An intelligent acoustic management system and method with real-time noise control

Abstract

The present invention relates to an intelligent acoustic management system and method with real-time noise control, and the intelligent acoustic management system and method of the present invention have many significant advantages. Based on a unique algorithm combined with multi-dimensional features collected by a sensor network, the location of the noise source can be accurately determined with small errors, providing a reliable basis for targeted noise reduction. The noise characteristics can be analyzed in real time, and the intelligent noise reduction equipment can be dynamically adjusted accordingly to effectively respond to various noise changes and significantly improve the acoustic environment. Accurate positioning and efficient control reduce equipment and energy waste, and improve equipment utilization and energy efficiency. The noise reduction effect is continuously optimized through closed-loop feedback. Fifth, user interaction is convenient, operation and management are convenient, and it helps to create a quiet and comfortable living and working environment.

Description

Technical Field

[0001] The present invention relates to the fields of acoustic technology, environmental monitoring and intelligent control, and in particular to an intelligent management system and method capable of locating noise sources in real time, analyzing noise characteristics and dynamically regulating noise reduction measures. Background Art

[0002] With the acceleration of industrialization and the continuous expansion of urbanization, noise pollution has become one of the environmental problems that need to be solved urgently worldwide. In the industrial field, the high-intensity noise generated by the continuous operation of various mechanical equipment in factory workshops, such as large compressors, generators, machine tools, etc., not only seriously affects the hearing health of workers, leading to hearing loss or even deafness, but also interferes with workers' attention, reduces work efficiency, and increases the rate of work errors. Long-term exposure to such an environment may also cause a series of health problems such as cardiovascular disease and neurasthenia. In urban environments, the number of traffic noise sources is increasing. The roar of vehicles such as cars, trains, and airplanes, as well as the huge noise generated by equipment such as pile drivers, cranes, and concrete mixers during construction, seriously interfere with the daily life of residents. Noise interferes with people's normal rest and sleep, makes people feel irritable and anxious, affects learning and work efficiency, and also has adverse effects on the normal operation of special places such as schools, hospitals, and scientific research institutions. The control method of active noise reduction in the existing technology is relatively simple, lacking in-depth analysis and precise regulation of the characteristics of the noise source. In terms of positioning accuracy, the algorithm is easily interfered by multiple factors in a complex environment. For example, in the presence of multipath propagation, reflected waves, background noise, and the accuracy limitation of the sensor itself, the positioning error is large. In actual tests, in a more complex industrial plant environment, the average positioning error can reach 5-8 meters. In urban areas with busy traffic, the error may be even greater, making it difficult to meet the needs of accurately attacking noise sources. In terms of noise reduction effect, the noise reduction ability of traditional passive sound insulation measures is limited, generally only 10-15dB of noise can be reduced, and the control effect on low-frequency noise is poor. However, due to the lack of accurate tracking and adaptation capabilities of simple active noise reduction technology to real-time changes in noise sources, the anti-phase sound waves it emits often cannot fully match the actual noise, resulting in unstable noise reduction effects, and it is difficult to achieve ideal noise reduction effects in complex noise environments. According to the prediction results of the model, relatively fixed control measures are adopted, such as adjusting the layout of sound insulation equipment, starting or shutting down some noise reduction equipment, etc., but these control measures lack real-time dynamic optimization capabilities and cannot be flexibly adjusted according to real-time changes in noise sources. Secondly, in terms of noise reduction control, since its control strategy is relatively fixed, once the characteristics of the noise source change (such as changes in noise frequency, intensity, position, etc.), the pre-established noise reduction strategy may not be able to respond effectively, thus affecting the noise reduction effect and failing to adapt to the complex and dynamic needs of noise pollution control. Systems based on traditional sensor networks and simple positioning algorithms are greatly interfered by environmental factors, and the algorithms themselves have limitations. Multipath effects, reflected waves, background noise and other factors in complex environments can seriously affect the measurement accuracy of signal arrival time difference or signal strength difference, resulting in large positioning errors. Moreover, these interference factors change dynamically in the actual environment, making the positioning results unstable and difficult to accurately determine the location of the noise source.The system based on machine learning model prediction is restricted by the quality of training data and the generalization ability of the model, and it is difficult to achieve high-precision positioning in actual complex scenarios. The inaccuracy and incompleteness of the training data will lead to inaccurate relationships learned by the model, and the model cannot effectively adapt to new environments or changes in noise sources, thus affecting positioning accuracy. The passive sound insulation measures of traditional systems have limited noise reduction capabilities and cannot effectively deal with noise of complex frequencies, and active noise reduction technology lacks precise matching. Simple active noise reduction technology does not fully consider the real-time characteristic changes of the noise source, and cannot adjust the parameters of the anti-phase sound wave according to the dynamic changes of the noise, resulting in unstable and unsatisfactory noise reduction effects. The fixed noise reduction strategy of the system predicted by the machine learning model cannot adapt to the real-time changes in the characteristics of the noise source. Once the frequency, intensity, position, etc. of the noise source change, the pre-set noise reduction strategy may not be able to respond effectively, and the noise reduction measures cannot be adjusted in time according to the actual situation, thus affecting the noise reduction effect and failing to meet the governance needs in complex dynamic noise environments. Due to inaccurate positioning and poor noise reduction effects, equipment and energy are wasted. In the case of large positioning errors, sound insulation or noise reduction equipment may be over-deployed in unnecessary areas, resulting in idleness and waste of equipment resources. At the same time, when noise reduction measures cannot effectively reduce noise, the equipment continues to run but cannot achieve the expected noise reduction target, consuming a lot of energy and low overall resource utilization efficiency. Both traditional systems and systems predicted by machine learning models have this problem of resource waste and cannot achieve optimal resource allocation. Most existing technologies lack the ability to intelligently perceive and automatically adapt to changes in noise environments. The algorithms and equipment parameters of traditional systems are relatively fixed, and they cannot automatically adjust the working mode according to environmental changes. Although the system predicted by the machine learning model has certain predictive capabilities, it lacks real-time monitoring and dynamic control mechanisms, cannot respond to dynamic changes in noise sources in a timely manner, lacks flexibility and intelligence in practical applications, and is difficult to cope with complex and changeable noise pollution situations. It is urgently needed to provide an intelligent acoustic management system and method with real-time noise control to achieve high-precision and high-stability noise source positioning, efficient and adaptive noise reduction effects, improve resource utilization efficiency, enhance the intelligence and adaptability of the system, so as to effectively solve the noise pollution problem in complex environments and create a quiet and comfortable living and working environment for people. Summary of the invention

[0003] The present invention provides an intelligent acoustic management system with real-time noise control, comprising:

[0004] Intelligent acoustic sensor network, composed of a large number of high-sensitivity, wide-band acoustic sensors, adopts a distributed layout to cover the monitoring area, has high-precision signal acquisition, preprocessing and self-organizing network communication capabilities, and the sensor nodes adopt low-power design;

[0005] The central processing unit is equipped with a high-performance processor, large-capacity memory and high-speed storage device, which runs the noise source location and control model based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm, the noise characteristic analysis program and the intelligent decision-making algorithm, and is used to receive and process data from the intelligent acoustic sensor network;

[0006] Dynamic control execution unit, including various types of intelligent noise reduction devices, can execute noise reduction control operations according to the instructions of the central processing unit and provide real-time feedback on the operation status of the equipment;

[0007] The communication network is used to build a stable and high-speed data transmission link between the intelligent acoustic sensor network, the central processing unit and the dynamic control execution unit, using a combination of wired and wireless communication methods;

[0008] The user interaction interface provides operators with an intuitive and convenient operating platform for system parameter setting, status monitoring, control command issuance, and querying system historical data and operation reports.

[0009] Furthermore, the acoustic sensor of the intelligent acoustic sensor network is specifically used for:

[0010] Collect noise signals in the environment in real time and convert them into digital signals;

[0011] Preprocess the collected signals, including filtering to remove interference signals, amplifying weak signals, and performing analog-to-digital conversion;

[0012] Through self-organizing network technology, it communicates with other sensor nodes, automatically senses changes in network topology and node failures, and adjusts the data transmission path in a timely manner.

[0013] Furthermore, the noise source location and control model based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm operated by the central processing unit is used for:

[0014] Receiving a noise signal feature vector preprocessed by an intelligent acoustic sensor network, wherein the feature vector includes a time domain feature, a frequency domain feature, and a time-frequency domain feature;

[0015] The feature vector is preliminarily processed using a quantum neural network to output a preliminary estimate of the noise source location;

[0016] The output of the quantum neural network is used as the input of the fuzzy inference system, and the estimation result of the noise source position is further optimized through the fuzzy inference system to achieve high-precision noise source positioning.

[0017] Furthermore, when training the quantum neural network-fuzzy inference system (QNN-FIS) fusion model, a quantum back propagation algorithm (QBP) is used for parameter optimization, specifically including:

[0018] Obtaining the measurement result of the quantum state through quantum measurement operation;

[0019] Calculating the gradient of a loss function with respect to a quantum gate parameter according to the measurement result, wherein the loss function is used to measure the difference between the predicted position and the actual position;

[0020] The quantum gate parameters are updated using the gradient descent method until the loss function converges or the preset number of training iterations is reached.

[0021] Further, the noise characteristic analysis program performed by the central processing unit is specifically used for:

[0022] For the located noise source, the frequency components of the noise signal are analyzed by fast Fourier transform (FFT), the sound pressure levels in different frequency bands are calculated, and the main noise frequency range is determined;

[0023] Analyze the statistical characteristics of time domain signals to determine the intensity variation pattern, stability, and periodicity of noise;

[0024] Time-frequency analysis technology is used to further observe the distribution of noise in the time-frequency domain and capture the transient characteristics and time-varying laws of noise.

[0025] Furthermore, the intelligent decision-making algorithm of the central processing unit is combined with a preset intelligent noise reduction strategy library to:

[0026] According to the noise characteristic analysis results, a suitable control strategy is selected from the intelligent noise reduction strategy library, wherein the intelligent noise reduction strategy library contains the optimal control strategy combination for different noise source types, locations, and characteristics;

[0027] The selected strategy is dynamically optimized and adjusted based on real-time monitoring data to generate a control plan and send instructions to the dynamic control execution unit to ensure that the noise reduction effect is always in the best state.

[0028] Furthermore, the various types of intelligent noise reduction devices of the dynamic control execution unit include:

[0029] Intelligent movable noise barriers that can achieve real-time adjustment of position and angle through motor drive and precise position control system;

[0030] Multi-band adaptive active noise reduction speaker array can automatically adjust the frequency, phase and amplitude of the emitted sound waves according to the real-time frequency characteristics of the noise to achieve precise noise reduction;

[0031] Intelligent ventilation system noise regulator can dynamically adjust the air volume, wind speed and ventilation path according to changes in ventilation noise.

[0032] Furthermore, the communication network preferentially uses wireless communication technology between the sensor network and the central processing unit, and selects wired communication or wireless communication between the central processing unit and the dynamic control execution unit according to the device distribution and communication requirements, so as to ensure the flexibility, real-time, reliability and stability of data transmission, and ensure the coordinated work between the various parts of the system.

[0033] On the other hand, the present invention also provides an intelligent acoustic management method with real-time noise source positioning and dynamic control functions, comprising the following steps:

[0034] Collect noise signals in the environment through intelligent acoustic sensor networks and perform preprocessing;

[0035] The noise source location and control model based on quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm is used to locate the noise source of the pre-processed noise signal.

[0036] The central processing unit performs noise characteristic analysis on the located noise source;

[0037] According to the noise characteristic analysis results, combined with the preset intelligent noise reduction strategy library, the intelligent decision-making algorithm of the central processing unit is used to generate a control plan and send instructions to the dynamic control execution unit;

[0038] The dynamic control execution unit performs noise reduction control operations according to the instructions, and the intelligent acoustic sensor network continuously monitors the noise changes and feeds the data back to the central processing unit;

[0039] The central processing unit evaluates the control effect based on the feedback information. If the expected effect is not achieved, the control strategy is further optimized and the noise reduction equipment parameters are adjusted again to form a closed-loop feedback optimization mechanism.

[0040] Furthermore, in the noise source locating step, it specifically includes:

[0041] Extracting a multi-dimensional feature vector of the preprocessed noise signal, including time domain features, frequency domain features, and time-frequency domain features;

[0042] The feature vector is input into the quantum neural network for preliminary processing, and a preliminary noise source location estimate is output;

[0043] The output of the quantum neural network is used as the input of the fuzzy inference system, and the estimation result of the noise source position is further optimized through the fuzzy inference system to achieve high-precision noise source positioning.

[0044] Beneficial effects:

[0045] 1. The present invention is based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm, combined with the multi-dimensional signal characteristics collected by the intelligent acoustic sensor network, and through quantum state processing and fuzzy inference optimization, the positioning accuracy is significantly improved, up to 1-3 meters, far exceeding the existing technology. This provides a reliable basis for precise noise reduction, can effectively take measures against noise sources, and avoid wasting resources in non-critical areas.

[0046] 2. Excellent noise reduction effect and adaptability: Through real-time noise characteristic analysis, based on the intelligent noise reduction strategy library and intelligent decision-making algorithm, the intelligent noise reduction equipment is dynamically adjusted. Whether it is stable low-frequency noise or sudden high-frequency noise, it can respond accurately, with significant and stable noise reduction effect. It can be adjusted in real time according to noise changes, adapt to complex dynamic environments, and effectively improve the acoustic environment.

[0047] 3. Efficient use of resources: Precise positioning and efficient noise reduction reduce unnecessary equipment deployment and energy consumption. Intelligent control enables equipment operating parameters to be adjusted as needed to avoid ineffective operation, improve equipment utilization and energy efficiency, reduce operating costs, and achieve optimal resource allocation.

[0048] 4. Strong intelligent adaptability: The system can intelligently sense changes in the noise environment and automatically adjust the working mode. The sensor network monitors in real time, the central processing unit makes dynamic decisions, and the execution unit precisely controls to form a closed-loop feedback optimization mechanism that can respond to dynamic changes in noise sources in a timely manner. It has high flexibility and intelligence and can effectively deal with various complex noise pollution situations.

[0049] 5. Improve user experience and management efficiency: The user interface provides convenient operation and comprehensive information display, which is convenient for operators to manage. Functions such as real-time viewing of key information, manual adjustment of strategies, and query of historical data not only improve the operability of the system, but also help users better understand the noise situation, assist in decision-making, and create a quieter and more comfortable environment for people. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of the algorithm flow. DETAILED DESCRIPTION

[0051] Example 1

[0052] 1. System architecture

[0053] 1. Smart Acoustic Sensor Network

[0054] It is composed of a large number of highly sensitive, wide-band acoustic sensors, which are distributed and cover the monitoring area. The sensors have high-precision signal acquisition capabilities and can collect noise signals in the environment in real time and convert them into digital signals. Each sensor also integrates an advanced signal processing chip, which can perform real-time pre-processing on the collected signals, including filtering to remove interference signals, amplifying weak signals, and performing analog-to-digital conversion. Sensor nodes communicate with each other through self-organizing network technology (such as ZigBee, Wirel essHART, etc.) to form an intelligent perception network that can automatically perceive changes in network topology, node failures, etc., and adjust the data transmission path in time to ensure reliable data transmission. At the same time, the sensor nodes also have a low-power design and adopt a sleep wake-up mechanism to reduce the overall power consumption of the system while ensuring the real-time data acquisition.

[0055] 2. Central Processing Unit (CPU)

[0056] As the core control and computing part of the system, it is equipped with high-performance processors, large-capacity memory and high-speed storage devices. The central processing unit runs the noise source location and control model, noise characteristic analysis program and intelligent decision-making algorithm based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm. It is responsible for receiving and processing massive data from the intelligent acoustic sensor network, performing real-time and in-depth analysis of acoustic signals, and realizing accurate location of noise sources, accurate evaluation of noise characteristics and intelligent formulation of control strategies.

[0057] 3. Dynamically control execution units

[0058] According to the instructions of the central processing unit, the corresponding noise reduction control operation is performed. The unit contains various types of intelligent noise reduction equipment, such as intelligent movable sound insulation barriers (which can achieve real-time adjustment of position and angle through motor drive and precise position control system), multi-band adaptive active noise reduction speaker array (which can automatically adjust the frequency, phase and amplitude of the emitted sound waves according to the real-time frequency characteristics of the noise to achieve precise noise reduction), intelligent ventilation system noise regulator (which can dynamically adjust the air volume, wind speed and ventilation path according to the changes in ventilation noise), etc. These devices are connected to the central processing unit through advanced communication interfaces (such as Ethernet, CAN bus, etc.), receive control instructions, and provide real-time feedback on the operating status of the equipment to ensure the accurate execution of the control operation.

[0059] 4. Communication Network

[0060] Build a stable and high-speed data transmission link between the intelligent acoustic sensor network, the central processing unit, and the dynamic control execution unit. Use a combination of wired and wireless communication methods. Between the sensor network and the central processing unit, wireless communication technology (such as Wi-Fi, 5G, etc.) is used first to ensure the flexibility and real-time performance of data transmission. Between the central processing unit and the dynamic control execution unit, wired communication (such as industrial Ethernet) or wireless communication (such as Bluetooth, ZigBee, etc.) can be selected according to the distribution and communication requirements of the equipment to ensure the reliability and stability of data transmission and the coordinated work between the various parts of the system.

[0061] 5. User Interface

[0062] Provide operators with an intuitive and convenient operating platform for system parameter setting, status monitoring and control command issuance. Through the graphical interface, operators can view key information such as the location of noise sources, noise intensity distribution, and the operating status of noise reduction equipment in real time, and present the real-time status of the noise environment and the working status of the system in a visual way. At the same time, operators can manually adjust the noise reduction strategy according to actual needs, such as temporarily adjusting equipment parameters under special circumstances or giving priority to noise problems in specific areas. In addition, the user interaction interface also provides the system's historical data query, report generation and data analysis functions to help operators better understand the system's operating effects and the changing trends of noise pollution, realize human-computer interaction functions, and improve the system's operability and management efficiency.

[0063] (II) Working Principle

[0064] 1. Noise source location

[0065] The noise source location is realized based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm. First, the noise signal collected by the intelligent acoustic sensor network is preprocessed to extract its multi-dimensional feature vector. These features include time domain features (such as the mean, variance, kurtosis, zero-crossing rate, etc. of the signal), frequency domain features (such as spectral center of gravity, spectral bandwidth, harmonic factor, power spectral density, etc.) and time-frequency domain features (such as time-frequency energy distribution after short-time Fourier transform, energy coefficient after wavelet packet decomposition, etc.). The extracted feature vector is input into the quantum neural network (QNN) for preliminary processing. Quantum neural network is a new computing model based on quantum mechanics principles and neural network structure. It uses quantum bits and quantum gate operations to process information and has stronger information processing capabilities and parallel computing capabilities. In QNN, quantum bits realize information encoding, transmission and processing through quantum gate operations. After calculations in multiple hidden layers, a preliminary noise source location estimate is output.

[0066] Then, the output of QNN is used as the input of fuzzy inference system (FIS). Fuzzy inference system is based on fuzzy logic theory and can handle uncertainty and fuzzy information. It maps the output of QNN to the actual noise source location space by defining fuzzy rules and membership functions, performs fuzzy reasoning and decision-making, and further optimizes the estimation result of the noise source location. For example, according to the characteristics of the noise signal collected by the sensor, QNN may output a preliminary location range, and FIS refines and accurately judges this range according to pre-defined fuzzy rules (such as "if the signal strength is within a certain range and the frequency characteristics conform to a certain pattern, the noise source may be located in a certain area"), and finally obtains a more accurate noise source location.

[0067] When training the QNN-FIS fusion model, the quantum back propagation algorithm (QBP) is used for parameter optimization. The QBP algorithm combines the measurement principle in quantum mechanics and the back propagation algorithm of traditional neural networks. It can effectively adjust the quantum gate parameters in the quantum neural network and the membership function parameters in the fuzzy reasoning system, and improve the positioning accuracy and generalization ability of the model.

[0068] 2. Noise characteristics analysis and control strategy formulation

[0069] The central processing unit performs real-time noise characteristic analysis on the located noise source. By performing fast Fourier transform (FFT) on the noise signal to analyze its frequency components, the sound pressure level of different frequency bands is calculated to determine the main noise frequency range; at the same time, the statistical characteristics of the time domain signal (such as mean, variance, peak value, etc.) are analyzed to determine the intensity change law, stability, and whether there is periodicity of the noise. In addition, time-frequency analysis techniques (such as wavelet transform) can be used to further observe the distribution of noise in the time-frequency domain and capture the transient characteristics and time-varying laws of noise.

[0070] According to the results of noise characteristic analysis, combined with the preset intelligent noise reduction strategy library, the intelligent decision-making algorithm is used to generate a control plan. The preset intelligent noise reduction strategy library is established based on a large amount of experimental data, theoretical analysis and expert experience, and contains the optimal control strategy combination for different noise source types, locations and characteristics. For example, for noise sources with low-frequency noise as the main and stable intensity, the strategy library may store strategies for preferentially adjusting the low-frequency band parameters of the multi-band adaptive active noise reduction speaker array and appropriately adjusting the position of the intelligent movable sound barrier to optimize the noise reduction effect; for high-frequency sudden noise sources, it may include strategies such as quickly starting the high-frequency band active noise reduction function and adjusting the ventilation system noise regulator to reduce the propagation of high-frequency noise. The intelligent decision-making algorithm selects the appropriate strategy in the strategy library according to the specific situation of the current noise source, and dynamically optimizes and adjusts it according to the real-time monitoring data. For example, if the frequency of the noise source changes, the decision algorithm will adjust the operating frequency of the active noise reduction speaker in time; if the noise intensity suddenly increases, the output power of the noise reduction equipment will be increased accordingly or the position of the sound barrier will be adjusted to ensure that the noise reduction effect is always in the best state.

[0071] 3. Dynamic control execution and feedback optimization

[0072] After receiving the instruction from the central processing unit, the dynamic control execution unit immediately executes the corresponding noise reduction control operation and accurately adjusts the working parameters of the intelligent noise reduction equipment. During the control process, the intelligent acoustic sensor network continuously monitors the noise changes in real time and feeds back the newly collected data to the central processing unit. The central processing unit evaluates the control effect based on the feedback information. If the expected effect is not achieved, the control strategy is further optimized and the noise reduction equipment parameters are adjusted again to form a closed-loop feedback optimization mechanism. For example, if it is found that the noise reduction in a certain area does not meet the preset target, the central processing unit will analyze the reasons. It may be that the equipment parameters are not adjusted accurately or the characteristics of the noise source have changed. Then, based on the analysis results, the control strategy is re-formulated and instructions are sent to the execution unit again.

[0073] Adjust the equipment parameters repeatedly to continuously improve the noise reduction effect until the preset requirements are met or the optimal state is reached.

[0074] In order to fully verify the advantages of the system and method of the present invention over the existing technology, a series of comparative efficiency tests were carried out. Three typical noise scenes, industrial plants, urban traffic arteries, and construction sites, were selected as test sites.

[0075] In each scenario, the system of the present invention (Scheme 1), the combination of traditional fixed noise reduction measures (Scheme 2), and the noise reduction system based on machine learning model prediction (Scheme 3) were tested. During the test, multiple high-precision noise monitoring points were evenly arranged in each scenario to record the noise data at different locations before and after noise reduction.

[0076] The results show that in the industrial plant scene, the noise reduction of Scheme 1 is 15-20dB higher than that of Scheme 2 and 10-15dB higher than that of Scheme 3 on average, and can adapt to changes in working conditions such as equipment start-up and stop in a short time; in urban traffic arteries, Scheme 1 can maintain stable and efficient noise reduction in different traffic flow periods, and the reduction in equivalent continuous sound level (Leq) is more obvious compared with Schemes 2 and 3; in construction sites, Scheme 1 has outstanding control effects on sudden high-intensity noise and can quickly reduce noise intensity, while Schemes 2 and 3 are slow to respond and have poor noise reduction effects. Through these comparative efficiency tests, the significant advantages of the present invention in terms of noise reduction effect, adaptability and resource utilization efficiency have been strongly demonstrated.

[0077] Example 2

[0078] (I) System architecture

[0079] 1. Smart Acoustic Sensor Network

[0080] It is composed of a large number of highly sensitive, wide-band acoustic sensors, which are distributed to cover the monitoring area. The sensors can collect noise signals with high precision and convert them into digital signals. The integrated microprocessor and wireless communication module can pre-process the collected signals (such as filtering, amplification, etc.), and then transmit the data to the central processing unit through the wireless communication network (such as ZigBee, Wi-Fi, etc.). It also adopts a low-power design to ensure real-time data collection while reducing system power consumption.

[0081] 2. Central Processing Unit (CPU)

[0082] As the core control and computing part of the system, it is equipped with high-performance processors, large-capacity memory and high-speed storage devices. It runs the noise source location and control model based on the evolutionary neural network-chaos optimization algorithm (ENN-COA), the noise characteristic analysis program and the intelligent decision-making algorithm, is responsible for receiving and processing the data from the sensor network, and performs real-time and in-depth analysis of the acoustic signal to achieve accurate location of the noise source, accurate characteristic evaluation and intelligent formulation of the control strategy.

[0083] 3. Dynamically control execution units

[0084] Noise reduction control operations are performed according to the instructions of the central processing unit, including a variety of intelligent noise reduction equipment, such as intelligent movable sound insulation barriers (the position and angle can be adjusted in real time), multi-band adaptive active noise reduction speaker arrays (which can automatically adjust the emission sound wave parameters according to the real-time frequency characteristics of the noise), and intelligent ventilation system noise regulators (which can dynamically adjust the air volume and wind speed according to the changes in ventilation noise). These devices are connected to the central processing unit through advanced communication interfaces (such as Ethernet, CAN bus, etc.), receive control instructions and provide real-time feedback on the operating status to ensure accurate execution of control.

[0085] 4. Communication Network

[0086] Build a stable high-speed data transmission link between the sensor network, central processing unit and dynamic control execution unit, and adopt a combination of wired and wireless communication methods. Wireless communication (such as Wi-Fi, 5G, etc.) is preferred between the sensor network and the central processing unit to ensure the flexibility and real-time performance of data transmission; wired (such as industrial Ethernet) or wireless (such as Bluetooth, ZigBee, etc.) communication methods are selected between the central processing unit and the dynamic control execution unit according to the equipment distribution and communication requirements to ensure the reliability and stability of data transmission and ensure the coordinated work of all parts of the system.

[0087] 5. User Interface

[0088] Provide operators with an intuitive and convenient operating platform for system parameter setting, status monitoring and control command issuance. Through the graphical interface, key information such as the location of the noise source, noise intensity distribution, and the operating status of the noise reduction equipment can be viewed in real time. The noise reduction strategy can also be manually adjusted according to actual needs. The system historical data can also be queried, reports can be generated, and data can be analyzed to help operators understand the system operation effect and noise pollution change trend, realize human-computer interaction, and improve system operability and management efficiency.

[0089] (II) Working Principle

[0090] 1. Noise source location

[0091] Noise source location is achieved based on the evolutionary neural network-chaos optimization algorithm (ENN-COA). First, the noise signal collected by the intelligent acoustic sensor network is preprocessed to extract multi-dimensional feature vectors, including time domain features (such as signal mean, variance, kurtosis, zero-crossing rate, etc.), frequency domain features (such as spectrum center of gravity, spectrum bandwidth, harmonic factor, power spectrum density, etc.) and time-frequency domain features (such as time-frequency energy distribution after short-time Fourier transform, energy coefficient after wavelet packet decomposition, etc.).

[0092] The extracted feature vector is input into the evolutionary neural network (ENN). ENN is a network structure that introduces the idea of ​​evolutionary algorithm based on the traditional neural network. Its parameters such as neuron connection weight can be continuously optimized through evolutionary algorithm. After the feature vector is input, it is processed by multiple hidden layers and outputs the preliminary estimate of the noise source location.

[0093] Then, based on the preliminary position estimate output by the ENN, the chaos optimization algorithm (COA) is used for further optimization. Based on the randomness and ergodicity of the chaotic system, the chaos optimization algorithm perturbs the preliminary estimate by generating a chaotic sequence, so that it can search for the optimal solution more comprehensively in the solution space. In the specific operation, a chaotic sequence is generated according to a preset chaotic mapping function (such as a Logistic mapping function), and the chaotic sequence is combined with the preliminary estimate to obtain a new position estimate. Through multiple iterations, the position estimate is continuously optimized, and finally a more accurate noise source position is obtained.

[0094] 2. Noise characteristics analysis and control strategy formulation

[0095] The central processing unit performs real-time noise characteristic analysis on the located noise source. By performing fast Fourier transform (FFT) on the noise signal to analyze its frequency components, the sound pressure level of different frequency bands is calculated to determine the main noise frequency range; the statistical characteristics of the time domain signal (such as mean, variance, peak value, etc.) are analyzed to determine the intensity change law, stability, and whether there is periodicity of the noise. In addition, time-frequency analysis technology (such as wavelet transform) can be used to further observe the distribution of noise in the time-frequency domain and capture the transient characteristics and time-varying laws of noise.

[0096] According to the results of noise characteristic analysis, combined with the preset intelligent noise reduction strategy library, the intelligent decision-making algorithm is used to generate the control strategy. The preset intelligent noise reduction strategy library is established based on a large amount of experimental data, theoretical analysis and expert experience, and contains the optimal control strategy combination for different noise source types, locations and characteristics. For example, for noise sources with low-frequency noise as the main and stable intensity, the strategy library may store strategies for preferentially adjusting the low-frequency band parameters of the multi-band adaptive active noise reduction speaker array and appropriately adjusting the position of the intelligent movable sound barrier to optimize the noise reduction effect; for high-frequency sudden noise sources, it may include strategies such as quickly starting the high-frequency band active noise reduction function and adjusting the ventilation system noise regulator to reduce the propagation of high-frequency noise. The intelligent decision-making algorithm selects the appropriate strategy in the strategy library according to the specific situation of the current noise source, and dynamically optimizes and adjusts it according to the real-time monitoring data. For example, if the frequency of the noise source changes, the decision algorithm will adjust the operating frequency of the active noise reduction speaker in time; if the intensity of the noise source suddenly increases, the output power of the noise reduction equipment will be increased accordingly or the position of the sound barrier will be adjusted to ensure that the noise reduction effect is always in the best state.

[0097] 3. Dynamic control execution and feedback optimization

[0098] After receiving the instruction from the central processing unit, the dynamic control execution unit immediately executes the corresponding noise reduction control operation and accurately adjusts the working parameters of the intelligent noise reduction equipment. During the control process, the intelligent acoustic sensor network continuously monitors the noise changes in real time and feeds back the newly collected data to the central processing unit. The central processing unit evaluates the control effect based on the feedback information. If the expected effect is not achieved, the control strategy is further optimized and the noise reduction equipment parameters are adjusted again to form a closed-loop feedback optimization mechanism. For example, if it is found that the noise reduction in a certain area does not reach the preset target, the central processing unit will analyze the reasons. It may be that the equipment parameters are not adjusted accurately or the noise source characteristics have changed. Then, based on the analysis results, the control strategy is re-formulated, and instructions are sent to the execution unit again to adjust the equipment parameters. This cycle is repeated to continuously improve the noise reduction effect until the preset requirements are met or the optimal state is reached.

[0099] (III) Detailed modeling and solution process

[0100] 1. Data collection and preprocessing stage

[0101] Intelligent acoustic sensor networks are deployed in different noise environment scenarios (such as industrial plants, urban traffic arteries, construction sites, etc.) to collect large-scale data. During the collection process, the sensor's location information, collection time, and environmental parameters (such as temperature, humidity, air pressure, etc., which will affect the speed of sound and thus the accuracy of noise source positioning) are recorded simultaneously.

[0102] The collected original acoustic signal is preprocessed, including removing the DC component and normalizing the signal to eliminate the baseline drift and dimensional difference in the signal; the signal is filtered using digital filtering technology (such as Butterworth filter, Chebyshev filter, etc.) to remove high-frequency interference and high-frequency fluctuations in the signal; signal enhancement processing is performed, such as using wavelet transform technology to decompose and reconstruct the signal to improve the signal-to-noise ratio, ensure the quality and reliability of the data, and provide an accurate data basis for subsequent model training and positioning calculations.

[0103] 2. Model training stage (taking ENN-COA algorithm as an example)

[0104] In the Evolutionary Neural Network (ENN) part, the ENN structure is constructed to determine parameters such as the number of neurons, the number of hidden layers, and the initialization method of connection weights. For example, the number of neurons can be set to 100-200, the number of hidden layers can be set to 2-3, and the connection weights can be randomly initialized.

[0105] The preprocessed acoustic signal feature vector is input into the ENN, and after calculation and processing in each hidden layer, a preliminary noise source position estimate is output. The error between the output value and the actual position value is calculated, and the mean square error (MSE) is used as the error measurement standard.

[0106] According to the error, the connection weight and other parameters of ENN are optimized using evolutionary algorithms (such as genetic algorithms, particle swarm optimization algorithms, etc.). In genetic algorithms, connection weight and other parameters are regarded as genes. Through operations such as selection, crossover, and mutation, the parameters are continuously optimized to make the output noise source position estimation value more accurate. In the particle swarm optimization algorithm, each particle represents a set of possible combinations of connection weight and other parameters. Through the interaction between particles and the self-update mechanism, the parameters are continuously optimized. This process is repeated until the error meets the preset requirements or the maximum number of iterations is reached, and the training of the ENN part is completed.

[0107] In the Chaotic Optimization Algorithm (COA) part, the parameters of the chaotic optimization algorithm are determined, such as the chaotic mapping function (such as the Logistic mapping function), the number of iterations, the disturbance intensity, etc. Based on the preliminary position estimate output by the ENN, a chaotic sequence is generated according to the chaotic mapping function. The chaotic sequence is combined with the preliminary position estimate to obtain a new position estimate. The error between the new position estimate and the actual position value is calculated, and the mean square error (MSE) is used as the error measurement standard. According to the error, the generation method of the chaotic sequence or the disturbance intensity and other parameters are adjusted. Through multiple iterations, the position estimate is continuously optimized until the error meets the preset requirements or reaches the maximum number of iterations, and the training of the COA part is completed. Finally, through the combination of the ENN-COA algorithm, more accurate noise source positioning is achieved.

[0108] V. Comparative Test

[0109] (I) Experimental setup

[0110] 1. Test scenario selection

[0111] 1. Select three typical noise scenes: industrial plants, urban traffic arteries, and construction sites as test locations.

[0112] 2. Test plan

[0113] 1. For each test scenario, the following three solutions are used for testing:

[0114] Solution 1: Acoustic management system and method of the present invention

[0115] According to the system design requirements, the intelligent acoustic sensor network is reasonably arranged in the test area to ensure full coverage of the noise source area. The sensor nodes are connected to the central processing unit through appropriate wireless communication methods to achieve real-time data transmission. The central processing unit runs the noise source positioning and control model and intelligent decision-making algorithm based on ENN-COA. The dynamic control execution unit controls the intelligent noise reduction equipment (such as intelligent movable sound insulation barriers, multi-band adaptive active noise reduction speaker arrays, intelligent ventilation system noise regulators, etc.) according to the instructions to perform dynamic noise reduction operations.

[0116] Option 2: Combination of traditional fixed noise reduction measures (noise barrier + muffler)

[0117] Install noise barriers of fixed height and angle around the test area, using sound insulation materials of conventional specifications and materials. Install silencers on noise-generating equipment (such as some equipment in factories and some machinery on construction sites), and the silencer parameters are pre-set according to the equipment type and cannot be adjusted. During the entire test process, the noise reduction measures remain fixed and do not have real-time monitoring and dynamic control functions.

[0118] Solution 3: Noise reduction system based on machine learning model prediction

[0119] The intelligent acoustic sensor network is also deployed, but it is only used to collect noise data regularly (such as every half an hour or one hour), without real-time analysis and positioning. Based on past experience, the staff manually adjusts the angle of the sound barrier (if adjustable), the working status of the muffler and other noise reduction measures in specific time periods (such as according to factory production shifts, peak traffic flow periods, etc.). The control process is relatively rough and lacks real-time and precision.

[0120] (II) Test results

[0121] 1. Comparison of positioning accuracy

[0122] In the industrial plant scenario, the average positioning accuracy of Scheme 1 can reach 1-3 meters, while Scheme 2 cannot accurately provide positioning accuracy data due to the lack of effective positioning means. The positioning accuracy of Scheme 3 is about 5-8 meters. In the urban traffic artery scenario, the positioning accuracy of Scheme 1 can still be maintained within the range of 1-3 meters, and Scheme 2 cannot accurately provide positioning accuracy. The positioning accuracy of Scheme 3 is about 6-9 meters. In the construction site scenario, the positioning accuracy of Scheme 1 is 1-3 meters, Scheme 2 cannot provide positioning accuracy data, and the positioning accuracy of Scheme 3 is about 4-7 meters. It can be seen that the system and method of the present invention can achieve high-precision noise source positioning in different scenarios, which is significantly better than the other two solutions.

[0123] 2. Noise reduction effect comparison

[0124] In the industrial plant scenario, after dynamic regulation, the average noise reduction of the equivalent continuous sound level (Leq) of Scheme 1 can reach 20-30dB, the average noise reduction of Scheme 2 is only 5-10dB, and the average noise reduction of Scheme 3 is about 8-12dB. In the urban traffic arterial scenario, Scheme 1 can maintain a good noise reduction effect in different traffic flow periods, with an average noise reduction of 15-25dB, 3-8dB, and 6-10dB in Scheme 2. In the construction site scenario, Scheme 1 can quickly and effectively regulate sudden high-intensity noise, with an average noise reduction of 18-28dB in the equivalent continuous sound level (Leq), 4-9dB in Scheme 2, and 7-11dB in Scheme 3. It can be seen that the system and method of the present invention have outstanding performance in noise reduction effect, can effectively reduce noise intensity, and are superior to traditional fixed noise reduction measures and noise reduction systems based on machine learning model prediction.

Claims

1. An intelligent acoustic management system with real-time noise control, characterized in that: include: Intelligent acoustic sensor network, composed of high-sensitivity, wide-bandwidth acoustic sensors, using a distributed layout to cover the monitoring area; The central processing unit is equipped with a high-performance processor, large-capacity memory and high-speed storage device, which runs the noise source location and control model based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm, the noise characteristic analysis program and the intelligent decision-making algorithm, and is used to receive and process data from the intelligent acoustic sensor network; The dynamic control execution unit includes an intelligent noise reduction device, which executes the noise reduction control operation according to the instructions of the central processing unit and provides real-time feedback on the operation status of the device; The communication network is used to build a stable and high-speed data transmission link between the intelligent acoustic sensor network, the central processing unit and the dynamic control execution unit, using a combination of wired and wireless communication methods; The user interaction interface provides operators with an intuitive and convenient operating platform for system parameter setting, status monitoring, control command issuance, and querying system historical data and operation reports.

2. The intelligent acoustic management system according to claim 1, characterized in that: The acoustic sensor of the intelligent acoustic sensor network is specifically used for: Collect noise signals in the environment in real time and convert them into digital signals; Preprocess the collected signals, including filtering to remove interference signals, amplifying weak signals, and performing analog-to-digital conversion; Through self-organizing network technology, it communicates with other sensor nodes, automatically senses changes in network topology and node failures, and adjusts the data transmission path in a timely manner.

3. The intelligent acoustic management system according to claim 1, characterized in that: The noise source location and control model based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm run by the central processing unit is used to: Receiving a noise signal feature vector preprocessed by an intelligent acoustic sensor network, wherein the feature vector includes a time domain feature, a frequency domain feature, and a time-frequency domain feature; The feature vector is preliminarily processed using a quantum neural network to output a preliminary estimate of the noise source location; The output of the quantum neural network is used as the input of the fuzzy inference system, and the estimation result of the noise source position is further optimized through the fuzzy inference system to achieve high-precision noise source positioning.

4. The intelligent acoustic management system according to claim 3, characterized in that: When training the quantum neural network-fuzzy inference system (QNN-FIS) fusion model, the quantum back propagation algorithm (QBP) is used to optimize parameters, specifically including: Obtaining the measurement result of the quantum state through quantum measurement operation; Calculating the gradient of a loss function with respect to a quantum gate parameter according to the measurement result, wherein the loss function is used to measure the difference between the predicted position and the actual position; The quantum gate parameters are updated using the gradient descent method until the loss function converges or the preset number of training iterations is reached.

5. The intelligent acoustic management system according to claim 1, characterized in that: The noise characteristic analysis program performed by the central processing unit is specifically used for: For the located noise source, the frequency components of the noise signal are analyzed by fast Fourier transform (FFT), the sound pressure levels in different frequency bands are calculated, and the main noise frequency range is determined; Analyze the statistical characteristics of time domain signals to determine the intensity variation pattern, stability, and periodicity of noise; Time-frequency analysis technology is used to further observe the distribution of noise in the time-frequency domain and capture the transient characteristics and time-varying laws of noise.

6. The intelligent acoustic management system according to claim 1, characterized in that: The intelligent decision-making algorithm of the central processing unit is combined with a preset intelligent noise reduction strategy library to: According to the noise characteristic analysis results, a suitable control strategy is selected from the intelligent noise reduction strategy library, wherein the intelligent noise reduction strategy library contains the optimal control strategy combination for different noise source types, locations, and characteristics; The selected strategy is dynamically optimized and adjusted based on real-time monitoring data to generate a control plan and send instructions to the dynamic control execution unit to ensure that the noise reduction effect is always in the best state.

7. The intelligent acoustic management system according to claim 1, characterized in that: The various types of intelligent noise reduction devices of the dynamic control execution unit include: Intelligent movable noise barriers that can achieve real-time adjustment of position and angle through motor drive and precise position control system; Multi-band adaptive active noise reduction speaker array can automatically adjust the frequency, phase and amplitude of the emitted sound waves according to the real-time frequency characteristics of the noise to achieve precise noise reduction; Intelligent ventilation system noise regulator can dynamically adjust the air volume, wind speed and ventilation path according to changes in ventilation noise.

8. The intelligent acoustic management system according to claim 1, characterized in that: The communication network preferentially uses wireless communication technology between the sensor network and the central processing unit, and selects wired communication or wireless communication between the central processing unit and the dynamic control execution unit according to the device distribution and communication requirements to ensure the flexibility, real-time, reliability and stability of data transmission, and to ensure the coordinated work between the various parts of the system.

9. An intelligent acoustic management method with real-time noise source location and dynamic control functions, characterized in that: The following steps are involved: Collect noise signals in the environment through intelligent acoustic sensor networks and perform preprocessing; The noise source location and control model based on the quantum neural network-fuzzy inference system (QNN-FIS) fusion algorithm operated by the central processing unit is used to locate the noise source of the pre-processed noise signal; The central processing unit performs noise characteristic analysis on the located noise source; According to the noise characteristic analysis results, combined with the preset intelligent noise reduction strategy library, the intelligent decision-making algorithm of the central processing unit is used to generate a control plan and send instructions to the dynamic control execution unit; The dynamic control execution unit performs noise reduction control operations according to the instructions, and the intelligent acoustic sensor network continuously monitors the noise changes and feeds the data back to the central processing unit; The central processing unit evaluates the control effect based on the feedback information. If the expected effect is not achieved, the control strategy is further optimized and the noise reduction equipment parameters are adjusted again to form a closed-loop feedback optimization mechanism.

10. The intelligent acoustic management method according to claim 9, characterized in that: In the noise source location step, it specifically includes: extracting a multi-dimensional feature vector of the pre-processed noise signal, including time domain features, frequency domain features and time-frequency domain features; The feature vector is input into the quantum neural network for preliminary processing, and a preliminary noise source location estimate is output; The output of the quantum neural network is used as the input of the fuzzy inference system, and the estimation result of the noise source position is further optimized through the fuzzy inference system to achieve high-precision noise source positioning.

Images