CN-117194888-B - Abnormal sound generation factor determining method and abnormal sound generation factor determining device

CN117194888BCN 117194888 BCN117194888 BCN 117194888BCN-117194888-B

Abstract

The disclosure provides an abnormal sound occurrence factor determining method and an abnormal sound occurrence factor determining device. In the characteristic correction processing, the audio signal acquired by the audio signal acquisition processing is corrected based on the acquired model information, so that the frequency characteristic of the acquired audio signal approaches the frequency characteristic of the learning audio signal. In the variable obtaining process, the corrected audio signal is input to the map, and the variable output from the map is obtained. In the factor determination process, the occurrence factor of the sound perceived by the microphone is determined based on the variable acquired by the variable acquisition process (fig. 5).

Inventors

TAKEUCHI SHINICHI
TIAN DUANCHUN
Dawa Guangfu
Zezhu Zhenwu
MURAKAMI RYO
LIU CHUANLIANG

Assignees

丰田自动车株式会社

Dates

Publication Date: 20260512
Application Date: 20230605
Priority Date: 20220607

Claims (9)

1. A method for determining occurrence factors of abnormal sounds, which determines occurrence factors of abnormal sounds, wherein, The abnormal sound occurrence factor determination method comprises the following steps: A storage circuit of an analysis device stores map data defining a map that receives as input a sound signal that is a signal related to a sound perceived by a microphone and outputs a variable related to a factor of occurrence of a sound in a vehicle, the map being obtained by performing machine learning, the sound signal input to the map being a learning sound signal when the machine learning is performed on the map, the microphone that perceives a sound represented by the learning sound signal being a learning microphone; executing, by an execution circuit of the analysis device, a sound signal acquisition process of acquiring the sound signal related to the sound perceived by the microphone; Obtaining model information by the execution circuit, wherein the model information is information related to a model of the microphone; when the acquired model information is different from the model information of the learning microphone, performing, by the execution circuit, a characteristic correction process of correcting the sound signal acquired by the sound signal acquisition process based on the acquired model information so that the frequency characteristic of the sound signal approaches the frequency characteristic of the learning sound signal; Executing, by the execution circuit, a variable acquisition process of acquiring a variable output from the map by inputting the sound signal acquired in the sound signal acquisition process to the map when the acquired model information is the same as the model information of the learning microphone, and acquiring a variable output from the map by inputting the sound signal corrected in the characteristic correction process to the map when the acquired model information is different from the model information of the learning microphone, and The execution circuit executes a factor determination process in which a factor of occurrence of sound perceived by the microphone is determined based on the variable acquired in the variable acquisition process.
2. The abnormal sound occurrence factor determination method according to claim 1, wherein, The storage circuit stores a plurality of the model information, The plurality of model information includes 1 st model information and 2 nd model information, The abnormal sound occurrence factor determining method further comprises the following steps: executing, by the execution circuit, a1 st characteristic correction process as the characteristic correction process in the case where the model information obtained is the 1 st model information, and When the obtained model information is the model 2 information, the execution circuit executes the characteristic 2 correction process as the characteristic correction process, In the 1 st characteristic correction process, the frequency characteristic of the audio signal is brought close to the frequency characteristic of the learning audio signal by correcting the acquired audio signal based on the 1 st model information, In the 2 nd characteristic correction process, the frequency characteristic of the audio signal is made close to the frequency characteristic of the learning audio signal by correcting the acquired audio signal based on the 2 nd model information.
3. The abnormal sound occurrence factor determination method according to claim 2, wherein, The abnormal sound occurrence factor determining method further comprises the following steps: in the case where the acquired model information is not stored in the storage circuit, Executing, by the execution circuit, the 1 st characteristic correction process and the 2 nd characteristic correction process; Inputting the sound signal corrected by the 1 st characteristic correction process to the map by the execution circuit, thereby obtaining the variable output from the map as a1 st output variable; Inputting the sound signal corrected by the 2 nd characteristic correction process to the map by the execution circuit, thereby obtaining the variable output from the map as a2 nd output variable; inputting the sound signal obtained by the sound signal obtaining process into the map by the execution circuit, thereby obtaining the variable outputted from the map as a 3 rd output variable, and A factor selection process is performed by the execution circuit, in which the sound generation factor is selected from among the sound generation factor based on the 1 st output variable, the sound generation factor based on the 2 nd output variable, and the sound generation factor based on the 3 rd output variable.
4. The abnormal sound occurrence factor determination method according to any one of claims 1 to 3, wherein, The factor determination process is the 1 st factor determination process, When the model of the microphone shown by the acquired model information is the same as the model of the learning microphone, The abnormal sound occurrence factor determining method further comprises the following steps: A reference variable obtaining process of obtaining a variable outputted from the map as a reference variable by inputting the audio signal obtained by the audio signal obtaining process into the map, and The execution circuit executes a2 nd factor determination process of determining a factor of occurrence of the sound perceived by the microphone based on the reference variable acquired by the reference variable acquisition process.
5. The abnormal sound occurrence factor determination method according to any one of claims 1 to 3, wherein, The execution circuit includes a1 st execution circuit provided in the vehicle or in a portable terminal held by an occupant of the vehicle, and a2 nd execution circuit provided outside the vehicle, The characteristic correction process, the variable acquisition process, and the factor determination process are executed by the 2 nd execution circuit.
6. An abnormal sound occurrence factor determination method includes: A storage circuit of an analysis device stores map data defining a map that receives a sound signal, which is a signal related to a sound perceived by a microphone, as an input, and outputs a variable related to a factor of occurrence of a sound in a vehicle, the map being obtained by performing machine learning, the sound signal input to the map being a learning sound signal when the machine learning is performed on the map, the microphone that perceives a sound represented by the learning sound signal being a learning microphone; executing, by an execution circuit of the analysis device, a sound signal acquisition process of acquiring the sound signal related to the sound perceived by the microphone; Obtaining model information by the execution circuit, wherein the model information is information related to a model of the microphone; A1 st characteristic correction process of correcting the frequency characteristic of the sound signal acquired by the sound signal acquisition process is executed by the execution circuit, and in the case where the model information of the microphone is 1 st model information different from the model information of the learning microphone, the 1 st characteristic correction process can bring the frequency characteristic of the sound signal close to the frequency characteristic of the learning sound signal; a2 nd characteristic correction process of correcting the frequency characteristic of the sound signal acquired by the sound signal acquisition process is executed by the execution circuit, and in the case where the model information of the microphone is 2 nd model information different from the model information of the learning microphone, the 2 nd characteristic correction process can bring the frequency characteristic of the sound signal close to the frequency characteristic of the learning sound signal; Executing, by the execution circuit, a variable obtaining process in which a variable output from the map is obtained as a1 st output variable by inputting the sound signal corrected by the 1 st characteristic correction process to the map, a variable output from the map is obtained as a2 nd output variable by inputting the sound signal corrected by the 2 nd characteristic correction process to the map, and a variable output from the map is obtained as a 3 rd output variable by inputting the sound signal obtained by the sound signal obtaining process to the map, and A factor selection process is performed by the execution circuit, in which the sound generation factor is selected from among the sound generation factor based on the 1 st output variable, the sound generation factor based on the 2 nd output variable, and the sound generation factor based on the 3 rd output variable.
7. The abnormal sound occurrence factor determination method according to claim 6, wherein, The execution circuit includes a1 st execution circuit provided in the vehicle or in a portable terminal held by an occupant of the vehicle, and a2 nd execution circuit provided outside the vehicle, The 1 st characteristic correction process, the 2 nd characteristic correction process, the variable acquisition process, and the factor selection process are executed by the 2 nd execution circuit.
8. An abnormal sound occurrence factor determination device determines occurrence factors of sounds perceived by a microphone, wherein, The abnormal sound generation factor determination device comprises an execution circuit and a storage circuit, The storage circuit stores map data defining a map that takes as input a sound signal that is a signal related to sound perceived by the microphone, and outputs a variable related to a factor of occurrence of sound in the vehicle, The mapping is obtained by implementing machine learning, The sound signal input to the map when machine learning is performed on the map is a learning sound signal, the microphone that senses the sound represented by the learning sound signal is a learning microphone, The execution circuit is configured to execute: A characteristic correction process of, when the model information of the microphone is different from the model information of the learning microphone, making the frequency characteristic of the sound signal related to the sound perceived by the microphone approximate to the frequency characteristic of the learning sound signal by correction corresponding to the model information as the information related to the model of the microphone; A variable obtaining process of obtaining a variable output from the map by inputting the sound signal related to the sound perceived by the microphone to the map when the model information of the microphone is the same as the model information of the learning microphone, and inputting the sound signal corrected by the characteristic correction process to the map when the model information of the microphone is different from the model information of the learning microphone, thereby obtaining a variable output from the map, and And a factor determination process of determining a factor of occurrence of sound perceived by the microphone based on the variable acquired by the variable acquisition process.
9. An abnormal sound occurrence factor determination device determines occurrence factors of sounds perceived by a microphone, wherein, The abnormal sound generation factor determination device comprises an execution circuit and a storage circuit, The storage circuit stores map data defining a map that outputs a variable relating to a factor of occurrence of sound in the vehicle by taking as input a sound signal that is a signal relating to sound perceived by the microphone, The mapping is obtained by implementing machine learning, The sound signal input to the map when machine learning is performed on the map is a learning sound signal, the microphone that senses the sound represented by the learning sound signal is a learning microphone, The execution circuit is configured to execute: A1 st characteristic correction process of correcting a frequency characteristic of the sound signal related to the sound perceived by the microphone, and in a case where model information of the microphone is 1 st model information different from model information of the learning microphone, the 1 st characteristic correction process can bring the frequency characteristic of the sound signal close to the frequency characteristic of the learning sound signal; a2 nd characteristic correction process of correcting a frequency characteristic of the sound signal, and in a case where the model information of the microphone is 2 nd model information different from the model information of the learning microphone, the 2 nd characteristic correction process can bring the frequency characteristic of the sound signal close to the frequency characteristic of the learning sound signal; A variable obtaining process of inputting the sound signal corrected by the 1 st characteristic correction process to the map to obtain a variable output from the map as a1 st output variable, and inputting the sound signal corrected by the 2 nd characteristic correction process to the map to obtain a variable output from the map as a2 nd output variable, and inputting the sound signal not corrected to the map to obtain a variable output from the map as a3 rd output variable, and A factor selection process of selecting a factor of occurrence of the sound from among a factor of occurrence of the sound based on the 1 st output variable, a factor of occurrence of the sound based on the 2 nd output variable, and a factor of occurrence of the sound based on the 3 rd output variable.

Description

Abnormal sound generation factor determining method and abnormal sound generation factor determining device Technical Field The present disclosure relates to an abnormal sound occurrence factor determination method and an abnormal sound occurrence factor determination device. Background Japanese patent application laid-open No. 2021-154816 uses a map obtained by performing machine learning for estimating a portion that is a factor of sound generated in a vehicle. Thus, techniques are disclosed for determining a location that is a factor of sound perceived by a microphone. In the method, the execution device obtains a variable output from the map by inputting a sound signal, which is a signal related to sound perceived by the microphone, and a state variable of a drive train device of the vehicle into the map. The execution device then determines the location that is a factor of the sound perceived by the microphone based on the variable output from the map. Disclosure of Invention According to one side of the present disclosure, example 1 of an abnormal sound occurrence factor determination method is provided. The abnormal sound generation factor determination method includes storing map data defining a map in a memory circuit of an analysis device. The map inputs a sound signal as a signal related to sound perceived by a microphone, and outputs a variable related to an occurrence factor of sound in a vehicle. The map is obtained by implementing machine learning. The sound signal input to the map when machine learning is performed on the map is a learning sound signal. The microphone that senses the sound represented by the learning sound signal is a learning microphone. The method further includes executing, by an execution circuit of the analysis device, a sound signal acquisition process in which the sound signal related to the sound perceived by the microphone is acquired, and acquiring, by the execution circuit, model information related to the model of the microphone. The method further includes executing, by the execution circuit, a characteristic correction process of correcting the sound signal acquired by the sound signal acquisition process based on the acquired model information so that a frequency characteristic of the sound signal approaches a frequency characteristic of the learning sound signal. The method further includes executing, by the execution circuit, a variable acquisition process in which the sound signal corrected by the characteristic correction process is input to the map to acquire a variable output from the map, and executing a factor determination process in which a factor of occurrence of the sound perceived by the microphone is determined from the variable acquired by the variable acquisition process. In the above abnormal sound generation factor determination method, the frequency characteristic of the sound signal related to the sound perceived by the microphone is corrected according to the model of the microphone. This can reduce the variation in frequency characteristics of the sound signal caused by the difference in model of the microphone that senses the sound. That is, the frequency characteristic of the audio signal input to the map can be made close to the frequency characteristic of the learning audio signal. Further, by inputting the sound signal thus corrected to the map, the occurrence factor of the sound perceived by the microphone is determined based on the variable output from the map. Therefore, the variation in the accuracy of determining the occurrence factor of the sound corresponding to the model of the microphone can be reduced. When machine learning is performed on the map, a microphone for acquiring a learning sound signal that is a sound signal input to the map is referred to as a learning microphone. The model of a microphone that senses sounds occurring in a vehicle may be different from the model of a learning microphone. The frequency characteristics of the model of the microphone are reflected on the audio signal. Therefore, when the model of the microphone that senses the sound generated in the vehicle is different from the model of the learning microphone, the frequency characteristic of the sound signal related to the sound sensed by the microphone will deviate from the frequency characteristic of the learning sound signal. Therefore, it is difficult to say that the accuracy of determining the occurrence location of the sound based on the variable outputted from the map is high. The above method reduces such a phenomenon. According to another aspect of the present disclosure, there is provided example 2 of the abnormal sound occurrence factor determination method. The abnormal sound generation factor determination method includes storing map data defining a map in a memory circuit of an analysis device. The map takes as input a sound signal that is a signal related to sound perceived by a microphone, and outputs a variable related to an oc