Abnormal noise diagnosis system

Abstract

To improve the accuracy of abnormal noise diagnosis in an abnormal noise diagnosis system.SOLUTION: The abnormal noise diagnosis system includes an inquiry information acquisition unit that acquires inquiry information about abnormal noise generated in a vehicle, a sound acquisition unit that acquires time series data of sounds emitted from the vehicle, a vehicle status acquisition unit that acquires vehicle status data that indicates the status of the vehicle in synchronization with the time series data of sounds, an arithmetic processing unit that obtains a spectrogram showing the relation between time, frequency, and sound pressure from the time series data of sounds, a pre-diagnosis processing unit that obtains an estimated time zone of occurrence of the abnormal noise on the basis of the inquiry information and the vehicle status data and assigns a weight indicating the possibility of abnormal noise occurring at each of multiple time zones within a diagnosis range of the spectrogram selected by a user, on the basis of the estimated time zone of occurrence, and a diagnosis unit that diagnoses the cause of the abnormal noise that occurred in the vehicle on the basis of the time series data of sounds, the diagnosis range, and the weights.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present disclosure relates to an abnormal sound diagnosis system that diagnoses abnormal sounds generated in a vehicle. [Background technology]

[0002] Conventionally, an abnormal sound identification device including a calculation device that can access a trained model of artificial intelligence and an output device is known (see, for example, Patent Document 1). The calculation device of this abnormal sound identification device identifies frequency-time data indicating a time change in the frequency spectrum of a sound recorded in a vehicle, and displays the identified frequency-time data on a monitor. In addition, a user of the abnormal sound identification device selects a first time range in which an abnormal sound is considered to be included and a second time range in which an abnormal sound is considered not to be included from the frequency-time data on the monitor. Next, the calculation device inputs the identified frequency-time data into the trained model. The trained model identifies one or more abnormal sounds occurring within the first time range as a first abnormal sound, and identifies one or more abnormal sounds occurring within the second time range as a second abnormal sound. Here, it is highly likely that no abnormal sound has occurred in the second time range, and that the second abnormal sound is an abnormal sound that has not been recognized by a person. Therefore, a first abnormal sound that matches the type of the second abnormal sound is highly likely to be an abnormal sound that has not been recognized by a person. Based on this, the computing device displays on the monitor the type of the first abnormal sound that does not match the type of the second abnormal sound, thereby making it possible to identify the type of abnormal sound contained in the sound recorded in the vehicle. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2023-012335 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the above-mentioned conventional abnormal sound identifying device, it is not necessarily easy to appropriately select the first time range in which an abnormal sound is considered to be included and the second time range in which an abnormal sound is considered not to be included. Furthermore, a user who has little experience using an abnormal sound identifying device tends to select a wide first time range and a narrow second time range. For this reason, in the above-mentioned abnormal sound identifying device, a sound that is not actually recognized as an abnormal sound may be erroneously determined as an abnormal sound by the trained model, which may reduce the accuracy of identifying the abnormal sound.

[0005] Therefore, a primary object of the present disclosure is to further improve the accuracy of abnormal sound diagnosis in an abnormal sound diagnosis system. [Means for solving the problem]

[0006] The abnormal sound diagnosis system disclosed herein diagnoses an abnormal sound generated in a vehicle, and includes an inquiry information acquisition unit, a sound acquisition unit, a vehicle state acquisition unit, a calculation processing unit, a pre-diagnosis processing unit, and a diagnosis unit. The inquiry information acquisition unit acquires inquiry information related to an abnormal sound generated in the vehicle. The sound acquisition unit acquires time series data of a sound emitted from the vehicle, and the vehicle state acquisition unit acquires vehicle state data indicating the state of the vehicle in synchronization with the time series data of the sound. The calculation processing unit acquires a spectrogram indicating the relationship between time, frequency, and sound pressure from the time series data of the sound. The pre-diagnosis processing unit acquires an estimated time period during which the abnormal sound occurs based on the inquiry information and the vehicle state data, and assigns a weight indicating the possibility that the abnormal sound has occurred to each of a plurality of time periods in the diagnosis range of the spectrogram selected by the user based on the estimated time period. The diagnosis unit diagnoses the cause of the abnormal sound generated in the vehicle based on the time series data of the sound, the diagnosis range, and the weight.

[0007] As a result, even if the user of the abnormal sound diagnosis system sets the time range of the spectrogram diagnosis range to be broad on the safe side, the diagnostic range of the spectrogram can be divided into multiple time periods based on the estimated occurrence time period and the likelihood of abnormal sound occurrence can be evaluated for each of the multiple time periods before the diagnosis unit diagnoses the abnormal sound. As a result, the abnormal sound diagnosis system of the present disclosure can further improve the accuracy of abnormal sound diagnosis by the diagnosis unit. [Brief description of the drawings]

[0008] [Figure 1] 1 is a schematic configuration diagram showing an abnormal sound diagnosis system according to the present disclosure. [Diagram 2] 4 is a flowchart showing a series of processes executed in a terminal constituting the abnormal sound diagnosis system of the present disclosure. [Diagram 3] FIG. 2 is an explanatory diagram showing an abnormal sound diagnosis procedure in the abnormal sound diagnosis system of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Next, an embodiment of the present disclosure will be described with reference to the drawings.

[0010] 1 is a schematic diagram showing an abnormal sound diagnosis system 1 according to the present disclosure. The abnormal sound diagnosis system 1 shown in the figure is for diagnosing the cause of abnormal sound generated in a vehicle V, such as a vehicle equipped with only an engine as a drive source, a hybrid vehicle (HEV, PHEV), or an electric vehicle (BEV, FCEV), and includes a mobile terminal 10 and a server 20 as a diagnostic device capable of exchanging information with the mobile terminal 10 through communication.

[0011] The mobile terminal 10 is used by a user of the abnormal sound diagnosis system 1, who is, for example, an operator at a vehicle dealership or a maintenance shop, when responding to a user (owner) of the vehicle V in which an abnormal sound has occurred, or when performing a reproduction test in which the vehicle V is driven (operated) on a roadway or a test bench to reproduce the abnormal sound. In this embodiment, the mobile terminal 10 is a smartphone including an SoC, a ROM, a RAM, a GPS module (location information acquisition unit), an auxiliary storage device (flash memory) M, a display unit 11, a communication module 12, a microphone (not shown), and the like. An abnormal sound diagnosis support application (program) is installed in the mobile terminal 10. As shown in FIG. 1, the mobile terminal 10 includes a medical interview information acquisition unit 13, a sound acquisition unit 14, a vehicle state acquisition unit 15, a calculation processing unit 16, a diagnosis preprocessing unit 17, and a display control unit 18, which are constructed by cooperation between the abnormal sound diagnosis support application (software) and hardware such as the SoC of the mobile terminal 10.

[0012] The display unit 11 of the mobile terminal 10 includes a touch panel type liquid crystal panel or an organic EL panel. The communication module 12 exchanges various information with the electronic control device of the vehicle V via short-distance wireless communication or a cable (dongle), and can exchange various information with the server 20 via a network such as the Internet. The medical inquiry information acquisition unit 13 acquires information indicating the state of the vehicle V when an abnormal sound occurs (hereinafter referred to as "medical inquiry information") provided by the user of the vehicle V via the display unit 11. The medical inquiry information includes vehicle type information, requests, date and time of occurrence, frequency of occurrence, type of sound, physical quantities that change when the vehicle V is traveling such as vehicle speed, driving state (driving state) of the vehicle V, warm-up effects in an engine-equipped vehicle, selection items selected by the driver while driving the vehicle V, driving environment information of the vehicle V, and the like.

[0013] The sound acquisition unit 14 acquires time axis data of sound (sound pressure) when the reproduction test is performed by the operator. The vehicle state acquisition unit 15 acquires information indicating the state of the vehicle V (hereinafter referred to as "vehicle state data") in synchronization with the acquisition of the time axis data of sound by the sound acquisition unit 14 when the reproduction test is performed. The vehicle state data includes a plurality of physical quantities (vehicle speed, engine speed, motor speed, etc.) corresponding to the items of the medical interview information. The calculation processing unit 16 executes an analysis process of the time axis data of sound acquired by the sound acquisition unit 14. The pre-diagnosis processing unit 17 executes narrowing down of the analysis results by the calculation processing unit 16 according to the selection of the operator, etc. The display control unit 18 controls the display unit 11. The medical interview information, the time axis data of sound, and the vehicle state data are all stored in the auxiliary storage device M.

[0014] The server 20 of the abnormal sound diagnosis system 1 is a computer (information processing device) including a CPU, ROM, RAM, input / output devices, a communication module, etc., and is installed and managed, for example, by an automobile manufacturer that manufactures the vehicle V. The server 20 is configured with an abnormal sound diagnosis unit 21 that diagnoses abnormal sounds generated in the vehicle V by cooperation between hardware such as a CPU and a pre-installed abnormal sound diagnosis application. The abnormal sound diagnosis unit 21 includes a neural network (convolutional neural network) configured by supervised learning (machine learning) so as to diagnose the cause of the abnormal sound generated in the vehicle V and the part that is the source of the abnormal sound based on the interview information and sound time axis data acquired by the mobile terminal 10. Furthermore, in the server 20, when it is found that a new abnormal sound has occurred in the vehicle V, re-learning is performed for the abnormal sound diagnosis unit 21 using the sound time axis data acquired for the new abnormal sound and the contents of each item of the interview information as teacher data.

[0015] Furthermore, the server 20 includes a storage device 22 that stores a database that stores information on a plurality of abnormal sounds that are known to occur in a vehicle for each vehicle model. The database stores information such as time axis data of the sound, the cause of the abnormal sound, the part that is the source of the sound, the contents of the medical interview information provided by the user, etc., and measures to eliminate the abnormal sound, for each of the plurality of abnormal sounds, by linking them to the information. The server 20 also updates the database based on information acquired from many vehicles including the vehicle V, and information on newly discovered abnormal sounds transmitted from automobile manufacturers (developers, etc.), vehicle dealers, repair shops, etc.

[0016] Next, the abnormal sound diagnosis procedure performed by the abnormal sound diagnosis system 1 will be described.

[0017] When a worker at a vehicle dealership or a repair shop is requested by a user of a vehicle V to eliminate an abnormal sound, the worker listens to the inquiry information from the user and inputs it into the mobile terminal 10, and then performs a reproduction test to obtain information necessary for diagnosing the abnormal sound. When performing the reproduction test, the worker connects the mobile terminal 10 to an electronic control device of the vehicle V, and places or fixes the mobile terminal 10 or an external microphone connected to the mobile terminal 10 in an appropriate position on the vehicle V. The worker also starts an abnormal sound diagnosis support application and turns on the start switch of the vehicle V. Accordingly, the mobile terminal 10 obtains information such as the vehicle identification number or chassis number of the vehicle V from the electronic control device. Furthermore, the worker taps a recording start button displayed on the display unit 11 and drives (operates) the vehicle V on a roadway or a test stand to reproduce the driving state in which the abnormal sound occurred based on the inquiry information from the user of the vehicle V.

[0018] While the vehicle V is running (operating), the sound acquisition unit 14 of the mobile terminal 10 acquires time axis data of the sound emitted from the vehicle V at predetermined time intervals (micro time) and stores the data in the auxiliary storage device M, and the vehicle state acquisition unit 15 acquires vehicle state data specified by the worker in response to medical interview information from the user of the vehicle V at predetermined time intervals (micro time) from the electronic control device of the vehicle V in synchronization with the acquisition of the sound time axis data by the sound acquisition unit 14 and stores the vehicle state data in the auxiliary storage device M. Then, when the worker taps the recording stop button displayed on the display unit 11, the acquisition of the sound time axis data and vehicle state data is completed, and the mobile terminal 10 executes a series of processes shown in FIG.

[0019] 2, after the reproduction test is completed, the arithmetic processing unit 16 of the mobile terminal 10 acquires time axis data of the sound from the auxiliary storage device M (step S100), and performs STFT (Short-Time Fourier Transform) on the acquired time axis data of the sound to acquire a spectrogram (acoustic spectrogram) showing the relationship between time, frequency, and sound pressure (step S110). The display control unit 18 of the mobile terminal 10 causes the display unit 11 to display the spectrogram acquired by the arithmetic processing unit 16 (step S120). In this embodiment, the spectrogram is a color map in which the horizontal axis is the time axis and the vertical axis is the frequency axis, and the sound pressure level is color-coded to show the relationship between time and sound pressure level for each frequency.

[0020] When the spectrogram is displayed on the display unit 11 of the mobile terminal 10, the operator selects (specifies) on the display unit 11 a range of the spectrogram that should be diagnosed (analyzed) by the abnormal sound diagnosis unit 21 (server 20) (hereinafter referred to as a "diagnosis range"). The diagnosis range specifies a time range and a frequency range that should be diagnosed (analyzed) by the abnormal sound diagnosis unit 21 (server 20). In response to the operator's operation on the screen, the pre-diagnosis processing unit 17 obtains the diagnosis range selected by the operator and instructs the display control unit 18 to display the diagnosis range on the display unit 11 (step S130). In step S130, the display control unit 18 displays the spectrogram on the display unit 11 so that only the diagnosis range selected by the operator is visible, for example by graying out the parts not selected by the operator.

[0021] Next, the pre-diagnosis processing unit 17 reads out the interview information acquired prior to the implementation of the reproduction test from the auxiliary storage device M (step S140), and determines whether the read interview information includes the "driving state" of the vehicle V, such as "acceleration", "deceleration", "constant speed driving", "slow driving", "turning", "braking", "engine running", and "EV driving" (step S150). If the interview information includes the driving state of the vehicle V (step S150: YES), the pre-diagnosis processing unit 17 reads out vehicle state data within the diagnosis range corresponding to the driving state from the auxiliary storage device M (step S160). The vehicle state data corresponding to the driving state is, for example, "vehicle speed" when the "driving state" is "acceleration," "deceleration," "constant speed driving," "slow driving," etc.; it is, for example, "steering angle" when the "driving state" is "turning," it is, for example, "brake pedal stroke" when the "driving state" is "braking," it is, for example, "engine rotation speed" when the "driving state" is "engine operating," and it is, for example, "motor rotation speed" when the "driving state" is "EV driving."

[0022] If vehicle condition data corresponding to the driving state is acquired in step S160 (step S170: YES), the pre-diagnosis processing unit 17 acquires an estimated occurrence time period of the abnormal noise based on the vehicle condition data corresponding to the driving state of the medical interview information (step S180). In step S180, if the "driving state" is, for example, "acceleration", the pre-diagnosis processing unit 17 acquires, as the estimated occurrence time period, a time period during which the vehicle V is accelerating, which is obtained based on the "vehicle speed", which is the vehicle condition data corresponding to the driving state. Furthermore, the pre-diagnosis processing unit 17 determines whether the estimated occurrence time period is within the time range of the diagnosis range selected by the operator (step S190).

[0023] If the estimated occurrence time zone is less than the time range of the diagnosis range (step S190: YES), the pre-diagnosis processing unit 17 assigns a weight indicating the possibility that an abnormal sound has occurred to each of the multiple time zones in the diagnosis range selected by the operator based on the estimated occurrence time zone (step S200). For example, as shown in FIG. 3, if the time range of the diagnosis range is divided into a time zone A that corresponds to (matches) the estimated occurrence time zone and a remaining time zone B ("driving state" = "deceleration") other than the estimated occurrence time zone, the pre-diagnosis processing unit 17 assigns, for example, a weight of "1" to the time zone A that corresponds to the estimated occurrence time zone as a weight indicating the possibility that an abnormal sound has occurred, and assigns a weight of "0" to the remaining time zone B. It goes without saying that the number of multiple time zones in the diagnosis range may be three.

[0024] After the process of step S200, when the worker taps an information transmission button displayed on the display unit 11, information necessary for diagnosing the abnormal sound is transmitted from the communication module 12 of the mobile terminal 10 to the server 20 (step S210). In this embodiment, the information transmitted from the mobile terminal 10 to the server 20 includes at least time axis data of the sound, medical interview information, and diagnostic range information including the time range of the diagnostic range, the frequency range, and weights for each of a plurality of time periods. When the information necessary for diagnosing the abnormal sound is transmitted from the mobile terminal 10 to the server 20, the abnormal sound diagnosis unit 21 of the server 20 diagnoses the cause of the abnormal sound generated in the vehicle V based on the information provided from the mobile terminal 10, and transmits the diagnosis result to the mobile terminal 10.

[0025] As shown in FIG. 3, the abnormal sound diagnosis unit 21 derives the abnormal sound name and its occurrence probability for, for example, a time period (time period A) corresponding to the estimated occurrence time period and a remaining time period (time period B) other than the estimated occurrence time period, and then generates a final diagnosis result by considering the weight for each time period. That is, in the example of FIG. 3, the diagnosis result for the time period with the weight "0" is excluded from the final diagnosis result. Also, the diagnosis for the time period with the weight "0" may be omitted. This makes it possible to reduce the diagnostic load on the abnormal sound diagnosis unit 21. Furthermore, the diagnosis result includes the cause of the abnormal sound generated in the vehicle V, the part that is the source of the abnormal sound, and the measures to eliminate the abnormal sound read from the storage device 22. Then, when the diagnosis result from the server 20 is received by the mobile terminal 10, the diagnosis result is displayed on the display unit 11 (step S220), and a series of processes executed by the mobile terminal 10 when diagnosing the abnormal sound is completed. By executing the process shown in FIG. 2, the worker can accurately explain the diagnosis result from the server 20 to the user of the vehicle V, etc., and promptly proceed with the abnormal sound countermeasure.

[0026] 2, if it is determined in step S150 that the medical interview information does not include the "driving state" of the vehicle V (step S150: NO), if vehicle state data corresponding to the driving state is not acquired in step S160 (step S170: NO), or if the estimated occurrence time period is equal to or longer than the time range of the diagnosis range selected by the operator (step S190: NO), the pre-diagnosis processing unit 17 assigns a weight of "1" indicating the possibility that an abnormal sound has occurred in the entire diagnosis range (step S205). In this case, after the processing of step S205, the processing of steps S210-S220 is executed as shown in FIG. 2, and the series of processing executed by the mobile terminal 10 is terminated.

[0027] As described above, the pre-diagnosis processing unit 17 of the abnormal sound diagnosis system 1 acquires the estimated occurrence time period of the abnormal sound based on the interview information and the vehicle condition data (S140-S180), and assigns a weight indicating the possibility that the abnormal sound has occurred to each of the multiple time periods in the diagnostic range of the spectrogram selected by the operator based on the estimated occurrence time period (S190: YES, S200). The abnormal sound diagnosis unit 21 of the abnormal sound diagnosis system 1 diagnoses the cause of the abnormal sound generated in the vehicle V based on the sound time series data, the diagnostic range, and the weight for each of the multiple time periods in the diagnostic range. As a result, even if the user of the abnormal sound diagnosis system 1 sets the time range of the diagnostic range of the spectrogram to a wide time range on the safe side, it is possible to divide the diagnostic range of the spectrogram into multiple time periods based on the estimated occurrence time period and evaluate the possibility of the abnormal sound having occurred for each of the multiple time periods before the abnormal sound diagnosis unit 21 diagnoses the abnormal sound. As a result, the abnormal sound diagnosis system 1 can further improve the accuracy of abnormal sound diagnosis by the abnormal sound diagnosis unit 21.

[0028] The values ​​set as the weights are not limited to "1" and "0". In other words, the values ​​set as the weights may be, for example, mutually different positive integers indicating that the larger the value, the more likely the abnormal sound has occurred. In this case, a diagnosis result may be generated taking into account the weight for each of a plurality of time periods in the diagnosis range. The abnormal sound diagnosis support application installed in the mobile terminal 10 may be installed in a tablet terminal, a laptop or desktop personal computer, or the like, and the tablet terminal or the like may be used instead of the mobile terminal 10. Furthermore, the abnormal sound diagnosis system 1 may be configured by a single information processing device such as a personal computer. [Industrial Applicability]

[0029] The invention disclosed herein is extremely useful for diagnosing abnormal noises occurring in a vehicle. [Explanation of symbols]

[0030] 1 abnormal sound diagnosis system, 10 mobile terminal, 13 interview information acquisition unit, 14 sound acquisition unit, 15 vehicle state acquisition unit, 16 calculation processing unit, 17 diagnosis pre-processing unit, 20 server, 21 abnormal sound diagnosis unit, V vehicle.

Claims

[Claim 1] An abnormal sound diagnosis system for diagnosing an abnormal sound generated in a vehicle, comprising: an inquiry information acquisition unit that acquires inquiry information regarding the abnormal noise occurring in the vehicle; A sound acquisition unit that acquires time series data of sounds emitted from the vehicle; a vehicle state acquisition unit that acquires vehicle state data indicating a state of the vehicle in synchronization with the time-series data of the sound; A processor that acquires a spectrogram showing the relationship between time, frequency, and sound pressure from the time-series data of the sound; a pre-diagnosis processing unit that acquires an estimated occurrence time period of the abnormal noise based on the interview information and the vehicle condition data, and assigns a weight indicating the possibility that the abnormal noise has occurred to each of a plurality of time periods in a diagnostic range of the spectrogram selected by a user based on the estimated occurrence time period; a diagnosis unit that diagnoses a cause of the abnormal sound generated in the vehicle based on the time-series data of the sound, the diagnosis range, and the weight; An abnormal sound diagnosis system comprising:

Images