WO2023151229A1 - 一种基于加速工况的实车主动降噪方法及系统 - Google Patents

一种基于加速工况的实车主动降噪方法及系统 Download PDF

Info

Publication number
WO2023151229A1
WO2023151229A1 PCT/CN2022/105938 CN2022105938W WO2023151229A1 WO 2023151229 A1 WO2023151229 A1 WO 2023151229A1 CN 2022105938 W CN2022105938 W CN 2022105938W WO 2023151229 A1 WO2023151229 A1 WO 2023151229A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
noise reduction
secondary channel
filter
error
Prior art date
Application number
PCT/CN2022/105938
Other languages
English (en)
French (fr)
Inventor
张士强
张程鹏
李�浩
李允�
曹蕴涛
王石
罗柏成
Original Assignee
中国第一汽车股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国第一汽车股份有限公司 filed Critical 中国第一汽车股份有限公司
Publication of WO2023151229A1 publication Critical patent/WO2023151229A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10K—SOUND-PRODUCING DEVICES; METHODS OR DEVICES FOR PROTECTING AGAINST, OR FOR DAMPING, NOISE OR OTHER ACOUSTIC WAVES IN GENERAL; ACOUSTICS NOT OTHERWISE PROVIDED FOR
    • G10K11/00—Methods or devices for transmitting, conducting or directing sound in general; Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
    • G10K11/16—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; DEAF-AID SETS; PUBLIC ADDRESS SYSTEMS
    • H04R3/00—Circuits for transducers, loudspeakers or microphones
    • H04R3/02—Circuits for transducers, loudspeakers or microphones for preventing acoustic reaction, i.e. acoustic oscillatory feedback

Definitions

  • the invention belongs to the field of automobile manufacturing; in particular, it relates to an active noise reduction method and system for real vehicles based on acceleration working conditions.
  • engine order noise is one of the main sources of interior noise in traditional fuel vehicles.
  • automobile noise control methods are mainly divided into passive noise control methods and active noise control methods.
  • Passive noise control has a good suppression effect on medium and high frequency noise by adding sound insulation materials and other methods, but due to the physical characteristics of low frequency noise, the passive control method is not effective in controlling low frequency noise.
  • Active noise reduction in the car is an application scenario of active noise control (Active Noise Control).
  • Secondary speakers (usually door speakers and subwoofers) emit secondary sounds with the same frequency as the original noise and a phase difference of 180°.
  • the signal is superimposed at the error microphone to generate a "quiet area" at the human ear to achieve the noise reduction effect.
  • the active noise reduction method Due to the physical characteristics of the system itself, the active noise reduction method has a better control effect on low-frequency noise, but it is not easy to control high-frequency noise. It is just a supplement to the passive noise control method, and does not need to add additional sound insulation materials. It has a high degree of integration and is in line with lightweight Therefore, the research and application of active noise reduction technology in automobiles is increasing.
  • Liu Jian from Harbin Institute of Technology used the LMS theory as the analysis basis to conduct a detailed and in-depth performance analysis of the narrowband ANC system based on the FxLMS algorithm.
  • Jidong Sun of Jilin University used the adaptive notch filter algorithm as the adaptive control algorithm, designed and developed a prototype of the adaptive controller for interior noise based on a digital signal processor (DSP), and established a single-channel adaptive control system for interior noise.
  • DSP digital signal processor
  • the active noise reduction technology based on the adaptive notch filter has a simple algorithm, a small amount of calculation, and has a good control effect on narrow-band noise such as engine order noise. It is also the most widely used active noise control method for engine order noise.
  • most of the related research on engine noise active noise reduction systems based on adaptive notch filters is to discuss the noise reduction effect and system stability of single-channel and multi-channel active noise reduction systems under steady-state conditions. There are relatively few studies on system debugging and effect optimization of noise reduction systems.
  • the present invention aims to solve one of the technical problems in the related art at least to a certain extent.
  • an object of the present invention is to propose a real vehicle active noise reduction method based on acceleration conditions, which can achieve the purpose of system debugging and effect optimization of the active noise reduction system under acceleration conditions.
  • the second purpose of the present invention is to propose an active noise reduction system for real vehicles based on acceleration conditions.
  • a third object of the present invention is to propose a computer device.
  • a fourth object of the present invention is to provide a non-transitory computer-readable storage medium.
  • a kind of real vehicle active noise reduction method based on acceleration working condition, described real vehicle active noise reduction method comprises the following steps:
  • S01 Arrange the error microphone, secondary speaker, CAN speed signal and ANC controller in the car;
  • an FIR filter with a filter length of 128 is used to identify the output sweep signal and the signal collected by the error microphone to obtain a secondary channel transfer function
  • S04 Based on S02 frequency sweep signal and error microphone acquisition signal, adopt FIR filter with a filter length of 1024, identify the output frequency sweep signal and the signal collected by error microphone to obtain the secondary channel transfer function, which is used to calculate convergence coefficient stability curve;
  • S05 Integrate the secondary channel transfer function obtained in S03 and the convergence coefficient stability curve obtained in S04 into the FxLMS algorithm program, and execute the noise reduction program to reduce the noise in the vehicle.
  • the S01 specifically includes setting the error microphone on the car frame; placing the secondary speaker in the car door, connecting the CAN speed signal to the OBD interface to monitor the engine speed; installing the ANC controller in the car.
  • the S02 is specifically, assuming that J secondary speakers and K error microphones are used in a multi-channel active noise reduction system, and there is a secondary channel between each secondary speaker and each error microphone, the entire The secondary channel transfer function of the system is represented by Hs(z);
  • x(n) is the reference signal formed according to the speed signal, Represents the secondary channel estimation of the system. There are J ⁇ K secondary channels in total.
  • the reference signal x(n) and its 90° phase shift signal are respectively convolved with the secondary channel estimation to obtain the filtered reference signals R 0 (n), R 1 (n) is the J ⁇ K dimensional matrix, namely:
  • the filter weight vectors W 1 and W 2 are two J ⁇ 1 dimensional vectors
  • the residual error signal vector E(n) is a K ⁇ 1 dimensional vector
  • Y(n) is the controller
  • the output secondary sound signal is a J ⁇ 1 dimensional vector
  • the iterative formula of the two adaptive weight vectors obtained by the FxLMS algorithm is:
  • the secondary sound signal output by the controller is:
  • the output sweep signal and the signal collected by the error microphone are identified to obtain the transfer function of the secondary channel, specifically,
  • the secondary channel transfer function for identifying the secondary channel is specifically, in the secondary channel identification model, x (n) represents the noise excitation in the secondary channel test process, that is, the excitation signal output by the controller; d( n) represents the voltage signal or sound pressure signal collected by the microphone received by the controller during the test; y(n) represents the output of the adaptive filter, that is, the response of the noise stimulus after passing through the filter; e(n) represents the response received by the controller
  • the weight vector of the adaptive filter is continuously updated and iterated according to the LMS algorithm, so that the residual error signal e(n) is constantly approaching 0, that is, let the filter output response y(n) continuously approach the voltage signal received by the controller; when the system converges and the residual error signal is close to 0, the adaptive filter weight vector coefficient can be equivalent to the secondary Channel impulse response function.
  • the calculation of the convergence coefficient stability curve in S04 is specifically, assuming that there are M secondary sound sources and L error microphones, and assuming that the complex component of the l error signal at the nth harmonic is denoted as E l ( ⁇ n ), the complex component of the mth secondary signal in this harmonic is recorded as W m ( ⁇ n ), then the error signal is
  • D l ( ⁇ n ) is the lth complex error signal caused by the primary sound source
  • C lm ( ⁇ n ) is the complex response from the mth secondary sound source to the lth error sensor at this frequency, in vector form
  • E( ⁇ n ) [E 1 ( ⁇ n ),E 2 ( ⁇ n ),...,E L ( ⁇ n )] T
  • W( ⁇ n ) [W 1 ( ⁇ n ),W 2 ( ⁇ n ),...,W M ( ⁇ n )] T
  • V(k) Q H (W(k)-W opt ) (16)
  • V(k) [1- ⁇ ] k V(0) (17)
  • the S05 is specifically, verifying the effect of identifying secondary channels of FIR filters with different filter lengths is specifically, selecting the length of the FIR filter 128 when identifying the secondary channels for algorithm integration;
  • the frequency resolution based on the stability boundary curve of the convergence coefficient is related to the selection of the length of the FIR filter.
  • df is the frequency resolution of the stability boundary curve
  • fs is the sampling rate used for secondary channel identification
  • Length_of_FIR is the length of the FIR filter used for secondary channel identification.
  • the secondary channel used to calculate the convergence coefficient stability boundary curve does not involve the amount of algorithm calculations. Relatively speaking, using a longer filter length is of great help to improve the accuracy of the stability boundary curve. If the filter length is short, the system will be unstable in some frequency ranges due to the resolution.
  • the present invention uses an FIR filter with a filter length of 1024 to identify the secondary channel for calculating the convergence coefficient stability boundary curve;
  • the change curve of the curve convergence coefficient is 1/5 of the stability boundary curve.
  • the noise reduction system includes
  • Error microphone used to collect the error signal and send it to the ANC controller for algorithm calculation
  • ANC controller calculates the operation to get the output signal to the secondary speaker through the power amplifier
  • CAN bus collect the engine speed to construct the reference signal of the active noise reduction system
  • the ANC controller is used to execute the FxLMS algorithm program.
  • a computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of any one of the methods described above when executing the computer program.
  • a non-transitory computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the methods described above are implemented.
  • the secondary channel identification two different filter lengths are used respectively, and the result of identifying the secondary channel with a 128-length FIR filter is used for system algorithm integration to reduce the calculation amount of the algorithm as much as possible;
  • the result of identifying the secondary channel with a FIR filter with a length of 1024 is used to calculate the convergence coefficient stability boundary curve, so that the calculated stability boundary curve has a higher resolution in the frequency domain, avoiding certain System divergence occurs in the engine speed range.
  • Fig. 5 is a curve diagram of the stability boundary of the convergence coefficient calculated by identifying the secondary channels of the FIR filters whose lengths are 128 and 1024 respectively according to the present invention.
  • a real vehicle active noise reduction method based on acceleration conditions comprising the following steps:
  • S01 Arrange the error microphone, secondary speaker, CAN speed signal and ANC controller in the car;
  • an FIR filter with a filter length of 128 is used to identify the output sweep signal and the signal collected by the error microphone to obtain a secondary channel transfer function
  • S04 Based on S02 frequency sweep signal and error microphone acquisition signal, adopt FIR filter with a filter length of 1024, identify the output frequency sweep signal and the signal collected by error microphone to obtain the secondary channel transfer function, which is used to calculate convergence coefficient stability curve;
  • S05 Integrate the secondary channel transfer function obtained in S03 and the convergence coefficient stability curve obtained in S04 into the FxLMS algorithm program, and execute the noise reduction program to reduce the noise in the vehicle.
  • the S01 specifically includes setting the error microphone on the vehicle frame; placing the secondary speaker in the vehicle door, connecting the CAN speed signal to the OBD interface, and monitoring the engine Speed; install the ANC controller in the car.
  • a real vehicle active noise reduction method based on acceleration conditions the S02 is specifically, the present invention proposes an active noise reduction system based on a multi-channel adaptive notch filter, that is, multiple secondary speakers are used to control multiple error microphone positions
  • the noise level of the multi-channel active noise reduction system is shown in Figure 2 as a block diagram of the secondary channel. Assuming that J secondary speakers and K error microphones are used in a multi-channel active noise reduction system, there is a secondary channel between each secondary speaker and each error microphone, and Hjk in Figure 2 represents the jth secondary The secondary channel between the loudspeaker and the kth error sensor, the secondary channel transfer function of the whole system is represented by Hs(z); the functional block diagram of the multi-channel adaptive notch filter is shown in Figure 3.
  • x(n) is the reference signal formed according to the speed signal, Represents the secondary channel estimation of the system. There are J ⁇ K secondary channels in total.
  • the reference signal x(n) and its 90° phase shift signal are respectively convolved with the secondary channel estimation to obtain the filtered reference signals R 0 (n), R 1 (n) is the J ⁇ K dimensional matrix, namely:
  • the filter weight vectors W 1 and W 2 are two J ⁇ 1 dimensional vectors
  • the residual error signal vector E(n) is a K ⁇ 1 dimensional vector
  • Y(n) is the controller
  • the output secondary sound signal is the J ⁇ 1 dimensional vector
  • the iterative formula of the two adaptive weight vectors obtained by the FxLMS algorithm is:
  • the secondary sound signal output by the controller is:
  • the S03 identifies the output sweep signal and the signal collected by the error microphone to obtain the secondary channel transfer function, specifically,
  • the secondary pathway identification model is shown in Figure 4.
  • x(n) represents the noise excitation during the secondary path test process, that is, the excitation signal output by the controller
  • d(n) represents the voltage signal collected by the microphone received by the controller during the test process or the sound pressure signal
  • y(n) represents the output of the adaptive filter, that is, the response of the noise excitation after passing through the filter
  • e(n) represents the residual error signal after the superposition of the voltage signal received by the controller and the output response of the filter
  • the weight vector of the adaptive filter is continuously updated and iterated according to the LMS algorithm, so that the residual error signal e(n) is continuously approaching 0, that is, the filter output response y(n) is continuously approaching The voltage signal received by the controller; when the system converges and the residual error signal is close to 0, the adaptive filter weight vector coefficients can be
  • D l ( ⁇ n ) is the lth complex error signal caused by the primary sound source
  • C lm ( ⁇ n ) is the complex response from the mth secondary sound source to the lth error sensor at this frequency, in vector form
  • E( ⁇ n ) [E 1 ( ⁇ n ),E 2 ( ⁇ n ),...,E L ( ⁇ n )] T
  • W( ⁇ n ) [W 1 ( ⁇ n ),W 2 ( ⁇ n ),...,W M ( ⁇ n )] T
  • H represents the Hermitian transpose (conjugate transpose) of a vector or matrix
  • E and W represent the complex error signal of L ⁇ 1 and the complex secondary sound signal of M ⁇ 1 respectively
  • a and B are respectively Positive definite weighting matrices of L ⁇ L and M ⁇ M
  • formula (6) can also be written as the sum of squares of the modulus of the unweighted error signal plus the sum of the squares of the modulus of the weighted secondary signal:
  • V(k) Q H (W(k)-W opt ) (16)
  • V(k) [1- ⁇ ] k V(0) (17)
  • the S05 is specifically that before the algorithm is debugged and run, the secondary channel identification results should be integrated into the algorithm program to filter the reference signal, generally using FIR filters, etc.
  • the effective secondary channel impulse response function In this application scenario, the selection of the filter length should take into account the noise reduction effect, system stability and algorithm computation.
  • the length of the FIR filter is selected to be 128 when the identification secondary channel is used for algorithm integration.
  • the frequency resolution (or the resolution that varies with the engine speed) based on the convergence coefficient stability boundary curve is related to the selection of the length of the FIR filter.
  • the longer the length of the secondary channel used to calculate the stability boundary curve, correspondingly, the more spectral lines of the stability boundary curve within the same frequency bandwidth range, the finer the frequency resolution, and the relationship can be written as:
  • df is the frequency resolution of the stability boundary curve
  • fs is the sampling rate used for secondary channel identification
  • Length_of_FIR is the length of the FIR filter used for secondary channel identification.
  • the secondary channel used to calculate the convergence coefficient stability boundary curve does not involve the amount of algorithm calculations. Relatively speaking, using a longer filter length is of great help to improve the accuracy of the stability boundary curve. If the filter length is short, the system will be unstable in some frequency ranges due to the resolution.
  • the present invention uses an FIR filter with a filter length of 1024 to identify the secondary channel for calculating the convergence coefficient stability boundary curve;
  • the size of the convergence coefficient not only affects the convergence speed of the active noise reduction system, but also affects the steady-state error when the system converges.
  • the convergence coefficient is too large, although the convergence speed increases, the corresponding steady-state error will increase. Affects the noise reduction effect to a certain extent. Therefore, based on the steady-state error of the system, when the present invention integrates the frequency-convergence coefficient curve, the change curve of the curve convergence coefficient is 1/5 of the stability boundary curve, taking into account factors such as system convergence speed, noise reduction effect, and steady-state error.
  • a real vehicle active noise reduction system based on acceleration conditions comprising
  • Error microphone used to collect the error signal and send it to the ANC controller for algorithm calculation
  • Secondary speaker ANC controller calculates the operation to get the output signal to the secondary speaker through the power amplifier; four door speakers are used as the secondary speaker,
  • CAN bus collect the engine speed to construct the reference signal of the active noise reduction system
  • the ANC controller is used to execute the FxLMS algorithm program.
  • a computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of any one of the methods described above when executing the computer program.
  • a non-transitory computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the methods described above are implemented.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Otolaryngology (AREA)
  • Signal Processing (AREA)
  • Soundproofing, Sound Blocking, And Sound Damping (AREA)
  • Fittings On The Vehicle Exterior For Carrying Loads, And Devices For Holding Or Mounting Articles (AREA)

Abstract

一种基于加速工况的实车主动降噪方法及系统。基于车内布置，ANC控制器播放的扫频信号通过次级扬声器播放后再通过误差麦克风的采集再次传递给ANC控制器；基于扫频信号和误差麦克风采集信号，采用滤波器长度为128的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数；基于扫频信号和误差麦克风采集信号，采用滤波器长度为1024的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数，用于计算收敛系数稳定性曲线；将次级通道传递函数和收敛系数稳定性曲线集成到FxLMS算法程序中，执行降噪程序对车辆内的噪声进行降低。

Description

一种基于加速工况的实车主动降噪方法及系统 技术领域
本发明属于汽车制造领域；具体涉及一种基于加速工况的实车主动降噪方法及系统。
背景技术
随着人们生活水平和生活质量的提高，人们愈发关注车辆乘坐舒适性及声学环境的静谧程度，甚至是车内声场的声品质控制和娱乐音的声场重放等等。而其中，发动机阶次噪声是传统燃油车车内噪声的主要来源之一。目前，汽车噪声控制方法主要分为被动噪声控制方法和主动噪声控制方法。被动噪声控制通过增加隔音材料等方法，对中高频噪声有较好的抑制效果，但是由于低频噪声的物理特性，被动控制方法对于低频噪声控制效果不佳。车内主动降噪即是噪声主动控制(Active Noise Control)的一种应用场景，通过次级扬声器(一般是车门扬声器和重低音扬声器)发出与原始噪声频率相同、相位相差180°的次级声音信号，在误差麦克风处叠加从而在人耳处产生一“静谧区域”，达到降噪效果。由于系统本身物理特性，主动降噪方法对低频噪声控制效果较好，反而不易控制高频噪声，恰好是被动噪声控制方法的补充，并且不用增加额外的隔音材料，集成化程度高，符合轻量化的发展趋势，因此主动降噪技术在汽车上的研究应用越来越多。
学者们对平稳窄带噪声信号的主动控制做了较多的研究，其中较有代表性的方法是将基于FxLMS算法的自适应陷波器应用于平稳窄带信号的主动控制。英国Southampton大学的Elliot和Boucher等人针对基于FxLMS算法的自适应陷波器，研究了多通道自适应前馈控制系统的特性表现。韩国仁 川国立大学的Young-Sup Lee等人通过理论和试验研究了不同长度的脉冲响应函数对主动噪声控制系统的影响。近年来，国内学者也越发关注对车内发动机噪声的主动控制，哈尔滨工业大学的刘剑以LMS理论为分析基础，对基于FxLMS算法的窄带ANC系统进行了详细深入的性能分析。吉林大学孙吉东采用自适应陷波器算法作为自适应控制算法，基于数字信号处理器(DSP)设计开发了车内噪声自适应控制器原理样机，建立了单通道的车内噪声自适应控制系统，并在一辆自主品牌轿车上进行了变速器空挡、发动机不同转速下的车内低频峰值噪声主动控制试验。
基于自适应陷波器的主动降噪技术算法简便、计算量小，并且对发动机阶次噪声等窄带噪声有较好的控制效果，也是现在针对发动机阶次噪声应用最为广泛的主动噪声控制方法。目前，基于自适应陷波器的发动机噪声主动降噪系统相关研究，多数是讨论稳态工况下单通道、多通道主动降噪系统的降噪效果和系统稳定性，针对加速工况下主动降噪系统的系统调试和效果优化的研究相对较少。
发明内容
本发明旨在至少在一定程度上解决相关技术中的技术问题之一。
为此，本发明的一个目的在于提出一种基于加速工况的实车主动降噪方法，该方法达到加速工况下主动降噪系统的系统调试和效果优化的目的。
本发明的第二个目的在于提出一种基于加速工况的实车主动降噪系统。
本发明的第三个目的在于提出一种计算机设备。
本发明的第四个目的在于提出一种非临时性计算机可读存储介质。
本发明通过以下技术方案实现：
一种基于加速工况的实车主动降噪方法，所述实车主动降噪方法包括以 下步骤：
S01：在车内布置误差麦克风、次级扬声器、CAN转速信号与ANC控制器；
S02：基于S01的车内布置，使ANC控制器播放的扫频信号通过次级扬声器播放后再通过误差麦克风的采集再次传递给ANC控制器；
S03：基于S02扫频信号和误差麦克风采集信号，采用滤波器长度为128的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数；
S04：基于S02扫频信号和误差麦克风采集信号，采用滤波器长度为1024的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数，用于计算收敛系数稳定性曲线；
S05：将S03得到的次级通道传递函数和S04得到的收敛系数稳定性曲线集成到FxLMS算法程序中，执行降噪程序对车辆内的噪声进行降低。
进一步的，所述S01具体为，将误差麦克风设置在车框上；将次级扬声器安放在车门内，将CAN转速信号连接在OBD接口处，监测发动机转速；将ANC控制器安装在车内。
进一步的，所述S02具体为，假设在多通道主动降噪系统中采用J个次级扬声器和K个误差麦克风，每个次级扬声器与每个误差麦克风之间都存在一条次级通道，整个系统的次级通道传函用Hs(z)表示；
x(n)为根据转速信号形成的参考信号，
Figure PCTCN2022105938-appb-000001
表示系统的次级通道估计，共有J×K条次级通道，参考信号x(n)及其90°相移信号分别与次级通道估计卷积后得到滤波参考信号R 0(n)、R 1(n)即J×K维矩阵，即有：
Figure PCTCN2022105938-appb-000002
多通道自适应陷波滤波系统中，滤波器权矢量W 1和W 2为两个J×1维向量，残余误差信号矢量E(n)为K×1维向量，Y(n)为控制器输出的次级声音信号即J×1维向量，由FxLMS算法得到两个自适应权重矢量的迭代公式为：
Figure PCTCN2022105938-appb-000003
所以，得到控制器输出次级声音信号有：
Y(n)＝x 0(n)W 0+x 1(n)W 1    (3)。
进一步的，所述S03将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数具体为，
所述对次级通道进行辨识次级通道传递函数具体为，在次级通路辨识模型中，x(n)代表次级通路测试过程中的噪声激励，也就是控制器输出的激励信号；d(n)代表测试过程中控制器接收到的麦克风采集得到电压信号或声压信号；y(n)代表自适应滤波器输出，即噪声激励经过滤波器后的响应；e(n)代表控制器接受到的电压信号与滤波器输出响应叠加后的残余误差信号；在次级通路辨识过程中，自适应滤波器的权矢量根据LMS算法不断进行更新迭代，使得残余误差信号e(n)不断逼近于0，也就是让滤波器输出响应y(n)不断逼近于控制器接收到的电压信号；当系统收敛，残余误差信号接近于0时，自适应滤波器权矢量系数即可以等效于次级通路脉冲响应函数。
进一步的，所述S04计算收敛系数稳定性曲线具体为，假设有M个次级声源和L个误差麦克风，假设第l个误差信号在第n个谐波的复数成分记为E l(ω n)，第m个次级信号在这个谐波的复数成分记为W m(ω n)，则误差信号为
Figure PCTCN2022105938-appb-000004
其中D l(ω n)是初级声源造成的第l个复数误差信号，C lm(ω n)是第m个次级声源到第l个误差传感器在该频率下的复数响应，向量形式有
E(ω n)＝D(ω n)+C(ω n)W(ω n)   (5)
其中
E(ω n)＝[E 1(ω n),E 2(ω n),...,E L(ω n)] T
D(ω n)＝[D 1(ω n),D 2(ω n),...,D L(ω n)] T
W(ω n)＝[W 1(ω n),W 2(ω n),...,W M(ω n)] T
Figure PCTCN2022105938-appb-000005
对于单频噪声来说目标函数写成
J＝E HAE+W HBW    (6)
其中上标H代表向量或矩阵的埃尔米特转置；E和W分别代表L×1的复数误差信号和M×1的复数次级声音信号，A和B分别是L×L和M×M的正定加权矩阵；式(6)也可以写成未加权误差信号模数平方和加上加权次级信号模数平方和：
J＝E HE+βW HW   (7)
结合式(5)目标函数可以写成变量W二次型的形式：
J＝D HD+W HC HD+D HCW+W H[C HC+βI]W   (8)
目标函数对于W实部(W R)和虚部(W I)的导数都是实数，所以可以定义复数梯度向量为：
Figure PCTCN2022105938-appb-000006
由于g的实部虚部相互独立，让g＝0设置J对于W R和W I的微分等于0，得到最优控制信号向量：
W opt＝-[C HC+βI] -1C HD   (10)
结合式(5)，复数梯度向量可写成：
g＝2[C HE+βW]   (11)
以与梯度向量反比的方向调整复数次级信号的实部和虚部，得到最速下降算法：
W(k+1)＝(1-αβ)W(k)-αC HE(k)   (12)
其中，α表示收敛系数。结合式(5)和式(10)迭代公式(12)写成：
(W(k+1)-W opt)＝[I-α(C HC+βI)](W(k)-W opt)   (13)
假设W(0)＝0，重复应用式(13)得到
W(k)-W opt＝-[I-α(C HC+βI)] kW opt   (14)
如果复数海塞矩阵写成复数酉矩阵的形式，标准化特征向量Q和特征值对角矩阵，Λ＝diag(λ 1,λ 2,...,λ M)，其中特征值都是实数，所以
C HC+βI＝QΛQ H   (15)
定义控制系统的主坐标为
V(k)＝Q H(W(k)-W opt)   (16)
所以式(14)写成
V(k)＝[1-αΛ] kV(0)   (17)
因为Λ是对角矩阵，控制系统主坐标的收敛是独立的，V(k)的第m个成分写成
Figure PCTCN2022105938-appb-000007
其中上述方程有效时需要保证-1＜1-αλ m＜1，得到基于收敛系数的稳定性条件：对于所有的m，0＜α＜2/λ m。
通过以上推导过程，结合式(15)和式(18)可以看到，在确定收敛系数α的值时，可通过次级通道传递函数矩阵计算收敛系数稳定性边界曲线。在汽车发动机转速变化时，阶次噪声频率也在变化，所以在该转速下满足系统稳定性的收敛系数限值也不同，即收敛系数稳定性边界值随发动机转速的变化而变化。
进一步的，所述S05具体为，验证不同滤波器长度的FIR滤波器辨识次级通道效果具体为，辨识次级通道用于算法集成时FIR滤波器长度选择128；
基于收敛系数稳定性边界曲线的频率分辨率与FIR滤波器长度的选择有关。用于计算稳定性边界曲线的次级通道长度越长，相对应的，相同频率带宽范围内稳定性边界曲线的谱线越多，频率分辨率越精细，其关系可以写成：
df＝fs/Length_of_FIR   (19)
其中，df为稳定性边界曲线的频率分辨率，fs为次级通道辨识使用的采样率，Length_of_FIR为次级通道辨识采用的FIR滤波器长度。与用于算法集成不同，计算收敛系数稳定性边界曲线所使用的次级通道不涉及算法运算量，相对来说，使用较长的滤波器长度对于提高稳定性边界曲线精确性具有较大帮助，滤波器长度如果较短会因为分辨率原因造成某些频率范围内系统不稳定的现象。本发明使用滤波器长度为1024的FIR滤波器辨识次级通道用于计算收敛系数稳定性边界曲线；
所以基于系统稳态误差，将曲线收敛系数变化曲线为稳定性边界曲线的1/5。
进一步的，所述降噪系统包括
误差麦克风：用来采集误差信号并给到ANC控制器进行算法计算；
次级扬声器：ANC控制器算运行得到输出信号经功放给到次级扬声器；
CAN总线：采集发动机转速来构造主动降噪系统的参考信号；
ANC控制器：ANC控制器用来执行FxLMS算法程序。
一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时实现上述任一项所述方法的步骤。
一种非临时性计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述任一项所述的方法的步骤。
本发明的有益效果是：
通过次级通道传递函数矩阵计算收敛系数稳定性边界曲线(在保证系统稳定性前提下，收敛系数上限随频率的变化曲线)，使得系统在发动机较宽转速范围内加速时同样能够达到良好的降噪效果，从而拓宽主动降噪系统的应用场景和范围。
在进行次级通道辨识时辨识两次，分别采用两种不同的滤波器长度，用长度为128的FIR滤波器辨识次级通道的结果用于系统算法集成，尽量降低算法计算量；
用长度为1024的FIR滤波器辨识次级通道的结果用于计算收敛系数稳定性边界曲线，使得计算出来的稳定性边界曲线在频率域上具备较高分辨率，避免加速工况下在某些发动机转速范围内出现系统发散的情况。
附图说明
附图1是本发明主动降噪系统结构组成示意图。
附图2是本发明多通道主动降噪系统次级通道框图。
附图3是本发明多通道自适应陷波滤波原理框图。
附图4是本发明次级通路辨识模型图。
附图5是本发明长度分别为128和1024的FIR滤波器辨识次级通道从而计算得到的收敛系数稳定性边界曲线图。
具体实施方式
下面将结合本发明实施例中的附图对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
实施例1
一种基于加速工况的实车主动降噪方法，所述实车主动降噪方法包括以下步骤：
S01：在车内布置误差麦克风、次级扬声器、CAN转速信号与ANC控制器；
S02：基于S01的车内布置，使ANC控制器播放的扫频信号通过次级扬声器播放后再通过误差麦克风的采集再次传递给ANC控制器；
S03：基于S02扫频信号和误差麦克风采集信号，采用滤波器长度为128的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数；
S04：基于S02扫频信号和误差麦克风采集信号，采用滤波器长度为1024的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数，用于计算收敛系数稳定性曲线；
S05：将S03得到的次级通道传递函数和S04得到的收敛系数稳定性曲线集成到FxLMS算法程序中，执行降噪程序对车辆内的噪声进行降低。
实施例2
一种基于加速工况的实车主动降噪方法，所述S01具体为，将误差麦克风设置在车框上；将次级扬声器安放在车门内，将CAN转速信号连接在OBD接口处，监测发动机转速；将ANC控制器安装在车内。
实施例3
一种基于加速工况的实车主动降噪方法，所述S02具体为，本发明提出主动降噪系统基于多通道自适应陷波器，即使用多个次级扬声器控制多个误差麦克风位置处的噪声水平，多通道主动降噪系统的次级通道框图如图2所示。假设在多通道主动降噪系统中采用J个次级扬声器和K个误差麦克风，每个次级扬声器与每个误差麦克风之间都存在一条次级通道，图2中Hjk表示第j个次级扬声器到第k个误差传感器之间的次级通道，整个系统的次级通道传函用Hs(z)表示；多通道自适应陷波滤波原理框图如图3所示。
x(n)为根据转速信号形成的参考信号，
Figure PCTCN2022105938-appb-000008
表示系统的次级通道估计，共有J×K条次级通道，参考信号x(n)及其90°相移信号分别与次级通道估计卷积后得到滤波参考信号R 0(n)、R 1(n)即J×K维矩阵，即有：
Figure PCTCN2022105938-appb-000009
多通道自适应陷波滤波系统中，滤波器权矢量W 1和W 2为两个J×1维向量，残余误差信号矢量E(n)为K×1维向量，Y(n)为控制器输出的次级声音信号即J×1维向量，由FxLMS算法得到两个自适应权重矢量的迭代公式为：
Figure PCTCN2022105938-appb-000010
所以，得到控制器输出次级声音信号有：
Y(n)＝x 0(n)W 0+x 1(n)W 1   (3)。
实施例4
一种基于加速工况的实车主动降噪方法，所述S03将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数具体为，
在主动降噪系统中需要考虑次级通道对控制系统的影响，因此需要对次级通路进行辨识。次级通路辨识模型如图4所示。在次级通路辨识模型中，x(n)代表次级通路测试过程中的噪声激励，也就是控制器输出的激励信号；d(n)代表测试过程中控制器接收到的麦克风采集得到电压信号或声压信号；y(n)代表自适应滤波器输出，即噪声激励经过滤波器后的响应；e(n)代表控制器接受到的电压信号与滤波器输出响应叠加后的残余误差信号；在次级通路辨识过程中，自适应滤波器的权矢量根据LMS算法不断进行更新迭代，使得残余误差信号e(n)不断逼近于0，也就是让滤波器输出响应y(n)不断逼近于控制器接收到的电压信号；当系统收敛，残余误差信号接近于0时，自适应滤波器权矢量系数即可以等效于次级通路脉冲响应函数。
实施例5
一种基于加速工况的实车主动降噪方法，所述S04计算收敛系数稳定性曲线具体为，在实车主动降噪系统中，假设有M个次级声源和L个误差麦克风，接下来从频域角度对窄带谐波噪声的主动控制进行分析；假设第l个误差信号在第n个谐波的复数成分记为E l(ω n)，第m个次级信号在这个谐波的复数成分记为W m(ω n)，则误差信号为
Figure PCTCN2022105938-appb-000011
其中D l(ω n)是初级声源造成的第l个复数误差信号，C lm(ω n)是第m个次级声源到第l个误差传感器在该频率下的复数响应，向量形式有
E(ω n)＝D(ω n)+C(ω n)W(ω n)   (5)
其中
E(ω n)＝[E 1(ω n),E 2(ω n),...,E L(ω n)] T
D(ω n)＝[D 1(ω n),D 2(ω n),...,D L(ω n)] T
W(ω n)＝[W 1(ω n),W 2(ω n),...,W M(ω n)] T
Figure PCTCN2022105938-appb-000012
对于单频噪声来说目标函数写成
J＝E HAE+W HBW   (6)
其中上标H代表向量或矩阵的埃尔米特转置(共轭转置)；E和W分别代表L×1的复数误差信号和M×1的复数次级声音信号，A和B分别是L×L和M×M的正定加权矩阵；式(6)也可以写成未加权误差信号模数平方和加上加权次级信号模数平方和：
J＝E HE+βW HW   (7)
结合式(5)目标函数可以写成变量W二次型的形式：
J＝D HD+W HC HD+D HCW+W H[C HC+βI]W   (8)
目标函数对于W实部(W R)和虚部(W I)的导数都是实数，所以可以定义复数梯度向量为：
Figure PCTCN2022105938-appb-000013
由于g的实部虚部相互独立，让g＝0设置J对于W R和W I的微分等于0，得到最优控制信号向量：
W opt＝-[C HC+βI] -1C HD   (10)
结合式(5)，复数梯度向量可写成：
g＝2[C HE+βW]   (11)
以与梯度向量反比的方向调整复数次级信号的实部和虚部，得到最速下降算法：
W(k+1)＝(1-αβ)W(k)-αC HE(k)   (12)
其中，α表示收敛系数。结合式(5)和式(10)迭代公式(12)写成：
(W(k+1)-W opt)＝[I-α(C HC+βI)](W(k)-W opt)   (13)
假设W(0)＝0，重复应用式(13)得到
W(k)-W opt＝-[I-α(C HC+βI)] kW opt   (14)
如果复数海塞矩阵写成复数酉矩阵的形式，标准化特征向量Q和特征值对角矩阵，Λ＝diag(λ 1,λ 2,...,λ M)，其中特征值都是实数，所以
C HC+βI＝QΛQ H   (15)
定义控制系统的主坐标为
V(k)＝Q H(W(k)-W opt)   (16)
所以式(14)写成
V(k)＝[1-αΛ] kV(0)   (17)
因为Λ是对角矩阵，控制系统主坐标的收敛是独立的，V(k)的第m个成分写成
Figure PCTCN2022105938-appb-000014
其中上述方程有效时需要保证-1＜1-αλ m＜1，得到基于收敛系数的稳定性条件：对于所有的m，0＜α＜2/λ m。
通过以上推导过程，结合式(15)和式(18)可以看到，在确定收敛系数α的值时，可通过次级通道传递函数矩阵计算收敛系数稳定性边界曲线。在汽车发动机转速变化时，阶次噪声频率也在变化，所以在该转速下满足系统稳定性的收敛系数限值也不同，即收敛系数稳定性边界值随发动机转速的变化而变化。
实施例6
一种基于加速工况的实车主动降噪方法，所述S05具体为，在算法调试运行前，次级通道辨识结果要集成到算法程序中对参考信号进行滤波处理，一般使用FIR滤波器等效次级通道脉冲响应函数。在这种应用场景下，滤波器长度的选择要考虑到降噪效果、系统稳定性和算法运算量。理论上，滤波器阶数越高，次级通道辨识结果越准确，越能反映次级扬声器到误差麦克风的频率响应；但是，滤波器长度越长，算法运算量会不断增加，且当滤波器长度大于某一值时，再增加FIR滤波器长度对降噪效果并没有明显的提升；为了兼顾算法运算量和辨识结果的准确性，辨识次级通道用于算法集成时FIR滤波器长度选择128；
基于收敛系数稳定性边界曲线的频率分辨率(或者是随发动机转速变化的分辨率)与FIR滤波器长度的选择有关。用于计算稳定性边界曲线的次级通道长度越长，相对应的，相同频率带宽范围内稳定性边界曲线的谱线越多，频率分辨率越精细，其关系可以写成：
df＝fs/Length_of_FIR   (19)
其中，df为稳定性边界曲线的频率分辨率，fs为次级通道辨识使用的采样率，Length_of_FIR为次级通道辨识采用的FIR滤波器长度。与用于算法集成不同，计算收敛系数稳定性边界曲线所使用的次级通道不涉及算法运算量，相对来说，使用较长的滤波器长度对于提高稳定性边界曲线精确性具有较大帮助，滤波器长度如果较短会因为分辨率原因造成某些频率范围内系统不稳定的现象。本发明使用滤波器长度为1024的FIR滤波器辨识次级通道用于计算收敛系数稳定性边界曲线；
收敛系数的大小不仅影响主动降噪系统的收敛速度，还会影响系统收敛时的稳态误差，当收敛系数过大时，虽然收敛速度增快，但是相应的稳态误差会有所增大，一定程度影响降噪效果。所以基于系统稳态误差，本发明在集成频率-收敛系数曲线时，将曲线收敛系数变化曲线为稳定性边界曲线的1/5，兼顾系统收敛速度、降噪效果和稳态误差等因素。
实施例7
一种基于加速工况的实车主动降噪系统，所述降噪系统包括
误差麦克风：用来采集误差信号并给到ANC控制器进行算法计算；
次级扬声器：ANC控制器算运行得到输出信号经功放给到次级扬声器；采用四个车门扬声器作为次级扬声器，
CAN总线：采集发动机转速来构造主动降噪系统的参考信号；
ANC控制器：ANC控制器用来执行FxLMS算法程序。
实施例8
一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时实现上述任一项所述方法的步骤。
实施例9
一种非临时性计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述任一项所述的方法的步骤。

Claims (9)

  1. 一种基于加速工况的实车主动降噪方法，其特征在于，所述实车主动降噪方法包括以下步骤：
    S01：在车内布置误差麦克风、次级扬声器、CAN转速信号与ANC控制器；
    S02：基于S01的车内布置，使ANC控制器播放的扫频信号通过次级扬声器播放后再通过误差麦克风的采集再次传递给ANC控制器；
    S03：基于S02扫频信号和误差麦克风采集信号，采用滤波器长度为128的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数；
    S04：基于S02扫频信号和误差麦克风采集信号，采用滤波器长度为1024的FIR滤波器，将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数，用于计算收敛系数稳定性曲线；
    S05：将S03得到的次级通道传递函数和S04得到的收敛系数稳定性曲线集成到FxLMS算法程序中，执行降噪程序对车辆内的噪声进行降低。
  2. 根据权利要求1所述一种基于加速工况的实车主动降噪方法，其特征在于，所述S01具体为，将误差麦克风设置在车框上；将次级扬声器安放在车门内，将CAN转速信号连接在OBD接口处，监测发动机转速；将ANC控制器安装在车内。
  3. 根据权利要求1所述一种基于加速工况的实车主动降噪方法，其特征在于，所述S02具体为，假设在多通道主动降噪系统中采用J个次级扬声器和K个误差麦克风，每个次级扬声器与每个误差麦克风之间都存在一条次级通道，整个系统的次级通道传函用Hs(z)表示；
    x(n)为根据转速信号形成的参考信号，
    Figure PCTCN2022105938-appb-100001
    表示系统的次级通道估计，共 有J×K条次级通道，参考信号x(n)及其90°相移信号分别与次级通道估计卷积后得到滤波参考信号R 0(n)、R 1(n)即J×K维矩阵，即有：
    Figure PCTCN2022105938-appb-100002
    多通道自适应陷波滤波系统中，滤波器权矢量W 1和W 2为两个J×1维向量，残余误差信号矢量E(n)为K×1维向量，Y(n)为控制器输出的次级声音信号即J×1维向量，由FxLMS算法得到两个自适应权重矢量的迭代公式为：
    Figure PCTCN2022105938-appb-100003
    所以，得到控制器输出次级声音信号有：
    Y(n)＝x 0(n)W 0+x 1(n)W 1   (3)。
  4. 根据权利要求3所述一种基于加速工况的实车主动降噪方法，其特征在于，所述S03将输出的扫频信号与误差麦克风采集到的信号进行辨识得到次级通道传递函数具体为，
    所述对次级通道进行辨识次级通道传递函数具体为，在次级通路辨识模型中，x(n)代表次级通路测试过程中的噪声激励，也就是控制器输出的激励信号；d(n)代表测试过程中控制器接收到的麦克风采集得到电压信号或声压信号；y(n)代表自适应滤波器输出，即噪声激励经过滤波器后的响应；e(n)代表控制器接受到的电压信号与滤波器输出响应叠加后的残余误差信号；在次级通路辨识过程中，自适应滤波器的权矢量根据LMS算法不断进行更新迭代，使得残余误差信号e(n)不断逼近于0，也就是让滤波器输出响应y(n)不断逼近于控制器接收到的电压信号；当系统收敛，残余误差信号接近于0时，自适应滤波器权矢量系数即可以等效于次级通路脉冲响应函数。
  5. 根据权利要求1所述一种基于加速工况的实车主动降噪方法，其特征在于，所述S04计算收敛系数稳定性曲线具体为，假设有M个次级声源和L个误差麦克风，假设第l个误差信号在第n个谐波的复数成分记为E l(ω n)，第m个次级信号在这个谐波的复数成分记为W m(ω n)，则误差信号为
    Figure PCTCN2022105938-appb-100004
    其中D l(ω n)是初级声源造成的第l个复数误差信号，C lm(ω n)是第m个次级声源到第l个误差传感器在该频率下的复数响应，向量形式有
    E(ω n)＝D(ω n)+C(ω n)W(ω n)   (5)
    其中
    E(ω n)＝[E 1(ω n),E 2(ω n),...,E L(ω n)] T
    D(ω n)＝[D 1(ω n),D 2(ω n),...,D L(ω n)] T
    W(ω n)＝[W 1(ω n),W 2(ω n),...,W M(ω n)] T
    Figure PCTCN2022105938-appb-100005
    对于单频噪声来说目标函数写成
    J＝E HAE+W HBW   (6)
    其中上标H代表向量或矩阵的埃尔米特转置；E和W分别代表L×1的复数误差信号和M×1的复数次级声音信号，A和B分别是L×L和M×M的正定加权矩阵；式(6)也可以写成未加权误差信号模数平方和加上加权次级信号模数平方和：
    J＝E HE+βW HW   (7)
    结合式(5)目标函数可以写成变量W二次型的形式：
    J＝D HD+W HC HD+D HCW+W H[C HC+βI]W   (8)
    目标函数对于W实部(W R)和虚部(W I)的导数都是实数，所以可以定义复数梯度向量为：
    Figure PCTCN2022105938-appb-100006
    由于g的实部虚部相互独立，让g＝0设置J对于W R和W I的微分等于0，得到最优控制信号向量：
    W opt＝-[C HC+βI] -1C HD   (10)
    结合式(5)，复数梯度向量可写成：
    g＝2[C HE+βW]   (11)
    以与梯度向量反比的方向调整复数次级信号的实部和虚部，得到最速下降算法：
    W(k+1)＝(1-αβ)W(k)-αC HE(k)   (12)
    其中，α表示收敛系数；结合式(5)和式(10)迭代公式(12)写成：
    (W(k+1)-W opt)＝[I-α(C HC+βI)](W(k)-W opt)   (13)
    假设W(0)＝0，重复应用式(13)得到
    W(k)-W opt＝-[I-α(C HC+βI)] kW opt   (14)
    如果复数海塞矩阵写成复数酉矩阵的形式，标准化特征向量Q和特征值对角矩阵，Λ＝diag(λ 1,λ 2,...,λ M)，其中特征值都是实数，所以
    C HC+βI＝QΛQ H   (15)
    定义控制系统的主坐标为
    V(k)＝Q H(W(k)-W opt)   (16)
    所以式(14)写成
    V(k)＝[1-αΛ] kV(0)   (17)
    因为Λ是对角矩阵，控制系统主坐标的收敛是独立的，V(k)的第m个成分写成
    Figure PCTCN2022105938-appb-100007
    其中上述方程有效时需要保证-1＜1-αλ m＜1，得到基于收敛系数的稳定性条件：对于所有的m，0＜α＜2/λ m；
    通过以上推导过程，结合式(15)和式(18)可以看到，在确定收敛系数α的值时，可通过次级通道传递函数矩阵计算收敛系数稳定性边界曲线；在汽车发动机转速变化时，阶次噪声频率也在变化，所以在该转速下满足系统稳定性的收敛系数限值也不同，即收敛系数稳定性边界值随发动机转速的变化而变化。
  6. 根据权利要求1所述一种基于加速工况的实车主动降噪方法，其特征在于，所述S05具体为，验证不同滤波器长度的FIR滤波器辨识次级通道效果具体为，辨识次级通道用于算法集成时FIR滤波器长度选择128；
    基于收敛系数稳定性边界曲线的频率分辨率与FIR滤波器长度的选择有关；用于计算稳定性边界曲线的次级通道长度越长，相对应的，相同频率带宽范围内稳定性边界曲线的谱线越多，频率分辨率越精细，其关系写成：
    df＝fs/Length_of_FIR   (19)
    其中，df为稳定性边界曲线的频率分辨率，fs为次级通道辨识使用的采样率，Length_of_FIR为次级通道辨识采用的FIR滤波器长度；与用于算法集成不同，计算收敛系数稳定性边界曲线所使用的次级通道不涉及算法运算量，相对来说，使用较长的滤波器长度对于提高稳定性边界曲线精确性具有较大帮助，滤波器长度如果较短会因为分辨率原因造成某些频率范围内系统不稳定的现象； 本发明使用滤波器长度为1024的FIR滤波器辨识次级通道用于计算收敛系数稳定性边界曲线；
    所以基于系统稳态误差，将曲线收敛系数变化曲线为稳定性边界曲线的1/5。
  7. 根据权利要求1所述一种基于加速工况的实车主动降噪系统，其特征在于，所述降噪系统包括
    误差麦克风：用来采集误差信号并给到ANC控制器进行算法计算；
    次级扬声器：ANC控制器算运行得到输出信号经功放给到次级扬声器；
    CAN总线：采集发动机转速来构造主动降噪系统的参考信号；
    ANC控制器：ANC控制器用来执行FxLMS算法程序。
  8. 一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1-6中任一项所述方法的步骤。
  9. 一种非临时性计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1-6中任一项所述的方法的步骤。
PCT/CN2022/105938 2022-02-14 2022-07-15 一种基于加速工况的实车主动降噪方法及系统 WO2023151229A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210131846.3 2022-02-14
CN202210131846.3A CN114677997A (zh) 2022-02-14 2022-02-14 一种基于加速工况的实车主动降噪方法及系统

Publications (1)

Publication Number Publication Date
WO2023151229A1 true WO2023151229A1 (zh) 2023-08-17

Family

ID=82072298

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/105938 WO2023151229A1 (zh) 2022-02-14 2022-07-15 一种基于加速工况的实车主动降噪方法及系统

Country Status (2)

Country Link
CN (1) CN114677997A (zh)
WO (1) WO2023151229A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114677997A (zh) * 2022-02-14 2022-06-28 中国第一汽车股份有限公司 一种基于加速工况的实车主动降噪方法及系统
CN115294953B (zh) * 2022-08-15 2023-05-05 浙江大学 一种多通道独立阶数滤波器的汽车车厢噪声主动控制方法
CN117690404B (zh) * 2024-02-04 2024-05-03 清华大学苏州汽车研究院（相城） 用于汽车的发动机噪声的主动控制方法及其装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108189724A (zh) * 2018-01-03 2018-06-22 上海工程技术大学 一种汽车反馈控制主动降噪颈枕
US20180190258A1 (en) * 2016-12-30 2018-07-05 Qualcomm Incorporated Adaptations for active noise cancellation inside a vehicle
CN111613202A (zh) * 2020-06-28 2020-09-01 清华大学苏州汽车研究院（相城） 一种用于车内降噪的主动控制系统及其控制方法
CN112509546A (zh) * 2020-10-22 2021-03-16 上海迪彼电子科技有限公司 一种针对汽车动态变化声场的主动降噪系统
CN113593517A (zh) * 2021-08-03 2021-11-02 深圳丽的声学科技有限公司 汽车智能分区主动降噪控制系统及方法
CN114677997A (zh) * 2022-02-14 2022-06-28 中国第一汽车股份有限公司 一种基于加速工况的实车主动降噪方法及系统

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
ATE402468T1 (de) * 2004-03-17 2008-08-15 Harman Becker Automotive Sys Geräuschabstimmungsvorrichtung, verwendung derselben und geräuschabstimmungsverfahren
JP4252074B2 (ja) * 2006-07-03 2009-04-08 政明 大熊 アクティブ消音装置におけるオンライン同定時の信号処理方法
EP2395501B1 (en) * 2010-06-14 2015-08-12 Harman Becker Automotive Systems GmbH Adaptive noise control
CN104038181B (zh) * 2014-06-05 2017-05-17 北京航空航天大学 一种基于nlms算法的自适应滤波器的构建方法
CN109448686A (zh) * 2018-12-13 2019-03-08 重庆邮电大学 基于次级通道在线辨识新算法交叉更新有源噪声控制系统
CN109859733A (zh) * 2019-01-02 2019-06-07 哈尔滨理工大学 基于fxlms算法的发动机噪声控制方法
CN111564151B (zh) * 2020-05-13 2022-09-23 吉林大学 一种车内发动机阶次噪声的窄带主动降噪优化系统
CN113257214A (zh) * 2021-02-04 2021-08-13 南京汉得利智能科技有限公司 一种用于风扇管道系统的主动降噪方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180190258A1 (en) * 2016-12-30 2018-07-05 Qualcomm Incorporated Adaptations for active noise cancellation inside a vehicle
CN108189724A (zh) * 2018-01-03 2018-06-22 上海工程技术大学 一种汽车反馈控制主动降噪颈枕
CN111613202A (zh) * 2020-06-28 2020-09-01 清华大学苏州汽车研究院（相城） 一种用于车内降噪的主动控制系统及其控制方法
CN112509546A (zh) * 2020-10-22 2021-03-16 上海迪彼电子科技有限公司 一种针对汽车动态变化声场的主动降噪系统
CN113593517A (zh) * 2021-08-03 2021-11-02 深圳丽的声学科技有限公司 汽车智能分区主动降噪控制系统及方法
CN114677997A (zh) * 2022-02-14 2022-06-28 中国第一汽车股份有限公司 一种基于加速工况的实车主动降噪方法及系统

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
CHEN, XIANGJUN: "Research on Self-adaptive Active Noise Control Technology in a Passenger Vehicle based on a Notch Delay LMS Algorithm", MASTER'S THESIS, 11 April 2019 (2019-04-11), CN, pages 1 - 69, XP009548257, DOI: 10.27151/d.cnki.ghnlu.2019.002522 *
ZHANG, SHIQIANG: "Active Control of Engine Noise In A Passenger Vehicle", MASTER'S THESIS, 8 June 2020 (2020-06-08), CN, pages 1 - 73, XP009548256, DOI: 10.27670/d.cnki.gcqdu.2020.002277 *

Also Published As

Publication number Publication date
CN114677997A (zh) 2022-06-28

Similar Documents

Publication Publication Date Title
WO2023151229A1 (zh) 一种基于加速工况的实车主动降噪方法及系统
EP3367378B1 (en) Active noise control using variable step-size adaptation
Jiang et al. Review of active noise control techniques with emphasis on sound quality enhancement
US9633645B2 (en) Adaptive noise control system with improved robustness
US10373600B2 (en) Active noise control system
JPH05265468A (ja) 能動型騒音制御装置
CN101552939A (zh) 车内声品质自适应主动控制系统和方法
CN108600894A (zh) 一种耳机自适应有源噪声控制系统及方法
WO2023124629A1 (zh) 一种车辆的主动降噪方法及设备、存储介质
CN107131054A (zh) 一种汽车车内主动降噪系统与装置
KR20200077397A (ko) 도로 소음 상쇄 시스템에서의 센서 소음층의 가청도 감소
CN114582312B (zh) 一种车内抗干扰自适应路噪主动控制方法及控制系统
CN103929704B (zh) 一种基于变换域的自适应声反馈消除的方法和系统
Patel et al. A hybrid multi-reference subband control strategy for active noise control headphones
Chen et al. A low-complexity multi-channel active noise control system using local secondary path estimation and clustered control strategy for vehicle interior engine noise
CN201395095Y (zh) 车内声品质自适应主动控制系统
CN113096629B (zh) 用于单通道反馈有源噪声控制系统的相对路径虚拟传感方法
Winberg et al. A new passive/active hybrid headset for a helicopter application
CN117594030B (zh) 一种主动降噪方法及相关装置
EP4224466A1 (en) Road noise cancellation shaping filters
Peng et al. Research on Application Technology of Local Sound Field Virtual Control Technology in Ship Cabin Noise Reduction
Atmojo et al. ANC Gen1: BBTA3-BPPT 1st Prototype of Active Noise Control for Vehicle Cabin Noise
Shi et al. An analysis of a feedback active noise control system using the remote microphone technique
Gui et al. Active noise control of refrigerator based on improved notch algorithm
Winberg Noise and vibration control of combustion engine vehicles

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22925588

Country of ref document: EP

Kind code of ref document: A1