Bone conduction signal self-adaptive enhancement system of acoustic-electro-mechanical sensor

Abstract

The invention relates to the technical field of electro-acoustic conversion and signal processing, and discloses a bone conduction signal adaptive enhancement system of an acoustic electro-mechanical sensor, which comprises a signal acquisition unit for synchronously acquiring voltage and current signals of a transducer driving loop; the impedance analysis unit is used for removing a static impedance component based on an equivalent circuit model and extracting a motional impedance component representing mechanical load characteristics; the self-adaptive equalization unit is used for determining damping characteristics according to the motional impedance component and adjusting the gain of the filter; and the displacement protection unit is used for estimating the displacement of the vibrating diaphragm of the transducer in real time according to the motional impedance component and preferentially executing dynamic range compression when the displacement is close to a physical limit. The problem of mechanical overload edge beating caused by gain compensation when bone conduction equipment is in a loose coupling low-damping state is solved, and the mechanical safety of a system is ensured while the sound quality is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to a bone conduction signal adaptive enhancement system of an electro-acoustic transducer, belonging to the technical field of electro-acoustic conversion and signal processing. BACKGROUND

[0002] The current bone conduction transducer belongs to an electro-acoustic device based on a contact conduction mechanism, which generates mechanical vibration through an electromechanical conversion component to drive the skull to realize audio signal transmission. The electro-acoustic response characteristics not only depend on the electromechanical parameters of the device itself, but also are highly coupled with the acoustic boundary conditions between the vibration panel and the human skin. The existing mainstream driving architecture usually regards the transducer as a constant equivalent circuit model load, adopts preset filter coefficients to cooperate with a constant voltage power amplifier open-loop driving, and the design logic presets the idealized and static state of the wearing state. In actual application scenarios, the wearer chews in the oral cavity, runs or jumps with vigorous limbs, or moves slightly in the wearing position, which causes the coupling stiffness and mechanical damping of the transducer and the skin contact interface to have millisecond-level nonlinear fluctuations. The dynamic time-varying nature of the acoustic coupling state destroys the original impedance matching state of the system, causes the radiation impedance to drift sharply and the system resonance frequency to change, and causes the output frequency response curve to be distorted and the listening consistency to be lost.

[0003] If the traditional improvement idea is followed, the thin film pressure sensor or contact microphone is added in the compact earphone cavity to monitor the wearing state, which occupies the precise stacking space of miniaturized devices, greatly increases the manufacturing cost and IPX level waterproof process difficulty, and it is difficult to solve the problem of sensor aging drift caused by long-term exposure to sweat environment; The conventional voltage limiting or fixed parameter equalization strategy cannot avoid the membrane displacement overload or compensate for the high-frequency energy loss in the loose coupling and low damping state due to the lack of real-time mechanical damping change sensing capability. The existing technology has the following disadvantages: lack of in-situ sensing capability of mechanical load based on electrical parameters of driving circuit, unable to obtain acoustic boundary condition characteristics under zero external sensor condition; Static signal processing logic cannot adapt to dynamic time-varying mechanical load, resulting in frequency response distortion and listening difference under complex working conditions; The voltage threshold protection mechanism cannot represent the real mechanical displacement affected by damping, and it is difficult to achieve adaptive balance of driving efficiency and mechanical safety. For example, the utility model patent with the authorization announcement number CN213030626U collects and analyzes the single soldier's electroencephalogram (EEG) signal to evaluate his cognitive state, intelligently matches the corresponding audio data, and uses bone conduction to implement auditory conditioning. This kind of system regards the bone conduction device only as an audio signal output carrier, and the core technology is concentrated in the upper application logic of audio content selection and intervention based on physiological signals. For the mechanical damping change, frequency response distortion and membrane mechanical overload of the bone conduction transducer itself under different wearing coupling states, there is no effective means to sense and compensate the mechanical load in-situ from the electrical driving circuit. When the wearing state changes from tight to loose, the system cannot guarantee that the sound signal is transmitted with consistent fidelity, and it is more likely to cause membrane displacement overrun and mechanical edge due to the low-frequency high-gain compensation demand, which limits the reliability of the application of such bone conduction system in dynamic complex environment.

[0004] Therefore, how to analyze the mechanical load characteristics only by using the driving circuit voltage and current signals, and to construct a closed-loop adaptive control strategy to real-time correct the driving signal, has become a technical problem to be solved by the present application. SUMMARY

[0005] To solve the problems raised in the background art, the technical solution of the present application is as follows: a bone conduction signal adaptive enhancement system of an electroacoustic transducer, comprising: A signal acquisition unit is connected to the driving circuit of the bone conduction transducer, and is used to synchronously acquire the port voltage signal and the circuit current signal of the transducer; An electromechanical parameter decoupling unit is used to process the port voltage signal and the circuit current signal based on a pre-set equivalent circuit model, remove the static impedance components generated by the voice coil direct current resistance and inductance, and extract the dynamic impedance components representing the load characteristics of the mechanical vibration system; An adaptive equalization unit is configured to determine a damping characteristic of a wearing coupling state according to a modulus variation of the motional impedance component, and generate a first gain adjustment coefficient for compensating for a frequency response loss according to the damping characteristic, and perform equalization processing on the audio signal; A displacement limiting unit is configured to prevent mechanical overload of the transducer caused by the first gain adjustment coefficient in a low damping state. The displacement limiting unit calculates a counter electromotive force according to the motional impedance component, performs time domain integration, obtains a real-time displacement estimation value of the diaphragm of the transducer, and constructs displacement limiting control logic: real-time monitoring of the real-time displacement estimation value, when the value exceeds a preset physical travel threshold, a second gain adjustment coefficient is generated, and dynamic range compression is performed on the audio signal processed by the adaptive equalization unit, and the amplitude of the transducer is limited within the linear mechanical travel range.

[0006] Preferably, the operation logic of the displacement limiting unit for performing time domain integration to obtain the real-time displacement estimation value conforms to the following relationship: wherein, is the real-time displacement estimation value, is the port voltage signal, is the loop current signal, is the direct current resistance of the voice coil, is the force factor constant of the transducer; the displacement limiting unit converts the electrical parameters into mechanical displacement parameters using the relationship, and uses the real-time displacement estimation value as an input variable of the displacement limiting control logic.

[0007] Preferably, the adaptive equalization unit includes a resonance frequency tracking module for compensating for mechanical resonance point drift caused by changes in wearing tightness; the resonance frequency tracking module scans the phase response of the motional impedance component in the low frequency band, and determines the frequency corresponding to the phase angle zero point as the real-time mechanical resonance frequency; the adaptive equalization unit is configured with at least one center frequency adjustable parameter filter, and the resonance frequency tracking module real-time aligns the center frequency of the parameter filter to the real-time mechanical resonance frequency.

[0008] Preferably, the resonance frequency tracking module is also used to perform quality factor adjustment logic; the quality factor adjustment logic determines the current mechanical damping ratio according to the amplitude of the motional impedance component at the real-time mechanical resonance frequency; when the mechanical damping ratio decreases, the resonance frequency tracking module increases the quality factor of the parameter filter to form a narrowband filtering characteristic; when the mechanical damping ratio increases, the resonance frequency tracking module reduces the quality factor of the parameter filter to form a wideband filtering characteristic.

[0009] Preferably, the electromechanical parameter decoupling unit comprises a temperature rise calibration module for eliminating the interference of voice coil temperature variation on the extraction of motional impedance; the temperature rise calibration module monitors the low-frequency variation trend of the real part of the instantaneous complex impedance calculated from the port voltage signal and the loop current signal, and determines the variation trend as the thermal resistance drift; the electromechanical parameter decoupling unit subtracts the thermal resistance drift from the instantaneous complex impedance when calculating the motional impedance component.

[0010] Preferably, the system further comprises a infrasonic pilot module for maintaining load detection when the amplitude of the driving signal is lower than a preset detection threshold; the infrasonic pilot module injects a single-frequency test signal with a frequency lower than the lower limit of human ear hearing into the transducer; the electromechanical parameter decoupling unit calculates the motional impedance component based on the response of the transducer to the single-frequency test signal, so that the adaptive equalization unit presets the filter parameters adapted to the current wearing state in the silent or weak signal state.

[0011] Preferably, the adaptive equalization unit is configured with a multi-band gain control strategy; when the modulus of the motional impedance component indicates that the coupling damping is lower than a preset damping threshold, the adaptive equalization unit increases the gain of the high-frequency band filter to compensate for the high-frequency attenuation, and increases the cutoff frequency of the low-frequency band filter to filter out invalid low-frequency components; when the modulus of the motional impedance component indicates that the coupling damping is higher than the preset damping threshold, the adaptive equalization unit maintains the high-frequency band gain and attenuates the gain of the resonance frequency band.

[0012] Preferably, the displacement limiting unit further comprises a predictive limiting logic; the predictive limiting logic calculates the first derivative of the real-time displacement estimate to obtain the vibration speed, and when the vibration speed indicates that the diaphragm is about to exceed the physical travel threshold, the active range compression is activated in advance before the real-time displacement estimate actually reaches the physical travel threshold.

[0013] Preferably, the electromechanical parameter decoupling unit extracts the motional impedance component by using quadrature demodulation logic; the electromechanical parameter decoupling unit performs multiplication operation and low-pass filtering on the port voltage signal and the loop current signal with reference sine signal and cosine signal respectively to obtain the real part and the imaginary part of the voltage and the current; the electromechanical parameter decoupling unit calculates the instantaneous complex impedance using the real part and the imaginary part, and separates out the motional impedance component by performing vector subtraction according to the equivalent circuit model.

[0014] Preferably, the system is integrated into the digital audio processing circuit of the bone conduction earphone; the signal acquisition unit comprises a sampling resistor and an analog-to-digital converter connected in series at the output end of the power amplifier; the electromechanical parameter decoupling unit, the adaptive equalization unit and the displacement limiting unit are algorithm modules running in the processor of the digital audio processing circuit; the system utilizes the computing resources of the digital audio processing circuit to realize closed-loop control of the bone conduction earphone without external physical sensors.

[0015] Compared with the prior art, the present application has the following advantages: 1. In the bone conduction signal of the acoustic-mechanical sensor, based on the in-situ sensing and response mechanism of acoustic boundary conditions according to motional impedance analysis, voltage and current signals are collected in real time in the drive circuit. The motional impedance components of the mechanical vibration system are separated and characterized using the equivalent circuit model of the transducer. A mechanical load feedback channel that depends only on electrical parameters is established. The bone conduction transducer is used as both an electroacoustic actuator and a mechanical sensor. The nonlinear fluctuation of the contact state between human skin and the transducer panel is mapped to the impedance spectrum characteristics that the drive circuit can identify. The controller dynamically reconstructs the coefficients of the front-end digital filter based on the characteristics. This allows the system to correct the distortion of the electro-mechanical-acoustic transfer function in real time when the tightness of the fit, the contact position, or the contact pressure changes randomly. The in-situ sensing method avoids the dependence of traditional acoustic compensation schemes on preset fixed curves. It ensures that the transducer always works in the optimal driving state that matches the current acoustic impedance under dynamically changing mechanical loads, and maintains frequency response flatness and auditory consistency in complex actual wearing environments.

[0016] 2. Based on the virtual displacement limitation and mechanical overload avoidance mechanism of back EMF integral, the system utilizes the back EMF information contained in the motional impedance to reconstruct the real-time mechanical displacement data of the transducer vibration system through circuit-level time-domain integral calculation. It constructs a voltage threshold-independent displacement-limited sidechain compression logic to solve the technical contradiction of bone conduction devices exceeding the physical limit in the pursuit of loudness compensation under loose coupling and low damping conditions. When the system recognizes the decrease in mechanical damping, it preemptively compresses the dynamic range of the drive signal based on the predicted real-time displacement value, strictly limiting the diaphragm motion stroke within the linear region. Based on the physical displacement protection strategy, it eliminates the risk of nonlinear distortion and component fatigue fracture caused by mechanical edge striking while ensuring sound pressure output, and achieves coordinated control of drive gain adjustment and mechanical safety boundary.

[0017] 3. A dual-timescale drift suppression mechanism with decoupled thermodynamic parameters introduces timescale separation logic for different physical effects during impedance analysis. This distinguishes between the slow DC resistance drift caused by voice coil temperature rise and the rapid mechanical impedance fluctuation caused by changes in wearing status. The low-frequency change trend of the real part of the complex impedance is locked and the thermal effect component is subtracted. This ensures that the feedback signal obtained by the adaptive control unit only reflects the changes in mechanical load characteristics. The parameter decoupling design cuts off the interference path of temperature changes on the acoustic compensation strategy, avoiding incorrect adjustment of audio equalization parameters when the voice coil heats up during long-term high-power operation. This ensures the continuous and reliable logic stability and compensation accuracy of the closed-loop control system during continuous operation. Attached Figure Description

[0018] Fig. 1 This is a flowchart of the closed-loop control logic for motional impedance feedback in this invention. Fig. 2 This is a comparison diagram of the frequency response characteristics of the present invention under different coupling states and after adaptive equalization; Fig. 3 This is a schematic diagram of the hardware architecture deployment and acoustic-electromechanical signal interaction principle of the system of the present invention. Detailed Implementation

[0019] The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0020] This invention provides an adaptive enhancement system for bone conduction signals of an acoustic-electromechanical sensor, comprising a precision sampling resistor and an analog-to-digital converter connected in series in the output circuit of a power amplifier. This architecture utilizes the electrical parameters of the drive circuit to achieve in-situ sensing of mechanical load, thereby forming a closed-loop control path. Logically, the system includes a signal acquisition unit, an electromechanical parameter decoupling unit, an adaptive equalization unit, and a displacement limiting unit. Each unit interacts with data via a register or memory data bus. The signal acquisition unit is connected to the transducer's drive circuit and synchronously acquires the transducer's port voltage signal at a sampling rate of not less than 48kHz through the series sampling resistor. With the loop current signal flowing through the sampling resistor To ensure impedance analysis accuracy, the phase delay between the voltage sampling channel and the current sampling channel is calibrated to the microsecond range. The electromechanical parameter decoupling unit is used to separate the motional impedance components characterizing the load characteristics of the mechanical vibration system from the electrical signals. This unit applies orthogonal demodulation logic to convert the acquired port voltage signals... With loop current signal Multiplying by the system-generated reference sine and cosine signals respectively, and filtering out harmonic components using a low-pass filter with a cutoff frequency below 20Hz, the real and imaginary parts of the voltage and current at the current moment are calculated. The internal temperature rise calibration module calculates the trend of the instantaneous complex impedance real part in the 0.1Hz to 1Hz frequency band, identifies this trend component as the thermal resistance drift caused by the voice coil temperature rise, and subtracts this drift in subsequent calculations. Based on this, the system uses a preset DC resistance including the voice coil. and voice coil inductor The equivalent circuit model of the transducer is obtained by subtracting the temperature-rise calibrated DC resistance from the total complex impedance. Components and the voice coil inductance The generated inductive reactance component Thus, the motional impedance component is extracted. The adaptive equalization unit is based on the motional impedance components. modulus Determine the current mechanical damping characteristics and dynamically adjust the audio filter parameters when the modulus... When the preset loose coupling threshold is exceeded, it indicates that the transducer is loosely connected to the skin, resulting in reduced mechanical damping. This unit increases the gain coefficient of the high-frequency shelf filter to compensate for high-frequency acoustic energy loss and simultaneously increases the cutoff frequency of the low-frequency high-pass filter to filter out invalid low-frequency components. When the modulus... When the value is below the preset tight coupling threshold, it indicates that the mechanical damping has increased. The unit maintains high-frequency gain and attenuates the gain in the resonant frequency band.

[0021] To compensate for the mechanical resonant frequency drift caused by changes in wearing position, the adaptive equalization unit is equipped with a resonant frequency tracking module and a parametric filter with an adjustable center frequency. This module scans the motional impedance components. The phase response in the 50Hz to 800Hz frequency band is used to determine the frequency corresponding to the zero-crossing point of the phase angle as the real-time mechanical resonance frequency. The system locks the center frequency of the parametric filter to the real-time mechanical resonant frequency in real time. The system adjusts the quality factor Q of the filter based on the impedance amplitude at that frequency: when the impedance amplitude indicates that the mechanical damping is lower than the preset damping threshold, the system increases the Q value to form narrowband filtering characteristics; when the mechanical damping is higher than the preset damping threshold, the system decreases the Q value to form wideband filtering characteristics. A displacement limiting unit is used to prevent mechanical overload under low-damping conditions. This unit utilizes the back electromotive force information contained in the motional impedance component to calculate the real-time displacement estimate of the transducer diaphragm using the following formula. : ,in, For port voltage signals, This is the loop current signal. The DC resistance of the voice coil. As the force factor constant of the transducer, the displacement limiting unit will also include the real-time displacement estimate. The first derivative of the velocity is used as the vibration velocity, and combined with the current displacement value, the future displacement trend is predicted. When the real-time displacement estimate is... When the predicted value exceeds the preset physical travel threshold, this unit takes priority over the gain adjustment of the adaptive equalization unit, generates an attenuation coefficient and activates the sidechain dynamic range compression logic to limit the amplitude of the drive signal and ensure that the transducer amplitude is within the linear mechanical travel range. The system also includes a subsonic pilot module, which is used to maintain load detection in a silent state. When the amplitude of the main audio signal is lower than the preset detection threshold of -60dBFS and lasts for a preset time, this module injects a single-frequency test signal with a frequency of 15Hz into the drive circuit. The electromechanical parameter decoupling unit calculates the motional impedance component based on the transducer's response to the single-frequency test signal, so that the adaptive equalization unit preloads filter parameters adapted to the current wearing state when there is no effective audio input.

[0022] Example 1: This example describes the response mechanism and technical effect of the bone conduction signal adaptive enhancement system of the above-mentioned acoustic-electromechanical sensor under non-steady-state and intense motion conditions. When the bone conduction device is in a high-intensity motion scenario, the contact interface between the transducer and the human skin undergoes periodic coupling loosening due to limb vibration, causing the mechanical damping ratio to fluctuate significantly in a short period of time. At the same time, the continuous high-power audio drive causes the voice coil temperature to rise. The resulting static resistance thermal drift and dynamic mechanical impedance changes are superimposed, forming a compound interference. Under this condition, the signal acquisition unit continuously and synchronously captures the port voltage signal of the drive circuit. With loop current signal The data stream is then fed into the electromechanical parameter decoupling unit. Addressing the engineering challenge of mixed thermal-mechanical parameter interference, the time-scale difference between the temperature rise effect and mechanical vibration in the frequency domain response is utilized to pinpoint the extremely low-frequency variation component below 1Hz in the real part of the instantaneous complex impedance. This component is identified as the thermal resistance drift caused by Joule heating and subtracted in real time. This step provides a thermally calibrated static resistance reference for subsequent processing. This eliminates baseline modulation interference from temperature fluctuations in mechanical load identification, ensuring that the system extracts only the motional impedance component that accurately characterizes the wearing coupling state. .

[0023] Based on this pure dynamic impedance component, the system resolves the technical contradiction between sound quality compensation and mechanical protection under low damping conditions. When vigorous exercise causes the wearer to loosen momentarily, the dynamic impedance modulus... A sharp increase indicates a decrease in mechanical damping, and the adaptive equalization unit then detects the mechanical resonant frequency. Frequency deviation caused by reduced boundary stiffness is addressed by the resonant frequency tracking module, which uses the drift trajectory of the motional impedance phase zero-crossing point to drive the parametric filter's center frequency to align in real time to the current frequency. The quality factor Q of the filter is increased based on the amplitude at that point; simultaneously, the displacement limiting unit utilizes the aforementioned thermally calibrated electrical parameters to perform time-domain integration. The voltage and current signals are converted into real-time displacement estimates of the transducer diaphragm. This unit uses the first derivative information of displacement to predict the motion trend of the diaphragm. When it is predicted that the diaphragm displacement is about to exceed the linear region of physical travel under low damping, this unit takes priority over the gain compensation logic of the adaptive equalization unit, activates the side chain dynamic range compression, and performs millisecond-level preventive attenuation of the drive signal. This mechanism uses the computing power of the circuit side to build a virtual mechanical limit fence, and prevents mechanical edge and component damage caused by loose coupling without relying on physical sensors, thereby achieving a dynamic balance between drive efficiency and system safety.

[0024] Example 2: This example constructs a performance verification experiment under non-ideal working conditions to quantify the robustness of the technical solution of this invention under dynamic load and thermal-mechanical combined interference. The experimental platform uses a SoundCheck audio analyzer as the core for signal excitation and acquisition. A precision sampling resistor is connected in series in the drive circuit to obtain the current signal in real time. To simulate the complex and variable mechanical boundary conditions in actual wear, a B&K4128C head and body simulator is introduced into the experiment. The contact pressure of the bone conduction transducer on the artificial ear part is dynamically adjusted by a servo pressure machine. In addition, a temperature control chamber is used to control the ambient temperature and combined with high-power continuous The experiment actively introduced voice coil thermal drift interference; two sets of comparative subjects were set up: the control group adopted the traditional constant gain open-loop drive scheme; the sample group of this invention integrates the above-mentioned adaptive enhanced closed-loop drive system. In order to comprehensively evaluate the system performance, the experiment designed gradient contact pressure test and limit displacement protection test. As the contact pressure decreased from 2.0N (tightly coupled reference) to 0.2N (loosely coupled limit), a sweep frequency signal of 20Hz to 20kHz was input to the two sets of transducers, and the resonant frequency shift and sound pressure level fluctuation at 1kHz were recorded in real time. Table 1 shows the key data of this gradient test.

[0025] Table 1: Comparison of Frequency Response Indicators under Contact Pressure Gradient Variation

[0026] Table 1 shows that as the contact pressure decreases, the mechanical resonance frequency of the control group shifts positively, with a maximum shift of 230Hz, resulting in a severe lack of low-frequency response; simultaneously, the sound pressure level at 1kHz decreases by 12.5dB, leading to a sharp drop in perceived loudness. In contrast, the sample group of this invention utilizes a resonance frequency tracking module to lock the mechanical resonance point in real time. The filtering parameters were aligned with the system, controlling the maximum frequency offset within 8Hz, and the sound pressure level fluctuation was only 1.2dB. This confirms that the system can effectively maintain frequency response consistency under a wide range of contact stiffness variations. Furthermore, the mechanical safety under loosely coupled, low-damping conditions was verified. With the contact pressure fixed at 0.2N and a 50Hz large dynamic pulse signal with an amplitude of 0dBFS input, the control group, without a displacement limiting mechanism, experienced a diaphragm displacement peak of 0.65mm, exceeding the physical limit of 0.5mm linear travel. The total harmonic distortion (THD) soared to 15.8%, accompanied by significant edge noise. The present invention's sample group, through a displacement limiting unit, calculates and predicts the displacement in real time. Upon detecting an overload trend, the sidechain dynamic range compression is activated first, precisely clamping the peak diaphragm displacement within a safe range of 0.48mm, controlling THD below 2.5%, and effectively avoiding mechanical overload. Finally, to verify the effectiveness of the temperature rise calibration module, the transducer was continuously operated at rated power for 1 hour, and the voice coil temperature decreased from 25°C. Rise to 65 Data records show that the control group experienced an error of approximately 3 dB in low-frequency equalization due to thermal drift of the resistance; while the sample group of this invention maintained the equalization parameter error within 0.2 dB by separating the thermal resistance component.

[0027] Example 3: This example combines Figs. 1 to 3 An adaptive enhancement system for bone conduction signals of an acoustic-electromechanical sensor is described, such as... Fig. 1 As shown, the system uses a bone conduction transducer to carry physical and mechanical loads and connects to the system through a drive circuit. The signal acquisition unit synchronously acquires port voltage and loop current signals to obtain electrical parameters. The data is transmitted to the electromechanical parameter decoupling unit, which removes static impedance based on the model and extracts the dynamic impedance component characterizing the mechanical load. This dynamic impedance component is split: on one hand, it is transmitted to the adaptive equalization unit to generate a first gain adjustment coefficient based on damping characteristics to compensate for frequency response loss; on the other hand, it is transmitted to the displacement limiting unit to estimate diaphragm displacement and perform dynamic range compression to prevent mechanical overload. Finally, the audio drive signal correction module integrates the gain adjustment coefficient and the displacement limiting control signal to superimpose equalization parameters and dynamic compression control on the drive signal. The resulting closed-loop correction signal is fed back to the bone conduction transducer, thereby completing the real-time response to changes in acoustic boundary conditions.

[0028] like Fig. 2 As shown in the figure, the x-axis represents frequency in Hz and the y-axis represents gain in dB, displaying a comparison of frequency response curves under three typical conditions: the solid line representing the tightly coupled state and the dashed line representing the loosely coupled state. It can be seen that in loose coupling, the low-frequency resonant peak is raised and attenuation occurs in the high-frequency range. The dotted line shows the frequency response after adaptive equalization, indicating that after processing by this system, the curve suppresses overshoot resonant energy at low frequencies and increases gain at high frequencies. Fig. 3 As shown, the system's deployment node is located within the bone conduction headphone terminal. The core control center is a digital audio processing chip (SoC / DSP). Within its execution environment, the core algorithm firmware layer integrates signal acquisition logic, electromechanical parameter decoupling logic, adaptive equalization controller, temperature rise calibration and thermal compensation, and displacement limitation protection logic. This chip interacts bidirectionally with the analog drive circuit. It acquires the loop current through a precision sampling resistor and inputs it to the chip along with voltage / current feedback signals. Simultaneously, the corrected drive signal output by the chip is amplified into a power signal by a power amplifier to drive the electroacoustic actuator of the bone conduction transducer, which serves as the physical node. During this process, the mechanical vibration / impedance load generated by the environmental node, namely the skin / skull load of the wearing part of the human body, directly acts on the transducer. The system senses the fluctuation of this physical load by monitoring the electrical changes in the drive loop, achieving closed-loop control without the need for external sensors.

[0029] Example 4: This example provides a standardized parameter calibration procedure and algorithm implementation path for the motional impedance characteristic mapping and adaptive filter parameter generation stages in the aforementioned bone conduction signal adaptive enhancement system. It addresses the parameter black box problem in the critical control chain from impedance measurement to filter coefficient update, specifically targeting the motional impedance magnitude. To establish the mapping relationship between the damping ratio and the coupling state, this system pre-constructs an impedance-damping lookup table based on large-scale offline test data. Built within a controlled acoustic laboratory environment, it utilizes a high-precision vibration table and force sensors to perform a full dynamic range impedance spectrum scan on a standard batch of bone conduction transducers. Specifically, the vibration table applies continuously varying mechanical damping loads from 0.1 N / m / s to 2.0 N / m / s to the transducers, with a step accuracy of no less than 0.05 N / m / s. Under each damping load level, the system synchronously records the transducer's impedance. The system responds and, after at least 100 repeated measurements and averaging, establishes an injective function relationship from physical damping to electrical impedance. During real-time operation, the system collects current data... The value is used to retrieve the corresponding instantaneous damping ratio from the lookup table using a bilinear interpolation algorithm. This value will serve as the core input variable for subsequent filter parameter calculations.

[0030] Secondly, regarding the core parameter of the adaptive filter, the center frequency... With quality factor The system dynamically adjusts according to the following determined transfer function reconstruction logic. To address the high-frequency attenuation problem caused by loose coupling, the system is defined based on the instantaneous damping ratio. The gain compensation function for the elevated filter is: ,in, The maximum compensation gain allowed by the system is typically set to 6dB to 9dB to prevent amplifier clipping. The reference damping ratio under standard tight coupling conditions is determined by the aforementioned offline calibration experiment, and is used when the real-time detected value is... At this time, the function outputs a positive gain value, which increases linearly or exponentially with decreasing damping ratio; simultaneously, to suppress the overshoot phenomenon of the resonance peak in the low-damped state, the system uses a second-order peak filter. The frequency band is subjected to notch filtering, its The value setting follows an inverse proportional adaptive rule: ,in, As the baseline quality factor, To prevent small positive numbers with a denominator of zero, this logic ensures that when the fit becomes loose, causing damping... When reduced, the filter The value automatically increases, forming a deep and narrow notch characteristic, precisely eliminating sharp resonance peak energy; conversely, in a tightly coupled, highly damped state, As the value decreases, the filter degenerates into a wideband fine-tuning mode, avoiding unnecessary spectral damage to the effective audio signal.

[0031] Example 5: To ensure the stability of the adaptive enhancement system of this invention under different batches of hardware, individual differences among wearers, and diverse acoustic environments, a standardized offline calibration and data filling procedure is performed when the system is first deployed or when major components are replaced. This is conducted in an acoustic testing chamber conforming to IEC60318-4 standards. A high-precision vibration table is used to apply a continuously linearly varying mechanical damping load from 0.1 N / m / s to 2.0 N / m / s to the bone conduction transducer, with a step accuracy set to 0.05 N / m / s. At each discrete damping load point, the system synchronously records the voltage and current response of the drive circuit, and obtains the corresponding dynamic impedance modulus through decoupling calculation. A reference mapping table of impedance modulus and physical damping is established based on this, and this offline-built lookup table serves as the core of the system's factory-preset firmware, eliminating the impedance characteristic dispersion caused by individual transducer manufacturing tolerances and ensuring the benchmark uniformity of wearing coupling state identification during online operation.

[0032] In addition, to address non-standardized wearing or unique anatomical differences that may occur in real-world usage scenarios, the system integrates a pre-deployment adaptive calibration procedure. When the device detects the first wear or a step change in wearing status that remains stable for more than 2 seconds, the calibration process is automatically triggered. The process injects a short-duration, weak broadband white noise excitation signal without the user's awareness, quickly scanning and locking the mechanical resonant frequency in the current wearing state. and dynamic impedance baseline The system uses this Using the zero point as a reference, the starting threshold and slope parameters of the gain compensation function in the adaptive equalization unit are dynamically adjusted so that subsequent gain adjustment actions are precisely adapted to the user's specific wearing characteristics. This on-site calibration mechanism effectively avoids the risk of overcompensation or undercompensation caused by individual user differences.

[0033] Example 6: This example describes a standardized engineering verification procedure for verifying the performance consistency and reliability of the above-mentioned adaptive enhancement system throughout its entire life cycle. It covers key stages from production line calibration to long-term aging testing. At the end of the production line, an initial parameter calibration procedure is performed. The assembled bone conduction headphones are placed in an automated test fixture equipped with a force feedback sensor. The standard contact pressure applied by the fixture is set to 1.5N. The system injects a logarithmic sweep signal with a frequency range of 20Hz to 20kHz and an amplitude of -20dBFS through the signal acquisition unit. Under this standard load, the port voltage and current responses are simultaneously acquired. The electromechanical parameter decoupling unit calculates the current motional impedance reference curve based on this response. And extract the peak frequency of the curve in the low-frequency resonance region. With peak impedance magnitude If the measured value deviates from the design specification center value by more than ±5%, the system will automatically correct the preset equivalent circuit model parameters. and The calculation continues until the error converges to within ±1%, and the final calibration parameters are stored in non-volatile memory as a static reference for subsequent adaptive calculations.

[0034] Next, the adaptive tracking capability verification procedure under dynamic load was executed. The test fixture was driven by a preset program to change the contact pressure between 0.5N and 2.0N in a sinusoidal manner at a frequency of 0.5Hz, simulating a slow human movement scenario. During this process, the system continuously input a 1kHz single-frequency test signal and output the current gain adjustment coefficient in real time. The verification system monitored the dynamic change trajectory of this coefficient and determined whether it remained synchronized with the phase change of the contact pressure, and whether the gain adjustment amplitude met the preset damping-gain mapping function requirements. If the gain response lag time exceeded 20ms or the adjustment amplitude error was greater than 1dB, the dynamic performance of the unit was deemed unqualified. Finally, the long-cycle aging and thermal stability verification procedure was performed, placing the earphone at 45°C. In a constant temperature environment, pink noise signals were continuously played at maximum power for 168 hours. During this period, the system automatically triggered a self-test process every hour, measuring the current static DC resistance by injecting a subsonic pilot signal. and compared with the initial calibration value The comparison and verification standards require the temperature rise calibration module to accurately identify the resistor drift.

[0035] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0036] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A bone conduction signal adaptive enhancement system for an acoustic electromechanical transducer, characterized by, The system comprises: a signal acquisition unit connected to the drive circuit of the bone conduction transducer, for synchronously acquiring the port voltage signal and the loop current signal of the transducer; an electromechanical parameter decoupling unit for processing the port voltage signal and the loop current signal based on a preset equivalent circuit model, removing the static impedance component generated by the direct current resistance and inductance of the voice coil, and extracting the motional impedance component representing the load characteristics of the mechanical vibration system; an adaptive equalization unit for determining the damping characteristics of the wearing coupling state according to the modulus variation of the motional impedance component, generating a first gain adjustment coefficient for compensating the frequency response loss according to the damping characteristics, and performing equalization processing on the audio signal; a displacement limiting unit for preventing mechanical overload of the transducer caused by the first gain adjustment coefficient in a low damping state; the displacement limiting unit calculates the back electromotive force based on the motional impedance component and performs time domain integration to obtain the real-time displacement estimation value of the diaphragm of the transducer, and constructs displacement limiting control logic: real-time monitoring of the real-time displacement estimation value, when the value exceeds the preset physical stroke threshold, a second gain adjustment coefficient is generated, and the audio signal processed by the adaptive equalization unit is executed dynamic range compression, limiting the amplitude of the transducer within the linear mechanical stroke range.

2. The bone conduction signal adaptive enhancement system of claim 1, wherein, The operation logic of the displacement limiting unit performing time domain integration to obtain the real-time displacement estimation value conforms to the following relationship: wherein, is the real-time displacement estimation value, is the port voltage signal, is the loop current signal, is the voice coil DC resistance, is the force factor constant of the transducer; the displacement limiting unit converts the electrical parameters into mechanical displacement parameters by using the relationship, and takes the real-time displacement estimation value as an input variable of the displacement limited control logic.

3. The bone conduction signal adaptive enhancement system of claim 1, wherein, The adaptive equalization unit comprises a resonance frequency tracking module for compensating for the mechanical resonance point drift caused by the tightness change of the wearing; the resonance frequency tracking module scans the phase response of the motional impedance component in the low frequency band, and determines the frequency corresponding to the zero crossing point of the phase angle as the real-time mechanical resonance frequency; the adaptive equalization unit is configured with at least one parametric filter with adjustable center frequency, and the resonance frequency tracking module aligns the center frequency of the parametric filter to the real-time mechanical resonance frequency in real time.

4. The bone conduction signal adaptive enhancement system of claim 3, wherein, The resonance frequency tracking module is also used to execute quality factor adjustment logic; the quality factor adjustment logic determines the current mechanical damping ratio according to the amplitude of the motional impedance component at the real-time mechanical resonance frequency; when the mechanical damping ratio decreases, the resonance frequency tracking module increases the quality factor of the parametric filter to form a narrowband filtering characteristic; when the mechanical damping ratio increases, the resonance frequency tracking module reduces the quality factor of the parametric filter to form a wideband filtering characteristic.

5. The bone conduction signal adaptive enhancement system of an acoustic electromechanical transducer according to claim 1, wherein, The electromechanical parameter decoupling unit comprises a temperature rise calibration module for eliminating the interference of voice coil temperature change on the extraction of motional impedance; the temperature rise calibration module monitors the low-frequency variation trend of the real part of the instantaneous complex impedance calculated from the port voltage signal and the loop current signal, and determines the variation trend as the thermal resistance drift; The electromechanical parameter decoupling unit subtracts the thermal resistance drift from the instantaneous complex impedance when calculating the motional impedance component.

6. The bone conduction signal adaptive enhancement system of claim 1, wherein, The system further comprises a infrasonic frequency pilot module for maintaining load detection when the amplitude of the drive signal is lower than the preset detection threshold; the infrasonic frequency pilot module injects a single-frequency test signal with a frequency lower than the lower limit of human hearing into the transducer; the electromechanical parameter decoupling unit calculates the motional impedance component based on the response of the transducer to the single-frequency test signal, so that the adaptive equalization unit presets the filter parameters adapted to the current wearing state in the silent or weak signal state.

7. The adaptive bone conduction signal enhancement system of claim 1, wherein, The adaptive equalization unit is configured with a multi-band gain control strategy; when the modulus of the motional impedance component indicates that the coupling damping is lower than a preset damping threshold, the adaptive equalization unit increases the gain of the high-frequency filter to compensate for high-frequency attenuation, and increases the cutoff frequency of the low-frequency filter to filter out invalid low-frequency components; when the modulus of the motional impedance component indicates that the coupling damping is higher than the preset damping threshold, the adaptive equalization unit maintains the high-frequency gain and attenuates the gain of the resonance frequency band.

8. The bone conduction signal adaptive enhancement system of claim 1, wherein, The displacement limiting unit further includes prediction clipping logic; the prediction clipping logic calculates the first derivative of the real-time displacement estimate to obtain a vibration speed, and when the vibration speed indicates that the diaphragm is about to exceed the physical travel threshold, the active range compression is activated in advance before the real-time displacement estimate actually reaches the physical travel threshold.

9. The bone conduction signal adaptive enhancement system of claim 1, wherein, The electromechanical parameter decoupling unit extracts the motional impedance component by using quadrature demodulation logic; the electromechanical parameter decoupling unit performs multiplication operation and low-pass filtering on the port voltage signal and the loop current signal with reference sine and cosine signals respectively to obtain the real and imaginary parts of the voltage and current; the electromechanical parameter decoupling unit calculates the instantaneous complex impedance by using the real and imaginary parts, and separates out the motional impedance component by performing vector subtraction according to the equivalent circuit model.

10. The bone conduction signal adaptive enhancement system of claim 1, wherein, The system is integrated in the digital audio processing circuit of the bone conduction earphone; the signal acquisition unit includes a sampling resistor and an analog-to-digital converter connected in series at the output end of the power amplifier; the electromechanical parameter decoupling unit, the adaptive equalization unit and the displacement limiting unit are algorithm modules running in the processor of the digital audio processing circuit.

Images