Active sound production system and method for optimizing door closing sound quality

Abstract

The invention discloses an active sound production system and method for optimizing door closing sound quality, and relates to the field of simulation testing, a multi-dimensional sensing array module comprises a distance sensor, a three-axis acceleration sensor, a distributed microphone array and a temperature and humidity sensor, the sensors synchronously collect data, the data are transmitted through a CAN bus, and the timeliness and consistency of the data are ensured; the environmental acoustic modeling module is used for carrying out noise spectrum analysis on microphone signals, evaluating a vehicle door motion state, processing distance and acceleration data, calculating a temperature correction coefficient, dynamically adjusting sound pressure parameters, periodically updating an environmental parameter matrix and reflecting acoustic characteristics under different working conditions; and the PID-fuzzy hybrid algorithm module adopts a PID and fuzzy logic fusion framework, presets a target frequency, a maximum sound pressure level and a door closing speed threshold value, and calculates frequency, sound pressure and speed deviation, and the fuzzy logic module dynamically adjusts PID parameters according to the noise level, weights and superposes a proportional term, an integral term and a differential term, and is used for regulating and controlling the door closing sound quality.

Description

Technical Field

[0001] This application relates to the field of simulation testing, and in particular to an active sound generation system for optimizing the sound quality of a door closing, an active sound generation method for optimizing the sound quality of a door closing, electronic devices, storage media, and simulation testing platforms. Background Technology

[0002] The sound quality of a car door closing is a core evaluation indicator of the overall NVH (noise, vibration, and harshness) performance of a vehicle, directly determining consumers' subjective perception of vehicle quality. With the upgrading of the automotive industry and the increasing demand for comfort from consumers, its optimization has become a key aspect of automotive R&D. Currently, technologies for improving the sound quality of door closing can be mainly divided into three categories: First, structural optimization, such as Chinese patent application publication number CN118660259A, which discloses an active sound generation method and system for improving the sound quality of car door closing, reducing vibration noise by enhancing door rigidity, optimizing door lock structure, and improving limiters; Second, material improvement, such as Chinese patent publication number CN103568998B, which discloses a vehicle door closing sound quality control device and method, using high-damping, sound-insulating materials to reduce door closing noise; Third, active noise cancellation, which uses microphones and speakers in the door or inside the vehicle to generate reverse sound waves to cancel out noise.

[0003] While existing technologies can improve sound quality to some extent, the following technical problems remain: First, poor adaptability to multiple vehicle models. Most technical solutions are developed for specific models and do not dynamically adjust parameters based on the structural differences and acoustic characteristics of different models, making them unsuitable for universal compatibility with multiple models. Second, insufficient environmental adaptability. Designed based on static environments, they do not consider the impact of temperature changes on the elasticity of sealing materials and the speed of sound propagation, nor do they avoid the interference of environmental noise on the listening experience, resulting in significant fluctuations in actual use. Third, a lack of personalized experience. Only a single door closing sound solution is provided, without adjustable options for users' different preferences such as "thickness" or "crispness," resulting in a lack of emotional and personalized features. Fourth, insufficient real-time noise reduction. Traditional active noise cancellation technologies have slow response times, making it difficult to capture and quickly cancel transient noises during the door closing process (such as door lock collision sounds and sealing strip squeezing sounds), resulting in unsatisfactory optimization effects. Summary of the Invention

[0004] The purpose of this invention is to provide an active sound generation system, an active sound generation method, an electronic device, a storage medium, and a simulation test platform for optimizing the sound quality of a door closing, thereby solving at least one of a number of technical problems.

[0005] Firstly, the compatibility with multiple vehicle models is poor. The technical solutions are mostly developed for specific vehicle models and do not dynamically adjust parameters based on the structural differences and acoustic characteristics of different vehicle models, making them unsuitable for universal compatibility with multiple vehicle models.

[0006] Secondly, it lacks environmental adaptability. Based on the static environment design, it does not consider the impact of temperature changes on the elasticity of sealing materials and the speed of sound propagation, nor does it avoid the interference of environmental noise on hearing. In actual use, the effect fluctuates greatly.

[0007] Third, it lacks a personalized experience, offering only a single door closing sound solution without designing adjustable options for users' different preferences such as "heavy" or "crisp," resulting in a lack of emotional and personalized elements.

[0008] Fourth, the noise reduction lacks real-time performance. Traditional active noise reduction technology has a slow response speed and is difficult to capture transient noises during the door closing process (such as door lock collision sound and sealing strip squeezing sound) and quickly cancel them out, resulting in unsatisfactory optimization effects.

[0009] This invention provides the following solution:

[0010] According to a first aspect of the present invention, an active sound generation system for optimizing the quality of a door closing sound is provided, the system comprising:

[0011] Multi-dimensional sensor array module: including distance sensor, three-axis accelerometer, distributed microphone array, temperature and humidity sensor. The sensors collect data synchronously and transmit the data through CAN bus to ensure data timeliness and consistency.

[0012] The environmental acoustic modeling module performs noise spectrum analysis on microphone signals, assesses the motion state of the car door, processes distance and acceleration data, calculates temperature correction coefficients, dynamically adjusts sound pressure parameters, and periodically updates the environmental parameter matrix to reflect the acoustic characteristics under different working conditions.

[0013] PID-Fuzzy Hybrid Algorithm Module: Adopts a fusion architecture of PID and fuzzy logic, presets target frequency, maximum sound pressure level, and door closing speed threshold, calculates frequency, sound pressure, and speed deviation, and the fuzzy logic module dynamically adjusts PID parameters according to noise level, weighting and superimposing proportional, integral, and derivative terms for the control of door closing sound quality.

[0014] Adaptive audio generation module: It uses a 16-bit / 44.1kHz sampling frequency to generate enhanced sound and noise reduction compensated by adaptive noise cancellation technology. After being synthesized by a digital signal processor, it is output to a power amplifier to drive the speaker, which is used for dual optimization of enhancement and noise reduction.

[0015] Acoustic feedback correction module: It uses the output sound signal collected by the microphone as feedback to calculate the frequency and sound pressure residual deviation. If the deviation exceeds the threshold, it automatically triggers the parameter self-learning mechanism to update the PID parameters, records the operating conditions to establish a mapping table, and loads matching optimization parameters in advance before the next door closing to verify the control effect.

[0016] Preferably, the multi-dimensional sensing array module includes:

[0017] The distance sensor uses a laser ranging module, installed next to the door latch, to collect the distance between the door and the locking point in real time, with a sampling frequency of 500Hz, and calculate the closing speed. The triaxial accelerometer is fixed to the reinforcing rib of the inner door panel, with a range of ±16g and a resolution of 1mg, and can detect the rate of change of door closing angle acceleration to capture motion characteristics. Four distributed microphone arrays (sensitivity -38dBV / Pa, frequency response 20Hz-20kHz) are placed in the A-pillar, B-pillar and door trim panels of the vehicle, and simultaneously collect ambient noise and door closing sound. The temperature and humidity sensor is installed next to the sealing strip, with a measurement range of -30℃ to 50℃ and 0-100%RH, and an accuracy of ±0.5℃ / ±3%RH, to compensate for the influence of the environment on acoustic parameters.

[0018] All sensors collect data synchronously at a sampling rate of 500Hz and transmit it efficiently via CAN bus to ensure data timeliness and consistency. The microphone array uses beamforming technology to focus on the area of ​​the door closing sound source, reducing environmental interference. Temperature and humidity sensors can specifically correct the effects of temperature on the elastic coefficient of the sealing strip and humidity on the speed of sound propagation, avoiding parameter deviations and outputting multi-dimensional raw data.

[0019] Preferably, the environmental acoustic modeling module includes:

[0020] Based on sensor data, a dynamic acoustic model is constructed. Its core functions are divided into three parts: First, environmental noise spectrum analysis, which uses Fast Fourier Transform (FFT) to decompose the microphone signal across its full bandwidth (10Hz-20kHz) to identify major frequency components such as street traffic noise (200-400Hz) and crowd noise (1000-2000Hz); second, door motion state assessment, which uses Kalman filtering to process distance and acceleration data, fits velocity / acceleration curves, and delineates key stages such as free closure, sealing strip contact, and locking; and third, temperature correction coefficient calculation, which combines temperature and humidity data with a material acoustic properties database to dynamically adjust the speed of sound under conditions ranging from -30℃ to 50℃, increasing the speed of sound by 0.6 m / s for every 1℃ increase in temperature (baseline 331.4 m / s).

[0021] The environmental acoustic modeling module updates the environmental parameter matrix every 10ms. This matrix contains 128 feature dimensions, including key information such as noise frequency, door closing speed, and temperature coefficient, which can comprehensively depict the dynamic changes in the environment of door closing sound propagation. Through high-frequency updates and multi-dimensional parameter integration, the model can accurately reflect the acoustic characteristics under different working conditions.

[0022] Preferably, the PID-fuzzy hybrid algorithm module includes:

[0023] It adopts an improved PID and fuzzy logic fusion architecture, with preset parameters set based on a large number of vehicle tests: target frequency range 300-500Hz, maximum sound pressure level 80dB, and door closing speed threshold 1.2m / s (to prevent sharp noise when closing the door at high speed); the algorithm calculates three core deviations in real time: frequency deviation Δf (actual main frequency - target main frequency, accuracy ±1Hz, actual main frequency extracted by peak detection algorithm), sound pressure deviation ΔSPL (A-weighted measured sound pressure - 80dB), and speed deviation Δv (actual speed obtained by differentiating distance data - 1.2m / s).

[0024] Frequency deviation calculation formula:

[0025]

[0026] In the formula: This indicates the frequency deviation, measured in Hz. It is the core indicator for judging the degree of deviation between the actual frequency of the door closing sound and the target frequency. A positive value means that the actual frequency is higher than the target value, and a negative value means that the actual frequency is lower than the target value. When its absolute value exceeds 50Hz, the system will activate the frequency correction mechanism to ensure that the frequency is stable in the comfortable range of 300-500Hz.

[0027] This represents the actual dominant frequency of the door closing sound, measured in Hz. It is extracted from the door closing sound spectrum collected by the microphone array using a peak detection algorithm (with the threshold set to 80% of the maximum amplitude of the spectrum). The accuracy can reach ±1Hz, directly reflecting the core frequency characteristics of the current door closing sound and serving as the basic data source for frequency optimization.

[0028] This indicates the target frequency of the door closing sound, measured in Hz. It is set based on the 300-500Hz comfort range verified by BMW engineers and a large amount of vehicle test data. The default value is 350Hz, which strictly meets the requirements of the "Automotive Acoustic Quality Standard" for the main frequency band of passenger car door closing sound and is the target benchmark for frequency optimization.

[0029] Formula for calculating sound pressure deviation:

[0030]

[0031] In the formula: SPL stands for Sound Pressure Deviation, measured in dB. It is used to quantify the difference between the actual sound pressure level of a door closing sound and the maximum permissible sound pressure level. A positive value indicates that the actual sound pressure level exceeds the limit, while a negative value indicates that the sound pressure level does not meet the target. When the deviation is greater than 5 dB, the system will trigger a compensation noise reduction enhancement mode to reduce the actual sound pressure level.

[0032] The sound pressure level of the door closing sound is expressed in dB. It is calculated using A-weighted weighting, which conforms to the characteristics of human hearing. The measurement time constant is set to 125ms. It is extracted from the sound signal collected by the microphone array and can truly reflect the sound pressure intensity actually perceived by the user, avoiding evaluation bias caused by differences in frequency sensitivity.

[0033] SP This indicates the maximum permissible sound pressure level of the door closing sound, measured in dB, with a value of 80 dB. It fully complies with the limits on abnormal peak values ​​of passenger car door closing sounds in the newly implemented "Automotive Acoustic Quality Standard" in 2025, and is the core threshold to ensure that the door closing sound quality does not exceed the standard.

[0034] The fuzzy logic module dynamically adjusts the PID parameters according to the ambient noise level: <60dB for low noise, 60-80dB for medium noise, and >80dB for high noise. In low noise conditions, the proportional coefficient Kp is reduced (0.1-0.2) to prevent overshoot, while in high noise conditions, Kp is increased (0.4-0.5) to accelerate the response. Simultaneously, the integral coefficient Ki (0.01-0.1) eliminates steady-state deviation, and the derivative coefficient Kd (0.05-0.2) suppresses sudden disturbances. The control output is generated by a weighted sum of the proportional term (adjusting the response intensity according to the current deviation), the integral term (accumulating historical deviations to eliminate long-term errors), and the derivative term (suppressing disturbances according to the rate of deviation change), ultimately outputting a precise control parameter vector.

[0035] PID control output:

[0036]

[0037] In the formula: Output represents the control output of the PID controller, which is the core parameter vector that guides the operation of the adaptive audio generation module. It includes specific adjustment instructions such as enhanced audio rate, noise reduction amplitude, and speaker gain, and directly determines the optimization direction and adjustment range of the audio signal.

[0038] This represents a proportionality coefficient, dynamically adjusted according to the ambient noise level (0.1-0.2 for low-noise environments, 0.3 for medium-noise environments, and 0.4-0.5 for high-noise environments). Its function is to adjust the coefficient based on the current deviation. The proportional output control quantity responds quickly to changes in deviation, avoiding sound distortion caused by over-adjustment in low noise conditions and preventing response lag from affecting the optimization effect in high noise conditions.

[0039] This represents the integral coefficient, with a value ranging from 0.01 to 0.1, and its function is to account for the deviation. Perform time integration, accumulate historical deviations and eliminate steady-state errors, such as long-term small sound pressure deviations of 2-3dB, to ensure that the quality of the door closing sound remains stable within the target range in the long term and to avoid the accumulation of deviations that lead to a gradual decline in quality.

[0040] This represents the differential coefficient, with a value ranging from 0.05 to 0.2, and its function is to account for the deviation. By differentiating the rate of change, we can predict the trend of deviation in advance, suppress sudden interference, such as sudden fluctuations in environmental noise or sudden speed changes caused by users closing doors forcefully, and avoid sound instability caused by large fluctuations in control quantity.

[0041] This represents the overall deviation vector, which consists of frequency deviations. Sound pressure deviation SPL, speed deviation The data consists of the difference between the actual closing speed and the 1.2m / s threshold. It reflects the comprehensive deviation of the closing sound frequency, sound pressure, motion state and target value in real time and is the core input data of PID control.

[0042] Indicates deviation The integral term over time has no unit and accumulates the total deviation from the start of the door closing action to the current moment. It is used to eliminate long-term static deviations and ensure that the system quality does not gradually deviate from the target range due to continuous small deviations (such as 1-2dB sound pressure deviation).

[0043] Indicates deviation The derivative term with respect to time, in units of 1 / s, reflects how quickly the deviation changes over time. It is used to respond in advance to changes in the deviation trend. For example, if the deviation increases rapidly, the control intensity can be increased in advance to reduce control lag and improve the system's ability to resist dynamic disturbances.

[0044] Preferably, the adaptive audio generation module includes:

[0045] The module generates two types of collaborative audio signals, "enhancement + noise reduction," based on algorithm control parameters. The sampling frequency is 16-bit / 44.1kHz (CD-quality), and it supports fine adjustment in 1Hz frequency steps and 0.5dB volume steps. The basic enhancement tone is based on an adjustable sine wave of 250-600Hz, superimposed with low-frequency harmonics of 50-150Hz (amplitude of 1 / 3 of the fundamental frequency) to enhance the sense of solidity when closing the door. When Δf>0 (actual main frequency is too high), the enhancement tone main frequency is reduced; when Δf<0, the enhancement tone main frequency is increased to ensure that the synthesized sound main frequency is stable at 300-500Hz. The noise reduction compensation uses adaptive noise cancellation technology to generate a sound wave with the opposite phase (180°±5°) and amplitude matching (error ≤2%) to the noise. When there is high noise, it focuses on canceling high-frequency noise of 1000-2000Hz.

[0046] After the two types of audio signals are synthesized by a digital signal processor (DSP), they are output to a power amplifier (output power 2×20W) to drive the two built-in speakers in the car door: the mid-bass unit is responsible for the 250-500Hz frequency band (frequency response 80-5000Hz), and the tweeter is responsible for the 500-600Hz frequency band (frequency response 500-20000Hz), achieving accurate sound reproduction; through precise synthesis by the DSP and frequency band division of the speakers, the enhancement of sound and noise reduction work together to improve the solidity of the door closing and effectively cancel environmental interference. At the same time, the fine adjustment function ensures that the sound parameters are highly matched with the target requirements, optimizing the door closing sound quality.

[0047] Preferably, the acoustic feedback correction module includes:

[0048] The control effect is verified through a closed-loop feedback mechanism. The output sound signal collected by the microphone array is used as the feedback source and compared with the target acoustic model (target spectrum, sound pressure level curve) of the environmental acoustic modeling module to calculate the frequency and sound pressure residual deviation (required to be ≤3dB). When the deviation exceeds the threshold three times in a row (e.g., ΔSPL_res>3dB), the parameter self-learning mechanism is automatically triggered: first, the deviation features (continuous high-frequency deviation, sudden pulse deviation, etc.) are extracted, and then the PID parameters are updated based on the gradient descent algorithm (Kp step size 0.05, Ki step size 0.01, Kd step size 0.02) to ensure the system control accuracy and avoid quality degradation caused by parameter drift.

[0049] Formula for calculating residual sound pressure deviation:

[0050]

[0051] In the formula: Representing the residual sound pressure deviation, measured in dB, it is a core evaluation metric of the acoustic feedback correction module, used to determine the final difference between the actual effect and the target effect after door closing sound optimization; its value is required to be... If the deviation exceeds the threshold three times in a row, the PID parameter self-learning mechanism will be automatically triggered to update the control parameters and improve the optimization accuracy.

[0052] The actual sound pressure level detected by feedback is expressed in dB. It is calculated from the speaker output sound signal collected by the microphone array after the door closing action is completed (when the door lock module triggers the locking signal). It can truly reflect the final sound pressure state of the door closing sound after optimization and is the direct basis for feedback correction, avoiding inaccurate evaluation due to signal loss during the optimization process.

[0053] This indicates the target sound pressure level for feedback correction, in dB, and is the base value compared to SP. Consistent (80dB), while also making minor adjustments based on current operating conditions (such as ambient temperature and user door-closing habits) (adjustment range). It serves as a dynamic benchmark for judging whether the optimization effect meets the standard, ensuring that the feedback correction always fits the actual use scenario;

[0054] The system synchronously records current operating conditions, including ambient temperature, door closing speed, and user closing force, and establishes an operating condition-parameter mapping table to provide a basis for rapid adjustment under different operating conditions. At the same time, it analyzes user door closing habits, such as 80% of users closing the door at a speed between 0.8-1.0 m / s. Before the next time that user closes the door, the system preloads matching and optimization parameters to reduce response latency. By adapting operating condition records to user habits, the system ensures long-term stable operation and improves the consistency of user experience.

[0055] By collecting door closing motion parameters (distance, acceleration) and environmental data of different vehicle models through multi-dimensional sensing, and combining them with a frequently updated environmental parameter matrix, the acoustic characteristics of different vehicle models can be dynamically adapted without modifying the vehicle structure or changing materials.

[0056] By combining temperature and humidity compensation with dynamic adjustment of PID parameters linked to noise level, stable sound quality is achieved under all operating conditions (-30℃ to 50℃, different noise environments), avoiding performance degradation caused by environmental fluctuations. This dual-dimensional adjustment logic of "real-time correction of environmental parameters + dynamic adaptation of algorithm parameters" is superior to existing static environment design schemes.

[0057] By finely adjusting audio parameters and adapting to user habits, the limitations of a single sound solution are broken, and the product is endowed with emotional design attributes. This personalized design of "adjustable parameters + habit memory" is a user experience optimization direction that has not been addressed by existing technologies.

[0058] Through a low-latency design that integrates high sampling rate sensing, fast algorithm computation, and instantaneous audio synthesis (total response latency ≤10ms), the system achieves real-time capture and cancellation of transient noise during door closing, solving the problems of "slow response and difficulty in handling transient noise" in traditional technologies. This closed-loop control of "real-time perception - fast computation - instantaneous cancellation" represents a breakthrough in efficiency optimization for active noise reduction technology.

[0059] According to a second aspect of the present invention, an active sound generation method for optimizing the quality of a door closing sound is provided. Based on the above-described system, the specific steps of the method are as follows:

[0060] S1. Real-time monitoring and data fusion: Step S1 is activated when the door opening angle exceeds 30°. The sensor array synchronously samples the data, and the preprocessed data through Kalman filtering, Fourier transform and beamforming algorithm is spatiotemporally registered. Multi-dimensional acoustic state vectors are fused to ensure the reliability and timeliness of multi-source data.

[0061] S2, Difference Calculation and Strategy Generation: Based on the acoustic state vector in step S1, calculate the frequency deviation and sound pressure deviation. When generating dynamic control parameters, combine the proportional, integral, and derivative three-term adjustment logic of the PID algorithm with the environment adaptation strategy for deviation analysis and control strategy formulation, and output the control parameter set.

[0062] S3. Dynamic sound generation control and feedback adjustment: The adaptive audio generation module synthesizes audio in real time according to the parameter set, and outputs it to the speaker. The acoustic feedback correction module starts to evaluate the effect. If the residual deviation exceeds 3dB, the preset parameter matrix is ​​finely adjusted. The user's door closing force is recorded and stored in the database. The database parameters are called the next time the user closes the door to optimize and adapt the door closing sound quality.

[0063] Preferably, the specific steps of real-time monitoring and data fusion in step S1 are as follows:

[0064] Step S11 is the system startup and data preprocessing stage. The trigger condition is that the door opening angle exceeds 30° (detected by the angle sensor at the door hinge). After activation, the sensor array synchronously collects data at a sampling rate of 500Hz. In the raw data preprocessing stage: the distance signal is filtered by Kalman (process noise covariance Q=0.01, observation noise covariance R=0.1) to remove mechanical vibration interference and obtain a smooth distance-time curve; the acceleration signal is processed by 1024-point Fast Fourier Transform (FFT) to extract characteristic frequencies, such as the 200-300Hz natural frequency of the door, to distinguish effective signals from noise; the microphone signal is focused on the door closing sound source by a delay-summation beamforming algorithm (spatial resolution ±5°) to suppress external noise crosstalk; the temperature and humidity signals are filtered by a 5-point moving average to eliminate instantaneous fluctuations.

[0065] S12, the environmental acoustic modeling module performs spatiotemporal registration on the preprocessed multi-source data, with a time synchronization accuracy of ±1ms and a spatial coordinate calibration error of ≤2cm, ensuring the consistency of data from different sensors in time and space dimensions; then, it fuses data such as distance, acceleration, noise spectrum, temperature and humidity to generate an acoustic state vector containing 128 feature dimensions, of which dimensions 1-10 are noise spectrum features and 11-20 are motion parameter features.

[0066] Preferably, the specific steps for difference calculation and strategy generation in step S2 are as follows:

[0067] S21. Based on acoustic state vector, deviation analysis is performed, and frequency deviation is calculated: The peak detection algorithm (threshold is 80% of the maximum amplitude) is used to extract the actual main frequency f0 from the door closing sound spectrum, and it is compared with the default target main frequency ftarget (350Hz) to obtain Δf. When Δf>50Hz (e.g., f0=410Hz), frequency correction is initiated, and the basic enhancement tone main frequency is adjusted first; when Δf<-50Hz, the enhancement tone main frequency and harmonic frequency are increased simultaneously to ensure that the main frequency returns to the target range; sound pressure level adjustment: The actual sound pressure level SPLact is calculated by A-weighted calculation (conforming to human hearing) (measurement time constant 125ms), and ΔSPL=SPLact-80dB is obtained. When ΔSPL>5dB (e.g., SPLact=86dB), the compensation noise reduction enhancement mode is triggered, and the noise reduction amplitude value is increased to reduce the actual sound pressure;

[0068] S22. When finally generating dynamic control parameters, the proportional, integral, and derivative adjustment logic of the PID algorithm is combined with the environment adaptation strategy of the fuzzy logic module. For example, in a medium-noise environment, Kp=0.3, Ki=0.05, and Kd=0.1 are set. The parameter set includes key information such as enhanced audio frequency, noise reduction amplitude, and speaker gain, which can be specifically adapted to different environments and motion conditions: in low noise, the focus is on fine adjustment of enhanced audio, and in high noise, the noise reduction effect is strengthened to ensure that the control strategy is highly matched with the actual needs.

[0069] Preferably, the specific steps of dynamic sound control and feedback adjustment in step S3 are as follows:

[0070] S31 and step S3 are the control strategy execution stage. The adaptive audio generation module synthesizes audio in real time according to the parameter set: the basic enhancement sound generates a sine wave at a set frequency (e.g., 350Hz), superimposed with a 100Hz low-frequency harmonic, and controls the sound pressure level at 75-80dB through 0.5dB step volume adjustment; the noise compensation and reduction generates an inverse sound wave in the 1000-2000Hz range according to the ambient noise spectrum; after the two types of signals are synthesized, they are output to the two speakers in the car door: the mid-bass unit is responsible for the 250-500Hz frequency band (frequency response 80-5000Hz), and the tweeter unit is responsible for the 500-600Hz frequency band (frequency response 500-20000Hz), to achieve accurate sound playback, and optimizes the quality after superimposing the actual door closing sound.

[0071] S32. The trigger signal for the completion of the door closing action is the door lock module locking signal (detecting that the bolt is fully engaged). At this time, the acoustic feedback correction module starts the effect evaluation: it collects the sound output from the speaker and compares it with the target acoustic model to calculate the residual deviation, such as Δfres=5Hz, ΔSPLres=2dB; if Δres>3dB (such as ΔSPLres=4dB), the preset parameter matrix is ​​finely adjusted proportionally. For every 10Hz increase in deviation, the target main frequency is adjusted by 1Hz. At the same time, the user's closing force (calculated from the peak acceleration) is recorded and stored in the habit database; the next time the user closes the door, the database parameters are directly called to reduce the calculation delay and stabilize and improve the user experience.

[0072] According to a third aspect of the present invention, an electronic device is provided, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0073] The memory stores a computer program that, when executed by the processor, causes the processor to perform steps such as an active sound generation method to optimize the quality of the door closing sound.

[0074] According to a fourth aspect of the present invention, a computer-readable storage medium is provided storing a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform steps such as an active sound generation method for optimizing the quality of a door closing sound.

[0075] According to a fifth aspect of the present invention, a simulation testing platform is provided, comprising:

[0076] Electronic devices for implementing steps of active sound generation methods, such as optimizing the quality of door closing sounds;

[0077] The processor runs programs, and when the programs are running, they execute steps such as active sound generation methods that optimize the quality of door closing sounds from data output by electronic devices.

[0078] Storage medium used to store programs that, when running, perform steps such as active sound generation methods to optimize the quality of door closing sounds based on data output from electronic devices.

[0079] The beneficial effects of this application through the above scheme are as follows:

[0080] 1. This invention constructs a comprehensive data acquisition system through a multi-dimensional sensor array. This array includes a distance sensor, a three-axis accelerometer, a distributed microphone array, and a temperature and humidity sensor, which can accurately collect motion parameters and environmental data when different vehicle models close their doors. Combined with an acoustic model that is dynamically updated every 10ms by an environmental acoustic modeling module, it can capture the structural differences between different vehicle models in real time, such as differences in door stiffness and door lock layout that lead to differences in acoustic characteristics. Dynamic adaptation to multiple vehicle models can be achieved without modifying the vehicle structure or changing materials. Temperature changes will change the elastic coefficient of the sealing material, resulting in changes in vibration and noise when the door is closed, while humidity affects the sound propagation speed. The sensor corrects these parameters in real time to avoid acoustic calculation errors. The hybrid algorithm of PID and fuzzy logic can also dynamically adjust the control parameters according to the environmental noise level. In low noise, the proportional coefficient is reduced to prevent over-adjustment of the sound, and in high noise, the proportional coefficient is increased to speed up the response. This solves the problem of unstable environmental fluctuation effects caused by static design in traditional technologies, and achieves stable sound quality in a temperature range of -30℃ to 50℃ and in various noise scenarios.

[0081] 2. This invention breaks through the limitations of traditional single-sound-based door closing schemes by combining an adaptive audio generation module and an acoustic feedback correction module. The adaptive audio generation module supports fine adjustment in 1Hz frequency steps and 0.5dB volume steps, generating different styles of basic enhanced sounds according to user preferences to meet the subjective needs of different users for door closing sounds. At the same time, the acoustic feedback correction module continuously records user door closing habits, including closing speed and force, to establish a dedicated user habit database. When the user closes the door again, the system does not need to recalculate parameters; it can directly preload matching optimized parameters, which not only reduces response latency but also ensures that the door closing sound always matches the user's habits, significantly improving the user's subjective experience satisfaction.

[0082] 3. This invention employs a low-latency design throughout the entire process. In the data acquisition stage, a multi-dimensional sensor array synchronously collects door-closing data at a high sampling rate of 500Hz, enabling rapid capture of transient noise during the closing process. In the data processing stage, Kalman filtering can remove mechanical vibration interference from the distance signal within 2ms, and Fast Fourier Transform can extract the characteristic frequency of the acceleration signal in a very short time, distinguishing between effective signals and noise, ensuring efficient data preprocessing. The PID-fuzzy hybrid algorithm calculates frequency, sound pressure, and speed deviation in real time, with a control command output delay of ≤3ms. The adaptive audio generation module can instantly synthesize compensation noise reduction that is out of phase and matches the amplitude of the transient noise, and quickly synthesizes audio through a digital signal processor and outputs it to the speaker. The total response delay of the entire control process is ≤10ms, which can track every transient noise node in the door-closing process in real time, achieving precise cancellation and ensuring that the door-closing sound always maintains a clear and comfortable auditory effect, solving the problem of "untimely capture and inaccurate cancellation" of transient noise in traditional technologies. Attached Figure Description

[0083] Figure 1 This is a flowchart of an active sound generation system for optimizing door closing sound quality provided in one or more embodiments of the present invention;

[0084] Figure 2 This is a flowchart of an active sound generation method for optimizing the quality of door closing sound provided in one or more embodiments of the present invention;

[0085] Figure 3 This is a flowchart illustrating the environmental acoustic modeling process according to a specific embodiment of the present invention.

[0086] Figure 4 This is a block diagram of an electronic device for an active sound generation method that optimizes the quality of door closing sound, provided by one or more embodiments of the present invention. Detailed Implementation

[0087] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0088] like Figures 1 to 3 As shown, this embodiment of the invention provides an active sound generation system for optimizing the sound quality of a door closing, the system comprising:

[0089] Multi-dimensional sensor array module: including distance sensor, three-axis accelerometer, distributed microphone array, temperature and humidity sensor. The sensors collect data synchronously and transmit the data through CAN bus to ensure data timeliness and consistency.

[0090] The environmental acoustic modeling module performs noise spectrum analysis on microphone signals, assesses the motion state of the car door, processes distance and acceleration data, calculates temperature correction coefficients, dynamically adjusts sound pressure parameters, and periodically updates the environmental parameter matrix to reflect the acoustic characteristics under different working conditions.

[0091] PID-Fuzzy Hybrid Algorithm Module: Adopts a fusion architecture of PID and fuzzy logic, presets target frequency, maximum sound pressure level, and door closing speed threshold, calculates frequency, sound pressure, and speed deviation, and the fuzzy logic module dynamically adjusts PID parameters according to noise level, weighting and superimposing proportional, integral, and derivative terms for the control of door closing sound quality.

[0092] Adaptive audio generation module: It uses a 16-bit / 44.1kHz sampling frequency to generate enhanced sound and noise reduction compensated by adaptive noise cancellation technology. After being synthesized by a digital signal processor, it is output to a power amplifier to drive the speaker, which is used for dual optimization of enhancement and noise reduction.

[0093] Acoustic feedback correction module: It uses the output sound signal collected by the microphone as feedback to calculate the frequency and sound pressure residual deviation. If the deviation exceeds the threshold, it automatically triggers the parameter self-learning mechanism to update the PID parameters, records the operating conditions to establish a mapping table, and loads matching optimization parameters in advance before the next door closing to verify the control effect.

[0094] The multi-dimensional sensing array module includes:

[0095] The distance sensor uses a laser ranging module, mounted next to the door latch, to collect the distance between the door and the lock point in real time with an accuracy of ±0.5mm and a sampling frequency of 500Hz, calculating the closing speed. The triaxial accelerometer is fixed to the reinforcing rib of the inner door panel, with a range of ±16g and a resolution of 1mg, which can detect the rate of change of door closing angle acceleration to capture motion characteristics. Four distributed microphone arrays (sensitivity -38dBV / Pa, frequency response 20Hz-20kHz) are placed in the A-pillar, B-pillar, and door trim panels to simultaneously collect ambient noise and door closing sounds. The temperature and humidity sensor is mounted next to the sealing strip, with a measurement range of -30℃ to 50℃ and 0-100%RH, and an accuracy of ±0.5℃ / ±3%RH, used to compensate for the influence of the environment on acoustic parameters.

[0096] The specific model of the laser ranging module is VL53L1X. Its ranging accuracy remains stable at ±0.5mm within a range of 5cm-4m, and the data output delay is ≤1ms at a sampling frequency of 500Hz. The door closing speed is calculated using the ratio of the distance difference between adjacent sampling points to the sampling interval, as shown in the formula: , where v is the closing speed (unit: m / s). This represents the distance value (in meters) to the nth sampling point. The distance value (in meters) for the (n-1)th sampling point is calculated to be 0.002 seconds based on a 500Hz sampling rate.

[0097] The triaxial accelerometer, model ADXL345, is fixed at the inner door panel reinforcing rib near the door lock at about 1 / 3 of its length. It is secured with M3 screws with a tightening torque of 2 N·m to ensure that the vibration transmission efficiency between the sensor and the door is ≥95% and the error of the detected angular acceleration change rate is controlled within ≤5%. The temperature and humidity sensor, model SHT30, has a measurement range of -30℃ to 50℃ and 0-100%RH. In a normal temperature environment of 25℃, its accuracy can reach ±0.3℃ / ±2%RH, which can more accurately compensate for the influence of the environment on acoustic parameters.

[0098] All sensors collect data synchronously at a 500Hz sampling rate and transmit it efficiently via the CAN bus to ensure data timeliness and consistency; the microphone array uses beamforming technology to focus on the sound source area of ​​the door closing, reducing environmental interference; the temperature and humidity sensors can specifically correct the effects of temperature on the elastic coefficient of the sealing strip and humidity on the speed of sound propagation, avoiding parameter deviations and outputting multi-dimensional raw data.

[0099] The CAN bus transmission parameters are set as follows: baud rate 500kbps, data frame format adopts standard frame (ID length 11 bits), sensor data is packaged into 8-byte data frames, of which distance data occupies 2 bytes, acceleration data occupies 3 bytes, microphone signal data occupies 2 bytes, and temperature and humidity data occupies 1 byte. The single frame data transmission delay is ≤5ms to ensure data timeliness. The element spacing of the microphone array beamforming technology is set to 12cm (optimized based on the actual installation positions of the A-pillar, B-pillar and door interior panels), and the beam pointing angle range is -25° to 25° (with the door latch as the sound source center), which can attenuate the ambient noise in non-target areas by ≥15dB.

[0100] The correction formula for the effect of temperature on the elastic coefficient of the sealing strip is as follows: , where k is the modified elastic coefficient of the sealing strip (unit: MPa). The baseline elastic modulus of the sealing strip (EPDM material) at 25℃ is 5.0 MPa, where T is the current ambient temperature (unit: ℃); the correction formula for humidity on the speed of sound propagation is... ,in The corrected sound velocity (unit: m / s) and RH represent relative humidity (unit: %), ensuring that the acoustic parameter calculation deviation is ≤2%;

[0101] Delay-summation beamforming delay calculation logic:

[0102] Sound source localization: With the center of the door lock as the target sound source point (three-dimensional coordinates (0, 20, 50) mm, see the spatial coordinate calibration table), the signal delay time is calculated by the spatial distance difference between each microphone and the sound source point.

[0103] Delay time calculation formula:

[0104]

[0105] Indicates the first Signal delay time per microphone (in seconds);

[0106] Indicates the first The straight-line distance (in meters) from each microphone to the sound source is calculated using spatial coordinates, and the formula is:

[0107] ( , , The coordinates of the sound source point, , , For the first (microphone coordinates)

[0108] This indicates the distance (in meters) from the microphone (auxiliary microphone next to the laser distance sensor) closest to the sound source.

[0109] The speed of sound after temperature and humidity correction (unit: m / s) is expressed by the following formula:

[0110] (T is ambient temperature, RH is relative humidity)

[0111] Delay compensation and signal superposition: based on calculations Delay compensation is applied to the signals collected by each microphone to ensure that the sound source signals of all microphones are aligned in time. The aligned signals are then summed using equal weighting to generate an audio signal focused on the door closing sound source area, with environmental noise attenuation in non-target areas ≥15dB.

[0112] Example: Taking a microphone on pillar A (coordinates (150, 200, 350) mm) as an example, with the sound source point at (0, 20, 50) mm, the calculation yields:

[0113] ,like (Auxiliary microphone distance), at 25℃ and RH=50% ,but (i.e., 1ms), the A-pillar microphone signal needs to be delayed by 1ms before being superimposed.

[0114] The environmental acoustic modeling module includes:

[0115] Based on sensor data, a dynamic acoustic model is constructed. Its core functions are divided into three parts: First, environmental noise spectrum analysis, which uses Fast Fourier Transform (FFT) to decompose the microphone signal across its full bandwidth (10Hz-20kHz) to identify major frequency components such as street traffic noise (200-400Hz) and crowd noise (1000-2000Hz); second, door motion state assessment, which uses Kalman filtering to process distance and acceleration data, fits velocity / acceleration curves, and delineates key stages such as free closure, sealing strip contact, and locking; and third, temperature correction coefficient calculation, which combines temperature and humidity data with a material acoustic properties database to dynamically adjust the speed of sound under conditions ranging from -30℃ to 50℃, increasing the speed of sound by 0.6 m / s for every 1℃ increase in temperature (baseline 331.4 m / s).

[0116] The specific parameters of the Fast Fourier Transform (FFT) are as follows: 1024 sampling points, Hanning window (sidelobe attenuation ≥31dB) window function, and a frequency resolution of 43.07Hz (calculated from a sampling rate of 44.1kHz). After decomposing the 10Hz-20kHz frequency band, the noise type is determined by the amplitude proportion—street traffic noise is identified when the amplitude in the 200-400Hz frequency band accounts for ≥30% of the total amplitude, and crowd noise is identified when the amplitude in the 1000-2000Hz frequency band accounts for ≥25% of the total amplitude, with an accuracy rate of ≥90%. When processing distance and acceleration data using Kalman filtering, the state vector is set to [distance, velocity, acceleration]. The process noise covariance matrix is:

[0117] Observation noise covariance matrix: The iteration update period is consistent with the sensor sampling period (2ms), and the error of the fitted velocity / acceleration curve is ≤3%.

[0118] The material acoustic properties database includes acoustic parameters for key door components (sealing strips, sheet metal, interior panels): the sound attenuation coefficient of the sealing strip (EPDM) is 0.1-0.3 dB / m at -30℃ to 50℃; the sound reflection coefficient of the door sheet metal (cold-rolled steel) is ≥0.8; and the sound absorption coefficient of the interior panels (PU foam) is 0.2-0.4 in the 250-600Hz frequency band; temperature correction coefficient. ,in This is a temperature correction factor, where T is the current ambient temperature (unit: °C), used to dynamically adjust the sound velocity calculation results to ensure that the sound velocity calculation accuracy is ≤0.5% under operating conditions of -30℃ to 50℃.

[0119] Material acoustic property parameter table

[0120] Material type Temperature range Speed ​​of sound (m / s) Sound attenuation coefficient (dB / m) Reflectance coefficient Absorption coefficient (250-600Hz band) EPDM sealing strip -30℃ to 50℃ 331.4-361.4 0.1-0.3 0.6 0.15-0.25 Cold-rolled steel car door sheet metal -30℃ to 50℃ 5100-5200 0.05-0.1 0.8-0.9 0.05-0.1 PU foam interior panels -30℃ to 50℃ 320-340 0.2-0.4 0.3-0.4 0.2-0.4 Tempered glass door -30℃ to 50℃ 5600-5700 0.03-0.08 0.9-0.95 0.02-0.05

[0121] Note: The speed of sound in the table is calculated based on 331.4 m / s. The speed of sound increases by 0.6 m / s for every 1°C increase in temperature. The sound attenuation coefficient, reflection coefficient, and absorption coefficient are all measured under standard acoustic testing conditions (atmospheric pressure 101.325 kPa).

[0122] The environmental acoustic modeling module updates the environmental parameter matrix every 10ms. This matrix contains 128 feature dimensions, including key information such as noise frequency, door closing speed, and temperature coefficient, which can comprehensively depict the dynamic changes of the door closing sound propagation environment. Through high-frequency updates and multi-dimensional parameter integration, the model can accurately reflect the acoustic characteristics under different working conditions.

[0123] The 128-dimensional environmental parameter matrix is ​​divided into the following dimensions: Dimensions 1-20 represent the noise amplitude (dB) of 20 equally divided frequency bands (each band is 1kHz) within the 10Hz-20kHz frequency range; Dimensions 21-30 represent the noise phase (°) of the corresponding frequency bands; Dimensions 31-40 represent the distance data (m) of 10 key time points (each 1ms) during the door closing process; Dimensions 41-50 represent the acceleration data (m / s²) at the corresponding time points; Dimensions 51-60 represent temperature (°C), humidity (%RH), temperature correction factor, and humidity correction factor; Dimensions 61-128 represent the characteristic values ​​of the door's motion state (including the duration of the free closing phase, the peak contact force of the sealing strip, the locking impact force, etc., totaling 68 dimensions). The data precision of each dimension is clearly marked to ensure the matrix is ​​reproducible.

[0124] The environmental parameter matrix update mechanism is based on the time synchronization signal of the vehicle CAN bus (accuracy ±0.1ms). When the bus timestamp is detected to have increased by 10ms cumulatively, the matrix update is triggered. The update process adopts a dual-buffer mechanism. One buffer is used to store the currently calculated matrix data, and the other buffer is used to output historical matrix data to avoid update interruption caused by data read and write conflicts. The update success rate is ≥99.9%. At the same time, the matrix data is stored in 16-bit binary format, and each dimension of the data is accompanied by a valid range identifier (such as the valid range of the temperature dimension -30℃ to 50℃), which facilitates the subsequent algorithm module to verify the validity of the data.

[0125] The PID-fuzzy hybrid algorithm module includes:

[0126] It adopts an improved PID and fuzzy logic fusion architecture, with preset parameters set based on extensive vehicle testing: target frequency range 300-500Hz (matching BMW's comfort range), maximum sound pressure level 80dB (compliant with the "Automotive Acoustic Quality Standard"), and door closing speed threshold 1.2m / s (to prevent sharp noise when closing doors at high speeds). The algorithm calculates three core deviations in real time: frequency deviation Δf (actual main frequency - target main frequency, accuracy ±1Hz, actual main frequency extracted by peak detection algorithm), sound pressure deviation ΔSPL (A-weighted measured sound pressure - 80dB), and speed deviation Δv (actual speed obtained by differentiating distance data - 1.2m / s).

[0127] The specific steps for extracting the actual dominant frequency using the peak detection algorithm are as follows: First, perform 1024-point FFT processing on the door closing sound signal collected by the microphone to obtain the spectrum data; Second, set the amplitude threshold to 80% of the maximum amplitude of the spectrum and filter out the peaks with amplitudes exceeding the threshold; Third, within the target frequency band of 300-500Hz, select the frequency corresponding to the maximum peak as the actual dominant frequency f0; if there is no peak within the target frequency band, select the peak closest to 350Hz (default target dominant frequency) as f0, ensuring that the Δf calculation accuracy is ±1Hz; The parameters for A-weighted sound pressure detection conform to the GB / T3241-2010 standard, the measurement time constant is set to 125ms, the number of sampling points is 50 (corresponding to a sampling time of 100ms), and the average value is taken as the actual sound pressure level SPLact, with a ΔSPL calculation error ≤0.5dB;

[0128] Details of speed deviation Δv calculation: The actual closing speed is obtained by differentiating the distance data using the five-point differentiation method, the formula is as follows: ,in Let d(n) be the actual velocity at the nth sampling point (in m / s), d(n) be the distance data at the nth sampling point (in m), and Δt be the sampling interval (0.002 s); velocity deviation Δv = -1.2 (unit: m / s). When Δv>0.3m / s, additional noise reduction enhancement logic is triggered to avoid sharp noise generated when closing the door at high speed. The trigger threshold of this logic has been verified by testing on more than 100 car models, with an adaptation rate of ≥95%.

[0129] The fuzzy logic module dynamically adjusts the PID parameters according to the ambient noise level: <60dB for low noise, 60-80dB for medium noise, and >80dB for high noise. In low noise conditions, the proportional coefficient Kp is reduced (0.1-0.2) to prevent overshoot, while in high noise conditions, Kp is increased (0.4-0.5) to accelerate the response. Simultaneously, the integral coefficient Ki (0.01-0.1) eliminates steady-state deviation, and the derivative coefficient Kd (0.05-0.2) suppresses sudden disturbances. The control output is generated by a weighted sum of the proportional term (adjusting the response intensity according to the current deviation), the integral term (accumulating historical deviations to eliminate long-term errors), and the derivative term (suppressing disturbances according to the rate of deviation change), ultimately outputting a precise control parameter vector.

[0130] The weighted summation formula for the control output is:

[0131]

[0132] Where e(t) is the comprehensive deviation vector ( The weighting coefficients of 0.6, 0.3, and 0.1 were determined by fitting the data using the Particle Swarm Optimization (PSO) algorithm with 100 iterations. The objective function was to minimize the control error. The specific dimensions of the control parameter vector were [enhanced audio frequency, enhanced audio amplitude, noise reduction frequency band, noise reduction amplitude, and speaker gain], totaling 5 dimensions. The numerical precision of each dimension was 1Hz, 0.1dB, 10Hz, 0.1dB, and 0.5dB, respectively, to ensure that the output parameters could be directly used for audio generation.

[0133] The adaptive audio generation module includes:

[0134] The module generates two types of collaborative audio signals, "enhancement + noise reduction," based on algorithm control parameters. The sampling rate is 16-bit / 44.1kHz (CD-quality), and it supports fine adjustment in 1Hz frequency steps and 0.5dB volume steps. The basic enhancement sound is based on an adjustable sine wave of 250-600Hz, superimposed with low-frequency harmonics of 50-150Hz (amplitude of 1 / 3 of the fundamental frequency) to enhance the sense of solidity when closing the door. When Δf>0 (actual main frequency is too high), the enhancement sound main frequency is reduced; when Δf<0, the main frequency is increased to ensure that the synthesized sound main frequency is stable at 300-500Hz. The noise reduction compensation uses adaptive noise cancellation technology to generate a sound wave with the opposite phase (180°±5°) and amplitude matching (error ≤2%) to the noise. When there is high noise, it focuses on canceling high-frequency noise of 1000-2000Hz.

[0135] Details of low-frequency harmonic superposition in the basic boost tone: The frequency of the low-frequency harmonics is 1 / 3 of the basic boost tone's main frequency (e.g., when the basic boost tone's main frequency is 300Hz, the harmonic frequency is 100Hz; when the main frequency is 450Hz, the harmonic frequency is 150Hz), and the amplitude is strictly controlled to be 1 / 3 of the fundamental frequency's amplitude (error ±0.1dB). The superposition phase is consistent with the fundamental frequency (phase difference ≤2°). The step control of frequency adjustment adopts Direct Digital Synthesis (DDS) technology, with the DDS chip model AD9850. The frequency control word calculation accuracy reaches 32 bits, ensuring that the frequency error is ≤0.1Hz when adjusting in 1Hz steps. The volume adjustment is achieved through a digital potentiometer (model MCP41010), with a resistance adjustment accuracy of 0.1% corresponding to a 0.5dB step, and a volume control error ≤0.05dB.

[0136] The specific implementation of adaptive noise cancellation technology (based on the LMS algorithm): The algorithm filter order is set to 32, the step size μ=0.001, and the convergence error threshold=0.0001. The calculation stops when the iteration error is less than the threshold. The generation of the phase-opposite sound wave is achieved through three steps: "noise phase detection - delay compensation - phase reversal". The noise phase detection adopts the Hilbert Transform, with a phase detection accuracy of ±2°. The delay compensation is calculated based on the installation distance between the microphone and the speaker (typical value 0.5m). The compensation time = distance / speed of sound, ensuring that the generated reverse sound wave is synchronized with the noise in time. The amplitude matching is achieved by comparing the noise amplitude and the reverse sound wave amplitude in real time and dynamically adjusting the reverse sound wave amplitude. The matching error is ≤1%, and the noise cancellation amount in the 1000-2000Hz frequency band is ≥15dB.

[0137] The two types of audio signals are synthesized by a digital signal processor (DSP) and output to a power amplifier (output power 2×20W) to drive the two built-in speakers in the car door: the mid-bass unit is responsible for the 250-500Hz frequency band (frequency response 80-5000Hz), and the tweeter is responsible for the 500-600Hz frequency band (frequency response 500-20000Hz), achieving accurate sound reproduction; through precise synthesis by the DSP and frequency band division of the speakers, the enhancement of sound and noise reduction are ensured to work together, which not only improves the solidity of closing the door, but also effectively cancels environmental interference. At the same time, the fine adjustment function ensures that the sound parameters are highly matched with the target requirements, optimizing the sound quality of closing the door.

[0138] The specific parameters of the digital signal processor (DSP) are as follows: Model: TI TMS320C6748, operating frequency: 456MHz, supports 16-bit / 44.1kHz audio sampling rate. During audio synthesis, linear interpolation is used to convert 500Hz sensor data into 44.1kHz audio data, with a total harmonic distortion (THD) of <0.01% during the conversion process. The DSP's audio synthesis process is as follows: First, read the control parameters output by the PID-fuzzy algorithm; second, generate digital signals for basic enhancement and noise reduction; third, mix the two signals (mixing ratio = enhancement:noise reduction = 7:3); fourth, convert the digital signals to analog signals through the DSP's built-in DAC (16-bit precision) and output them to the power amplifier. The entire synthesis process has a delay of ≤5ms.

[0139] Detailed specifications of the power amplifier and speaker: The power amplifier model is TDA7388, with an output power of 2×20W (8Ω load), a frequency response of 20Hz-20kHz (±1dB), a signal-to-noise ratio ≥90dB, and a total harmonic distortion (THD) <0.05% (at 1kHz and 10W output); the mid-bass speaker model is PioneerTS-G1620F, with a rated power of 20W, impedance of 4Ω, a frequency response range of 80-5000Hz (±3dB), and sensitivity... The frequency response is 90dB / W / m; the tweeter model is PioneerTS-T110, with a rated power of 15W, impedance of 4Ω, frequency response range of 500-20000Hz (±3dB), and sensitivity of 92dB / W / m; the installation distance between the two is set to 15cm, and the installation positions are the lower part (mid-bass) and the upper part (tweeter) of the door trim panel, respectively, to ensure that the sound coverage uniformity in the 250-600Hz frequency band is ≥90%, and the amplitude difference at the frequency band junction (500Hz) is ≤1dB.

[0140] The acoustic feedback correction module includes:

[0141] The control effect is verified through a closed-loop feedback mechanism. The output sound signal collected by the microphone array is used as the feedback source and compared with the target acoustic model (target spectrum, sound pressure level curve) of the environmental acoustic modeling module to calculate the frequency and sound pressure residual deviation (required to be ≤3dB). When the deviation exceeds the threshold three times in a row (e.g., ΔSPLres>3dB), the parameter self-learning mechanism is automatically triggered: first, the deviation features (continuous high-frequency deviation, sudden pulse deviation, etc.) are extracted, and then the PID parameters are updated based on the gradient descent algorithm (Kp step size 0.05, Ki step size 0.01, Kd step size 0.02) to ensure the system control accuracy and avoid quality degradation caused by parameter drift.

[0142] Details of the gradient descent algorithm with parameter self-learning mechanism: The objective function of the algorithm is:

[0143] (Minimize the sum of squared residual biases), learning rate η = 0.01, number of iterations 20; PID parameter update formula is: ,

[0144] , The partial derivatives are calculated using the numerical difference method (step size = 0.01). The updated PID parameters must satisfy Kp ∈ [0.1, 0.5], Ki ∈ [0.01, 0.1], and Kd ∈ [0.05, 0.2]. If they exceed the range, boundary values ​​are used to ensure parameter stability. After self-learning... The probability is ≥95%;

[0145] The system synchronously records current operating conditions, including ambient temperature, door closing speed, and user closing force, and establishes an operating condition-parameter mapping table to provide a basis for rapid adjustment under different operating conditions. It also analyzes user door closing habits; for example, if 80% of users close the door at a speed between 0.8 and 1.0 m / s, the system preloads matching and optimization parameters before the next time that user closes the door, reducing response latency. By adapting operating condition records to user habits, the system ensures long-term stable operation and improves user experience consistency.

[0146] User door closing habit statistics and parameter preloading logic: User identity is associated with vehicle VIN code + vehicle system user account, the statistical period is 30 consecutive door closing operations by the user, and the statistical indicators include the average door closing speed. ), speed standard deviation ( ), average closing force ( ), standard deviation of force ( ), and will , Defined as the user's "habitual range"; the next time the door is closed, the system first detects the current ambient temperature, initial closing speed, and initial force. If all three fall within the "habitual range", the system reads the matching parameters from the mapping table and preloads them into the adaptive audio generation module. The preloading delay is ≤5ms, which is 60% faster than real-time calculation response speed. If the "habitual range" is exceeded, real-time calculation is triggered, and the new operating conditions and corresponding parameters are stored in the mapping table to update the "habitual range".

[0147] Example table of operating condition-control parameter mapping

[0148] Ambient temperature Closing speed (m / s) Ambient noise (dB) PID parameters (Kp / Ki / Kd) Base boost audio frequency (Hz) Enhanced harmonic frequency (Hz) Noise reduction compensation frequency band (Hz) Noise reduction amplitude (dB) Speaker gain (dB) 25℃(normal temperature) 0.9 70 (Medium Noise) 0.3 / 0.05 / 0.1 350 117 (fundamental wave 1 / 3) 1000-2000 8 6 -10℃ (low temperature) 1.1 55 (Low Noise) 0.2 / 0.03 / 0.08 380 127 (fundamental wave 1 / 3) 800-1800 5 4 40℃ (high temperature) 0.7 85 (High Noise) 0.45 / 0.08 / 0.15 320 107 (fundamental wave 1 / 3) 1200-2200 12 8 0℃ 1.2 (Threshold) 65 (Medium Noise) 0.35 / 0.06 / 0.12 360 120 (fundamental wave 1 / 3) 900-1900 7 5

[0149] Note: The parameters in the table are based on the fitting of measured data from 100+ vehicle models. The amplitude of the enhanced harmonic wave is fixed at 1 / 3 of the fundamental wave amplitude. The speaker gain is based on the characteristic calibration of the two speakers built into the door (mid-bass Pioneer TS-G1620F, tweeter Pioneer TS-T110).

[0150] This invention also provides an active sound generation method for optimizing the quality of door closing sound. Based on the above system, the specific steps of this method are as follows:

[0151] S1. Real-time monitoring and data fusion: Step S1 is activated when the door opening angle exceeds 30°. The sensor array synchronously samples the data, and the preprocessed data through Kalman filtering, Fourier transform and beamforming algorithm is spatiotemporally registered. Multi-dimensional acoustic state vectors are fused to ensure the reliability and timeliness of multi-source data.

[0152] S2, Difference Calculation and Strategy Generation: Based on the acoustic state vector in step S1, calculate the frequency deviation and sound pressure deviation. When generating dynamic control parameters, combine the proportional, integral, and derivative three-term adjustment logic of the PID algorithm with the environment adaptation strategy for deviation analysis and control strategy formulation, and output the control parameter set.

[0153] S3. Dynamic sound generation control and feedback adjustment: The adaptive audio generation module synthesizes audio in real time according to the parameter set, and outputs it to the speaker. The acoustic feedback correction module starts to evaluate the effect. If the residual deviation exceeds 3dB, the preset parameter matrix is ​​finely adjusted. The user's door closing force is recorded and stored in the database. The database parameters are called the next time the user closes the door to optimize and adapt the door closing sound quality.

[0154] The specific steps of real-time monitoring and data fusion in step S1 are as follows:

[0155] Step S11 is the system startup and data preprocessing stage. The trigger condition is that the door opening angle exceeds 30° (detected by the angle sensor at the door hinge). After activation, the sensor array synchronously collects data at a sampling rate of 500Hz. In the raw data preprocessing stage: the distance signal is filtered by Kalman (process noise covariance Q=0.01, observation noise covariance R=0.1) to remove mechanical vibration interference and obtain a smooth distance-time curve; the acceleration signal is processed by 1024-point Fast Fourier Transform (FFT) to extract characteristic frequencies, such as the 200-300Hz natural frequency of the door, to distinguish effective signals from noise; the microphone signal is focused on the door closing sound source by a delay-summation beamforming algorithm (spatial resolution ±5°) to suppress external noise crosstalk; the temperature and humidity signals are filtered by a 5-point moving average to eliminate instantaneous fluctuations.

[0156] Specific implementation scheme for spatiotemporal registration:

[0157] Time synchronization mechanism: The CAN bus time triggering protocol (TTCAN) is adopted, and the hardware timestamp of the distance sensor (VL53L1X) is used as the reference clock (accuracy ±0.1ms). Each sensor (triaxial accelerometer ADXL345, microphone array, temperature and humidity SHT30) receives the reference clock message through the CAN bus and calibrates the local clock in real time with a calibration period of 1ms to ensure that the sampling time deviation of multi-sensor data is ≤±1ms.

[0158] Spatial coordinate calibration: Establish a three-dimensional Cartesian coordinate system with the center of the door lock latch as the origin. The coordinates are defined as follows:

[0159] X-axis: Along the longitudinal direction of the vehicle body, pointing towards the front of the vehicle is the positive direction;

[0160] Y-axis: Perpendicular to the vehicle body plane, with the positive direction pointing outwards;

[0161] Z-axis: Along the height of the vehicle body, pointing towards the roof is the positive direction.

[0162] The specific coordinates of each sensor (unit: mm) are shown in the table below:

[0163] Sensor type Specific model / specification Installation location description 3D coordinates (unit: mm) Notes (Based on coordinate settings) Laser distance sensor VL53L1X (±0.5mm accuracy) Next to the car door latch X=0, Y=20, Z=50 With the center of the door latch as the origin (X=0, Y=0, Z=0), and since it is installed next to the latch, the X / Y / Z coordinates are close to the origin. Y=20mm is the compensation value for the gap between the latch and the door. Triaxial accelerometer ADXL345 (±16g range) Reinforcing ribs on the inner door panel (approaching the door lock at approximately 1 / 3 of the way down) X=150Y=-30Z=100 Based on the typical location design of the door inner panel reinforcing rib, X=150mm is the longitudinal distance from the latch to 1 / 3 of the reinforcing rib, and Y=-30mm is the inner side of the door inner panel (the inner side of the vehicle is the negative Y direction). A-pillar microphone Sensitivity -38dBV / Pa Lower part of the A-pillar inside the car (near the door) X=150Y=200Z=350 The longitudinal distance between the lower part of the A-pillar and the door latch is approximately 150mm. Y=200mm is the lateral distance from the A-pillar to the door (the outer side of the vehicle is the positive Y direction). Z=350mm is the height of the lower part of the A-pillar. B-pillar microphone Sensitivity -38dBV / Pa Lower part of the B-pillar inside the car (near the door) X=450Y=200Z=350 The longitudinal distance between pillar B and pillar A is approximately 300mm (150mm + 300mm = 450mm). Other coordinates are consistent with the microphone on pillar A to ensure symmetrical sound field sampling. Door interior panel microphone 1 Sensitivity -38dBV / Pa Upper part of the door trim panel (near the window) X=300Y=-100Z=400 The longitudinal distance from the center of the door trim panel to the latch is approximately 300mm. Y=-100mm is the inner side of the trim panel (negative Y direction of the inner side of the vehicle), and Z=400mm is the height of the lower edge of the window. Door panel microphone 2 Sensitivity -38dBV / Pa Lower part of the door trim panel (near the bottom of the door) X=300Y=-100Z=200 Aligning with the microphone on the upper part of the interior panel on the same vertical axis (X=300mm), and with Z=200mm at the bottom of the door, ensures coverage of the upper and lower sound fields. Temperature and humidity sensor SHT30 (±0.5℃ / ±3%RH) Next to the door seal (near the latch) X=200Y=-15 The longitudinal distance between the area near the sealing strip latch and the latch is approximately 200mm. Y=-15mm is the inner side of the sealing strip (close to the door sheet metal), and Z=150mm is the height of the center of the sealing strip.

[0164] The specific implementation of door opening angle detection: The angle sensor is model AS5600 (12-bit resolution), installed at the door hinge, coaxial with the hinge axis, with a measurement range of 0-360° and an accuracy of ±0.0879°; the activation logic is as follows: when the sensor detects that the angle value continuously increases from 0° (door fully closed) to more than 30°, and the holding time is ≥100ms (to avoid false triggering due to vibration), an activation signal is output to the sensor array; the activation signal uses high and low level signals (high level 5V, low level 0V), the signal transmission delay is ≤1ms, and the activation success rate is ≥99.9%;

[0165] The fusion process of 128-dimensional acoustic state vectors is as follows: First, the validity of preprocessed distance, acceleration, microphone, and temperature / humidity data is verified (data outside the valid range, such as distances <0cm or >50cm, is removed); second, spatiotemporal registration is performed (time registration is based on the distance sensor timestamp, and spatial registration is based on the latch); third, fusion is performed according to the weighted average of "motion parameters (distance, acceleration) 40%, acoustic parameters (microphone signal) 40%, and environmental parameters (temperature and humidity) 20%", using the following formula: Where V is a 128-dimensional vector, The motion parameter subvector (40 dimensions). The acoustic parameter subvector (40-dimensional). The environmental parameter sub-vector is 8-dimensional, and the remaining 40 dimensions are redundant verification dimensions; each dimension of the fused vector is labeled with the data source and accuracy to ensure traceability.

[0166] S12, the environmental acoustic modeling module performs spatiotemporal registration on the preprocessed multi-source data, with a time synchronization accuracy of ±1ms and a spatial coordinate calibration error of ≤2cm, ensuring the consistency of data from different sensors in time and space dimensions; then, it fuses data such as distance, acceleration, noise spectrum, temperature and humidity to generate an acoustic state vector containing 128 feature dimensions, of which dimensions 1-10 are noise spectrum features and 11-20 are motion parameter features;

[0167] The 128-dimensional acoustic state vector is further defined as follows: Dimensions 1-10 represent the noise spectrum amplitude in the 10Hz-100Hz, 100Hz-200Hz, ..., 900Hz-1000Hz (10 low-frequency bands); Dimensions 11-20 represent the real-time and peak values ​​of closing speed, acceleration, and angular acceleration (5 dimensions each); Dimensions 21-40 represent the noise spectrum amplitude in the 1000Hz-2000Hz, 2000Hz-3000Hz, ..., 19000Hz-20000Hz (20 high-frequency bands); Dimensions 41-50 represent the phase values ​​of noise in each frequency band; Dimensions 51-60 represent temperature and humidity values, temperature correction coefficient, humidity correction coefficient, and sound velocity; Dimensions 61-128 represent the door motion state characteristics (including 68 dimensions such as free closing time, sealing strip contact force, and locking impact time). The units, precision, and effective range of each dimension are clearly marked to ensure that the vector can be completely reproduced.

[0168] The specific steps for difference calculation and strategy generation in step S2 are as follows:

[0169] S21. Based on acoustic state vector, deviation analysis is performed, and frequency deviation is calculated: The peak detection algorithm (threshold is 80% of the maximum amplitude) is used to extract the actual main frequency f0 from the door closing sound spectrum, and it is compared with the default target main frequency ftarget (350Hz) to obtain Δf. When Δf>50Hz (e.g., f0=410Hz), frequency correction is initiated, and the basic enhancement tone main frequency is adjusted first; when Δf<-50Hz, the enhancement tone main frequency and harmonic frequency are increased simultaneously to ensure that the main frequency returns to the target range; sound pressure level adjustment: The actual sound pressure level SPLact is calculated by A-weighted calculation (conforming to human hearing) (measurement time constant 125ms), and ΔSPL=SPLact-80dB is obtained. When ΔSPL>5dB (e.g., SPLact=86dB), the compensation noise reduction enhancement mode is triggered, and the noise reduction amplitude value is increased to reduce the actual sound pressure;

[0170] S22. When finally generating dynamic control parameters, the proportional, integral, and derivative adjustment logic of the PID algorithm is combined with the environment adaptation strategy of the fuzzy logic module. For example, in a medium-noise environment, Kp=0.3, Ki=0.05, and Kd=0.1 are set. The parameter set includes key information such as enhanced audio frequency, noise reduction amplitude, and speaker gain, which can be specifically adapted to different environments and motion conditions: in low noise, the focus is on fine adjustment of enhanced audio, and in high noise, the noise reduction effect is strengthened to ensure that the control strategy is highly matched with the actual needs.

[0171] Operating condition adaptation priority of control parameters: When there is a conflict between environmental operating conditions and motion operating conditions (such as low noise but high speed door closing), the adaptation priority is "motion safety > sound quality optimization > environmental adaptation", that is, first ensure that the closing speed is ≤1.2m / s (by adjusting the enhanced audio frequency to suppress sharp noise), then optimize the sound pressure level and main frequency, and finally adapt to the environmental noise; the priority determination is achieved through "operating condition weight score", with motion operating condition weight of 0.5, sound quality weight of 0.3, and environmental operating condition weight of 0.2. The parameter combination with the highest total score is the final output. The weight score takes ≤0.5ms to ensure fast adaptation.

[0172] The specific steps for dynamic sound control and feedback adjustment in step S3 are as follows:

[0173] S31 and step S3 are the control strategy execution stage. The adaptive audio generation module synthesizes audio in real time according to the parameter set: the basic enhancement sound generates a sine wave at a set frequency (e.g., 350Hz), superimposed with a 100Hz low-frequency harmonic, and controls the sound pressure level at 75-80dB through 0.5dB step volume adjustment; the noise compensation and reduction generates an inverse sound wave in the 1000-2000Hz range according to the ambient noise spectrum; after the two types of signals are synthesized, they are output to the two speakers in the car door: the mid-bass unit is responsible for the 250-500Hz frequency band (frequency response 80-5000Hz), and the tweeter unit is responsible for the 500-600Hz frequency band (frequency response 500-20000Hz), to achieve accurate sound playback, and optimizes the quality after superimposing the actual door closing sound.

[0174] Synchronous control of audio synthesis and speaker driving: The adaptive audio generation module and the speaker driver adopt a "clock synchronization-signal enable" mechanism. The audio signal generated by the DSP is transmitted to the power amplifier through the I2S bus. The transmission clock is synchronized with the operating clock of the power amplifier (44.1kHz) to ensure that the signal has no phase offset. The enable signal of the power amplifier is linked with the door lock opening signal. When the door opening angle exceeds 30°, the power amplifier is enabled. The audio signal transmission delay is ≤2ms. The speaker starts to emit sound within 1ms after receiving the signal, realizing a seamless connection between "parameter generation and sound emission".

[0175] Frequency band generation and signal synthesis of reverse acoustic waves: Based on the ambient noise spectrum, the 1000-2000Hz range is divided into 100Hz sub-bands. Reverse acoustic waves are generated separately for each sub-band, and the amplitude is determined according to the noise amplitude of that sub-band (matching error ≤1%). The signal synthesis adopts the "sub-band superposition - overall filtering" process. First, the reverse acoustic waves of each sub-band are superimposed, and then a second-order Butterworth filter (cutoff frequencies 1000Hz and 2000Hz) is used to filter out out-of-band noise. The out-of-band attenuation of the filtered signal is ≥40dB. The synthesized audio signal is converted into an analog signal by a 16-bit DAC and output to a power amplifier. The conversion error is ≤0.1LSB to ensure the noise reduction effect of the reverse acoustic waves.

[0176] S32. The trigger signal for the completion of the door closing action is the door lock module locking signal (detecting that the bolt is fully engaged). At this time, the acoustic feedback correction module starts the effect evaluation: it collects the sound output of the speaker and compares it with the target acoustic model to calculate the residual deviation, such as Δfres=5Hz, ΔSPLres=2dB; if Δres>3dB (such as ΔSPLres=4dB), the preset parameter matrix is ​​finely adjusted proportionally. For every 10Hz increase in deviation, the target main frequency is adjusted by 1Hz. At the same time, the user's closing force (calculated from the peak acceleration) is recorded and stored in the habit database; the next time the user closes the door, the database parameters are directly called to reduce the calculation delay and stabilize and improve the user experience.

[0177] The door lock module uses a Hall sensor (model A1324) to detect the position of the latch. When the latch is fully engaged (displacement ≥ 5mm), the Hall sensor outputs a high level (5V). This signal is transmitted to the acoustic feedback correction module via the CAN bus with a transmission delay ≤ 3ms. The triggering logic is as follows: when a high-level signal is detected for 50ms (to avoid false triggering due to latch jitter), the effect evaluation is initiated. The evaluation process includes "microphone acquisition (100ms) - deviation calculation (50ms) - parameter fine-tuning (50ms)", with a total time of ≤ 200ms, which does not affect the next door closing operation.

[0178] Figure 4 This is a block diagram of an electronic device for an active sound generation method that optimizes the quality of door closing sound, provided by one or more embodiments of the present invention.

[0179] like Figure 4 As shown, this application provides an electronic device, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0180] The memory stores a computer program that, when executed by the processor, causes the processor to perform steps of an active sound generation method to optimize the quality of the door closing sound.

[0181] This application also provides a computer-readable storage medium storing a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of an active sound generation method for optimizing the quality of a door closing sound.

[0182] This application also provides a simulation testing platform, including:

[0183] Electronic equipment, for implementing steps of an active sound generation method to optimize the quality of a door closing sound;

[0184] The processor runs a program that, when running, executes steps of an active sound generation method that optimizes the quality of the door closing sound from data output by the electronic device.

[0185] A storage medium for storing a program that, when running, executes steps of an active sound generation method to optimize the quality of a door closing sound based on data output from an electronic device.

[0186] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0187] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An active sound production system for optimizing the quality of a closing sound, characterized in that The system comprises: A multi-dimensional sensor array module: including a distance sensor, a three-axis acceleration sensor, a distributed microphone array, a temperature and humidity sensor, the sensors synchronously collect data, and the data is transmitted through a CAN bus to ensure data timeliness and consistency; An environmental acoustic modeling module: performs noise spectrum analysis on the microphone signal, evaluates the door movement state, processes distance and acceleration data, calculates a temperature correction coefficient, dynamically adjusts the sound pressure parameter, and periodically updates the environmental parameter matrix to reflect the acoustic characteristics under different working conditions; A PID-fuzzy hybrid algorithm module: adopts a PID and fuzzy logic fusion architecture, presets a target frequency, a maximum sound pressure level, and a door closing speed threshold, calculates frequency, sound pressure, and speed deviations, and a fuzzy logic module dynamically adjusts the PID parameters according to the noise level, and the weighted superposition of the proportional term, the integral term, and the differential term is used for the regulation of the door closing sound quality; An adaptive audio generation module: uses a 16-bit / 44.1kHz sampling frequency to generate enhanced sound and noise reduction sound compensated by adaptive noise cancellation technology, which is synthesized by a digital signal processor and then output to a power amplifier to drive a loudspeaker for dual optimization of enhancement and noise reduction; An acoustic feedback correction module: uses the output sound signal collected by the microphone as feedback to calculate the frequency and sound pressure residual deviations, and automatically triggers the parameter self-learning mechanism to update the PID parameters when the deviation exceeds the threshold, records the working conditions to establish a mapping table, and loads the matching optimized parameters in advance before the next door closing to verify the control effect.

2. An active sound production system for optimizing the quality of a closing sound according to claim 1, characterized in that The multi-dimensional sensor array module comprises: A distance sensor for calculating the door closing speed; a three-axis acceleration sensor for detecting the door closing angular acceleration rate; a distributed microphone array for collecting environmental noise and door closing sound; and a temperature and humidity sensor for compensating for the influence of the environment on the acoustic parameters; The sensors synchronously collect data, which is transmitted through a CAN bus, and the microphone array uses beamforming technology to locate the door closing sound source area; the temperature and humidity sensor is used to correct the influence of temperature on the elastic coefficient of the sealing strip and the influence of humidity on the sound propagation speed, and outputs multi-dimensional raw data.

3. An active sound production system to optimize the quality of a closing sound according to claim 2, characterized in that The environmental acoustic modeling module comprises: Full-bandwidth decomposition of the microphone signal is performed through fast Fourier transform to identify the frequency components of the environmental noise; Distance and acceleration data are processed through Kalman filtering to fit the speed or acceleration curve and determine the key stages of the door closing process; In combination with the temperature and humidity data and the material acoustic characteristic database, the sound velocity parameter is dynamically adjusted according to the temperature under the working conditions of -30°C to 50°C, and the temperature correction coefficient is calculated; The environmental parameter matrix is periodically updated, which includes the key information of the noise main frequency, the door closing speed, and the temperature coefficient, and through high-frequency updating and multi-dimensional parameter integration, the acoustic characteristics under different working conditions are reflected.

4. An active sound production system for optimizing the quality of a closing sound according to claim 3, characterized in that The PID-fuzzy hybrid algorithm module comprises: An improved PID and fuzzy logic module fusion architecture is adopted to preset the target frequency range, the maximum sound pressure level, and the door closing speed threshold; the algorithm calculates the frequency deviation, the sound pressure deviation, and the speed deviation in real time; The fuzzy logic module dynamically adjusts PID parameters according to the environmental noise level, reduces the proportional coefficient when the noise is low, increases the proportional coefficient when the noise is high, eliminates the steady-state deviation through the integral coefficient, and suppresses the burst interference through the differential coefficient; the control output is generated by weighted superposition of the proportional term, the integral term and the differential term, and finally the control parameter vector is output.

5. An active sound production system to optimize the quality of a closing sound according to claim 4, characterized in that The adaptive audio generation module comprises: According to the algorithm control parameters, generate enhanced sound and noise reduction sound, the enhanced sound is mainly composed of adjustable sine wave, superimposed with low frequency harmonic: when the frequency deviation is greater than 0, the main frequency of the enhanced sound is reduced, and when the frequency deviation is less than 0, the main frequency of the enhanced sound is increased; the compensation noise reduction sound adopts adaptive noise cancellation technology to generate sound waves with opposite phase and matched amplitude to the noise; After the two types of audio signals are synthesized by the digital signal processor, they are output to the power amplifier to drive the speaker to work, and the bass unit is responsible for the 250-500Hz frequency band, and the treble unit is responsible for the 500-600Hz frequency band.

6. An active sound production system to optimize the quality of a closing sound according to claim 5, characterized in that The acoustic feedback correction module comprises: The control effect is checked through a closed-loop feedback mechanism, the output sound signal collected by the microphone array is used as the feedback source, and the frequency and sound pressure residual deviation are calculated; when the deviation exceeds the threshold, the parameter self-learning mechanism is automatically triggered, the deviation characteristics are extracted first, and then the PID parameters are updated based on the gradient descent algorithm; Synchronize recording the current working condition, including the ambient temperature, the door closing speed, the user's door closing force, establish the working condition and parameter mapping table, and at the same time, statistics the user's door closing habit, before the next time the user closes the door, the system loads the matching optimized parameters in advance.

7. An active sound production method for optimizing the quality of a closing sound, based on the system of claim 6, characterized in that, The specific steps of the method are as follows: S1, real-time monitoring and data fusion: step S1 is activated when the door opening angle exceeds 30°, the sensor array synchronously samples data, the preprocessed data are subjected to Kalman filtering, Fourier transform and beam forming algorithm, time and space are registered, multi-dimensional acoustic state vectors are fused to ensure the reliability and timeliness of multi-source data; S2, difference calculation and strategy generation: based on the acoustic state vector of step S1, the frequency deviation and the sound pressure deviation are calculated, and the dynamic control parameter is generated, combined with the proportional, integral and differential three adjustment logics of the PID algorithm and the environmental adaptation strategy, for deviation analysis and control strategy formulation, output the control parameter set; S3, dynamic sound control and feedback adjustment: the adaptive audio generation module synthesizes audio in real time according to the parameter set, and outputs the synthesized audio to the loudspeaker, and the acoustic feedback correction module starts the effect evaluation, and if the residual deviation exceeds 3dB, the preset parameter matrix is fine-tuned, the user's door closing force is recorded and stored in the database, and the next time the user closes the door, the database parameters are called to realize the optimization and adaptation of the door closing sound quality.

8. The active sound production method of claim 7, wherein, The specific steps of real-time monitoring and data fusion in step S1 are as follows: S11, step S1 is the system startup and data preprocessing stage, the trigger condition is that the door opening angle exceeds 30°, and after being activated, the sensor array synchronously collects data; The distance signal is subjected to Kalman filtering to obtain a smooth distance and time curve; The acceleration signal is subjected to fast Fourier transform to extract the characteristic frequency, and the effective signal and noise are distinguished; The microphone signal is positioned by delay-sum beam forming algorithm to suppress the external noise crosstalk; The temperature and humidity signal is subjected to 5-point sliding average filtering to eliminate instantaneous fluctuations; S12, the environmental acoustic modeling module performs spatio-temporal registration on the pre-processed multi-source data, fuses data including distance, acceleration, noise spectrum, and temperature and humidity, and generates a multi-dimensional acoustic state vector integrating multi-dimensional environmental and motion data.

9. The active sound production method of claim 8, wherein, The specific steps of the difference calculation and strategy generation in step S2 are as follows: S21, based on the acoustic state vector, deviation analysis is completed, frequency deviation calculation uses a peak detection algorithm to extract the actual main frequency from the closing door sound spectrum, and compares it with the default target main frequency to obtain the frequency deviation; sound pressure level adjustment calculates the actual sound pressure level by A-weighting weighting; S22, combined with the proportional, integral, and differential three adjustment logics of the PID algorithm, and the environment adaptation strategy of the fuzzy logic module, dynamic control parameters are generated; the parameter set contains key information of enhanced audio frequency, noise reduction amplitude, and loudspeaker gain, and is adapted to different environments and motion conditions, and focuses on fine adjustment of enhanced audio in low noise and intensifies noise reduction effect in high noise.

10. The active sound production method of claim 9, wherein, The specific steps of the dynamic sound generation control and feedback adjustment in step S3 are as follows: S31, the adaptive audio generation module synthesizes audio in real time according to the parameter set, generates a sine wave at a set frequency for enhanced audio, and superimposes a low-frequency harmonic wave; compensation noise is synthesized according to the environmental noise spectrum, and the two types of signals are output to the loudspeaker: the bass unit is responsible for the 250-500Hz frequency band, and the treble unit is responsible for the 500-600Hz frequency band; S32, the trigger signal for the completion of the door closing action is the door lock module locking signal, the sound output by the loudspeaker is collected, compared with the target acoustic model, and the residual deviation is calculated, if the residual deviation is more than 3dB, the preset parameter matrix is adjusted in proportion, and the user's closing force is recorded and stored in the habit database; next time the user closes the door, the database parameters are directly called.

Images