Downsampling method

Abstract

To accurately perform downsampling of running data.SOLUTION: A downsampling method of running data with machine learning includes the steps of: acquiring the running data; calculating a predetermined feature amount on the running data; setting a time length or a position of a window so that the rate of the running data to each of running regions is almost equal; producing cutout data from the running data using the window; calculating an error between appearance frequency of the feature amount on the running data and appearance frequency of the feature amount on the cutout data; and selecting the cutout data so that the error is a predetermined value or less.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a downsampling method. [Background technology]

[0002] If travel data is collected while a vehicle is traveling and stored, it can be used for various information processing. Travel data stored over a long period of time will be a huge amount of data, so there is a demand for a technology to compress the travel data. For example, Patent Document 1 discloses an invention that compresses travel data by outputting vehicle speed data at the time when the vehicle speed becomes equal to a set vehicle speed and at the time when the change in vehicle speed becomes an inflection point. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2008-108247 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the conventional technology, the accuracy of the vehicle speed data after compression is not sufficient. In order to efficiently compress the driving data, it is desirable to shorten the sampling time as much as possible, but if the sampling time is too short, there is a possibility that the accuracy will decrease.

[0005] In view of the above-mentioned technical problems, one aspect of the present disclosure aims to perform downsampling of driving data with high accuracy. [Means for solving the problem]

[0006] A downsampling method according to one aspect of the present disclosure includes the steps of acquiring driving data, calculating predetermined features for the driving data, setting the time length or position of a window so that the proportion of driving data for each driving area is approximately the same, generating cut-out data from the driving data using the window, calculating the error between the frequency of occurrence of the features of the driving data and the frequency of occurrence of the features of the cut-out data, and selecting the cut-out data so that the error is equal to or less than a predetermined value. Effect of the Invention

[0007] According to one aspect of the present disclosure, downsampling of driving data can be performed with high accuracy. [Brief description of the drawings]

[0008] [Figure 1] 1 is a flowchart illustrating an example of a downsampling method. [Diagram 2] FIG. 11 is a diagram illustrating an example of a feature amount. [Diagram 3] 1 is a flowchart illustrating an example of a failure symptom detection method. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functional configurations are denoted by the same reference numerals, and redundant description will be omitted.

[0010] [Embodiment] An embodiment of the present disclosure is a downsampling method for driving data using artificial intelligence (AI). In this embodiment, the downsampling method calculates a predetermined feature amount for the entire driving data, compares the appearance frequency of the feature amount of the entire driving data with the appearance frequency of the feature amount of cut-out data cut out from the driving data, and selects cut-out data such that an error in the appearance frequency is equal to or less than a predetermined value.

[0011] In the downsampling method, when extracting cut-out data from the driving data, a window is used in which at least one of the time length and the position is set so that the proportion of driving data for each driving region is approximately the same. The driving region may be set by the vehicle speed and the required driving force, or may be set by the vehicle speed and the acceleration. The window may be set by fixing the total window time, or may be set after extending the total window time. At this time, when the error in the occurrence frequency is large, the proportion of the time to be extended may be increased. The window may be set so that the error in the sum of the feature amounts is equal to or less than a predetermined value.

[0012] Various indices can be used for the frequency of occurrence error, such as the root mean squared error (RMSE), the mean squared error (MSE), the mean absolute error (MAE), the mean absolute percentage error (MAPE), etc.

[0013] <Driving data> In this embodiment, the processing target is driving data acquired from a drive unit used in a hybrid vehicle. This type of drive unit is disclosed in, for example, Reference 1.

[0014] [Reference 1] JP 2008-232287 A

[0015] The drive device disclosed in Reference 1 is configured as follows. The drive device has a plurality of multi-plate clutches. Each multi-plate clutch has a cancel chamber. A plurality of multi-plate clutches may share one cancel chamber.

[0016] Among the multiple multi-plate clutches, one multi-plate clutch is configured to have a clutch drum, a clutch piston, a clutch hub, a return spring, and a cancel plate. The clutch drum is a bottomed cylindrical member connected to the first intermediate shaft so as to rotate integrally, and a plurality of ring-shaped friction plates are splined to its inner circumferential surface so as to rotate integrally. The clutch piston is disposed in the clutch drum, and forms a hydraulic chamber between the clutch drum and itself. The clutch hub has a cylindrical portion facing the inner circumferential surface of the cylindrical portion of the clutch drum, and is a cylindrical member connected to the second intermediate shaft so as to rotate integrally, and a plurality of ring-shaped friction mating plates are splined to its outer circumferential surface so as to rotate integrally.

[0017] The driving data is output from a control device (ECU; Electronic Control Unit) that controls the operation of the drive device. Data such as vehicle speed, access opening, shift position, oil temperature of the power transmission mechanism, and mode select switch are input to the control device. The control device controls the operation of the drive device based on the input data. For example, the torque of the motor generator, the electric oil pump, line pressure, and engagement pressure are output from the control device. Multiple engagement pressures are output corresponding to each friction material in the drive device.

[0018] In this embodiment, the travel data is uploaded to an external storage device via a wireless communication network such as a mobile phone network. The external storage device may be, for example, a cloud storage service provided by cloud computing. Travel data from various vehicles is uploaded to the external storage device, and big data is constructed.

[0019] <Downsampling method> The downsampling method in this embodiment will be described with reference to Fig. 1. Fig. 1(A) is a flowchart showing an example of the downsampling method.

[0020] The downsampling method in this embodiment is executed by an information processing device realized by a computer, etc. The information processing device includes a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), a HDD (Hard Disk Drive), an input device, an output device, an external I / F (Interface), and a communication I / F, each of which is interconnected by a system bus.

[0021] In step S101, the information processing device acquires driving data acquired in each driving mode. The driving mode may be, for example, an urban mode, a highway mode, a mountain mode, etc.

[0022] In step S102, the information processing device combines the driving data of each driving mode acquired in step S101 to generate one driving data. Hereinafter, the combined driving data will be referred to as "total data." When evaluating parts, it is necessary to create an evaluation mode taking into account all of the ways in which each part is used, so it is necessary to combine the driving data of each driving mode. The purpose of the evaluation may be, for example, calculation of a damage rate.

[0023] In step S103, the information processing device performs machine learning and clustering on the entire data generated in step S102.

[0024] In step S104, the information processing device calculates a predetermined feature amount (explanatory variable) for the entire data clustered in step S103. The explanatory variable may be, for example, the number of engagements of the friction material for shifting, the engagement torque, the oil temperature, the relative rotation speed during idling, etc.

[0025] In step S105, the information processing device extracts N pieces of data with a time series length of TL from the entire data by using a window with a variable time length. Hereinafter, the extracted running data will be referred to as "extracted data".

[0026] In step S106, the information processing device calculates the frequency of appearance of the feature amount calculated in step S104 for each piece of extracted data extracted in step S105.

[0027] In step S107, the information processing device calculates the error between the occurrence frequency of the feature of the entire data and the occurrence frequency of the feature of the cut-out data. Next, the information processing device determines whether the error is less than a predetermined value. If the error is less than the predetermined value (YES), the information processing device proceeds to step S108. On the other hand, if the error is equal to or greater than the predetermined value (NO), the information processing device returns to step S105.

[0028] After returning to step S105, the information processing device changes at least one of the window length TL or the window position so that the ratio of driving data for each driving area becomes approximately the same, and cuts out N pieces of data again. In this way, the information processing device repeats cutting out cut-out data while adjusting at least one of the window length TL or the window position until the error in the occurrence frequency of the feature becomes less than a predetermined value.

[0029] In step S108, the information processing device selects the N pieces of cut data cut in step S105 as compressed driving data.

[0030] FIG. 1B is a flowchart showing an example of the data extraction process (step S105 in FIG. 1A).

[0031] In step S111, the information processing device determines whether the counter Sd is greater than a predetermined value S1. If the counter Sd is greater than the predetermined value S1 (YES), the information processing device proceeds to step S115. On the other hand, if the counter Sd is equal to or less than the predetermined value S1 (NO), the information processing device proceeds to step S112.

[0032] In step S112, the information processing device fixes (maintains) the total time of the window.

[0033] In step S113, the information processing device changes at least one of the time length of each window or the position of each window so that the proportion of driving data for each driving area becomes approximately the same. The time length of each window may be uniform or may differ. The information processing device increases the number of windows N as the counter Sd increases. For example, after N, the number of windows is set to N+1. Here, since the total time of the windows is fixed, the time length of each window becomes shorter. In step S114, 1 is added to the counter Sd.

[0034] In step S115, the information processing device determines whether the counter Sb is greater than the predetermined value S2. If the counter Sb is greater than the predetermined value S2 (YES), the information processing device proceeds to step S119. On the other hand, if the counter Sb is equal to or less than the predetermined value S2 (NO), the information processing device proceeds to step S116.

[0035] In step S116, the information processing device extends the total time of the window by X%, where X is a predetermined positive number. In this embodiment, the sampling time cannot be shortened, but this is a constraint that should be tolerated in order to reduce the error. The percentage X% by which the total time is changed may be set according to the magnitude of the error. For example, when the error is large, X can be set to a large value.

[0036] In step S117, the information processing device changes at least one of the time length of each window or the position of each window so that the proportion of driving data for each driving area is approximately the same. The information processing device increases the number of windows N as the counter Sb increases. Here, since the total window time has been extended, a new window corresponding to the extended time is set. In step S118, the information processing device adds 1 to the counter Sb.

[0037] In step S119, the information processing device ends the downsampling method.

[0038] An example of the feature amount in this embodiment will be described with reference to Fig. 2. Fig. 2(A) is a diagram showing an example of the entire data. In the entire data shown in Fig. 2(A), the horizontal axis represents time and the vertical axis represents the engagement torque of the friction material. In this example, the cut-out data is cut out in windows W1 to W4 shown in Fig. 2(A).

[0039] Fig. 2(B) is a diagram showing an example of the feature quantities of the entire data and the feature quantities of the cut-out data. In Fig. 2(B), the occurrence frequency of the feature quantities of the entire data and the feature quantities of the cut-out data for the engagement torque is compared. For each feature quantity, the horizontal axis represents the value of the feature quantity, and the vertical axis represents the frequency. The smaller the error in the distribution of the occurrence frequency of the feature quantities between the entire data and the cut-out data, the more similar the features of the entire data and the cut-out data are.

[0040] FIG. 2(C) is a diagram showing a first example of the distribution of the total data. Since the actual total data is in the tens of thousands of data, 100 pieces of data are shown in one plot in FIG. 2(C). The total data shown in FIG. 2(C) has the required driving force (representative accelerator opening) on ​​the vertical axis and the vehicle speed on the horizontal axis. In FIG. 2(C), the data is divided into nine driving regions A11-A33, but the number of regions can be set arbitrarily. When setting each window, for example, when shortening the time to 1 / 20, the proportion of data in each driving region A11-A33 is the same, and the time is set to 1 / 20.

[0041] Fig. 2(D) is a diagram showing a second example of the distribution of driving data. In the overall data shown in Fig. 2(D), the vertical axis represents acceleration and the horizontal axis represents vehicle speed. In Fig. 2(D), the data is divided into 15 driving regions A41 to A83, but the number of regions may be set arbitrarily. Fig. 2(C) and Fig. 2(D) may be multiplied to divide the data into 135 driving regions.

[0042] <Method for detecting signs of failure> 3 is a flowchart showing an example of a failure sign detection method according to this embodiment. The failure sign detection method is an example of information processing to which the downsampling method according to this embodiment is applied.

[0043] In step S201, the information processing device acquires driving data of a vehicle to be detected from big data. In step S202, the information processing device extracts the driving data acquired in step S201 using the downsampling method (see FIG. 1) in this embodiment.

[0044] In step S203, the information processing device calculates a predetermined damage rate Ss based on the cut-out data cut in step S202. Since the cut-out data is compact due to downsampling compared to the original driving data, the calculation of the damage rate Ss can be performed in a short time.

[0045] Next, the information processing device multiplies the damage rate Ss by a predetermined magnification to calculate the current damage rate Sr. The predetermined magnification can be calculated by (actual driving time or driving distance) / (driving time or driving distance of the cut-out data).

[0046] In step S204, the information processing device displays the current damage rate Sr calculated in step S203.

[0047] In step S205, the information processing device compares the current damage rate Sr with a specified value and determines whether the current damage rate Sr is significantly larger than the travel distance or travel time. If the damage rate Sr is equal to or greater than the specified value (YES), the information processing device proceeds to step S206. On the other hand, if the damage rate Sr is less than the specified value (NO), the information processing device skips steps S206 to S207 and ends the failure sign detection method.

[0048] In step S206, the information processing device determines that a malfunction symptom has been detected. In step S207, the information processing device notifies the user of the vehicle of the malfunction symptom. Specifically, the information processing device notifies the user of the vehicle that a malfunction is likely to occur in the near future.

[0049] <Effects> The downsampling method in this embodiment selects the cut-out data so that the error between the frequency of occurrence of the feature of the entire driving data and the frequency of occurrence of the feature of the cut-out data is equal to or less than a predetermined value. The cut-out data selected by the downsampling method has similar characteristics to the entire driving data. Therefore, according to this embodiment, downsampling of the driving data can be performed with high accuracy. In one aspect, according to this embodiment, a failure sign can be detected with high accuracy.

[0050] Although the embodiments of the present invention have been described in detail above, the present invention is not limited to these embodiments, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.

Claims

[Claim 1] Acquiring driving data; Calculating a predetermined feature amount for the running data; Setting a time length or a position of the window so that the ratio of the traveling data for each traveling area is approximately the same; generating cut-out data from the traveling data using the window; calculating an error between an appearance frequency of the feature amount of the traveling data and an appearance frequency of the feature amount of the cut-out data; selecting the extracted data so that the error is equal to or smaller than a predetermined value; The downsampling method has the following structure:

Images