CN112992182B - Vehicle wind noise level testing system and testing method thereof - Google Patents
Vehicle wind noise level testing system and testing method thereof Download PDFInfo
- Publication number
- CN112992182B CN112992182B CN202110185549.2A CN202110185549A CN112992182B CN 112992182 B CN112992182 B CN 112992182B CN 202110185549 A CN202110185549 A CN 202110185549A CN 112992182 B CN112992182 B CN 112992182B
- Authority
- CN
- China
- Prior art keywords
- wind noise
- time domain
- module
- noise level
- domain signal
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
- 238000012360 testing method Methods 0.000 title claims abstract description 39
- 238000004364 calculation method Methods 0.000 claims abstract description 35
- 238000001228 spectrum Methods 0.000 claims abstract description 31
- 238000012545 processing Methods 0.000 claims abstract description 24
- 238000013528 artificial neural network Methods 0.000 claims abstract description 23
- 238000013461 design Methods 0.000 claims abstract description 13
- 238000000034 method Methods 0.000 claims abstract description 11
- 239000010410 layer Substances 0.000 claims description 79
- 239000013598 vector Substances 0.000 claims description 35
- 241000282414 Homo sapiens Species 0.000 claims description 18
- 238000001914 filtration Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 16
- 210000005069 ears Anatomy 0.000 claims description 10
- 238000010183 spectrum analysis Methods 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 5
- 238000010998 test method Methods 0.000 claims description 3
- 239000011229 interlayer Substances 0.000 claims description 2
- 238000003062 neural network model Methods 0.000 abstract description 10
- 238000013507 mapping Methods 0.000 abstract description 4
- 238000011156 evaluation Methods 0.000 description 13
- 230000006870 function Effects 0.000 description 7
- 230000000694 effects Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000005070 sampling Methods 0.000 description 4
- 230000003595 spectral effect Effects 0.000 description 4
- 101150056353 SPL2 gene Proteins 0.000 description 3
- 101150042817 NFS1 gene Proteins 0.000 description 2
- 101100126298 Rickettsia conorii (strain ATCC VR-613 / Malish 7) iscS gene Proteins 0.000 description 2
- 101150114492 SPL1 gene Proteins 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M17/00—Testing of vehicles
- G01M17/007—Wheeled or endless-tracked vehicles
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/18—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being spectral information of each sub-band
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
- G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a vehicle wind noise level test system which comprises a calculation module and a data acquisition module which are connected with each other, wherein the data acquisition module respectively acquires a wind noise time domain signal and a wind noise level value, the calculation module comprises a wind noise time domain signal processing module and a neural network calculation module, the wind noise time domain signal processing module converts the wind noise time domain signal into a wind noise Mel frequency coefficient, the input layer of the neural network calculation module is the wind noise Mel frequency coefficient, and the output layer is the wind noise level value. The method comprises the steps of converting a time domain signal of wind noise into a Mel frequency coefficient, establishing a mapping relation between the Mel frequency coefficient and a wind noise level value by using a neural network model, and predicting the wind noise level value according to collected wind noise frequency spectrum data, so that multiple design schemes of the same vehicle type are compared, the efficiency of vehicle wind noise design is greatly improved, and the prediction precision is high.
Description
Technical Field
The invention relates to the technical field of vehicle wind noise testing, in particular to a vehicle wind noise level testing system and a testing method thereof.
Background
The evaluation of the wind noise level of the vehicle generally adopts the combination of objective test and objective evaluation. The objective test is to collect and analyze the vehicle wind noise data to obtain wind noise frequency spectrum data; and (4) performing objective evaluation, namely assigning the wind noise level of the vehicle by a professional based on objective auditory characteristics of human ears. The objective test plays a significant role in the aspects of vehicle wind noise data comparison, problem analysis and the like; the objective evaluation is also indispensable, and the level of the vehicle wind noise is finally judged by a user based on the auditory characteristics of human ears. Therefore, the objective evaluation of the wind noise of the vehicle is complementary to the objective evaluation, and the objective evaluation is not enough.
Since the objective evaluation of the wind noise of the vehicle cannot be obtained by a certain criterion, the objective evaluation value is difficult to obtain and predict from the wind noise spectrum data. On the other hand, the wind noise frequency spectrum data only considers the perception characteristic of human body to sound loudness, but does not consider the perception characteristic of human ears to sound frequency.
Mel is the unit of subjective guanyin height, mel frequency is extracted based on human ear auditory characteristics, and has consistency with human objective evaluation, and it is nonlinear relation with Hz (Hertz) frequency. Therefore, a mapping relation exists between the objective data of the vehicle wind noise and the objective evaluation data, and the relation is complex and nonlinear. In order to accurately predict the vehicle wind noise level, wind noise spectrum data obtained through testing needs to be converted into Mel spectrum data, and a relationship between the Mel spectrum data and an objective evaluation value is established, so that the vehicle wind noise level is predicted.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a vehicle wind noise level testing system and a testing method thereof.
In order to achieve the purpose, the invention provides a vehicle wind noise level testing system, which is characterized in that: the wind noise measuring device comprises a calculation module and a data acquisition module which are connected with each other, wherein the data acquisition module respectively acquires a wind noise time domain signal and a wind noise level value, the calculation module comprises a wind noise time domain signal processing module and a neural network calculation module, the wind noise time domain signal processing module is used for converting the wind noise time domain signal into a wind noise Mel frequency coefficient, an input layer of the neural network calculation module is the wind noise Mel frequency coefficient, and an output layer of the neural network calculation module is the wind noise level value.
Further, the wind noise time domain signal processing module comprises a frequency spectrum analysis module, and the frequency analysis module is used for outputting a wind noise energy spectrum according to the wind noise time domain signal.
Further, the spectrum analysis module comprises a sound pressure level calculation module, and the sound pressure level calculation module is used for outputting the wind noise pressure level according to the wind noise time domain signal.
Further, the spectrum analysis module comprises an energy spectrum calculation module, and the energy spectrum calculation module is used for outputting the wind noise energy spectrum according to the wind noise pressure level.
Further, the wind noise time domain signal processing module comprises a Mel filtering module, and the Mel filtering module is used for outputting a noise Mel filtering value according to the wind noise energy spectrum.
Further, the wind noise time domain signal processing module comprises a discrete cosine transform module, and the discrete cosine transform module is used for outputting a wind noise Mel frequency coefficient according to the wind noise Mel filtering value.
Further, the relationship between the output layer and the middle layer of the neural network computing module is as follows:
O＝W 2 *F(y)
where O is the output layer vector, W 2 And F (y) is an activation function, and y is an intermediate layer vector.
Further, the relation between the input layer and the middle layer of the neural network computing module is as follows:
y＝||W 1 -x||·*b 1
where y is the intermediate layer vector, x is the input layer vector, W 1 Is a weight coefficient matrix of the input layer and the intermediate layer, b 1 Is a bias matrix of the input layer vector and the intermediate layer vector.
The invention also provides a test method based on the vehicle wind noise level test system, which is characterized by comprising the following steps: under the same test road condition and working condition, collecting wind noise time domain signals and wind noise level values of vehicles of the same vehicle type and multiple design schemes, processing the wind noise time domain signals through a wind noise time domain signal processing module to obtain a wind noise Mel frequency coefficient, taking the wind noise Mel frequency coefficient as an input layer, taking the wind noise level value as an output layer, setting model parameters in a neural network computing module, changing the test road condition and the working condition, collecting the wind noise time domain signals of the vehicles of the same vehicle type and multiple design schemes again, then obtaining the wind noise Mel frequency coefficient, and inputting the wind noise Mel frequency coefficient into a neural network computing module to obtain a wind noise score prediction value.
Further, the method for determining the wind noise level value comprises the step of assigning the vehicle wind noise level based on the auditory characteristics of human ears.
The invention has the beneficial effects that: according to the method, the time domain signal of the wind noise is converted into the Mel frequency coefficient, the neural network model is utilized to establish the mapping relation between the Mel frequency coefficient and the wind noise level value, the wind noise level value can be predicted according to the collected wind noise frequency spectrum data, multiple design schemes of the same vehicle type are compared according to the wind noise level value, the design scheme with the optimal wind noise level is selected, the efficiency of vehicle wind noise design is greatly improved, and the prediction precision is high.
Drawings
FIG. 1 is a schematic structural diagram of a vehicle wind noise level testing system according to the present invention.
FIG. 2 is a flow chart of a method for testing the wind noise level of a vehicle according to the present invention.
FIG. 3 is a graph of Mel frequency versus Hz frequency.
Fig. 4 is a schematic diagram of a triangular filter bank in the present embodiment.
Detailed Description
The following detailed description is provided to further explain the claimed embodiments of the present invention and to enable those skilled in the art to more clearly understand the claims. The scope of the invention is not limited to the following specific examples. It is within the purview of one skilled in the art to effect the invention in variations of the embodiments described below including what is claimed herein and other embodiments.
As shown in fig. 1 to 4, a vehicle wind noise level testing system includes a calculation module and a data collection module which are connected to each other, the data collection module respectively collects a wind noise time domain signal and a wind noise level value, the calculation module includes a wind noise time domain signal processing module and a neural network calculation module, the wind noise time domain signal processing module converts the wind noise time domain signal into a wind noise Mel frequency coefficient, an input layer of the neural network calculation module is the wind noise Mel frequency coefficient, and an output layer is the wind noise level value. The Mel frequency is consistent with the auditory characteristics of human ears, and a mapping relation between the Mel frequency coefficient and the wind noise level value can be established by utilizing a neural network model, so that the wind noise level value is predicted.
In this embodiment, the wind noise time domain signal processing module includes a frequency spectrum analysis module, an input end of the frequency analysis module is a wind noise time domain signal, and an output end of the frequency analysis module is a wind noise energy spectrum. The frequency spectrum analysis module comprises a sound pressure level calculation module, and the sound pressure level calculation module outputs a wind noise pressure level according to a wind noise time domain signal; the frequency spectrum analysis module comprises an energy spectrum calculation module which outputs an air noise energy spectrum according to the air noise pressure level.
The calculation process of the sound pressure level calculation module is as follows:
applying a window function to the collected wind noise time domain signal, and carrying out FFT analysis to obtain a wind noise frequency spectrum
p(k)＝FFT[x(t)]
In the formula, p (k) is a frequency spectrum of a k-th spectral line of the wind noise, x (t) is a wind noise time domain signal, wherein t is time, x is wind noise pressure, and x (t) represents the change relation of the wind noise pressure along with time.
And taking a model of the wind noise frequency spectrum to obtain a wind noise amplitude spectrum A (k).
Converting an amplitude spectrum of wind noise into a wind noise pressure level
Wherein SPL1 (k) is the sound pressure level of the k-th spectral line, po is reference sound pressure, and 2 x 10 are taken -5 Pa。
The calculation process of the energy spectrum calculation module is as follows:
the sound pressure level of the wind noise is converted into a weighted sound pressure level.
SPL2(k)＝SPL1(k)-L(k)
SPL2 (k) is the weighted sound pressure level of the kth spectral line, L (k) is the attenuation value of the weighted characteristic curve at a certain frequency, and A is the weighted sound pressure level in the embodiment.
The weighted sound pressure level SPL2 (k) is squared to obtain an energy spectrum E (k).
In this embodiment, the wind noise time domain signal processing module includes a Mel filtering module, an input end of the Mel filtering module is a wind noise energy spectrum, and an output end of the Mel filtering module is a wind noise Mel filtering value.
The calculation process of the Mel filtering module is as follows:
the sound level heard by human ear is not linear with the frequency of sound, and the Mel frequency scale is more suitable for the auditory characteristic of human ear. The specific relationship between Mel frequency and actual frequency can be expressed by the following formula:
mel＝2595*tog 10 (1+f/700)
wherein Mel is Mel frequency, and f is Hz frequency.
The Mel scale has a high resolution at low frequencies (Hz) and a low resolution at high frequencies (Hz), which are consistent with the auditory characteristics of the human ear. Similar to the critical band division, the Mel triangular filter bank needs to be designed as follows.
Wherein, hm (k) is the amplitude of the kth spectral line of the mth filter, f (M) is the center frequency of the mth filter, M is more than or equal to 1 and less than or equal to M, and M is the number of the filters.
The endpoint frequency and center frequency f (m) of each filter of the Mel triangular filter bank are calculated as: the lower frequency limit of the noise of interest, e.g., 500Hz, is selected and its corresponding mel value, denoted mel _ min, is determined. According to the sampling frequency of the wind noise data and the Shannon sampling theorem, the upper limit of the frequency can be determined, if the sampling frequency is 10000Hz, the upper limit of the frequency is 10000/2=5000Hz, and the corresponding mel value is obtained and is recorded as mel _ max. The center frequency f (j) of each filter of the Mel triangular filter bank is:
wherein, N is the data block length of the wind noise time domain signal when FFT analysis is carried out; fs is the wind noise sampling frequency, the left end point frequency is f (j-1), and the right end point frequency is f (j + 1).
Finally, the wind noise Mel filtered value F (m) is
F(m)＝∑ k E(k)*H m (k)，1≤m≤M
In this embodiment, the wind noise time domain signal processing module includes a discrete cosine transform module, an input end of the discrete cosine transform module is a wind noise Mel filter value, and an output end of the discrete cosine transform module is a wind noise Mel frequency coefficient.
The discrete cosine transform module is calculated as follows:
first, a discrete cosine transform coefficient w (t) is calculated
Then calculate the Mel frequency coefficient C
Wherein, C (t) is the t-th Mel frequency coefficient.
The discrete cosine transform is carried out to eliminate the correlation among Mel frequency filtering values F (m), so that wind noise characteristic parameters are independent from each other, and the accuracy of wind noise prediction scoring of testers in a subsequent neural network model is improved.
In this embodiment, the neural network computing module adopts a GRNN neural network model, and is composed of three layers, where the first layer is an input layer, and n1 nodes are provided in total, and correspond to n1 input parameters. The second layer is an intermediate layer (hidden layer), n2 nodes are provided in total, and n2 is equal to the sample number of sample data. The third layer is an output layer, n3 nodes are provided in total, and the output layer is determined by the response actually required by the system.
Let input layer vector x = (x) 1 ,x 2 ,...,x n1 ) T Intermediate layer vector y = (y) 1 ，y 2 ，...，y n2 ) T Output layer vector o = (o) 1 ，o 2 ，...，o n3 ) T 。
In this embodiment, the relationship between the output layer and the intermediate layer of the neural network computing module is as follows:
o＝W 2 *F(y)
where o is the output layer vector, W 2 And F (y) is an activation function, and y is an intermediate layer vector.
In this embodiment, the relationship between the input layer and the intermediate layer of the neural network computing module is as follows:
y＝||W 1 -x||·*b 1
where y is the interlayer vector, x is the input layer vector, W 1 As a matrix of weighting coefficients for the input and intermediate layers, b 1 Is the offset matrix of the input layer vector and the intermediate layer vector,. Is the product of the quantities, i.e. the product of each corresponding element in the matrix. And | | is the Euclidean distance between two vectors, and the calculation formula is as follows.
Where L is the dimension of the vector.
The intermediate layer vector y can be expressed in the form
||W 1 -x||.*b 1
＝diag((W 1 -ones(n2，1)*x T )*(W 1 -ones(n2，1)*x T ) T ).^(0.5)
*b 1
Wherein, diag represents a column vector formed by elements on the main diagonal of the matrix; T representing a transposition; a represents the quantity squared (i.e., the power of each corresponding element in the matrix); is the product of quantities, i.e. the product of each corresponding element in the matrix. ones (n 2, 1) represents a column vector having a dimension n2, each of which has a value of 1, and n2 is the number of intermediate layer nodes.
The activation function F (y) is typically a gaussian function, as follows:
in this embodiment, the number n1 of nodes in the input layer is equal to the number of wind noise characteristic parameters, that is, the number M of filters, the number n3=1 in the output layer is a subjective evaluation value of the vehicle wind noise level, and the number n2 of nodes in the hidden layer is a vehicle sample amount R.
Determining parameters of the neural network model, and making a weight coefficient matrix W of the input layer and the intermediate layer 1 ＝x T Let the weighting coefficient matrix W of the output layer and the middle layer 2 = o, let the offset coefficient b of each node of the input layer vector and the intermediate layer vector 1i Equal to 0.8326 and an activation function F (y) of 0.5.
The GRNN neural network model has few parameters needing to be adjusted, has no training process, avoids complicated and tedious mathematical calculation and has high prediction speed; the GRNN neural network model has a simple structure and good fitting effect and prediction effect; the GRNN neural network model has a good prediction effect on small sample data, a large number of wind noise data samples are not required to be obtained, and the BP neural network model can obtain a good prediction effect only by training the model through a large number of samples.
The test method of the vehicle wind noise level test system comprises the following steps:
1. under the same test road condition and working condition, collecting wind noise time domain signals and corresponding R wind noise level values of vehicles with R design schemes of the same vehicle type. The input layer vector x is a matrix of R columns and M rows and the output layer vector o is a row vector of dimension R. The wind noise level value is assigned by a tester based on objective auditory characteristics of human ears, and the objective auditory characteristics of human ears refer to objective reflection of loudness, frequency and audio regularity of sound by human beings, so that the wind noise level comprehensively reflects the loudness, frequency and audio regularity of wind noise waves.
2. And processing the wind noise time domain signal through a wind noise time domain signal processing module to obtain a wind noise Mel frequency coefficient. The specific processing process comprises the steps of obtaining a wind noise pressure level by using a sound pressure level calculation module according to a wind noise time domain signal, obtaining a wind noise energy spectrum by using an energy spectrum module, obtaining a wind noise Mel filtering value by using a Mel filtering module, and obtaining a wind noise Mel frequency coefficient by using a discrete cosine transform module.
3. Taking wind noise Mel frequency coefficient as an input layer, taking a wind noise level value as an output layer, bringing the wind noise level value into a neural network computing module, setting model parameters in the neural network computing module, wherein the model parameters comprise a weight coefficient matrix of the input layer and an intermediate layer and a weight coefficient matrix W of the output layer and the intermediate layer 2 = o, offset coefficient b of each node of input layer vector and intermediate layer vector 1i And activate function F (y).
4. And (3) changing the tested road condition and working condition, collecting wind noise time domain signals of the vehicles of the same vehicle type with the R design schemes again, repeating the step (2) to obtain a wind noise Mel frequency coefficient, and inputting the wind noise Mel frequency coefficient into the neural network computing module, wherein at the moment, because the model parameters are set, an output layer vector can be directly obtained to be used as a wind noise score prediction value.
After the tester obtains the wind noise level values of the R vehicles under different testing road conditions and working conditions, the optimal design scheme can be selected according to the wind noise level values of different design schemes, the condition that the tester needs to make scores for many times under different testing road conditions and working conditions is avoided, and the whole testing efficiency is improved.
Claims (10)
1. The utility model provides a vehicle wind noise level test system which characterized in that: the wind noise control system comprises a calculation module and a data acquisition module which are connected with each other, wherein the data acquisition module respectively acquires a wind noise time domain signal and a wind noise level value, the calculation module comprises a wind noise time domain signal processing module and a neural network calculation module, the wind noise time domain signal processing module is used for converting the wind noise time domain signal into a wind noise Mel frequency coefficient, the input layer of the neural network calculation module is the wind noise Mel frequency coefficient, and the output layer is the wind noise level value; the wind noise level value is evaluated by a tester based on objective auditory characteristics of human ears, and the objective auditory characteristics of human ears refer to the objective reflection of human loudness, frequency and audio regularity to sound.
2. The vehicle wind noise level testing system of claim 1, wherein: the wind noise time domain signal processing module comprises a spectrum analysis module, and the spectrum analysis module is used for outputting a wind noise energy spectrum according to the wind noise time domain signal.
3. The vehicle wind noise level testing system of claim 2, wherein: the frequency spectrum analysis module comprises a sound pressure level calculation module which is used for outputting the wind noise pressure level according to the wind noise time domain signal.
4. The vehicle wind noise level testing system of claim 2, wherein: the spectrum analysis module comprises an energy spectrum calculation module, and the energy spectrum calculation module is used for outputting an air noise energy spectrum according to an air noise pressure level.
5. The vehicle wind noise level testing system of claim 1, wherein: the wind noise time domain signal processing module comprises a Mel filtering module, and the Mel filtering module is used for outputting a wind noise Mel filtering value according to a wind noise energy spectrum.
6. The vehicle wind noise level testing system of claim 1, wherein: the wind noise time domain signal processing module comprises a discrete cosine transform module, and the discrete cosine transform module is used for outputting wind noise Mel frequency coefficients according to a wind noise Mel filtering value.
7. The vehicle wind noise level testing system of claim 1, wherein: the relation between the output layer and the middle layer of the neural network computing module is as follows:
O＝W 2 *F(y)
where O is the output layer vector, W 2 F (y) is the weight coefficient matrix of the output layer and the middle layer, and is the activation function, and y is the middle layer vector.
8. The vehicle wind noise level testing system of claim 1, wherein: the relation between the input layer and the middle layer of the neural network computing module is as follows:
y＝||W 1 -x||·b 1
where y is the interlayer vector, x is the input layer vector, W 1 Is a weight coefficient matrix of the input layer and the intermediate layer, b 1 Is the offset matrix of the input layer vector and the intermediate layer vector, | | | | is the euclidean distance between the two vectors.
9. A test method based on the vehicle wind noise level test system of any one of claims 1 to 8, characterized in that: under the same test road condition and working condition, collecting wind noise time domain signals and wind noise level values of vehicles of the same vehicle type and multiple design schemes, processing the wind noise time domain signals through a wind noise time domain signal processing module to obtain a wind noise Mel frequency coefficient, taking the wind noise Mel frequency coefficient as an input layer and the wind noise level values as an output layer, setting model parameters in a neural network computing module, changing the test road condition and the working condition, collecting the wind noise time domain signals of the vehicles of the same vehicle type and the multiple design schemes again, then obtaining the wind noise Mel frequency coefficient and inputting the wind noise Mel frequency coefficient into a neural network computing module to obtain the wind noise level values.
10. The method for testing a vehicle wind noise level test system according to claim 9, wherein: the method for determining the wind noise level value comprises the step of assigning the vehicle wind noise level based on the auditory characteristics of human ears.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110185549.2A CN112992182B (en) | 2021-02-10 | 2021-02-10 | Vehicle wind noise level testing system and testing method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110185549.2A CN112992182B (en) | 2021-02-10 | 2021-02-10 | Vehicle wind noise level testing system and testing method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112992182A CN112992182A (en) | 2021-06-18 |
| CN112992182B true CN112992182B (en) | 2023-03-31 |
Family
ID=76393979
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110185549.2A Active CN112992182B (en) | 2021-02-10 | 2021-02-10 | Vehicle wind noise level testing system and testing method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112992182B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114544194B (en) * | 2022-01-25 | 2023-06-23 | 东风汽车集团股份有限公司 | Vehicle road noise evaluation method based on spectrum analysis |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5568404A (en) * | 1994-10-24 | 1996-10-22 | Ford Motor Company | Method and system for predicting sound pressure levels within a vehicle due to wind noise |
| CN104361887A (en) * | 2014-11-20 | 2015-02-18 | 哈尔滨工业大学 | Quick acoustic event detection system under traveling noise environment |
| CN105841797B (en) * | 2016-03-15 | 2019-07-02 | 中南大学 | A kind of Car window electric machine extraordinary noise detection method and device based on MFCC and SVM |
| CN207147752U (en) * | 2017-07-14 | 2018-03-27 | 中汽研（常州）汽车工程研究院有限公司 | Automotive interior wind is made an uproar experimental apparatus for testing |
| JP7052552B2 (en) * | 2018-05-17 | 2022-04-12 | トヨタ自動車株式会社 | Wind noise analysis device and wind noise analysis method |
| CN108896323A (en) * | 2018-08-31 | 2018-11-27 | 重庆长安汽车股份有限公司 | A kind of automobile wind is made an uproar the test method of leakage sound |
| CN110688712A (en) * | 2019-10-11 | 2020-01-14 | 湖南文理学院 | Evaluation index for objective annoyance degree of automobile wind vibration noise sound quality and calculation method thereof |
| CN110990752B (en) * | 2019-10-31 | 2023-03-28 | 重庆长安汽车股份有限公司 | Method for evaluating level of noise in automobile during driving process of automobile |
| CN112067117B (en) * | 2020-08-14 | 2022-10-28 | 中国第一汽车股份有限公司 | Method for evaluating automobile wind noise performance |
| CN112161815A (en) * | 2020-09-07 | 2021-01-01 | 东风汽车集团有限公司 | Vehicle road noise subjective evaluation value prediction method |
| CN112131662A (en) * | 2020-09-17 | 2020-12-25 | 东风汽车集团有限公司 | Passenger car wind noise subjective evaluation objective quantification method |
-
2021
- 2021-02-10 CN CN202110185549.2A patent/CN112992182B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN112992182A (en) | 2021-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109682561B (en) | Method for automatically detecting free vibration response of high-speed railway bridge to identify mode | |
| CN108417228A (en) | Voice tone color method for measuring similarity under instrument tamber migration | |
| CN101426169B (en) | Time-domain tracking filter quickly detecting acoustic response parameter of sounding body and system | |
| CN108663115A (en) | A kind of car inside idle noise objective quantification evaluation method | |
| CN101426168B (en) | Sounding body abnormal sound detection method and system | |
| CN102857852B (en) | Method for processing playback array control signal of loudspeaker of sound-field quantitative regeneration control system | |
| CN108630209B (en) | Marine organism identification method based on feature fusion and deep confidence network | |
| CN103546853A (en) | Speaker abnormal sound detecting method based on short-time Fourier transformation | |
| CN105136280A (en) | System and method for testing single noise quality in multi-source noise environment | |
| CN105371945B (en) | Remove the method and device of noise in hall impulse response measurement acoustical signal | |
| CN113763986B (en) | Abnormal sound detection method for air conditioner indoor unit based on sound classification model | |
| CN105301099A (en) | Food crispness detection method | |
| CN110596247B (en) | Ultrasonic structure health monitoring method in temperature change environment | |
| CN112052712B (en) | Power equipment state monitoring and fault identification method and system | |
| CN113901379B (en) | Real-time data dynamic online quick processing method for edge terminal | |
| CN112992182B (en) | Vehicle wind noise level testing system and testing method thereof | |
| CN106997765A (en) | The quantitatively characterizing method of voice tone color | |
| CN113188797B (en) | Bearing fault diagnosis method based on microphone array | |
| CN116698986B (en) | Sound absorbing material quality control system and method based on tuned band-pass filter | |
| US6453253B1 (en) | Impulse response measuring method | |
| CN107785025A (en) | Noise remove method and device based on room impulse response duplicate measurements | |
| CN112735468A (en) | MFCC-based automobile seat motor abnormal noise detection method | |
| CN117473263A (en) | Automatic recognition method and system for frequency and damping ratio of bridge vibration monitoring | |
| CN111855125B (en) | Quantitative evaluation method for automobile door lock closing sound tremor degree | |
| WO2023060542A1 (en) | Bearing fault detection method and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |