Wireless microphone intelligent control system and method

Abstract

The present invention relates to the field of audio processing technology, and discloses a wireless microphone intelligent control system and method, wherein a wireless microphone intelligent control method includes: collecting multi-point acoustic data through a distributed microphone network to construct a spatial acoustic topology model; constructing an adaptive notch filter group to achieve howling suppression based on the topology analysis results; monitoring environmental changes in real time through a distributed acoustic sensor network to dynamically optimize suppression parameters; integrating a lightweight convolutional neural network on the microphone device to identify acoustic events; and adopting a layered edge computing architecture to achieve low-latency acoustic event processing and howling prevention. The present invention effectively solves the technical problems of unstable howling suppression, weak environmental adaptability and lack of spatial perception in complex environments.

Description

Technical Field

[0001] The present invention relates to the technical field of audio processing, and more particularly to a wireless microphone intelligent control system and method. Background Art

[0002] Wireless microphone systems are widely used in large conference centers, multi-function halls, lecture halls, commercial venues and other scenarios, providing speakers and performers with flexible and convenient sound pickup solutions.

[0003] However, in complex acoustic environments, traditional wireless microphone systems face three major technical challenges:

[0004] Complex and variable acoustic feedback paths make howling difficult to control. In large spaces, sound can travel from the speakers back to the microphones via multiple paths, forming an acoustic feedback loop. Traditional systems struggle to accurately model the interactions of these complex feedback paths and multi-point howling. Using only single-frequency detection and suppression methods results in unstable howling suppression and limited gain margin.

[0005] Dynamic spatial configurations lead to frequent changes in acoustic characteristics. Factors such as wall movement, furniture rearrangement, and changes in audience density in multi-purpose venues can significantly impact the acoustic environment. Traditional systems require frequent recalibration and struggle to adapt to these changes in real time. This requires re-identifying acoustic feedback paths whenever the environment changes, resulting in reduced system availability.

[0006] Traditional systems lack the ability to collaborate across multiple devices and understand spatial context. Most systems process signals at the single-device level, unable to effectively integrate information from multiple pickup points. They lack overall acoustic spatial awareness, making it difficult to accurately identify and locate acoustic anomalies in complex environments. Their functionality is limited to basic sound pickup.

[0007] The above technical problems limit the application effect and function expansion of wireless microphone systems in complex acoustic environments. An intelligent wireless microphone control system is needed that can accurately model complex acoustic feedback paths, quickly adapt to changes in dynamic acoustic environments, and have multi-device collaborative perception capabilities to improve the system's stability, adaptability, and functional diversity. Summary of the Invention

[0008] The present invention provides a wireless microphone intelligent control system and method, which solves the technical problems of wireless microphone systems in related arts such as unstable howling suppression, weak environmental adaptability and lack of spatial perception ability in complex acoustic environments.

[0009] The present invention provides a wireless microphone intelligent control method, comprising the following steps:

[0010] Multiple wireless microphones and fixed microphones distributed in space are used to collect the first type of multi-point acoustic data, construct a spatial acoustic transfer function matrix, and use singular value decomposition to analyze the characteristics of the acoustic feedback path to form a spatial acoustic topology model.

[0011] Based on the results of spatial acoustic topology analysis, an adaptive notch filter bank is constructed and optimized to achieve precise howling suppression.

[0012] The second type of acoustic data is collected in real time through a distributed acoustic sensor network to construct a three-dimensional sound field distribution, detect changes in the spatial acoustic environment, and generate optimized suppression parameters for the changed area;

[0013] A lightweight convolutional neural network is integrated into each microphone device, combined with multi-microphone information fusion technology to achieve the recognition and location of acoustic anomalies;

[0014] A layered edge computing architecture and energy-saving communication strategy are used to process acoustic data in layers, achieving low-latency acoustic event processing and howling prevention.

[0015] In a preferred embodiment, the step of constructing a spatial acoustic transfer function matrix includes:

[0016] During the system initialization phase, the excitation signal is output through each speaker in turn;

[0017] The received signal is recorded simultaneously by all microphones;

[0018] For each speaker-microphone combination, the transfer function is calculated to form a spatial acoustic transfer function matrix.

[0019] In a preferred embodiment, the step of constructing the adaptive notch filter bank includes:

[0020] Based on the potential howling frequency points and their risk indexes identified by acoustic topology analysis, a second-order notch filter unit is constructed for each potential howling frequency point;

[0021] According to the howling risk index, the attenuation factor of each notch filter is optimized;

[0022] Construct a multi-stage linkage notch filter structure to deal with the mutual influence between howling frequencies;

[0023] Predictive howling suppression is achieved, suppressing howling before it is fully formed.

[0024] In a preferred embodiment, the steps of constructing the distributed acoustic sensor network include:

[0025] Use the fixed microphone on the charging station as the reference node and the wireless microphone as the dynamic node;

[0026] Each wireless microphone is equipped with an inertial measurement unit to sense its own motion state;

[0027] estimating the relative position between the wireless microphone and the reference node based on the time of flight or the received signal strength indication;

[0028] Data transmission and synchronization between nodes are achieved through wireless communication networks.

[0029] In a preferred embodiment, the step of detecting changes in the spatial acoustic environment includes:

[0030] Periodically updating the spatial acoustic transfer function matrix and calculating the rate of change of the spatial acoustic transfer function matrix and the matrix at the previous moment;

[0031] Use singular value decomposition to track the direction of change of acoustic properties;

[0032] Based on the components of the singular vectors, the spatial region where the acoustic environment changes is determined;

[0033] When the rate of change exceeds a preset threshold, the adaptive update mechanism is triggered.

[0034] In a preferred embodiment, the structure of the lightweight convolutional neural network includes:

[0035] Input layer, used to receive feature vectors;

[0036] Two convolutional layers and a pooling layer to extract acoustic features;

[0037] Two fully connected layers for classifying acoustic events;

[0038] Weight pruning technology and fixed-point quantization are used to reduce the parameter scale, and the overall model size is controlled within 500KB.

[0039] In a preferred embodiment, the multi-microphone information fusion technology includes:

[0040] Calculate the spatial position of the sound source based on the time difference and signal strength ratio;

[0041] Based on the principle of temporal and spatial consistency, it associates the same type of events identified by multiple devices;

[0042] Use a confidence-weighted voting mechanism to integrate event classification results from multiple devices;

[0043] Construct a spatial heat map of acoustic events and identify high-activity areas as points of interest for potential abnormal events.

[0044] In a preferred embodiment, the layered edge computing architecture includes:

[0045] The device layer, consisting of a low-power microprocessor built into the wireless microphone, is responsible for signal acquisition and preliminary feature extraction;

[0046] The edge aggregation layer, consisting of edge servers integrated with charging stations, is responsible for multi-device data integration and model updates;

[0047] The cloud layer is used for long-term data storage and model optimization;

[0048] A hierarchical message queue mechanism and an adaptive task migration algorithm are used between layers for communication and resource optimization.

[0049] In a preferred embodiment, the energy-saving communication strategy includes:

[0050] Adopting an event-driven communication mechanism, data upload is triggered only when a potential event is detected;

[0051] Dynamically adjust data collection frequency according to environmental complexity;

[0052] Implement data compression transmission to transmit only valid features instead of original audio data;

[0053] Establish an optimized wireless communication protocol between the wireless microphone and the charging base, supporting multi-device synchronization and low-power transmission.

[0054] In a preferred embodiment, a wireless microphone intelligent control system is used to execute a wireless microphone intelligent control method, including:

[0055] Multiple wireless microphone modules, each equipped with a microphone array, a signal processing unit and a wireless communication module;

[0056] Charging cradle module, integrating fixed microphone array and edge computing server;

[0057] Spatial acoustic topology analysis module, used to construct and analyze spatial acoustic transfer function matrix;

[0058] Adaptive howling suppression module, used to achieve precise howling control based on topology analysis;

[0059] Distributed acoustic event detection module, used to realize multi-device collaborative acoustic anomaly event identification and location.

[0060] The beneficial effects of the present invention are:

[0061] The present invention effectively solves the howling problem in complex acoustic environments by constructing a spatial acoustic topology model and an adaptive notch filter bank, and achieves accurate howling prediction and suppression.

[0062] The present invention adopts the adaptive technology of partitioned acoustic characteristics to enhance the environmental adaptability of the system;

[0063] This invention utilizes distributed collaborative acoustic event detection technology to expand the functionality of wireless microphone systems from simple human-computer interaction tools to an environmental safety monitoring platform. In typical office or conference environments, the system improves the accuracy of acoustic event classification and event location, automatically recording audio evidence of unusual acoustic events and providing precise location information.

[0064] This invention utilizes a layered edge computing architecture to achieve low-power, low-latency distributed intelligent processing. The system rationally distributes tasks across different layers, extending the battery life of a single wireless microphone while reducing the latency of howling and acoustic event detection, meeting the needs of real-time interaction.

[0065] The present invention also achieves device collaboration and system scalability. Multiple microphone devices form a collaborative network. As the number of devices increases, the system's acoustic spatial perception and event localization accuracy continue to improve. It also supports the dynamic addition or removal of devices without complex reconfiguration. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a flow chart of a wireless microphone intelligent control method of the present invention. DETAILED DESCRIPTION

[0067] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. Furthermore, features described in some examples may be combined in other examples.

[0068] At least one embodiment of the present invention discloses a wireless microphone intelligent control method, such as Figure 1 As shown, the following steps are included:

[0069] Step 1: Collect first-type multi-point acoustic data using multiple wireless microphones and fixed microphones distributed in space, construct a spatial acoustic transfer function matrix, analyze the acoustic feedback path characteristics using singular value decomposition, and form a spatial acoustic topology model;

[0070] The specific implementation includes the following sub-steps:

[0071] Step 1.1, multi-point acoustic data collection;

[0072] Using wireless microphone arrays distributed in the space and condenser microphones on the charging base, acoustic signals are collected simultaneously at multiple key locations in the space.

[0073] The acquisition system includes:

[0074] Wireless microphones, each equipped with a high-precision ADC (analog-to-digital converter) with a sampling rate of 48kHz and a quantization accuracy of 24 bits, are used to capture sound signals at different locations in space;

[0075] A condenser microphone integrated into the charging cradle, with the same sampling rate as the wireless microphone, is used to provide acoustic data at a fixed reference point;

[0076] The data recorded at each acquisition point includes time domain signals and the corresponding frequency domain representation ,in Indicates the index number of the collection point, the value range is .

[0077] Step 1.2, acoustic transfer function matrix construction;

[0078] Construct a spatial acoustic transfer function matrix based on acoustic data collected at multiple points .

[0079] This matrix describes the acoustic propagation characteristics between any two points in space, and its mathematical expression is:

[0080] ;

[0081] in, Indicates from point Arrive The acoustic transfer function, represents the complex variable in the frequency domain, and are the indexes of the signal source and receiving point respectively, Indicates the total number of wireless microphones in the system, Represents the total number of fixed microphones on the charging base. The dimension of the entire matrix is , which includes the acoustic transfer relationship between all microphones and speakers.

[0082] The transfer function is estimated by:

[0083] During the system initialization phase, an excitation signal (such as white noise or a frequency sweep signal) is output through each speaker in turn;

[0084] The received signal is recorded by all microphones simultaneously;

[0085] For each speaker-microphone combination, calculate the transfer function:

[0086] ;

[0087] in, Indicates that the to the microphone The acoustic transfer function describes the characteristics of the sound signal propagating from the loudspeaker to the microphone in space; Indicates the speaker Representation of the output excitation signal in the frequency domain; Indicates microphone Representation of the received response signal in the frequency domain; is a complex variable, representing a complex variable in the frequency domain, used to describe the frequency characteristics of the signal; is the index number of the speaker, indicating a specific speaker position; The index number of the microphone, indicating a specific microphone position.

[0088] Step 1.3, singular value decomposition analysis of acoustic feedback path characteristics;

[0089] The acoustic transfer function matrix constructed Perform singular value decomposition (SVD) to analyze the main characteristics of the acoustic feedback path:

[0090] ;

[0091] in, is the acoustic transfer function matrix, which describes the acoustic propagation characteristics between points in space; is the left singular vector matrix whose column vectors represent the main acoustic modes in the microphone reception space; is the right singular vector matrix whose column vectors represent the main spatial modes of sound emitted by the loudspeaker; It is a singular value diagonal matrix, and the elements on the diagonal represent the energy of the corresponding acoustic propagation path; Representation matrix The conjugate transpose of Each complex element in is conjugated and then the matrix is ​​transposed; the singular value decomposition decomposes the complex acoustic transfer function matrix into orthogonal acoustic modes, which facilitates the identification of the main acoustic feedback path.

[0092] By analyzing:

[0093] The size of the singular value reflects the energy contribution of different acoustic feedback paths;

[0094] The singular vector corresponding to the maximum singular value represents the spatial distribution of the main acoustic feedback path;

[0095] The distribution characteristics of singular values ​​reveal the complexity of the acoustic environment.

[0096] Step 1.4, extraction of key features of feedback path;

[0097] Based on the singular value decomposition results, the key features of the acoustic feedback path are extracted:

[0098] Before extraction The acoustic feedback path corresponding to the largest singular value, where The value is determined adaptively according to the cumulative contribution rate, and the minimum cumulative contribution of 90% is taken. value;

[0099] For each main path, extract its frequency response characteristics, focusing on analyzing its amplitude-frequency characteristics and phase-frequency characteristics;

[0100] Identify potential howling frequencies, i.e., frequencies where the transfer function gain approaches or exceeds unity gain and the phase approaches an integer multiple of 2π;

[0101] Calculate the howling risk index for each potential howling frequency point:

[0102] ;

[0103] in, Indicates the The howling risk index of each frequency point. The larger the value, the higher the possibility of howling at that frequency point. Indicates the The gain value of a frequency point reflects the amplification factor of the acoustic transfer function at that frequency point; Indicates the The phase value of a frequency point describes the phase characteristics of the acoustic signal at that frequency point; Indicates that the phase value is normalized to During the cycle; Indicates rounding the normalized phase value to the nearest integer value; Indicates the phase and the nearest integer multiple the degree of deviation; Indicates that the phase is close to an integer multiple The closer the value is to 1, the easier it is to generate howling.

[0104] The output spatial acoustic topology model contains the following core information:

[0105] Complete acoustic transfer function matrix ;

[0106] Characterization of the main acoustic feedback paths;

[0107] List of potential howling frequencies and their risk index;

[0108] Spatial distribution characteristics of acoustic feedback paths.

[0109] This model provides a basis for subsequent adaptive howling suppression and spatial acoustic environment change detection.

[0110] Step 2: Based on the results of spatial acoustic topology analysis, an adaptive notch filter bank is constructed and optimized to achieve precise howling suppression.

[0111] The specific implementation includes the following sub-steps:

[0112] Step 2.1, adaptive notch filter bank construction;

[0113] Based on the potential howling frequency points and their risk indexes identified in step 1.4, an adaptive notch filter bank is constructed. The transfer function of the filter bank is expressed as:

[0114] ;

[0115] in, is the total transfer function of the adaptive notch filter bank, which represents the processing response of the system to the input signal; Indicates that all notch filter units are cascaded and multiplied to combine the effects of multiple filters; The number of potential howling frequencies that need to be suppressed determines the number of notch filter units included in the filter bank; For the The normalized angular frequency of the potential howling frequency point, , is the howling frequency, is the sampling frequency; is the unit delay operator, which means the signal is delayed by one sampling period; is a double-unit delay operator, indicating that the signal is delayed by two sampling periods; Determines the center frequency position of the notch filter, that is, the target frequency of howling suppression; To control the attenuation factor of the notch filter bandwidth, , The closer it is to 1, the narrower the notch filter bandwidth is, and the smaller the impact on useful signals near the target frequency is; Indicates The zero point at a frequency completely eliminates the signal at that frequency; Indicates proximity The frequency pole controls the bandwidth and depth of the notch; each potential howling frequency point corresponds to a second-order notch filter unit, and the second-order structure can accurately control the suppression effect of a specific frequency.

[0116] Step 2.2, filter parameter optimization based on risk index;

[0117] According to the howling risk index , optimize the parameters of each notch filter:

[0118] Howling Risk Index The higher the value, the corresponding attenuation factor The closer the setting is to 1 (narrowband notch), the less impact on the useful signal;

[0119] Risk index below the threshold The frequency point where notch filtering is not applied to avoid unnecessary signal distortion is:

[0120] Indicates the system's preset risk index threshold. Frequencies below this threshold are considered to have a low likelihood of howling and do not require suppression.

[0121] Attenuation Factor and risk index The relationship can be expressed as:

[0122] ;

[0123] in, Indicates the The attenuation factor of a notch filter determines the bandwidth characteristics of the filter;

[0124] Indicates the preset minimum attenuation factor, corresponding to a wider filter bandwidth, used for frequencies with lower risks;

[0125] Indicates the preset maximum attenuation factor, corresponding to a narrower filter bandwidth, used for frequencies with higher risks;

[0126] Indicates the The current howling risk index of each frequency point; Indicates the maximum risk index value among all frequency points detected by the system; Indicates normalizing the risk index to the interval [0, 1] for linear mapping of the attenuation factor.

[0127] Step 2.3, multi-stage linkage notch filter structure is realized;

[0128] Construct a multi-stage linkage notch filter structure to deal with the mutual influence between howling frequencies:

[0129] The filter group is divided into multiple cascaded sub-filter groups, each group is responsible for processing the howling in a specific frequency band;

[0130] Establish a linkage mechanism between notch filters: When howling at one frequency point is suppressed, it may cause changes in the acoustic characteristics of other frequencies. The system updates the risk index of each frequency point in real time.

[0131] Implement gradual adjustment of filter parameters to avoid audio distortion caused by sudden parameter changes.

[0132] Step 2.4, predictive howling suppression algorithm;

[0133] Predictive howling suppression is achieved, suppressing howling before it is fully formed:

[0134] Based on the temporal dynamic changes of the spatial acoustic topology model, the howling development trend index is calculated:

[0135] ;

[0136] in, Represents the howling development trend index, which is used to quantify the rate of change of howling risk; It represents the derivative operation with respect to time, and is used to calculate the rate of change of the risk index over time; Indicates the The howling risk index of a frequency point reflects the possibility of howling at that frequency point; When it is a positive value, it means that the risk of howling is increasing. The larger the value, the faster the howling is formed. A negative value indicates that the risk of howling is decreasing and the system may be stabilizing naturally. This index is used for predictive howling control, enabling the system to take suppression measures before howling is fully formed.

[0137] for For frequencies greater than the preset threshold, the corresponding notch filter is activated in advance;

[0138] The suppression strength is increased, that is, as As the value of increases, the depth of the notch gradually increases.

[0139] The filter can be implemented using an efficient IIR (Infinite Impulse Response) structure, whose difference equation is expressed as:

[0140] ;

[0141] in, Indicates the current time The filter output signal; Indicates the current time Input signal; Represents the input signal at the previous moment, that is, the input delayed by one sampling period; Represents the input signal at the first two moments, that is, the input delayed by two sampling periods; Represents the output signal at the previous moment, that is, the output delayed by one sampling period; Represents the output signal at the first two moments, that is, the output delayed by two sampling periods; 、 、 Represent the feedforward coefficients of the filter at the current moment, the previous moment, and the previous two moments, respectively, and are used to process the current and historical input signals; 、 Respectively represent the feedback coefficients of the filter at the previous moment and the previous two moments, which are used to process the historical output signal;

[0142] The coefficients correspond to the filter transfer function parameters:

[0143] : Weight coefficient of the current input signal;

[0144] : Weight coefficient of the input signal at the previous moment, where is the normalized angular frequency of the target howling frequency point;

[0145] : Weight coefficient of the input signal at the first two moments;

[0146] : Weight coefficient of the output signal at the previous moment, where To control the attenuation factor of the notch filter bandwidth;

[0147] : The weight coefficient of the output signal in the first two moments, which is determined by the attenuation factor The square of is determined.

[0148] The final output is a signal processed by adaptive howling suppression filtering, which has minimal howling risk while maintaining the sound quality and clarity of the original signal.

[0149] Step 3: Collect the second type of acoustic data in real time through a distributed acoustic sensor network, construct a three-dimensional sound field distribution, detect changes in the spatial acoustic environment, and generate optimized suppression parameters for the changed area;

[0150] The specific implementation includes the following sub-steps:

[0151] Step 3.1, construction of distributed acoustic sensor network;

[0152] Build a distributed acoustic sensor network consisting of wireless and fixed microphones:

[0153] Use M fixed microphones on the charging base as reference nodes, whose positions are fixed and known;

[0154] Use N wireless microphones as dynamic nodes, communicating with the reference node via a wireless communication network (such as low-power Bluetooth 5.0 or a proprietary wireless protocol);

[0155] Each wireless microphone is equipped with an IMU (Inertial Measurement Unit), which includes a three-axis accelerometer and a three-axis gyroscope to sense its own motion state;

[0156] Estimate the relative position between the wireless microphone and the reference node based on TOF (Time of Flight) or RSSI (Received Signal Strength Indication).

[0157] The specific implementation of this distributed acoustic sensor network includes:

[0158] Sensor node type and configuration:

[0159] Fixed reference node: Each charging station integrates 4 to 8 omnidirectional condenser microphones, with a sampling rate of 48kHz and a signal-to-noise ratio of >70dB, in a cost-optimized configuration;

[0160] Dynamic mobile node: Each wireless microphone integrates 2 to 4 directional condenser microphones, with a sampling rate of 48kHz, a signal-to-noise ratio of >65dB, and a power-optimized configuration;

[0161] Each node is equipped with an independent signal processing unit, including a low-power DSP and storage buffer, which can independently complete preliminary data processing.

[0162] Network topology and communication protocol:

[0163] Using a star network topology, the charging station is the central node and all wireless microphones are sub-nodes;

[0164] Uses the improved low-power Bluetooth 5.0 protocol, supporting broadcast, multicast, and point-to-point communication modes;

[0165] Design a dynamic time slot allocation algorithm to dynamically adjust the communication bandwidth according to the transmission priority;

[0166] Implement communication error detection and automatic error correction mechanism to ensure communication reliability in complex electromagnetic environments.

[0167] Node synchronization and positioning method:

[0168] Periodically send synchronization signals, and each node calculates and corrects the clock deviation based on the received timestamp;

[0169] The spatial position of the wireless microphone is determined by combining TOF, RSSI, and signal angle estimation using the trilateration principle.

[0170] Use IMU data for trajectory tracking and inertial navigation to maintain positioning accuracy during communication interruptions;

[0171] The Kalman filter algorithm is applied to fuse multi-source positioning data to improve positioning stability and accuracy.

[0172] The distributed acoustic sensor network exhibits the following characteristics in practical applications:

[0173] In a typical 100 square meter conference room environment, the positioning accuracy can reach ±0.3 meters;

[0174] Supports up to 16 wireless microphones to be connected simultaneously, and can be flexibly expanded;

[0175] The location information update rate of each node is 10Hz, meeting the real-time tracking requirements;

[0176] The network synchronization accuracy is better than 1 millisecond, ensuring the time consistency of multi-point acoustic data;

[0177] The power consumption of a single wireless microphone node is less than 50mW (standby state) and 150mW (active state).

[0178] Step 3.2, reconstructing the three-dimensional sound field distribution by acoustic holography;

[0179] Using multi-point acoustic data collected by a distributed sensor network, acoustic holography technology is used to reconstruct the three-dimensional sound field distribution in space:

[0180] The spatial Fourier transform algorithm is used to convert discrete acoustic sampling point data into continuous sound field distribution;

[0181] The mathematical model of sound field reconstruction is expressed as:

[0182] ;

[0183] in, Indicates a point The sound pressure distribution at is the position vector of any point in space; From the source Arrive Green's function, which represents the sound wave from the source point Propagate to the receiving point transfer characteristics; It represents the normal vibration velocity at the source point, that is, the vibration velocity of the sound source surface perpendicular to the surface direction; is the sound source surface, indicating the integration area; is a small area element on the surface of the sound source; the entire integral represents all the sound source surface points to any point in space The superposition contribution of the sound pressure at ;

[0184] For discrete sampling points, the above formula can be approximated as:

[0185] ;

[0186] in, Indicates a point Sound pressure distribution at ; Indicates the number of wireless microphones; Indicates the number of fixed microphones; Indicates the microphone index, from 1 to the total number of microphones; is the acoustic holography weight coefficient, which is used to adjust the contribution weights of different microphone sampling points; From the microphone position To any point in space Green's function, describing the propagation characteristics of sound waves; For sampling points The sound pressure value measured at the microphone The strength of the collected sound signal.

[0187] Step 3.3, real-time detection of acoustic environment changes;

[0188] Based on continuously collected acoustic data, real-time detection of changes in the spatial acoustic environment:

[0189] Periodically update the spatial acoustic transfer function matrix , calculate the rate of change of the matrix with respect to the previous moment:

[0190] ;

[0191] in, Represents the Frobenius norm of a matrix, which is used to calculate the square root of the sum of the squares of the matrix elements and measure the overall difference between two matrices; Indicates the current time The spatial acoustic transfer function matrix of ; Indicates the previous moment The spatial acoustic transfer function matrix of ; is the sampling time interval, i.e., the time difference between two acoustic characteristic measurements; A quantitative indicator of the change in the acoustic environment at the current moment. A larger value indicates a more significant change in the environment.

[0192] Use singular value decomposition to track the main directions of change in acoustic properties:

[0193] ;

[0194] in, Indicates the current time The spatial acoustic transfer function matrix of ; Indicates the previous moment The spatial acoustic transfer function matrix of ; It is an orthogonal matrix, whose column vectors represent the spatial distribution characteristics of the acoustic environment changes; It is a diagonal matrix, and the elements on the diagonal are singular values, indicating the strength of each main change direction; yes The conjugate transposed matrix of , whose row vectors represent the frequency distribution characteristics of the acoustic environment changes. The main column vectors and The larger singular values ​​in can determine the spatial region and frequency range where the acoustic environment changes most significantly.

[0195] based on and The main components of , determine the spatial region where the acoustic environment changes;

[0196] When the rate of change Exceeding the preset threshold , triggers the adaptive update mechanism.

[0197] Step 3.4, regionalized adaptive parameter optimization;

[0198] Dynamically generate optimized suppression parameters for detected acoustic change areas:

[0199] Divide the space into acoustic regions, and the acoustic characteristics in each region are approximately the same;

[0200] For each area , calculate its acoustic characteristic change index:

[0201] ;

[0202] in, Indicates the region index number, ranging from 1 to ; A quantitative indicator that represents the degree of change in the acoustic characteristics of the area. A larger value indicates a more significant change. Indicates the area Sum the transfer functions of all microphone-speaker pairs in the Indicates microphone and speakers A pair of combinations; Indicates area The set of all microphone-speaker pairs contained within; Indicates the current time microphone and speakers Acoustic transfer function between Indicates the previous moment The same transfer function of It represents the absolute difference between the transfer functions at two moments in time and is used to quantify the degree of change in the acoustic characteristics of this particular microphone-speaker pair;

[0203] According to the change index , assigning different adaptive parameters to each region:

[0204] High-variability areas: Update the acoustic topology model more frequently and use a more aggressive howl prediction threshold;

[0205] Low-change areas: Reduce the update frequency, maintain the existing suppression parameters, and reduce the computational burden;

[0206] A smooth transition mechanism is introduced to prevent parameter mutations at region boundaries.

[0207] This zoned adaptive system automatically adjusts system parameters to maintain efficient howling suppression performance when spatial acoustic conditions change (such as occupant movement, door and window openings, and spatial reconfiguration), eliminating the need for recalibration. This significantly reduces computing resource consumption while allowing the system to focus on areas of the acoustic environment that are truly changing, improving response speed and algorithm efficiency.

[0208] Step 4: Integrate a lightweight convolutional neural network on each microphone device and combine it with multi-microphone information fusion technology to realize the recognition and location of acoustic abnormal events;

[0209] The specific implementation includes the following sub-steps:

[0210] Step 4.1, single-device lightweight acoustic event detection;

[0211] Integrate a lightweight acoustic event detector into each wireless microphone device:

[0212] Preprocess the audio signal, including framing, windowing, and feature extraction. Each frame is 20ms long and the frame shift is 10ms.

[0213] Extract Mel-Frequency Cepstral Coefficients (MFCC), spectral centroid, spectral flux, zero crossing rate and other acoustic features to form a feature vector ;

[0214] Use a lightweight convolutional neural network to analyze the feature vector. The network structure is:

[0215] Input layer: receives feature vector , with dimensions of 128× ,in is the number of features;

[0216] Convolutional layer 1: 32 Convolution kernel, ReLU activation function, step size of 1, padding of 1, output feature map dimension of 128× ×32;

[0217] Pooling layer 1: Max pooling, with a stride of 2, reduces the output feature map dimension by half to 64× ( / 2)×32;

[0218] Convolutional layer 2: 64 Convolution kernel, ReLU activation function, step size of 1, padding of 1, output feature map dimension of 64×( / 2)×64;

[0219] Pooling layer 2: Max pooling, with a stride of 2, reduces the output feature map dimension by half to 32× ( / 4)×64;

[0220] Fully connected layer 1: 128 neurons, ReLU activation;

[0221] Fully connected layer 2: output layer, neurons ( is the number of acoustic event categories to be detected), Softmax activation;

[0222] The training method of this lightweight convolutional neural network:

[0223] Supervised learning using annotated acoustic event datasets;

[0224] The cross entropy loss function is used to evaluate the difference between the model prediction and the true label;

[0225] The Adam optimizer was used to adjust the network parameters. The initial learning rate was set to 0.001, and the learning rate was reduced to 0.1 every 50 epochs.

[0226] Use Dropout technology (with a probability of 0.5) to prevent overfitting;

[0227] Use batch normalization techniques to improve training stability and convergence speed;

[0228] Model compression and optimization methods:

[0229] Use weight pruning technology to remove connections that have little contribution to the output, achieving a compression rate of 40%;

[0230] Use 8-bit fixed-point quantization to convert floating-point weights into integers to reduce storage requirements;

[0231] Use knowledge distillation technology to guide lightweight model learning with large pre-trained models;

[0232] In specific applications, specific classifiers are selectively activated according to scene characteristics to reduce the amount of computation.

[0233] The network output can be expressed as:

[0234] ;

[0235] in, represents the acoustic feature vector of the input; Represents the processing result of the convolutional neural network on the input features; Represents the weight matrix of the first fully connected layer; Represents the bias vector of the first fully connected layer; Represents the rectified linear unit activation function, which is used to introduce nonlinear characteristics; Represents the weight matrix of the second fully connected layer (output layer); Represents the bias vector of the second fully connected layer; represents the normalized exponential function, which converts the output into a probability distribution; Represents the final classification result, that is, the probability distribution of each acoustic event category.

[0236] The system can identify 15 acoustic event categories, including: human speech; coughing; applause; laughter; knocking; keyboard tapping; door opening and closing; chair moving; shuffling of papers; glass breaking; mobile phone ringing; alarms; appliance humming; unusual screams; and background music. These categories cover both normal acoustic events commonly found in conference rooms, classrooms, offices, and other settings, as well as unusual acoustic events that require special attention.

[0237] The network uses quantization technology to reduce the parameter size, and the overall model size is controlled within 500KB, which is suitable for deployment on resource-constrained microphone devices.

[0238] When applied in a large conference room environment, this lightweight convolutional neural network can detect 10 common acoustic events in real time, including speech, applause, knocks, door opening and closing, and glass breaking, with an average detection accuracy of 92% and a single detection latency of less than 50 milliseconds. Furthermore, through incremental learning, the model can gradually adapt to and learn new acoustic event types in specific scenarios, demonstrating its strong adaptability.

[0239] Step 4.2, multi-microphone sound source localization;

[0240] Based on the time delay and signal strength difference, calculate the spatial position of the sound source:

[0241] Calculate the time difference between the sound arriving at different microphones:

[0242] ;

[0243] in, Indicates that the sound reaches the microphone and microphone The time difference between Indicates that the sound reaches the microphone time point; Indicates that the sound reaches the microphone time point;

[0244] when When it is a positive value, it means that the sound reaches the microphone first After reaching the microphone ;

[0245] when When it is a negative value, it means that the sound reaches the microphone first After reaching the microphone ;

[0246] This time difference is a key input parameter for sound source localization calculation;

[0247] Based on the known microphone position information and time difference, the sound source position is calculated using the hyperbola localization method:

[0248] ;

[0249] in, is the sound source position vector, which represents the coordinates of the sound source to be found in three-dimensional space; and Microphone and The position vector represents the known coordinates of the two microphones in three-dimensional space; is the speed of sound, which is approximately 340 m / s (in air at standard atmospheric pressure and 20°C); For the sound to reach the microphone and time difference; Indicates the sound source to the microphone The Euclidean distance, or the length of the sound wave's path, is the mathematical relationship that describes the sound source's location on a hyperbola (a hyperboloid in three-dimensional space) with the two microphones as foci.

[0250] Combining time difference and intensity ratio information, a weighted positioning model is constructed to improve positioning accuracy:

[0251] ;

[0252] in, is the estimated sound source position vector, which represents the final calculated coordinates of the sound source in three-dimensional space; Indicates finding the minimum value of the following expression value; Indicates that all microphones Perform summation; is the weight coefficient of the positioning model, which is related to the signal quality. The better the signal quality, the greater the weight. is the sound source position vector, which is a variable in the optimization process; and Microphone and The position vector of Indicates the sound source to the microphone The Euclidean distance of is the speed of sound, usually about 340 m / s; For the sound to reach the microphone and The entire formula indicates that the optimal sound source position is determined by minimizing the sum of squares of the time difference estimation errors of all microphone pairs.

[0253] Step 4.3: Multi-device information fusion and collaborative detection;

[0254] Integrate detection results from multiple microphone devices to improve the accuracy of acoustic event recognition:

[0255] Based on the principle of temporal and spatial consistency, the same type of events identified by multiple devices are associated, and if the difference in event timestamps is less than the threshold And the spatial position difference is less than the threshold , are considered the same event;

[0256] Use a confidence-weighted voting mechanism to integrate event classification results from multiple devices:

[0257] ;

[0258] in, is the final event category, which indicates the acoustic event type determined by the system after integrating the detection results of multiple microphone devices; Indicates finding the category that makes the following expression reach the maximum value ; Indicates that all The contributions of each microphone device are accumulated; Indicates the number of the microphone device, from 1 to ; For devices For categories The predicted probability of A microphone device determines whether the current acoustic event belongs to the category The confidence level of is in the range of [0, 1]; Device weight, reflecting the microphone device The reliability and importance of a device in the overall system are typically dynamically adjusted based on factors such as the device's signal quality and historical accuracy. This formula describes a weighted voting mechanism that comprehensively considers the detection results of multiple microphone devices and selects the category with the largest weighted probability as the final judgment.

[0259] For low-confidence areas, increase the weight of neighboring devices to improve the overall judgment accuracy.

[0260] Step 4.4, construction of acoustic event spatial heat map;

[0261] Based on the multi-device collaborative detection results, a spatial heat map of acoustic events is constructed:

[0262] Divide the space into a three-dimensional grid, each grid point Represents a position in space at time The acoustic event heat at

[0263] The heat calculation formula is:

[0264] ;

[0265] in, Represents coordinate points in space In time The heat value of the acoustic event at ; Indicates that all The contributions of each microphone device are accumulated; Indicates the number of the microphone, from 1 to ; Indicates a microphone device The weight coefficient reflects the reliability and importance of the equipment in the overall system; Represents a spatial point to the microphone The Euclidean distance of Indicates on microphone The signal strength detected at Represents the distance attenuation function, which describes how the sound intensity decreases with distance while taking into account the original signal strength The function is usually a nonlinear attenuation model, reflecting the physical characteristics of sound waves propagating in space;

[0266] Use a time decay factor to make the impact of historical events gradually weaken over time:

[0267] ;

[0268] in, Represents spatial coordinates At the time point The heat value of the acoustic event at ; is the time decay coefficient, which controls the rate at which the impact of historical events weakens. A larger value indicates a faster decay of the impact of historical events. is the current system time; Indicates the time from when the event occurred To current time time difference; is an exponential decay function, whose value decreases rapidly as the time difference increases, so that the influence of earlier events on the current heat map gradually weakens. The entire formula describes how to dynamically update the acoustic event heat map as time progresses, ensuring that the heat map can reflect the latest acoustic activity status.

[0269] Heatmap threshold segmentation is used to identify high-activity areas as points of interest for potential abnormal events.

[0270] When this distributed collaborative detection system detects high-confidence abnormal acoustic events (such as breaking glass, unusual screams, etc.), it can automatically record audio evidence of the event and provide precise location information, expanding the function of the wireless microphone system from a simple human-computer interaction tool to an environmental safety monitoring platform.

[0271] Step 5: Use a layered edge computing architecture and energy-saving communication strategy to perform hierarchical processing on acoustic data to achieve low-latency acoustic event processing and howling prevention;

[0272] The specific implementation includes the following sub-steps:

[0273] Step 5.1: Building a hierarchical computing architecture;

[0274] Establish a three-layer edge computing architecture and reasonably distribute computing tasks:

[0275] The first layer (device layer): The low-power microprocessor built into the wireless microphone is mainly responsible for signal acquisition, preliminary feature extraction and event candidate detection;

[0276] The second layer (edge ​​aggregation layer): The edge server integrated in the charging station has medium computing power and is responsible for multi-device data integration, acoustic topology model update and collaborative event verification;

[0277] The third layer (cloud layer): optional cloud servers for long-term data storage, model optimization, and complex algorithm training.

[0278] The specific implementation of this layered edge computing architecture includes:

[0279] Hardware platform configuration:

[0280] Device layer: Each wireless microphone integrates an ARM Cortex-M4F or equivalent performance microcontroller with a main frequency of 80MHz, 256KB on-chip RAM, and 1MB flash memory;

[0281] Edge convergence layer: The charging station integrates a high-performance SoC processor, such as an ARM Cortex-A53 quad-core processor with a main frequency of 1.5GHz, 2GB of RAM, and 16GB of eMMC storage;

[0282] Cloud layer: Standard cloud server instances, with computing resources configured on demand.

[0283] Software architecture implementation:

[0284] The device layer uses a real-time operating system (RTOS), such as FreeRTOS, to ensure time determinism for critical tasks. The edge aggregation layer runs a lightweight Linux system, providing flexible task scheduling and network connectivity.

[0285] Inter-layer communication uses a layered message queue mechanism, and high-priority messages can interrupt low-priority transmissions;

[0286] The data flow adopts pipeline processing mode, and each processing unit works in parallel to reduce end-to-end delay.

[0287] Resource allocation strategy:

[0288] Dynamically adjust the allocation of computing resources at each layer based on task type and system load status;

[0289] Dedicated computing resources are reserved for critical tasks (such as howling prediction) to ensure worst-case performance;

[0290] Implement an adaptive task migration algorithm to intelligently distribute computing tasks between layers and optimize overall performance;

[0291] Set resource usage limits to prevent a single task from taking up too many resources and increasing system latency.

[0292] In actual application scenarios, this layered edge computing architecture exhibits the following characteristics:

[0293] The end-to-end delay of howling prediction and suppression is less than 10 milliseconds, meeting the needs of real-time audio processing;

[0294] The end-to-end latency of acoustic event detection is controlled within 100 milliseconds, allowing users to perceive it as a real-time response.

[0295] The system can independently complete more than 90% of its functions without cloud connection. Only advanced analysis and model updates require cloud connection.

[0296] The edge aggregation layer can process data streams from up to 16 wireless microphones simultaneously, keeping the average CPU load below 30%;

[0297] The battery life of wireless microphones is 30% longer than traditional solutions thanks to the rational distribution of computing tasks.

[0298] Step 5.2, computing task allocation and scheduling;

[0299] Optimize computing task allocation based on task characteristics and real-time requirements:

[0300] The tasks performed by the device layer include:

[0301] Audio signal preprocessing (sampling rate conversion, noise reduction, etc.);

[0302] Basic feature extraction (MFCC, spectral features, etc.);

[0303] Lightweight event classification (using the convolutional neural network in 4.1);

[0304] Preliminary howling detection (based on simplified frequency domain analysis);

[0305] The tasks performed by the edge aggregation layer include:

[0306] Multi-device information fusion and collaborative detection;

[0307] Acoustic transfer function matrix update and analysis;

[0308] Execution of complex howling prediction algorithms;

[0309] Adaptive filter parameter optimization;

[0310] The tasks performed by the cloud layer include:

[0311] Model training and optimization;

[0312] Historical data analysis;

[0313] System performance evaluation;

[0314] Step 5.3, energy-saving communication strategy;

[0315] Implement efficient energy-saving communication methods to reduce system power consumption:

[0316] Adopting event-driven communication mode, data upload is triggered only when potential events or significant changes in the acoustic environment are detected;

[0317] Adaptive sampling rate technology is used to dynamically adjust the data collection frequency according to the complexity of the environment:

[0318] Silent environment: reduce sampling rate and processing frequency;

[0319] Complex or changing environments: Increase sampling rate and processing frequency;

[0320] Implement data compression transmission, transmitting only valid features instead of raw audio data. The compression algorithm is selected based on the balance between computational complexity and compression efficiency.

[0321] Establish an optimized wireless communication protocol between the wireless microphone and the charging base, supporting multi-device synchronization and low-power transmission.

[0322] Step 5.4, system collaborative optimization algorithm;

[0323] Achieve collaborative optimization among subsystems:

[0324] Build a status sharing module to enable each device to understand the overall system status, such as the working mode of the charging station and other microphones, battery level, etc.

[0325] Introducing a dynamic task migration algorithm to migrate computing tasks from high-load devices to low-load devices when the load is unbalanced;

[0326] Build a priority processing queue to ensure that critical tasks (such as howling prevention) receive priority processing resources;

[0327] Implement an adaptive sleep strategy and implement strategic sleep for some devices or functional modules based on usage scenarios and battery status.

[0328] This layered architecture approach fully leverages the advantages of edge computing and rationally distributes computing tasks among different layers. This not only ensures the system's real-time response capabilities to critical events (such as potential howling and abnormal acoustic events), but also effectively reduces the system's overall power consumption and communication bandwidth requirements, enabling the wireless microphone system to operate stably for a long time and provide users with reliable sound pickup and environmental monitoring services.

[0329] Application examples of this implementation:

[0330] To verify the effectiveness of this implementation, we deployed the wireless microphone intelligent control system in a multi-functional conference room and conducted field tests and data collection. The following is a detailed implementation process and verification of the system's effectiveness in actual application.

[0331] The test site is a 300-square-meter multi-functional conference hall with the following features:

[0332] Movable partition walls can adjust the size and shape of the space according to needs;

[0333] The ceiling is made of reflective material, the floor is carpeted, and the surrounding walls are partly glass and partly sound-absorbing material;

[0334] The standard configuration has 4 main speakers and 2 auxiliary speakers;

[0335] Daily activities in the space include meetings, lectures, small music performances and many other uses.

[0336] The following devices are deployed in this scenario:

[0337] 8 wireless microphones, each equipped with 3 microphone units and built-in signal processing unit;

[0338] 2 charging cradles, each with 6 integrated fixed condenser microphones and edge computing servers;

[0339] Connects to the main console of the conference room audio system.

[0340] Example of building a spatial acoustic topology model:

[0341] During the system initialization phase, test signals were sent to the environment to acquire spatial acoustic characteristics. Table 1 shows the main eigenvalues ​​of the acoustic transfer function between some measurement points and the characteristics of the main acoustic feedback paths determined.

[0342] Table 1: Main eigenvalues ​​of spatial acoustic transfer function and feedback path characteristics;

[0343]

[0344] Based on this data, the system constructed a complete spatial acoustic topology model for subsequent howling prediction and suppression. After performing singular value decomposition on the acquired complete transfer function matrix, four primary acoustic feedback paths were identified, which together contributed 92.7% of the total feedback energy.

[0345] Adaptive howling suppression example:

[0346] Based on the spatial acoustic topology model, the system automatically identifies potential howling frequencies and configures the corresponding adaptive notch filter parameters. Table 2 shows these parameters and their adaptive adjustments after spatial changes.

[0347] Table 2: Adaptive notch filter parameter configuration and adjustment;

[0348]

[0349] Acoustic event detection and localization example:

[0350] During the system's operation in the conference hall, it detected and located a variety of acoustic events. Table 3 shows some typical acoustic events identified by the system and their characteristics.

[0351] Table 3: Typical acoustic events detected and their characteristics;

[0352]

[0353] The system's spatial heat map of acoustic events helps staff monitor the entire conference hall's acoustic conditions in real time, improving response times to unusual situations. In one test, the system successfully detected the sound of breaking glass in the back row of the auditorium and provided accurate location information in less than 100 milliseconds.

[0354] Edge computing performance examples:

[0355] Under different load conditions, the computing resource usage and delay performance of each layer of the system are shown in Table 4.

[0356] Table 4: Edge computing architecture performance indicators;

[0357]

[0358] Technical effect verification:

[0359] To verify the effectiveness of this implementation, we compared the performance of traditional howling suppression technology and this solution in various scenarios. Tables 5 and 6 show two key technical effects of the system: howling prediction and suppression, and spatial variation adaptability.

[0360] Table 5: Comparison of howling prediction and suppression effects;

[0361]

[0362] Table 6: Comparison of adaptability to changes in spatial acoustic conditions;

[0363]

[0364] Table 7: Comprehensive evaluation of the overall system performance;

[0365]

[0366] The above real-world application examples demonstrate the significant advantages of this wireless microphone intelligent control method in terms of howling prediction and suppression, as well as environmental adaptability. The system not only detects potential howling 2-3 dB in advance, increasing the system gain margin by 4.5 dB, but also maintains over 95% suppression efficiency despite varying spatial acoustic conditions, simultaneously empowering wireless microphone systems with new environmental safety monitoring capabilities.

[0367] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A wireless microphone intelligent control method, characterized in that: The following steps are involved: Multiple wireless microphones and fixed microphones distributed in space are used to collect the first type of multi-point acoustic data, construct a spatial acoustic transfer function matrix, and use singular value decomposition to analyze the characteristics of the acoustic feedback path to form a spatial acoustic topology model. Based on the results of spatial acoustic topology analysis, an adaptive notch filter bank is constructed and optimized to achieve precise howling suppression. The second type of acoustic data is collected in real time through a distributed acoustic sensor network to construct a three-dimensional sound field distribution, detect changes in the spatial acoustic environment, and generate optimized suppression parameters for the changed area; A lightweight convolutional neural network is integrated into each microphone device, combined with multi-microphone information fusion technology to achieve the recognition and location of acoustic anomalies; A layered edge computing architecture and energy-saving communication strategy are used to process acoustic data in layers, achieving low-latency acoustic event processing and howling prevention.

2. A wireless microphone intelligent control method according to claim 1, characterized in that: The step of constructing the spatial acoustic transfer function matrix includes: During the system initialization phase, the excitation signal is output through each speaker in turn; The received signal is recorded simultaneously by all microphones; For each speaker-microphone combination, the transfer function is calculated to form a spatial acoustic transfer function matrix.

3. The wireless microphone intelligent control method according to claim 1, characterized in that: The steps of constructing the adaptive notch filter bank include: Based on the potential howling frequency points and their risk indexes identified by acoustic topology analysis, a second-order notch filter unit is constructed for each potential howling frequency point; According to the howling risk index, the attenuation factor of each notch filter is optimized; Construct a multi-stage linkage notch filter structure to deal with the mutual influence between howling frequencies; Predictive howling suppression is achieved, suppressing howling before it is fully formed.

4. The wireless microphone intelligent control method according to claim 1, characterized in that: The steps of constructing the distributed acoustic sensor network include: Use the fixed microphone on the charging station as the reference node and the wireless microphone as the dynamic node; Each wireless microphone is equipped with an inertial measurement unit to sense its own motion state; estimating the relative position between the wireless microphone and the reference node based on the time of flight or the received signal strength indication; Data transmission and synchronization between nodes are achieved through wireless communication networks.

5. The wireless microphone intelligent control method according to claim 1, characterized in that: The step of detecting changes in the spatial acoustic environment includes: Periodically updating the spatial acoustic transfer function matrix and calculating the rate of change of the spatial acoustic transfer function matrix and the matrix at the previous moment; Use singular value decomposition to track the direction of change of acoustic properties; Based on the components of the singular vectors, the spatial region where the acoustic environment changes is determined; When the rate of change exceeds a preset threshold, the adaptive update mechanism is triggered.

6. The wireless microphone intelligent control method according to claim 1, characterized in that: The structure of the lightweight convolutional neural network includes: Input layer, used to receive feature vectors; Two convolutional layers and a pooling layer to extract acoustic features; Two fully connected layers for classifying acoustic events; Weight pruning technology and fixed-point quantization are used to reduce the parameter scale, and the overall model size is controlled within 500KB.

7. The wireless microphone intelligent control method according to claim 1, characterized in that: The multi-microphone information fusion technology includes: Calculate the spatial position of the sound source based on the time difference and signal strength ratio; Based on the principle of temporal and spatial consistency, it associates the same type of events identified by multiple devices; Use a confidence-weighted voting mechanism to integrate event classification results from multiple devices; Construct a spatial heat map of acoustic events and identify high-activity areas as points of interest for potential abnormal events.

8. The wireless microphone intelligent control method according to claim 1, characterized in that: The layered edge computing architecture includes: The device layer, consisting of a low-power microprocessor built into the wireless microphone, is responsible for signal acquisition and preliminary feature extraction; The edge aggregation layer, consisting of edge servers integrated with charging stations, is responsible for multi-device data integration and model updates; The cloud layer is used for long-term data storage and model optimization; A hierarchical message queue mechanism and an adaptive task migration algorithm are used between layers for communication and resource optimization.

9. The wireless microphone intelligent control method according to claim 1, characterized in that: The energy-saving communication strategy includes: Adopting an event-driven communication mechanism, data upload is triggered only when a potential event is detected; Dynamically adjust data collection frequency according to environmental complexity; Implement data compression transmission to transmit only valid features instead of original audio data; Establish an optimized wireless communication protocol between the wireless microphone and the charging base, supporting multi-device synchronization and low-power transmission.

10. A wireless microphone intelligent control system, used to execute a wireless microphone intelligent control method according to any one of claims 1 to 9, characterized in that: include: Multiple wireless microphone modules, each equipped with a microphone array, a signal processing unit and a wireless communication module; Charging cradle module, integrating fixed microphone array and edge computing server; Spatial acoustic topology analysis module, used to construct and analyze spatial acoustic transfer function matrix; Adaptive howling suppression module, used to achieve precise howling control based on topology analysis; Distributed acoustic event detection module, used to realize multi-device collaborative acoustic anomaly event identification and location.

Images