The frequency band segmentation of the high-frequency gain adopts a critical band division method, and the initial adjustment parameter verification is completed through an artificial ear simulator. In actual application, the system continuously monitors the environmental noise level and the user's head movement state, and dynamically adjusts the parameter settings. When a strong noise environment is detected, the gain weight of the air conduction path is automatically increased to enhance the speech clarity; in a quiet environment, the bone conduction path parameters are optimized to provide a more natural sound quality. The system also establishes a personal hearing preference model based on long-term user data, and continuously optimizes the parameter configuration through machine learning algorithms. All parameter adjustment processes ensure smooth transition to avoid auditory discomfort, ensuring that users can obtain stable auditory experience in different use scenarios.

[0064] Referring to Figure 3 In the experimental study of the frequency characteristics of the mechanical impedance matching coefficient of the bilateral ear canal, the results are based on the frequency response analysis technology of the mechanical impedance matching coefficient. In the specific implementation, the mechanical impedance matching coefficient of the left ear canal is represented by a blue curve, and that of the right ear canal is represented by an orange curve. The two curves respectively quantify the impedance matching coefficient variation trend in the linear frequency range of 200Hz to 4000Hz. The reference frequency points are identified by dashed lines, with 1500Hz for the left ear and 1700Hz for the right ear, corresponding to the peak value region of the matching coefficient of each ear canal. The coordinate system uses linear scales, with key frequency points marked at 500Hz intervals on the horizontal axis, and the vertical axis covers a matching coefficient range of 0.4 to 0.9, with grid auxiliary lines to ensure data reading accuracy. Data analysis shows that the left ear canal reaches a matching coefficient peak of about 0.7 in the mid-frequency band (800-1500Hz), indicating optimal impedance matching in this region; the right ear canal presents a more gentle attenuation characteristic in the high-frequency band (2000-4000Hz), with a matching coefficient of about 0.7 at 1700Hz, reflecting the stability of high-frequency transmission. In the parameterization process, the mechanical impedance matching coefficient is obtained through experimental measurement, and the closer its value is to 1, the higher the vibration coupling efficiency of the vibrator and the ear canal cartilage, and the shape difference of the frequency response curve can be used for calibration of the personalized acoustic model.

[0065] Example 3: In-ear bone-aided hearing dual vibrator adjustment method In the stage of real-time monitoring of the user's bilateral auditory feedback signal, the vibration waveform feedback by the vibrator is collected by a bone conduction microphone. The bone conduction microphone uses a piezoelectric acceleration sensor with a sensitivity of 10 mV / g and a frequency response range of 20 Hz to 10 kHz. During the vibration waveform collection process, the bone conduction microphone is directly coupled to the radiation surface of the bone conduction vibrator. A charge amplifier is used to convert the piezoelectric signal to a voltage signal. The charge amplifier gain is set to 1 V / pC, the low-frequency cutoff frequency is 0.1 Hz, and the high-frequency cutoff frequency is 20 kHz. The sampling system uses a 24-bit sigma-delta analog-to-digital converter with a sampling rate of 48 kHz and a dynamic range of 110 dB. The collected voltage signal is processed by digital filtering. First, a notch filter is used to eliminate 50 Hz power frequency interference. Then, a digital bandpass filter is used to retain the effective frequency band of 20 Hz-10 kHz.

[0066] The harmonic distortion component and the energy decay rate in the waveform are extracted. The harmonic distortion analysis uses a frequency spectrum analysis method based on fast Fourier transform. The continuously collected vibration waveform data is processed in frames. The frame length is 1024 sampling points, and the frame shift is 512 points. A Hanning window is used to reduce spectral leakage. For each frame of data, a 4096-point FFT transform is performed to calculate the fundamental component amplitude and the amplitude of each harmonic component (n=2,3,...,10). The total harmonic distortion is calculated using the following formula:

[0067] ;

[0068] where: represents the total harmonic distortion, represents the fundamental voltage effective value, represents the nth harmonic voltage effective value. The energy decay rate is obtained by analyzing the envelope of the vibration waveform. The Hilbert transform is used to extract the waveform envelope. The envelope is logarithmically fitted, and the decay time constant is calculated, which is defined as the time required for the amplitude to decay to 1 / e of the initial value.

[0069] The residual sound pressure level in the ear canal is monitored by an air conduction microphone, which uses a pre-polarized MEMS microphone with a sensitivity of -38 dBV / Pa and a frequency range of 20 Hz-20kHz. The sound pressure level measurement uses equivalent continuous A-weighted sound level with an integration time of 1 second and an update rate of once per second. The sound wave interference strength of the air conduction path is calculated using a dual microphone method, with the main microphone located at the entrance of the ear canal and the reference microphone near the eardrum. The interference strength is evaluated by calculating the cross-correlation function between the two microphone signals, using the Welch method with a segment length of 1024 points, Hanning window weighting, and an overlap rate of 50%. The harmonic distortion component, energy decay rate, and sound wave interference strength are quantified as auditory feedback signals using a normalization process. The harmonic distortion is normalized to a value in the range of 0-1, with a mapping of 0 when THD is less than 1% and a mapping of 1 when THD is greater than 10%, with linear interpolation for intermediate values. The energy decay rate is normalized based on the time constant When is less than 1 ms, it is mapped to 1, and is greater than 10 ms, it is mapped to 0. The sound wave interference strength is normalized using the maximum value of the cross-correlation function, with a mapping of 1 for complete coherence and a mapping of 0 for complete incoherence.

[0070] The signal conditioning circuit of the bone conduction microphone includes a programmable gain amplifier with a gain range of 20-60 dB, which is automatically adjusted according to the input signal amplitude. Real-time calculation of the harmonic distortion component uses a sliding window spectral analysis method, with an update rate of once every 10 ms for the harmonic distortion value. The calculation of the energy decay rate introduces temperature compensation, which corrects the changes in material acoustic properties with temperature based on readings from a temperature sensor installed on the surface of the vibrator. The sound pressure level measurement of the air conduction microphone uses a true RMS detection circuit with an adjustable integration time constant to meet different measurement requirements. The calculation of the sound wave interference strength takes into account the multi-path propagation effect, with contributions from direct sound and reflected sound separated by multi-channel cross-correlation analysis. The time synchronization of the auditory feedback signals uses a hardware trigger signal, with left and right ear data synchronized to the microsecond level using a precision clock chip. The calibration of the bone conduction microphone uses a standard vibration table, with 9-point calibration in the acceleration range of 0.1 g to 10 g to establish the voltage-acceleration conversion relationship. The calibration of the air conduction microphone uses an acoustic calibrator, with calibration at frequency points of 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, 4000 Hz, and 8000 Hz.

[0071] The monitoring of harmonic distortion components includes two levels of total harmonic distortion and individual harmonic distortion, which reflect the overall distortion characteristics of the vibration waveform and the distortion of specific frequency components, respectively. The analysis of energy decay rate uses a multi-exponential model fitting, and simultaneously calculates the time constants of fast and slow decay components. The spatial characteristics of acoustic interference intensity are obtained by analyzing the coherence of different frequency components, providing frequency-selective interference characteristic information. The quality evaluation of auditory feedback signals includes signal-to-noise ratio checking, and when the signal-to-noise ratio is lower than 20 dB, the data frame is automatically discarded and the reacquisition process is triggered. The mechanical installation of bone conduction microphones uses vibration isolation design, and a silicone buffer layer is set between the sensor and the shell to reduce external vibration interference. The acoustic design of air conduction microphones includes acoustic damping net and waterproof film, providing flat frequency response characteristics. The frequency domain analysis of harmonic distortion components uses Zoom-FFT technology to achieve higher frequency resolution in the frequency band of interest. The calculation of energy decay rate introduces decay mode recognition algorithm to automatically identify linear decay, exponential decay and other decay modes.

[0072] The data transmission of auditory feedback signals uses CRC-16 error detection coding to ensure the reliability of data transmission. The temperature compensation of bone conduction microphones is realized by lookup table, and in the temperature range of-10℃ to +60℃, a compensation coefficient is set every 5℃. The frequency response compensation of air conduction microphones is realized by digital filter bank, which provides a frequency response flatness of ±0.5 dB in the range of 20 Hz-20 kHz. The monitoring of harmonic distortion components includes the separate statistics of even and odd harmonics, providing more detailed information for the evaluation of vibrator working state. The calculation of energy decay rate includes decay starting point detection, which uses change point detection algorithm to accurately identify the time of decay start. The analysis of acoustic interference intensity includes interference spectrum feature extraction, calculating parameters such as center frequency and bandwidth of interference spectrum. The cache management of auditory feedback signals uses double buffering mechanism to realize continuous data acquisition and seamless processing. The long-term stability monitoring of bone conduction microphones is realized by regular self-checking, and the sensor self-checking program is automatically executed once a day.

[0073] The directional characteristics of air conduction microphones are digitally compensated by measuring the frequency response at different incident angles. The time-varying characteristics of harmonic distortion components are obtained by calculating the first derivative of THD, reflecting the dynamic characteristics of distortion degree changes. The adaptive calculation of energy decay rate automatically adjusts the analysis parameters according to signal strength, ensuring the analysis accuracy in weak signal conditions. The environmental adaptation of sound wave interference intensity is realized through background noise spectrum estimation, automatically adjusting the detection threshold under different environmental noise levels. The data fusion of auditory feedback signals uses a Kalman filter-based method to dynamically adjust the fusion weight according to the signal-to-noise ratio of each parameter. The mechanical coupling state of bone conduction microphones is evaluated by monitoring the installation resonance frequency, and a maintenance prompt is issued when the resonance frequency deviation exceeds 10% of the nominal value. The acoustic environment recognition of air conduction microphones automatically identifies different acoustic environments such as quiet, noisy, and reverberation by analyzing the statistical characteristics of the noise spectrum. The abnormal detection of harmonic distortion components uses a statistical process control-based method to establish the normal fluctuation range of THD, triggering an early warning when it exceeds the range.

[0074] The multi-scale analysis of energy decay rate simultaneously calculates short-term (millisecond level) and long-term (second level) decay characteristics, fully characterizing the energy decay behavior of the oscillator. The frequency dependence of sound wave interference intensity is obtained by calculating the inter-coherence function in different frequency bands, providing 1 / 3 octave resolution for interference characteristic analysis. The data compression of auditory feedback signals uses a lossy compression algorithm based on wavelet transform, retaining the main signal features at a compression ratio of 4:1. Fault diagnosis of bone conduction microphones is achieved by monitoring the bias voltage and output noise level, allowing real-time detection of sensor operating status. The sound pressure level measurement of air conduction microphones uses a time-weighted integration method, with an integration time constant that can be set to 125ms or 1s according to measurement requirements. Real-time calculation of sound wave interference intensity uses a recursive estimation algorithm to reduce computational complexity and memory requirements. The time stamp synchronization of auditory feedback signals uses the IEEE1588 precision time protocol to achieve sub-millisecond time synchronization accuracy. Mechanical impedance matching of bone conduction microphones is achieved by monitoring the load impedance changes, allowing real-time adjustment of driving parameters to optimize energy transmission efficiency.

[0075] The display of harmonic distortion components uses time-frequency analysis techniques to show the variation of harmonic distortion over time through spectrograms. The trend analysis of energy decay rate uses a sliding window averaging method, with the window length adaptively adjusted within 1-10 seconds according to signal characteristics. The spatial distribution characteristics of sound wave interference intensity are estimated by beamforming technology, providing directional information of the sound field. The long-term recording of auditory feedback signals uses a circular storage method, saving complete data for the past 24 hours for historical trend analysis. The installation torque monitoring of bone conduction microphones is achieved through a miniature torque sensor, ensuring that the contact pressure between the sensor and the vibrator is within the optimal range. The sound calibration of air conduction microphones uses a multi-frequency point calibration method, providing accurate sound pressure level measurements in the main audio frequency band. The traceability analysis of harmonic distortion components locates the main link of distortion by comparing the differences between the original waveform and the reconstructed waveform. The environmental compensation of energy decay rate considers the comprehensive influence of temperature, humidity, and air pressure, and uses a multiple regression model for real-time correction. The time-varying characteristics of sound wave interference intensity are captured through short-time coherence analysis, with an analysis window length of 100 ms and an overlap rate of 75%, balancing the requirements of time resolution and frequency resolution. The wireless transmission of auditory feedback signals uses an adaptive frequency hopping mechanism to achieve reliable data transmission in the 2.4 GHz frequency band. The life prediction of bone conduction microphones is based on the comprehensive evaluation of running time, operating temperature, and vibration intensity, providing early warning of sensor performance degradation. The dustproof design of air conduction microphones uses a nano oil-repellent coating to effectively prevent earwax and dust accumulation from affecting measurement accuracy.

[0076] Example 4: refer to Figure 4 In the initial adjustment parameter adjustment stage according to the auditory feedback signal, the difference in harmonic distortion components between the two sides is compared. The difference in harmonic distortion components is calculated using a difference algorithm based on frequency domain analysis. The harmonic distortion data of the left and right ears is first processed for time alignment, using a precise clock chip to achieve time synchronization with a synchronization accuracy of 10 microseconds. The time-aligned data is analyzed using a sliding window, with a window length of 500 milliseconds and a Blackman-Harris window function to reduce spectral leakage. The difference calculation of harmonic distortion components includes the difference of fundamental components and the difference of second to tenth harmonic components, with each harmonic component difference represented by the root mean square value. The difference analysis algorithm calculates the relative difference of the corresponding harmonic amplitudes of the left and right ears in real time, as well as the harmonic phase difference, comprehensively evaluating the imbalance of harmonic distortion components.

[0077] The vibration frequency of the bone conduction oscillator is corrected to balance the energy of both harmonics. The correction process employs a fuzzy proportional-integral-derivative (FID) control algorithm. The FID controller comprises three parts: a fuzzification interface, a fuzzy inference engine, and a defuzzification interface. The fuzzification interface converts the differences in harmonic distortion components into fuzzy linguistic variables, defining three fuzzy sets: "small," "medium," and "large." The fuzzy inference engine performs inference based on 25 fuzzy rules, with the rule format being "if the difference is small and the rate of change is small, then the frequency adjustment amount is small." The defuzzification interface uses the centroid method to convert the fuzzy output into a precise frequency correction amount, with the correction range limited to ±2Hz and the correction speed adjustable. The phase delay of the air-conducting loudspeaker is adjusted based on the intensity of acoustic interference, using an adaptive filtering algorithm. The order of the adaptive filter is set to 32, and the step size factor is adjustable from 0.001 to 0.01. The phase delay adjustment process calculates the optimal phase delay value in real time to minimize the acoustic interference intensity. The adjustment algorithm includes an LMS (Least Mean Square) adaptive algorithm, which iteratively updates the filter coefficients to gradually approach the optimal phase delay. The phase delay adjustment range covers 0-180 degrees, with a resolution of 0.1 degrees and an adjustment period of 10 milliseconds.

[0078] The dynamic range of vibration amplitude is optimized based on the energy decay rate, employing a multi-segment compression and expansion technique. The dynamic range is divided into eight frequency bands, each undergoing independent compression and expansion. The compression ratio is dynamically adjusted according to the energy decay rate, using a 2:1 compression ratio for rapid decay and a 1.5:1 compression ratio for slow decay. The expansion threshold is adaptively set based on historical amplitude data to ensure the accuracy of dynamic range optimization. The amplitude optimization algorithm monitors the input signal level in real time and automatically adjusts the gain coefficient to ensure stable output amplitude. Optimized dual-sided oscillator control parameters are generated and stored in a structured data format. The data structure includes a file header, parameter body, and checksum. The file header records the parameter version number, generation timestamp, and device identifier. The parameter body stores specific parameter values ​​such as vibration frequency, phase delay, and amplitude range. The checksum uses the CRC32 algorithm to ensure data integrity. The parameter update mechanism employs a differential update method, transmitting only changed parameter values ​​to improve transmission efficiency.

[0079] The real-time monitoring of harmonic distortion component difference adopts a multi-thread architecture, with a data acquisition thread responsible for obtaining raw data, a processing thread for difference calculation, and a control thread for parameter adjustment. The vibration frequency correction introduces a temperature compensation mechanism, which monitors the environmental temperature in real time through a temperature sensor and corrects the frequency temperature drift. The phase delay adjustment considers individual differences and customizes the adjustment strategy according to the user's ear canal characteristics. The amplitude dynamic range optimization includes amplitude limiting protection, which automatically enables soft limiting function when detecting amplitude abnormalities. The verification of optimized bilateral vibrator control parameters uses a multi-layer verification mechanism. The first layer of verification checks the reasonableness of the parameter range, ensuring that the parameter value is within the preset safe range. The second layer of verification verifies the logical consistency between parameters to avoid contradictory parameter settings. The third layer of verification verifies the parameter effect through simulation testing to predict the system response after parameter adjustment. The verification process is fully automated, and when the verification fails, it automatically reverts to the previous set of qualified parameters. Refer to Table 1 for a detailed correspondence between sound wave interference intensity and phase delay adjustment parameters.

[0080] Table 1: Detailed correspondence table between sound wave interference intensity and phase delay adjustment parameters

[0081]

[0082] The long-term statistical analysis of harmonic distortion component difference records the last 1000 groups of difference data and identifies the difference trend through time series analysis. The effect evaluation of vibration frequency correction adopts a combination of objective indicators and subjective evaluation, with objective indicators including harmonic distortion improvement and subjective evaluation collected through user feedback. The stability test of phase delay adjustment includes stress testing to verify the robustness of the adjustment algorithm under extreme working conditions. The adaptive learning function of amplitude dynamic range optimization continuously optimizes parameter settings based on usage habits. The quality management system for generating optimized bilateral vibrator control parameters includes parameter version control, change record, and rollback mechanism. Each parameter version saves complete configuration information, and the change record details the reason and effect of each parameter adjustment. The rollback mechanism ensures quick recovery to a stable state in case of parameter adjustment failure. The parameter management interface provides parameter import and export functions, supporting batch operations and fine adjustment of individual parameters.

[0083] The alarm system of harmonic distortion component difference sets multiple thresholds, triggering different levels of alarms according to the severity of the difference. First-level alarm prompts slight imbalance, second-level alarm prompts obvious imbalance, and third-level alarm prompts serious imbalance. The safety protection of vibration frequency correction includes frequency jump detection, which automatically pauses correction when frequency anomalies are detected. The abnormal processing mechanism of phase delay adjustment monitors the adjustment effect, and starts the fault diagnosis process when the adjustment effect does not meet the expectations. The compatibility test of the optimized bilateral vibrator control parameters covers different age groups, from teenagers to the elderly, to test the parameter effect. The personalized parameter setting records the optimal parameter combination of each user, establishing a personal hearing characteristic model. The migration learning function of control parameters allows quick adaptation to new users, shortening the parameter optimization time. The parameter adjustment log records detailed operation records for subsequent analysis and optimization. The real-time performance of the optimized bilateral vibrator control parameters is achieved through optimization algorithms, with key algorithms optimized at the assembly level, with a calculation delay controlled within 5 milliseconds. The parameter transmission protocol uses a custom real-time transmission protocol that supports error retransmission and flow control. The parameter storage uses EEPROM memory to ensure that parameters are not lost after power failure. The parameter backup mechanism automatically backs up parameter settings periodically to prevent data loss.

[0084] The visual analysis of harmonic distortion component difference provides time-frequency graph display, which intuitively shows the time-varying characteristics of the difference. The historical data trend analysis of vibration frequency correction helps optimize the correction strategy. The precision calibration of phase delay adjustment is performed periodically using a standard signal source. The parameter adaptive adjustment of amplitude dynamic range optimization automatically optimizes parameter settings according to the use environment. The verification experiment design of the optimized bilateral vibrator control parameters includes a control group and an experimental group, using a double-blind test method to ensure the objectivity of the results. The optimization of the parameter adjustment algorithm uses genetic algorithms to automatically find the optimal parameter combination. The safety boundary of the control parameters is determined based on a large amount of experimental data, and the long-term tracking of the parameter adjustment effect is achieved through a cloud platform to continuously optimize the parameter settings.

[0085] In the parameter coordination constraint phase, a complete rule system is established. The parameter coordination constraint rules include six core constraint conditions. The frequency range constraint sets the bone conduction vibrator working frequency band to 200 Hz to 4 kHz, the air conduction speaker working frequency band to 100 Hz to 8 kHz, and the overlapping interval to 400 Hz to 2 kHz. The gain limit constraint requires that the maximum gain of the bone conduction vibrator in the overlapping area not exceed 60 dB, the maximum gain of the air conduction speaker not exceed 55 dB, and the bilateral gain difference be controlled within 3 dB. The phase tolerance constraint limits the phase deviation to within ±5 degrees, and the time delay difference constraint requires that the bilateral time delay difference be less than 10 microseconds. The harmonic distortion constraint ensures that the total harmonic distortion is less than 5%, and the intermodulation distortion constraint requires that the intermodulation distortion product be less than -40 dB.

[0086] The solution of the parameter coordination constraint rule adopts a linear programming method, and uses a simplex algorithm to iteratively calculate the feasible solution space. Each parameter variable defines a clear numerical range, the frequency parameter is in units of hertz, the gain parameter is in units of decibels, the phase parameter is in units of degrees, the time delay parameter is in units of microseconds, and the distortion parameter is expressed in percentage or decibels. The constraint solver continuously monitors the inequality constraint condition, and automatically adjusts the relaxation variable when a constraint conflict is detected. The parameter update mechanism re-evaluates the constraint condition every 100 milliseconds and dynamically optimizes the inequality boundary range. When the control parameters of the bilateral vibrator trigger the frequency band overlap interval, the system starts the energy redistribution algorithm based on the auditory masking effect. The energy redistribution algorithm divides the 400-2000 Hz overlap interval into 24 one-third octave subbands, and independently calculates the bone conduction weight and air conduction weight for each subband. The weight value ranges from 0 to 1, and satisfies the constraint condition that the sum of the weights is always 1. The weight calculation is based on the instantaneous sound pressure level, spectral characteristics and auditory masking threshold of the signal in the subband, and the auditory masking threshold is calculated in real time by a psychoacoustic model. The system dynamically adjusts the weight distribution strategy according to the masking effect strength, increasing the air conduction weight to 0.7 in strong masking effect, and increasing the bone conduction weight to 0.6 in weak masking effect.

[0087] The implementation of the energy redistribution algorithm includes three core links. The signal analysis link performs a 256-point short-time Fourier transform on the input signal using a Hanning window function with a 50% overlap rate. The weight calculation link combines a lookup table method and a bilinear interpolation technique, and pre-stores the optimal weight combination for typical acoustic scenarios. The weight application link is implemented through a digital filter bank, with each subband corresponding to an independent bandpass filter whose coefficients are dynamically updated according to the real-time weight value. The entire algorithm execution period is controlled within 20 milliseconds to ensure real-time system response. The head rotation angle data is collected by a nine-axis inertial measurement unit, which integrates a three-axis gyroscope, a three-axis accelerometer, and a three-axis magnetometer with a sampling frequency of 100 Hz and a data output frequency of 50 Hz. The sensor data is represented using quaternion notation, and data fusion is performed using an extended Kalman filter algorithm to output pitch angle, roll angle, and yaw angle data with measurement accuracies of 1°, 1°, and 3°, respectively. The data processing process includes three steps: data preprocessing, coordinate transformation, and angle compensation. The data preprocessing uses median filtering to remove impulse noise and sliding average filtering to smooth high-frequency fluctuations. Coordinate transformation converts device coordinate system data to world coordinate system, and angle compensation corrects geomagnetic interference through hard iron compensation and soft iron compensation algorithms.

[0088] The update of the personalized acoustic transfer model is based on head rotation angle data and uses a head-related transfer function library for spatial audio processing. The transfer function library contains transmission function data for 360 azimuth angles and elevation angles ranging from -40 to +90 degrees. The model update algorithm uses a k-nearest neighbor algorithm to select the optimal transfer function, ensures phase continuity through spherical linear interpolation, and completes model reconstruction every 50 milliseconds. The parameter reset process uses a gradual transition algorithm to calculate the weighted average of the new and old parameter sets every 20 milliseconds, with a transition time constant of 100 milliseconds. The frequency parameters use logarithmic interpolation, and the sound pressure level difference is strictly controlled within 6 dB to ensure smooth transition of the auditory experience. The frequency band overlap interval detection is achieved through real-time spectral monitoring, with power spectral density calculations for bone conduction and air conduction signals every 10 milliseconds. The power spectral density calculation uses the Welch method with a 256-point segment length, Hanning window weighting, and an overlap rate of 50%. When the correlation coefficient of the bone conduction and air conduction power spectrum in the overlapping frequency band exceeds 0.7, the system determines that there is effective overlap and dynamically adjusts the interval boundaries, with an adjustment amplitude limited to ±10% of the center frequency. The detection algorithm integrates spectral calibration, noise threshold, and confidence evaluation modules, with pink noise reference signals used for spectral calibration and noise thresholds adjusted adaptively based on environmental noise levels.

[0089] The weight adjustment strategy is based on the Bark scale critical band division principle, which divides the auditory frequency band into 24 sub-bands. The weight adjustment algorithm includes fast response and fine adjustment modes, with the fast response mode suitable for transient signal processing and a response time of 20 milliseconds, and the fine adjustment mode optimized for steady-state signals with a response time of 100 milliseconds. The weight smoothing process uses a first-order inertia link with a time constant adjustable between 10 and 100 milliseconds. The personalized acoustic transfer model uses an incremental learning strategy for updating, with new data weights set to 0.7 and historical data weights set to 0.3. The model parameters are updated using the recursive least squares method with a forgetting factor of 0.95. The system safety monitoring mechanism includes multiple protection measures, including amplitude limiters to prevent output signal overload, phase continuity to avoid phase jumps, and spectral equalizers to maintain frequency response flatness. The safety monitoring algorithm analyzes the output signal characteristics in real time and switches to a safe mode within 5 milliseconds in abnormal situations. Parameter coordination constraint rules are stored in Extensible Markup Language format, supporting dynamic loading and online updating. The rule library maintains metadata such as rule identifiers, effective times, and constraint conditions, and checks for updates daily through a secure Hypertext Transfer Protocol connection server. Rule parsing uses a recursive descent algorithm based on Backus-Normalform-defined syntax rules.

[0090] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting; it is not intended to exclude myriad other embodiments of the present application that other inventors can develop based on the same general inventive concepts embodied by the described embodiments. That is, although the present application is described in terms of particular embodiments and implementations, it is to be understood that the terminology used is for the purpose of descriptive clarity and that it should be taken in its broadest possible sense. For example, the terms "a", "an", and "the" include both singular and plural referents unless the context clearly dictates otherwise. The terms "comprises", "comprising", "includes", "including" and the like can be used in conjunction with the term "consisting of to include the elements or steps listed after such conjunctive language, but not to the exclusion of other elements or steps. The singular forms "a", "an" and "the" include plural referents unless the context clearly dictates otherwise.

[0091] While the embodiments of the application have been shown and described, it is to be understood that the application is not limited to these embodiments. Rather, it is the intent that all variations and modifications, which fall within the spirit and broad scope of the underlying principles and application disclosed herein, are to be regarded as within the scope of the application. Various embodiments of the application have been described in fulfillment of the various objectives of the application. It should be recognized that numerous modifications and adaptations thereof will be apparent to those skilled in the art. For example, although the application has been described in the context of a single user, the application can be used in a multi-user environment. Although the application has been described in the context of a single computer, the application can be used in a networked environment. Although the application has been described in the context of a single computer, the application can be used in a multi-computer environment. Although the application has been described in the context of a single computer, the application can be used in a multi-computer environment. Although the application has been described in the context of a single computer, the application can be used in a multi-computer environment. Although

Claims

1. A method for bilateral vibrator adjustment in in-ear bone air conduction auxiliary hearing aids, characterized in that, include: The system acquires structural feature data of both ear canals of the user and collects the contact pressure distribution of ear canal cartilage and the three-dimensional geometric parameters of ear canal morphology through miniature sensors. A personalized acoustic transmission model is constructed based on the structural feature data. The personalized acoustic transmission model includes bone conduction acoustic wave transmission path features and air conduction acoustic wave transmission path features. The initial adjustment parameters of the two-sided oscillator are generated based on the personalized acoustic transmission model. The initial adjustment parameters include the vibration frequency and vibration amplitude of the bone conduction oscillator and the frequency response compensation value of the air conduction loudspeaker. Real-time monitoring of the user's bilateral auditory feedback signals, including dynamic changes in bone conduction sensing intensity and air conduction sensing intensity; The initial adjustment parameters are dynamically adjusted based on the auditory feedback signal to generate optimized bilateral oscillator control parameters, thereby achieving coordinated matching of bilateral bone air conduction sound output.

2. The method for bilateral resonator adjustment of in-ear bone conduction auxiliary hearing aids according to claim 1, characterized in that, The acquisition of structural feature data of the user's bilateral ear canals includes: The pressure distribution in the ear canal cartilage region is measured using a contact pressure sensor array, and the peak pressure location and pressure decay gradient in the pressure distribution are extracted. Data on the axial radius of curvature and radial cross-sectional area changes of the ear canal were collected using a miniature 3D scanning module. The peak pressure location, pressure decay gradient, axial radius of curvature, and radial cross-sectional area variation data are integrated into an ear canal structural feature dataset.

3. The method for bilateral resonator adjustment of in-ear bone conduction auxiliary hearing aids according to claim 2, characterized in that, The construction of a personalized acoustic transmission model based on the structural feature data includes: The mechanical impedance matching coefficient of the bone conduction oscillator is calculated based on the ear canal structural feature dataset. The mechanical impedance matching coefficient reflects the vibration coupling efficiency between the oscillator and the ear canal cartilage. The acoustic resonant frequency and attenuation characteristics of the air conduction path in the ear canal are extracted, and the air conduction frequency response compensation curve is generated by combining the axial radius of curvature and radial cross-sectional area variation data. The mechanical impedance matching coefficient and the air conduction frequency response compensation curve are correlated and mapped to form the personalized acoustic transmission model.

4. The method for bilateral resonator adjustment of in-ear bone conduction auxiliary hearing aids according to claim 3, characterized in that, The initial adjustment parameters for generating the dual-sided oscillator based on the personalized acoustic transmission model include: The reference vibration frequency and vibration amplitude of the bone conduction oscillator are set according to the mechanical impedance matching coefficient, so that the vibration energy is concentrated at the peak pressure position of the ear canal cartilage. The high-frequency gain of the air-conducting loudspeaker is adjusted according to the air conduction frequency response compensation curve to compensate for the sound wave phase shift caused by the axial curvature of the ear canal. The reference vibration frequency, vibration amplitude, and high-frequency gain are combined to form the initial adjustment parameters.

5. The method for bilateral resonator adjustment of in-ear bone conduction auxiliary hearing aids according to claim 4, characterized in that, The real-time monitoring of the user's bilateral auditory feedback signals includes: The vibration waveform fed back by the oscillator is acquired by a bone conduction microphone, and the harmonic distortion components and energy decay rate in the waveform are extracted. The residual sound pressure level in the ear canal is monitored by an air conduction microphone, and the sound wave interference intensity of the air conduction path is calculated. The harmonic distortion components, energy decay rate, and acoustic interference intensity are quantized into the auditory feedback signal.

6. The method for bilateral resonator adjustment of in-ear bone conduction auxiliary hearing aids according to claim 5, characterized in that, The step of dynamically adjusting the initial adjustment parameters based on the auditory feedback signal includes: By comparing the differences in harmonic distortion components on both sides, the vibration frequency of the bone conduction oscillator is corrected to balance the harmonic energy on both sides. The phase delay of the air-conducting loudspeaker is adjusted according to the intensity of the sound wave interference to suppress the sound wave cancellation effect in the ear canal; Based on the energy decay rate, the dynamic range of the vibration amplitude is optimized, and the optimized control parameters of the dual-sided oscillator are generated.

7. The method for bilateral resonator adjustment of in-ear bone conduction auxiliary hearing aids according to claim 6, characterized in that, Also includes: Establish a parameter coordination constraint rule for the two-sided oscillator, which restricts the frequency band overlap between the bone conduction oscillator and the air conduction loudspeaker; When the optimized dual-side oscillator control parameters trigger the frequency band overlap interval, the frequency band weights of the bone conduction and air conduction are redistributed.

8. The method for bilateral resonator adjustment of in-ear bone conduction auxiliary hearing aids according to claim 7, characterized in that, Also includes: The personalized acoustic transmission model is updated based on the user's head rotation angle data, which is collected by an inertial measurement unit. The optimized dual-sided oscillator control parameters are reset based on the updated personalized acoustic transmission model to adapt to changes in the sound source orientation.

9. An in-ear bone conduction-assisted hearing aid bilateral vibrator adjustment system, used to implement the in-ear bone conduction-assisted hearing aid bilateral vibrator adjustment method as described in any one of claims 1-8, characterized in that, include: The ear canal feature acquisition module is used to acquire structural feature data of both ear canals of the user; An acoustic modeling module is used to construct a personalized acoustic transmission model based on the structural feature data. The parameter generation module is used to generate the initial adjustment parameters of the dual-sided oscillator based on the personalized acoustic transmission model. The feedback monitoring module is used to monitor the user's bilateral auditory feedback signals in real time; The dynamic adjustment module is used to dynamically adjust the initial adjustment parameters based on the auditory feedback signal to generate optimized dual-sided oscillator control parameters.

10. An in-ear bone conduction hearing aid, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the bilateral vibrator adjustment method for in-ear bone air conduction auxiliary hearing aids as described in any one of claims 1 to 8.

Images