Vehicle diagnosis system

Abstract

To provide a vehicle diagnostic system that can make a diagnosis including determination of a type of an anomaly from sound data recorded.SOLUTION: A vehicle diagnostic system includes a processing circuit 110 and a storage device 120. The storage device 120 stores data of a learned model and probability distribution data obtained by generating pieces of generated data by the learned model using pieces of sound data recorded using vehicles with identified types of anomalies and obtaining, for each of the types of anomalies, a probability distribution of a loss variable. The processing circuit 110 is configured to execute a loss calculation process that generates the pieces of the generated data using pieces of diagnostic sound data and obtains data of the probability distribution of the loss variable in the pieces of the generated data as sample data and a diagnostic process that determines which probability distribution of an anomaly the sample data matches and outputs a diagnosis result indicating that a type of anomaly corresponding to the probability distribution determined to match the sample data has occurred in a target vehicle 10.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] This invention relates to a vehicle diagnostic system.

Background Art

[0002] Patent Document 1 discloses a noise determination device that determines whether specific abnormal noises are included in vehicle running sounds by using sound data obtained by recording vehicle running sounds generated as the vehicle runs. This noise determination device calculates the area of a portion exceeding a threshold value in a waveform obtained by frequency-analyzing the sound data. Then, this noise determination device compares the area of the portion exceeding the threshold value with a predetermined determination value to determine the presence or absence of abnormal noises.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The noise determination device disclosed in Patent Document 1 can determine the presence or absence of abnormal noises, but cannot distinguish the types of abnormal noises. There is a need for a vehicle diagnostic system that can perform a diagnosis including discrimination of the type of abnormality for a target vehicle by using sound data obtained by recording sounds emitted from the target vehicle, which is the vehicle to be diagnosed.

Means for Solving the Problems

[0005] Hereinafter, means for solving the above problems and their operational effects will be described. A vehicle diagnosis system for solving the above problems diagnoses the target vehicle by using sound data obtained by recording sounds emitted from the target vehicle, which is the vehicle to be diagnosed. This vehicle diagnosis system includes a processing circuit and a storage device. The storage device stores data of a learned model trained by supervised learning so as to generate the operation data from the training sound data by using the training sound data recorded while operating a reference vehicle without any abnormality and the operation data indicating the operation status of the reference vehicle collected simultaneously with the training sound data. The storage device also stores probability distribution data obtained for each type of abnormality, which is a probability distribution indicating the magnitude of the error of each generated data from the operation data collected simultaneously with the sound data, by generating a plurality of generated data, which are the operation data generated by the learned model using a plurality of sound data recorded using a plurality of vehicles for which the types of abnormalities can be specified respectively.

[0006] The processing circuit executes a loss calculation process of generating a plurality of generated data by the learned model using a plurality of diagnostic sound data recorded while operating the target vehicle, and obtaining data of the probability distribution of the loss variable in the plurality of generated data as sample data. The processing circuit executes a diagnosis process of determining which abnormality the sample data conforms to in the probability distribution by using the sample data and the probability distribution data, and outputting a diagnosis result indicating that an abnormality of the type corresponding to the probability distribution determined to conform has occurred in the target vehicle.

Advantages of the Invention

[0007] The above vehicle diagnosis system can perform a diagnosis including discrimination of the type of abnormality from the recorded sound data.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Embodiments for Carrying Out the Invention

[0009] Hereinafter, an embodiment of the vehicle diagnosis system will be described with reference to FIGS. 1 to 6. <Configuration of the Vehicle Diagnosis System> As shown in FIG. 1, this vehicle diagnosis system includes a data center 100 and a data acquisition device 300. The data center 100 is communicably connected to the data acquisition device 300 via a communication network 200. As shown in FIG. 1, the data center 100 includes a storage device 120 in which a program is stored and a processing circuit 110. The processing circuit 110 executes various processes by executing the program stored in the storage device 120. The data center 100 includes a communication device 130.

[0010] The data acquisition device 300 is, for example, a personal computer. The data acquisition device 300 includes a storage device 320 in which a program is stored and a processing circuit 310. The processing circuit 310 executes various processes by executing the program stored in the storage device 320. The data acquisition device 300 includes a communication device 330. In this embodiment, the data acquisition device 300 is connected to the data center 100 by wireless communication via the communication network 200. The data acquisition device 300 includes a display device 340 for displaying information. The data acquisition device 300 includes a microphone 350.

[0011] When diagnosing the target vehicle 10, which is the vehicle to be diagnosed, using this vehicle diagnostic system, the microphone 350 is installed at a predetermined position with respect to the target vehicle 10. The data acquisition device 300 is connected to the vehicle control unit 20 of the target vehicle 10. Then, an operator operates the target vehicle 10 to run the target vehicle 10 in a measurement driving pattern. The measurement driving pattern is a driving pattern suitable for acquiring data for diagnosis and is a predetermined driving pattern over a predetermined time. When the target vehicle 10 is being run in the measurement driving pattern in this way, the data acquisition device 300 records sound with the microphone 350. The data acquisition device 300 acquires operation data indicating the operation status of the target vehicle 10 simultaneously with the recording of the sound data.

[0012] The vehicle control unit 20 controls each part of the target vehicle 10. Various sensors for detecting the state of the target vehicle 10 are connected to the vehicle control unit 20. When the data acquisition device 300 is connected to the vehicle control unit 20, the data acquisition device 300 can acquire information on the target vehicle 10 through the vehicle control unit 20.

[0013] <Overview of Diagnosis by Vehicle Diagnostic System> As described above, in this vehicle diagnostic system, when diagnosing the target vehicle 10, the data acquisition device 300 is connected to the vehicle control unit 20 of the target vehicle 10. Then, while the target vehicle 10 is in operation, the data acquisition device 300 records sound with the microphone 350. The data acquisition device 300 transmits data including the recorded sound data to the data center 100. Then, the data center 100 diagnoses the target vehicle 10 using the received data.

[0014] The data acquisition device 300 records, as diagnostic sound data, the sound data recorded by the microphone 350 while operating the target vehicle 10 in a measurement driving pattern over a predetermined time in the storage device 320. The data acquisition device 300 stores the operation data collected simultaneously with the sound data in the storage device 320. For example, when diagnosing the transmission of the target vehicle 10, the operation data is the engine rotational speed NE, the input rotational speed Nin, the output rotational speed Nout, and the gear ratio. Then, the data acquisition device 300 stores, as a data set for a predetermined time, the operation data including the diagnostic sound data collected in this way in the storage device 320.

[0015] The data acquisition device 300 extracts a data set for a predetermined time stored in the storage device 320 while changing the extraction start time for each piece of data within a window Tw, which has a time width shorter than the predetermined time, and shapes the extracted data into diagnostic data. That is, the data acquisition device 300 extracts data for a period shorter than the predetermined time and shapes the extracted data into diagnostic data. In the data shaping process for shaping the diagnostic data, the data acquisition device 300 converts the diagnostic sound data into a mel spectrogram and treats it as image data. The vertical axis of the mel spectrogram is frequency expressed in the mel scale, and the horizontal axis is the time axis. In the mel spectrogram, intensity is represented by color. Lower intensity portions are represented by darker blue colors, and higher intensity portions are represented by brighter red colors. The sound data included in one data set for the predetermined time becomes one mel spectrogram for the predetermined time. The data acquisition device 300 transmits the shaped diagnostic data to the data center 100. In this vehicle diagnostic system, a predetermined number of data sets are used to diagnose one target vehicle 10. Therefore, an operator operates the target vehicle 10 in a measurement driving pattern and repeats the task of creating data sets a predetermined number of times to create a predetermined number of diagnostic data. The data acquisition device 300 transmits the predetermined number of diagnostic data to the data center 100. Upon receiving the diagnostic data, the data center 100 stores the diagnostic data in the storage device 120. The data center 100 then executes the routine shown in FIG. 2 to diagnose the target vehicle 10. This routine is executed by the processing circuit 110 of the data center 100. In the following description, the step numbers of each process are indicated by adding an "S" before the number.

[0016] 2, the data center 100 executes a loss calculation process (S100) and a diagnosis process (S200) to diagnose the target vehicle 10. The loss calculation process (S100) is a process of inputting diagnostic sound data included in the diagnostic data into a trained model to calculate a loss variable LOS. The loss variable LOS is a value indicating the magnitude of the error in the generated data generated by the trained model.

[0017] The trained model is a model trained by supervised learning using a large amount of training data, including training sound data and operational data collected simultaneously with the training sound data, so that operational data can be generated from the training sound data. For example, the data center 100 handles diagnostic sound data as image data. For this reason, the data center 100 uses a model that partially uses the image classification model ResNet-18. ResNet-18 is a pre-trained image classification model trained on the ImageNet dataset. ResNet-18 has been trained on over one million image data and can classify input images into 1,000 categories. The trained model stored in the storage device 120 of the data center 100 is a model obtained by transfer learning of the pre-trained ResNet-18. This trained model is a model obtained by replacing the classification output layer of ResNet-18 with a neural network MLP and training this neural network MLP by supervised learning. Training involves adjusting the weights and biases of each layer of this neural network MLP. The training sound data is sound data recorded while a reference vehicle, which is a vehicle in a state that serves as a reference for diagnosis, is operated in a measurement driving pattern. In this example, the reference vehicle is a vehicle that has been subjected to a certain amount of break-in after manufacture, has undergone sufficient maintenance inspection, and has been confirmed to have no abnormalities.

[0018] The learned model is optimized to generate operation data from the sound emitted from the reference vehicle. Therefore, when diagnostic sound data emitted from the target vehicle 10 in a state different from that of the reference vehicle is input, the data on the operation status cannot be correctly generated. That is, if the state of the target vehicle 10 deviates from the state of the reference vehicle, a deviation occurs between the operation data stored in the data set as correct data and the generated data. A large deviation indicates that the state of the target vehicle 10 deviates from the state of the reference vehicle by that much. As described above, the reference vehicle is a vehicle in a normal state. Therefore, in this vehicle diagnosis system, the abnormality of the target vehicle 10 is diagnosed based on this deviation, that is, the error of the generated data. The way the error appears varies depending on the type of abnormality.

[0019] FIG. 3 is a graph showing operational data and generated data generated using a trained model. The graph in FIG. 3 shows only one of the multiple variables included in the operational data. In other words, it shows only one of the objective variables, which are the output of the trained model. In FIG. 3, the operational data, which is the correct data, is shown with a solid line. In FIG. 3, the generated data is shown with a square symbol. In FIG. 3, time ts is the start time of the measurement driving pattern, and time tf is the end time of the measurement driving pattern. In other words, FIG. 3 shows the operational data and generated data in one dataset. The trained model outputs one value of the generated data in window Tw for each variable based on the data within that window Tw, which has been formatted into a list. The error ERR indicates the deviation of the generated data from the operational data, which is the correct data, i.e., the error. The greater the deviation of the state of the target vehicle 10 from the state of the reference vehicle, the greater the error ERR. Therefore, this vehicle diagnosis system calculates a loss variable LOS, which is the sum of the errors ERR of all variables included in the generated data generated from one dataset, through a loss calculation process. Therefore, one loss variable LOS is calculated for one data set. The loss variable LOS only needs to indicate the magnitude of the error. Therefore, the loss variable LOS does not have to be a simple sum of the errors ERR. For example, as a preprocessing step for calculating the loss variable LOS, a process of eliminating outliers from multiple errors ERR can be performed, and the sum after excluding the outliers can be used as the loss variable LOS.

[0020] FIG. 4 shows an example of the probability distribution of the loss variable LOS according to the presence or absence of an abnormality and the type of abnormality. FIG. 4(a) shows the probability distribution of the loss variable LOS in a test vehicle in which an abnormality has occurred in component D. FIG. 4(b) shows the probability distribution of the loss variable LOS in a test vehicle in which an abnormality has occurred in component C. FIG. 4(c) shows the probability distribution of the loss variable LOS in a test vehicle in which an abnormality has occurred in component B. FIG. 4(d) shows the probability distribution of the loss variable LOS in a test vehicle in which an abnormality has occurred in component A. FIG. 4(e) shows the probability distribution of the loss variable LOS in a reference vehicle in which no abnormality has occurred. These probability distributions are obtained using a plurality of sound data recorded by preparing a plurality of test vehicles in which the type of abnormality is specified and repeatedly performing test runs in a measurement driving pattern a plurality of times for each test vehicle. For example, 10 test vehicles in which an abnormality has occurred in component D are prepared, and test runs are repeatedly performed a plurality of times for the 10 test vehicles to record a plurality of sound data. Using the sound data thus recorded, the probability distribution of the loss variable LOS when an abnormality has occurred in component D is obtained. The other probability distributions according to the type of abnormality and the probability distribution when no abnormality has occurred are also obtained in the same manner.

[0021] As shown in FIG. 4, the probability distribution of the loss variable LOS changes according to the presence or absence of an abnormality and the type of abnormality. The learned model is optimized to generate operation data of a reference vehicle in a state where no abnormality has occurred from the sound data. Therefore, the probability distribution of the loss variable LOS when no abnormality has occurred has an average of 0.

[0022] The storage device 120 of the data center 100 stores probability distribution data obtained for each type of abnormality of the probability distribution of the loss variable LOS and probability distribution data for the reference vehicle.

[0023] This vehicle diagnostic system performs diagnostic processing using the probability distribution data for the reference vehicle stored in the storage device 120 and the probability distribution data. Hereinafter, the contents of the loss calculation process and the diagnostic process will be described with reference to a flowchart.

[0024] <Calculation Process of Loss> As described above, when the data center 100 receives a predetermined number of diagnostic data, the received diagnostic data is stored in the storage device 120. Then, the processing circuit 110 of the data center 100 executes the routine shown in FIG. 2.

[0025] As shown in FIG. 2, when starting this routine, the processing circuit 110 first executes a loss calculation process (S100). FIG. 5 is a flowchart showing the flow of processing in the loss calculation process routine. When starting this routine, the processing circuit 110 reads one piece of diagnostic data (S110). Then, the processing circuit 110 sequentially inputs a plurality of lists included in the read diagnostic data into the learned model to calculate data on the operating status respectively (S120). That is, the processing circuit 110 generates a plurality of pieces of generated data. Next, for all the generated data, the processing circuit 110 calculates the magnitude of the error ERR in the generated data for each variable in the operating data collected simultaneously with the diagnostic sound data (S130). Then, the processing circuit 110 stores and records the calculated error ERR in the storage device 120 (S130).

[0026] Next, the processing circuit 110 calculates a loss variable LOS (S140). Then, the processing circuit 110 stores and records the calculated loss variable LOS in the storage device 120 (S140). After recording the loss variable LOS, the processing circuit 110 determines whether the calculation of the loss variable LOS has been executed a predetermined number of times (S150). This process is a process of determining whether the calculation of the loss variable LOS for all of the predetermined number of diagnostic data has been completed.

[0027] When it is determined that the processing circuit 110 has not executed the calculation of the loss variable LOS for the default number of times (S150: NO), the processing circuit 110 returns the process to S110. Then, the processing circuit 110 reads one piece of diagnostic data for which the loss variable LOS has not been calculated (S110) and executes the processes from S120 to S140. When it is determined that the processing circuit 110 has executed the calculation of the loss variable LOS for the default number of times (S150: YES), the processing circuit 110 advances the process to S160. That is, the processing circuit 110 repeats the processes from S110 to S140 until the calculation of the loss variable LOS for all of the default number of pieces of diagnostic data is completed. Then, when the calculation of the loss variable LOS for all of the default number of pieces of diagnostic data is completed, the processing circuit 110 advances the process to S160.

[0028] The processing circuit 110 records, as sample data, the data of the probability distribution of the calculated default number of loss variables LOS in the storage device 120 (S160). Then, the processing circuit 110 ends this routine.

[0029] As shown in FIG. 2, when the loss calculation process (S100) ends, the processing circuit 110 executes a diagnosis process (S200). <Regarding the diagnosis process> In the diagnosis process, the processing circuit 110 outputs a diagnosis result using the sample data, the probability distribution data, and the probability distribution data for the reference vehicle. Specifically, the processing circuit 110 determines which probability distribution the sample data corresponds to for any abnormality, and outputs a diagnosis result indicating that the type of abnormality corresponding to the determined probability distribution has occurred in the target vehicle 10. Further, the processing circuit 110 also determines whether the sample data matches the probability distribution data for the reference vehicle, and if it is determined that they match, outputs a diagnosis result indicating that no abnormality has occurred in the target vehicle 10.

[0030] FIG. 6 is a flowchart showing the flow of processing in the diagnosis process routine. When starting this routine, the processing circuit 110 reads the sample data recorded in the storage device 120 (S210).

[0031] Next, the processing circuit 110 performs a statistical hypothesis test (S220) in order to perform a diagnosis including not only determination of the presence or absence of an abnormality in the target vehicle 10 but also discrimination of the type of abnormality. Specifically, the processing circuit 110 first selects one probability distribution. That is, it selects either one of the probability distributions corresponding to the plurality of types of abnormalities included in the probability distribution data or the probability distribution for the reference vehicle. Then, the processing circuit 110 sets the selected probability distribution as the population. The processing circuit 110 determines whether the sample data fits the probability distribution in the population by performing a statistical hypothesis test. In the statistical hypothesis test, a null hypothesis that is contrary to the alternative hypothesis that the sample data does not fit the population is set. And in the statistical hypothesis test, if this null hypothesis cannot be rejected, it is determined that the sample data fits the population. In this case, the null hypothesis is the hypothesis that the sample data fits the population. There are types such as t-test, F-test, chi-square test, etc. for the statistical hypothesis test depending on the type of test statistic, and which method to adopt may be determined according to the distribution of the data.

[0032] In this example, for example, the significance level is set to 0.05. The significance level may be set to 0.01. In this example, the processing circuit 110 obtains a test statistic from the sample data. Then, the processing circuit 110 calculates a p-value, which is the probability that the test statistic takes a value equal to the test statistic calculated from the sample data when the null hypothesis is assumed to be correct, and compares the p-value with the significance level. The processing circuit 110 rejects the null hypothesis when the p-value is smaller than the significance level. And the processing circuit 110 determines that the sample data does not fit the population. On the other hand, when the p-value is greater than or equal to the significance level, the processing circuit 110 cannot reject the null hypothesis, so it determines that the sample data fits the population.

[0033] Next, the processing circuit 110 records the result of the statistical hypothesis test in the storage device 120 (S230). Then, the processing circuit 110 determines whether the tests for all of the probability distributions corresponding to the plurality of types of abnormalities included in the probability distribution data and the probability distribution for the reference vehicle have been completed (S240).

[0034] When the processing circuit 110 determines that not all the tests are completed (S240: NO), the processing circuit 110 returns the process to S220. Then, the processing circuit 110 performs a statistical hypothesis test (S220) using the probability distribution in which the tests are not completed as a new population, and records the result in the storage device 120 (S230). When the processing circuit 110 determines that all the tests are completed (S240: YES), the processing circuit 110 advances the process to S250. That is, the processing circuit 110 repeats the processes of S220 and S230 until the tests for all the probability distributions corresponding to the plurality of types of abnormalities included in the probability distribution data and the probability distribution for the reference vehicle are completed. In this way, the processing circuit 110 performs a statistical hypothesis test for each population using each probability distribution corresponding to the plurality of types of abnormalities included in the probability distribution data or the probability distribution for the reference vehicle as a population. Thereby, the processing circuit 110 determines whether the sample data fits the probability distribution in each population.

[0035] Then, when the tests for all the probability distributions corresponding to the plurality of types of abnormalities included in the probability distribution data and the probability distribution for the reference vehicle are completed, the processing circuit 110 advances the process to S250.

[0036] Next, the processing circuit 110 outputs a diagnosis result of the target vehicle 10. Specifically, the processing circuit 110 outputs a diagnosis result indicating that an abnormality corresponding to the population for which it is determined that the sample data fits, with reference to the test results recorded in the storage device 120, has occurred in the target vehicle 10. Further, when it is determined that the data fits the probability distribution data for the reference vehicle, the processing circuit 110 outputs a diagnosis result indicating that no abnormality has occurred in the target vehicle 10.

[0037] For example, if the processing circuitry 110 determines that the sample data fits the probability distribution of the loss variable LOS in a test vehicle in which an abnormality has occurred in part B, it outputs a diagnosis result that an abnormality has occurred in part B. For example, if the processing circuitry 110 determines that the sample data fits the probability distribution of the loss variable LOS in a reference vehicle, it outputs a diagnosis result that no abnormality has occurred in the target vehicle 10.

[0038] After outputting the diagnostic result in this way, the processing circuit 110 ends this routine. That is, the processing circuit 110 ends the diagnostic process and ends the series of routines shown in FIG.

[0039] The data acquisition device 300 receives the diagnostic result and displays the received diagnostic result on the display device 340 as the diagnostic result of the target vehicle 10 . <Operation of this embodiment> The processing circuit 110 generates generated data by restoring the operational data from the diagnostic data using the trained model. Then, the processing circuit 110 executes a loss calculation process (S100) to calculate a loss variable LOS in the generated data and create sample data.

[0040] In the diagnosis process (S200), the processing circuit 110 determines which of the population sets the sample data applies to, and outputs a diagnosis result corresponding to the population set that the processing circuit 110 determines the sample data applies to.

[0041] The vehicle diagnosis system uses a loss variable LOS calculated by comparing generated data calculated using a trained model with operational data. The loss variable LOS reflects the difference in the state between the target vehicle 10 and the reference vehicle. That is, the loss variable LOS also reflects characteristics according to the type of abnormality occurring in the target vehicle 10. Therefore, the probability distribution of the loss variable LOS reflects characteristics according to the type of abnormality occurring in the target vehicle 10.

[0042] Therefore, this vehicle diagnostic system can identify the type of abnormality by comparing the probability distribution of the loss variable LOS of the target vehicle 10 with the probability distribution data obtained for each type of abnormality of the probability distributions of the loss variables LOS of a plurality of vehicles for which the types of abnormalities have been identified.

[0043] <Effects of this Embodiment> (1) According to the vehicle diagnostic system described above, it is possible to perform a diagnosis including discrimination of the type of abnormality from the recorded sound data.

[0044] (2) In the storage device 120, in addition to the probability distribution data, data of the probability distribution for the reference vehicle is stored. In the diagnostic process, the processing circuit 110 also determines whether the sample data matches the data of the probability distribution for the reference vehicle. Then, when the processing circuit 110 determines that the sample data matches the data of the probability distribution for the reference vehicle, it outputs a diagnostic result indicating that no abnormality has occurred in the target vehicle 10. Thereby, the vehicle diagnostic system can diagnose that no abnormality has occurred in the target vehicle 10.

[0045] (3) The learned model generates generated data from diagnostic sound data including image data of a spectrogram obtained by frequency analyzing the sound data. The vehicle diagnostic system uses the image data obtained by frequency analyzing the sound data. Therefore, according to this vehicle diagnostic system, it is possible to efficiently extract the features included in the sound data and perform the loss calculation process.

[0046] [[ID=1~17]] <Modification Example> This embodiment can be implemented with the following modifications. This embodiment and the following modification examples can be implemented in combination with each other as long as there is no technical contradiction.

[0047] The data used as diagnostic data does not have to be a mel spectrogram. For example, a spectrogram obtained by wavelet transforming sound data may be used. A spectrogram obtained by short-time Fourier transforming sound data may also be used. It is not necessary to convert sound data into image data. For example, features may be extracted from the sound data itself and used as diagnostic data. In this case, there is no need to use ResNet-18, which handles image data, as a trained model. Although a model using transfer learning of ResNet-18 is shown as an example, the model structure is not limited to this. The trained model only needs to be able to output generated data based on diagnostic data.

[0048] The vehicle diagnostic system may be composed of only the data acquisition device 300. In this case, the memory device 320 stores the trained model data, the probability distribution data, and the probability distribution data for the reference vehicle. In this case, the processing circuit 310 of the data acquisition device 300 executes the loss calculation process and the diagnostic process.

[0049] The data acquisition device 300 itself does not have to be equipped with a microphone 350. Sound data can be acquired from an external device and loss calculation processing can be performed. Loss calculation processing can also be performed using multiple sound data recorded using multiple microphones 350.

[0050] In the above embodiment, as an example of diagnosing the transmission of the target vehicle 10, an example was shown in which the operation data included data on the rotational speed of a rotating shaft in the powertrain. The vehicle diagnostic system may also diagnose the target vehicle 10 by diagnosing other units in the target vehicle 10. For example, to diagnose an engine, data on the engine rotational speed NE, ignition timing, and engine load factor may be used as operation data. For example, to diagnose a drive motor, data on the rotational speed of the motor's output shaft may be used as operation data. To diagnose a four-wheel drive vehicle, data on the rotational speed of each drive wheel may be used as operation data.

[0051] ·In the diagnostic process, an example in which the processing circuit 110 performs a statistical hypothesis test was shown. In the diagnostic process, it suffices to be able to determine, using the sample data and the probability distribution data, to which probability distribution of which population the sample data conforms. For example, it may be determined whether the average value in the sample data is closest to the average value in the probability distribution of each population, and it may be determined that the sample data conforms to the probability distribution with the closest average value.

[0052] ·In the above embodiment, the vehicle diagnostic system includes the processing circuit 110 and the storage device 120 and executes software processing. However, this is merely an example. For example, the vehicle diagnostic system may include a dedicated hardware circuit (such as an ASIC etc.) that processes at least a part of the software processing executed in the above embodiment. That is, the vehicle diagnostic system may have any of the following configurations (A) to (C). (A) The vehicle diagnostic system includes an execution device that executes all processing according to a program and a storage device that stores the program. That is, the vehicle diagnostic system includes a software execution device. (B) The vehicle diagnostic system includes an execution device that executes a part of the processing according to a program and a storage device. Further, the vehicle diagnostic system includes a dedicated hardware circuit that executes the remaining processing. (C) The vehicle diagnostic system includes a dedicated hardware circuit that executes all processing. The software execution device and / or the dedicated hardware circuit may be plural. That is, the above processing may be executed by a processing circuitry including at least one of one or more software execution devices and one or more dedicated hardware circuits. The storage device that stores the program, that is, the computer-readable medium, includes any available medium that can be accessed by a general-purpose or dedicated computer.

Explanation of Signs

[0053] 10... Target vehicle, 20... Vehicle control unit, 100... Data center, 110... Processing circuit, 120... Storage device, 130... Communication device, 200... Communication network, 300... Data acquisition device, 310... Processing circuit, 320... Storage device, 330... Communication device, 340... Display device, 350... Microphone

Claims

1. A vehicle diagnosis system for diagnosing a target vehicle by using sound data obtained by recording sounds emitted from the target vehicle which is a vehicle to be diagnosed, comprising a processing circuit and a storage device, Data of a learned model trained by supervised learning to generate the operation data from the training sound data by using training data including the training sound data recorded while operating a reference vehicle without any abnormality and the operation data indicating the operation status of the reference vehicle collected simultaneously with the training sound data, A plurality of generated data which are the operation data generated by the learned model by using a plurality of sound data recorded by using a plurality of vehicles for which types of abnormalities can be specified respectively, and probability distribution data of loss variables indicating the magnitudes of errors of the respective generated data from the operation data collected simultaneously with the sound data, obtained for each type of abnormality, are stored in the storage device, A loss calculation process for generating a plurality of the generated data by the learned model by using a plurality of diagnostic sound data recorded while operating the target vehicle, and obtaining data of the probability distribution of loss variables in the plurality of generated data as sample data, A diagnostic process for determining which probability distribution the sample data fits for any abnormality by using the sample data and the probability distribution data, and outputting a diagnostic result indicating that an abnormality of the type corresponding to the probability distribution determined to fit has occurred in the target vehicle, The vehicle diagnosis system in which the processing circuit executes the above.

2. In the storage device, in addition to the probability distribution data, data of the probability distribution for the reference vehicle is also stored, In the diagnostic process, the processing circuit also determines whether the sample data fits the data of the probability distribution for the reference vehicle, and when it is determined that it fits, outputs a diagnostic result indicating that no abnormality has occurred in the target vehicle The vehicle diagnosis system according to Claim 1.

3. In the diagnostic process, the processing circuit performs a statistical hypothesis test for each of the probability distributions corresponding to the plurality of types of abnormalities included in the probability distribution data as a population to determine whether the sample data fits the probability distribution in each population, and outputs a diagnostic result indicating that an abnormality corresponding to the population determined to fit the sample data has occurred in the target vehicle The vehicle diagnosis system according to claim 1 or claim 2.

4. The learned model generates the generated data from the diagnostic sound data including image data of a spectrogram obtained by frequency-analyzing the sound data. The vehicle diagnosis system according to claim 1.

5. The loss variable is the sum of errors of all variables included in the generated data generated from the sound data. The vehicle diagnosis system according to claim 1.

Images