Audio routing error intelligent detection and multi-mode prompting system and method of wireless earphone

Abstract

The invention relates to the technical field of audio detection, and discloses an audio routing error intelligent detection and multi-mode prompting system and method for a wireless earphone, and the method comprises the steps: S1, outputting a detection starting enable signal, and executing a step S2; s2, audio signals of a built-in main microphone of the earphone are collected, and when average signal energy is lower than a dynamic detection threshold value, the step S3 is executed; s3, when the motion frequency domain energy is higher than a motion threshold value, determining that the microphone routing is wrong and executing a step S4, and when the motion frequency domain energy is lower than the motion threshold value and the low energy duration is smaller than a first preset duration, returning to execute the step S2; s4, when the audio signal meets the mute condition and lasts for a second preset duration, detecting a voice gap and executing the step S5; and S5, independently routing the audio signal to a Bluetooth uplink encoder. According to the invention, two low-signal states of routing configuration error and environment quietness are effectively distinguished, and the false alarm rate is reduced.

Description

Technical Field

[0001] This invention relates to the field of audio detection technology, and in particular to a system and method for intelligent detection and multimodal alerting of audio routing errors in wireless headphones. Background Technology

[0002] Wireless Bluetooth headsets are widely used in voice communication scenarios. When users make calls using devices such as mobile phones and computers, errors in operating system settings or improper application permission configurations often cause the calling application to incorrectly select the headset's built-in microphone and use the device's built-in microphone instead. This microphone routing misconfiguration can lead to a decrease in call quality, including increased ambient noise, lower volume, and lack of noise cancellation, but users often fail to notice it immediately.

[0003] Existing technologies primarily employ LED indicator lights or simple alert sounds, but these have significant drawbacks. LED indicators become completely ineffective when the user is wearing headphones because they are out of sight. Alert sounds only provide feedback for user actions and lack automatic detection and proactive alarm mechanisms for system routing configuration errors. More critically, existing solutions cannot intelligently distinguish between microphone routing errors and situations where the user is in a quiet environment; simple energy detection leads to a high false alarm rate. Furthermore, existing alert sounds are directly mixed into the audio stream and transmitted to the other party, interfering with the call and compromising user privacy. Additionally, the lack of a continuous reminder mechanism means the system stops alerting if the user doesn't address the issue promptly, resulting in persistent problems without the user's knowledge, severely impacting call experience and communication efficiency. Summary of the Invention

[0004] This invention provides an intelligent detection and multimodal alert system and method for audio routing errors in wireless headphones. By combining Bluetooth call status, wear detection, microphone energy analysis, and three-axis accelerometer motion feature recognition, this invention achieves high-precision automatic detection of microphone routing errors, effectively distinguishes between two low signal states: routing configuration errors and quiet environments, and reduces the false alarm rate.

[0005] In a first aspect, the present invention provides a method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones, the method comprising: S1. Monitor the hands-free protocol connection status flag and the headset wearing status flag and output a detection start enable signal, then execute step S2. S2. Acquire the audio signal from the built-in main microphone of the earphone, calculate the average signal energy and dynamic detection threshold, and when the average signal energy is lower than the dynamic detection threshold, execute step S3. S3. Calculate the motion frequency domain energy. When the motion frequency domain energy is higher than the motion threshold, confirm that the microphone routing is incorrect and execute step S4. When the motion frequency domain energy is lower than the motion threshold and the duration of low energy is less than the first preset duration, return to execute step S2. S4. When the audio signal meets the silence condition and continues for a second preset duration, a speech gap is detected and step S5 is executed. S5. Generate a digital signal for a prompt tone and drive the speaker to play it. The audio signal is independently routed to the Bluetooth uplink encoder, and the LED is driven to light up.

[0006] In conjunction with the first aspect, in a first implementation of the first aspect of the present invention, step S1 includes: The connection management event of the baseband layer is parsed by the Bluetooth protocol stack to identify the establishment status of the synchronous connection-oriented link. When the synchronous connection-oriented link is in an active state, a call connection status flag is generated. Infrared light is emitted by an infrared sensor and infrared signals reflected from the ear canal are received. The intensity value of the infrared signal is compared with a preset intensity threshold to obtain a wearing status flag. The capacitance value of the sensor electrode is detected by a capacitive sensor, and the capacitance value is compared with a preset capacitance threshold to obtain the wearing status flag bit; When both the call connection status flag and the wearing status flag are high, a detection start enable signal is output and step S2 is executed.

[0007] In conjunction with the first aspect, in a second implementation of the first aspect of the present invention, step S2 includes: The audio signal from the main microphone is amplified by a preamplifier and then input into an analog-to-digital converter for quantization. A detection window is set and the average signal energy within the detection window is calculated. After the call is established, the ambient noise floor reference value of the main microphone is calculated, and the ambient noise floor reference value is added to a preset offset to obtain the dynamic detection threshold. The average signal energy is compared with the dynamic detection threshold. When the average signal energy is lower than the dynamic detection threshold, the detection window is marked as a low-energy window. When a preset number of consecutive detection windows are marked as low-energy windows, a preliminary abnormality flag is output and step S3 is executed.

[0008] In conjunction with the first aspect, in a third implementation of the first aspect of the present invention, step S3 includes: The X-axis acceleration components, Y-axis acceleration components, and Z-axis acceleration components are acquired by a triaxial accelerometer and then synthesized into a net acceleration signal. Apply a Hanning window function to the net acceleration signal and perform a fast Fourier transform to calculate the frequency domain power spectral density; The energy of the frequency domain power spectral density is integrated to obtain the motion frequency domain energy, and the motion frequency domain energy is compared with the motion threshold. When the motion frequency domain energy is higher than the motion threshold, output a routing error confirmation flag and execute step S4. When the motion frequency domain energy is lower than the motion threshold and the duration of low energy is less than the first preset duration, return to execute step S2.

[0009] In conjunction with the first aspect, in a fourth implementation of the first aspect of the present invention, the X-axis acceleration component, Y-axis acceleration component, and Z-axis acceleration component are acquired by a triaxial accelerometer, and a net acceleration signal is synthesized, including: The X-axis acceleration component, Y-axis acceleration component, and Z-axis acceleration component are acquired by a triaxial accelerometer and transmitted to the main control chip. The average values ​​of the X-axis acceleration component, the Y-axis acceleration component, and the Z-axis acceleration component within a preset time period are calculated as DC bias values. The corresponding DC bias values ​​are then subtracted from the X-axis acceleration component, the Y-axis acceleration component, and the Z-axis acceleration component to obtain the debiased X-axis acceleration, the debiased Y-axis acceleration, and the debiased Z-axis acceleration. The net acceleration signal is synthesized based on the debiased X-axis acceleration, the debiased Y-axis acceleration, and the debiased Z-axis acceleration.

[0010] In conjunction with the first aspect, in a fifth implementation of the first aspect of the present invention, step S4 includes: The audio signal is segmented into frames to obtain multiple audio frames, and the short-time frame energy and short-time frame zero-crossing rate are calculated. The energy threshold is calculated based on the environmental noise baseline and the preset energy offset, and the lower threshold and upper threshold of the zero crossing rate are set. When the short-time frame energy is higher than the energy threshold and the short-time frame zero-crossing rate is between the lower threshold and the upper threshold of the zero-crossing rate, the audio frame is marked as a voice activity frame; otherwise, the audio frame is marked as a silent frame. The number of consecutive frames marked as silent is counted and the corresponding silence duration is calculated. When the silence duration reaches the second preset duration, a playback enable signal is output and step S5 is executed.

[0011] In conjunction with the first aspect, in a sixth implementation of the first aspect of the present invention, when the short-time frame energy is higher than an energy threshold and the short-time frame zero-crossing rate is between the lower threshold and the upper threshold of the zero-crossing rate, the audio frame is marked as a speech activity frame; otherwise, the audio frame is marked as a silence frame, including: The short-time frame energy of each audio frame is compared with an energy threshold; When the energy of the short-time frame is higher than the energy threshold, it is determined that the audio frame meets the energy condition; The short-time frame zero-crossing rate of each audio frame is compared with the lower threshold of the zero-crossing rate and the upper threshold of the zero-crossing rate. When the short-time frame zero-crossing rate is greater than the lower threshold of the zero-crossing rate and less than the upper threshold of the zero-crossing rate, it is determined that the audio frame satisfies the zero-crossing rate condition. When both the energy condition and the zero-crossing rate condition are met, the audio frame is marked as a voice activity frame; when neither the energy condition nor the zero-crossing rate condition is met, the audio frame is marked as a silent frame.

[0012] In conjunction with the first aspect, in the seventh implementation of the first aspect of the present invention, step S5 includes: A sinusoidal carrier signal is generated by a digital signal processor. A trapezoidal amplitude envelope is applied to the sinusoidal carrier signal to obtain a single prompt tone unit. Two consecutive prompt tone units are generated and a silence segment of a preset duration is inserted between them to obtain a digital prompt tone signal. The digital signal of the prompt tone is volume normalized to make its amplitude reach the preset ratio of the call volume, and a local playback path mark is set for the digital signal of the prompt tone in the audio mixer; The digital signal of the prompt tone is weighted and mixed with the Bluetooth downlink audio signal to obtain a mixed audio signal. The audio signal of the main microphone is independently routed to the input of the Bluetooth uplink encoder to avoid entering the uplink. The mixed audio signal is input to the digital-to-analog converter and driven by the power amplifier to play the speaker. Configure the color parameters of the LED and set the flashing sequence to synchronize the flashing start time of the LED with the playback start time of the digital signal of the prompt tone.

[0013] In conjunction with the first aspect, in the eighth implementation of the first aspect of the present invention, the method for intelligent detection and multimodal prompting of audio routing errors in the wireless earphone further includes: The average signal energy is continuously monitored, and the detection process is terminated when the average signal energy of a preset number of consecutive detection windows is higher than the dynamic detection threshold. When the average signal energy remains below the dynamic detection threshold and the call connection status flag and the wearing status flag remain unchanged, the periodic reminder process is initiated. Read the current value of the reminder count counter. When the current value of the reminder count counter is less than or equal to the first count threshold, set the waiting time interval to the first duration. When the value of the reminder count counter is greater than the first count threshold and less than or equal to the second count threshold, set the waiting time interval to the second duration. When the value of the reminder count counter is greater than the second count threshold, set the waiting time interval to the third duration. Start the timer to load the waiting time interval and start the countdown. When the timer countdown ends, the reminder count counter is incremented, and step S4 is executed again. When a voice gap is detected, step S5 is executed again.

[0014] Secondly, the present invention provides an intelligent detection and multimodal prompting system for audio routing errors in wireless headphones, wherein the intelligent detection and multimodal prompting system for audio routing errors in wireless headphones is used to execute any of the methods for intelligent detection and multimodal prompting of audio routing errors in wireless headphones described in the present invention.

[0015] The technical solution provided by this invention combines Bluetooth call status, wear detection, microphone energy analysis, and triaxial accelerometer motion feature recognition to achieve high-precision automatic detection of microphone routing errors. This effectively distinguishes between two low-signal states: routing configuration errors and quiet environments, reducing the false alarm rate. A dynamic threshold calculation method adjusts the detection threshold in real time based on ambient noise, enhancing adaptability and detection accuracy in different noise environments. Motion frequency domain energy analysis technology captures the head vibration characteristics of the user while speaking, providing a reliable auxiliary confirmation basis for routing error determination. Path marking and signal separation technology in the audio mixer enables localized playback of prompt tones, ensuring that the prompt tones are played only through the speaker and not transmitted to the other party via the Bluetooth uplink encoder, protecting call privacy and avoiding interference with the other party. An intelligent timing control mechanism based on VAD voice activity detection accurately identifies user speech gaps through joint judgment of short-time frame energy and zero-crossing rate, ensuring that prompt tones play during user speech gaps, avoiding interruptions to normal calls. The dual-modal audio-visual synchronous alert solution combines auditory and visual stimulation to ensure that users can effectively perceive problems whether they are wearing the device (through in-ear alerts) or not (through LED lights), achieving full-scene coverage. The intelligent attenuation periodic alert strategy uses segmented, progressively increasing waiting time intervals to ensure timely notification of problems while avoiding excessive disturbance to the user, reflecting a human-centered design philosophy.

[0016] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of an embodiment of the intelligent detection and multimodal prompting method for audio routing errors in wireless headphones according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The terms "comprising" and "having," and any variations thereof, used in the embodiments of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0021] To facilitate understanding of this embodiment, a detailed description of the intelligent detection and multimodal prompting method for audio routing errors in wireless headphones, as disclosed in this embodiment of the invention, will be provided first. For example... Figure 1 As shown, this method includes the following steps: S1. Monitor the hands-free protocol connection status flag and the headset wearing status flag and output a detection start enable signal, then execute step S2. S2. Collect the audio signal from the built-in main microphone of the earphone, calculate the average signal energy and dynamic detection threshold. When the average signal energy is lower than the dynamic detection threshold, execute step S3. S3. Calculate the motion frequency domain energy. When the motion frequency domain energy is higher than the motion threshold, confirm that the microphone routing is incorrect and execute step S4. When the motion frequency domain energy is lower than the motion threshold and the duration of low energy is less than the first preset duration, return to execute step S2. S4. When the audio signal meets the silence condition and continues for a second preset duration, a speech gap is detected and step S5 is executed. S5 generates a digital signal for a prompt tone and drives the speaker to play it. The audio signal is independently routed to the Bluetooth uplink encoder, while simultaneously driving the LED to light up.

[0022] In one specific embodiment, step S1 includes: The connection management events of the baseband layer are parsed through the Bluetooth protocol stack to identify the establishment status of the synchronous connection-oriented link. When the synchronous connection-oriented link is in an active state, a call connection status flag is generated. The infrared sensor emits infrared light and receives the infrared signal reflected from the ear canal. The intensity value of the infrared signal is compared with a preset intensity threshold to obtain the wearing status flag. The capacitance value of the sensor electrode is detected by a capacitive sensor, and the capacitance value is compared with a preset capacitance threshold to obtain the wearing status flag bit; When both the call connection status flag and the wear status flag are high, the detection start enable signal is output and step S2 is executed.

[0023] Specifically, at the Bluetooth protocol stack level, connection management events transmitted from the baseband layer are parsed. These events include operation commands or status flags such as connection requests, connection establishment, connection interruption, and connection release. By parsing connection management events, it is determined whether a Synchronous Connection-Oriented Audio (SCO) link has been established, which is the basic channel for audio transmission in the headset. When an active SCO link is detected, it indicates that the headset has entered a call or voice communication state, and a high-level call connection status flag is generated as a basis for determining call activity. Simultaneously, a wearing status determination mechanism is constructed. Using an infrared sensor module, infrared light signals of preset intensity are periodically emitted towards the earbud outlet. After the infrared light enters the ear canal, it is reflected back by the ear canal wall. The reflected signal is captured by the photodetector at the receiving end and converted into an electrical signal of corresponding intensity. The actual intensity of the infrared signal is compared with a set intensity threshold. If the intensity is higher than the threshold, it indicates that there is a stable target on the reflection path, that is, the user is wearing the headset, and a high-level wearing status flag is generated. Simultaneously, a capacitive contact detection mechanism is configured, utilizing conductive electrodes on the earphone body to detect minute capacitance changes when it comes into contact with the skin. When the capacitance value around the electrodes is higher or lower than a preset capacitance threshold, it indicates that the earphone is in contact with the human body surface, determining that the user has worn the earphone, and generating a high-level wearing status flag. Through logical judgment, when the call connection status flag and any wearing status flag are simultaneously at a high level, it indicates that the earphone is both in a call and being worn by the user. At this time, the detection start condition is triggered, a detection start enable signal is output, and the audio signal anomaly analysis and routing error detection steps are entered, i.e., step S2 is executed.

[0024] In one specific embodiment, step S2 includes: The audio signal from the main microphone is amplified by a preamplifier and then input into an analog-to-digital converter for quantization. A detection window is set and the average signal energy within the detection window is calculated. After the call is established, the ambient noise floor reference value of the main microphone is calculated, and the ambient noise floor reference value is added to the preset offset to obtain the dynamic detection threshold. The average signal energy is compared with the dynamic detection threshold. When the average signal energy is lower than the dynamic detection threshold, the detection window is marked as a low-energy window. When a preset number of consecutive detection windows are marked as low-energy windows, a preliminary abnormality flag is output and step S3 is executed.

[0025] Specifically, energy analysis is performed on the audio signal acquired by the main microphone to determine if there are signal anomalies during the current call due to the microphone not being properly routed. A preamplifier amplifies the analog audio signal acquired by the main microphone to improve the dynamic range and signal-to-noise ratio. The preamplified signal is then input to an analog-to-digital converter (ADC) to convert the continuous-time analog signal into a discrete-time digital signal sequence at a specified sampling rate. A fixed-duration detection window is set, for example, every 3 seconds as an analysis cycle. Within each detection window, all sampled data are squared and summed point by point. The average signal energy within that time period is then calculated based on the sample size, reflecting the current input signal strength level of the microphone. During the initial few seconds after the call is established, the main microphone signal is continuously acquired, and the average noise energy during this period is calculated as an ambient noise floor baseline, reflecting the typical intensity level of background noise when the user is not speaking. The ambient noise floor baseline is added to a preset energy offset to form the dynamic detection threshold for the current call environment. The energy offset is set to a fixed gain value of 3 to 5 dB, or determined according to an empirical coefficient based on actual conditions. After each detection window is completed, the average signal energy calculated within the detection window is compared with the dynamic detection threshold. If the average energy is lower than the dynamic detection threshold, the current window is marked as a low-energy window. A low-energy window indicates that the microphone has not collected a valid voice signal, or that the current audio path has failed to transmit a signal to the main microphone. To prevent misjudgment caused by occasional silence or short pauses, a sliding window mechanism is used for continuous multi-window judgment. When the number of consecutive windows reaches a preset threshold (e.g., 3 consecutive windows) and all are marked as low-energy windows, it is considered that the main microphone is abnormally silent during the call, and a preliminary abnormality flag is output, indicating that there is a configuration error in the current audio input path. The process then enters the auxiliary confirmation stage and executes step S3.

[0026] After the call is established, the ambient noise floor reference value of the main microphone is calculated. The ambient noise floor reference value is added to a preset offset to obtain the dynamic detection threshold. The process also includes: setting the number of historical detection windows corresponding to the sliding window duration; maintaining a circular queue that stores the average signal energy of the most recent number of historical detection windows; storing the average signal energy of the current detection window in the circular queue and removing the oldest historical average signal energy value after each calculation; summing all historical average signal energy values ​​in the circular queue and dividing by the number of historical detection windows to obtain the average energy of the sliding window; squaring the differences between all historical average signal energy values ​​in the circular queue and the average energy of the sliding window and summing them; dividing the sum by the number of historical detection windows and taking the square root to obtain the standard deviation of the sliding window energy; adding the average energy of the sliding window to the product of a preset empirical coefficient and the standard deviation of the sliding window energy to obtain the adaptive dynamic detection threshold; comparing the average signal energy of the current detection window with the adaptive dynamic detection threshold; marking the current detection window as a low-energy window when the average signal energy is lower than the adaptive dynamic detection threshold; and outputting a preliminary abnormality flag and executing step S3 when a preset number of consecutive detection windows are marked as low-energy windows.

[0027] Before outputting a preliminary anomaly flag and executing step S3 when a predetermined number of consecutive detection windows are all marked as low-energy windows, the process further includes: calculating the total number of detection windows marked as low-energy windows within the most recent predetermined historical duration and dividing it by the total number of detection windows within that duration to obtain the low-energy window percentage; multiplying the low-energy window percentage by a predetermined first weighting coefficient to obtain the energy anomaly confidence component; calculating the duration for which the wearing status flag bit remains high within the most recent predetermined historical duration and dividing it by that historical duration to obtain the wearing stability percentage; multiplying the wearing stability percentage by a predetermined second weighting coefficient to obtain the wearing stability confidence component; and reading the duration of the call connection status flag bit. When the call stability duration threshold is reached, the call stability flag is set to true and the call stability confidence component is assigned a preset third weight coefficient. When the duration of the state does not exceed the call stability duration threshold, the call stability flag is set to false and the call stability confidence component is assigned zero. The energy anomaly confidence component, the wearing stability confidence component, and the call stability confidence component are weighted and summed to obtain the comprehensive anomaly confidence. The comprehensive anomaly confidence is compared with the preset confidence threshold. When the comprehensive anomaly confidence is higher than the confidence threshold, a high confidence anomaly flag is output and step S3 is executed to perform motion frequency domain energy analysis. When the comprehensive anomaly confidence is lower than the confidence threshold, it is determined to be a low confidence anomaly and the process returns to step S2 to continue monitoring.

[0028] In one specific embodiment, step S3 includes: The X-axis acceleration components, Y-axis acceleration components, and Z-axis acceleration components are acquired by a triaxial accelerometer and then synthesized into a net acceleration signal. Apply a Hanning window function to the net acceleration signal and perform a fast Fourier transform to calculate the frequency domain power spectral density; The energy of the motion frequency domain is obtained by integrating the power spectral density in the frequency domain, and then compared with the motion threshold. When the motion frequency domain energy is higher than the motion threshold, output a routing error confirmation flag and execute step S4. When the motion frequency domain energy is lower than the motion threshold and the duration of low energy is less than the first preset duration, return to execute step S2.

[0029] Specifically, acceleration components in the X, Y, and Z axes are collected from a triaxial accelerometer at a set sampling rate (e.g., 100Hz). The net acceleration signal is synthesized by summing the squares of these three components at each sampling point and taking the square root, reflecting the total acceleration variation across the three spatial dimensions. To extract the periodic vibration characteristics exhibited by the user during speech, a Hanning window function is applied to the net acceleration signal within a preset analysis window to reduce spectral leakage. A Fast Fourier Transform (FFT) is then performed to transform the time-domain signal to the frequency domain, obtaining the frequency-domain power spectral density distribution. The power spectral density reflects the energy distribution characteristics of each frequency component, with characteristic vibration frequency components forming in the range of 50Hz to 200Hz during speech. The power spectral density within this frequency range is integrated to calculate its total energy, which is defined as the motion frequency-domain energy, representing the frequency-domain intensity of head or jaw movements. The motion frequency domain energy is compared with a preset motion threshold. If the motion frequency domain energy is higher than the threshold, it is assumed that the user is indeed making sounds during the call. However, due to the persistently low energy state of the microphone signal, it indicates that the audio routing is not configured correctly. Therefore, a high-level routing error confirmation flag is output, and the process proceeds to the next step, i.e., step S4. Conversely, if the current motion frequency domain energy is lower than the motion threshold, it is further determined whether the duration of the low energy state is less than a first preset duration, such as 5 seconds. If the duration does not exceed the threshold, it is considered a normal quiet state or a short period of silence. The system exits the current path and returns to step S2 to re-acquire and judge the audio signal energy.

[0030] The motion frequency domain energy is obtained by integrating the power spectral density in the frequency domain, including: dividing the power spectral density into three sub-bands: low-frequency, mid-frequency, and high-frequency. The low-frequency band corresponds to the frequency range of slight head swaying characteristics, the mid-frequency band corresponds to the frequency range of jaw opening and closing movements, and the high-frequency band corresponds to the frequency range of vocal cord vibration bone conduction characteristics. The power spectral density of each of the three sub-bands is then integrated to obtain the low-frequency, mid-frequency, and high-frequency energy components. The energy ratio sequence of the low-frequency, mid-frequency, and high-frequency energy components is calculated, and the low-frequency energy is further divided... The ratio of the energy quantity to the mid-frequency energy component is recorded as the first energy ratio, the ratio of the mid-frequency energy component to the high-frequency energy component is recorded as the second energy ratio, and the ratio of the low-frequency energy component to the high-frequency energy component is recorded as the third energy ratio. The first energy ratio, the second energy ratio, and the third energy ratio are compared with their respective preset energy ratio thresholds. When the first energy ratio, the second energy ratio, and the third energy ratio are all within their respective preset energy ratio threshold ranges, it is determined that a speech motion feature pattern has been detected. The low-frequency energy component, the mid-frequency energy component, and the high-frequency energy component are weighted and summed to obtain the motion frequency domain energy.

[0031] In one specific embodiment, X-axis acceleration components, Y-axis acceleration components, and Z-axis acceleration components are acquired using a triaxial accelerometer, and a net acceleration signal is synthesized, including: The X-axis acceleration components, Y-axis acceleration components, and Z-axis acceleration components are acquired by a triaxial accelerometer and transmitted to the main control chip. The average values ​​of the X-axis acceleration component, Y-axis acceleration component, and Z-axis acceleration component within a preset time period are calculated as DC bias values. The corresponding DC bias values ​​are then subtracted from the X-axis acceleration component, Y-axis acceleration component, and Z-axis acceleration component to obtain the debiased X-axis acceleration, debiased Y-axis acceleration, and debiased Z-axis acceleration. The net acceleration signal is synthesized based on the debiased X-axis acceleration, debiased Y-axis acceleration, and debiased Z-axis acceleration.

[0032] Specifically, raw acceleration information in each direction is obtained from a three-axis accelerometer embedded inside the wireless earphone. The accelerometer collects the X-axis, Y-axis, and Z-axis acceleration components in real time at a set sampling frequency, and transmits the data through I... 2Bus interfaces such as C or SPI sequentially transmit the three-axis acceleration component data to the sensor data processing unit of the headphone's main control chip. Upon receiving the raw acceleration data, the main control chip performs DC bias correction on the acceleration component of each axis to remove constant offsets introduced by device static errors, sensor attitude, or gravity component coupling. Within a preset static time period (e.g., 500 milliseconds or 1 second), the acceleration sampling data for the X, Y, and Z axes are accumulated and averaged to obtain the average acceleration values ​​for each axis as the DC bias values. The X-axis acceleration component at each real-time sampling point is subtracted from the X-axis DC bias value to obtain the debiased X-axis acceleration data. The same debiasing operation is repeated for the Y and Z-axis acceleration components to obtain the debiased Y-axis acceleration and debiased Z-axis acceleration, respectively. The main control chip synthesizes the three-axis debiased acceleration components into a one-dimensional net acceleration signal by summing the squares and taking the square root. The net acceleration amplitude is obtained by summing the squares of the debiased X, Y, and Z accelerations and then taking the square root.

[0033] In one specific embodiment, step S4 includes: The audio signal is segmented into frames to obtain multiple audio frames, and the short-time frame energy and short-time frame zero-crossing rate are calculated. The energy threshold is calculated based on the environmental noise baseline and the preset energy offset, and the lower threshold and upper threshold of the zero crossing rate are set. When the short-time frame energy is higher than the energy threshold and the short-time frame zero-crossing rate is between the lower zero-crossing rate threshold and the upper zero-crossing rate threshold, the audio frame is marked as a speech activity frame; otherwise, the audio frame is marked as a silent frame. The number of consecutive frames marked as silent is counted and the corresponding silence duration is calculated. When the silence duration reaches the second preset duration, a playback enable signal is output and step S5 is executed.

[0034] Specifically, continuous audio data is processed by frame segmentation using a sliding window approach. The continuous audio signal is divided into multiple short frames of equal length along the time axis, for example, with a frame length of 20 to 30 milliseconds and a frame shift of 10 milliseconds or less, to ensure sufficient temporal overlap between frames and improve the continuity and recognizability of speech features. Within each audio frame, its short-time frame energy and zero-crossing rate are calculated. The short-time frame energy characterizes the intensity level of the audio signal in that frame, obtained by summing the squares of all samples within the frame and then normalizing. The zero-crossing rate represents the frequency at which the audio signal crosses zero, reflecting the frequency structure and degree of change of the signal, calculated by counting the number of symbol changes between adjacent samples. Based on the ambient noise baseline value of the main microphone obtained at the initial stage of call establishment, and combined with an empirically set energy offset, an additive operation is performed to obtain the dynamic energy threshold adapted for the current detection period. Two empirical values ​​are also set as the lower and upper limits of the zero-crossing rate to eliminate false interference from low-frequency silent segments or high-frequency background noise. For each frame, a judgment is made. If the short-time frame energy of an audio frame is greater than the dynamic energy threshold and its zero-crossing rate is between the set upper and lower thresholds, then the frame is considered to meet the typical speech production conditions, that is, it has both effective energy and a reasonable frequency structure, and the audio frame is marked as a speech activity frame. If either condition is not met, the frame is marked as a silent frame, indicating that the user has not spoken during this period or the audio routing is abnormal. During the processing, the number of consecutive frames marked as silent frames is continuously counted, and the silence duration is calculated by accumulating the duration of each frame. When the continuous silence duration reaches the set second preset duration threshold, such as 200 milliseconds or 300 milliseconds, it is considered that the user is currently in a speech gap window. At this point in time when the user is not disturbed, a prompt mechanism is triggered, a playback enable signal is output, and the process proceeds to step S5.

[0035] In one specific embodiment, when the short-time frame energy is higher than the energy threshold and the short-time frame zero-crossing rate is between the lower zero-crossing rate threshold and the upper zero-crossing rate threshold, the audio frame is marked as a speech activity frame; otherwise, the audio frame is marked as a silence frame, including: The short-time frame energy of each audio frame is compared with an energy threshold; When the energy of a short frame is higher than the energy threshold, the audio of that frame is determined to meet the energy condition. The short-time frame zero-crossing rate of each audio frame is compared with the lower zero-crossing rate threshold and the upper zero-crossing rate threshold. When the short-time frame zero-crossing rate is greater than the lower threshold of zero-crossing rate and less than the upper threshold of zero-crossing rate, the audio frame is determined to meet the zero-crossing rate condition. When both the energy condition and the zero-crossing rate condition are met, the audio frame is marked as a speech activity frame; when neither the energy condition nor the zero-crossing rate condition is met, the audio frame is marked as a silent frame.

[0036] Specifically, the short-time frame energy of the current frame audio is compared with a dynamically set energy threshold. If the short-time frame energy is higher than the energy threshold, the frame audio is deemed to meet the energy condition, indicating that the frame contains sufficient energy components to support the basic intensity requirements of the user's voice. Simultaneously, the zero-crossing rate of the frame is calculated, and the short-time frame zero-crossing rate is compared with preset lower and upper zero-crossing rate thresholds. If the result shows that the zero-crossing rate of the frame audio is greater than the lower threshold and less than the upper threshold, it indicates that the frame possesses typical speech frequency characteristics, i.e., it does not belong to low-frequency stationary noise or high-frequency jitter interference, and the frame is deemed to meet the zero-crossing rate condition. After completing the dual evaluation of the energy condition and the zero-crossing rate condition, the evaluation results are logically judged. If the frame simultaneously meets both the energy condition and the zero-crossing rate condition, it is marked as a speech activity frame, indicating that the frame audio highly originates from the user's normal speech signal and should be considered an active part of the call. However, if either condition is not met, whether due to insufficient energy or abnormal frequency structure, the frame audio is marked as a silent frame, indicating that the frame does not possess speech characteristics or that there is a configuration abnormality in the acquisition link.

[0037] In one specific embodiment, step S5 includes: A sinusoidal carrier signal is generated by a digital signal processor. A trapezoidal amplitude envelope is applied to the sinusoidal carrier signal to obtain a single prompt tone unit. Two consecutive prompt tone units are generated and a silence segment of preset duration is inserted between them to obtain a digital prompt tone signal. The digital signal of the prompt tone is volume-normalized to make its amplitude reach the preset ratio of the call volume, and a local playback path mark is set for the digital signal of the prompt tone in the audio mixer; The digital signal of the prompt tone is weighted and mixed with the Bluetooth downlink audio signal to obtain a mixed audio signal. The audio signal of the main microphone is independently routed to the input of the Bluetooth uplink encoder to avoid entering the uplink. The mixed audio signal is input to the digital-to-analog converter and driven by the power amplifier to play the speaker. Configure the color parameters of the LED and set the flashing sequence to synchronize the start time of the LED flashing with the start time of the digital signal of the prompt tone.

[0038] Specifically, the digital signal processor inside the earphone generates a basic sinusoidal carrier signal with a defined frequency parameter, set at around 800Hz. To ensure a smooth start and end boundary for the alert tone, a trapezoidal amplitude envelope function is applied to the sinusoidal carrier. Gradual amplitude changes are created along the rising and falling edges of the envelope function, making the entire alert tone unit sound more natural and avoiding abrupt sound impacts, resulting in a single alert tone unit. A second alert tone unit with the same parameters is constructed, and a preset duration silence segment, such as a 100-millisecond all-zero signal segment, is inserted between the two units, forming a "alert tone—silence—alert tone" combination structure on the timeline, resulting in a sequence of digital alert tone signals. The digital alert tone signals undergo volume normalization processing. Through normalization and scaling operations, their amplitude is adjusted to a preset proportion of the current call volume, set to 50% to 70%. After volume processing, a local playback path marker is added to the alert tone signal in the audio mixer. This marker instructs the audio routing module to restrict the alert tone signal to the headset's local playback path during signal forwarding, preventing it from being encoded into the Bluetooth uplink audio signal. Within the audio mixer, the alert tone signal and the Bluetooth downlink call audio signal are weighted and mixed using a linear superposition strategy to generate a mixed audio signal. This mixed audio signal is then input to a digital-to-analog converter (DAC), which drives the headset's internal speaker via a power amplifier for sound output, enabling local playback of the alert tone. Simultaneously, visual cues are configured. By setting the color parameters of the LEDs (e.g., selecting amber RGB values ​​of 255,191,0) and their flashing timing parameters (e.g., alternating on and off every 200 milliseconds), and aligning the LED flashing start time with the start time of the alert tone digital signal playback using synchronization control logic, a synchronized audio-visual cues are achieved. Users can directly perceive routing anomalies through hearing while wearing the headset, and receive system alerts through visual signals when not wearing it.

[0039] The system weighted mixes the digital alert tone signal and the Bluetooth downlink audio signal, and independently routes the main microphone's audio signal to the Bluetooth uplink encoder input to avoid entering the uplink. This includes: writing a local playback path marker field to the header of the digital alert tone signal data packet in the audio mixer and setting the marker value to a preset local playback identifier; writing a downlink playback path marker field to the header of the Bluetooth downlink audio signal data packet and setting the marker value to a preset downlink playback identifier; and writing an uplink transmission path marker field to the header of the main microphone's audio signal data packet and setting the marker value to a preset uplink transmission identifier. In the downlink mixing channel of the audio router, the path marker field of each input signal data packet is read, and when the marker value of the path marker field is either the local playback identifier or the downlink playback identifier, the path marker field is selected. When the identification code is obtained, the signal is routed to the downlink mixer for weighted summation to obtain a mixed audio signal. In the uplink transmission channel of the audio router, the path tag field of each input signal data packet is read. When the tag value of the path tag field is the uplink transmission identification code, the signal is routed to the input of the Bluetooth uplink encoder. When the tag value of the path tag field is not the uplink transmission identification code, the signal is blocked from entering the Bluetooth uplink encoder. The mixed audio signal is input to the digital-to-analog converter for digital-to-analog conversion to obtain an analog mixed audio signal. The analog mixed audio signal is input to the power amplifier for power amplification and then drives the speaker to play. At the same time, the audio signal of the main microphone is input to the Bluetooth uplink encoder via a separate uplink transmission channel for encoding and then transmitted to the peer device through the synchronous connection link.

[0040] Before the mixed audio signal is input into the digital-to-analog converter and driven by the power amplifier to drive the speaker for playback, the method further includes: before the digital prompt signal starts playing, reading the volume gain coefficient of the current Bluetooth downlink audio signal and storing it as the original volume gain coefficient; calculating the product of the original volume gain coefficient and a preset volume attenuation ratio to obtain the attenuated volume gain coefficient; linearly transitioning the volume gain coefficient of the Bluetooth downlink audio signal from the original volume gain coefficient to the attenuated volume gain coefficient; during the playback of the digital prompt signal, maintaining the volume gain coefficient of the Bluetooth downlink audio signal at the attenuated volume gain coefficient; when weighting the digital prompt signal and the Bluetooth downlink audio signal, setting the mixing weight coefficient of the digital prompt signal to a first preset weight value and setting the mixing weight coefficient of the Bluetooth downlink audio signal to a second preset weight value, wherein the first preset weight value is greater than the second preset weight value; after the digital prompt signal finishes playing, linearly restoring the volume gain coefficient of the Bluetooth downlink audio signal from the attenuated volume gain coefficient to the original volume gain coefficient; and after completing the volume restoration process, inputting the mixed audio signal into the digital-to-analog converter and driven by the power amplifier to drive the speaker for playback.

[0041] In one specific embodiment, the method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones further includes: The average signal energy is continuously monitored. When the average signal energy of a preset number of consecutive detection windows is higher than the dynamic detection threshold, the detection process is terminated. When the average signal energy remains below the dynamic detection threshold and the call connection status flag and the wearing status flag remain unchanged, the periodic reminder process is initiated. Read the current value of the reminder count counter. When the current value of the reminder count counter is less than or equal to the first count threshold, set the waiting time interval to the first duration. When the value of the reminder count counter is greater than the first count threshold but less than or equal to the second count threshold, set the waiting time interval to the second duration. When the value of the reminder count counter is greater than the second count threshold, set the waiting time interval to the third duration. Start the timer to load the waiting time interval and start the countdown. When the timer countdown ends, the reminder count counter is incremented, and step S4 is executed again. When a voice gap is detected, step S5 is executed again.

[0042] Specifically, after entering the detection process, the average signal energy of the main microphone is continuously monitored, and the average signal energy is compared in real time with the dynamic detection threshold set based on the ambient noise floor and energy offset to form a continuous state analysis judgment. When the average signal energy in several consecutive detection windows is found to be higher than the dynamic detection threshold, and the number of consecutive windows reaches a preset threshold, such as 3 or 5 consecutive sampling windows showing that the microphone signal is at normal call strength, the audio routing status is considered to be normal, and the user's voice behavior is correctly captured. At this time, there is no need to continue to execute redundant detection tasks, and the audio routing error detection process is terminated to reduce power consumption and release processing resources. If continuous monitoring finds that the average signal energy is always lower than the dynamic detection threshold, and the call connection status flag remains at a valid high level, it indicates that the SCO link is still in an active state. At the same time, the wearing status flag also remains at a high level, confirming that the user is still wearing the headset for a call. At this time, the system enters the periodic reminder process to send multimodal prompts to the user. After entering the reminder process, the current value of the reminder count counter is read to determine the time interval of the next reminder cycle. When the counter value is less than or equal to the first threshold (e.g., 3 times), the current waiting interval is set to the first duration (e.g., 30 seconds) for high-frequency initial reminders. When the counter value is greater than the first threshold but less than or equal to the second threshold (e.g., 6 times), the waiting interval is adjusted to the second duration (e.g., 45 seconds) to reduce interference frequency. When the number of reminders exceeds the second threshold, it indicates that the user has not yet addressed the issue, and the waiting time is further extended to the third duration (e.g., 60 seconds) to enter a low-frequency but continuous prompting phase. After setting the waiting duration, the timer module loads the currently set waiting interval and starts the countdown process. At the end of the countdown, the reminder count counter is incremented once to record the completed prompt cycle and update the status. Simultaneously, step S4 in the detection process is re-entered to determine whether the user is currently in a voice gap state. If the user is detected to be in a voice gap state, i.e., no voice activity frame appears briefly, step S5 is immediately executed, i.e., a prompt tone signal is generated and played through the local speaker, while simultaneously driving the LED to flash synchronously, completing a complete multimodal prompting operation.

[0043] The process includes: reading the current value of the reminder count counter; setting a first waiting time interval when the current value of the reminder count counter is less than or equal to the first count threshold; setting a second waiting time interval when the value of the reminder count counter is greater than the first count threshold but less than or equal to the second count threshold; setting a third waiting time interval when the value of the reminder count counter is greater than the second count threshold; and further including: reading the current value of the reminder count counter and comparing it with a preset maximum reminder count threshold; setting a corresponding waiting time interval based on the range of the current value of the reminder count counter when the current value of the reminder count counter is less than the maximum reminder count threshold; and stopping the periodic reminder process and switching to passive monitoring when the current value of the reminder count counter reaches the maximum reminder count threshold. In passive monitoring mode, step S2 continues to monitor the average signal energy. When the average signal energy of a preset number of consecutive detection windows is higher than the dynamic detection threshold, the problem is determined to be resolved, the reminder count counter is cleared, and the detection process is terminated. When the call connection status flag or the wearing status flag is low, the detection scenario is determined to be invalid, the reminder count counter is cleared, and the detection mode is exited. A reminder history queue is maintained to store the timestamp of each reminder and the corresponding reminder count counter value. After each execution of step S5, the current system timestamp and the incremented value of the reminder count counter are written into the reminder history queue. When the detection process is terminated, the contents of the reminder history queue are used to statistically analyze the frequency and duration of microphone routing errors.

[0044] The above describes the method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to embodiments of the present invention. The following describes the system for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to embodiments of the present invention. One embodiment of the system for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to embodiments of the present invention includes the following: the system is used to execute the method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones as described in any one of the embodiments.

[0045] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0046] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0047] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones, characterized in that, include: S1. Monitor the hands-free protocol connection status flag and the headset wearing status flag and output a detection start enable signal, then execute step S2. S2. Acquire the audio signal from the built-in main microphone of the earphone, calculate the average signal energy and dynamic detection threshold, and when the average signal energy is lower than the dynamic detection threshold, execute step S3. S3. Calculate the motion frequency domain energy. When the motion frequency domain energy is higher than the motion threshold, confirm that the microphone routing is incorrect and execute step S4. When the motion frequency domain energy is lower than the motion threshold and the duration of low energy is less than the first preset duration, return to execute step S2. S4. When the audio signal meets the silence condition and continues for a second preset duration, a speech gap is detected and step S5 is executed. S5. Generate a digital signal for a prompt tone and drive the speaker to play it. The audio signal is independently routed to the Bluetooth uplink encoder, and the LED is driven to light up.

2. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 1, characterized in that, Step S1 includes: The connection management event of the baseband layer is parsed by the Bluetooth protocol stack to identify the establishment status of the synchronous connection-oriented link. When the synchronous connection-oriented link is in an active state, a call connection status flag is generated. Infrared light is emitted by an infrared sensor and infrared signals reflected from the ear canal are received. The intensity value of the infrared signal is compared with a preset intensity threshold to obtain a wearing status flag. The capacitance value of the sensor electrode is detected by a capacitive sensor, and the capacitance value is compared with a preset capacitance threshold to obtain the wearing status flag bit; When both the call connection status flag and the wearing status flag are high, a detection start enable signal is output and step S2 is executed.

3. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 2, characterized in that, Step S2 includes: The audio signal from the main microphone is amplified by a preamplifier and then input into an analog-to-digital converter for quantization. A detection window is set and the average signal energy within the detection window is calculated. After the call is established, the ambient noise floor reference value of the main microphone is calculated, and the ambient noise floor reference value is added to a preset offset to obtain the dynamic detection threshold. The average signal energy is compared with the dynamic detection threshold. When the average signal energy is lower than the dynamic detection threshold, the detection window is marked as a low-energy window. When a preset number of consecutive detection windows are marked as low-energy windows, a preliminary abnormality flag is output and step S3 is executed.

4. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 3, characterized in that, Step S3 includes: The X-axis acceleration components, Y-axis acceleration components, and Z-axis acceleration components are acquired by a triaxial accelerometer and then synthesized into a net acceleration signal. Apply a Hanning window function to the net acceleration signal and perform a fast Fourier transform to calculate the frequency domain power spectral density; The energy of the frequency domain power spectral density is integrated to obtain the motion frequency domain energy, and the motion frequency domain energy is compared with the motion threshold. When the motion frequency domain energy is higher than the motion threshold, output a routing error confirmation flag and execute step S4. When the motion frequency domain energy is lower than the motion threshold and the duration of low energy is less than the first preset duration, return to execute step S2.

5. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 4, characterized in that, The X-axis acceleration components, Y-axis acceleration components, and Z-axis acceleration components are acquired using a triaxial accelerometer and synthesized into a net acceleration signal, including: The X-axis acceleration component, Y-axis acceleration component, and Z-axis acceleration component are acquired by a triaxial accelerometer and transmitted to the main control chip. The average values ​​of the X-axis acceleration component, the Y-axis acceleration component, and the Z-axis acceleration component within a preset time period are calculated as DC bias values. The corresponding DC bias values ​​are then subtracted from the X-axis acceleration component, the Y-axis acceleration component, and the Z-axis acceleration component to obtain the debiased X-axis acceleration, the debiased Y-axis acceleration, and the debiased Z-axis acceleration. The net acceleration signal is synthesized based on the debiased X-axis acceleration, the debiased Y-axis acceleration, and the debiased Z-axis acceleration.

6. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 5, characterized in that, Step S4 includes: The audio signal is segmented into frames to obtain multiple audio frames, and the short-time frame energy and short-time frame zero-crossing rate are calculated. The energy threshold is calculated based on the environmental noise baseline and the preset energy offset, and the lower threshold and upper threshold of the zero crossing rate are set. When the short-time frame energy is higher than the energy threshold and the short-time frame zero-crossing rate is between the lower threshold and the upper threshold of the zero-crossing rate, the audio frame is marked as a voice activity frame; otherwise, the audio frame is marked as a silent frame. The number of consecutive frames marked as silent is counted and the corresponding silence duration is calculated. When the silence duration reaches the second preset duration, a playback enable signal is output and step S5 is executed.

7. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 6, characterized in that, When the short-time frame energy is higher than the energy threshold and the short-time frame zero-crossing rate is between the lower threshold and the upper threshold of the zero-crossing rate, the audio frame is marked as a speech activity frame; otherwise, the audio frame is marked as a silent frame, including: The short-time frame energy of each audio frame is compared with an energy threshold; When the energy of the short-time frame is higher than the energy threshold, it is determined that the audio frame meets the energy condition; The short-time frame zero-crossing rate of each audio frame is compared with the lower threshold of the zero-crossing rate and the upper threshold of the zero-crossing rate. When the short-time frame zero-crossing rate is greater than the lower threshold of the zero-crossing rate and less than the upper threshold of the zero-crossing rate, it is determined that the audio frame satisfies the zero-crossing rate condition. When both the energy condition and the zero-crossing rate condition are met, the audio frame is marked as a voice activity frame; when neither the energy condition nor the zero-crossing rate condition is met, the audio frame is marked as a silent frame.

8. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 7, characterized in that, Step S5 includes: A sinusoidal carrier signal is generated by a digital signal processor. A trapezoidal amplitude envelope is applied to the sinusoidal carrier signal to obtain a single prompt tone unit. Two consecutive prompt tone units are generated and a silence segment of a preset duration is inserted between them to obtain a digital prompt tone signal. The digital signal of the prompt tone is volume normalized to make its amplitude reach the preset ratio of the call volume, and a local playback path mark is set for the digital signal of the prompt tone in the audio mixer; The digital signal of the prompt tone is weighted and mixed with the Bluetooth downlink audio signal to obtain a mixed audio signal. The audio signal of the main microphone is independently routed to the input of the Bluetooth uplink encoder to avoid entering the uplink. The mixed audio signal is input to the digital-to-analog converter and driven by the power amplifier to play the speaker. Configure the color parameters of the LED and set the flashing sequence to synchronize the flashing start time of the LED with the playback start time of the digital signal of the prompt tone.

9. The method for intelligent detection and multimodal prompting of audio routing errors in wireless headphones according to claim 8, characterized in that, The method for intelligent detection and multimodal prompting of audio routing errors in the wireless headphones also includes: The average signal energy is continuously monitored, and the detection process is terminated when the average signal energy of a preset number of consecutive detection windows is higher than the dynamic detection threshold. When the average signal energy remains below the dynamic detection threshold and the call connection status flag and the wearing status flag remain unchanged, the periodic reminder process is initiated. Read the current value of the reminder count counter. When the current value of the reminder count counter is less than or equal to the first count threshold, set the waiting time interval to the first duration. When the value of the reminder count counter is greater than the first count threshold and less than or equal to the second count threshold, set the waiting time interval to the second duration. When the value of the reminder count counter is greater than the second count threshold, set the waiting time interval to the third duration. Start the timer to load the waiting time interval and start the countdown. When the timer countdown ends, the reminder count counter is incremented, and step S4 is executed again. When a voice gap is detected, step S5 is executed again.

10. A system for intelligent detection and multimodal alerting of audio routing errors in wireless headphones, characterized in that, The method for intelligent detection and multimodal prompting of audio routing errors in a wireless headset as described in any one of claims 1-9.

Images