Brake noise detection method, device, and program

Abstract

To enable the presence of multiple types of brake noise to be inspected and brake noise to be identified during a test run in a finished vehicle inspection process.SOLUTION: A brake noise detection device which is part of a finished vehicle inspection device comprises: a sound acquisition unit 10 for acquiring the sound generated in the vicinity of wheels during travelling of the vehicle under test and generating sound data; a vehicle signal acquisition unit 11 for acquiring information on the driving condition of the vehicle such as a brake signal and a vehicle speed signal; a vehicle state determination unit 12 for determining correspondence to a plurality of specific vehicle driving conditions on the basis of the acquired vehicle signals; and a brake noise determination unit 13 for analyzing the sound data and extracting a sound that can be a candidate for multiple types of brake noise, and determining the type of brake noise on the basis of a vehicle driving condition when the sound that is a candidate for brake noise is generated.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a brake noise detection technique for detecting abnormal noise generated near a brake device while a vehicle such as an automobile is traveling. [Background technology]

[0002] For example, abnormal noises may occur during driving when a baffle plate installed to protect the disc rotor of a brake system comes into contact with the disc rotor. Furthermore, when the brake system is activated, an abnormality in the brake pads that come into contact with the disc rotor may cause an abnormal noise known as "brake squeal." For example, in the final stage of the finished vehicle inspection process on an automobile production line, inspectors use their ears to determine whether or not abnormal noises are present, a so-called sensory inspection. However, it is generally difficult for inspectors to distinguish between abnormal noises caused by such different factors and make an accurate judgment.

[0003] Patent Document 1 describes an operating sound inspection method for inspecting finished motors, in which the operating sound of the motor being inspected is acquired via a microphone, and the operating sound data for, for example, two seconds is divided into multiple blocks and then subjected to a fast Fourier transform for frequency analysis. The frequency characteristics of the operating sound of the motor being inspected are compared with the frequency characteristics of a good motor that has been determined in advance to be a good product, and whether the operating sound is abnormal is determined based on the difference between the two. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-118717 Summary of the Invention [Problem to be solved by the invention]

[0005] The technology of Patent Document 1 mentioned above can determine whether the motor under test is defective as a whole, but cannot distinguish between multiple abnormal noises caused by different factors, as in the case of a brake device. [Means for solving the problem]

[0006] The brake noise detection method according to the present invention comprises: The sound generated around the wheels of the target vehicle while it is running is acquired and used as sound data. This sound data is analyzed to extract multiple types of brake noise candidates, Acquire information on the vehicle's driving state and determine whether the information corresponds to a plurality of specific vehicle driving states; The type of abnormal brake noise is determined based on the vehicle operating conditions at the time when the noise that is a candidate for abnormal brake noise is generated. [Effects of the Invention]

[0007] According to this invention, there is no need to collect sounds generated near the wheel section multiple times. By detecting abnormal sounds in association with the vehicle driving state, when an abnormal sound is detected, it is possible to identify which of several types of abnormal brake sounds it is, and false detection of abnormal brake sounds can be suppressed. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a functional block diagram of a first embodiment in which the present invention is applied to a finished vehicle inspection process. [Figure 2] 4 is a flowchart showing the flow of processing in the first embodiment. [Figure 3] FIG. 10 is a functional block diagram of a second embodiment. [Figure 4] 10 is a flowchart showing the flow of processing in a second embodiment. [Figure 5] FIG. 10 is a functional block diagram of a third embodiment. [Figure 6] 10 is a flowchart showing the flow of processing according to a third embodiment. [Figure 7] FIG. 4 is an explanatory diagram showing a display example on a display unit. DETAILED DESCRIPTION OF THE INVENTION

[0009] An embodiment of the present invention will now be described. The brake noise detection of this embodiment is performed, for example, as an inspection to confirm that no abnormal noise is generated from the brake system during the finished vehicle inspection process, which is the final stage of an automobile production line. Generally, during the finished vehicle inspection process, an inspector test-drives the finished vehicle to be inspected on free rollers, inspecting numerous items such as the engine, meters, and brakes. During this test run on the free rollers, an abnormal noise inspection is performed based on sounds generated near the wheels during an appropriate inspection period (e.g., several seconds to 10-odd seconds) that includes acceleration and braking. This brake noise detection device is configured as part of the inspection equipment used in the finished vehicle inspection process.

[0010] 1 shows a functional block diagram of the brake noise detection device of the first embodiment. The brake noise detection device of the first embodiment is configured to include a sound acquisition unit 10, a vehicle signal acquisition unit 11, a vehicle state determination unit 12, and an abnormal brake noise determination unit 13.

[0011] The sound acquisition unit 10 includes a microphone that acquires sounds generated from the wheel section and converts them into electrical signals, i.e., sound data, and a recording unit that temporarily stores the sound data. The microphone is installed outside the vehicle so that it can collect sounds from the wheel section of the vehicle running on free rollers. The microphone's directivity and frequency characteristics are selected according to the position relative to the vehicle being measured and the required frequency band. Typically, microphones with directivity toward each wheel section of the vehicle are used. A microphone array or the like may be used to localize the sound source and obtain sound data from which noise has been removed. For example, this microphone may be an existing microphone installed to collect horn sounds during the finished vehicle inspection process. As described above, the abnormal sound inspection is performed while the vehicle is running for, for example, about 10 seconds, and the sound acquisition unit 10 acquires a series of sound data having a duration of, for example, 10 seconds.

[0012] The vehicle signal acquisition unit 11 acquires various information indicating the vehicle's operating state, i.e., vehicle signals, in parallel with the acquisition of sounds by the sound acquisition unit 10. The vehicle signal acquisition unit 11 includes multiple components, such as a brake signal acquisition unit 11-1 and an acceleration signal acquisition unit 11-2, that individually acquire several pieces of information. The brake signal acquisition unit 11-1 acquires a brake signal, which is an ON / OFF signal indicating that the driver (inspector) has operated the brakes (depressed the brake pedal). The acceleration signal acquisition unit 11-2 acquires an accelerator pedal position signal or a vehicle speed signal to determine the vehicle acceleration. These vehicle signals can be acquired, for example, from a computer system on the vehicle via CAN communication or the like. Alternatively, detection devices may be temporarily attached to the brake pedal, accelerator pedal, etc. for inspection, and the position signals, etc., may be acquired from these detection devices via wired or wireless communication.

[0013] The vehicle state determination unit 12 determines whether the vehicle operating state corresponds to a plurality of predetermined specific vehicle operating states based on the vehicle signals acquired by the vehicle signal acquisition unit 11. For example, if the brake signal is ON, it determines that the vehicle operating state satisfies the brake squeal generation condition, and if the brake signal is OFF, it determines that the vehicle operating state does not satisfy the brake squeal generation condition. Note that "brake squeal" refers to an abnormal sound (a sound different from normal brake operation sounds) that occurs when a friction member comes into contact with a rotating member during braking (in the case of disc brakes, the brake pad comes into contact with the disc rotor), and can be expressed as onomatopoeia such as "squeak," "beep," or "chirp."

[0014] In addition, the brake operating sound (which can be expressed as a "goo" sound) that occurs when the brakes are applied at low vehicle speeds is a normal sound, so when the brake signal changes from OFF to ON at low vehicle speeds, it is determined that the vehicle operating state satisfies the exclusion conditions for brake squeal detection.Since the sound that occurs when the brakes are applied at relatively high vehicle speeds is abnormal "brake squeal," when the brake signal changes from OFF to ON at relatively high vehicle speeds, it is determined that the vehicle operating state does not satisfy the exclusion conditions for brake squeal detection.

[0015] The abnormal brake noise determination unit 13 performs frequency analysis on the sound data acquired by the sound acquisition unit 10, for example, using a short-time Fourier transform (STFT), and then extracts multiple types of candidate abnormal brake noises. Finally, the type of abnormal brake noise is determined based on the vehicle operating conditions at the time the candidate abnormal brake noise is generated. In one embodiment, the abnormal brake noises are classified into the aforementioned "brake squeal" and "baffle plate contact noise," which is an abnormal noise generated by interference or vibration between a rotating member and a non-rotating member, regardless of brake operation. The latter abnormal noise can be expressed as onomatopoeia such as "sniffing," "clank," "rattling," or "chat." For example, when the baffle plate contacts the disc rotor, "baffle plate contact noise" such as "sniffing" or "clank" occurs. If a component near the brake device is not properly secured, so-called rattle noises such as "rattling" or "chattle" may occur. In the following description, the latter abnormal noise, including rattle noise, will be referred to as "baffle plate contact noise."

[0016] In the first embodiment, the brake noise judgment unit 13 simultaneously processes a series of sound data to extract sounds that are candidates for multiple types of brake noise, and judges the type of brake noise for each of the extracted multiple brake noise candidates based on whether or not the vehicle is in a specific operating state.

[0017] 2 is a flowchart showing the processing flow of the brake noise detection device of the first embodiment. First, in step 101, acquisition of sound via a microphone is started, and at the same time, acquisition of vehicle signals is started in step 102. In the next step 103, the acceleration state of the vehicle is determined from changes in the vehicle speed signal or the accelerator pedal position signal.

[0018] In step 104, a short-time Fourier transform (STFT) is performed on the acquired sound data to analyze the frequency characteristics, thereby obtaining the frequency-sound pressure level characteristics (for example, a spectrum with frequency on the horizontal axis and sound pressure level on the vertical axis) for each short time period.

[0019] The next steps 105 and 106 are executed substantially simultaneously for each short time period of the short-time Fourier transform in order to extract candidates for two types of brake abnormal noise in parallel.

[0020] On the other hand, in step 105, peaks in a relatively low predetermined frequency band (e.g., 2000 to 3000 Hz) are detected to extract candidates for baffle plate contact noise. Then, in step 107, it is determined whether more than a predetermined number of peaks (e.g., 5) are included. Peak detection can be performed using an appropriate known method. For example, the result of a short-time Fourier transform is differentiated twice, and the part with a negative minimum value is determined to be a peak. Alternatively, a frequency having a sound pressure level equal to or greater than a preset threshold value may be determined to be a peak. If more than a predetermined number of peaks are included, the process proceeds to step 109, where the sound at that time is determined to be a candidate for baffle plate contact noise (abbreviated as BP contact noise in the figure). Then, the process proceeds to step 111, which will be described later. If the number of detected peaks is small in the determination in step 107, the process proceeds to step 121, where it is determined to be a normal running noise. Baffle plate contact noises (including rattle noises) such as "tinkling," "shaking," "clacking," and "rattling" span a relatively wide frequency range and generally include multiple (e.g., 10 or more) frequency peaks.

[0021] In step 106, peaks in a relatively high frequency band (e.g., 10 kHz or higher) are detected to extract candidates for brake squeal. Then, in step 108, it is determined whether fewer than a predetermined number of peaks (e.g., five) are included. If fewer than the predetermined number of peaks are included, the process proceeds to step 110, where the sound is determined to be a candidate for brake squeal. The process then proceeds to step 111, which will be described later. If the determination in step 108 is NO, the process proceeds to step 120, where the sound is determined to be a normal driving sound. Abnormal sounds such as "squeak," "peep," and "chirp," i.e., brake squeal, generally have a small number of peaks, approximately one to three, in a relatively high frequency range. For example, a relatively high frequency band of 10 kHz or higher corresponds to the harmonics of the fundamental frequency of these brake squeals. By focusing on the harmonics of the brake squeal sound, it becomes easier to distinguish it from other noises in the factory where finished vehicle inspections are performed.

[0022] The number of peaks that serve as the threshold in step 107 and step 108 may be different from each other.

[0023] In step 111, it is determined whether two types of abnormal noise candidates have been detected by the processing up to that point. Note that in the illustrated example, two types of abnormal brake noise, brake squeal and baffle plate contact noise, are targeted, but it is also possible to extract candidates for other types of abnormal brake noise, and in this case, the processing for extracting candidate abnormal brake noises is also executed simultaneously in parallel with steps 105 and 106.

[0024] If the determination in step 111 is YES, the process proceeds to step 112, where it is determined whether the brake signal is OFF. If the brake signal is OFF, the process proceeds to step 113, where it is finally determined that the abnormal noise is baffle plate contact noise.

[0025] If the brake signal is ON, the process proceeds from step 112 to step 114, where it is determined that the noise is a candidate for brake squeal, and then in step 115 it is determined whether the vehicle speed is equal to or greater than a predetermined threshold. If the vehicle speed is equal to or greater than the threshold, the process proceeds to step 116, where it is finally determined that the noise is abnormal brake squeal. If the vehicle speed is less than the threshold, the process proceeds to step 122, where it is determined that the noise is normal driving noise, i.e., normal brake operation noise.

[0026] If the determination in step 111 is NO, the process proceeds to step 119, where it is determined which type of abnormal noise the abnormal noise candidate is. For example, it is determined whether it is the baffle plate contact noise candidate in step 109. If the determination is YES, the process proceeds from step 119 to step 113, where it is finally determined that the abnormal noise is baffle plate contact noise.

[0027] If the determination in step 119 is NO, the process proceeds to step 117, where it is determined whether the vehicle speed is equal to or greater than a predetermined threshold and the brake signal is ON. If the determination here is YES, the process proceeds to step 118, where it is finally determined that the brake squeal is abnormal. If the vehicle speed is less than the threshold or the brake signal is OFF, the process proceeds to step 123, where it is determined that the brake squeal is normal driving noise, i.e., normal braking noise.

[0028] The above processing is repeatedly executed at time intervals equal to the time interval of the short-time Fourier transform while acquiring a sound having a duration of, for example, 10 seconds. Note that the above-described processing of analyzing sound data and determining abnormal sounds can be performed in parallel with acquiring sound via a microphone. Alternatively, the processing may be performed after acquiring all of a series of sound data having a duration of, for example, 10 seconds.

[0029] As described above, according to the above embodiment, by acquiring sounds near the wheel section via a microphone during test runs in the completed vehicle inspection process, it is possible to detect and distinguish between two different types of abnormal noise, namely, baffle plate contact noise and brake squeal, thereby enabling efficient abnormal brake noise inspection. Furthermore, by determining abnormal noise taking into account the brake signal and vehicle speed, false detection of abnormal brake noise is reduced.

[0030] Next, a brake noise detection device according to a second embodiment will be described with reference to Figures 3 and 4. The following mainly describes the differences from the first embodiment. The brake noise detection device according to the second embodiment analyzes sound data at each time corresponding to a specific vehicle operating state related to brake noise during a series of test runs, and extracts sounds that are candidates for the corresponding type of brake noise.

[0031] As shown in the block diagram of Figure 3, the brake noise detection device of the second embodiment, like the first embodiment, includes a sound acquisition unit 10, a vehicle signal acquisition unit 11 (brake signal acquisition unit 11-1, acceleration signal acquisition unit 11-2, etc.), a vehicle state determination unit 12, and a brake noise determination unit 13, and further includes a model switching unit 20.

[0032] In the second embodiment, the abnormal brake noise judgment unit 13 sequentially extracts multiple types of candidate brake noises from a series of sound data and performs abnormal noise judgment using the corresponding abnormal noise detection model for each. The model switching unit 20 switches between abnormal noise detection models based on the vehicle signals acquired by the vehicle signal acquisition unit 11. In one embodiment, switching is performed between a first abnormal noise detection model suitable for detecting baffle plate contact noise, including rattle noise, and a second abnormal noise detection model suitable for detecting brake squeal, depending on the vehicle operating state determined by the ON / OFF status of the brake signal and the acceleration / deceleration of the vehicle.

[0033] 4 is a flowchart showing the processing flow of the brake noise detection device of the second embodiment. First, in step 101, acquisition of sound via a microphone is started, and at the same time, acquisition of vehicle signals is started in step 102. In the next step 103, the acceleration state of the vehicle is determined from changes in the vehicle speed signal or the accelerator pedal position signal.

[0034] Next, the process proceeds to step 201, where it is determined whether the vehicle is in an accelerating state. Specifically, if the vehicle speed change rate is equal to or greater than 0, or if the accelerator opening is equal to or greater than a predetermined threshold, it is determined that the vehicle is in an accelerating state, and the determination result in step 201 is YES. The process in step 201 corresponds to the model switching unit 20.

[0035] If the determination in step 201 is YES, the process proceeds from step 201 to step 104A, where a short-time Fourier transform (STFT) is performed on the acquired sound data to analyze the frequency characteristics. This allows for the acquisition of frequency-sound pressure level characteristics (e.g., a spectrum with frequency on the horizontal axis and sound pressure level on the vertical axis) for each short time period. In the next step 202, appropriate preprocessing is applied to the output signal of the short-time Fourier transform. For example, logarithmic processing is performed to ensure that even small frequency peaks are not overlooked, making it suitable for detecting baffle plate contact noise. Then, in step 105, as in the first embodiment, peaks in a relatively low, predetermined frequency band (e.g., 2000 to 3000 Hz) are detected to extract candidates for baffle plate contact noise, and in step 107, it is determined whether or not the number of peaks included is greater than a predetermined number (e.g., five). If the determination in step 107 is NO, the process proceeds to step 204, where it is finally determined that baffle plate contact noise has been detected. If the determination in step 107 is NO, the process returns to step 103 and repeats the process.

[0036] Therefore, while the vehicle is in an accelerating state, it is repeatedly determined whether the acquired sound includes the baffle plate contact sound.

[0037] On the other hand, if the determination in step 201 is NO, the process proceeds from step 201 to step 104B, where a short-time Fourier transform (STFT) is performed on the acquired sound data to analyze the frequency characteristics. Next, in step 106, as in the first embodiment, peaks in a relatively high frequency band (for example, 10 kHz or higher) are detected to extract candidates for brake squeal. Then, in step 108, it is determined whether there are fewer than a predetermined number of peaks (for example, five). If there are fewer than the predetermined number of peaks, the process proceeds to step 203, where it is determined whether the brake signal is ON. If the brake signal is ON, the process proceeds to step 205, where it is finally determined that the abnormal noise is brake squeal.

[0038] If the determination in step 108 is NO, or if the determination in step 203 is NO, the process returns to step 103 and the process is repeated.

[0039] In other words, when the vehicle changes from an accelerating state to a decelerating state (when the vehicle speed change rate becomes negative or when the accelerator opening becomes less than the threshold value), the first abnormal noise detection model for baffle plate contact noise is switched to the second abnormal noise detection model for brake squeal, and thereafter, it is repeatedly determined whether the acquired sound contains brake squeal.

[0040] For example, if it is assumed that brake squeal occurs in conjunction with braking, switching the abnormal noise detection model after the brake signal is turned ON may result in a delay in processing, which may prevent accurate detection of brake squeal that occurs in the early stages of braking.In many cases, a decrease in accelerator opening and a corresponding decrease in vehicle speed occur before the driver applies the brakes, so by starting brake squeal detection at the timing of detecting these, as in the above embodiment, brake squeal can be reliably detected.In other words, in the above embodiment, a reversal of vehicle speed or accelerator opening from an accelerating state to a decelerating state is recognized as a sign of brake application.

[0041] As described above, according to the second embodiment, the abnormal noise detection model is switched in response to changes in the vehicle operating conditions during test runs in the finished vehicle inspection process, and abnormal noise is detected in a manner that corresponds to the vehicle operating conditions, so that, similar to the first embodiment, it is possible to detect and distinguish between both baffle plate contact noise and brake squeal during a series of test runs, and brake abnormal noise inspection can be performed efficiently. Furthermore, by performing abnormal noise judgment based on specific vehicle operating conditions, false detection of abnormal brake noise is reduced.

[0042] Next, a brake noise detection device according to a third embodiment will be described with reference to Figures 5 and 6. In order to reduce false detections, the brake noise detection device according to the third embodiment does not analyze the sound or detect abnormal noise during periods when the sound pressure level of the original sound is low.

[0043] 5, the abnormal brake noise detection device of the third embodiment, like the first embodiment, includes a sound acquisition unit 10, a vehicle signal acquisition unit 11 (such as a brake signal acquisition unit 11-1 and an acceleration signal acquisition unit 11-2), a vehicle state determination unit 12, and an abnormal brake noise determination unit 13, and further includes a sound pressure monitoring unit 30. The illustrated example also includes a display unit 31 for displaying the inspection results.

[0044] The sound pressure monitoring unit 30 monitors the sound pressure level of the sound acquired by the sound acquisition unit 10 via the microphone, and if the sound pressure level is below a predetermined threshold, outputs an instruction to the brake abnormal noise judgment unit 13 to skip the above-mentioned processes such as the short-time Fourier transform process and peak extraction. The sound pressure level is judged at short time intervals in the short-time Fourier transform. That is, the sound data acquired via the microphone is divided into fixed time intervals, and the sound pressure level is monitored to see if it exceeds a predetermined threshold (e.g., 10 dB). The sound pressure level to be compared with the threshold can be any appropriate value, such as the average value for the divided period, the maximum value within the period, or the median value within the period. By ignoring sounds in sections with low sound pressure levels in this way, false detection is suppressed.

[0045] The display unit 31 is a display means for displaying the inspection results to workers in the vicinity of the inspection device, the inspector driving the vehicle, the manager, etc. For example, it is configured to include a liquid crystal display, an organic EL display, etc. Furthermore, if some kind of notification sound or voice is involved, it is configured to include a sound source, amplifier, speaker, etc. for generating and emitting the sound.

[0046] 6 is a flowchart showing the processing flow of the brake noise detection device of the third embodiment. As in the first embodiment, first, in step 101, acquisition of sound via a microphone is started. In the next step 301, the sound pressure level is calculated at predetermined short time intervals, and in step 302, it is determined whether the sound pressure level exceeds a predetermined threshold (for example, 10 dB). If the sound pressure level exceeds the predetermined threshold, the process proceeds to step 102.

[0047] The processing from step 102 onwards is the same as in the first embodiment described above, so a description thereof will be omitted.

[0048] If the sound pressure level is equal to or lower than the threshold value, the process does not proceed to step 102 and subsequent steps, but returns to step 101 to repeat the acquisition of sound.

[0049] Next, an example of the display on the display unit 31 will be described with reference to Fig. 7. In the display example of Fig. 7, for example, the magnitude of the sound pressure level in a relatively low predetermined frequency band (for example, 2000 to 3000 Hz) for detecting baffle plate contact noise and the magnitude of the sound pressure level in a relatively high frequency band (for example, 10 kHz or higher) for detecting brake squeal are shown by the size of a circle, and the circle is colored red, green, or black to indicate whether or not an abnormal noise is present (whether it is OK or NG).

[0050] The left half of the display screen shows the inspection results for the left front wheel, and the right half shows the inspection results for the right front wheel. At the top, the words "NG" and "OK" are displayed, respectively, as symbols 701 and 702. At the top, graphs 703 and 704 of the spectrum obtained as a result of the short-time Fourier transform are displayed. The contents of these graphs 703 and 704 change over time. A row of circles 705 and 706 labeled "BP contact noise" indicates the sound pressure level (e.g., three levels) in the 2000-3000 Hz range, which is the target for detecting baffle plate contact noise. Red indicates NG (baffle plate contact noise included) and green indicates OK (baffle plate contact noise not included). Circles currently undergoing evaluation are displayed in black. These circles are arranged in chronological order, with each corresponding to a separate short-time Fourier transform result. In the illustrated example, all of the baffle plate contact noises are OK (green) circles.

[0051] The rows of circles 707 and 708 at the bottom labeled "Brake Squeal" indicate the sound pressure level of, for example, 10 kHz or higher, which is the target for brake squeal detection, by the size of the circle (for example, in three stages), and also indicate sounds that are NG (including brake squeal) in red and sounds that are OK (not including brake squeal) in green. Circles that are currently being evaluated are black. These circles are arranged in chronological order, and each one corresponds to the result of the short-time Fourier transform at each individual time. In the illustrated example, two circles, 707a and 707b, are red (NG), and the others are green (OK). Also, in the illustrated example, the circles 707 and 708 related to brake squeal are displayed only while the brake signal is ON, and therefore only three circles are arranged.

[0052] Such displays on the display unit 31 allow inspectors to intuitively understand the level of sound pressure and whether the test result was OK. For example, the color of the circle changes from green to red during the test, allowing inspectors to easily recognize that an abnormal noise has occurred.

[0053] The graphic used to indicate the magnitude of the sound pressure level and the test results is not limited to the circle described above, but may be a polygon, star, or other suitable shape. Different graphics may be used for baffle plate contact noise and brake squeal.

[0054] While the present invention has been described above as an embodiment in which it is applied to the inspection of brake systems during the inspection process of completed vehicles, the present invention is not limited to this and can be applied in a variety of ways. For example, the present invention can easily inspect for abnormal brake noise during inspection and maintenance at a dealership or repair shop. It is also possible to inspect for abnormal brake noise while the vehicle is actually traveling on the road.

[0055] Furthermore, in the above-described embodiment, abnormal sound candidates are extracted by performing short-time Fourier transform processing and then finding peaks in a predetermined frequency band, but other well-known abnormal sound detection methods such as a deep auto-encoding Gaussian mixture model, which is one of the machine learning models for unsupervised anomaly detection, may also be used. [Explanation of symbols]

[0056] 10...Sound acquisition section 11...Vehicle signal acquisition unit 12...Vehicle condition determination unit 13...Brake noise detection unit 20...Model switching section 30...Sound pressure monitoring section 31...Display section

Claims

1. The sound generated around the wheels of the target vehicle while it is running is acquired and used as sound data. This sound data is analyzed to extract multiple types of brake noise candidates, Acquire information on the vehicle's driving state and determine whether the information corresponds to a plurality of specific vehicle driving states; determining the type of abnormal brake noise based on the vehicle driving state at the time when the noise that is a candidate for abnormal brake noise occurs; How to detect abnormal brake noise.

2. From a series of sound data, multiple types of sounds that are candidates for abnormal brake noise are extracted in parallel, determining the type of abnormal brake noise for each of the extracted plurality of candidates for abnormal brake noise based on whether the vehicle is in a specific driving state; The brake noise detection method according to claim 1 .

3. Analyzing the sound data for each time period corresponding to a plurality of specific vehicle operating states, and extracting sounds that are candidates for the corresponding type of abnormal brake noise. The brake noise detection method according to claim 1 .

4. The information on the vehicle's driving state includes information on brake operation and information on vehicle speed or accelerator opening. The brake noise detection method according to claim 1 .

5. The information on the vehicle's operating state includes information on the vehicle speed or accelerator pedal position, When the vehicle speed or accelerator opening changes from an accelerating state to a decelerating state, it is considered a sign of braking. The brake noise detection method according to claim 3 .

6. The types of brake noise include a first type of noise that occurs due to interference or vibration between a rotating member and a non-rotating member regardless of brake operation, and a second type of noise that occurs due to sliding of a friction member during brake operation. The brake noise detection method according to claim 1 .

7. Perform short-time Fourier transform processing of the sound data, If a number of peaks greater than a predetermined number are included in a relatively low predetermined frequency band, the noise is extracted as a candidate for a first allophone; If a relatively high predetermined frequency band contains a number of peaks that is less than a second predetermined number, the noise is extracted as a candidate for a second allophone. The brake noise detection method according to claim 6.

8. Monitor the sound pressure level of the captured sound, If the sound pressure level is below a certain level, the sound data will not be analyzed. The brake noise detection method according to claim 1 .

9. determining the sound pressure level in the predetermined relatively low frequency band and the sound pressure level in the predetermined relatively high frequency band, The display shows the sound pressure levels of each sound as figures arranged at regular time intervals. whether or not the first allophone and the second allophone are included in the sound at each time is indicated by the color of the graphic. The brake noise detection method according to claim 7.

10. a sound acquisition unit that acquires sounds generated near the wheels while the target vehicle is running and generates sound data; a vehicle state determination unit that acquires information on a vehicle driving state and determines whether the information corresponds to a plurality of specific vehicle driving states; an abnormal brake noise determination unit that analyzes the sound data to extract sounds that are candidates for a plurality of types of abnormal brake noise, and determines the type of abnormal brake noise based on the vehicle operating state at the time the candidate abnormal brake noise is generated; and A brake noise detection device comprising:

11. 10. A brake noise detection program that causes a computer system constituting a vehicle inspection device to execute the brake noise detection method according to claim 1.

Images