Vehicle Evaluation System

Abstract

To provide a vehicle evaluation system capable of determining the level of the state of a target vehicle from recorded sound data and evaluating the target vehicle.SOLUTION: A vehicle evaluation system includes a processing circuit 110 and a storage device 120. The storage device 120 stores: data on a learned model; and probability distribution data in which a plurality of pieces of generated data are generated by the learned model using a plurality of pieces of sound data recorded using a plurality of vehicles for which evaluation ranks are respectively identified and probability distribution of a loss variable is obtained for each evaluation rank. The processing circuit 110 executes: loss calculation processing that generates the plurality of pieces of generated data using a plurality of pieces of sound data for evaluation and obtains the data on the probability distribution of the loss variable in the plurality of pieces of generated data as sample data; and evaluation processing that determines which probability distribution for evaluation rank matches the sample data, and outputs the evaluation rank corresponding to the probability distribution determined to match as an evaluation rank of a target vehicle 10.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a vehicle evaluation system. [Background technology]

[0002] Patent Document 1 discloses an abnormal sound detection device that analyzes sound data recorded from a vehicle to determine whether or not an abnormal sound is occurring. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-222189 Summary of the Invention [Problem to be solved by the invention]

[0004] To evaluate a vehicle, it is necessary not only to determine whether or not abnormal noise occurs, but also to distinguish the level of the vehicle. Therefore, a vehicle evaluation system suitable for evaluating vehicles is required. [Means for solving the problem]

[0005] The means for solving the above problems and their effects will be described below. A vehicle evaluation system for solving the above problem evaluates a target vehicle using sound data recorded from the target vehicle, which is a vehicle to be evaluated. The vehicle evaluation system includes a processing circuit and a storage device. The storage device stores data of a trained model trained by supervised learning to generate operation data from training sound data, including training sound data recorded while the reference vehicle is operating in a state that serves as a reference for evaluation, and operation data indicating the operating status of the reference vehicle collected simultaneously with the training sound data. The storage device also stores probability distribution data, which is the operation data generated by the trained model using multiple sound data recorded using multiple vehicles, each of which has an evaluation rank indicating the level of the vehicle's condition, and which calculates, for each evaluation rank, a probability distribution of a loss variable indicating the magnitude of error of each generated data from the operation data collected simultaneously with the sound data.

[0006] The processing circuit generates a plurality of pieces of generated data by the trained model using a plurality of pieces of evaluation sound data recorded while the target vehicle is in operation, and executes a loss calculation process to obtain probability distribution data of loss variables in the plurality of pieces of generated data as sample data. The processing circuit executes an evaluation process to determine, using the sample data and the probability distribution data, for which probability distribution the sample data applies, and to output the evaluation rank corresponding to the probability distribution that the sample data applies to as the evaluation rank of the target vehicle. [Effects of the Invention]

[0007] The vehicle evaluation system described above can determine the level of the condition of the target vehicle from the recorded sound data and evaluate the target vehicle. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a schematic diagram showing an embodiment of a vehicle evaluation system. [Figure 2]FIG. 2 is a flowchart showing the flow of processing related to the evaluation of a target vehicle. [Figure 3] FIG. 3 is a graph showing generated data, operational data, and an error in the generated data. [Figure 4] 4(a) to 4(e) are graphs showing examples of probability distributions of loss variables. [Figure 5] FIG. 5 is a flowchart showing the flow of the loss calculation process. [Figure 6] FIG. 6 is a flowchart showing the process flow in the evaluation process. DETAILED DESCRIPTION OF THE INVENTION

[0009] An embodiment of a vehicle evaluation system will be described below with reference to FIGS. <Vehicle evaluation system configuration> As shown in FIG. 1, this vehicle evaluation system includes a data center 100 and a data acquisition device 300. The data center 100 is communicatively 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 the program stored in the storage device 120 to perform various processes. 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 the program stored in the storage device 320 to perform various processes. 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 that displays information. The data acquisition device 300 includes a microphone 350.

[0011] When using this vehicle evaluation system to evaluate a target vehicle 10, which is a vehicle to be evaluated, a microphone 350 is installed at a predetermined position relative to the target vehicle 10. A data acquisition device 300 is connected to the vehicle control unit 20 of the target vehicle 10. An operator then operates the target vehicle 10 to operate the target vehicle 10 in a measurement driving pattern. The measurement driving pattern is a driving pattern suitable for acquiring data for evaluation, and is a predetermined driving pattern spanning a predetermined period of time. While the target vehicle 10 is operating in this measurement driving pattern, the data acquisition device 300 records sound with the microphone 350. The data acquisition device 300 acquires operational data indicating the operating status of the target vehicle 10 simultaneously with recording the sound data.

[0012] The vehicle control unit 20 controls each part of the target vehicle 10. Various sensors that detect 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 about the target vehicle 10 through the vehicle control unit 20.

[0013] <Outline of evaluation by the vehicle evaluation system> As described above, in this vehicle evaluation system, when evaluating 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 operating, the data acquisition device 300 records sound with the microphone 350. The data acquisition device 300 transmits data including data on the recorded sound to the data center 100. Then, the data center 100 evaluates the target vehicle 10 using the received data.

[0014] The data acquisition device 300 records sound data collected by the microphone 350 while the target vehicle 10 is operating in a measurement driving pattern for a predetermined time in the storage device 320 as evaluation sound data. The data acquisition device 300 stores operational data collected simultaneously with the sound data in the storage device 320. For example, when evaluating the transmission of the target vehicle 10, the operational data includes the engine rotation speed NE, the input rotation speed Nin, the output rotation speed Nout, and the gear ratio. The data acquisition device 300 then stores the collected operational data, including the evaluation sound data, in the storage device 320 as a data set for a predetermined time.

[0015] The data acquisition device 300 extracts data from a data set of 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 evaluation data. That is, the data acquisition device 300 extracts data for a period shorter than the predetermined time and shapes the data into evaluation data. In the data shaping process for shaping the evaluation data, the data acquisition device 300 converts the evaluation sound data into a mel spectrogram and handles 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 of a predetermined time becomes one mel spectrogram of the predetermined time. The data acquisition device 300 transmits the shaped evaluation data to the data center 100. In this vehicle evaluation system, a predetermined number of data sets are used to evaluate one target vehicle 10. To this end, 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 evaluation data. The data acquisition device 300 transmits the predetermined number of evaluation data to the data center 100. Upon receiving the evaluation data, the data center 100 stores the evaluation data in the storage device 120. The data center 100 then executes the routine shown in FIG. 2 to evaluate 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 an evaluation process (S200) to evaluate the target vehicle 10. The loss calculation process (S100) is a process of inputting evaluation sound data included in the evaluation data into a trained model and calculating a loss variable LOS. The loss variable LOS is a value indicating the magnitude of 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 evaluation sound data as image data. Therefore, 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 evaluation, is operated in a measurement driving pattern. In this example, a vehicle that has been subjected to a certain amount of break-in after manufacture, has undergone sufficient maintenance and inspection, and has been confirmed to have no abnormalities is used as the reference vehicle. In other words, the reference vehicle is a vehicle in extremely good condition with almost no deterioration.

[0018] The trained model is optimized to generate operational data from sounds emitted from a reference vehicle. Therefore, if evaluation sound data emitted from a target vehicle 10 in a state different from that of the reference vehicle is input, it is not possible to correctly generate operational status data. In other words, if the state of the target vehicle 10 deviates from the state of the reference vehicle, a discrepancy occurs between the operational data stored in the dataset as correct data and the generated data. A large discrepancy indicates that the state of the target vehicle 10 deviates that much from the state of the reference vehicle. As described above, the reference vehicle is a vehicle in extremely good condition with almost no deterioration. Therefore, this vehicle evaluation system evaluates the target vehicle 10 based on this discrepancy, i.e., the error in the generated data. The manner in which the error appears varies depending on the level of the state of the target vehicle 10.

[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 evaluation 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] Figure 4 shows an example of the probability distribution of the loss variable LOS according to the evaluation rank indicating the level of the vehicle condition. This vehicle evaluation system determines the evaluation rank of the target vehicle 10 through evaluation processing and outputs the evaluation rank. The vehicle evaluation system selects an evaluation rank from five levels of evaluation rank: S rank, A rank, B rank, C rank, and D rank. S rank is the highest evaluation rank of the five evaluation ranks. And D rank is the lowest evaluation rank of the five evaluation ranks. The evaluations decrease in order from S rank, A rank, B rank, C rank, and D rank.

[0021] Figure 4(a) shows the probability distribution of the loss variable LOS for a D-rank test vehicle. Figure 4(b) shows the probability distribution of the loss variable LOS for a C-rank test vehicle. Figure 4(c) shows the probability distribution of the loss variable LOS for a B-rank test vehicle. Figure 4(d) shows the probability distribution of the loss variable LOS for an A-rank test vehicle. Figure 4(e) shows the probability distribution of the loss variable LOS for an S-rank test vehicle. These probability distributions were obtained using multiple sound data recorded by preparing multiple test vehicles with specific evaluation ranks and repeatedly running test runs using a measurement driving pattern with each test vehicle. For example, ten D-rank test vehicles are prepared, and multiple test runs are run with each of the ten test vehicles multiple times to record multiple sound data. The probability distribution of the loss variable LOS for a D-rank test vehicle is obtained using the sound data recorded in this way. Probability distributions for test vehicles of other evaluation ranks can also be obtained using test vehicles of the respective evaluation ranks in a similar manner.

[0022] As shown in Figure 4, the probability distribution of the loss variable LOS changes depending on the evaluation rank. The trained model is optimized to generate operational data of a reference vehicle with little degradation from sound data. Therefore, the probability distribution of the loss variable LOS for an S rank with little degradation has an average of 0.

[0023] The storage device 120 of the data center 100 stores probability distribution data of the loss variable LOS calculated for each evaluation rank. This vehicle evaluation system performs evaluation processing using probability distribution data stored in the storage device 120.

[0024] The loss calculation process and the evaluation process will be described below with reference to flowcharts. <About loss calculation processing> As described above, when the data center 100 receives a predetermined number of evaluation data, it stores the received evaluation data in the storage device 120. Then, the processing circuit 110 of the data center 100 executes the routine shown in FIG.

[0025] As shown in FIG. 2, when this routine starts, the processing circuit 110 first executes a loss calculation process (S100). 5 is a flowchart showing the flow of processing in a loss calculation processing routine. When this routine starts, the processing circuit 110 reads one piece of evaluation data (S110). Then, the processing circuit 110 sequentially inputs multiple lists included in the read evaluation data into the trained model to calculate data on the operating status for each of them (S120). That is, the processing circuit 110 generates multiple pieces of generated data. Next, the processing circuit 110 calculates the magnitude of the error ERR in the generated data for each variable in the operation data collected simultaneously with the evaluation sound data for all of the generated 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 the 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 or not the calculation of the loss variable LOS has been performed a predetermined number of times (S150). This process is a process for determining whether or not the calculation of the loss variable LOS for all of the predetermined number of evaluation data has been completed.

[0027] If the processing circuit 110 determines that the calculation of the loss variable LOS has not been performed the predetermined number of times (S150: NO), the processing circuit 110 returns the process to S110. Then, the processing circuit 110 reads one piece of evaluation data for which the loss variable LOS has not been calculated (S110) and executes the processes of S120 to S140. If the processing circuit 110 determines that the calculation of the loss variable LOS has been performed the predetermined number of times (S150: YES), the processing circuit 110 proceeds to the process of S160. That is, the processing circuit 110 repeats the processes of S110 to S140 until the calculation of the loss variable LOS for all of the predetermined number of evaluation data is completed. Then, when the calculation of the loss variable LOS for all of the predetermined number of evaluation data is completed, the processing circuit 110 proceeds to the process of S160.

[0028] The processing circuit 110 records the calculated probability distribution data of the predetermined number of loss variables LOS as sample data 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) is completed, the processing circuit 110 executes the evaluation process (S200). <About evaluation processing> In the evaluation process, the processing circuit 110 uses the sample data and the probability distribution data to output an evaluation result. Specifically, the processing circuit 110 determines which probability distribution for which evaluation rank the sample data applies, and outputs the evaluation rank corresponding to the probability distribution that the sample data applies to as the evaluation rank of the target vehicle 10.

[0030] 6 is a flowchart showing the flow of processing in the evaluation processing routine. When this routine starts, 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 to determine the evaluation rank of the target vehicle 10 (S220). Specifically, the processing circuit 110 selects one of the probability distributions corresponding to each evaluation rank included in the probability distribution data. Then, the processing circuit 110 sets the selected probability distribution as a population set. The processing circuit 110 performs a statistical hypothesis test to determine whether the sample data fits the probability distribution in the population set. In the statistical hypothesis test, a null hypothesis is set that is opposite to the alternative hypothesis that the sample data does not fit the population set. Then, in the statistical hypothesis test, if the null hypothesis cannot be rejected, it is determined that the sample data fits the population set. In this case, the null hypothesis is a hypothesis that the sample data fits the population set. There are various types of statistical hypothesis tests, such as a t-test, an F-test, and a chi-square test, depending on the type of test statistic, and which method to use can be determined depending on the distribution of the data.

[0032] In this example, the significance level is set to 0.05, for example. The significance level may also be set to 0.01. In this example, the processing circuit 110 calculates a test statistic from the sample data. Then, the processing circuit 110 calculates a p-value, which is the probability that the test statistic will be 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. If the p-value is smaller than the significance level, the processing circuit 110 rejects the null hypothesis. Then, the processing circuit 110 determines that the sample data does not fit the population. On the other hand, if the p-value is equal to or greater than the significance level, the processing circuit 110 cannot reject the null hypothesis and therefore determines that the sample data fits the population.

[0033] Next, the processing circuit 110 records the results of the statistical hypothesis testing in the storage device 120 (S230). Then, the processing circuit 110 determines whether or not testing has been completed for all of the probability distributions corresponding to the evaluation ranks included in the probability distribution data (S240).

[0034] If the processing circuit 110 determines that all tests have not been completed (S240: NO), the processing circuit 110 returns the process to S220. Then, the processing circuit 110 performs a statistical hypothesis test using the untested probability distribution as a new population (S220) and records the result in the storage device 120 (S230). If the processing circuit 110 determines that all tests have been completed (S240: YES), the processing circuit 110 proceeds to S250. That is, the processing circuit 110 repeats the processes of S220 and S230 until tests are completed for all probability distributions corresponding to each evaluation rank included in the probability distribution data. In this way, the processing circuit 110 performs a statistical hypothesis test for each population using each probability distribution corresponding to each evaluation rank included in the probability distribution data as a population. In this way, the processing circuit 110 determines whether the sample data fits the probability distribution in each population.

[0035] Then, when the processing circuit 110 completes testing all of the probability distributions corresponding to the evaluation ranks included in the probability distribution data, it proceeds to S250. Next, the processing circuitry 110 outputs the evaluation result of the target vehicle 10. Specifically, the processing circuitry 110 refers to the test results recorded in the storage device 120 and outputs, as the evaluation rank of the target vehicle 10, an evaluation rank corresponding to the population to which the sample data is determined to apply.

[0036] For example, if the processing circuit 110 determines that the sample data fits the probability distribution of the loss variable LOS for a test vehicle ranked B, it outputs an evaluation result indicating that the test vehicle is ranked B.

[0037] After outputting the evaluation result in this way, the processing circuit 110 ends this routine. That is, the processing circuit 110 ends the evaluation process and ends the series of routines shown in FIG.

[0038] The data acquisition device 300 receives the evaluation result and displays the received evaluation result on the display device 340 as the evaluation result of the target vehicle 10 . <Operation of this embodiment> The processing circuit 110 generates generated data by restoring the operational data from the evaluation 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.

[0039] In the evaluation process (S200), the processing circuit 110 determines which of the population sets the sample data falls into, and outputs an evaluation result corresponding to the population set to which the sample data falls.

[0040] The vehicle evaluation system uses a loss variable LOS calculated by comparing generated data calculated using a trained model with operational data. The loss variable LOS reflects differences in the state of the target vehicle 10 and the reference vehicle. That is, the probability distribution of the loss variable LOS reflects characteristics according to the state of the target vehicle 10. Therefore, the evaluation rank of the target vehicle 10 can be determined by comparing the probability distribution of the loss variable LOS of the target vehicle 10 with probability distribution data obtained by calculating the probability distribution of the loss variable LOS for multiple vehicles with identified evaluation ranks for each evaluation rank.

[0041] <Effects of this embodiment> (1) The vehicle evaluation system described above can determine the level of the condition of the target vehicle 10 from the recorded sound data and evaluate the target vehicle 10.

[0042] (2) The trained model generates generated data from evaluation sound data that includes image data of a spectrogram obtained by frequency-analyzing sound data. The vehicle evaluation system uses image data obtained by frequency-analyzing sound data. Therefore, this vehicle evaluation system can efficiently extract features contained in the sound data and perform loss calculation processing.

[0043] (3) The vehicle evaluation system outputs the evaluation results by applying them to a preset evaluation rank. Therefore, this evaluation system makes it easier to understand the relative level of the condition of the target vehicle 10 in the used car market.

[0044] <Example of change> This embodiment can be modified as follows: This embodiment and the following modifications can be combined and implemented within the scope of technical compatibility.

[0045] The data used as evaluation data does not have to be a mel spectrogram. For example, a spectrogram obtained by wavelet transforming sound data can be used. A spectrogram obtained by short-time Fourier transforming sound data can also be used. It is not necessary to convert sound data into image data. For example, features can be extracted from the sound data itself and used as evaluation data. In that 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 has been 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 the evaluation data.

[0046] In the above vehicle evaluation system, an example is shown in which a vehicle in extremely good condition is used as the reference vehicle to evaluate the target vehicle 10. The reference vehicle does not necessarily have to be a vehicle in good condition. For example, a vehicle in extremely poor condition with a low rating can also be used as the reference vehicle. The loss variable LOS in the above evaluation process is a value that indicates the degree of deviation between the condition of the target vehicle 10 and that of the reference vehicle. Therefore, if an aging vehicle with an extremely low rating is used as the reference vehicle, the smaller the loss variable LOS, the lower the rating. The target vehicle 10 can also be evaluated using the probability distribution of such loss variable LOS.

[0047] The vehicle evaluation system may be composed of only the data acquisition device 300. In this case, the trained model data and the probability distribution data are stored in the storage device 320. In this case, the processing circuit 310 of the data acquisition device 300 executes the loss calculation process and the evaluation process.

[0048] 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.

[0049] In the above embodiment, as an example of evaluating 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 evaluation system may also evaluate the target vehicle 10 by evaluating other units in the target vehicle 10. For example, to evaluate an engine, data on the engine rotational speed NE, ignition timing, and engine load factor may be used as operation data. For example, to evaluate a drive motor, data on the rotational speed of the motor's output shaft may be used as operation data. To evaluate a four-wheel drive vehicle, data on the rotational speed of each drive wheel may be used as operation data.

[0050] In the evaluation process, an example has been shown in which the processing circuit 110 performs a statistical hypothesis test. In the evaluation process, it is sufficient to use sample data and probability distribution data to determine which of the population's probability distributions the sample data applies to. For example, it may be determined whether the mean value of the sample data is closest to the mean value of the probability distribution of each population, and it may be determined that the sample data applies to the probability distribution with the closest mean value.

[0051] In the above embodiment, the vehicle evaluation system includes a processing circuit 110 and a storage device 120 to execute software processing. However, this is merely an example. For example, the vehicle evaluation system may include a dedicated hardware circuit (e.g., an ASIC) that processes at least part of the software processing executed in the above embodiment. That is, the vehicle evaluation system may have any of the following configurations (A) to (C). (A) The vehicle evaluation system includes an execution device that executes all processing in accordance with a program and a storage device that stores the program. That is, the vehicle evaluation system includes a software execution device. (B) The vehicle evaluation system includes an execution device that executes part of the processing in accordance with the program and a storage device. Furthermore, the vehicle evaluation system includes a dedicated hardware circuit that executes the remaining processing. (C) The vehicle evaluation system includes a dedicated hardware circuit that executes all processing. There may be multiple software execution devices and / or dedicated hardware circuits. That is, the above processing may be executed by a processing circuitry that includes at least one of one or more software execution devices and one or more dedicated hardware circuits. A storage device, i.e., a computer-readable medium, that stores a program includes any available medium accessible by a general-purpose or dedicated computer. [Explanation of symbols]

[0052] 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 evaluation system that evaluates a target vehicle using sound data that records sounds emitted from the target vehicle, the vehicle being evaluated, comprising: a processing circuit; and a storage device; Data of a trained model trained by supervised learning to generate the operation data from the training sound data using training data including training sound data recorded while operating a reference vehicle in a state that serves as a reference for evaluation and operation data that indicates the operation status of the reference vehicle and that was collected simultaneously with the training sound data; and a plurality of generated data, which is the operation data generated by the trained model using a plurality of sound data recorded using a plurality of vehicles, each of which has an evaluation rank indicating the level of the vehicle condition, and probability distribution data, which is a probability distribution of a loss variable indicating the magnitude of error of each generated data from the operation data collected simultaneously with the sound data, calculated for each evaluation rank; is stored in the storage device, a loss calculation process that generates a plurality of pieces of generated data by the trained model using a plurality of pieces of evaluation sound data recorded while the target vehicle is in operation, and obtains, as sample data, data on the probability distribution of loss variables in the plurality of pieces of generated data; an evaluation process of determining, using the sample data and the probability distribution data, to which probability distribution for which evaluation rank the sample data applies, and outputting the evaluation rank corresponding to the probability distribution to which the sample data applies as the evaluation rank of the target vehicle; The processing circuit executes the vehicle evaluation system.

2. In the evaluation process, the processing circuit determines whether or not the sample data fits into the probability distribution in each of the population sets by performing a statistical hypothesis test on each of the population sets, each of which corresponds to the plurality of evaluation ranks included in the probability distribution data, and outputs the evaluation rank corresponding to the population set to which the sample data is determined to fit as the evaluation rank of the target vehicle. The vehicle evaluation system according to claim 1 .

3. The trained model generates the generated data from evaluation sound data including image data of a spectrogram obtained by frequency analysis of the sound data. The vehicle evaluation system according to claim 1 .

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

5. The operation data includes data on the rotation speed of a rotating shaft in the powertrain. The vehicle evaluation system according to claim 1 .

Images