Self-adaptive bone conduction hearing aid device, system and method based on environment intelligent recognition

Abstract

A self-adaptive bone conduction hearing aid device, system and method based on environment intelligent identification, the hearing aid device comprises a connector beam, a right sounding device and a left sounding device, the right sounding device and the left sounding device are respectively installed on two sides of the connector beam, and hair clasp comb teeth are arranged on the connector beam. The hair clasp structure is arranged on the connecting head beam, so that the hair clasp structure can assist in fixing the hair clasp structure on the head of a user, and the hair clasp structure is stable, firm and attractive. According to the invention, the bone conduction vibrator and the pickup microphone are arranged in a split manner, and a double-microphone structure is adopted, so that the problem of mechanical coupling squeal can be solved, and intelligent recognition can be more effectively carried out on a scene. The architecture provided by the invention realizes the 1t; the total delay of 8ms is crucial to user perception authenticity and prevention of an echo effect. Through deep fusion of the architecture and the algorithm, full-link intelligent optimization from environment perception to sound output is realized, and psychological and physiological use barriers of a user are broken through hardware design.

Description

Technical Field

[0001] This invention discloses a hearing aid device, particularly an adaptive bone conduction hearing aid device, system, and method based on intelligent environmental recognition, belonging to the field of personal audio electronic product technology. Background Technology

[0002] Hearing aids are electronic assistive devices designed for people with hearing loss. Their core function is to improve auditory experience and enhance verbal communication abilities through precise sound processing and amplification. The core working principle of a hearing aid involves a microphone collecting ambient sound and converting it into an electrical signal. A digital signal processor then performs noise reduction, filtering, and gain adjustment on the electrical signal. An amplifier precisely amplifies the processed signal according to the degree of hearing loss. The receiver restores the electrical signal to an acoustic signal, which is then transmitted into the ear canal and finally to the brain for processing through the auditory system.

[0003] Currently, hearing aids generally come in two main forms: air conduction hearing aids and bone conduction hearing aids. Air conduction hearing aids (the most common type) conduct sound to the inner ear through the air and include behind-the-ear, in-the-canal, deep-in-the-canal, and body-worn types, suitable for the vast majority of people with hearing loss. Among them, behind-the-ear hearing aids have high power and are suitable for all types of hearing loss; in-the-canal / deep-in-the-canal hearing aids are highly discreet and suitable for mild to moderate hearing loss; body-worn hearing aids are cost-effective and easy to operate, making them suitable for the elderly. Bone conduction hearing aids conduct sound by vibrating the skull, without needing to be inserted into the ear canal. They are suitable for people with conductive / mixed hearing loss, sensitive or deformed ear canals, and have excellent high-frequency speech recognition performance in noisy environments.

[0004] Traditional hearing aids often suffer from unreasonable design structures, (1) acoustic feedback interference: since the bone conduction vibrator and the pickup microphone are in the same physical shell, mechanical coupling howling is easily induced when high gain compensation occurs; (2) wearing comfort and psychological barriers: the existing device structure is fixed and cannot adapt to different head shapes and pressures, and the appearance has a clear medical device feel, which causes users to have a "stigma" and problems such as not being comfortable to wear or falling off easily; (3) high learning cost: relying on smartphone applications for fitting, mode and volume adjustment poses a huge barrier for elderly users under the digital divide; (4) weak environmental adaptability: traditional devices lack localized intelligent scene recognition and cannot dynamically adjust compression parameters in noisy environments (such as restaurants and streets), resulting in poor speech clarity; (5) contradiction between latency and computing power: relying on cloud AI will lead to serious audio latency (echo), while local processing is often limited by DSP memory. Summary of the Invention

[0005] In view of the shortcomings of the existing hearing aids mentioned above, which have unreasonable design structures, the present invention provides an adaptive bone conduction hearing aid device, system and method based on intelligent environmental recognition. It has a headband structure on the connecting headband, which can help fix the device on the user's head, which is both stable and secure, and also aesthetically pleasing.

[0006] The technical solution adopted by the present invention to solve its technical problem is: an adaptive bone conduction hearing aid based on intelligent environmental recognition. The hearing aid includes a connecting headband, a right-side sound-emitting device and a left-side sound-emitting device. The right-side sound-emitting device and the left-side sound-emitting device are respectively installed on both sides of the connecting headband, and a hairband comb is provided on the connecting headband.

[0007] An adaptive bone conduction hearing aid system based on intelligent environmental recognition is disclosed. The system includes a hearing aid processing unit and a logic judgment module, which are connected via an I2C bus. The hearing aid processing unit is connected to two microphones, and the logic judgment module is connected to an encoder, a switch, an indicator light, and a bone conduction vibrator.

[0008] An adaptive bone conduction hearing aid method based on intelligent environmental recognition, the method comprising: (1) Pre-processing of the hearing aid, including: Microphone acquisition: Two microphones convert ambient sound into electrical signals, which are then amplified by the hearing aid processing unit. AFC: The hearing aid processing unit eliminates feedback through phase cancellation; 16-channel WDRC: Employs 16-channel high-precision compression to compensate for specific frequency losses; I2S transmission: The processed audio is transmitted as a digital stream to the logic judgment module; (2) The logic judgment module performs context processing, including: AI-based contextual noise reduction: After receiving the I2S signal, it uses its DSP to classify the environment and dynamically suppress background noise; Mixing: The audio received by the wireless Bluetooth chip from a mobile phone call or TV transmitter is combined with the audio from the hearing aid processing unit. Bone conduction oscillator drive: Based on the physical characteristics of the bone conduction oscillator, low-frequency enhancement and peak limiting are performed, and the output control signal drives the bone conduction oscillator to vibrate and produce sound.

[0009] The technical solution adopted by the present invention to solve its technical problem further includes: The connecting headband includes an elastic connecting piece and an outer sleeve. The outer sleeve is fitted over the outside of the elastic connecting piece. The hairband comb teeth are fixedly installed on the hairband connecting part. The hairband connecting part is arranged parallel to the elastic connecting piece. A strip groove is provided on the outer sleeve, and the hairband comb teeth extend from the strip groove.

[0010] The headband connection part adopts a magnetic or snap-on structure. Identification contacts are set on the headband connection part. After the Bluetooth DSP obtains the identification information of the ornament through the identification contacts, it automatically calls the corresponding acoustic preset parameter set from memory.

[0011] The right-side and left-side sound-emitting devices each include a front cover, a rear cover, a connecting part, a control board, a microphone, and a bone conduction sound-emitting mechanism. The front cover and the rear cover are fastened together to form a sound-emitting device housing. One end of the connecting part is installed inside the sound-emitting device housing, and the other end is installed together with the connecting head beam. The control board and the bone conduction sound-emitting mechanism are fixedly installed inside the sound-emitting device housing, and the bone conduction sound-emitting mechanism is electrically connected to the control board. The microphone is fixedly installed on the sound-emitting device housing and is electrically connected to the control board.

[0012] The top of the front cover is fixedly provided with an installation ring, the connecting part is inserted into the installation ring, and the connecting part is provided with one or more limiting parts.

[0013] The control panel is equipped with a volume regulator, and the front cover has a volume adjustment knob connected to the volume regulator.

[0014] The control board is equipped with a charging interface, and the housing of the sound-generating device is equipped with a switch. The switch is a slide switch, and corresponding switching devices are set at corresponding positions on the control board.

[0015] The bone conduction sound generation mechanism includes a fitting sleeve, a fitting sleeve support frame, a bone conduction vibrator fixing component, a bone conduction vibrator, a resonant cavity, a sealing cover, and a resonant cavity cover. The fitting sleeve support frame is fixedly embedded in the rear cover, the fitting sleeve is fixedly installed on the fitting sleeve support frame, the sealing cover is fixedly installed on the front end of the fitting sleeve support frame, the resonant cavity cover is fixedly installed on the rear end of the fitting sleeve support frame, the resonant cavity is fixedly disposed inside the fitting sleeve support frame, the bone conduction vibrator fixing component is fixedly installed on the resonant cavity, the bone conduction vibrator is disposed inside the resonant cavity, and the bone conduction vibrator is electrically connected to the control board through an electrode connection wire.

[0016] The beneficial effects of this invention are as follows: The invention incorporates a headband structure on the connecting headband, which helps to secure the device to the user's head, ensuring stability and a sleek appearance. The invention separates the bone conduction transducer and microphone, employing a dual-microphone structure, which not only solves the problem of mechanical coupling feedback but also enables more effective intelligent scene recognition. The architecture of this invention achieves a total latency of <8ms, which is crucial for user perception of realism and preventing the "echo effect." Through deep integration of architecture and algorithms, this invention achieves end-to-end intelligent optimization from environmental perception to sound output, and breaks down psychological and physiological barriers to user use through hardware design.

[0017] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0018] Figure 1 This is a three-dimensional structural diagram of the present invention.

[0019] Figure 2 This is a schematic diagram of the decomposed state structure of the present invention.

[0020] Figure 3 This is a three-dimensional structural diagram of the sound-generating device on the right side in this invention.

[0021] Figure 4 This is a schematic diagram of the exploded structure of the sound-generating device on the right side in this invention.

[0022] Figure 5 This is a two-dimensional structural diagram of the right-side sound-generating device in this invention from a second perspective.

[0023] Figure 6 This is a second-view exploded view structural diagram of the right-side sound-generating device in this invention.

[0024] Figure 7 This is a block diagram of the main circuit of the present invention.

[0025] Figure 8 This is a flowchart of scene recognition and judgment based on signal-to-noise ratio and spectral flatness in this invention.

[0026] In the diagram, 1-connecting headband, 11-elastic connecting piece, 12-outer cover, 121-strip groove, 13-hairband comb teeth, 14-hairband connecting part, 2-right side sound-generating device, 21-front cover, 211-mounting ring, 22-rear cover, 23-connecting part, 231-limiting part, 24-battery, 25-control board, 251-volume adjuster, 252-volume adjustment knob, 253-charging interface, 254-switch, 261-fitting sleeve, 262-fitting sleeve support frame, 263-bone conduction vibrator fixing part, 264-bone conduction vibrator, 265-resonance cavity, 266-sealing cover, 267-resonance cavity cover, 268-electrode connecting wire, 27-microphone, 3-left side sound-generating device. Detailed Implementation

[0027] This embodiment is a preferred embodiment of the present invention. All other embodiments that are the same as or similar to this embodiment in principle and basic structure are within the protection scope of the present invention.

[0028] Please refer to the appendix for details. Figure 1 To be continued Figure 6This invention mainly protects an adaptive bone conduction hearing aid based on intelligent environmental recognition, which mainly includes a connecting headband 1, a right-side sound-emitting device 2, and a left-side sound-emitting device 3. The right-side sound-emitting device 2 and the left-side sound-emitting device 3 are respectively installed on both sides of the connecting headband 1. The connecting headband 1 is provided with a hairband comb 13, which can help fix the invention on the user's head. On the one hand, it can make the use more stable and less likely to fall off. On the other hand, it can also be used as a hair accessory without losing its appearance.

[0029] In this embodiment, the connecting headband 1 mainly includes an elastic connecting piece 11 and an outer sleeve 12. The elastic connecting piece 11 is made of an elastic material, such as steel sheet or carbon fiber sheet. The outer sleeve 12 is fitted over the elastic connecting piece 11 and is made of a flexible material to prevent rigid contact from causing discomfort to the user, making it more comfortable to use. The headband comb teeth 13 are fixedly installed on the headband connecting part 14, which is arranged parallel to the elastic connecting piece 11. The outer sleeve 12 has a strip groove 121, from which the headband comb teeth 13 extend. Alternatively, the headband comb teeth 13 can be integrally formed with the elastic connecting piece 11, that is, the headband comb teeth 13 can be directly generated on the elastic connecting piece 11. In this embodiment, the headband connection part 14 adopts a magnetic structure, which can be attracted together with the elastic connecting piece 11, making it easy to replace. The headband connection part 14 can be equipped with identification contacts, which can be electrically connected to the control board 25. When different attributes (such as "outdoor type" or "elegant type") of the headband comb teeth 13 are connected, the system automatically switches the corresponding acoustic parameter theme through the identification circuit.

[0030] In this embodiment, the right-side sound-emitting device 2 and the left-side sound-emitting device 3 have the same structure, and are symmetrically installed at both ends of the connecting head beam 1. The following detailed description will take the right-side sound-emitting device 2 as an example.

[0031] In this embodiment, the right-side sound-generating device 2 mainly includes a front cover 21, a rear cover 22, a connecting part 23, a control board 25, a microphone 27, and a bone conduction sound-generating mechanism. The front cover 21 and the rear cover 22 are fastened together to form the outer shell of the sound-generating device. One end of the connecting part 23 is installed inside the outer shell of the sound-generating device, and the other end is installed together with the connecting head beam 1. The control board 25 and the bone conduction sound-generating mechanism are fixedly installed inside the outer shell of the sound-generating device, and the bone conduction sound-generating mechanism is electrically connected to the control board 25. The microphone 27 is fixedly installed on the outer shell of the sound-generating device and is electrically connected to the control board 25.

[0032] In this embodiment, a mounting ring 211 is fixedly provided on the top of the front cover 21, and the connecting part 23 is inserted into the mounting ring 211. The connecting part 23 is provided with one or more limiting parts 231, which can easily adjust the length of the connecting part 23 inserted into the connecting head beam 1. The connecting part 23 and the mounting ring 211 adopt a multi-directional ball joint, which supports adjustment in the XYZ axis direction. By changing the lever arm angle, precise control of skull pressure can be achieved.

[0033] In this embodiment, a battery 24 is fixedly installed inside the housing of the sound-generating device. The battery 24 can power the right-side sound-generating device 2. The battery 24 can be a lithium battery, but other types of batteries can also be used in practice. In this embodiment, batteries are respectively installed inside the right-side sound-generating device 2 and the left-side sound-generating device 3, and the two sound-generating devices are powered independently. In practice, a single battery can also be used to power both sides simultaneously.

[0034] In this embodiment, a volume regulator 251 is installed on the control board 25, and a volume adjustment knob 252 is provided on the front cover 2. The volume adjustment knob 252 is connected to the volume regulator 251. The volume regulator 251 can be a sliding rheostat or an encoder. The output volume is adjusted by rotating the volume regulator 251 through the volume adjustment knob 252.

[0035] In this embodiment, a charging interface 253 is installed on the control board 25 for charging the battery 24, and a switch 254 is installed on the housing of the sound-generating device. The switch 254 is a slide switch, and corresponding switching devices are set at corresponding positions on the control board 25. The power on / off operation can be realized through the switch 254.

[0036] In this embodiment, the bone conduction sound generation mechanism mainly includes a fitting sleeve 261, a fitting sleeve support frame 262, a bone conduction vibrator fixing component 263, a bone conduction vibrator 264, a resonance cavity 265, a sealing cover 266, and a resonance cavity cover 267. The fitting sleeve support frame 262 is fixedly embedded in the rear cover 22, serving a fixing and supporting function. The fitting sleeve 261 is fixedly installed on the fitting sleeve support frame 262 for conforming to the user's skin. The fitting sleeve 261 can be made of silicone material, allowing the user to... For greater comfort, the sealing cover 266 is fixedly installed at the front end of the fitting sleeve support frame 262 (i.e., the end near the fitting sleeve 261), the resonant cavity cover 267 is fixedly installed at the rear end of the fitting sleeve support frame 262, the resonant cavity 265 is fixedly installed inside the fitting sleeve support frame 262, the bone conduction oscillator fixing component 263 is fixedly installed on the resonant cavity 265, the bone conduction oscillator 264 is installed inside the resonant cavity 265, and the bone conduction oscillator 264 is electrically connected to the control board 25 through the electrode connection line 268.

[0037] In this embodiment, the microphone 27 is embedded in the front cover 21. The present invention uses a dual silicon microphone array to acquire the original ambient sound signal.

[0038] Please refer to the appendix for details. Figure 7This invention also protects an adaptive bone conduction hearing aid system based on intelligent environmental recognition. This acoustic processing system employs a dual-chip collaborative approach, connecting pre- and post-stage chips. The pre-stage is a hearing aid processing unit, which can be defined as a pre-stage DSP, used as a slave device primarily responsible for task execution. In this embodiment, the MA1180 hearing aid chip can be used, which has a built-in 16-channel WDRC, static feedback suppression, and EQ compensation module, responsible for basic hearing aid logic. The post-stage is a logic judgment module, which can be defined as a post-stage DSP or a Bluetooth DSP, used as a master device primarily responsible for perception and decision-making tasks. In this embodiment, a HiFi4 chip can be used. The DSP, integrated into the wireless SOC chip, runs the offline AI algorithm logic layer, responsible for feature extraction and context classification. The DSP also implements an anti-feedback algorithm combining a PPFB filter bank and dynamic frequency shifting. The front-end DSP performs hardware adaptive filtering (static AFC), while the Bluetooth DSP performs dynamic frequency shifting (DFS). This two-stage cascaded approach achieves anti-feedback. Two microphones are connected to the front-end hearing aid chip, while the rear-end wireless SOC chip connects to the encoder, switch, indicator lights, and bone conduction transducer. When the MA1180 hearing aid chip's internal algorithm detects severe and uncontrollable feedback, it immediately sends an interrupt signal to the Bluetooth chip, which then instantly performs a hard-mute at its output.The total link latency (End-to-End Latency) of the system of this invention is less than 8ms. To be precise, the slight fluctuation of algorithm overhead when switching between different environmental modes will usually control the total link latency between 5.5ms and 8ms. The specific allocation and composition of this total latency are as follows: (1) Hardware processing latency of the front-end DSP (MA1180) (about 1.5ms - 2.0ms). This part of the latency includes the time for microphone analog signal to digital signal (ADC), the hardware acceleration processing time of 16-channel WDRC (wide dynamic range compression), and the calculation time of internal static feedback suppression (AFC). Since the MA1180 uses dedicated hardware logic circuits (Hard-wired Logic) to execute the algorithm, its processing latency is extremely low and fixed; (2) Bus transmission latency between the two chips (about 0.5ms - 1.0ms). The audio stream processed by the front-end DSP is transmitted through I2S (Inter-IC). The digital audio bus synchronously transmits to the Bluetooth DSP. Due to the use of a high sampling rate (such as 48kHz) and synchronous transmission mechanism, the transmission delay at this physical layer is almost negligible, mainly due to the slight delay caused by buffering; (3) Bluetooth DSP (HiFi4 core) algorithm processing delay (about 3.0ms - 4.0ms). This part is the most complex part of the system logic, including the sub-band decomposition of the polyphase filter bank (PPFB), the calculation of offline AI environment recognition logic, and dynamic frequency shift (DFS) and adaptive noise reduction processing. The Bluetooth DSP adopts a "non-blocking" algorithm architecture and efficient PPFB filter, avoiding the long frame delay (Block Delay) caused by the traditional FFT algorithm. Thus, while ensuring the accuracy of AI recognition, the calculation time is compressed to within 4ms. Traditional hearing aids using FFT are prone to a delay of >10ms, while the PPFB solution is the core mathematical basis that can achieve <8ms; (4) Post-stage output and conversion delay (about 0.5ms - The total delay (1.0ms) includes the digital-to-analog (DAC) processing time after signal synthesis, as well as the response time for driving the bone conduction oscillator to emit mechanical vibrations. In actual operation, the total delay of each module is approximately between 5.5ms and 8.0ms, which is far below the echo threshold perceived by the human ear (typically 15ms-20ms), ensuring the real-time nature and authenticity of bone conduction auditory feedback.

[0039] The Bluetooth DSP can "reverse reconstruct" the 16-channel WDRC parameters of the front-end DSP via the I2C control bus, including: (1) Environmental feature extraction and perception: The Bluetooth DSP receives the digital audio stream processed by the front-end DSP in real time through the I2S bus. The Bluetooth DSP uses its integrated HiFi4 DSP computing power to perform time-frequency analysis through polyphase filter bank (PPFB), decompose the audio signal into multiple sub-bands, and calculate the spectral flatness (SFM), root mean square energy (RMS), energy distribution of the current environment, signal-to-noise ratio and zero crossover rate (ZCR) in real time to construct the statistical feature vector of the current acoustic environment.

[0040] (2) Offline AI logic performs dynamic decision-making and parameter mapping: The local running offline AI algorithm logic layer (including statistical models or lightweight neural network models) classifies the current environment into a preset situational category mode (e.g., quiet, speech, traffic, or noisy environment) based on the extracted feature vectors. Subsequently, the Bluetooth DSP retrieves the optimized gain compensation curve and 16-channel compression strategy that match the scene from its local memory.

[0041] (3) I2C Reverse Reconstruction Execution: The Bluetooth DSP, acting as the Master device, sends a register rewrite instruction set to the Slave device of the preceding DSP with logical address MA1180 via the I2C control bus. This instruction set directly rewrites the register values ​​of the 16-channel WDRC (Wide Dynamic Range Compression) controller inside the preceding DSP, dynamically modifying core parameters including the compression threshold, compression ratio, attack time, and release time of each channel.

[0042] (4) Adaptive processing closed-loop effect: Through this I2C reverse reconstruction mechanism, the 16-channel actuator of the front-end DSP can switch from "general hearing aid mode" to "specific scene optimization mode" in milliseconds based on the judgment result of the Bluetooth DSP. The compression behavior changes from "static fixed" to "dynamic adaptive". This architecture decouples perception (Bluetooth DSP) and execution (front-end DSP), which not only gives full play to the low latency processing advantage of the front-end DSP, but also gives the system the ability to adjust offline intelligently. Thus, it can perform millisecond-level fine reconstruction of parameters for complex acoustic scenes without increasing the computing power burden of the front-end DSP.

[0043] In this embodiment, the front-end and back-end are connected via an I2C bus, and the gain parameter set of the front-end chip can be reconstructed in real time based on the logic judgment result.

[0044] Please refer to the appendix for details. Figure 8 This invention also protects an adaptive bone conduction hearing aid method based on intelligent environmental recognition, which mainly includes the following steps: (1) The pre-processing of the preamplifier (i.e., MA1180) (which can be called: front-end) mainly includes: Microphone acquisition: Two microphones convert ambient sound into electrical signals, which are then amplified by the low-noise PGA inside the hearing aid chip (MA1180).

[0045] AFC (Adaptive Feedback Suppression): Since the mechanical vibration of the bone conduction oscillator is easily transmitted back to the microphone, the hearing aid chip (MA1180) eliminates feedback at this stage through phase cancellation.

[0046] 16-channel WDRC: Compared to traditional methods, this invention uses 16-channel high-precision compression, which can more finely divide the frequency (each segment has a narrower bandwidth) and provide extremely smooth gain compensation for the specific frequency loss of the elderly (such as high-frequency steep drop-off), avoiding the harshness caused by sudden changes in sound.

[0047] I2S transmission: The processed "pure hearing aid sound" is transmitted as a digital stream to the wireless Bluetooth chip on time with low latency.

[0048] (2) Scenario & Driver processing of wireless Bluetooth chips, which mainly includes: AI-powered contextual noise reduction: After receiving the I2S signal, it uses its DSP computing power to classify the environment (such as canteen, road, home) and dynamically suppress background noise.

[0049] Mixing: At this time, the audio received by the wireless Bluetooth chip from the mobile phone call or TV transmitter is seamlessly blended with the hearing aid audio sent by the hearing aid chip (MA1180).

[0050] Bone conduction oscillator drive: Based on the physical characteristics of bone conduction oscillators, low-frequency enhancement and peak limiting are performed to protect the oscillator from mechanical impact due to overload.

[0051] This invention features a headband structure on the connecting headband, which helps to secure the device to the user's head, ensuring stability and a sleek appearance. The invention separates the bone conduction transducer and microphone, employing a dual-microphone structure. This not only solves the problem of mechanical coupling feedback but also enables more effective intelligent scene recognition. The architecture achieves a total latency of <8ms, which is crucial for user perception of realism and preventing the "echo effect." Through deep integration of architecture and algorithms, this invention achieves end-to-end intelligent optimization from environmental perception to sound output, and breaks down psychological and physiological barriers to user experience through hardware design.

Claims

1. An adaptive bone conduction hearing aid based on intelligent environmental recognition, characterized in that: The hearing aid device includes a connecting headband (1), a right-side sound-emitting device (2) and a left-side sound-emitting device (3). The right-side sound-emitting device (2) and the left-side sound-emitting device (3) are respectively installed on both sides of the connecting headband (1). The connecting headband (1) is provided with hairband comb teeth (13).

2. The adaptive bone conduction hearing aid based on intelligent environmental recognition according to claim 1, characterized in that: The connecting headband (1) includes an elastic connecting piece (11) and an outer sleeve (12). The outer sleeve (12) is fitted on the outside of the elastic connecting piece (11). The hairband comb teeth (13) are fixedly installed on the hairband connecting part (14). The hairband connecting part (14) is arranged parallel to the elastic connecting piece (11). A strip groove (121) is provided on the outer sleeve (12), and the hairband comb teeth (13) extend from the strip groove (121).

3. The adaptive bone conduction hearing aid based on intelligent environmental recognition according to claim 2, characterized in that: The hairband connection part (14) adopts a magnetic or snap-on structure. Identification contacts are set on the hairband connection part (14). After the Bluetooth DSP obtains the identification information of the decorative piece through the identification contacts, it automatically calls the corresponding acoustic preset parameter set from the memory.

4. The adaptive bone conduction hearing aid based on intelligent environmental recognition according to claim 1, characterized in that: The right-side sound-generating device (2) and the left-side sound-generating device (3) respectively include a front cover (21), a rear cover (22), a connecting part (23), a control board (25), a microphone (27) and a bone conduction sound-generating mechanism. The front cover (21) and the rear cover (22) are fastened together to form the outer shell of the sound-generating device. One end of the connecting part (23) is installed inside the outer shell of the sound-generating device, and the other end is installed together with the connecting head beam (1). The control board (25) and the bone conduction sound-generating mechanism are fixedly installed inside the outer shell of the sound-generating device, and the bone conduction sound-generating mechanism is electrically connected to the control board (25). The microphone (27) is fixedly installed on the outer shell of the sound-generating device and is electrically connected to the control board (25).

5. The adaptive bone conduction hearing aid based on intelligent environmental recognition according to claim 4, characterized in that: The front cover (21) is fixedly provided with an installation ring (211) on the top, and the connecting part (23) is inserted into the installation ring (211). The connecting part (23) is provided with one or more limiting parts (231).

6. The adaptive bone conduction hearing aid based on intelligent environmental recognition according to claim 4, characterized in that: The control panel (25) is equipped with a volume regulator (251), and the front cover (2) is provided with a volume adjustment knob (252), which is connected to the volume regulator (251).

7. The adaptive bone conduction hearing aid based on intelligent environmental recognition according to claim 4, characterized in that: The control board (25) is equipped with a charging interface (253), and the housing of the sound-generating device is equipped with a switch (254). The switch (254) is a sliding switch, and the corresponding switching device is set at the corresponding position on the control board (25).

8. The adaptive bone conduction hearing aid based on intelligent environmental recognition according to claim 4, characterized in that: The bone conduction sound generation mechanism includes a fitting sleeve (261), a fitting sleeve support frame (262), a bone conduction vibrator fixing component (263), a bone conduction vibrator (264), a resonance cavity (265), a sealing cover (266), and a resonance cavity cover (267). The fitting sleeve support frame (262) is fixedly embedded in the rear cover (22), the fitting sleeve (261) is fixedly installed on the fitting sleeve support frame (262), and the sealing cover (266) is fixedly installed on the fitting sleeve support frame (265). The front end of the sleeve support frame (262) and the resonant cavity cover (267) are fixedly installed at the rear end of the sleeve support frame (262). The resonant cavity (265) is fixedly installed inside the sleeve support frame (262). The bone conduction oscillator fixing part (263) is fixedly installed on the resonant cavity (265). The bone conduction oscillator (264) is installed inside the resonant cavity (265). The bone conduction oscillator (264) is electrically connected to the control board (25) through the electrode connection line (268).

9. An adaptive bone conduction hearing aid system based on environmental intelligent recognition, disposed within any one of the adaptive bone conduction hearing aids based on environmental intelligent recognition as described in any one of claims 1 to 8, characterized in that: The system includes a hearing aid processing unit and a logic judgment module, which are connected via an I2C bus. The hearing aid processing unit is connected to two microphones, and the logic judgment module is connected to an encoder, a switch, an indicator light, and a bone conduction vibrator.

10. An adaptive bone conduction hearing aid method based on environmental intelligent recognition, implemented using an adaptive bone conduction hearing aid device based on environmental intelligent recognition as described in any one of claims 1 to 8, characterized in that: The method includes: (1) Pre-processing of the hearing aid, including: Microphone acquisition: Two microphones convert ambient sound into electrical signals, which are then amplified by the hearing aid processing unit. AFC: The hearing aid processing unit eliminates feedback through phase cancellation, and the pre-amplifier DSP performs hardware adaptive filtering to perform static AFC; 16-channel WDRC: Employs 16-channel high-precision compression to compensate for specific frequency losses; I2S transmission: The processed audio is transmitted as a digital stream to the logic judgment module; (2) The logic judgment module performs context processing, including: AI Contextual Noise Reduction: Upon receiving the I2S signal, the DSP performs environmental classification and dynamically suppresses background noise. The Bluetooth DSP performs dynamic frequency shifting. The Bluetooth DSP receives the digital audio stream processed by the front-end DSP in real time via the I2S bus. Utilizing its integrated HiFi4 DSP computing power, it performs time-frequency analysis through a polyphase filter bank to calculate the energy distribution, spectral flatness, and signal-to-noise ratio of the current environment. The offline algorithm logic layer classifies the current environment into a preset contextual mode based on the extracted features. Subsequently, the Bluetooth DSP retrieves the gain compensation curve and compression strategy matching the mode from its local memory. The Bluetooth DSP sends a specific instruction set to the front-end DSP with logic address MA1180 via the I2C control bus. This instruction set directly rewrites the register values ​​of the 16-channel WDRC controller inside the front-end DSP, including the compression threshold, compression ratio, and time constant of each channel. Mixing: The audio received by the wireless Bluetooth chip from a mobile phone call or TV transmitter is combined with the audio from the hearing aid processing unit. Bone conduction oscillator drive: Based on the physical characteristics of the bone conduction oscillator, low-frequency enhancement and peak limiting are performed, and the output control signal drives the bone conduction oscillator to vibrate and produce sound.

Images