AI Bluetooth earphone sound leakage prevention control method and system based on big data

Abstract

The invention belongs to the technical field of electro-acoustic conversion, and provides an AI Bluetooth headset sound leakage prevention control method and system based on big data, and the method comprises the steps: carrying out the active sound leakage prevention test of a plurality of Bluetooth headsets of the same model, determining the sound leakage prevention inhibition rate of the Bluetooth headsets with different electric quantities through the monitoring of the electric quantity attenuation process of the Bluetooth headsets, and determining the sound leakage prevention inhibition rate of the Bluetooth headsets with different electric quantities. Judging whether the sound leakage prevention test of the Bluetooth headset is qualified or not; if the test is not qualified, extracting Bluetooth earphones with unqualified sound leakage prevention tests, screening and analyzing the Bluetooth earphones, determining unqualified sound leakage prevention earphones, performing unqualified proportion analysis on unqualified sound leakage prevention syllables of the unqualified sound leakage prevention earphones, and identifying remarkable unqualified nodes in the unqualified sound leakage prevention syllables; according to the invention, the method effectively compensates for the attenuation of the sound leakage prevention inhibition rate in the case of low electric quantity, avoids the sudden enhancement of the sound leakage, avoids the audio distortion problem caused by excessive compensation, and reduces the attenuation risk of the sound leakage prevention effect in the scene of low electric quantity.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of electro-acoustic conversion, and in particular to an AI Bluetooth earphone leakage sound control method and system based on big data. BACKGROUND

[0002] With the rapid development of wireless audio technology, Bluetooth earphones have been widely used in daily commuting, office work, sports and other multi-scene scenarios due to their portability and wireless advantages. The leakage sound performance, as a core indicator affecting user privacy protection and auditory experience, has attracted increasing market attention in terms of stability and full power scenario adaptability.

[0003] However, the existing Bluetooth earphone leakage sound control method lacks precise monitoring and quantitative analysis of the multi-node in the full power attenuation range, which cannot achieve quality screening and consistency control of batch products on the production side, nor can it avoid the sudden increase in leakage sound caused by power changes in user full power use scenarios. At the same time, the existing technology cannot accurately identify the core weak link of the leakage sound control function and the suppression weak power range of rapid performance decay, resulting in a lack of targeted data support for subsequent optimization. In addition, the quantitative correlation between the gain value and the leakage sound suppression rate is not established, and a reasonable upper limit constraint for the gain value is not set, which may result in insufficient compensation to alleviate low power leakage sound decay, or excessive compensation leading to audio distortion, and the stability, reliability and adaptability of the overall leakage sound control are insufficient.

[0004] Therefore, the present application provides an AI Bluetooth earphone leakage sound control method and system based on big data. SUMMARY

[0005] In order to make up for the deficiencies of the prior art and solve at least one technical problem raised in the background art.

[0006] In a first aspect, the present application provides an AI Bluetooth earphone leakage sound control method based on big data, comprising: actively testing a plurality of Bluetooth earphones of the same model for leakage sound control, determining the leakage sound suppression rate of Bluetooth earphones at different power levels by monitoring the power decay process of the Bluetooth earphones, and determining whether the leakage sound test of the Bluetooth earphones is qualified; If the test is not qualified, extract the Bluetooth earphones that fail the leakage sound test and perform screening analysis to determine the unqualified leakage sound earphones, perform unqualified proportion analysis on the unqualified leakage sound nodes of the unqualified leakage sound earphones, and identify the significant unqualified nodes in the unqualified leakage sound nodes; Based on the significant unqualified nodes in the leakage sound nodes, analyze the rate of change of the leakage sound suppression rate of the significant unqualified nodes with the remaining power of the earphones, and determine the suppression weak power range of the leakage sound performance; When the Bluetooth headset power attenuates to the weak power interval of the suppression, based on the dynamic gain compensation algorithm, the correlation between the gain value and the leakage sound suppression rate is combined, the leakage sound attenuation value is adjusted by gradient compensation of the gain value, and sudden enhancement of the leakage sound at low power is avoided.

[0007] Preferably, the determination process of the leakage sound suppression rate of the different power Bluetooth headset is: The power attenuation process of the Bluetooth headset from full power to low power is monitored in real time, the interval between the power attenuation process of the Bluetooth headset from full power to low power is recorded as a power analysis interval, and the power analysis interval is evenly divided into multiple power test nodes according to the power attenuation gradient; The microphone at the front end of the sound level meter is used as a sensor to convert the air pressure fluctuation caused by the leakage sound into an electrical signal, record the leakage sound pressure level of each power test node, set a reference leakage sound pressure level, and perform difference operation on the reference leakage sound pressure level and the leakage sound pressure level of the power test node to obtain the leakage sound pressure level change value of the power test node; If the leakage sound pressure level change value of the power test node is greater than 0, the corresponding power test node is recorded as an effective leakage sound prevention node; otherwise, the corresponding power test node is recorded as an invalid leakage sound prevention node; The leakage sound pressure level change value of the effective leakage sound prevention node is processed by ratio to obtain the leakage sound suppression rate of the effective leakage sound prevention node.

[0008] Preferably, the process of judging whether the leakage sound test of the Bluetooth headset is qualified is: If the leakage sound suppression rate of the effective leakage sound prevention node is less than the leakage sound suppression rate standard value, the corresponding effective leakage sound prevention node is recorded as an unqualified leakage sound prevention node; The number of qualified leakage sound prevention nodes in all power test nodes is calculated, and the leakage sound performance value is recorded; If the leakage sound performance value is greater than or equal to the leakage sound performance threshold value, it indicates that the leakage sound test of the Bluetooth headset is qualified; otherwise, it indicates that the leakage sound test of the Bluetooth headset is unqualified.

[0009] Preferably, the determination process of the unqualified leakage sound prevention earphone is: The Bluetooth headset with unqualified leakage sound test is recorded as a test unqualified earphone, and based on any test unqualified earphone, the invalid leakage sound prevention node number ratio of the test unqualified earphone in the power analysis interval is calculated, and the fault representation value is recorded; If the fault representation value is less than the fault representation standard value, the corresponding test unqualified earphone is recorded as an unqualified leakage sound prevention earphone.

[0010] Preferably, the process of identifying the significant unqualified node in the unqualified leakage sound prevention node is: extracting all unqualified leakage sound nodes of all unqualified leakage sound earphones, and integrating into an unqualified leakage sound node sequence according to the order of the electric quantity attenuation; Based on any unqualified leakage sound node, the number of unqualified leakage sound nodes in all unqualified leakage sound earphones is calculated, and the unqualified coefficient of the unqualified leakage sound node is recorded. If the unqualified coefficient of the test unqualified leakage sound node is greater than or equal to the unqualified coefficient threshold, the corresponding unqualified leakage sound node is recorded as a significant unqualified node.

[0011] Preferably, the weak inhibition electric quantity interval determination process of the leakage sound performance is: Integrating all significant unqualified nodes into a significant unqualified node sequence according to the order of the electric quantity attenuation, taking two adjacent significant unqualified nodes in the significant unqualified node sequence as a node analysis window, determining the falling window in the node analysis window, taking the absolute value of the inhibition rate change value of the falling window to obtain the inhibition rate decay value, and performing ratio calculation on the inhibition rate decay value and the electric quantity change corresponding to the falling window to obtain the change rate value; The interval corresponding to the electric quantity change is recorded as the electric quantity interval. If the change rate value is greater than the change rate threshold, the corresponding electric quantity interval is recorded as the weak inhibition electric quantity interval.

[0012] Preferably, the determination process of the falling window is: Based on any node analysis window, the residual electric quantity of all significant unqualified nodes is obtained, the residual electric quantity of the latter significant unqualified node in the node analysis window is subtracted from the residual electric quantity of the former significant unqualified node to obtain the electric quantity change, and the leakage sound inhibition rate of the latter significant unqualified node in the node analysis window is subtracted from the leakage sound inhibition rate of the former significant unqualified node to obtain the inhibition rate change value. If the inhibition rate change value is less than 0, it indicates that the leakage sound inhibition rate in the corresponding node analysis window shows a downward trend, and the corresponding node analysis window is recorded as a falling window.

[0013] Preferably, the process of gradient compensation for the gain value is: Divide the weak inhibition electric quantity interval by a fixed electric quantity to obtain gradient test nodes, and quickly retrieve the leakage sound inhibition rate and natural gain data of each gradient test node through a big data platform; Taking the absolute value of the difference between the leakage sound inhibition rates of adjacent gradient test nodes, the gradient inhibition rate decay value is obtained. Integrate all gradient inhibition rate attenuation values into a gradient inhibition rate attenuation sequence according to the order of power attenuation, and the AI reinforcement learning model autonomously learns the balance rule between gain compensation and audio distortion based on historical compensation data, optimizes the correlation between the gain value and the anti-leakage sound inhibition rate in real time, processes the ratio of the first gradient inhibition rate attenuation value in the gradient inhibition rate attenuation sequence and the correlation to obtain a compensation gain value, and processes the compensation gain value and the natural gain value to obtain a compensated gain value. Compare and analyze the compensated gain value and the preset gain value upper limit to generate a continue compensation signal. Based on the continue compensation signal, record the compensation gain value as a pre-compensation gain value, determine a new compensation gain value according to the power attenuation in the weak power interval and according to the gradient inhibition rate attenuation sequence, process the difference between the new compensation gain value and the pre-compensation gain value to obtain a gradient compensation value, and perform gradient compensation on the compensated gain value based on the gradient compensation value and compare the compensated gain value with the preset gain value upper limit until the gradient inhibition rate attenuation sequence ends or the gain compensation stops.

[0014] Preferably, the generation process of the continue compensation signal is as follows: If the compensated gain value is less than the preset gain value upper limit, compensation is continued, and a continue compensation signal is generated.

[0015] In the second aspect, the application further provides an AI Bluetooth earphone anti-leakage sound control system based on big data, which comprises: An anti-leakage sound test analysis module: actively test the anti-leakage sound of multiple Bluetooth earphones of the same type, monitor the power attenuation process of the Bluetooth earphones, determine the anti-leakage sound inhibition rate of Bluetooth earphones with different power, and judge whether the anti-leakage sound test of the Bluetooth earphones is qualified. A significant unqualified identification module: if the test is unqualified, extract the Bluetooth earphones with unqualified anti-leakage sound and perform screening analysis to determine the anti-leakage sound unqualified earphones, perform unqualified proportion analysis on the unqualified anti-leakage sound nodes of the anti-leakage sound unqualified earphones, and identify the significant unqualified nodes in the unqualified anti-leakage sound nodes. An interval determination module: based on the significant unqualified nodes in the anti-leakage sound nodes, analyze the rate of change of the anti-leakage sound inhibition rate of the significant unqualified nodes with the remaining power of the earphones, and determine the inhibition weak power interval of the anti-leakage sound performance. A gain compensation module: when the power of the Bluetooth earphones attenuates to the inhibition weak power interval, based on a dynamic gain compensation algorithm, combining the correlation between the gain value and the anti-leakage sound inhibition rate, adjust the anti-leakage sound attenuation value by gradient compensation of the gain value to avoid sudden increase of the leakage sound at low power.

[0016] The application has the following advantages: The application guarantees that users obtain stable and qualified anti-leakage sound experience in full power use scenarios through precise multi-node testing of the full power attenuation range of Bluetooth earphones, avoids sudden increase of leakage sound caused by power change, and provides reliable basis for subsequent identification of weak anti-leakage sound suppression range and optimization of anti-leakage sound strategy; for unqualified earphones, the significantly unqualified nodes are accurately identified by integrating unqualified anti-leakage sound node sequences and calculating unqualified coefficients, and the core weak link of the anti-leakage sound function is effectively located, which not only takes into account the quality control needs of the production end and the actual use experience of users, but also provides targeted support for subsequent anti-leakage sound strategy optimization and product performance improvement.

[0017] Based on adjacent window analysis of the significantly unqualified node sequence, the application accurately locates the weak power range of the anti-leakage sound suppression rate by calculating the rate of change of the anti-leakage sound suppression rate with power, provides a clear range of action and data support for subsequent targeted compensation, and under the constraint of the upper limit of the preset gain value, gradient gain compensation is performed on the weak range by combining the gain value and the anti-leakage sound suppression rate correlation coefficient determined by the experiment, which effectively makes up for the attenuation of the anti-leakage sound suppression rate at low power, avoids sudden increase of leakage sound, and at the same time avoids the problem of audio distortion caused by excessive compensation, reducing the risk of anti-leakage sound effect attenuation in low power scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0018] The application will be further described below with reference to the accompanying drawings.

[0019] Figure 1 is the acquisition step flowchart of the AI Bluetooth earphone anti-leakage sound control method based on big data of the embodiment of the application; Figure 2 is the system block diagram in the AI Bluetooth earphone anti-leakage sound control system based on big data of the embodiment of the application. DETAILED DESCRIPTION

[0020] In order to make the technical means, creative features, purposes and effects realized by the application easy to understand, the application will be further described below with reference to the specific embodiments. Embodiment 1

[0021] Please refer to Figure 1 The AI Bluetooth earphone anti-leakage sound control method based on big data of the embodiment of the application includes the following steps: Step one: actively test the anti-leakage sound of multiple Bluetooth earphones of the same type, monitor the power attenuation process of the Bluetooth earphones, determine the anti-leakage sound suppression rate of Bluetooth earphones at different power, and judge whether the anti-leakage sound test of the Bluetooth earphones is qualified; In the anechoic chamber, the full-power earphones are fixed on the artificial head model, the earphone wearing angle and pressure are checked to ensure consistent fit; Real-time monitoring of the Bluetooth headset from full power to low power (until the automatic shutdown) power attenuation process, the Bluetooth headset from full power to low power between the interval of power attenuation process is recorded as the power analysis interval, the power analysis interval is evenly divided into multiple power test nodes according to the power attenuation gradient, the earphone power is sequentially reduced to the power test node by the controllable discharge device, and the earphone power is sequentially reduced to the power test node by the controllable discharge device. Rest for 10 minutes to ensure stable power and avoid the influence of discharge heating on performance; Turn on the active anti-leakage sound function, play 1kHz pure tone, and use the microphone in front of the sound level meter as a sensor to convert the air pressure fluctuation caused by the leakage sound into an electrical signal. Record the leakage sound pressure level of each power test node at 1m; Set the reference leakage sound pressure level, and perform difference operation on the reference leakage sound pressure level and the leakage sound pressure level of the power test node to obtain the leakage sound pressure level change value of the power test node; It should be noted that the reference leakage sound pressure level is the leakage sound pressure level collected by the Bluetooth test sound when the active anti-leakage sound function is turned off. The leakage sound pressure level is a direct quantitative indicator for judging the degree of leakage. The higher the leakage sound pressure level value, the more obvious the leakage, and the worse the suppression effect of the anti-leakage sound function. Conversely, the leakage is less obvious, and the anti-leakage effect is better; If the leakage sound pressure level change value of the power test node is greater than 0, it means that the anti-leakage of the power test node is effective, and the corresponding power test node is recorded as an effective anti-leakage node; If the leakage sound pressure level change value of the power test node is less than or equal to 0, it means that the anti-leakage of the power test node is ineffective, and the corresponding power test node is recorded as an ineffective anti-leakage node; Based on the effective anti-leakage node, the leakage sound pressure level change value of the effective anti-leakage node is processed by ratio with the reference leakage sound pressure level to obtain the anti-leakage suppression rate of the effective anti-leakage node; The anti-leakage suppression rate standard value is set by the person skilled in the art according to historical experience. If the anti-leakage suppression rate of the effective anti-leakage node is less than the anti-leakage suppression rate standard value, it means that the anti-leakage effect of the effective anti-leakage node is unqualified, and the corresponding effective anti-leakage node is recorded as an unqualified anti-leakage node. If the anti-leakage suppression rate of the effective anti-leakage node is greater than or equal to the anti-leakage suppression rate standard value, it means that the anti-leakage effect of the effective anti-leakage node is qualified, and the corresponding effective anti-leakage node is recorded as a qualified anti-leakage node; Calculate the number of qualified anti-leakage nodes in all power test nodes, and record it as the anti-leakage performance value; If the anti-leakage performance value is greater than or equal to the anti-leakage performance threshold value, it means that the anti-leakage test of the Bluetooth headset is qualified; If the anti-leakage performance value is less than the anti-leakage performance threshold value, it means that the anti-leakage test of the Bluetooth headset is unqualified; It should be noted that the noise leakage performance value reflects the stability and reliability of the headphone's noise leakage prevention function under different battery levels. It intuitively presents the headphone's ability to consistently meet noise leakage prevention standards from full charge to low battery, providing a basis for subsequent identification of weak suppression areas and optimization of noise leakage prevention strategies. At the same time, the noise leakage performance value transforms the abstract noise leakage prevention performance into specific percentage data, which is convenient for the production end to conduct quality screening and consistency control of batch products, and can also ensure that users have a stable and qualified noise leakage prevention experience in full battery usage scenarios, avoiding the problem of sudden increase in noise leakage due to changes in battery level. Step 2: If the test fails, extract the Bluetooth headphones that fail the sound leakage test and perform screening and analysis to identify the headphones that fail the sound leakage test. Analyze the failure rate of the failure points of the headphones that fail the sound leakage test to identify the significant failure points among the failure points. Bluetooth headsets that fail the sound leakage prevention test are recorded as unqualified headsets. Based on any unqualified headset, the percentage of invalid sound leakage prevention nodes in the power analysis interval of the unqualified headset is calculated and recorded as the fault characterization value. If the fault characterization value is greater than or equal to the fault characterization standard value, the corresponding unqualified earphone will be recorded as a faulty earphone. If the fault characterization value is less than the fault characterization standard value, the corresponding unqualified headphones will be recorded as unqualified headphones for sound leakage prevention. It should be noted that the fault characterization standard values ​​were set by those skilled in the art based on historical experience; Extract all the faulty sound leakage nodes from the headphones that fail to meet the sound leakage prevention standards, and integrate them into a faulty sound leakage node sequence according to the order of power decay. It should be noted that an invalid anti-leakage node indicates that the anti-leakage circuit has failed and requires hardware repair or replacement. This falls under the category of faults and is not within the scope of normal performance optimization. A non-compliant anti-leakage node indicates that the anti-leakage system is basically working, but the effect is not up to standard. The sequence of non-compliant nodes directly reflects the change law of the anti-leakage suppression rate with power decay. Therefore, we will analyze the non-compliant anti-leakage nodes. Based on any unqualified sound leakage prevention node, calculate the percentage of times the unqualified sound leakage prevention node appears in all unqualified sound leakage prevention headphones, and record it as the unqualified coefficient of the unqualified sound leakage prevention node. If the failure coefficient of a failed sound leakage prevention node is greater than or equal to the failure coefficient threshold, the corresponding failed sound leakage prevention node will be recorded as a significantly failed node. If the failure coefficient of a failed sound leakage prevention node is less than the failure coefficient threshold, the corresponding failed sound leakage prevention node is recorded as a non-significantly failed node. It should be noted that by limiting the statistical sample to headphones that fail to prevent sound leakage, the non-compliance coefficient avoids interference from qualified products in the analysis results, ensuring that the data can truly reflect the common failure of non-compliance products in preventing sound leakage at specific power levels. At the same time, by specifying the frequency of occurrence of each non-compliance point in the form of quantitative proportion, it is helpful to assist the production end in identifying batch quality defects and improve the consistency and stability of the product's sound leakage performance. The technical solution of this invention is as follows: Multiple Bluetooth earphones of the same model undergo active sound leakage prevention testing. By monitoring the battery decay process of the Bluetooth earphones, the sound leakage suppression rate of Bluetooth earphones with different battery levels is determined, and the pass / fail status of the sound leakage prevention test is judged. If the test fails, the Bluetooth earphones that fail the sound leakage prevention test are extracted and screened for analysis to identify the earphones that fail the sound leakage prevention test. The failure rate of the failure points in the failure points of the earphones is analyzed to identify the significant failure points among the failure points. This invention achieves precise multi-node testing of Bluetooth earphones across the entire battery decay range, thus realizing batch processing at the production end. This process ensures quality screening and consistency control of mass-produced products, guaranteeing users a stable and qualified sound leakage prevention experience across all battery usage scenarios. It also prevents sudden increases in sound leakage due to changes in battery level and provides a reliable basis for identifying weak areas in sound leakage suppression and optimizing sound leakage prevention strategies. For headphones that fail the test, by integrating the sequence of unqualified sound leakage prevention nodes and calculating the unqualified coefficient, significant unqualified nodes are accurately identified, effectively locating the core weak points of the sound leakage prevention function. This approach balances the quality control needs of the production end with the actual user experience, and provides targeted support for subsequent optimization of sound leakage prevention strategies and improvement of product performance. Example 2

[0022] Please see Figure 1 As shown in the embodiment of the present invention, the AI ​​Bluetooth headset anti-leakage control method based on big data further includes the following steps: Step 3: Based on the significantly unqualified nodes in the anti-leakage nodes, analyze the rate of change of the anti-leakage suppression rate of the significantly unqualified nodes with the remaining power of the headphones, and determine the weak power range of the anti-leakage performance. All significantly unqualified nodes are integrated into a significantly unqualified node sequence according to the order of power decay, and two adjacent significantly unqualified nodes in the significantly unqualified node sequence are used as a node analysis window; Based on any node analysis window, obtain the remaining power of all significantly unqualified nodes. Subtract the remaining power of the next significantly unqualified node from the remaining power of the previous significantly unqualified node in the node analysis window to obtain the power change. Subtract the sound leakage suppression rate of the next significantly unqualified node from the sound leakage suppression rate of the previous significantly unqualified node in the node analysis window to obtain the suppression rate change value. If the change in the inhibition rate is less than 0, it means that the sound leakage inhibition rate in the corresponding node analysis window is showing a downward trend, and the corresponding node analysis window is recorded as the decreasing window. If the change in the inhibition rate is greater than or equal to 0, it means that the sound leakage inhibition rate in the corresponding node analysis window shows a non-decreasing trend, and the corresponding node analysis window is recorded as a non-decreasing window. The absolute value of the suppression rate change in the falling window is taken to obtain the suppression rate decay value. The ratio of the suppression rate decay value to the change in the amount of electricity corresponding to the falling window is calculated to obtain the change rate value. The interval corresponding to the change in electricity consumption is denoted as the electricity consumption interval; It should be noted that the rate of change value represents the degree to which the suppression rate changes as the headphone battery power decreases. The faster the rate of change, the more severe the decrease in the sound leakage suppression rate within the corresponding battery range, and the weaker the sound leakage prevention performance of the headphone. If the rate of change is greater than the rate of change threshold, the corresponding energy range is recorded as the weak energy range for suppression. If the rate of change is less than or equal to the rate of change threshold, the corresponding energy range is recorded as the non-suppressed weak range. Step 4: When the Bluetooth headset battery depletes to the weak suppression range, based on the dynamic gain compensation algorithm, combined with the correlation coefficient between the gain value and the leakage suppression rate, the leakage attenuation value is adjusted by gradient compensation of the gain value to avoid a sudden increase in leakage when the battery is low. By dividing the weak power range with a fixed power interval, gradient test nodes are obtained, and the sound leakage suppression rate and natural gain data of each gradient test node are quickly retrieved through a big data platform. The absolute value of the difference between the sound leakage suppression rates of adjacent gradient test nodes is used to obtain the gradient suppression rate attenuation value. All gradient suppression rate attenuation values ​​are integrated into a gradient suppression rate attenuation sequence according to the power attenuation order. The AI ​​reinforcement learning model learns the balance law between gain compensation and audio distortion based on historical compensation data, and optimizes the correlation coefficient between gain value and anti-leakage suppression rate in real time. The first gradient suppression rate attenuation value in the gradient suppression rate attenuation sequence is compared with the correlation coefficient to obtain the compensation gain value. The compensated gain value is summed with the natural gain value to obtain the compensated gain value. If the compensated gain value is less than the preset upper limit of the gain value, compensation will continue and a continued compensation signal will be generated. If the compensated gain value is greater than or equal to the preset upper limit of gain value, then gain compensation will stop. Based on the continued compensation signal, the compensation gain value is recorded as the previous compensation gain value. According to the power attenuation in the weak power range, a new compensation gain value is determined according to the gradient suppression rate attenuation sequence. The difference between the new compensation gain value and the previous compensation gain value is processed to obtain the gradient compensation value. Based on the gradient compensation value, the compensated gain value is gradient compensated and compared with the preset upper limit of the gain value until the gradient suppression rate attenuation sequence ends or the gain compensation is stopped. It should be noted that gradient compensation is applied to the gain value based on the gradient compensation value, thereby improving the sound leakage suppression rate. The purpose of gradient compensation is to avoid audio distortion caused by one-time compensation. While ensuring that the audio is not distorted, the sound leakage suppression rate is improved by adjusting the gain value. For example, the weak power range is suppressed as [35% power, 15% power]. To avoid audio distortion, the upper limit of the gain value is set to 15.6dB. Due to the upper limit of the gain value, unlimited compensation is not possible. 30% battery level (5% reduction in gradient suppression rate compared to 35% battery level): Compensation gain 1dB, natural gain 11dB, compensated gain 12dB. The compensated gain is less than the upper limit of 15.6dB, so continue to compensate. 25% battery (5% decrease from 30% battery): The cumulative gradient suppression rate decreases by 6%, requiring an additional 1dB of gain compensation. That is, the gradient compensation value is 1dB, and the compensated gain value is 12dB. The compensated gain value is less than the upper limit of the gain value of 15.6dB, so compensation continues. 20% battery (5% decrease from 25% battery): The cumulative gradient suppression rate decreases by 9%, requiring an additional 1dB compensation, i.e., the gradient compensation value is 1dB. The compensated gain value is 12dB. The compensated gain value is less than the upper limit of the gain value of 15.6dB, so compensation continues. 15% battery level (5% decrease from 20% battery level): The cumulative gradient suppression rate decreases by 12%, requiring an additional 1.2dB compensation. That is, the gradient compensation value is 1.2dB, and the compensated gain value is 12.2dB. The compensated gain value is less than the upper limit of the gain value of 15.6dB, so compensation continues. At 15% battery level, the noise reduction rate decreases by 12% (from 35% at full charge to 23% at 15% battery level). The natural gain drops to 8dB due to insufficient power supply (from 12dB at full charge, a decrease of 4dB). Experiments determined that the correlation coefficient between gain value and sound leakage suppression rate is 0.9%. Experimental data shows that "for every 1dB increase in gain, the sound leakage suppression rate increases by 0.9%". 15% power needs to compensate for 12% of the suppression rate attenuation value, and theoretically, it needs to compensate 12%÷0.9%≈13.3dB. Based on the continuity of gradient compensation, the previous gradient test node (20% battery) has already accumulated 3dB compensation (natural gain 9dB→12dB). The 15% battery level has a 3% greater attenuation of the sound leakage suppression rate than the 20% battery level (20% battery level attenuation value 9% → 15% battery level attenuation value 12%), requiring an additional compensation of approximately 0.2dB (3%÷0.9%≈0.33dB, taking 0.2dB for simplification). Therefore, on the basis of the 3dB compensation at 20% battery level, an additional 1.2dB is added (cumulative 4.2dB), making the natural gain 8dB + compensation 4.2dB = 12.2dB. The 12.2dB gain does not exceed the upper limit of 15.6dB, while simultaneously increasing the sound leakage suppression rate from 23% to 33%, achieving effective compensation. The technical solution of this invention is as follows: Based on significantly defective nodes in the anti-leakage node array, the rate of change of the anti-leakage suppression rate of the significantly defective nodes with the remaining battery power of the earphone is analyzed to determine the weak battery range for anti-leakage performance. When the Bluetooth earphone battery power decays to the weak battery range, based on the dynamic gain compensation algorithm and combined with the correlation coefficient between the gain value and the anti-leakage suppression rate, the gain value is gradient-compensated to adjust the anti-leakage attenuation value, avoiding a sudden increase in leakage when the battery is low. Based on the adjacent window analysis of the significantly defective node sequence, this invention calculates the rate of change of the anti-leakage suppression rate with the battery power to accurately locate the weak battery range where the anti-leakage performance decays rapidly, providing a clear scope and data support for subsequent targeted compensation. Relying on the dynamic gradient gain compensation algorithm and combined with the experimentally determined correlation coefficient between the gain value and the anti-leakage suppression rate, gradient gain compensation is performed on the weak range under the constraint of a preset upper limit of the gain value, effectively compensating for the attenuation of the anti-leakage suppression rate when the battery is low, avoiding a sudden increase in leakage, and avoiding audio distortion caused by over-compensation, thus reducing the risk of attenuation of the anti-leakage effect in low battery scenarios. Example 3

[0023] Based on the same inventive concept as the AI ​​Bluetooth headset anti-leakage control method based on big data in the foregoing embodiments, such as Figure 2 As shown, this application provides an AI Bluetooth headset anti-leakage control system based on big data, wherein the system includes the following modules: Anti-leakage test analysis module: Active anti-leakage test is performed on multiple Bluetooth headphones of the same model. By monitoring the battery decay process of the Bluetooth headphones, the anti-leakage suppression rate of Bluetooth headphones with different battery levels is determined, and it is judged whether the anti-leakage test of the Bluetooth headphones is qualified. Significant non-compliance identification module: If the test fails, extract the Bluetooth headphones that fail the sound leakage test and perform screening and analysis to determine the headphones that fail the sound leakage test. Analyze the non-compliance ratio of the non-compliance sound leakage nodes of the headphones that fail the sound leakage test and identify the significant non-compliance nodes among the non-compliance sound leakage nodes. The interval determination module analyzes the rate at which the sound leakage suppression rate of the significantly unqualified nodes changes with the remaining battery power of the headphones, based on the significantly unqualified nodes in the sound leakage prevention nodes, and determines the weak battery power range of the sound leakage prevention performance. Gain compensation module: When the Bluetooth headset's battery level drops to the range where sound leakage suppression is weak, the module uses a dynamic gain compensation algorithm and the correlation coefficient between the gain value and the sound leakage suppression rate to adjust the sound leakage attenuation value by performing gradient compensation on the gain value, thus avoiding a sudden increase in sound leakage when the battery is low.

[0024] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A big data-based AI Bluetooth headset anti-leakage control method, characterized in that: include: Active sound leakage prevention test was conducted on multiple Bluetooth headsets of the same model. By monitoring the battery decay process of the Bluetooth headsets, the sound leakage suppression rate of Bluetooth headsets with different battery levels was determined, and the sound leakage prevention test of the Bluetooth headsets was judged to be qualified. If the test fails, extract the Bluetooth headphones that fail the sound leakage test and screen them to identify the headphones that fail the sound leakage test. Analyze the failure rate of the failure points of the headphones that fail the sound leakage test and identify the significant failure points among the failure points. Based on the significantly unqualified nodes in the anti-leakage nodes, the rate of change of the anti-leakage suppression rate of the significantly unqualified nodes with the remaining power of the headphones is analyzed to determine the weak power range of the anti-leakage performance. When the Bluetooth headset's battery level drops to the range where sound leakage suppression is weak, a dynamic gain compensation algorithm is used. This algorithm combines the correlation coefficient between the gain value and the sound leakage suppression rate. By performing gradient compensation on the gain value, the sound leakage attenuation value is adjusted to avoid a sudden increase in sound leakage when the battery is low.

2. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 1, characterized in that: The process for determining the sound leakage suppression rate of Bluetooth earphones with different battery levels is as follows: The power decay process of the Bluetooth headset from full charge to low charge is monitored in real time. The interval between the power decay process of the Bluetooth headset from full charge to low charge is recorded as the power analysis interval. The power analysis interval is evenly divided into multiple power test nodes according to the power decay gradient. Using the microphone at the front end of the sound level meter as a sensor, the air pressure fluctuation caused by the leakage sound is converted into an electrical signal. The leakage sound pressure level of each power test node is recorded, a reference leakage sound pressure level is set, and the difference between the reference leakage sound pressure level and the leakage sound pressure level of the power test node is calculated to obtain the change value of the leakage sound pressure level of the power test node. If the change in sound pressure level of the power test node is greater than 0, the corresponding power test node is recorded as an effective sound leakage prevention node; otherwise, the corresponding power test node is recorded as an invalid sound leakage prevention node. The sound pressure level change of the effective sound leakage prevention node is compared with the reference sound pressure level to obtain the sound leakage suppression rate of the effective sound leakage prevention node.

3. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 2, characterized in that: The process for determining whether a Bluetooth headset passes the sound leakage test is as follows: If the sound leakage suppression rate of an effective sound leakage prevention node is less than the standard value of the sound leakage suppression rate, the corresponding effective sound leakage prevention node will be recorded as an unqualified sound leakage prevention node. Calculate the percentage of qualified anti-leakage nodes among all power test nodes, and record it as the anti-leakage performance value; If the sound leakage performance value is greater than or equal to the sound leakage performance threshold, the Bluetooth headset passes the sound leakage test; otherwise, the Bluetooth headset fails the sound leakage test.

4. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 3, characterized in that: The process for determining the substandard headphone with poor sound leakage prevention is as follows: Bluetooth headsets that fail the sound leakage prevention test are recorded as unqualified headsets. Based on any unqualified headset, the percentage of invalid sound leakage prevention nodes in the power analysis interval of the unqualified headset is calculated and recorded as the fault characterization value. If the fault characterization value is less than the fault characterization standard value, the corresponding unqualified headphones will be recorded as unqualified headphones for sound leakage prevention.

5. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 4, characterized in that: The process for identifying significantly defective nodes among the substandard sound leakage prevention nodes is as follows: Extract all the faulty sound leakage nodes from the headphones that fail to meet the sound leakage prevention standards, and integrate them into a faulty sound leakage node sequence according to the order of power decay. Based on any unqualified sound leakage prevention node, calculate the percentage of times the unqualified sound leakage prevention node appears in all unqualified sound leakage prevention headphones, and record it as the unqualified coefficient of the unqualified sound leakage prevention node. If the failure coefficient of a non-compliant sound leakage prevention node is greater than or equal to the failure coefficient threshold, the corresponding non-compliant sound leakage prevention node is recorded as a significantly non-compliant node.

6. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 5, characterized in that: The process for determining the weak electrical range for the sound leakage prevention performance is as follows: All significantly non-compliant nodes are integrated into a significantly non-compliant node sequence according to the order of power decay. The two adjacent significantly non-compliant nodes in the significantly non-compliant node sequence are used as a node analysis window to determine the falling window in the node analysis window. The absolute value of the suppression rate change value of the falling window is taken to obtain the suppression rate decay value. The ratio of the suppression rate decay value to the power change corresponding to the falling window is calculated to obtain the change rate value. The interval corresponding to the change in electricity consumption is denoted as the electricity consumption interval; If the rate of change is greater than the rate of change threshold, the corresponding energy range is recorded as the weak energy range for suppression.

7. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 6, characterized in that: The process of determining the descending window is as follows: Based on any node analysis window, obtain the remaining power of all significantly unqualified nodes. Subtract the remaining power of the next significantly unqualified node from the remaining power of the previous significantly unqualified node in the node analysis window to obtain the power change. Subtract the sound leakage suppression rate of the next significantly unqualified node from the sound leakage suppression rate of the previous significantly unqualified node in the node analysis window to obtain the suppression rate change value. If the change in the inhibition rate is less than 0, it means that the sound leakage inhibition rate in the corresponding node analysis window is showing a downward trend, and the corresponding node analysis window is recorded as the decreasing window.

8. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 6, characterized in that: The process of gradient compensation for the gain value is as follows: By dividing the weak power range with a fixed power interval, gradient test nodes are obtained, and the sound leakage suppression rate and natural gain data of each gradient test node are quickly retrieved through a big data platform. The absolute value of the difference between the sound leakage suppression rates of adjacent gradient test nodes is used to obtain the gradient suppression rate attenuation value. All gradient suppression rate attenuation values ​​are integrated into a gradient suppression rate attenuation sequence according to the power attenuation order. The AI ​​reinforcement learning model learns the balance law between gain compensation and audio distortion based on historical compensation data, and optimizes the correlation coefficient between gain value and anti-leakage suppression rate in real time. The first gradient suppression rate attenuation value in the gradient suppression rate attenuation sequence is compared with the correlation coefficient to obtain the compensation gain value. The compensated gain value is summed with the natural gain value to obtain the compensated gain value. The compensated gain value is compared and analyzed with the preset upper limit of the gain value to generate a further compensation signal; Based on the continued compensation signal, the compensation gain value is recorded as the pre-compensation gain value. According to the power attenuation in the weak power range, a new compensation gain value is determined according to the gradient suppression rate attenuation sequence. The new compensation gain value is subtracted from the pre-compensation gain value to obtain the gradient compensation value. Based on the gradient compensation value, gradient compensation is performed on the compensated gain value and compared with the preset upper limit of the gain value until the gradient suppression rate attenuation sequence ends or the gain compensation stops.

9. The AI ​​Bluetooth headset anti-leakage control method based on big data according to claim 8, characterized in that: The process of generating the continued compensation signal is as follows: If the compensated gain value is less than the preset upper limit of the gain value, compensation will continue and a continued compensation signal will be generated.

10. A big data-based AI Bluetooth headset anti-leakage control system, characterized in that, The system is used to perform the method according to any one of claims 1-9, the system comprising: Anti-leakage test analysis module: Active anti-leakage test is performed on multiple Bluetooth headphones of the same model. By monitoring the battery decay process of the Bluetooth headphones, the anti-leakage suppression rate of Bluetooth headphones with different battery levels is determined, and it is judged whether the anti-leakage test of the Bluetooth headphones is qualified. Significant non-compliance identification module: If the test fails, extract the Bluetooth headphones that fail the sound leakage test and perform screening and analysis to determine the headphones that fail the sound leakage test. Analyze the non-compliance ratio of the non-compliance sound leakage nodes of the headphones that fail the sound leakage test and identify the significant non-compliance nodes among the non-compliance sound leakage nodes. The interval determination module analyzes the rate at which the sound leakage suppression rate of the significantly unqualified nodes changes with the remaining battery power of the headphones, based on the significantly unqualified nodes in the sound leakage prevention nodes, and determines the weak battery power range of the sound leakage prevention performance. Gain compensation module: When the Bluetooth headset's battery level drops to the range where sound leakage suppression is weak, the module uses a dynamic gain compensation algorithm and the correlation coefficient between the gain value and the sound leakage suppression rate to adjust the sound leakage attenuation value by performing gradient compensation on the gain value, thus avoiding a sudden increase in sound leakage when the battery is low.

Images