JP7559702B2 - Diagnostic device for diagnosing vehicle noise - Google Patents

Description

The technology disclosed in this specification relates to a diagnostic device that diagnoses sounds generated by a vehicle.

In general, it is desirable that, when a vehicle is running, no abnormal sounds, i.e. sounds that should not occur if the vehicle is properly manufactured, occur. In this regard, Patent Document 1 describes a diagnostic device that diagnoses sounds generated by a vehicle. This diagnostic device includes a microphone and a processing circuit connected to the microphone. The processing circuit executes a process of continuously recording the generated sounds and a process of Fourier transforming the recorded generated sounds. Based on this Fourier transform data, abnormal sounds are detected and identified by machine learning such as deep learning.

JP 2019-164107 A

In the above-mentioned diagnostic device, abnormal sounds are detected and identified by machine learning, and the process cannot be understood from the outside, which may lead to the so-called "black box problem." For example, the diagnosis results by machine learning are not necessarily complete, and may sometimes lead to misguided diagnosis results. In such cases, if the diagnosis results by machine learning differ from the expected results based on the user's own experience, the user may feel distrustful of the diagnostic device. In particular, abnormal sounds include, for example, abnormal sounds with constant frequency and magnitude (hereinafter referred to as "steady sounds") and abnormal sounds with at least one of the frequency and magnitude changing from moment to moment (hereinafter referred to as "fluctuating sounds"). The user can also perceive the change in frequency or magnitude of fluctuating sounds, and they often cause discomfort to the user compared to steady sounds. When such fluctuating sounds are occurring, if a diagnosis result that seems to be misidentified as a steady sound, contrary to the user's expectations, is shown, the user will feel strong distrust.

In consideration of the above situation, this specification provides technology that can reliably detect fluctuating sounds contained in vehicle generated sounds and more accurately diagnose the generated sounds.

The technology disclosed in this specification is embodied in a diagnostic device that diagnoses sounds generated by a vehicle. This diagnostic device includes a microphone and a processing circuit connected to the microphone. The processing circuit executes the following processes: recording the generated sound using the microphone for a predetermined time; repeatedly performing a Fourier transform on the generated sound recorded for the predetermined time at a predetermined cycle; counting the number of times that sound pressure rises across a predetermined threshold for each frequency in a series of Fourier transform data repeatedly created at the predetermined cycle; determining that the generated sound contains a fluctuating sound when the counted number is equal to or greater than a predetermined lower limit and equal to or less than an upper limit; and presenting actual data or recorded sound prepared in association with the fluctuating sound when it is determined that the generated sound contains the fluctuating sound.

In the above diagnostic device, while recording the generated sound of the vehicle, the Fourier transform of the recorded generated sound is repeatedly performed for a predetermined time to create a series of Fourier transform data. When the generated sound of the vehicle includes a fluctuating sound, in the series of Fourier transform data, a sequence of Fourier transform data in which the sound pressure exceeds the threshold and a sequence of Fourier transform data in which the sound pressure is below the threshold appear repeatedly at a relatively gentle frequency. On the other hand, when the generated sound of the vehicle includes a stationary sound, only Fourier transform data in which the sound pressure exceeds the threshold appears continuously, or a sequence of Fourier transform data in which the sound pressure exceeds the threshold and a sequence of Fourier transform data in which the sound pressure is below the threshold appear repeatedly at a frequency that is imperceptible to humans. For this reason, in the series of Fourier transform data, the number of times that the sound pressure rises across a predetermined threshold is counted, and when the counted number is equal to or greater than a predetermined lower limit and equal to or less than an upper limit, it can be determined that the generated sound includes a fluctuating sound. This makes it possible to reliably detect fluctuating sounds contained in sounds generated by the vehicle, and when fluctuating sounds are detected, it is possible to present performance data (e.g., data based on performance in another vehicle) or recorded sounds that are prepared in association with the fluctuating sounds.

1 is a schematic diagram showing the configuration of a diagnostic device 10. 5 is a flowchart illustrating an example of a diagnostic process executed by a processing circuit 16. 13 shows Fourier transform data in which the sound pressure does not exceed the threshold T at any frequency. 1 shows Fourier transform data where the sound pressure exceeds a threshold T at a certain frequency.

The diagnostic device 10 will be described with reference to the drawings. The diagnostic device 10 is a device that diagnoses sounds generated by a vehicle. The diagnostic device 10 is mounted on, for example, a mobile phone (e.g., a smartphone), a tablet computer, a personal digital assistant, etc. However, in another embodiment, the diagnostic device 10 may be mounted on a vehicle.

As shown in FIG. 1, the diagnostic device 10 includes a microphone 12 and an A/D converter 14. The microphone 12 is a device that collects ambient sounds. The A/D converter 14 can convert analog signals into digital signals. The A/D converter 14 is electrically connected to the microphone 12 and converts the sounds collected by the microphone 12 into digital signals. Therefore, when the microphone 12 is disposed in the vehicle cabin, the microphone 12 can collect sounds related to the vehicle that are perceived by passengers in the vehicle cabin, that is, sounds generated by the vehicle. Here, the sounds generated by the vehicle include sounds that are generated even if the vehicle is properly manufactured (i.e., normal sounds) and sounds that should not be generated if the vehicle is properly manufactured (i.e., abnormal sounds). The sounds generated by the vehicle collected by the microphone 12 are converted into audio data, which is a digital signal, by the A/D converter 14. The specific configuration of the microphone 12 is not particularly limited as long as it is capable of collecting ambient sounds.

As shown in FIG. 1, the diagnostic device 10 further includes a processing circuit 16. The processing circuit 16 includes a memory 18 and a processor 20, and executes diagnostic processing to diagnose sounds generated by the vehicle. The memory 18 is, for example, a non-volatile memory, and stores programs and data related to the diagnostic processing. The data stored in the memory 18 includes, for example, parameters such as a threshold value T used in the diagnostic processing, as well as performance data and recorded sounds that are presented to the user according to the diagnostic results. Although not limited to this, the programs and data stored in the memory 18 are installed in the diagnostic device 10 from a server on the Internet as one or more packaged applications.

The processing circuit 16 is communicatively connected to each of the microphone 12 and the A/D converter 14, and can monitor and control the operation of the microphone 12 and the A/D converter 14. For example, the processing circuit 16 can collect the sound generated by the vehicle using the microphone 12, and record (record) the audio data output from the A/D converter 14 in the memory 18. The processor 20 uses the programs and data stored in the memory 18 to perform various processes, described below, on the audio data recorded in the memory 18, thereby detecting the fluctuating sound contained in the sound generated by the vehicle.

As shown in FIG. 1, the diagnostic device 10 further includes a touch panel 22. The touch panel 22 is a display device that displays various information, and is also an operation device that accepts operations by the user. That is, the user can input various instructions to the diagnostic device 10 by operating the touch panel 22. The instruction input from the touch panel 22 is transmitted to the processing circuit 16. For example, when the user inputs an instruction to start execution of the diagnostic process to the touch panel 22, the processing circuit 16 starts a series of processes related to the diagnostic process using the microphone 12 and the A/D converter 14. Note that the touch panel 22 is an example of a user interface and is not particularly limited. In other embodiments, the diagnostic device 10 may include a display, a speaker, an operation button, etc. in other forms, or may accept instructions from the user by voice input using the microphone 12. In addition, the diagnostic process by the diagnostic device 10 is not limited to an instruction by the user, but may be instructed via communication from outside, or may be scheduled in advance.

Next, the diagnostic process executed by the processing circuit 16 will be described with reference to the flowchart shown in FIG. 2. As described above, the processing circuit 16 starts executing the diagnostic process, for example, in response to an instruction from a user. First, in step S10, the processing circuit 16 uses the microphone 12 to record the sound generated by the vehicle for a predetermined time. This predetermined time is not particularly limited, but can be, for example, 10 seconds. Next, in step S12, the processing circuit 16 repeatedly executes a Fourier transform on the generated sound recorded for the predetermined time at a predetermined cycle. This predetermined cycle is not particularly limited, but can be, for example, the system frequency of the processor 20. As an example, the system frequency in this embodiment is 22 kHz.

Through the above process, a series of Fourier transform data is created for the generated sound recorded over a predetermined time period. As shown in Figs. 3 and 4, in each Fourier transform data, the sound pressure for each frequency of the generated sound is described at each timing. Then, as shown in Fig. 4, if the sound pressure exceeds the threshold T at any frequency, it is determined that there is a high probability that the vehicle generated sound contains an abnormal sound. Here, the abnormal sound contained in the vehicle generated sound is classified into a stationary sound whose frequency and magnitude are constant, and a fluctuating sound whose frequency or magnitude at least changes from moment to moment. For example, if all of the series of Fourier transform data are data as shown in Fig. 4, that is, if the sound pressure exceeds the threshold T at a specific frequency, it can be determined that the vehicle generated sound contains a stationary sound. On the other hand, for example, if a series of Fourier transform data includes a series of Fourier transform data (see Fig. 4) whose sound pressure exceeds the threshold T and a series of Fourier transform data (see Fig. 3) whose sound pressure is below the threshold T, it can be determined that the vehicle generated sound contains a fluctuating sound.

The user can also perceive changes in frequency or volume of fluctuating sounds, and compared to steady sounds, fluctuating sounds often cause discomfort to the user. When such fluctuating sounds are generated, and a diagnosis result is presented that indicates that the sound is mistaken for a steady sound, contrary to the user's expectations, this creates a strong sense of distrust in the user. The diagnostic device 10 of this embodiment can reliably detect fluctuating sounds contained in sounds generated by the vehicle by further executing the processing described below on the series of Fourier transform data.

Returning to FIG. 2, in step S14, the processing circuit 16 counts the number of times N that the sound pressure rises across a predetermined threshold T for each frequency in a series of Fourier transform data repeatedly created over a predetermined period. If the vehicle has an engine, the low frequency range A shown in FIG. 3 and FIG. 4 may be excluded from the target to be processed. This is because the engine operation sound, which is a normal sound, is prominently detected in the low frequency range A. The frequency at which the engine operation sound is detected is determined by the number of cylinders and the engine speed. For example, when the speed of a six-cylinder engine is 6000 rpm, the frequency of the engine operation sound is 300 Hz, so that peak components originating from the engine operation sound are prominently detected at frequencies below 300 Hz. Therefore, in step S14, the low frequency range A in which peak components originating from the engine operation sound are detected is excluded, and the number of times N that the sound pressure rises across the predetermined threshold T in the other frequency ranges is counted. The predetermined threshold T may be a different value for each frequency, or may be constant regardless of frequency.

In step S16, the processing circuit 16 judges whether the number of times N counted in step S14 is 6 or more and 100 or less. YES in step S16, i.e., the number of times N counted in step S14 is 6 or more and 100 or less, means that in a series of Fourier transform data, a string of Fourier transform data in which the sound pressure exceeds the threshold T and a string of Fourier transform data in which the sound pressure is below the threshold T appear repeatedly with a relatively gentle frequency. The user can perceive such gentle fluctuations in sound pressure as fluctuating sound. Therefore, if YES in step S16, the processing circuit 16 judges that the sound generated by the vehicle contains fluctuating sound (step S18). Then, the processing circuit 16 presents the actual data or recorded sound prepared in association with the fluctuating sound (step S20), and ends the diagnosis process. Note that the actual data means various data related to the fluctuating sound detected in another vehicle (for example, the state when the fluctuating sound occurred, such as the driving speed and road surface condition, and the cause of the fluctuating sound, etc.).

If the answer is NO in step S16, the processing circuit 16 proceeds to step S22. In step S22, the processing circuit 16 determines whether the number of times N counted in step S14 is 0. YES in step S22, i.e., the number of times N counted in step S14 is 0, means that the series of Fourier transform data does not include a string of Fourier transform data in which the sound pressure exceeds the threshold value T. Therefore, if the answer is YES in step S22, the processing circuit 16 determines that the sounds generated by the vehicle do not include abnormal sounds (step S24) and ends the diagnosis process.

If step S22 is NO, the processing circuit 16 proceeds to step S26. In step S26, the processing circuit 16 determines whether the number of times N that the sound pressure exceeded the predetermined threshold T in step S14 is 1, and whether the sound pressure continues to exceed the predetermined threshold T over the predetermined recorded time (e.g., 10 seconds). YES in step S26, i.e., the number of times N counted in step S14 is 1, and the sound pressure continues to exceed the predetermined threshold T over the predetermined recorded time, means that the sound pressure at a specific frequency exceeds the threshold T for all of the series of Fourier transform data. Therefore, if step S26 is YES, the processing circuit 16 determines that the vehicle-generated sound includes a stationary sound (step S30) and ends the diagnosis process.

If the answer is NO in step S26, the processing circuit 16 proceeds to step S28. In step S28, the processing circuit 16 determines whether the number of times N counted in step S14 is greater than 100. YES in step S28, i.e., the number of times N counted in step S14 is greater than 100, means that in a series of Fourier transform data, a string of Fourier transform data whose sound pressure exceeds the threshold T and a string of Fourier transform data whose sound pressure is below the threshold T appear repeatedly at a frequency that is imperceptible to humans. In this way, when a string of Fourier transform data whose sound pressure exceeds the threshold T and a string of Fourier transform data whose sound pressure is below the threshold T appear repeatedly at a frequency that is imperceptible to humans, the human ear recognizes it as a stationary sound. Therefore, if the answer is YES in step S28, the processing circuit 16 determines that the sound generated by the vehicle includes a stationary sound (step S30) and ends the diagnosis process.

If the answer is NO in step S28, the diagnostic device 10 returns to the process of step S10. In other words, if the number of times N counted in step S14 is 1 or more and 5 or less, the diagnostic device 10 restarts the diagnostic process by executing the above series of processes again, since it is highly likely that the recorded audio data contains noise sounds other than sounds generated by the vehicle.

In the case where a string of Fourier transform data in which the sound pressure exceeds the threshold T and a string of Fourier transform data in which the sound pressure is below the threshold T appear repeatedly, the specific number of times N for distinguishing whether or not the repetition occurs at a frequency imperceptible to humans is not particularly limited. That is, the upper limit of the number of times N in step S16 and the lower limit of the number of times N in step S28 are not particularly limited. In addition, the specific number of times N for distinguishing whether or not the recorded audio data contains noise is also not particularly limited. That is, the lower limit of the number of times N in step S16 is also not particularly limited.

As described above, the diagnostic method implemented by the diagnostic device 10 of this embodiment involves recording the generated sound of the vehicle while repeatedly performing a Fourier transform of the recorded generated sound over a predetermined time to create a series of Fourier transform data. When the generated sound of the vehicle includes a fluctuating sound, in the series of Fourier transform data, a sequence of Fourier transform data in which the sound pressure exceeds the threshold T and a sequence of Fourier transform data in which the sound pressure is below the threshold T appear repeatedly at a relatively gentle frequency. On the other hand, when the generated sound of the vehicle includes a stationary sound, only Fourier transform data in which the sound pressure exceeds the threshold T appear consecutively, or a sequence of Fourier transform data in which the sound pressure exceeds the threshold T and a sequence of Fourier transform data in which the sound pressure is below the threshold T appear repeatedly at a frequency that is imperceptible to humans. For this reason, in the series of Fourier transform data, the number N of times that the sound pressure rises across a predetermined threshold T is counted, and when the counted number N is equal to or greater than a predetermined lower limit value and equal to or less than an upper limit value, it can be determined that the generated sound includes a fluctuating sound. This makes it possible to reliably detect fluctuating sounds contained in sounds generated by the vehicle, and when fluctuating sounds are detected, it is possible to present performance data (e.g., data based on performance in another vehicle) or recorded sounds that are prepared in association with the fluctuating sounds.

In the above embodiment, the microphone 12 does not necessarily have to be provided in the diagnostic device 10. That is, in another embodiment, the microphone 12 may be provided separately from the diagnostic device 10. In this case, the microphone 12 may be disposed inside the vehicle, and the diagnostic device 10 may be disposed at a location away from the vehicle. By connecting the microphone 12 and the diagnostic device 10 so that they can communicate with each other, the processing circuit 16 can execute the same diagnostic processing as described above.

Although several specific examples have been described in detail above, these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and variations of the specific examples given above. The technical elements described in this specification or drawings exhibit technical usefulness either alone or in combination.

10: Diagnostic device 12: Microphone 14: A/D converter 16: Processing circuit 18: Memory 20: Processor 22: Touch panel A: Range N: Number of times T: Threshold

Claims (1)

A diagnostic device for diagnosing a sound generated by a vehicle, comprising:
A microphone and a processing circuit connected to the microphone,
The processing circuitry includes:
A process of recording the generated sound for a predetermined period of time using the microphone;
A process of repeatedly performing a Fourier transform on the generated sound recorded over the predetermined time period at a predetermined cycle;
A process of counting the number of times that the sound pressure rises above a predetermined threshold for each frequency in the series of Fourier transform data repeatedly generated at the predetermined period;
A process of determining that the generated sound includes a fluctuating sound when the counted number of times is equal to or greater than a predetermined lower limit value and equal to or less than an upper limit value;
When it is determined that the generated sound includes the fluctuating sound, a process of presenting performance data or a recorded sound prepared in association with the fluctuating sound is executed.
Diagnostic equipment.

Images