Inter-floor noise active suppression system and method

Abstract

The invention discloses an inter-floor noise active suppression system and method.The system comprises a noise monitoring module, a signal processing module, a sound wave emission module and an intelligent control module, and the noise monitoring module, the sound wave emission module and the intelligent control module are all electrically connected with the signal processing module; the noise monitoring module comprises a broadband condenser microphone array, a vibration acceleration sensor and a beam forming processor. Through combination of acoustic monitoring, signal processing and sound wave emission technologies, upstairs noise can be monitored in real time, reverse sound waves with opposite phases and equal amplitudes are generated, so that the noise is effectively counteracted, and through combination of the multi-mode acoustic sensor array and a reverse sound wave noise reduction algorithm, real-time monitoring and quick response to floor noise are realized, and the noise reduction efficiency is improved. The real-time performance and accuracy of the noise reduction effect are ensured, and a quiet living and working environment is created for residents and office staff.

Description

Technical Field

[0001] The present invention relates to the technical field of inter - floor noise suppression, and particularly to an active inter - floor noise suppression system and method. Background Art

[0002] With the acceleration of the urbanization process, the intensive construction of high - rise residential buildings and office buildings has led to an increasingly serious problem of inter - floor noise. Traditional noise reduction measures, such as adding sound insulation layers and using sound - insulating materials, can alleviate the noise problem to a certain extent, but they are often accompanied by limitations such as complex construction and space occupation. To solve the above problems, the present application proposes an active inter - floor noise suppression system and method. Summary of the Invention

[0003] Based on the technical problems existing in the background art, the present invention proposes an active inter - floor noise suppression system and method.

[0004] An active inter - floor noise suppression system proposed by the present invention includes a noise monitoring module, a signal processing module, a sound wave emission module, and an intelligent control module. The noise monitoring module, the sound wave emission module, and the intelligent control module are all electrically connected to the signal processing module;

[0005] The noise monitoring module includes a broadband capacitive microphone array, a vibration acceleration sensor, and a beamforming processor;

[0006] The signal processing module includes an FPGA unit, an embedded AI coprocessor, and an adaptive filtering algorithm library. The FPGA unit is connected to the embedded AI coprocessor and the adaptive filtering algorithm library through a fiber optic interface;

[0007] The sound wave emission module includes a distributed loudspeaker array system and a power amplifier module.

[0008] Preferably, the broadband capacitive microphone array is used for air - borne sound wave acquisition, acoustic fingerprint extraction, and environmental sound field mapping. When acquiring air - borne sound waves, it adopts a 10Hz - 20kHz wide - frequency response design, covering the entire frequency band of building noise (including low - frequency structure - borne sound and high - frequency air - borne sound). Each array node integrates 4 MEMS microphones arranged in a tetrahedron to achieve 360° sound field sampling; when extracting acoustic fingerprints, through multi - channel synchronous sampling with a sampling rate ≥48kHz, it captures the time - frequency characteristics of noise events, providing raw data for subsequent noise classification. Its time - frequency characteristics include the rising - edge characteristics of impulse noise and the harmonic structure of steady - state noise; environmental sound field mapping combines the array topology structure to estimate the azimuth / elevation angle of the sound source, and the spacing of the array topology structure is dynamically adjusted according to the λ / 2 criterion;

[0009] The vibration acceleration sensor is used for structural vibration monitoring, vibration-acoustic field coupling analysis, and solid sound source localization. For structural vibration monitoring, a triaxial piezoelectric accelerometer with a frequency response of 0.5 Hz - 5 kHz and a sensitivity of 100 mV / g is embedded in the floor contact surface to monitor the vibration acceleration signal of the solid sound transmission path. For vibration-acoustic field coupling analysis, the intrinsic mode function (IMF) of the vibration signal is extracted through Hilbert-Huang transform, and a transfer function model of vibration energy and air-borne sound radiation is established. For solid sound source localization, based on the time difference of arrival (TDOA) algorithm of vibration waves and combined with the sensor network topology, the coordinates of the vibration source are calculated.

[0010] The beamforming processor is used for three-dimensional sound field reconstruction, multi-source separation, and sound track prediction. For three-dimensional sound field reconstruction, a frequency-domain beamforming algorithm is adopted to generate an acoustic intensity cloud map based on the microphone array data, with a resolution of 5° * 5° at 1 kHz. For multi-source separation, through blind source separation (BSS) technology and combined with vibration sensing data, the mixed noise sources of air-borne sound and structure-borne sound are distinguished. For sound track prediction, a Kalman filter is used to track the moving sound source and predict the movement trajectory in the next 200 ms.

[0011] Preferably, the FPGA unit is used for real-time data processing, synchronously managing the sampling / transmission timing of the microphone array, vibration sensors, and speaker array, generating a precision clock signal, receiving the raw data through a fiber optic interface, and distributing it to the embedded AI co-processor and the adaptive filtering module.

[0012] The embedded AI co-processor is used for noise feature extraction, intelligent classification and decision-making of noise, and dynamically selecting the optimal filtering algorithm.

[0013] The adaptive filtering algorithm library is used for dynamic parameter optimization, nonlinear compensation, and multi-path effect correction. When performing dynamic parameter optimization, a variable step-size FxLMS algorithm is adopted to update the coefficients of the 256th-order FIR filter in real time. Nonlinear compensation is based on the Volterra series model to compensate for speaker distortion and reduce harmonic interference. Multi-path effect correction combines the building acoustic transfer function matrix H(z) to compensate for the multi-path delay and attenuation of sound wave propagation.

[0014] Preferably, the distributed speaker array system is used for generating three-dimensional beams, multi-band collaborative transmission, and spatial sound field reconstruction. When generating three-dimensional beams, acoustic material units are adopted to modulate and generate highly directional sound beams to accurately cover the noise source area and support dynamic beam adjustment to adapt to the tracking of moving sound sources. When performing multi-band collaborative transmission, low-frequency units of 50 - 800 Hz are used to suppress structure-borne sound, and high-frequency units of 800 Hz - 20 kHz are used to cancel air-borne sound. When performing spatial sound field reconstruction, the floor is divided into honeycomb-shaped sound control sub-domains, with the minimum unit of a single sub-domain being 0.5 × 0.5 m, and each sub-domain speaker is independently driven to achieve local noise cancellation.

[0015] The power amplifier module is used for high-fidelity signal amplification and non-linear distortion suppression.

[0016] Preferably, the intelligent control module has a multi-parameter feedback adjustment mechanism, which is used to automatically adjust the working parameters of the acoustic wave emission module according to the changes in the noise environment to ensure the optimization of the noise reduction effect.

[0017] The present invention also proposes a method for active suppression of inter-floor noise, including the following steps:

[0018] S1: Acoustic modeling and system deployment: The noise monitoring module constructs a building geometric model based on BIM / laser point cloud data, and solves the sound field equation by the finite element-statistical energy hybrid method (FE-SEA);

[0019] S2: Noise monitoring and signal processing: The beamforming processor realizes blind source separation of acoustic vibration signals through the FastICA algorithm, solves the mixing matrix A and the independent sound sources S, separates the noise components in the airborne sound and the structure vibration, distinguishes the impact sound and the speech sound upstairs, and provides a pure signal source for subsequent classification and cancellation;

[0020] S3: Acoustic wave emission and dynamic cancellation: Combining the signals separated in S2 and the sound field equation in S1, the acoustic wave emission module calculates the optimal driving weights of the speaker array through distributed beamforming weights, so that the emitted acoustic waves form destructive interference in the target area, realizing a highly directional sound beam, accurately covering the noise source area, and applying a reverse force through the piezoelectric actuator to control the structure vibration, suppressing the structure-borne sound caused by the structure vibration, canceling the structure-borne sound caused by the floor vibration, suppressing the low-frequency structure sound, and compensating for the deficiency of the airborne sound cancellation;

[0021] S4: Intelligent optimization and scenario adaptation: The signal processing module adopts a reward function reinforcement learning strategy to dynamically optimize the parameters, quantify the balance between the noise reduction effect and the energy consumption, and combine with the intelligent control module to guide the optimization direction of the control strategy, minimizing the power consumption while maximizing the noise reduction amount, and supporting scenario-based configuration at the same time.

[0022] Preferably, in the above S1, the logical steps for solving the sound field equation are as follows:

[0023] S101: In the low-frequency band, the Helmholtz equation is used for numerical simulation of the building sound field, describing the propagation law of low-frequency acoustic waves in the building structure, simulating the spatial distribution of the sound pressure field, accurately predicting the propagation path of low-frequency noise in the floor and wall structures, and providing basic sound field data for sensor deployment and reverse acoustic wave design;

[0024] The formula used in the Helmholtz equation is: Where is the Laplacian operator, representing the second-order spatial differential of the sound pressure field; p(r) is the sound pressure at position r; k is the wave number, k = ω / c, where ω is the angular frequency and c is the speed of sound; ω is the angular frequency, ω = 2πf, and f is the frequency; ρ0 is the air density; q(r) is the sound source term, representing the sound source intensity per unit volume; j is the imaginary unit, j 2 = -1;

[0025] S102: Perform the energy transfer of the building sound field in the high-frequency band, apply the Statistical Energy Analysis (SEA) equation to analyze the energy transfer and dissipation of high-frequency sound waves in the building subsystem, predict the energy distribution of high-frequency noise, and optimize the high-frequency cancellation strategy of the speaker array;

[0026] The formula used in its Statistical Energy Analysis (SEA) equation is: wη i E i +∑ j≠i wβ ij (E i -E j ) = P in,i , where η i is the loss factor (dimensionless) of the i-th subsystem, characterizing the energy dissipation ability; E i is the vibration energy of the i-th subsystem; β ij is the coupling loss factor (dimensionless) between subsystems i and j, reflecting the energy transfer efficiency; P in,i is the power input to the i-th subsystem; ω is the angular frequency;

[0027] S103: Perform impulse response verification, obtain the actual acoustic transfer function, calibrate the simulation model, and ensure that the sound field prediction results are consistent with the real environment. The formula used for impulse response verification is: where h(t) is the impulse response function in the time domain; IFFT is the inverse fast Fourier transform; S in (f) is the spectrum of the input signal (MLS sequence); S out (f) is the spectrum of the output signal (measurement microphone signal); * is the complex conjugate operator; ∈ is the regularization factor to prevent the denominator from being zero.

[0028] Preferably, in the above S2, the formulas used to solve the mixing matrix A and the independent sound source S are: where X air is the air-borne sound signal observation matrix; X vib is the vibration signal observation matrix; A is the mixing matrix, describing the propagation path characteristics from the sound source to the sensor; S is the independent sound source signal matrix; N is the additive noise matrix.

[0029] Preferably, in the above S3, the formula used for calculating the distributed beamforming weights is: where \(w\) is the complex weight vector of the loudspeaker array; \(R\) is the noise covariance matrix, reflecting the spatial correlation of the noise; \(d\) is the steering vector, representing the transfer function from the sound source to each loudspeaker; \(H\) is the conjugate transpose operator;

[0030] The formula used when the piezoelectric actuator applies a reverse force to control the structure vibration is: where \(F\) control is the reverse force applied by the piezoelectric actuator; \(\alpha\) is the force coupling coefficient, calibrated by the structural parameters; \(u\) is the floor vibration displacement; is the vibration acceleration; \(dx\) is the integration variable along the floor propagation direction.

[0031] Preferably, in the step S4, the expression of the reward function is: where \(r_t\) is the reward value at time \(t\); \(e\) pre is the residual noise energy before noise reduction; \(e\) post \(||\) is the residual noise energy after noise reduction; \(\lambda\) is the energy consumption penalty factor (dimensionless), balancing the noise reduction effect and power consumption; \(P\) amp is the total power consumption of the power amplifier.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] By combining acoustic monitoring, signal processing and acoustic wave emission technologies, the present invention can monitor the noise upstairs in real time and generate reverse acoustic waves with opposite phases and equal amplitudes, thereby effectively canceling the noise. And through the combination of the multi-modal acoustic sensor array and the reverse acoustic wave noise reduction algorithm, the real-time monitoring and rapid response of the floor noise are realized, ensuring the real-time performance and accuracy of the noise reduction effect, and creating a quiet living and working environment for residents and office workers. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a block diagram of an active floor noise suppression system proposed by the present invention;

[0035] Figure 2 is a flowchart of an active floor noise suppression method proposed by the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0036] The present invention will be further described below in conjunction with specific embodiments.

[0037] Embodiment

[0038] Refer to Figure 1-2, in this embodiment, an active floor noise suppression system is proposed, which includes a noise monitoring module, a signal processing module, an acoustic wave emission module, and an intelligent control module. The noise monitoring module, the acoustic wave emission module, and the intelligent control module are all electrically connected to the signal processing module. Among them, the intelligent control module has a multi-parameter feedback adjustment mechanism, which is used to automatically adjust the working parameters of the acoustic wave emission module (such as acoustic wave frequency, amplitude, emission angle, etc.) according to the change of the noise environment to ensure the optimization of the noise reduction effect;

[0039] The noise monitoring module includes a broadband capacitive microphone array, a vibration acceleration sensor, and a beamforming processor;

[0040] Among them, the broadband capacitive microphone array is used for air acoustic wave acquisition, acoustic fingerprint extraction, and environmental sound field mapping. When collecting air acoustic waves, it adopts a wide frequency response design of 10Hz - 20kHz, covering the full frequency band of building noise (including low-frequency structure sound and high-frequency air sound). Each array node integrates 4 MEMS microphones arranged in a tetrahedron to achieve 360° sound field sampling; when extracting acoustic fingerprints, through multi-channel synchronous sampling, its sampling rate ≥ 48kHz, capturing the time-frequency characteristics of noise events, providing raw data for subsequent noise classification. Its time-frequency characteristics include the rising edge characteristics of impulse noise and the harmonic structure of steady-state noise; environmental sound field mapping combines the array topology structure to estimate the azimuth / elevation angle of the sound source. The spacing of the array topology structure is dynamically adjusted according to the λ / 2 criterion;

[0041] The vibration acceleration sensor is used for structural vibration monitoring, vibration-sound field coupling analysis, and solid sound source localization. For structural vibration monitoring, a three-axis piezoelectric accelerometer with a frequency response of 0.5Hz - 5kHz and a sensitivity of 100mV / g is embedded in the floor contact surface to monitor the vibration acceleration signal of the solid sound transmission path. Its vibration acceleration signal can be the impact spectrum of footsteps and the characteristic frequency of equipment vibration; vibration-sound field coupling analysis extracts the intrinsic mode function (IMF) of the vibration signal through Hilbert-Huang transform and establishes a transfer function model of vibration energy and air sound radiation; solid sound source localization is based on the time difference of arrival (TDOA) algorithm of vibration waves. Combining the sensor network topology, the coordinates of the vibration source are solved, and the accuracy of its coordinates is ±0.3m;

[0042] The beamforming processor is used for three-dimensional sound field reconstruction, multi-source separation, and sound track prediction. For three-dimensional sound field reconstruction, a frequency-domain beamforming algorithm is adopted to generate a sound intensity cloud map according to the microphone array data. The resolution at 1kHz reaches 5° * 5°; multi-source separation is through blind source separation (BSS) technology, combined with vibration sensing data, to distinguish the mixed noise sources of air sound and structure sound (such as separating the voices and impacts of people upstairs at the same time); sound track prediction performs Kalman filtering tracking on moving sound sources (such as the sound of dragging furniture) and predicts the movement trajectory in the next 200ms;

[0043] The signal processing module includes an FPGA unit, an embedded AI coprocessor, and an adaptive filtering algorithm library. The FPGA unit is connected to the embedded AI coprocessor and the adaptive filtering algorithm library through a fiber optic interface;

[0044] Among them, the FPGA unit is used for real-time data processing, synchronously managing the sampling / transmission timings of the microphone array, vibration sensor, and speaker array, generating a precision clock signal with a jitter <5ps, receiving raw data through the fiber optic interface, and distributing it to the embedded AI coprocessor and the adaptive filtering module;

[0045] The embedded AI coprocessor is used for noise feature extraction, intelligent classification and decision-making of noise, and dynamically selecting the optimal filtering algorithm (such as preferentially using the psychoacoustic masking model for human voices);

[0046] The adaptive filtering algorithm library is used for dynamic parameter optimization, nonlinear compensation, and multipath effect correction. When performing dynamic parameter optimization, the variable step-size FxLMS algorithm is adopted (the step size μ = 0.001 - 0.1 is dynamically adjusted), the coefficients of the 256-order FIR filter are updated in real time. Nonlinear compensation is based on the Volterra series model to compensate for speaker distortion and reduce harmonic interference. Multipath effect correction combines the building acoustic transfer function matrix H(z) to compensate for the multipath delay and attenuation of sound wave propagation;

[0047] The sound wave emission module includes a distributed speaker array system and a power amplifier module;

[0048] Among them, the distributed speaker array system is used for generating three-dimensional beams, multi-band collaborative emission, and spatial sound field reconstruction. When generating three-dimensional beams, acoustic material units are used to modulate and generate highly directional sound beams (the beam width of the sound beam is ±15°), accurately covering the noise source area, and supporting dynamic beam adjustment to adapt to mobile sound source tracking (such as the sound of dragging furniture). When performing multi-band collaborative emission, low-frequency units in the range of 50 - 800Hz are used to suppress structure-borne sound, and high-frequency units in the range of 800Hz - 20kHz are used to cancel airborne sound. When performing spatial sound field reconstruction, the floor is divided into honeycomb-shaped sound control sub-domains, the minimum unit of a single sub-domain is 0.5×0.5m, and each sub-domain speaker is independently driven to achieve local noise cancellation;

[0049] The power amplifier module is used for high-fidelity signal amplification and nonlinear distortion suppression.

[0050] This embodiment also proposes a method for active suppression of inter-floor noise, including the following steps:

[0051] S1: Acoustic modeling and system deployment: The noise monitoring module constructs a building geometric model based on BIM / laser point cloud data, and solves the sound field equation through the finite element-statistical energy hybrid method (FE-SEA);

[0052] The logical steps for solving the sound field equation are as follows:

[0053] S101: In the low-frequency band (20 - 500 Hz), the Helmholtz equation is used for numerical simulation of the building sound field to describe the propagation law of low-frequency sound waves in the building structure, simulate the spatial distribution of the sound pressure field, accurately predict the propagation path of low-frequency noise in the floor and wall structures, and provide basic sound field data for sensor deployment and reverse acoustic wave design;

[0054] The formula used in the Helmholtz equation is: Where is the Laplace operator, representing the second-order spatial differential of the sound pressure field; p(r) is the sound pressure at position r (unit: Pa); k is the wave number, k = ω / c, where ω is the angular frequency and c is the speed of sound; ω is the angular frequency, ω = 2πf, f is the frequency; ρ0 is the air density (unit: kg / m 3 ); q(r) is the sound source term, representing the sound source intensity per unit volume (unit: m 3 / s 2 ); j is the imaginary unit, j 2 = -1;

[0055] S102: Conduct energy transfer of the building sound field in the high-frequency band (500 - 20 kHz), apply the Statistical Energy Analysis (SEA) equation to analyze the energy transfer and dissipation of high-frequency sound waves in building subsystems (such as floors, doors, and windows), predict the energy distribution of high-frequency noise, and optimize the high-frequency cancellation strategy of the speaker array;

[0056] The formula used in the Statistical Energy Analysis (SEA) equation is: wη i E i +∑ j≠i wβ ij (E i -E j ) = P in,i , where η i is the loss factor (dimensionless) of the i-th subsystem, characterizing the energy dissipation ability; E i is the vibration energy of the i-th subsystem (unit: J); β ij is the coupling loss factor (dimensionless) between subsystems i and j, reflecting the energy transfer efficiency; P in,i is the power input to the i-th subsystem (unit: W); ω is the angular frequency;

[0057] S103: Conduct impulse response verification, obtain the actual acoustic transfer function, calibrate the simulation model, and ensure that the sound field prediction results are consistent with the real environment. The formula used for impulse response verification is: Where h(t) is the time-domain impulse response function; IFFT is the inverse fast Fourier transform; Sin (f) is the spectrum of the input signal (MLS sequence); S out (f) is the spectrum of the output signal (measured microphone signal); * is the complex conjugate operator; ∈ is the regularization factor (unit: Pa 2 / Hz), which prevents the denominator from being zero;

[0058] S2: Noise monitoring and signal processing: For spatio-temporal alignment of multi-modal data through the FPGA unit, the beamforming processor realizes blind source separation of acoustic and vibration signals through the FastICA algorithm, solves the mixing matrix A and the independent sound sources S, separates the noise components in the airborne sound and structure vibration, distinguishes the impact sound upstairs (dominated by structure vibration) and the speech sound (dominated by airborne sound), and provides a pure signal source for subsequent classification and cancellation;

[0059] The formulas used when solving the mixing matrix A and the independent sound sources S are as follows: where X air is the observation matrix of the airborne sound signal; X vib is the observation matrix of the vibration signal; A is the mixing matrix, which describes the propagation path characteristics from the sound source to the sensor; S is the independent sound source signal matrix; N is the additive noise matrix;

[0060] S3: Acoustic wave emission and dynamic cancellation: Combining the signals separated in S2 and the sound field equation in S1, the acoustic wave emission module calculates the optimal driving weights of the speaker array through distributed beamforming weights, so that the emitted acoustic waves form destructive interference in the target area, realizes a highly directional sound beam (beam width ±15°), accurately covers the noise source area, and controls the structure vibration by applying a reverse force through the piezoelectric actuator, suppresses the structure-borne sound caused by the structure vibration, cancels the structure-borne sound caused by the floor vibration, suppresses the low-frequency structure sound (such as footsteps), and compensates for the deficiency of the airborne sound cancellation;

[0061] The formula used for calculating the distributed beamforming weights is as follows: where w is the complex weight vector of the speaker array; R is the noise covariance matrix, which reflects the spatial correlation of the noise; d is the steering vector, which represents the transfer function from the sound source to each speaker; H is the conjugate transpose operator;

[0062] The formula used when the piezoelectric actuator applies a reverse force to control the structure vibration is as follows: where F control is the reverse force applied by the piezoelectric actuator (unit: N); α is the force coupling coefficient (unit: N·s 2 / m), which is calibrated by the structure parameters; u is the floor vibration displacement (unit: m); is the vibration acceleration (unit: m / s 2); dx is the integration variable along the floor slab propagation direction;

[0063] S4: Intelligent optimization and scenario adaptation: The signal processing module adopts a reward function reinforcement learning strategy to dynamically optimize parameters, quantify the balance between noise reduction effect and energy consumption, and combine with the intelligent control module to guide the optimization direction of the control strategy, minimizing power consumption while maximizing the noise reduction amount, and also supporting scenario-based configuration (such as the medical device whitelist shielding the MRI characteristic frequency of 128 kHz);

[0064] The expression of the reward function is as follows: where rt is the reward value at time t; e pre is the residual noise energy before noise reduction (unit: Pa 2 ) ; e post || is the residual noise energy after noise reduction (unit: Pa 2 ) ; λ is the energy consumption penalty factor (dimensionless), balancing the noise reduction effect and power consumption; P amp is the total power consumption of the power amplifier;

[0065] In addition, this method is not only applicable to residential buildings, but also can be widely applied to various building scenarios such as office buildings, schools, libraries, hospitals, etc., and has a broad market application prospect;

[0066] This embodiment combines acoustic monitoring, signal processing and sound wave emission technologies, can monitor the upstairs noise in real time and generate reverse sound waves with opposite phases and equal amplitudes, thereby effectively canceling the noise, and through the combination of the multi-modal acoustic sensor array and the reverse sound wave noise reduction algorithm, realizes the real-time monitoring and rapid response to the floor noise, ensuring the real-time and accuracy of the noise reduction effect, and creating a quiet living and working environment for residents and office workers.

[0067] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. An active noise suppression system between floors, characterized in that: It includes a noise monitoring module, a signal processing module, a sound wave transmitting module and an intelligent control module, and the noise monitoring module, the sound wave transmitting module and the intelligent control module are all electrically connected to the signal processing module; The noise monitoring module includes a wideband capacitive microphone array, a vibration acceleration sensor and a beamforming processor; The signal processing module includes an FPGA unit, an embedded AI coprocessor and an adaptive filtering algorithm library, and the FPGA unit is connected to the embedded AI coprocessor and the adaptive filtering algorithm library through an optical fiber interface; The sound wave emission module includes a distributed speaker array system and a power amplifier module.

2. The active noise suppression system between floors according to claim 1 is characterized in that: The broadband capacitive microphone array is used for air sound wave collection, acoustic fingerprint extraction and environmental sound field mapping. A 10Hz-20kHz broadband response design is adopted for air sound wave collection to cover the full frequency band of building noise. Each array node integrates 4 MEMS microphones arranged in a tetrahedron to achieve 360° sound field sampling. Acoustic fingerprint extraction uses multi-channel synchronous sampling with a sampling rate of ≥48kHz to capture the time-frequency characteristics of noise events and provide raw data for subsequent noise classification. Its time-frequency characteristics include the rising edge characteristics of impulse noise and the harmonic structure of steady-state noise. The environmental sound field mapping is combined with the array topology to realize the estimation of the azimuth / elevation angle of the sound source. The spacing of the array topology is dynamically adjusted according to the λ / 2 criterion. The vibration acceleration sensor is used for structural vibration monitoring, vibration-acoustic field coupling analysis and solid sound source positioning. The structural vibration monitoring uses a triaxial piezoelectric accelerometer with a frequency response of 0.5Hz-5kHz and a sensitivity of 100mV / g, which is embedded in the floor contact surface to monitor the vibration acceleration signal of the solid sound transmission path; the vibration-acoustic field coupling analysis extracts the inherent mode function of the vibration signal through the Hilbert-Huang transform, and establishes a transfer function model of vibration energy and air sound radiation; the solid sound source positioning is based on the vibration wave arrival time difference algorithm, combined with the sensor network topology, to solve the vibration source coordinates; The beamforming processor is used for three-dimensional sound field reconstruction, multi-source separation and sound track prediction. The three-dimensional sound field reconstruction adopts a frequency domain beamforming algorithm to generate a sound intensity cloud map based on microphone array data, with a resolution of 5°*5° at 1kHz; multi-source separation uses blind source separation technology combined with vibration sensor data to distinguish between mixed noise sources of air sound and structure sound; sound track prediction performs Kalman filtering tracking on mobile sound sources to predict the movement trajectory of the next 200ms.

3. The active noise suppression system between floors according to claim 1 is characterized in that: The FPGA unit is used for real-time data processing, synchronously managing the sampling / transmission timing of the microphone array, vibration sensor and speaker array, generating a precise clock signal, and receiving raw data through an optical fiber interface and distributing it to the embedded AI coprocessor and adaptive filtering module; The embedded AI coprocessor is used for noise feature extraction, intelligent noise classification and decision making, and dynamic selection of the optimal filtering algorithm; The adaptive filtering algorithm library is used for dynamic parameter optimization, nonlinear compensation and multipath effect correction. The variable step size FxLMS algorithm is used for dynamic parameter optimization to update the 256-order FIR filter coefficients in real time. The nonlinear compensation is based on the Volterra series model to compensate for the speaker distortion and reduce harmonic interference. The multipath effect correction is combined with the architectural acoustic transfer function matrix H(z) to compensate for the multipath delay and attenuation of sound wave propagation.

4. The active noise suppression system between floors according to claim 1 is characterized in that: The distributed speaker array system is used to generate three-dimensional beams, multi-band collaborative transmission and spatial sound field reconstruction. When generating three-dimensional beams, acoustic material units are used to modulate and generate highly directional sound beams, accurately cover the noise source area, and support dynamic beam adjustment to adapt to mobile sound source tracking. When transmitting in multi-band collaborative mode, a low-frequency unit of 50-800 Hz is used to suppress structure-borne sound, and a high-frequency unit of 800 Hz-20 kHz is used to offset airborne sound. When reconstructing the spatial sound field, the floor is divided into honeycomb sound control subdomains, and the minimum unit of a single subdomain is 0.5×0.5 m. Each subdomain speaker is driven independently to achieve local noise cancellation. The power amplifier module is used for high-fidelity signal amplification and nonlinear distortion suppression.

5. The active noise suppression system between floors according to claim 1 is characterized in that: The intelligent control module is equipped with a multi-parameter feedback adjustment mechanism for automatically adjusting the working parameters of the sound wave emission module according to changes in the noise environment to ensure the optimization of the noise reduction effect.

6. A method for actively suppressing inter-floor noise, characterized in that: The following steps are involved: S1: Acoustic modeling and system deployment: The noise monitoring module constructs the building geometry model based on BIM / laser point cloud data and solves the acoustic field equations through the finite element-statistical energy hybrid method (FE-SEA); S2: Noise monitoring and signal processing: The beamforming processor uses the FastICA algorithm to achieve blind source separation of acoustic and vibration signals, solve the mixing matrix A and the independent sound source S, separate the noise components in airborne sound and structural vibration, distinguish between the impact sound and the voice upstairs, and provide a pure signal source for subsequent classification and cancellation; S3: Sound wave emission and dynamic cancellation: Combining the separated signal in S2 and the sound field equation of S1, the sound wave emission module calculates the optimal driving weights of the speaker array through distributed beamforming weights, so that the emitted sound waves form destructive interference in the target area, realize highly directional sound beams, accurately cover the noise source area, and apply reverse force through piezoelectric actuators to control structural vibration, suppress solid-borne sound caused by structural vibration, cancel solid-borne sound caused by floor vibration, suppress low-frequency structure-borne sound, and compensate for the lack of airborne sound cancellation; S4: Intelligent optimization and scenario adaptation: The signal processing module adopts a reward function reinforcement learning strategy, dynamically optimizes parameters, quantifies the balance between noise reduction effect and energy consumption, and combines with the intelligent control module to guide the optimization direction of the control strategy, maximizing the amount of noise reduction while minimizing power consumption, while supporting scenario-based configuration.

7. A method for actively suppressing inter-floor noise according to claim 6, characterized in that: In S1, the logical steps for solving the acoustic field equation are as follows: S101: The Helmholtz equation is used in the low-frequency band to perform numerical simulation of the building sound field, describe the propagation law of low-frequency sound waves in the building structure, simulate the spatial distribution of the sound pressure field, accurately predict the propagation path of low-frequency noise in the floor and wall structure, and provide basic sound field data for sensor deployment and reverse sound wave design; The formula used by the Helmholtz equation is: in is the Laplace operator, which represents the spatial second-order differential of the sound pressure field; p(r) is the sound pressure at position r; k is the wave number, k = ω / c, where ω is the angular frequency and c is the speed of sound; ω is the angular frequency, ω = 2πf, f is the frequency; ρ0 is the air density; q(r) is the sound source term, which represents the sound source intensity per unit volume; j is the imaginary unit, j 2 = -1; S102: Conduct high-frequency building sound field energy transfer, apply statistical energy analysis equations, analyze the energy transfer and dissipation of high-frequency sound waves in building subsystems, predict the energy distribution of high-frequency noise, and optimize the high-frequency cancellation strategy of the speaker array; The formula used in the statistical energy analysis (SEA) equation is: wη i E i +∑ j≠i wβ ij (E i -E j )=P in,i , where η i is the loss factor (dimensionless) of the ith subsystem, representing the energy dissipation capacity; E i is the vibration energy of the ith subsystem; β ij is the coupling loss factor between subsystems i and j (dimensionless), reflecting the energy transfer efficiency; P in,i is the power input to subsystem i; ω is the angular frequency; S103: perform impulse response verification, obtain the actual acoustic transfer function, calibrate the simulation model, and ensure that the sound field prediction result is consistent with the actual environment. The formula used for impulse response verification is: Where h(t) is the time domain impulse response function; IFFT is the inverse fast Fourier transform; S in (f) is the spectrum of the input signal (MLS sequence); S out (f) is the spectrum of the output signal (measured microphone signal); * is the complex conjugate operator; ∈ is the regularization factor to prevent the denominator from being zero.

8. The method for actively suppressing inter-floor noise according to claim 6, characterized in that: In S2, the formula used to solve the mixing matrix A and the independent sound source S is: Where X air is the air sound signal observation matrix; X vib is the vibration signal observation matrix; A is the mixing matrix, which describes the propagation path characteristics from the sound source to the sensor; S is the independent sound source signal matrix; N is the additive noise matrix.

9. The method for actively suppressing inter-floor noise according to claim 6, characterized in that: In S3, the formula used for calculating the distributed beamforming weight is: Where w is the complex weight vector of the speaker array; R is the noise covariance matrix, reflecting the spatial correlation of the noise; d is the steering vector, representing the transfer function from the sound source to each speaker; H is the conjugate transpose operator; The formula used when a piezoelectric actuator applies an opposing force to control the vibration of a structure is: where F control is the opposing force applied by the piezoelectric actuator; α is the force coupling coefficient, which is calibrated by the structural parameters; u is the vibration displacement of the floor; is the vibration acceleration; dx is the integral variable along the propagation direction of the floor.

10. The method for actively suppressing inter-floor noise according to claim 6, characterized in that: In S4, the expression of the reward function is: Where rt is the reward value at time t; e pre is the residual noise energy before noise reduction; e post || is the residual noise energy after noise reduction; λ is the energy penalty factor (dimensionless), which balances the noise reduction effect and power consumption; P amp is the total power consumption of the power amplifier.

Images