Multi-area independent sound field control method and system for automobile sound equipment

Abstract

The invention discloses an automobile audio multi-area independent sound field control method and system, and relates to the technical field of audio control, and the method comprises the steps: interacting a cabin management end of a target vehicle, obtaining the real-time state information of human body interaction equipment, carrying out the human body state sensing based on the real-time state information, obtaining a human body state sensing result, and transmitting the human body state sensing result to a control terminal; comprising a human body existence state and a human body behavior state. And by taking the human body state sensing result as a constraint condition, matching a preset sound field control strategy and carrying out strategy initialization to generate a reference sound field control strategy. And constructing a simulation control model based on the internal environment structure information and the sound layout information of the target vehicle, and performing path tracking-based control optimization on the reference sound field control strategy. And implementing multi-area independent sound field control according to an optimization result, and recording control parameters and corresponding human body existence states. Therefore, the technical effects of providing a multi-dimensional personalized control path and improving the sound field control accuracy are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of audio control, and in particular to a method and system for controlling multi-zone independent sound fields of automobile audio. Background Art

[0002] Traditional car audio systems focus on optimizing the overall sound field, usually adjusting the sound effects through fixed parameters or simple user settings. However, as consumers' demand for personalized and intelligent experience increases, a single sound field control method can no longer meet the diverse needs of different users in different scenarios. In particular, in terms of multi-zone independent sound field control, existing technologies have not yet provided an effective solution. Summary of the invention

[0003] The present invention provides a method and system for controlling multi-zone independent sound fields of automobile audio systems, so as to solve the technical problems in the prior art of single control form, lack of personalized real-time adjustment, and influence on control accuracy, and realize the technical effect of providing multi-dimensional personalized control paths and improving the accuracy of sound field control.

[0004] In a first aspect, the present invention provides a method for controlling a multi-zone independent sound field of a car audio system, wherein the method for controlling a multi-zone independent sound field of a car audio system comprises: The cockpit management end of the interactive target vehicle obtains the real-time status information of the human interaction device.

[0005] Based on the acquired real-time status information, human body status perception is performed to obtain human body status perception results, wherein the human body status perception results include human body existence status and human body behavior status.

[0006] Taking the human body state perception result as a constraint condition, matching a preset sound field control strategy and initializing the strategy, the output is a reference sound field control strategy.

[0007] Based on the internal environment structure information and the audio layout information of the target vehicle, a simulation control model is constructed, and the reference sound field control strategy is optimized based on path tracking.

[0008] Based on the control optimization results, multi-zone independent sound field control is performed, and the control parameters and the corresponding human presence status are recorded.

[0009] In a feasible implementation, the cockpit management terminal of the interactive target vehicle obtains the real-time status information of the human interaction device, including: The seat pressure distribution information of the seat area is collected through the pressure distribution sensor assembly.

[0010] The restraint state sensor assembly is activated based on the pressure distribution data, the seat inclination data and the seat belt state data are acquired, and the data are output as seat restraint information.

[0011] According to the seat pressure distribution information and in combination with the privacy authorization status, the acoustic collection component is activated to obtain the partition voice information.

[0012] The seat pressure distribution information, the seat restraint information and the partition voice information are used as the real-time status information.

[0013] In a feasible implementation, human body state perception is performed based on the acquired real-time state information to obtain a human body state perception result, wherein the human body state perception result includes a human body existence state and a human body behavior state, including: The seat pressure distribution information is mapped to a preset seat grid coordinate system to generate a pressure value distribution matrix.

[0014] The centroid coordinates of the pressure value distribution matrix in the horizontal and vertical dimensions are calculated to obtain the pressure center offset.

[0015] The seat belt height, seat cushion pitch angle and seat back pitch angle in the seat restraint information are synchronously acquired.

[0016] The pressure center offset, the safety belt height, the seat cushion pitch angle and the seat back pitch angle are input into a convolutional neural network model trained by transfer learning, and the human body posture classification result and the auditory center height are output as the human body existence state.

[0017] In a feasible implementation, human body state perception is performed based on the acquired real-time state information to obtain a human body state perception result, wherein the human body state perception result includes a human body existence state and a human body behavior state, including: The partitioned speech data is subjected to short-time Fourier transform, the frequency spectrum center of gravity and root mean square amplitude in the 0.5-4kHz frequency band are extracted, and a two-dimensional feature vector is constructed.

[0018] Based on a preset excitement grading strategy, the excitement level corresponding to the two-dimensional feature vector is determined.

[0019] The fluctuation variance of the pressure distribution data within a continuous preset time window is calculated, and when the fluctuation variance exceeds a preset jitter threshold, the current excitement level is increased by one level.

[0020] When the weighted sum of the seat back pitch angle and the seat cushion pitch angle is greater than a preset threshold, an attenuation factor is applied to the current excitement level and output as the human behavior state.

[0021] In a feasible implementation, the human body state perception result is used as a constraint condition, a preset sound field control strategy is matched and the strategy is initialized, and the output is a reference sound field control strategy, including: According to the human body existence state and human body behavior state, a preset sound field control strategy is matched in the HRTF database.

[0022] The human body weight data is estimated based on the human body state sensing result, and the low-frequency gain is adjusted according to the human body weight data to compensate for vibration absorption.

[0023] A multivariate regression model based on sample sound and image data is constructed, and the optimal sound and image height is predicted by combining the human body state perception result and the human body weight data, wherein the sample sound and image data includes the associated stored sample weight, sample auditory center height, sample human body posture classification result and sample optimal sound and image height.

[0024] In a feasible implementation, based on the internal environment structure information and the sound layout information of the target vehicle, a simulation control model is constructed, and the reference sound field control strategy is optimized based on path tracking, including: The pre-stored three-dimensional model of the vehicle interior structure and the acoustic material property parameters are loaded to construct a simulation control model, and when a change in the opening and closing state of the window is detected, the boundary reflection condition of the simulation control model is dynamically adjusted.

[0025] Access the environmental temperature and humidity sensor data in real time, and correct the calculated value of the sound wave propagation speed accordingly.

[0026] A reflection number constraint is set, a propagation path tracking of the reference sound field control strategy is performed according to the reflection number constraint, and a uniformity index and a consistency index of the sound field are calculated based on the propagation path tracking result.

[0027] The weighted values ​​of the uniformity index and the consistency index are used as optimization targets, and the reference sound field control strategy is iteratively optimized in combination with an optimization algorithm.

[0028] In a feasible implementation, the present invention further includes: The human body's existence status is memorized and stored in association with control parameters to build a control record library.

[0029] Statistical analysis is performed on the cumulative triggering duration, cumulative triggering frequency, and cumulative triggering frequency of each human body existence state.

[0030] According to the preset normal judgment constraint set, the cumulative trigger duration, the cumulative trigger frequency and the cumulative trigger frequency are respectively subjected to threshold judgment. When any two or three of them meet the threshold, multiple normal control plans are constructed according to the corresponding control records.

[0031] In a feasible implementation, it also includes: dynamically adjusting the normal control plan based on the active control record and a preset update strategy, wherein the update strategy includes a periodic strategy and a trigger strategy.

[0032] In a feasible implementation, the present invention further includes: The adjustment frequency and the adjustment frequency of dynamically adjusting the normal control plan are accumulated.

[0033] When it is detected that either the adjustment frequency or the adjustment frequency exceeds the adjustment limit within a unit time, a dynamic adjustment record corresponding to the accumulated time window in the historical adjustment record is extracted.

[0034] Based on the dynamic adjustment record, a parameter offset sequence is extracted for fitting analysis, a strategy optimization function is generated, and the preset sound field control strategy is updated.

[0035] Based on the updated sound field control strategy, the control parameter adjustment amount of the normal control plan is recalculated and updated.

[0036] In a second aspect, the present invention further provides a multi-zone independent sound field control system for a car audio system, wherein the multi-zone independent sound field control system for a car audio system comprises: The status acquisition module is used to interact with the cockpit management end of the target vehicle to obtain the real-time status information of the human interaction device.

[0037] The human body state perception module is used to perform human body state perception based on the acquired real-time state information, and obtain human body state perception results, wherein the human body state perception results include human body existence state and human body behavior state.

[0038] The strategy initialization module is used to match the preset sound field control strategy and initialize the strategy based on the human body state perception result as a constraint condition, and output it as a reference sound field control strategy.

[0039] The simulation optimization module is used to construct a simulation control model based on the internal environment structure information and the audio layout information of the target vehicle, and perform control optimization based on path tracking on the reference sound field control strategy.

[0040] The sound field control module is used to perform multi-zone independent sound field control according to the control optimization results, and record the control parameters and the corresponding human body presence status.

[0041] The present invention discloses a method and system for controlling a multi-zone independent sound field of an automobile audio system, comprising: a cockpit management terminal of an interactive target vehicle, obtaining real-time status information of a human body interactive device, and based on the information, performing human body status perception to obtain a human body presence status and a human body behavior status; using a human body status perception result as a constraint condition, matching and initializing a preset sound field control strategy to generate a reference sound field control strategy; combining the internal environment structure information and the audio layout information of the target vehicle, constructing a simulation control model, and performing control optimization based on path tracking on the reference sound field control strategy; finally, realizing multi-zone independent sound field control according to the optimization result, and recording control parameters and corresponding human body presence status. The method and system for controlling a multi-zone independent sound field of an automobile audio system disclosed by the present invention solve the technical problems of single control form, lack of personalized real-time adjustment, and influence on control accuracy, and realize the technical effect of providing a multi-dimensional personalized control path and improving the accuracy of sound field control. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 The present invention is a flowchart of a method for controlling multi-zone independent sound fields of a car audio system.

[0043] Figure 2 The present invention is a structural schematic diagram of a multi-zone independent sound field control system for car audio.

[0044] Explanation of reference numerals: state acquisition module 11 , human body state perception module 12 , strategy initialization module 13 , simulation optimization module 14 , sound field control module 15 . DETAILED DESCRIPTION

[0045] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention are shown in the drawings, rather than all of them.

[0046] Embodiment 1, as Figure 1 The present invention is a flowchart of a method for controlling a multi-zone independent sound field of a car audio system, wherein the method for controlling a multi-zone independent sound field of a car audio system comprises: S100: The cockpit management terminal of the interactive target vehicle obtains the real-time status information of the human body interactive device.

[0047] Specifically, the real-time status information includes seat pressure distribution information, seat restraint information and partition voice information, which is used for matching human state perception and sound field control strategy. Through multimodal data acquisition (pressure distribution, seat restraint, partition voice), it can break through the accuracy limitations of traditional single sensors, help to more accurately perceive the state of people in the car, and adjust the sound field control strategy accordingly, providing each passenger with a personalized audio experience.

[0048] In some embodiments, the cockpit management terminal of the interactive target vehicle obtains the real-time status information of the human body interactive device, including: The seat pressure distribution information of the seat area is collected through the pressure distribution sensor component; based on the pressure distribution data, the restraint status sensor component is activated to obtain the seat inclination data and the seat belt status data, and output as the seat restraint information; according to the seat pressure distribution information and in combination with the privacy authorization status, the acoustic collection component is activated to obtain the partition voice information; the seat pressure distribution information, the seat restraint information and the partition voice information are converted into the real-time status information.

[0049] Specifically, the pressure distribution sensor component is used to measure the pressure distribution in the seat area, and can capture detailed pressure information of the contact between the human body and the seat. The restraint state sensor component is used to detect the physical state of the seat, such as the seat inclination and seat belt state, and provide auxiliary data on the human behavior state. The acoustic collection component is used to collect voice information from different areas in the car, that is, to support partitioned voice recognition.

[0050] Specifically, before collecting voice information, it is necessary to obtain the user's privacy authorization to ensure the legality of data collection.

[0051] Specifically, first, the pressure distribution data of the seat area is collected through the pressure distribution sensor component, so as to detect the seat occupancy status (no one / single person / multiple people), identify the sitting posture of the occupants (normal sitting posture, tilted, half-lying, etc.), and analyze the body characteristics (weight distribution, posture changes) in the subsequent steps; then, the restraint status sensor component is activated to perform seat inclination detection and seat belt status detection, where the seat inclination includes the inclination of the seat cushion and the seat back, and the seat belt status includes the seat belt height and the seat belt usage; at the same time, with the user's authorization, the partitioned voice collection component is activated according to the seat pressure distribution information and the privacy authorization status to capture the occupant's voice information for subsequent emotional perception (judging the passenger status based on voice characteristics, such as fatigue, tension, excitement).

[0052] Through multi-modal data collection (pressure distribution, seat restraints, partitioned voice), the collection and integration of real-time status information provides comprehensive data support for the dynamic matching of sound field control strategies, improving the personalization and accuracy of sound field control. At the same time, the introduction of privacy authorization status ensures the legality and security of data collection.

[0053] S200: Performing human body state perception based on the acquired real-time state information to obtain a human body state perception result, wherein the human body state perception result includes a human body existence state and a human body behavior state.

[0054] Specifically, the human presence state refers to static information such as the position, posture and auditory center height of the people in the car; the human behavior state refers to the dynamic behavior of the people in the car, such as voice activity, excitement level, etc.

[0055] For example, in a smart cockpit, the driver's sitting posture changes are detected by the pressure distribution sensor, the seat inclination and seat belt status are obtained by the seat restraint sensor, and the driver's voice commands and excitement are recognized by the acoustic collection component. After these data are integrated, they can accurately perceive the driver's status and adjust the sound field control strategy accordingly to provide the driver with a personalized audio experience.

[0056] In some embodiments, human body state perception is performed based on the acquired real-time state information to obtain a human body state perception result, wherein the human body state perception result includes a human body existence state and a human body behavior state, including: The seat pressure distribution information is mapped to a preset seat grid coordinate system to generate a pressure value distribution matrix; the centroid coordinates of the pressure value distribution matrix in the horizontal and vertical dimensions are calculated to obtain the pressure center offset; the seat belt height, seat cushion pitch angle and seat back pitch angle in the seat constraint information are simultaneously obtained; the pressure center offset, the seat belt height, the seat cushion pitch angle and the seat back pitch angle are input into a convolutional neural network model trained by transfer learning, and the human body posture classification result and the auditory center height are output as the human body existence state.

[0057] Specifically, the seat grid coordinate system is a virtual coordinate system used to divide the seat surface into multiple grid points, where each grid point corresponds to the measurement value of a pressure sensor. The pressure value distribution matrix is ​​a matrix generated by mapping the seat pressure distribution information to the seat grid coordinate system, which is used to represent the pressure value of each grid point. The centroid coordinates are the coordinates of the center point of the pressure distribution matrix in the horizontal and vertical dimensions, which are used to determine the position of the pressure center. The pressure center offset is the offset of the centroid coordinates of the pressure distribution matrix relative to the preset reference point, which is used to judge changes in human body posture.

[0058] Specifically, first, the seat pressure distribution information is mapped, that is, the seat pressure data obtained by the pressure distribution sensor is mapped to the preset seat grid coordinate system to generate a pressure value distribution matrix to analyze the human body contact area and force conditions; then, the horizontal (X-axis) and vertical (Y-axis) centroid coordinates of the pressure value distribution matrix are calculated to obtain the X-axis offset of the pressure center (used to determine the left and right tilt) and the Y-axis offset of the pressure center (used to determine the front and rear tilt or semi-reclining state); at the same time, the seat restraint information is obtained synchronously, among which the seat belt height is collected to analyze the user's height and safety status, and the seat cushion pitch angle is collected to determine the user's pelvic inclination, and analyze whether it is tilted or lying. The seat back pitch angle is collected and combined with the pressure center offset to comprehensively judge the sitting posture (such as sitting upright, semi-reclining, and fully lying).

[0059] Furthermore, a convolutional neural network (CNN) model trained using transfer learning is used: Through transfer learning technology, fine-tuning is performed on the basis of the pre-trained convolutional neural network model, so that the model has the ability to classify human postures. The model input data includes pressure center offset, seat belt height, seat cushion pitch angle, and seat back pitch angle, and outputs human posture classification results (such as sitting upright, tilting, half-lying, and fully lying) and auditory center height (calculated based on head position to optimize the in-car sound effect). For example, in a smart cockpit, when the driver adjusts the seat posture, the pressure distribution sensor detects the pressure change and generates a pressure value distribution matrix. By calculating the center of mass coordinates, it is determined whether the driver's sitting posture has changed. At the same time, the seat restraint sensor provides seat belt height and seat angle information, which are jointly input into the convolutional neural network model, and the model outputs the driver's posture classification (such as sitting upright, tilted, relaxed, etc.) and auditory center height.

[0060] In the above steps, the combination of pressure center offset and seat constraint information provides more comprehensive data support for human posture classification. The convolutional neural network model trained by transfer learning improves the accuracy of human posture classification, so that it can dynamically adapt to the user's state changes and provide a reliable basis for the personalized adjustment of the sound field control strategy.

[0061] In some embodiments, human body state perception is performed based on the acquired real-time state information to obtain a human body state perception result, wherein the human body state perception result includes a human body existence state and a human body behavior state, including: Performing a short-time Fourier transform on the partitioned voice data, extracting the frequency spectrum center of gravity and the root mean square amplitude in the 0.5-4kHz frequency band, and constructing a two-dimensional feature vector; determining the excitement level corresponding to the two-dimensional feature vector based on a preset excitement grading strategy; calculating the fluctuation variance of the pressure distribution data in a continuous preset time window, and when the fluctuation variance exceeds a preset jitter threshold, raising the current excitement level by one level; when the weighted sum of the chair back pitch angle and the seat cushion pitch angle is greater than a preset threshold, applying an attenuation factor to the current excitement level, and outputting it as the human behavior state.

[0062] Specifically, the short-time Fourier transform is used to decompose a non-stationary signal into a series of locally stationary signals. By performing Fourier transform on the signal in a continuous time window, the spectrum information in the window can be obtained. Among them, the center of gravity of the spectrum indicates the "center" position of the spectrum distribution, reflecting the balance point of the signal energy distribution. For speech data, the center of gravity of the spectrum in the 0.5-4kHz frequency band is often used to reflect the energy distribution of vowels and consonants in the language. The root mean square amplitude is an amplitude quantitative indicator obtained by taking the square root of the average of the squares of the signal amplitudes. It is used to reflect the energy level and stability of the signal, that is, to quantify the strength of the signal.

[0063] Specifically, the excitement grading strategy is a preset rule or model used to map the input feature vector to the corresponding excitement level, such as low, medium, high, and extremely high. The pressure distribution data fluctuation variance is used to measure the pressure changes caused by the occupant's sitting posture stability or body movements. A larger fluctuation variance usually indicates an abnormal position change. When the fluctuation variance of the pressure distribution data exceeds the jitter threshold, it can be considered that the current occupant state fluctuates, and the excitement level is adjusted accordingly.

[0064] Specifically, first, real-time voice data is obtained from the voice acquisition module of each region to ensure that the data quality and sampling frequency meet the analysis requirements (for example, the sampling rate is 16kHz or higher), and the partitioned voice data is divided into frames of fixed length (for example, 20-40ms), and each frame of data is Fourier transformed using the short-time Fourier transform method to obtain a frequency domain representation. Then, the center of gravity and root mean square amplitude of the spectrum in the frequency band are calculated and combined into a two-dimensional feature vector, which reflects the energy distribution and stability of the voice signal in the 0.5-4kHz frequency band. Then, according to the preset excitement grading strategy (such as the pre-trained mapping function), the two-dimensional feature vector is mapped to the excitement level. For example, for a frame of voice data, the calculated two-dimensional feature vector is [2.5kHz, 0.8], corresponding to a higher center of gravity of the spectrum and a larger RMS value, which can be judged as "medium excitement" according to the preset rules.

[0065] Specifically, the seat pressure distribution data is preprocessed to generate a sequence of pressure values ​​within a continuous time window, and the variance of the pressure data is calculated within a preset time window; when the pressure data fluctuation variance exceeds the jitter threshold, the current excitement level is increased by one level to reflect the incentive response caused by the unstable state of the occupant.

[0066] Furthermore, the seat back pitch angle and seat cushion pitch angle of the seat are collected respectively, and weighted summed according to the preset weights to reflect the overall posture change of the occupant on the seat. If the weighted sum exceeds the preset threshold, it can be considered that the occupant is in a relatively relaxed or expected state of relaxation, so an attenuation factor is applied to the current excitement level (for example, the excitement is reduced by one level or multiplied by a coefficient less than 1).

[0067] The above steps can accurately capture the characteristics of human voice activity by extracting the spectral features of the voice signal through short-time Fourier transform. Combined with the fluctuation variance of the pressure distribution data and the seat angle information, the excitement level can be dynamically adjusted to achieve accurate perception of the human behavior state. This effectively integrates multi-source data such as voice, pressure and posture, improves the accuracy and real-time performance of in-vehicle human behavior state detection, and provides a solid data foundation for sound field control.

[0068] S300: using the human body state perception result as a constraint condition, matching a preset sound field control strategy and initializing the strategy, and outputting a reference sound field control strategy.

[0069] In some embodiments, the human body state perception result is used as a constraint condition, a preset sound field control strategy is matched and the strategy is initialized, and the output is a reference sound field control strategy, including: According to the human body existence state and human body behavior state, a preset sound field control strategy is matched in the HRTF database; based on the human body state perception result, the human body weight data is estimated, and the low-frequency gain is adjusted according to the human body weight data to compensate for vibration absorption; a multivariate regression model based on the sample sound and image data is constructed, and the optimal sound and image height is predicted by combining the human body state perception result and the human body weight data, wherein the sample sound and image data includes the associated stored sample weight, sample auditory center height, sample human body posture classification result and sample optimal sound and image height.

[0070] Specifically, the HRTF database, or Head-Related Transfer Function database, is used to store acoustic characteristic data related to the human head and ears to help achieve three-dimensional sound positioning. The sound field control strategy, including parameters such as audio signal gain, balance, delay, and virtual sound image positioning, is used to optimize the in-car auditory experience.

[0071] Specifically, the multivariate regression model is used to calculate and predict the optimal sound and image height by combining the human body state perception results and weight data, wherein the sample sound and image data is a data set including sample weight, sample auditory center height, sample human body posture classification results and sample optimal sound and image height, which is used to train the multivariate regression model.

[0072] Specifically, first, according to the human body state (such as posture classification results and auditory center height) and human behavior state (such as excitement level), the corresponding sound field control strategy is matched in the HRTF database. For example, if the occupant is detected to be "sitting upright" and the auditory center height is 1.2m, the sound field configuration parameters (speaker delay, surround sound level and balance parameters) corresponding to this state are selected in the HRTF database. Then, according to the seat pressure distribution information in the human body state perception result, the occupant weight is estimated using a statistical algorithm or a pre-trained model, and the low-frequency gain is adjusted according to the weight data to compensate for vibration absorption. For example, if the weight is large (for example, >70kg), it is considered that the occupant has a strong ability to absorb low-frequency vibrations, and the low-frequency gain needs to be increased to compensate.

[0073] Furthermore, a multivariate regression model based on sample sound and image data is constructed, the current human body state perception results and weight data are input into the regression model, the optimal sound and image height is predicted, and the initialized reference sound field control strategy is output.

[0074] For example, in the smart cockpit, the most suitable sound field control strategy is found in the HRTF database based on the driver's posture classification results and the height of the hearing center. If the driver is heavier, the low-frequency gain is increased to compensate for more vibration absorption. At the same time, the optimal sound image height is predicted through a multivariate regression model to ensure the accuracy of sound effect positioning.

[0075] By combining the human body state perception results and the HRTF database, the optimal sound field control strategy can be dynamically matched to ensure the personalization and accuracy of the sound effects. At the same time, the low-frequency gain compensation mechanism effectively solves the problem of sound effect differences caused by different body weights and improves the consistency of the sound effects. The introduction of the multivariate regression model makes the prediction of the optimal sound image height more accurate, further enhancing the positioning effect of the three-dimensional sound effects.

[0076] S400: Based on the internal environment structure information and the audio layout information of the target vehicle, a simulation control model is constructed to perform control optimization based on path tracking on the reference sound field control strategy.

[0077] Specifically, the simulation control model is a digital virtual model that is used to simulate the in-vehicle sound field environment, thereby optimizing the sound field in combination with the physical structure and acoustic characteristics of the vehicle. Path tracking is used to track the propagation path of sound waves in the vehicle and analyze acoustic phenomena such as reflection and diffraction, thereby reflecting the performance of the baseline sound field control strategy in the internal environment of the target vehicle.

[0078] Specifically, the sound field uniformity index is used to measure the uniformity of the sound field intensity in spatial distribution. The sound field consistency index is used to measure the temporal consistency of the sound field in different areas. By building a simulation control model and combining it with path tracking technology, the sound field control strategy can be optimized in real time to ensure the stability and consistency of the sound field under complex working conditions.

[0079] In some embodiments, based on the internal environment structure information and the sound layout information of the target vehicle, a simulation control model is constructed, and the reference sound field control strategy is optimized based on path tracking, including: The pre-stored three-dimensional model of the vehicle interior structure and the acoustic material property parameters are loaded to construct a simulation control model, and when the window opening and closing state changes are detected, the boundary reflection conditions of the simulation control model are dynamically adjusted; the environmental temperature and humidity sensor data are accessed in real time, and the calculated value of the sound wave propagation speed is corrected accordingly; the reflection number constraint is set, and the propagation path of the benchmark sound field control strategy is tracked according to the reflection number constraint, and the uniformity index and consistency index of the sound field are calculated based on the propagation path tracking results; the weighted values ​​of the uniformity index and the consistency index are used as optimization targets, and the benchmark sound field control strategy is iteratively optimized in combination with the optimization algorithm.

[0080] Specifically, the three-dimensional model of the vehicle interior structure is a three-dimensional digital model that represents the interior space of the vehicle, covering structural information such as the interior layout, cockpit, windows, dashboard, etc., including geometric parameters (size, shape, position) and material properties (density, sound absorption coefficient, reflection coefficient, etc.). Among them, the acoustic material property parameters are used to describe the acoustic properties of the surfaces of various components in the vehicle, such as sound absorption rate, reflectivity and transmission loss, which directly affect the propagation, reflection and attenuation of sound waves in the vehicle.

[0081] Specifically, first, load the three-dimensional model data of the target vehicle's internal environment structure from the pre-stored database, including the geometric information and installation position of each component, and load the acoustic parameters of each interior material, such as sound absorption and reflectivity, to use acoustic simulation software (such as COMSOL, EASE, etc.) to establish a simulation control model for in-vehicle sound wave propagation. Optionally, when a change in the window opening and closing state is detected, adjust the reflection conditions of the window part in the simulation model in real time (such as the reflectivity changes from higher to lower or vice versa) to reflect the actual changes in the in-vehicle sound field. Then, obtain the current in-vehicle ambient temperature and humidity data through the on-board temperature and humidity sensor. According to the changes in temperature and humidity, use the sound velocity correction formula (for example: the sound velocity is positively correlated with the temperature) to correct the calculated value of the sound wave propagation velocity in the simulation model to ensure that the model accurately reflects the actual sound field propagation characteristics.

[0082] Furthermore, according to the acoustic requirements inside the vehicle, an upper limit on the number of sound wave reflections is preset (for example, a maximum of 3 reflections are allowed), and then in the simulation control model, the path of the sound waves under the benchmark sound field control strategy is tracked, the propagation path of the sound waves in the vehicle is recorded, and the uniformity indicators (such as the standard deviation of the sound pressure level in each area) and consistency indicators (such as the average value of the output delay and phase difference of different speakers) are calculated based on the tracking results.

[0083] Furthermore, the uniformity index and the consistency index are set according to preset weights and the weighted summation result is set as the optimization objective function. The optimization algorithm (such as gradient descent, genetic algorithm or reinforcement learning) is used to iteratively adjust the sound field control parameters (such as speaker gain, delay, equalization parameters) so that the above-mentioned optimization objective function reaches the optimal state, and the optimized reference sound field control strategy is output for the dynamic control of the actual sound system.

[0084] The above method steps, by building a simulation control model and combining it with path tracking technology, can optimize the sound field control strategy in real time to ensure the stability and consistency of the sound field under complex working conditions. Among them, the dynamic adjustment of boundary reflection conditions and real-time environmental correction enables the model to adapt to changes in the in-vehicle environment and improve the adaptability and accuracy of sound field control. The application of the optimization algorithm further improves the sound field uniformity and consistency indicators to improve the user experience.

[0085] S500: Perform multi-zone independent sound field control according to the control optimization result, and record the control parameters and the corresponding human body presence status.

[0086] Specifically, first, according to the optimized sound field control strategy, the sound field parameters of different areas in the car, such as volume, frequency response and sound image positioning, are independently adjusted to meet the needs of different passengers. Then, the control parameters of each area, including volume, frequency response, sound image positioning, etc., are recorded, and the recorded control parameters are associated with the corresponding human presence state and stored to build a control record library for subsequent analysis and optimization.

[0087] For example, in a smart cockpit, the sound field parameters of the driver and passenger areas are adjusted based on the optimization results. The driver area may require stronger low-frequency sound effects to enhance the driving experience, while the passenger area may require softer mid- and high-frequency sound effects to provide a comfortable auditory environment. The control parameters of each area are recorded and stored in association with the human presence status such as the posture of the driver and passenger, the height of the hearing center, etc.

[0088] Through multi-zone independent sound field control, each passenger can be provided with a personalized sound experience and the comfort of the in-car sound field can be improved. The association between control parameters and human presence is recorded to provide data support for subsequent strategy optimization and user experience improvement. This data-driven optimization method not only improves the intelligence level of sound field control, but also enhances its self-learning ability, enabling it to continuously adapt to changes in user needs.

[0089] In some embodiments, the method for controlling multi-zone independent sound fields of a car audio system further includes: The human body existence state is memorized and stored in association with the control parameters to build a control record library; the cumulative trigger duration, cumulative trigger frequency and cumulative trigger frequency of each human body existence state are statistically analyzed; according to the preset normal judgment constraint set, the cumulative trigger duration, the cumulative trigger frequency and the cumulative trigger frequency are threshold judged respectively, and when any two or three of them meet the threshold, multiple normal control plans are constructed according to the corresponding control records.

[0090] Specifically, the control record library is a database used to store the human body status and corresponding control parameter data during historical interactions. The records include timestamps, human body status, application control parameters, environmental information, etc., providing data support for subsequent statistical analysis.

[0091] Specifically, the cumulative trigger duration refers to the cumulative duration of a human body state (such as roll) being detected and triggering the relevant control strategy within the preset time window. The cumulative trigger frequency refers to the number of times a human body state triggers detection within the preset time window, reflecting the frequency of occurrence of the state. The cumulative trigger frequency refers to the average number of times a human body state is triggered per unit time, which is usually equal to the cumulative trigger frequency divided by the observation time.

[0092] Specifically, the normal discrimination constraint set is a preset threshold set used to determine whether the human presence state is a typical human presence state that often occurs, including specific standards for cumulative triggering duration, frequency, and frequency. In other words, the normal discrimination constraint set is used to determine whether the human presence state can be regarded as a commonly used human presence state. The normal control plan is a preset control strategy constructed based on the control records that meet the normal discrimination constraints, which is used to achieve fast and automated sound field control adjustments to improve the response efficiency of the control.

[0093] Exemplarily, the preset normality discrimination constraint set is set according to historical data and expert experience, for example: cumulative trigger duration threshold: 25 minutes; cumulative trigger frequency threshold: 4 times; cumulative trigger frequency threshold: 0.15 times / minute.

[0094] Specifically, first, the collected human body state is associated with the corresponding control parameters and stored to build a control record library for subsequent statistics and strategy construction; then the data in the control record library is statistically analyzed to calculate the cumulative triggering time (total duration), cumulative triggering frequency (number of occurrences) and cumulative triggering frequency per unit time of each human body state. For example, for the "rolling" state, the cumulative triggering time is 30 minutes in an observation period; the cumulative triggering frequency is 5 times; the cumulative triggering frequency is 0.17 times / minute (5 times / 30 minutes). Then, each statistical indicator is judged. When any two or three indicators meet or exceed the preset threshold, the human body state is considered to be a normal state. For example, the statistical results of the above-mentioned "rolling" state are a cumulative triggering time of 30 minutes, a frequency of 5 times, and a frequency of 0.17 times / minute, all of which exceed the preset thresholds, and meet the normal judgment conditions.

[0095] Optionally, when it is detected that any two or three types of indicators meet the threshold, it means that the control record is suitable for most scenarios or has high-frequency usage characteristics. The corresponding "normal control plan" can be constructed and stored in a special control plan library. In subsequent use, these normal control plans can be preferentially or automatically called, thereby improving user experience and sound field management efficiency.

[0096] Furthermore, based on the control records that meet the threshold, multiple normal control plans are constructed, each plan corresponding to a common human body state. For example, for the "rolling" state, the plan may include measures such as adjusting the left and right balance of the car audio, adjusting the speaker delay, and optimizing the sound field distribution.

[0097] In some embodiments, the method for controlling multi-zone independent sound fields of a car audio system further includes: dynamically adjusting a normal control plan based on active control records and a preset update strategy, wherein the update strategy includes a periodic strategy and a trigger strategy.

[0098] Specifically, active control records are control parameters and effect data recorded during the execution of sound field control, which are used for subsequent strategy adjustments; normal control plans are preset sound field control strategies, which are used for rapid application in common scenarios to improve response speed.

[0099] Among them, the periodic strategy is a strategy for updating the normal control plan at fixed time intervals. The trigger strategy is used to trigger the update of the normal control plan when a specific condition or event is detected.

[0100] Specifically, first, when executing sound field control, the control parameters and corresponding human presence status when the user actively adjusts the sound field are synchronously recorded to form an active control record; then, according to the preset update strategy, based on the preset time interval (such as weekly or monthly), or when specific conditions are detected (such as user feedback, environmental changes or increased adjustment frequency), the update process is triggered, such as adjusting the parameters in the normal control plan through the analysis results of the active control record (such as concentrated values) to adapt to new needs or environmental changes.

[0101] In some embodiments, the method for controlling multi-zone independent sound fields of a car audio system further includes: Accumulate the adjustment frequency and adjustment frequency of dynamic adjustment of the normal control plan; when it is detected that either the adjustment frequency or the adjustment frequency exceeds the adjustment limit within a unit time, extract the dynamic adjustment record corresponding to the cumulative time window in the historical adjustment record; based on the dynamic adjustment record, extract the parameter offset sequence for fitting analysis, generate a strategy optimization function and update the preset sound field control strategy; based on the updated sound field control strategy, recalculate and update the control parameter adjustment amount of the normal control plan.

[0102] Specifically, historical adjustment records are control parameter change data recorded during long-term operation when the normal control plan is dynamically adjusted, including the time of each adjustment, parameter values ​​before and after the adjustment, and environmental status information. The cumulative time window is used to count the fixed time intervals of historical adjustment records, and the cumulative adjustment frequency and frequency are calculated within this window as the basis for subsequent analysis. The parameter offset sequence represents the change sequence of each control parameter (such as volume, balance, etc.) in the normal control plan relative to the baseline value within the cumulative time window, reflecting the trend and amplitude of dynamic adjustment.

[0103] Specifically, the active or passive adjustment operations of the normal control plan are continuously monitored, and each change of the sound field control parameters is regarded as an adjustment, that is, when the actual sound field effect is detected to deviate from the user's expectations, and the user actively adjusts the sound mode or volume through the vehicle interface, voice commands, etc., the adjustment frequency and adjustment frequency in the corresponding statistical counter are increased. Then, the adjustment limit is set, for example, the adjustment frequency threshold is 10 times / hour and the adjustment frequency threshold is 15 times / hour. When it is detected that the adjustment frequency or adjustment frequency in a unit time exceeds the preset adjustment limit, the dynamic adjustment record corresponding to the cumulative time window of the current trigger period is selected from the historical adjustment record database. Then, the control parameters before and after each adjustment in the selected records are compared to generate a parameter offset sequence, and the extracted parameter offset sequence is used as a data sample, and the least squares method, curve fitting or regression analysis method is used to fit a mathematical function reflecting the adjustment trend.

[0104] Furthermore, the strategy optimization function is used to adjust the original preset sound field control strategy and update the control parameters to make the strategy more in line with the current actual environment and user behavior. At the same time, the updated sound field control strategy is used as a new benchmark to recalculate the adjustment amount of each control parameter (such as volume, balance, delay, etc.), and the updated adjustment amount is applied to the normal control plan to ensure that the vehicle audio system can adapt to the occupant status and environmental changes in real time and dynamically.

[0105] By monitoring the frequency and frequency of adjustment, abnormal or frequent changes in strategy adjustment can be detected in time, and the sound field control strategy can be automatically optimized to reduce manual intervention. This self-optimization mechanism not only improves the intelligence level and adaptability of the system, but also enhances the stability and accuracy of sound field control, ensuring the best sound experience under different working conditions, further improving user experience and reliability.

[0106] In summary, the method for controlling multi-zone independent sound fields of car audio provided by the present invention has the following technical effects: Through the cockpit management terminal of the interactive target vehicle, the real-time status information of the human interaction device is obtained, and the human state is perceived based on the information to obtain the human presence state and human behavior state; with the human state perception result as a constraint condition, the preset sound field control strategy is matched and initialized to generate a benchmark sound field control strategy; combining the internal environment structure information and audio layout information of the target vehicle, a simulation control model is constructed, and the benchmark sound field control strategy is optimized based on path tracking; finally, multi-zone independent sound field control is realized according to the optimization results, and the control parameters and the corresponding human presence state are recorded, so as to provide a multi-dimensional personalized control path and improve the technical effect of the accuracy of sound field control.

[0107] Embodiment 2, as Figure 2 This is a schematic diagram of the structure of a multi-zone independent sound field control system for car audio of the present invention. For example, Figure 1 The flowchart of the method for controlling the multi-zone independent sound field of a car audio system of the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0108] Based on the same concept as the method for controlling a multi-zone independent sound field of a car audio system in the above embodiment, the present invention also provides a multi-zone independent sound field control system for a car audio system, comprising: The status acquisition module 11 is used to interact with the cockpit management terminal of the target vehicle and obtain the real-time status information of the human body interaction device.

[0109] The human body state perception module 12 is used to perform human body state perception based on the acquired real-time state information, and obtain human body state perception results, wherein the human body state perception results include human body existence state and human body behavior state.

[0110] The strategy initialization module 13 is used to match the preset sound field control strategy and perform strategy initialization based on the human body state perception result as a constraint condition, and output a reference sound field control strategy.

[0111] The simulation optimization module 14 is used to construct a simulation control model based on the internal environment structure information and the sound layout information of the target vehicle, and perform control optimization based on path tracking on the reference sound field control strategy.

[0112] The sound field control module 15 is used to perform multi-region independent sound field control according to the control optimization result, and record the control parameters and the corresponding human body presence status.

[0113] The status acquisition module 11 includes: The seat pressure distribution information acquisition unit is used to acquire the seat pressure distribution information of the seat area through the pressure distribution sensor assembly.

[0114] The seat restraint information acquisition unit is used to activate the restraint state sensor component based on the pressure distribution data, acquire seat inclination data and seat belt state data, and output them as seat restraint information.

[0115] The partition voice information acquisition unit is used to activate the acoustic collection component to acquire the partition voice information according to the seat pressure distribution information and the privacy authorization status.

[0116] The real-time status information integration unit is used to integrate the seat pressure distribution information, the seat restraint information and the partition voice information into the real-time status information.

[0117] The human body state sensing module 12 includes: The pressure value distribution matrix generating unit is used to map the seat pressure distribution information to a preset seat grid coordinate system to generate a pressure value distribution matrix.

[0118] The pressure center offset calculation unit is used to calculate the centroid coordinates of the pressure value distribution matrix in the horizontal and vertical dimensions to obtain the pressure center offset.

[0119] The seat restraint information synchronous acquisition unit is used to synchronously acquire the seat belt height, seat cushion pitch angle and seat back pitch angle in the seat restraint information.

[0120] The human posture classification result output unit is used to input the pressure center offset, the safety belt height, the seat cushion pitch angle and the seat back pitch angle into a convolutional neural network model trained by transfer learning, and output the human posture classification result and the auditory center height as the human body existence state.

[0121] The human body state sensing module 12 further includes: The partitioned speech data feature extraction unit is used to perform short-time Fourier transform on the partitioned speech data, extract the frequency spectrum center of gravity and root mean square amplitude in the 0.5-4kHz frequency band, and construct a two-dimensional feature vector.

[0122] The excitement level determination unit is used to determine the excitement level corresponding to the two-dimensional feature vector based on a preset excitement grading strategy.

[0123] The excitement level adjustment unit is used to calculate the fluctuation variance of the pressure distribution data within a continuous preset time window, and when the fluctuation variance exceeds a preset jitter threshold, the current excitement level is increased by one level.

[0124] The human behavior state output unit is used to apply an attenuation factor to the current excitement level and output it as the human behavior state when the weighted sum of the seat back pitch angle and the seat cushion pitch angle is greater than a preset threshold.

[0125] Wherein, the strategy initialization module 13 includes: The simulation control model building unit is used to load the pre-stored three-dimensional model of the vehicle interior structure and the acoustic material property parameters, build the simulation control model, and dynamically adjust the boundary reflection conditions of the simulation control model when the window opening and closing state changes are detected.

[0126] The sound wave propagation velocity correction unit is used to access the environmental temperature and humidity sensor data in real time and correct the calculated value of the sound wave propagation velocity accordingly.

[0127] The sound field uniformity and consistency index calculation unit is used to set the reflection number constraint, perform the propagation path tracking of the reference sound field control strategy according to the reflection number constraint, and calculate the uniformity index and consistency index of the sound field based on the propagation path tracking result.

[0128] The reference sound field control strategy optimization unit is used to iteratively optimize the reference sound field control strategy by taking the weighted values ​​of the uniformity index and the consistency index as optimization targets and combining with an optimization algorithm.

[0129] The simulation optimization module 14 includes: The sound field control strategy matching unit is used to match the preset sound field control strategy in the HRTF database according to the human body existence state and human body behavior state.

[0130] A low-frequency gain adjustment unit is used to estimate human body weight data based on the human body state perception result, and adjust the low-frequency gain according to the human body weight data to compensate for vibration absorption.

[0131] The optimal sound and image height prediction unit is used to construct a multivariate regression model based on sample sound and image data, and combine the human body state perception result and the human body weight data to predict the optimal sound and image height, wherein the sample sound and image data includes the associated stored sample weight, sample auditory center height, sample human body posture classification result and sample optimal sound and image height.

[0132] In some implementations, the car audio multi-zone independent sound field control system further includes: The control record library construction unit is used to memorize the human body's existence status and store it in association with the control parameters to construct the control record library.

[0133] The human body existence state statistical analysis unit is used to statistically analyze the cumulative triggering duration, cumulative triggering frequency and cumulative triggering frequency of each human body existence state.

[0134] The normal control plan generation unit is used to perform threshold judgment on the cumulative trigger duration, the cumulative trigger frequency and the cumulative trigger frequency according to a preset normal judgment constraint set. When any two or three of them meet the threshold, multiple normal control plans are constructed according to the corresponding control records.

[0135] In some implementations, the car audio multi-zone independent sound field control system further includes: The normal control plan adjustment unit is used to dynamically adjust the normal control plan based on the active control record and a preset update strategy, wherein the update strategy includes a periodic strategy and a trigger strategy.

[0136] In some implementations, the car audio multi-zone independent sound field control system further includes: The adjustment frequency and frequency accumulation unit is used to accumulate the adjustment frequency and adjustment frequency for dynamically adjusting the normal control plan.

[0137] The historical adjustment record extraction unit is used to extract the dynamic adjustment record corresponding to the accumulated time window in the historical adjustment record when it is detected that either the adjustment frequency or the adjustment frequency exceeds the adjustment limit within a unit time.

[0138] The strategy optimization function generation and updating unit is used to extract the parameter offset sequence for fitting analysis based on the dynamic adjustment record, generate a strategy optimization function and update the preset sound field control strategy.

[0139] The normal control plan control parameter updating unit is used to recalculate and update the control parameter adjustment amount of the normal control plan based on the updated sound field control strategy.

[0140] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to a car audio multi-zone independent sound field control system described in embodiment two. For the sake of brevity of the specification, they will not be further elaborated here.

[0141] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the above-mentioned embodiments. It should be understood that those skilled in the art can still modify the technical solutions recorded in the above-mentioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A method for controlling multi-zone independent sound fields of a car audio system, characterized in that: include: The cockpit management terminal of the interactive target vehicle obtains the real-time status information of the human interaction device; Based on the acquired real-time status information, human body status perception is performed to obtain a human body status perception result, wherein the human body status perception result includes a human body existence status and a human body behavior status; Taking the human body state perception result as a constraint condition, matching a preset sound field control strategy and initializing the strategy, and outputting a reference sound field control strategy; Based on the internal environment structure information and the sound layout information of the target vehicle, a simulation control model is constructed to perform control optimization based on path tracking on the reference sound field control strategy; Based on the control optimization results, multi-zone independent sound field control is performed, and the control parameters and the corresponding human presence status are recorded.

2. A method for controlling multi-zone independent sound fields of car audio according to claim 1, characterized in that: The cockpit management end of the interactive target vehicle obtains real-time status information of the human interaction device, including: collecting seat pressure distribution information of the seat area through a pressure distribution sensor assembly; activating a restraint state sensor assembly based on the pressure distribution data, acquiring seat inclination data and seat belt state data, and outputting the data as seat restraint information; According to the seat pressure distribution information and in combination with the privacy authorization status, the acoustic collection component is activated to obtain the partition voice information; The seat pressure distribution information, the seat restraint information and the partition voice information are used as the real-time status information.

3. A method for controlling multi-zone independent sound fields of a car audio system as claimed in claim 2, characterized in that: Based on the acquired real-time status information, human body status perception is performed to obtain a human body status perception result, wherein the human body status perception result includes a human body existence status and a human body behavior status, including: Mapping the seat pressure distribution information to a preset seat grid coordinate system to generate a pressure value distribution matrix; Calculating the centroid coordinates of the pressure value distribution matrix in the horizontal and vertical dimensions to obtain the pressure center offset; Synchronously acquiring the seat belt height, seat cushion pitch angle and seat back pitch angle in the seat restraint information; The pressure center offset, the safety belt height, the seat cushion pitch angle and the seat back pitch angle are input into a convolutional neural network model trained by transfer learning, and the human body posture classification result and the auditory center height are output as the human body existence state.

4. A method for controlling multi-zone independent sound fields of a car audio system as claimed in claim 3, characterized in that: Based on the acquired real-time status information, human body status perception is performed to obtain a human body status perception result, wherein the human body status perception result includes a human body existence status and a human body behavior status, including: Performing short-time Fourier transform on the partitioned speech data, extracting the frequency spectrum center of gravity and root mean square amplitude in the 0.5-4kHz frequency band, and constructing a two-dimensional feature vector; Based on a preset excitement grading strategy, determining the excitement level corresponding to the two-dimensional feature vector; Calculating the fluctuation variance of the pressure distribution data within a continuous preset time window, and when the fluctuation variance exceeds a preset jitter threshold, raising the current excitement level by one level; When the weighted sum of the seat back pitch angle and the seat cushion pitch angle is greater than a preset threshold, an attenuation factor is applied to the current excitement level and output as the human behavior state.

5. A method for controlling multi-zone independent sound fields of a car audio system as claimed in claim 4, characterized in that: Taking the human body state perception result as a constraint condition, matching the preset sound field control strategy and initializing the strategy, and outputting the benchmark sound field control strategy, including: According to the human body existence state and human body behavior state, matching the preset sound field control strategy in the HRTF database; estimating human body weight data based on the human body state sensing result, and adjusting low-frequency gain according to the human body weight data to compensate for vibration absorption; A multivariate regression model based on sample sound and image data is constructed, and the optimal sound and image height is predicted by combining the human body state perception result and the human body weight data, wherein the sample sound and image data includes the associated stored sample weight, sample auditory center height, sample human body posture classification result and sample optimal sound and image height.

6. A method for controlling multi-zone independent sound fields of a car audio system as claimed in claim 5, characterized in that: Based on the internal environment structure information and the sound layout information of the target vehicle, a simulation control model is constructed, and the reference sound field control strategy is optimized based on path tracking, including: Loading the pre-stored three-dimensional model of the vehicle interior structure and acoustic material property parameters, constructing a simulation control model, and dynamically adjusting the boundary reflection conditions of the simulation control model when detecting a change in the window opening and closing state; Real-time access to environmental temperature and humidity sensor data, and corresponding correction of the calculated value of sound wave propagation speed; Setting a reflection number constraint, performing propagation path tracing of the reference sound field control strategy according to the reflection number constraint, and calculating a uniformity index and a consistency index of the sound field based on the propagation path tracing result; The weighted values ​​of the uniformity index and the consistency index are used as optimization targets, and the reference sound field control strategy is iteratively optimized in combination with an optimization algorithm.

7. A method for controlling multi-zone independent sound fields of car audio as claimed in claim 1, characterized in that: Also includes: Memorize the human body's state and store it in association with control parameters to build a control record library; Statistically analyze the cumulative triggering duration, cumulative triggering frequency and cumulative triggering frequency of each human body state; According to the preset normal judgment constraint set, the cumulative trigger duration, the cumulative trigger frequency and the cumulative trigger frequency are respectively subjected to threshold judgment. When any two or three of them meet the threshold, multiple normal control plans are constructed according to the corresponding control records.

8. A method for controlling multi-zone independent sound fields of a car audio system as claimed in claim 7, characterized in that: Also includes: According to the active control records, the normal control plan is dynamically adjusted based on a preset update strategy, wherein the update strategy includes a periodic strategy and a trigger strategy.

9. A method for controlling multi-zone independent sound fields of a car audio system as claimed in claim 8, characterized in that: Also includes: Accumulating the adjustment frequency and the frequency of dynamic adjustment of the normal control plan; When it is detected that either the adjustment frequency or the adjustment frequency exceeds the adjustment limit within a unit time, a dynamic adjustment record corresponding to the accumulated time window in the historical adjustment record is extracted; Based on the dynamic adjustment record, extract the parameter offset sequence for fitting analysis, generate a strategy optimization function and update the preset sound field control strategy; Based on the updated sound field control strategy, the control parameter adjustment amount of the normal control plan is recalculated and updated.

10. A multi-zone independent sound field control system for car audio, characterized in that: A method for controlling a multi-zone independent sound field of a car audio system according to any one of claims 1 to 9, comprising: The status acquisition module is used to interact with the cockpit management terminal of the target vehicle and obtain the real-time status information of the human interaction device; A human body state perception module, used to perceive the human body state based on the acquired real-time state information, and obtain a human body state perception result, wherein the human body state perception result includes a human body existence state and a human body behavior state; A strategy initialization module, used to match a preset sound field control strategy and perform strategy initialization based on the human body state perception result as a constraint condition, and output a reference sound field control strategy; A simulation optimization module, for constructing a simulation control model based on the internal environment structure information and the sound layout information of the target vehicle, and performing control optimization based on path tracking on the reference sound field control strategy; The sound field control module is used to perform multi-zone independent sound field control according to the control optimization results, and record the control parameters and the corresponding human body presence status.

Images