Singing expiration strength stability training system for breath flow control

Abstract

The invention belongs to the technical field of vocal music training, and provides a singing expiration strength stability training system for breath flow control, comprising: a hardware sensing module for collecting airflow dynamic data, respiratory muscle group electrophysiological data and respiratory kinematics data of a user; the data processing control center is used for receiving and processing the airflow dynamic data, the respiratory muscle group electrophysiological data and the respiratory kinematics data to obtain a processing result; the multi-mode feedback module outputs real-time feedback to the user through visual, auditory and tactile multiple channels based on the processing result; according to the method, heterogeneous data of airflow dynamics, electromyography and respiratory kinematics are synchronously collected and fused, song audios are analyzed in combination with an artificial intelligence algorithm to generate a scene-based target curve, and a personalized training baseline is established based on accurate decision and multi-modal sensory collaborative feedback of a fuzzy logic algorithm; the efficient, accurate and personalized improvement of the stability of the exhaling force of singing is realized.

Description

Technical Field

[0001] This invention belongs to the field of vocal training technology, specifically a singing exhalation strength stability training system for breath flow control. Background Technology

[0002] The core expressive power of singing comes from the precise and stable control of exhalation force, which directly determines the presentation effect of pitch maintenance, timbre uniformity, phrase extension length and dynamic changes in dynamics. It is a key ability barrier for singers to advance from beginner to professional.

[0003] However, traditional singing breathing training mainly relies on the teacher's auditory judgment and experience guidance, making the training process entirely subjective. Teachers can only indirectly infer whether the exhalation force is stable by listening to the singer's voice, and cannot obtain objective data such as airflow output, muscle exertion, and breathing coordination. This makes it difficult to accurately pinpoint the core root cause of unstable exhalation and cannot provide singers with quantifiable evidence of progress. Singers, on the other hand, can only adjust their breathing based on their own perception, lacking a clear understanding of whether their strength is up to standard or whether fluctuations exceed reasonable ranges. Feedback is delayed and vague, and error correction relies entirely on trial and error, resulting in low training efficiency and varying effects from person to person. It heavily depends on the accumulated experience of both teachers and students.

[0004] With the development of sensor technology, a number of singing breathing training devices based on monitoring a single physical parameter have emerged. These devices mostly use airflow sensors to collect data on the velocity or pressure of exhaled airflow, or only monitor the amplitude of chest and abdominal movements through a breathing belt, without involving the fusion analysis of electrophysiological signals of the respiratory muscle groups and multi-dimensional data. Their core flaw lies in the fact that they can only capture the surface phenomenon of expiratory force output, failing to connect it to deeper causes such as muscle exertion patterns and respiratory coordination. For example, when a singer experiences unstable exhalation, the device can only show excessive flow rate fluctuations, but cannot determine whether the problem stems from insufficient abdominal support, abnormal tension in the intercostal muscles, or an imbalance in the phase of chest and abdominal movements. This leads to a training approach that only knows the phenomenon but not the underlying cause, making it difficult to optimize expiratory control ability at its root.

[0005] Furthermore, the training content of existing training equipment is severely disconnected from actual singing scenarios and lacks personalized adaptation capabilities. Most of these devices only preset basic training modes such as long tones and fixed resistance, using uniform target parameters without considering the differentiated requirements of expiratory force for different musical styles and phrases. They also cannot automatically analyze the audio characteristics of the target song and dynamically generate the corresponding target expiratory force curve, resulting in the common problem of stable breath during training but unbalanced breath during performance. On the other hand, these devices do not establish personalized baselines based on individual user physiological characteristics, but instead use uniform assessment thresholds and training standards, ignoring the differences in lung function, muscle strength, and breathing habits among different singers. For example, there are significant differences in basic expiratory ability between novice singers and professional singers. Using the same deviation tolerance and muscle state assessment standards will lead to frustration for novices due to excessive difficulty, and professional singers will not be able to improve effectively due to insufficient difficulty, resulting in extremely poor adaptability. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a singing exhalation strength stability training system for breath flow control, thereby resolving the issues raised in the background art, such as the lack of objective basis for subjective assessment in singing exhalation training, the disconnect between training scenarios and actual singing, and the lack of targeted and simplistic feedback decisions.

[0007] A singing expiratory strength stability training system for breath flow control, comprising:

[0008] The hardware sensing module is used to collect users' airflow dynamics data, respiratory muscle electrophysiological data, and respiratory kinematics data.

[0009] The data processing and control center is used to receive and process the airflow dynamics data, respiratory muscle electrophysiological data, and respiratory kinematics data to obtain processing results.

[0010] A multimodal feedback module is used to output real-time feedback to the user through multiple channels of vision, hearing and touch based on the processing results;

[0011] The data processing control center includes:

[0012] The real-time comparison unit is used to generate a real-time expiratory force curve of the user based on the airflow dynamics data, dynamically compare the real-time expiratory force curve with the target expiratory force curve, and provide auxiliary evaluation of the user's respiratory muscle activity pattern and coordination based on the respiratory muscle electrophysiological data and respiratory kinematic data.

[0013] The feedback decision unit is used to integrate the dynamic comparison results and the auxiliary evaluation results, generate corresponding guidance instructions, and control the multimodal feedback module to output corresponding guidance information.

[0014] Preferably, the hardware sensing module further includes:

[0015] The airflow monitoring unit, by wearing airflow and air pressure sensors over the user's mouth and nose, is used to monitor the flow rate and pressure of exhaled airflow and acquire the airflow dynamics data;

[0016] The electromyography (EMG) activity monitoring unit is attached to the user's expiratory muscle group via a surface EMG sensor array to monitor the level of muscle electrical activity and acquire electrophysiological data of the respiratory muscle group.

[0017] The respiratory motion monitoring unit, which uses a breathing belt sensor wrapped around the user's chest and abdomen, is used to monitor the amplitude of respiratory movements of the chest and abdomen and acquire the respiratory kinematic data.

[0018] The audio acquisition unit is used to acquire the audio signal of the user's singing.

[0019] Preferably, the data processing control center further includes:

[0020] The feature extraction unit is used to extract feature parameters for quantifying expiratory stability from the airflow dynamics data, respiratory muscle electrophysiological data, and respiratory kinematics data. The feature parameters include at least one of the following: standard deviation of airflow velocity, fluctuation range of airflow pressure, area integral difference between the target expiratory force curve and the actual expiratory force curve, ratio of mean abdominal electromyographic activity to mean chest electromyographic activity, and chest-abdominal respiratory movement coordination coefficient.

[0021] The personalized baseline generation unit is used to establish a personalized expiratory force stability baseline based on the user's expiratory data in a calm state.

[0022] The target curve generation unit is used to generate and call the target expiratory force curve.

[0023] Preferably, the target curve generation unit includes:

[0024] A preset curve library is used to store predefined target curves corresponding to different singing techniques;

[0025] The intelligent analysis unit analyzes the audio information of the target song and dynamically generates the corresponding target expiratory force curve based on the length, pitch, and intensity characteristics of the musical phrases.

[0026] Preferably, the multimodal feedback module further includes:

[0027] A visual feedback unit is used to generate target path graphic elements and real-time tracking graphic elements on the display interface, which correspond to the target expiratory force curve and the real-time expiratory force curve, respectively; when the real-time tracking graphic element deviates from the target path graphic element, a visual warning effect is triggered.

[0028] An auditory feedback unit is used to play the target pitch and generate real-time guiding sound effects, wherein the harmony of the real-time guiding sound effects is positively correlated with the stability of the user's exhalation force, and to broadcast voice guidance information generated based on the guiding instructions.

[0029] The tactile feedback unit is used to generate a first steady-state tactile signal when the deviation between the real-time expiratory force curve and the target expiratory force curve is within a preset tolerance range, and to generate a second dynamic warning tactile signal when the deviation between the real-time expiratory force curve and the target expiratory force curve exceeds the preset tolerance range, via an array of tactile generators on the wearable device.

[0030] Preferably, the second dynamic warning tactile signal indicates the direction and degree of deviation of the exhalation force by changing the vibration intensity and frequency of the array tactile generator and activating oscillators at different positions in the array tactile generator.

[0031] Preferably, the auxiliary evaluation performed in the real-time comparison unit includes:

[0032] Based on the characteristics of electromyographic amplitude or frequency changes in the electrophysiological data of the respiratory muscle groups, the activation level and fatigue state of the expiratory muscle groups are assessed.

[0033] The coordination of thoracic and abdominal respiratory movements is assessed based on the phase and amplitude relationship of the chest and abdominal motion signals in the respiratory kinematics data.

[0034] Preferably, the assessment of the activation level and fatigue state of the expiratory muscle group includes: when the electromyographic amplitude of a specific muscle group is continuously higher than a preset threshold, it is determined to be abnormal tension; when the median frequency of the electromyographic signal changes with a preset trend over time, it is determined to be premature fatigue.

[0035] The assessment of the coordination of chest and abdominal breathing movements includes: calculating the phase difference between chest and abdominal movement signals; when the phase difference exceeds a preset range, it is determined to be a movement incoordination.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. This invention constructs a digital model for evaluating the stability of expiratory force in singing by simultaneously collecting and fusing three types of heterogeneous data: airflow dynamics, electromyography, and respiratory kinematics. It also adopts a sing-along mode and uses intelligent algorithms to automatically parse and map the song audio into a target expiratory force curve, thus achieving seamless integration between training content and real singing scenarios.

[0038] 2. This invention integrates three pieces of information—the degree of deviation, muscle state, and coordination—through a feedback decision-making unit, and uses a fuzzy logic algorithm to generate the most targeted and differentiated guidance strategy. Furthermore, it is not simply a collection of various feedback methods, but rather a feedback scheme that complements information and coordinates senses: vision provides a macroscopic trajectory, hearing provides harmony cues and voice guidance, and touch provides private and immediate directional fine-tuning. Together, these three elements constitute an efficient immersive learning environment. Through multimodal feedback collaboration, abstract data deviations are transformed into intuitive perceptions, enabling rapid feedback reception and improved error correction efficiency.

[0039] 3. This invention uses an intelligent sing-along mode to dynamically generate phrase-level target curves based on the target song. The training content directly addresses actual singing needs, and the preset curve library covers basic techniques such as long notes, crescendo and diminuendo, taking into account both basic training and scenario applications. The personalized baseline generation unit models based on the user's calm exhalation data, and both the target curve and evaluation threshold are adapted to individual physiological characteristics, making it suitable for singers of different levels. Attached Figure Description

[0040] Figure 1 This is a block diagram of the singing exhalation strength stability training system of the present invention. Detailed Implementation

[0041] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0042] Example: This invention provides a singing expiratory strength stability training system for breath flow control, including a hardware sensing module, a data processing control center, and a multimodal feedback module; each module is sequentially connected to form a closed-loop training system.

[0043] The hardware sensing module is used to collect airflow dynamics data, respiratory muscle electrophysiological data, respiratory kinematics data, and singing audio signals during the user's singing training process; specifically, it includes:

[0044] ① The airflow monitoring unit uses the SFM3000 high-precision airflow and pressure sensor, which is worn at the user's mouth and nose. The sensor has a range of 0-50L / min (flow rate) and 0-10kPa (pressure), a sampling rate of 100Hz, and an accuracy of ±2%. It collects the flow rate and pressure of exhaled air in real time and outputs airflow dynamics data.

[0045] ② The electromyography (EMG) activity monitoring unit uses a Delsys Trigno wireless surface EMG sensor array, which contains four sensors that are attached to the user's expiratory muscle groups, including the rectus abdominis, transverse abdominis, external intercostal muscles, and scalene muscles. The sampling rate is 2000 Hz and the resolution is 16 bits. It is used to monitor the amplitude and frequency changes of muscle electrical activity and output the electrophysiological data of the respiratory muscle groups.

[0046] ③ The respiratory motion monitoring unit uses an elastic strain gauge breathing belt sensor wrapped around the user's chest and abdomen respectively. It has a sampling rate of 50Hz and a sensitivity of 0.1mm. By sensing changes in the circumference of the chest and abdomen, it acquires data on the amplitude and phase relationship of respiratory motion and outputs respiratory kinematic data.

[0047] ④ The audio acquisition unit uses an omnidirectional condenser microphone, integrated into a mouth and nose mask or wearable device, to acquire the audio signal of the user's singing.

[0048] The aforementioned multi-sensor data streams are analyzed under a unified timeline by using hardware synchronization pulse signals or software timestamp interpolation algorithms, which facilitates subsequent coordinated evaluation.

[0049] Data Processing Control Center: Receives all data transmitted from the hardware sensing module, and generates precise guidance instructions through feature extraction, baseline modeling, target generation, real-time comparison, and decision analysis. Specifically, this includes:

[0050] ① Feature Extraction Unit: Based on the LabVIEW data processing platform, the unit synchronously aligns the collected airflow dynamics data, respiratory muscle electrophysiological data, and respiratory kinematics data, and extracts feature parameters for quantifying expiratory stability. The feature parameters include the standard deviation of airflow velocity, the fluctuation range of airflow pressure, the area integral difference between the target expiratory force curve and the actual expiratory force curve, the ratio of the mean of abdominal electromyographic activity to the mean of chest electromyographic activity, the median frequency of electromyographic signals, the coordination coefficient of thoracic and abdominal respiratory movements, and the phase difference of thoracic and abdominal movement signals.

[0051] ② Personalized Baseline Generation Unit: Upon first use, users complete three sets of steady exhalations in a calm state (each set lasting 10 seconds) under system guidance. Based on the collected airflow dynamics data, electromyography data, and respiratory motion data, the system calculates the mean and standard deviation of each characteristic parameter to establish a user-specific "expiratory force stability baseline" as a benchmark for subsequent training evaluation. In addition to establishing the baseline, the system also guides users to perform device calibration upon first use. For example, it performs zero-point calibration of the airflow sensor in a no-airflow state; ensures good contact between each electrode and the skin, and performs electromyography sensor contact impedance detection when the impedance value is below the set threshold; and guides users to perform a maximum inhalation and exhalation to calibrate the individual maximum range of chest and abdominal movement and to calibrate the tightness of the breathing belt.

[0052] ③ The target curve generation unit has two working modes: preset mode and intelligent follow-along mode, which generates a target expiratory force curve:

[0053] i. Preset curve library: Stores predefined target curves for different singing techniques, including long tone stability curve (constant flow rate, fluctuation ≤5%), crescendo and diminuendo curve (flow rate changes according to linear / exponential law), and staccato interval curve (flow rate starts and stops quickly, peak value is stable).

[0054] ii. Intelligent Analysis Unit: An audio analysis model trained using the Python TensorFlow framework. After the user selects a target song, the system automatically parses the song's audio information, splits the musical phrases using a phrase detection algorithm (based on energy and spectral entropy), and extracts the length, pitch change (based on fundamental frequency F0 detection), and intensity (based on root mean square energy) features of each phrase. Combined with the breathing requirements corresponding to the song's style (bel canto / pop / folk), the system dynamically generates the target expiratory force curve for each phrase.

[0055] ④ Real-time comparison unit: Based on airflow dynamics data, it generates the user's real-time expiratory force curve through moving average filtering, and uses a dynamic time warping algorithm to dynamically compare the real-time expiratory force curve with the target expiratory force curve, calculates the deviation value (deviation rate = |real-time value - target value| / target value × 100%); and provides auxiliary assessment of the user's respiratory muscle activity patterns and coordination based on respiratory muscle electrophysiological data and respiratory kinematic data.

[0056] Among them, the auxiliary assessment is based on the electrophysiological data of the respiratory muscle group and the respiratory kinematic data, and includes two core assessments:

[0057] i. Muscle status assessment: When the electromyography amplitude of a specific muscle group is monitored to be higher than 150% of the personalized baseline (preset threshold) for 3 seconds, it is judged as abnormal tension; when the median frequency of the electromyography signal continues to decrease over time (slope ≤ -0.5Hz / s), it is judged as premature muscle fatigue.

[0058] ii. Respiratory Coordination Assessment: The phase difference between chest and abdominal motion signals is calculated using a cross-correlation algorithm. When the phase difference exceeds ±30° (preset range), it is determined to be an incoordination of chest and abdominal respiratory movements.

[0059] ⑤ The feedback decision unit, based on the deviation values ​​compared in real time (divided into slight deviation ≤5%, moderate deviation 5%-15%, and severe deviation >15%) and the auxiliary evaluation results, uses a fuzzy logic algorithm to generate corresponding guidance instructions, the specific rules of which are as follows:

[0060] i) Slight deviation + normal muscle condition + normal coordination: Maintain the current feedback mode, no additional voice guidance required;

[0061] ii. Moderate deviation (too strong / too weak) + normal muscle condition: Generate voice commands such as reducing exhalation force / increasing abdominal support, and adjust multimodal feedback signals simultaneously;

[0062] iii. Severe deviation + abnormal muscle tension: Generate voice commands such as relaxing intercostal muscles / reducing shoulder exertion, and strengthen tactile warning signals;

[0063] iv. Coordination abnormalities: Generate voice commands to adjust chest and abdominal coordination, and overlay coordination prompts in the visual feedback.

[0064] The multimodal feedback module, based on guidance instructions from the data processing control center, provides real-time feedback to users through visual, auditory, and tactile channels, achieving immersive guidance. Specifically, it includes:

[0065] ① The visual feedback unit generates target path graphic elements and real-time tracking graphic elements on the display interface, corresponding to the target expiratory force curve and the real-time expiratory force curve, respectively. When the real-time tracking graphic element deviates from the target path graphic element, a visual warning effect is triggered. The visual warning effect can be manifested as follows: for slight deviation, the edge of the target path turns yellow; for moderate deviation, the corresponding edge of the deviation side turns red; for severe deviation, the edge flashes and a deviation value prompt pops up. At the same time, an RGB color mapping is used to generate a muscle activity heat map, with red indicating high electromyographic amplitude / abnormal tension, green indicating normal, and blue indicating insufficient activity. The waveform of chest and abdominal movement amplitude and the coordination coefficient value are displayed simultaneously.

[0066] ② Auditory feedback unit, outputting via built-in speaker or wireless headphones; playing target pitch by playing pre-stored pure tones (long tones) or accompaniment / original vocal phrases of the target song (intelligent sing-along mode) for users to sing along; generating real-time guidance sound effects, the harmony of which is positively correlated with the stability of the user's exhalation force; using sine wave synthesis for background harmony, when exhalation force is stable, the background harmony frequency is harmonious with the target pitch; with moderate deviation, the harmony is slightly detuned; with severe deviation, the harmony is distorted and a prompt sound is superimposed; and broadcasting voice guidance information generated based on guidance commands, using a TTS speech synthesis engine to broadcast concise commands generated by the feedback decision unit;

[0067] ③ The haptic feedback unit, through an array of haptic generators on the wearable device, has the following specific feedback logic:

[0068] When the deviation between the real-time expiratory force curve and the target expiratory force curve is within the preset tolerance range, a first steady-state tactile signal is generated.

[0069] When the deviation between the real-time expiratory force curve and the target expiratory force curve exceeds a preset tolerance range, a second steady-state tactile signal is generated. This second dynamic warning tactile signal indicates the direction and degree of expiratory force deviation by changing the vibration intensity and frequency of the array-type tactile generator and activating oscillators at different positions within the generator. The vibration intensity is positively correlated with the deviation rate, and the vibration frequency increases with the duration of the deviation. For example, if the expiratory force is too strong, the right motor is activated; if it is too weak, the left motor is activated. Furthermore, the preset tolerance range and judgment threshold are dynamically adjusted by the system or teacher based on the user's skill level.

[0070] The workflow of this system, taking the training of musical phrases from a target song as an example:

[0071] 1. Users wear devices that integrate various sensors and complete device connection and calibration via an app;

[0072] 2. When a user selects a target song in the app, the system automatically analyzes the song audio, extracts the first training phrase, and dynamically generates the corresponding target expiratory strength curve.

[0073] 3. The system plays the accompaniment (target tone) of the musical phrase, and the user sings along with the accompaniment. The hardware sensing module simultaneously collects airflow dynamics data, electromyography data, respiratory motion data, and singing audio.

[0074] 4. The data processing and control center receives 100 sets of data per second, extracts feature parameters, generates a real-time expiratory force curve, dynamically compares it with the target curve, and simultaneously assesses muscle status and respiratory coordination.

[0075] When the user exhales too forcefully and the rectus abdominis muscle is normal, the right side of the target path in the visual feedback unit turns red, the auditory feedback unit produces a slightly detuned sound effect, the tactile feedback unit activates the right motor, and at the same time the voice broadcasts "reduce the exhalation force".

[0076] 5. After the musical phrase is sung, the system generates a training report, which displays the stability score, deviation rate distribution, muscle exertion assessment and coordination coefficient, and provides improvement suggestions, such as: extending the exhalation holding time and improving the stability of long notes.

[0077] As shown above, when a user begins singing while wearing the device, the hardware sensing module collects physiological signals. The data processing and control center converts these signals into real-time curves and compares them with the target for evaluation. This generates decision commands that drive the multimodal feedback module to provide the user with immediate and precise guidance through visual, auditory, and tactile channels. The user adjusts their state according to the guidance, thus forming an efficient biofeedback loop that rapidly improves the user's control over exhalation force.

[0078] It should be understood that numerous specific implementation decisions can be made during the development of any actual implementation method, and in any engineering or design project. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, the development effort will be a routine task in design, manufacturing, and production without requiring extensive experimentation.

[0079] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A singing expiratory strength stability training system for breath flow control, characterized in that, include: The hardware sensing module is used to collect users' airflow dynamics data, respiratory muscle electrophysiological data, and respiratory kinematics data. The data processing and control center is used to receive and process the airflow dynamics data, respiratory muscle electrophysiological data, and respiratory kinematics data to obtain processing results. A multimodal feedback module is used to output real-time feedback to the user through multiple channels of vision, hearing and touch based on the processing results; The data processing control center includes: The real-time comparison unit is used to generate a real-time expiratory force curve of the user based on the airflow dynamics data, dynamically compare the real-time expiratory force curve with the target expiratory force curve, and provide auxiliary evaluation of the user's respiratory muscle activity pattern and coordination based on the respiratory muscle electrophysiological data and respiratory kinematic data. The feedback decision unit is used to integrate the dynamic comparison results and the auxiliary evaluation results, generate corresponding guidance instructions, and control the multimodal feedback module to output corresponding guidance information.

2. The singing expiratory strength stability training system for breath flow control as described in claim 1, characterized in that, The hardware sensing module also includes: The airflow monitoring unit, by wearing airflow and air pressure sensors over the user's mouth and nose, is used to monitor the flow rate and pressure of exhaled airflow and acquire the airflow dynamics data; The electromyography (EMG) activity monitoring unit is attached to the user's expiratory muscle group via a surface EMG sensor array to monitor the level of muscle electrical activity and acquire electrophysiological data of the respiratory muscle group. The respiratory motion monitoring unit, which uses a breathing belt sensor wrapped around the user's chest and abdomen, is used to monitor the amplitude of respiratory movements of the chest and abdomen and acquire the respiratory kinematic data. The audio acquisition unit is used to acquire the audio signal of the user's singing.

3. The singing expiratory strength stability training system for breath flow control as described in claim 1, characterized in that, The data processing control center also includes: The feature extraction unit is used to extract feature parameters for quantifying expiratory stability from the airflow dynamics data, respiratory muscle electrophysiological data, and respiratory kinematics data. The feature parameters include at least one of the following: standard deviation of airflow velocity, fluctuation range of airflow pressure, area integral difference between the target expiratory force curve and the actual expiratory force curve, ratio of mean abdominal electromyographic activity to mean thoracic electromyographic activity, and thoracic and abdominal respiratory movement coordination coefficient. The personalized baseline generation unit is used to establish a personalized expiratory force stability baseline based on the user's expiratory data in a calm state. The target curve generation unit is used to generate and call the target expiratory force curve.

4. The singing expiratory strength stability training system for breath flow control as described in claim 3, characterized in that, The target curve generation unit includes: A preset curve library is used to store predefined target curves corresponding to different singing techniques; The intelligent analysis unit analyzes the audio information of the target song and dynamically generates the corresponding target expiratory force curve based on the length, pitch, and intensity characteristics of the musical phrases.

5. The singing expiratory strength stability training system for breath flow control as described in claim 1, characterized in that, The multimodal feedback module also includes: A visual feedback unit is used to generate target path graphic elements and real-time tracking graphic elements on the display interface, which correspond to the target expiratory force curve and the real-time expiratory force curve, respectively; when the real-time tracking graphic element deviates from the target path graphic element, a visual warning effect is triggered. An auditory feedback unit is used to play the target pitch and generate real-time guiding sound effects, wherein the harmony of the real-time guiding sound effects is positively correlated with the stability of the user's exhalation force, and to broadcast voice guidance information generated based on the guiding instructions. The tactile feedback unit is used to generate a first steady-state tactile signal when the deviation between the real-time expiratory force curve and the target expiratory force curve is within a preset tolerance range, and to generate a second dynamic warning tactile signal when the deviation between the real-time expiratory force curve and the target expiratory force curve exceeds the preset tolerance range, via an array of tactile generators on the wearable device.

6. The singing expiratory strength stability training system for breath flow control as described in claim 5, characterized in that: The second dynamic warning tactile signal indicates the direction and degree of deviation of the exhalation force by changing the vibration intensity and frequency of the array tactile generator and activating oscillators at different positions in the array tactile generator.

7. The singing expiratory strength stability training system for breath flow control as described in claim 1, characterized in that, The auxiliary evaluation performed in the real-time comparison unit includes: Based on the characteristics of electromyographic amplitude or frequency changes in the electrophysiological data of the respiratory muscle groups, the activation level and fatigue state of the expiratory muscle groups are assessed. The coordination of thoracic and abdominal respiratory movements is assessed based on the phase and amplitude relationship of the chest and abdominal motion signals in the respiratory kinematics data.

8. The singing expiratory strength stability training system for breath flow control as described in claim 7, characterized in that, The assessment of the activation level and fatigue state of the expiratory muscle groups includes: when the electromyographic amplitude of a specific muscle group is continuously higher than a preset threshold, it is determined to be abnormal tension; when the median frequency of the electromyographic signal changes with a preset trend over time, it is determined to be premature fatigue. The assessment of the coordination of chest and abdominal breathing movements includes: calculating the phase difference between chest and abdominal movement signals; when the phase difference exceeds a preset range, it is determined to be a movement incoordination.

Images