Information processing device

Abstract

To provide an information processing device capable of calculating an index value in a short period of time by extracting extraction data having obtained the feature of the entire original data.SOLUTION: A processing device of an information processing device executes retrieval processing which includes a first step (S120) of calculating a relative frequency distribution of original data, a second step (S125) of setting a plurality of time windows for cutting out a partial period data of the original data, a third step (S130) of cutting out data from the original data, a fourth step (S140) of calculating a relative frequency distribution of extraction data, and a fifth step (S145) for calculating an error between the relative frequency distribution of the original data and the relative frequency distribution of the extraction data, and repeatedly executes an attempt from the second step to the fifth step, changing settings of the plurality of time windows. The processing device calculates an index value of damage of a friction material of a transmission by using the extraction data whose error becomes equal to or smaller than the threshold value (S160).SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] This invention relates to an information processing apparatus.

Background Art

[0002] Patent Document 1 discloses an information processing apparatus that reduces the size of analysis data by compressing original data for analysis. The original data for analysis is data collected over a predetermined period using sensors mounted on a vehicle.

[0003] The information processing apparatus disclosed in Patent Document 1 compresses data by extracting data from the original data, including data acquired when a certain vehicle speed is reached and data acquired at the inflection point of the vehicle speed.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] The above information processing apparatus extracts data focusing only on the vehicle speed. Therefore, the above information processing apparatus cannot extract data according to the characteristics of data other than the vehicle speed. There is a need for an information processing apparatus that can obtain extraction data that captures the characteristics of the entire original data including a plurality of feature amounts.

Means for Solving the Problems

[0006] An information processing device for solving the above problem acquires original data created by collecting data over a predetermined period using multiple sensors mounted on a vehicle and calculates an index value indicating the magnitude of damage accumulated in the friction material of the transmission. This information processing device includes a processing unit that performs processing. The original data includes, as feature quantities, data on the amount of heat generated by the friction material of the transmission and data on the engagement frequency of the friction material of the transmission. In this information processing device, the search process performed by the processing unit includes a first step of calculating the relative frequency distribution in the original data for each of the multiple feature quantities included in the original data. The search process includes a second step of setting multiple time windows that extract data for a portion of the original data such that the sum of the periods of all time windows is shorter than the predetermined period. The search process includes a third step of extracting data from the original data using the multiple time windows. The search process includes a fourth step of calculating the relative frequency distribution in the extracted data, which is obtained by combining the data extracted using the multiple time windows, for each of the feature quantities. The search process includes a fifth step of calculating the error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data. After performing the first step, the processing device performs the search process, which involves repeatedly performing the trials from the second to the fifth step by changing the settings of the multiple time windows, to extract the extracted data in which the error is less than or equal to a threshold. The processing device calculates the index value using the extracted data in which the error is less than or equal to a threshold.

[0007] In one embodiment of the information processing device, the device performs clustering, which is a machine learning method, to classify the data in each interval, obtained by dividing the original data into fixed-period intervals, into a predetermined number of clusters. In the second step, the device sets the multiple time windows such that the difference between the ratio of each cluster in the extracted data and the ratio of each cluster in the entire original data is less than or equal to a threshold. [Effects of the Invention]

[0008] This information processing device can calculate index values ​​with the same accuracy as when using the original data, using extracted data that is smaller in volume than the original data. Therefore, this information processing device can achieve both a reduction in data volume and maintenance of accuracy, and can calculate index values ​​in a shorter time compared to when using the original data. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 is a schematic diagram showing the relationship between a data center, which is one embodiment of an information processing device, a vehicle, and an information processing terminal. [Figure 2] Figure 2 is a graph showing the original data, where (a) shows the change in the amount of heat generated by the friction material, and (b) shows the change in the engagement frequency of the friction material. [Figure 3] Figure 3 is a flowchart showing the processing flow performed by the processing unit in the data center. [Figure 4] Figure 4 is a graph showing an example of clustering the original data using two features. [Figure 5] Figure 5 is a graph showing an example of the relative frequency distribution of heat generation from friction materials in the original data. [Figure 6] Figure 6 is a graph showing an example of the relative frequency distribution of engagement frequency of friction materials in the original data. [Modes for carrying out the invention]

[0010] Below, a data center 500, which is one embodiment of an information processing device, will be described with reference to Figures 1 to 6. <Configuration of the information processing system> Figure 1 shows the configuration of an information processing system including a data center 500. As shown in Figure 1, the data center 500 communicates with the vehicles 10 via a communication network 400. The data center 500 also communicates with information processing terminals 600 via the communication network 400. The data center 500 communicates with multiple vehicles 10 and multiple information processing terminals 600 via the communication network 400.

[0011] <Data Center 500 Configuration> As shown in Figure 1, the data center 500 includes a processing unit 510. The data center 500 also includes a storage device 520 and a communication device 530. The processing unit 510 includes a CPU that executes processing according to a program and a ROM in which the program is stored. The storage device 520 stores a large amount of data. The communication device 530 is implemented as hardware such as a network adapter, various communication software, or a combination thereof. The communication device 530 enables wired or wireless communication via the communication network 400.

[0012] The data center 500 may be configured using multiple computers. For example, the data center 500 may be configured using multiple server devices. <Vehicle 10 configuration> Each of the multiple vehicles 10 is equipped with a communication device 80. These communication devices 80 are implemented as hardware such as network adapters, various communication software, or a combination thereof. These communication devices 80 are configured to enable wired or wireless communication via a communication network 400.

[0013] Each vehicle 10 is equipped with an engine 20 and an automatic transmission 30. For example, the automatic transmission 30 is a planetary gear type transmission. The automatic transmission 30 includes a friction engagement device 31 and a hydraulic control circuit 32. The friction engagement device 31 changes the combination of planetary gear trains that transmit power by engaging or disengaging a plurality of friction engagement elements. Thereby, the automatic transmission 30 forms a plurality of shift stages with different gear ratios. The friction engagement elements are, for example, clutches and brakes. These friction engagement elements each have a friction material. The hydraulic control circuit 32 controls the hydraulic pressure supplied to each friction engagement element of the friction engagement device 31.

[0014] The vehicle 10 includes an engine control device 40 and a transmission control device 50. The engine control device 40 controls the engine 20. The transmission control device 50 controls the automatic transmission 30 by controlling the hydraulic control circuit 32.

[0015] The engine control device 40 and the transmission control device 50 are equipped with various sensors that collect information on each part of the vehicle 10. In each vehicle 10, driving data is collected from these various sensors. From each vehicle 10, the driving data is transmitted to the data center 500 by the communication device 80. For example, the driving data including the driving distance, position information, and vehicle speed of each vehicle 10 is transmitted from each vehicle 10 to the data center 500. Also, the driving data includes various data indicating the state of the automatic transmission 30 acquired by the transmission control device 50 of the vehicle 10. The identification information that identifies each vehicle 10 is also transmitted from each vehicle 10 to the data center 500 together with the driving data.

[0016] The data center 500 stores the driving data in the storage device 520 together with the received identification information. Thus, the driving data of a plurality of vehicles 10 is accumulated in the storage device 520 of the data center 500.

[0017] <Configuration of the information processing terminal 600> The information processing terminal 600 includes a processing device 610, a storage device 620, and a communication device 630. The processing device 610 includes a CPU that executes processing according to a program, and a ROM in which the program is stored. The storage device 620 stores data. The communication device 630 is implemented as hardware such as a network adapter, various communication software, or a combination thereof. The communication device 630 realizes wired or wireless communication via the communication network 400. The information processing terminal 600 is, for example, a personal computer.

[0018] <Regarding the analysis of the running data of the vehicle 10> The information processing terminal 600 is used for the task of analyzing the running data. When analyzing the running data, an instruction to execute the analysis is sent from the information processing terminal 600 to the data center 500. The processing device 510 of the data center 500 that has received the instruction performs analysis using a part of the vast amount of running data stored in the storage device 520 of the data center 500. The running data to be used is selected from the vast amount of running data stored in the storage device 520 according to the purpose of the analysis.

[0019] For example, the processing device 510 calculates the load on a specific part of a specific vehicle 10 based on the running data of the specific vehicle 10. The processing device 510 estimates the damage accumulated in that part based on the calculated load. For example, the processing device 510 calculates an index value indicating the magnitude of the damage accumulated in the friction material of the automatic transmission 30 of a specific vehicle 10 based on the running data of the specific vehicle 10. The processing device 510 of the data center 500 outputs the calculated result by sending it to the information processing terminal 600. The received result is displayed on the information processing terminal 600 that has received the result.

[0020] To perform such analysis, the processing device 510 analyzes a large amount of running data collected over a long period. Since the processing device 510 needs to perform a vast amount of calculations, the analysis takes a long time.

[0021] Therefore, it is conceivable to extract data that captures the characteristics of the entire original data from the large amount of original driving data. If such extracted data can be obtained, the processing unit 510 can perform analysis in a shorter time by using the extracted data. For example, when estimating the damage to the friction material after 100,000 hours of driving, the processing unit 510 estimates the damage using 20,000 hours of extracted data extracted from 100,000 hours of original data. Then, the processing unit 510 calculates an index value for the damage to the friction material after 100,000 hours of driving by multiplying the index value calculated from the 20,000 hours of extracted data by 5.

[0022] Figure 2 shows an example of the original data. The original data shown in Figure 2 is 100,000 hours of driving data for one vehicle 10. The original data shown in Figure 2 includes, as features, the amount of heat generated by the friction material for which the damage index value is calculated, and the engagement frequency of the friction material.

[0023] Figure 2(a) shows the change in the amount of heat generated by the friction material over 100,000 hours. The amount of heat generated by the friction material can be calculated based on the relative rotational speed, which is the difference between the input rotational speed and the output rotational speed at the friction engagement element, the torque transmitted by the friction engagement element, and the shift time of the automatic transmission 30. The amount of heat generated is calculated by the transmission control device 50. Alternatively, the relative rotational speed, the torque transmitted by the friction engagement element, and the shift time of the automatic transmission 30 may be transmitted from the vehicle 10 to the data center 500, where the amount of heat generated can be calculated.

[0024] Figure 2(b) shows the change in engagement frequency of the friction material over 100,000 hours. Engagement frequency indicates the number of times the friction material engaged per unit of time. The engagement frequency is calculated by the transmission control device 50. Alternatively, data recording the timing of engagement of each friction material may be transmitted from the vehicle 10 to the data center 500, where the engagement frequency is calculated. The engagement frequency may also be data indicating the interval between gear changes using the same friction material.

[0025] The amount of heat generated and the engagement frequency are correlated with the damage to the friction material of vehicle 10. The processing unit 510 of the data center 500 estimates the damage to the friction material from the engagement frequency and driving data that includes the engagement frequency as features.

[0026] Extracted data is created by cutting out data from the original data using multiple time windows. In Figure 2, three time windows—the first time window W_1, the second time window W_2, and the third time window W_3—are shown as examples of multiple time windows, each represented by a dashed line. The start and end dates of each time window are set so that they do not overlap. In this example, 20,000 hours of driving data are extracted as data. Therefore, the start and end dates of each time window are set so that the total length of the periods of all time windows is 20,000 hours.

[0027] Data center 500 searches for start and end date settings for each time window that represent extraction patterns for extracting data that captures the characteristics of the entire original data. Data Center 500 extracts data from the original data using the extraction patterns found through exploration. Data Center 500 then performs analysis using the extracted data.

[0028] <Search process for extraction patterns> Figure 3 is a flowchart showing the sequence of processes related to the extraction pattern search process. This sequence of processes is performed by the processing unit 510 of the data center 500.

[0029] As shown in Figure 3, the processing unit 510 acquires original data in the processing of step S100. The original data is a portion of the driving data selected from the vast amount of driving data stored in the storage device 520 of the data center 500 according to the purpose of the analysis. For example, the original data for calculating an index value indicating the magnitude of damage accumulated in a specific friction material of the automatic transmission 30 of one vehicle 10 is the driving data of the target vehicle 10 over a predetermined period, selected from the vast amount of driving data of multiple vehicles 10. For example, when estimating the damage to the friction material after 100,000 hours of driving, the original data is the driving data of the target vehicle 10 over a predetermined period.

[0030] In step S110, the processing unit 510 assigns labels to the original data by clustering. Specifically, the processing unit 510 divides the original data into fixed-period intervals. The length of the interval for dividing the original data is, for example, several minutes. Then, the processing unit 510 performs clustering, a machine learning method that classifies the data in each interval into a predetermined number of clusters. The clustering algorithm used is, for example, the k-means method. The k-means method is a clustering algorithm that classifies data into a predetermined number of clusters. The clustering algorithm is not limited to the k-means method.

[0031] The original data includes driving data collected under different conditions, such as driving in urban areas, driving in suburban areas, and driving on highways. By performing clustering, the driving data included in the original data can be classified into clusters of driving data with similar characteristics. The number of clusters to be classified can be arbitrarily set depending on the content of the analysis.

[0032] Figure 4 is a graph showing an example of clustering the original data into four clusters using the k-means method, with two features included in the original data as explanatory variables. For example, the two features are the amount of heat generated and the engagement frequency, as shown in Figure 2. In Figure 4, each data point in each section of the original data is represented by a single point. When performing clustering, the processing unit 510 uses representative values ​​of the explanatory variables in the data for each section. For example, the processing unit 510 uses the mean value of the features in the data for each section as the representative value. The processing unit 510 may also use the moving average of the features over multiple consecutive time-series sections as the representative value.

[0033] In Figure 4, these points are shown on a two-dimensional space with the first feature FV_a and the second feature FV_b as the coordinate axes. Figure 4 is an example of the original data being clustered into four clusters: the first cluster M_1, the second cluster M_2, the third cluster M_3, and the fourth cluster M_4. In Figure 4, the boundaries of the four clusters are shown by solid lines. In Figure 4, the centroids of each cluster are indicated by white triangles. Centroid cgM_1 is the centroid of the first cluster M_1. Centroid cgM_2 is the centroid of the second cluster M_2. Centroid cgM_3 is the centroid of the third cluster M_3. Centroid cgM_4 is the centroid of the fourth cluster M_4.

[0034] Figure 4 shows an example with two explanatory variables, but the number of explanatory variables is not limited to two. For example, if the original data contains three features, the processing unit 510 may use these three features as explanatory variables to perform clustering. In that case, the processing unit 510 will cluster the original data in a three-dimensional coordinate space.

[0035] The processing unit 510 then assigns labels to the original data indicating the clustering results. Specifically, it assigns a label to each data point, which was represented as a point in coordinate space, to identify the classified cluster. In this way, the processing unit 510 creates the original data with labels.

[0036] Next, in step S120, the processing unit 510 calculates the relative frequency distribution of the original data. As mentioned above, the original data contains multiple features. The processing unit 510 calculates the relative frequency distribution of each feature in the original data.

[0037] A frequency distribution classifies data into multiple classes and represents the distribution of the number of data points in each class. Relative frequency indicates what percentage of the total sum of frequencies a particular class represents.

[0038] Figure 5 shows the relative frequency distribution for calorific value in the original data shown in Figure 2. In this relative frequency distribution, the calorific value in the original data is divided into m classes from 1 to m, and the relative frequency distribution is shown for each class.

[0039] Figure 6 shows the relative frequency distribution of engagement frequency in the original data shown in Figure 2. In this relative frequency distribution, the engagement frequency in the original data is divided into m classes from 1 to m, and the relative frequency distribution is shown.

[0040] In step S120, the processing unit 510 calculates this relative frequency distribution for each feature contained in the original data. The number of classes in the relative frequency distribution of each feature is the same.

[0041] For example, as shown in Figure 2, if the original data includes two features, heat generation and engagement frequency, the processing unit 510 calculates the relative frequency distribution of each of these two features.

[0042] Next, in the processing of step S125, the processing unit 510 sets multiple time windows in order to extract extracted data from the original data. Figure 2 shows an example of multiple time windows, consisting of three time windows W_1, W_2, and W_3. In the example shown in Figure 2, the duration of each time window is equal. As shown in Figure 2, the data extracted by each extraction window is the data for each feature over the same period.

[0043] In step S125, the processing unit 510 randomly sets multiple time windows such that the sum of the periods of all time windows is shorter than the default period, which is the period of the entire original data. As will be described later, the processing unit 510 combines all the data extracted using the multiple time windows set here to create extracted data. The sum of the periods of all time windows is a value that determines the capacity of the extracted data. Therefore, the sum of the periods of all time windows is set in advance.

[0044] For example, each time the processing unit 510 executes the process in step S125, it randomly sets the number of time windows, the start date of each time window, and the end date of each time window. At this time, the processing unit 510 sets each time window so that they do not overlap. In this way, the processing unit 510 randomly sets multiple time windows so that the sum of the periods of all time windows equals a predetermined period. In the process in step S125, the processing unit 510 may set multiple time windows by fixing the period of each time window to a constant value, as shown in Figure 2. In the process in step S125, the processing unit 510 may set multiple time windows by fixing the number of time windows to a constant value.

[0045] In addition to the requirements described above, when the processing unit 510 sets multiple time windows through step S125, it sets multiple time windows such that the difference between the ratio of each cluster in the extracted data and the ratio of each cluster in the original data as a whole is less than or equal to a threshold.

[0046] In this way, by setting multiple time windows through the process in step S125, an extraction pattern for extracting data from the original data is determined. Once the extraction pattern is determined, the processing unit 510 proceeds to step S130.

[0047] In step S130, the processing unit 510 extracts data from the original data according to the determined extraction pattern. In other words, in step S130, the processing unit 510 extracts data from the original data according to multiple set time windows. Then, the processing unit 510 combines all the data extracted according to the multiple time windows to create extracted data.

[0048] In the next step, S140, the processing unit 510 calculates the relative frequency distribution of the extracted data. The processing unit 510 calculates the relative frequency distribution of the extracted data in the same way as the method used to calculate the relative frequency distribution in step S120. That is, in the processing of step S140, the processing unit 510 calculates the relative frequency distribution of the extracted data for each feature. At this time, the processing unit 510 makes the number of classes in the relative frequency distribution of each feature the same as the relative frequency distribution in step S120.

[0049] For example, as shown in Figure 2, if the original data includes two features, heat generation and engagement frequency, the processing unit 510 calculates the relative frequency distribution of each of these two features in step S140.

[0050] Next, in step S145, the processing unit 510 calculates the error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data. For example, the processing unit 510 calculates the Mean Absolute Error (MAE). The Mean Absolute Error (MAE) is expressed by the following formula.

[0051]

number

[0052] In the above formula, "n" is the number of features. "m" is the number of classes in the relative frequency distribution. "Y" is the frequency of the corresponding feature in the corresponding class in the original data. "y" is the frequency of the corresponding feature in the corresponding class in the extracted data.

[0053] As shown in the formula above, the processing unit 510 calculates the error as the sum of the frequency errors in each class for each feature between the relative frequency distribution in the entire original data and the relative frequency distribution in the extracted data.

[0054] Once the error is calculated, the processing unit 510 proceeds to step S150. In step S150, the processing unit 510 determines whether the calculated error is less than or equal to a threshold. The threshold is a value used to determine whether extracted data having a relative frequency distribution similar to the relative frequency distribution in the original data has been extracted using the set extraction pattern. The magnitude of this threshold is pre-set so that, based on the error being less than or equal to the threshold, it can be determined that extracted data having a relative frequency distribution similar to the relative frequency distribution in the original data has been extracted.

[0055] If, in the process of step S150, it is determined that the error is below a threshold (step S150: YES), the processing unit 510 proceeds to step S160. In step S160, the processing unit 510 calculates a target index value using the extracted data created in the most recent step S130. Here, an index value indicating the magnitude of damage accumulated in the friction material is calculated. For example, the processing unit 510 calculates the fatigue damage degree as an index value indicating the magnitude of damage accumulated in the friction material.

[0056] The fatigue damage level is an index value that represents the percentage of accumulated fatigue, assuming that the damage to the friction material follows the linear cumulative damage law, also known as Minor's law, with the fatigue leading to damage set as "1". Here, the damage inflicted on the friction material over a certain period is calculated from the amount of heat generated and the engagement frequency. Then, the magnitude of the damage that leads to damage to the friction material due to a single load input is set as "1", and the calculated percentage of damage is used as the fatigue index value. By repeating this process, the calculated fatigue index values ​​are accumulated to calculate the fatigue damage level, which is the percentage of accumulated fatigue relative to the fatigue that leads to damage. When the fatigue damage level reaches "1", it means that damage has occurred, and the calculated fatigue damage level is a value between "0" and "1".

[0057] Here, since the fatigue damage level is calculated using extracted data, which is part of the original data, the processing unit 510 converts the calculated fatigue damage level to a size corresponding to the original data and calculates the fatigue damage level as an index value. For example, if the original data is 100,000 hours of driving data and the extracted data is 20,000 hours of driving data, the calculated fatigue damage level is multiplied by 5 to obtain the fatigue damage level as an index value.

[0058] On the other hand, if the processing in step S150 determines that the error is greater than the threshold (step S150: NO), the processing unit 510 returns to step S125. Then, the processing unit 510 executes the search process from step S125 to step S145 again.

[0059] In this way, the processing unit 510 repeatedly executes the search process in steps S125 to S145 by changing the settings of multiple time windows, and extracts data from the original data in which the error is below a threshold. Then, the processing unit 510 calculates an index value using the extracted data. Once the index value is calculated, the processing unit 510 proceeds to step S170.

[0060] In step S170, the processing unit 510 determines whether the index value is equal to or greater than a default value. The default value is a value used to predict that the likelihood of damage occurring is high based on the index value being equal to or greater than a default value. For example, here, for example, "0.9" can be set as the default value for the fatigue damage degree. In this case, it is possible to predict that the likelihood of damage occurring is high based on the fact that 90% of the fatigue leading to damage has been reached.

[0061] In step S170, if it is determined that the index value is greater than or equal to a predetermined value (step S170: YES), the processing unit 510 proceeds to step S180. In step S180, the processing unit 510 outputs the index value and the failure prediction. Specifically, the processing unit 510 sends the index value and the failure prediction to the information processing terminal 600 that sent the instruction requesting analysis.

[0062] Failure prediction is, for example, a message indicating that a failure has been predicted. In this way, the processing unit 510 issues a notification indicating that a failure has been predicted if the calculated index value is greater than or equal to a predetermined value. Failure prediction may also be information about the lifespan until a failure occurs. For example, if the index value is the fatigue damage degree calculated using extracted data extracted from 100,000 hours of original data, the processing unit 510 calculates the running time until the fatigue damage degree reaches "1" and outputs it as lifespan information. Lifespan information may also be converted to running distance based on the running distance of 100,000 hours and output.

[0063] In step S170, if it is determined that the index value is less than the default value (step S170: NO), the processing unit 510 proceeds to step S190. In step S190, the processing unit 510 outputs the index value. Specifically, the processing unit 510 sends the index value to the information processing terminal 600 that sent the instruction requesting analysis.

[0064] When the processing in step S180 or step S190 is executed, the processing unit 510 terminates this series of processes. <Operation of this embodiment> The data center 500, which is an information processing device in this embodiment, acquires original data created by collecting data over a predetermined period of time using multiple sensors mounted on the vehicle 10, and calculates an index value indicating the magnitude of damage accumulated in the friction material of the automatic transmission 30.

[0065] The data center 500 includes a processing unit 510 that performs processing. The original data includes data on the amount of heat generated by the friction material of the automatic transmission 30 and data on the engagement frequency of the friction material of the automatic transmission 30 as features. In this data center 500, the search process performed by the processing unit 510 includes a first step (step S120) of calculating the relative frequency distribution in the original data for each of the multiple features included in the original data. The search process includes a second step (step S125) of setting up multiple time windows to extract data from a portion of the original data such that the sum of the periods of all time windows is shorter than a predetermined period. The search process includes a third step (step S130) of extracting data from the original data using the multiple time windows. The search process includes a fourth step (step S140) of calculating the relative frequency distribution for each feature in the extracted data obtained by combining all the data extracted using the multiple time windows. The search process includes a fifth step (step S145) in which the error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data is calculated. After executing the first step, the processing unit 510 executes a search process that repeatedly performs trials from the second to the fifth step by changing the settings of multiple time windows. The processing unit 510 then extracts data in which the error is below a threshold (step S150: YES). The processing unit 510 calculates an index value using the extracted data in which the error is below a threshold (step S160).

[0066] According to this Data Center 500, it is possible to obtain extracted data that captures the characteristics of the entire original data, including multiple features. Therefore, this Data Center 500 can calculate index values ​​with the same accuracy as when using the original data, even though the extracted data has a smaller volume than the original data.

[0067] <Effects of this embodiment> (1) According to the data center 500, which is an information processing device of this embodiment, it is possible to achieve both a reduction in the amount of data and an improvement in the accuracy of calculating index values.

[0068] (2) According to the data center 500, which is an information processing device of this embodiment, the index value can be calculated in a shorter time compared to when using the original data. (3) The processing unit 510 performs clustering, a machine learning technique that classifies the data from each interval of the original data into a predetermined number of clusters (step S110). Then, in the second step of the search process (step S125), the processing unit 510 sets multiple time windows such that the difference between the ratio of each cluster in the extracted data and the ratio of each cluster in the entire original data is less than or equal to a threshold.

[0069] Multiple intervals classified into the same cluster are intervals with similar characteristics. In the search process described above, the settings output from the processing unit 510 are such that the difference in the ratio of the entire original data to each cluster is below a threshold, and the relative frequency distribution of each feature is similar, allowing for the extraction of relevant data.

[0070] Therefore, the search process performed by the data center 500 described above can find settings that allow for the acquisition of extracted data that more closely resembles the characteristics of the entire original data. (4) The processing unit 510 terminates the search process when it has extracted one data point whose error is below the threshold, and calculates an index value using the extracted data point whose error is below the threshold. Therefore, the data center 500 can calculate the index value as soon as it has extracted one data point whose error is below the threshold, and output the results quickly.

[0071] (5) If the calculated indicator value is greater than or equal to a predetermined value (step S170: YES), the processing unit 510 issues a notification indicating that it has predicted the occurrence of a failure. Therefore, the data center 500 can notify the user that a failure has been predicted before the failure occurs.

[0072] (6) The processing unit 510 calculates the degree of fatigue damage as an indicator value. As a result, the data center 500 can inform the user how much time it has before failure occurs.

[0073] <Example of changes> This embodiment can be implemented with the following modifications. This embodiment and the following modifications can be combined with each other to the extent that they do not contradict each other technically.

[0074] As examples of features, the heat generated by the friction material for which the damage index value is calculated and the engagement frequency of the friction material were used as examples, but the index value may also be calculated by including oil temperature data as a feature.

[0075] In the above embodiment, an example was shown in which the information processing device is implemented as a data center 500. An example was shown in which the calculation of the index value is performed in the data center 500. In contrast, the above information processing device may be implemented as an information processing terminal 600. In this case, the calculation of the index value is performed by the processing device 610 of the information processing terminal 600. The above information processing device may also be implemented as a control device of the vehicle 10. In this case, the calculation of the index value can also be performed by the control device of the vehicle 10. For example, the calculation of the index value can also be performed by the transmission control device 50 of the vehicle 10.

[0076] The above embodiment shows an example of extracting one data point and calculating an index value. In contrast, multiple data points may be extracted, and the final index value may be determined using multiple index values ​​calculated from each data point. For example, the minimum value, maximum value, mode, and mean may be used as the final index value. Alternatively, multiple index values ​​may be output.

[0077] • In the above embodiment, an example was shown in which a notification is given that a failure has been predicted when the index value is greater than or equal to a predetermined value. This may be omitted. After calculating the index value, only the processing in step S190 may be executed to output only the index value.

[0078] • While fatigue damage level was used as an example of an indicator value to be calculated, the indicator value to be calculated is not limited to fatigue damage level. • Damage index values ​​for multiple friction materials may be calculated. Alternatively, the relative frequency distribution for each feature of each friction material may be calculated, and a selection pattern may be searched to minimize the error in the relative frequency distribution for all features.

[0079] • For each friction material, a extraction pattern may be searched to minimize the relative frequency distribution of the features. Then, an index value may be calculated for each friction material. The invention can also be applied to transmissions equipped with friction materials other than the automatic transmission 30 shown in the above embodiment, such as the forward / reverse switching mechanism of a stepped automatic transmission or continuously variable transmission installed in a hybrid vehicle, and the clutch of a manual transmission.

[0080] The example shown illustrates how the processing unit 510 sets multiple time windows so that the difference between the ratio of each cluster in the original data and the ratio of each cluster in the extracted data is less than or equal to a threshold. However, the processing unit 510 may set multiple time windows without such constraints. In this case, the clustering step S110 may be omitted.

[0081] The method for determining the time window setting in the cropping pattern does not have to be random. The setting of the time window in the cropping pattern can be changed according to a pre-defined rule, and the trial can be repeated.

[0082] The error calculated in step S145 is not limited to the mean absolute error (MAE). For example, the processing unit 510 may calculate the mean squared error as the error. The processing unit 510 may also calculate the root mean squared error as the error.

[0083] This example demonstrates the use of extracted data created by combining all the extracted data. Alternatively, extracted data can be created by combining only a portion of the extracted data. [Explanation of Symbols]

[0084] 10...Vehicle, 20...Engine, 30...Automatic transmission, 31...Friction engagement device, 32...Hydraulic control circuit, 40...Engine control device, 50...Transmission control device, 80...Communication device, 400...Communication network, 500...Data center, 510...Processing device, 520...Storage device, 530...Communication device, 600...Information processing terminal, 610...Processing device, 620...Storage device, 630...Communication device

Claims

1. This information processing device acquires original data created by collecting data over a predetermined period using multiple sensors mounted on the vehicle, and calculates an index value indicating the magnitude of damage accumulated in the transmission friction material. It includes a processing unit that performs processing, The aforementioned original data includes, as feature quantities, data on the amount of heat generated by the friction material of the transmission and data on the engagement frequency of the friction material of the transmission. The aforementioned processing device The process includes: a first step of calculating the relative frequency distribution in the original data for each of the multiple features contained in the original data; a second step of setting multiple time windows to extract data for a portion of the original data such that the sum of the periods of all time windows is shorter than the predetermined period; a third step of extracting data from the original data using the multiple time windows; a fourth step of calculating the relative frequency distribution in the extracted data obtained by combining the data extracted using the multiple time windows, for each of the features; and a fifth step of calculating the error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data; wherein after executing the first step, the process of repeating the trials from the second to fifth steps by changing the settings of the multiple time windows is executed to extract the extracted data for which the error is less than or equal to a threshold. The process involves calculating the index value using the extracted data in which the error is below a threshold, and performing the following: Information processing device.

2. The processing unit performs clustering, a machine learning technique, which classifies the data from each interval obtained by dividing the original data into fixed period intervals into a predetermined number of clusters. In the second step, the processing device sets the multiple time windows such that the difference between the ratio of each cluster in the extracted data and the ratio of each cluster in the original data as a whole is less than or equal to a threshold. The information processing apparatus according to claim 1.

3. The processing device terminates the search process when it has extracted one data item whose error is below the threshold, and calculates the index value using the extracted data item whose error is below the threshold. The information processing apparatus according to claim 1.

4. If the calculated index value is greater than or equal to a predetermined value, the processing device will issue a notification indicating that a malfunction has been predicted. The information processing apparatus according to claim 1.

5. The processing apparatus calculates the degree of fatigue damage as the index value. The information processing apparatus according to claim 1.

Images