JP2010128661A - Method and apparatus for guessing cause of failure, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of guessing a cause of a failure, which improves efficiency in analyzing a failure cause based on a plurality of logs output from each of apparatuses. <P>SOLUTION: The method of guessing a failure cause is executed by a computer, and includes: a feature amount calculation procedure of synchronizing log data individually output in the plurality of apparatuses mounted on the same vehicle on the basis of output time, and calculating a feature amount of each log data per unit time; a unit time extraction procedure of extracting the unit time to which the feature amount having appearance frequency smaller than a threshold value belongs; a feature amount integration procedure of integrating each feature amount of the extracted unit time and each feature amount of the unit time just before the unit time; a case data extraction procedure of extracting case data most similar to a pattern of the integrated feature amount from a plurality of case data recorded in a storage device in association with the failure cause; and an output procedure of outputting information showing the failure cause associated with the extracted case data. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a failure cause estimation method, a failure cause estimation device, and a program, and more particularly, to a failure cause estimation method, a failure cause estimation device, and a program that analyze a failure cause based on log data.

In order to analyze the cause of failure of various devices mounted on automobiles and the like, logs relating to device operations are recorded. However, when there are many items (variables) recorded in the log or when the log recording time becomes long, the amount of log data increases. As a result, it becomes difficult for humans to analyze a large amount of log data in a short time. Therefore, conventionally, log data is compared with past case data, and similar case data is searched to automatically determine the cause of failure, countermeasures, and the like.
JP 2003-195934 A JP 2008-197912 A

  However, each device mounted on the vehicle is highly independent, and a log is recorded independently for each device. Therefore, there has been a problem that it is difficult to efficiently analyze the cause or the like based on a plurality of independent log data regarding a malfunction in an operation realized by cooperation or cooperation of a plurality of devices.

  The present invention has been made in view of the above points, and is a failure cause estimation method and a failure cause estimation device that can efficiently analyze failure causes based on a plurality of logs individually output from each device. And to provide a program.

  Therefore, in order to solve the above problem, the failure cause estimation method synchronizes the log data output individually in a plurality of devices mounted on the same vehicle according to the output time, and the feature amount of each log data for each unit time. A feature amount calculation procedure for calculating a unit time, a unit time extraction procedure for extracting a unit time to which a feature amount whose appearance frequency is lower than a threshold belongs, each feature amount of the extracted unit time, and a unit time immediately before the unit time Extract the case data that is most similar to the integrated feature value pattern from the feature value integration procedure that integrates each feature value and multiple case data that is recorded in the storage device in association with the cause of failure. And an output procedure for outputting information indicating the cause of failure associated with the extracted case data.

  In such a failure cause estimation method, failure cause analysis based on a plurality of logs individually output from each device can be made more efficient.

  Failure analysis based on a plurality of logs individually output from each device can be made efficient.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram for explaining an outline of a failure cause estimation apparatus according to an embodiment of the present invention.

  The failure cause estimation device 10 is a computer having a function of estimating the cause of failure of a device (on-vehicle device) of the automobile 20. As shown in the figure, the failure cause inference apparatus 10 is based on log data (log file) in which logs that are output independently (individually) from various in-vehicle devices are recorded. Guess. In the present embodiment, the cause of failure refers to a concept including information indicating a failure location. In the present embodiment, an example in which a failed device is estimated from the devices included in the in-vehicle system of the automobile 20 will be described as the cause of the failure.

  FIG. 2 is a diagram illustrating a schematic configuration example of an in-vehicle system mounted on the automobile according to the present embodiment. In the figure, the in-vehicle system 2 includes devices such as an operation panel 21, an air conditioner unit 22, a DVD unit 23, and a TV unit 24. The operation panel 21 and other devices are connected by a communication cable.

  The operation panel 21 is a device for operating devices such as the air conditioner unit 22, the DVD unit 23, and the TV unit 24, and is disposed at a position where it can be operated from the driver's seat of the automobile 20, for example. The operation panel 21 includes a liquid crystal display 211, operation switches 212, and the like. The liquid crystal display 211 is a display device that displays a screen (operation screen or the like) related to the device that is the operation target. The operation switch 212 is a switch for operating the DVD unit 23, for example. Note that although a plurality of operation switches are arranged on the general operation panel 21, they are omitted for convenience of explanation in the present embodiment.

  The air conditioner unit 22 is a so-called air conditioner. The DVD unit 23 is a so-called DVD player. The TV unit 24 is a so-called TV tuner.

  The in-vehicle system 2 will be described in more detail. FIG. 3 is a diagram illustrating a detailed configuration example of the in-vehicle system according to the present embodiment. In FIG. 3, the same parts as those of FIG.

  In the figure, the operation unit 21 includes a liquid crystal display 211, an operation switch 212, a clock 213, a log recording device 214, a processing device 215, a display switch 216, and the like.

  The clock 213 is a so-called clock and measures time. The log recording device 214 is a recording device that records a log related to the operation unit 21. For convenience, it is assumed that the log related to the operation unit 21 in the present embodiment is a log related to the operation on the operation switch 212. That is, the operation history of the operation switch 212 is recorded in the log recording device 214. The processing device 215 executes processing for controlling the operation of the operation unit 21 and communication with other devices based on the program. The processing device 215 also records information related to the operation of the operation switch 212, information embedded in the program, and the like in the log recording device 214 as a log together with the time input from the clock 213. The processing device 215 further transmits a command corresponding to the operation of the operation switch 212 to the DVD unit 23 by CAN (Controller Area Network) communication or switches the display switch 216. The display switch 216 is a switch for switching a device to be displayed on the liquid crystal display 211. In the figure, a state in which the DVD unit 23 is connected is shown. Therefore, in the state shown in FIG. 8, a screen related to the DVD unit 23 is displayed on the liquid crystal display 211.

  The air conditioner unit 22 includes a clock 221, a log recording device 222, a screen output unit 223, a processing device 224, and the like. The clock 221 is a clock. The log recording device 222 is a recording device that records a log related to the air conditioner unit 22. The screen output unit 223 outputs screen display information related to the air conditioner unit 22 to the liquid crystal display 211. The processing device 224 executes processing for controlling the entire air conditioner based on the program. The processing device 224 also records information related to the air conditioner unit 22, information embedded in the program, and the like in the log recording device 222 as a log together with the time input from the clock 231.

  The DVD unit 23 includes a clock 231, a log recording device 232, a screen output unit 233, a processing device 234, and the like. The TV unit 24 includes a clock 241, a log recording device 242, a screen output unit 243, a processing device 244, and the like. Since the function of each component is obvious from the function of the component of the same name in the air conditioner unit 22, the description is omitted.

  As is clear from FIG. 3, in the in-vehicle system 2, logs are recorded and managed individually or independently for each device.

In the figure, an example is shown in which a clock is provided for each device, but the synchronization of each clock is guaranteed. Alternatively, a common timepiece may be provided for a plurality of devices. Next, details of the failure cause estimation device 10 will be described. FIG. 4 is a diagram illustrating a hardware configuration example of the failure cause inference apparatus according to the embodiment of the present invention. 4 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, a display device 106, and an input connected to each other via a bus B. Device 107.

  A program that realizes processing in the failure cause inference apparatus 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 on which the program is recorded is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 implements functions related to the failure cause estimation device 10 in accordance with a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network. The display device 106 displays a GUI (Graphical User Interface) or the like by a program. The input device 107 includes a keyboard and a mouse, and is used for inputting various operation instructions.

  FIG. 5 is a diagram illustrating a functional configuration example of the failure cause estimation apparatus according to the embodiment of the present invention. In the figure, a failure cause estimation apparatus 10 includes a feature amount calculation unit 11, a unit time extraction unit 12, a feature amount integration unit 13, a cause estimation unit 14, an estimation result comparison unit 15, an output unit 16, a known case DB 17, and the like. Have. Each of these units is software realized by processing that the program causes the CPU 101 to execute.

  The feature amount calculation unit 11 synchronizes each log data individually output from each device (operation panel 21, air conditioner unit 22, DVD unit 23, TV unit 24) of the automobile 20 based on the output time (output time). In this embodiment, the feature amount calculation unit 11 calculates the feature amount of each log data every 30 seconds. An example in which the number of log recordings is adopted as the feature amount will be described. However, the feature amount may be calculated based on other information included in the log data, such as the length of the log and the value of the parameter recorded in the log.

  The log data is acquired in advance from the log recording device of each device and recorded in the auxiliary storage device 102. The transfer method of the log data output from each device to the failure cause estimation device 10 (auxiliary storage device 102) is not limited to a predetermined method. It may be via a portable recording medium, or via wireless communication or satellite communication. Alternatively, a cable connected to an ODB2 (On Board Diagnostics Phase 2) connector of the automobile 20 may be used.

  The unit time extraction unit 12 extracts (determines) a unit time (time zone) indicating a combination (pattern) of feature amounts having a relatively low appearance frequency.

  The feature amount integration unit 13 integrates (adds) the extracted feature amounts of the unit time and the feature amounts of the immediately preceding unit time.

  The cause inference unit 14 extracts case data similar to the integrated feature amount pattern from the case data stored in the known case DB 17, and estimates the cause of the failure based on the extracted case data.

  The known case DB 17 is a database that manages a correspondence between a failure cause in a known (for example, past) case and a feature amount pattern related to past log data when the failure occurs. The known case DB 17 is recorded in the auxiliary storage device 102, for example. However, the known case DB 17 may be recorded in another computer or storage device connected via the network.

  When a plurality of case data are extracted by the cause estimation unit 14, the estimation result comparison unit 15 determines the priority (priority order) of the plurality of case data based on the feature amount of the log data that is the extraction source. judge.

  The output unit 16 outputs information indicating the cause of failure estimated by the cause estimating unit 14 to an output device (for example, the display device 106 or a printer (not shown)).

  Hereinafter, the processing procedure of the failure cause estimation apparatus 10 will be described. FIG. 6 is a flowchart for explaining the processing procedure of the failure cause estimating apparatus.

  In step S <b> 101, the feature amount calculation unit 11 reads the log data of each device transferred to the auxiliary storage device 102 into the memory device 103.

  FIG. 7 is a diagram illustrating an example of log data on the operation panel. The log data 210 shown in the figure has items such as date, time, and switch log. The date is the date when the log is output. The time is the time when the log is output. The date and time are acquired from the clock 213. The switch log is a log output according to the operation of the operation switch 212. That is, each time the operation switch 212 is operated, for example, a log (switch log) indicating the operation content is recorded in the log recording device 214 of the operation panel 21 together with the date and time.

  FIG. 8 is a diagram illustrating an example of log data of the air conditioner unit. The log data 220 shown in the figure has items such as date, time, air conditioner log, and state detection log. The date is the date when the log is output. The time is the time when the log is output. The air conditioner log is a log output in response to some operation of the air conditioner unit 22 (regardless of whether it is a normal operation or an abnormal operation). The state detection log is a log that is output to indicate that a predetermined state is detected by the program by the designer of the program that controls the processing device 224 of the air conditioner unit 22. The designer's intention is, for example, the idea that it will be effective for analysis when a failure occurs. The figure shows a state detection log that is output when the air conditioner unit 22 is in the state S1, and a state detection log that is output when the air conditioner unit 22 is in the state S2. The state may be defined by, for example, the value of a variable in the program, or may be defined by the presence or absence of a signal or event from the device.

  FIG. 9 is a diagram illustrating an example of log data of the DVD unit. The log data 230 shown in the figure has items such as date, time, DVD log, and status detection log. The meaning of the date, time, and state detection log is the same as that of the air conditioner log 220. However, the figure shows a state detection log that is output when the DVD unit 23 is in the state P3 and a state detection log that is output when the DVD unit 23 is in the state P4. The DVD log is a log output in response to some operation of the DVD unit 23 (regardless of normal operation or abnormal operation).

  FIG. 10 is a diagram illustrating an example of log data of the TV unit. The log data 240 shown in the figure has items such as date, time, TV log, and status detection log. The meaning of the date, time, and state detection log is the same as that of the air conditioner log 220 and the like. The figure shows a state detection log that is output when the TV unit 24 is in the state P5 and a state detection log that is output when the TV unit 24 is in the state P6. The TV log is a log output according to some operation of the TV unit 24 (regardless of normal operation or abnormal operation).

  In step S101, the log data shown in FIGS. 7 to 10 is read.

  Subsequently, the feature amount calculation unit 11 synchronizes the log data 210 to 240 with time (output time) (S102). Conceptually, each log data is arranged with the time axis aligned.

  Subsequently, the feature amount calculation unit 11 calculates a feature amount (number of outputs) for each unit time of 30 seconds for each of the log data 210 to 240 (S103). Here, since the output time of each log data is synchronized, the unit time of each log data becomes the same time zone (the start time and the end time coincide). As a result, feature amount data as shown in FIG. 11 is generated in the memory device 103.

  FIG. 11 is a diagram showing the number of times each log data is output per unit time. The feature amount data 250 shown in the figure includes date, start time, end time, switch log output count, air conditioner log output count, DVD log output count, TV log output count, and status detection log output count (S1 to S6). Have items.

  The start time and end time are the start time or end time of each unit time. The switch log output count, the air conditioner log output count, the DVD log output count, and the TV log output count are the respective log output counts per unit time. In the figure, the air-conditioner log output count, DVD log output count, and TV log output count are limited to the output count of the log (retry log) that is output when a retry is performed, not all logs. Has been. This is because retry execution is considered to be closely related to failure. However, what log feature amount (output count) is calculated may be appropriately selected in consideration of the contents of the collected logs. The retry means that a certain operation fails and the same operation is repeated. Whether the log is a retry log is determined based on the contents of the log. That is, it is recorded in the log that retry is being executed.

  The state detection log output count is the output count of each state detection log (S1 to S6) per unit time.

  Subsequently, the unit time extraction unit 12 calculates the appearance frequency of the combination (pattern) of the log output count for each unit time based on the feature amount data 250, and the appearance frequency is relatively low (for example, a predetermined value) The unit time to which the pattern (lower than the threshold value) belongs is extracted (S103). For example, in the feature amount data 250, the pattern of the first record is “0000000”, and the appearance frequency of the pattern is relatively high. On the other hand, the pattern of the second record is “0100100000”, and the appearance frequency of the pattern is relatively low. Therefore, the second record is an extraction target.

  As a result, the unit time (time zone) shown in FIG. 12 is extracted. FIG. 12 is a diagram illustrating an example of a unit time extracted by the unit time extraction unit.

  In the figure, this is a unit time for extracting a row where 1 is an extraction target. In the figure, seven unit times are extracted. The first extraction target is as described above. The pattern of the number of output times of the second extraction target is “0010001100”. The pattern of the number of output times of the third extraction target is “0000000100”. The pattern of the number of output times of the fourth extraction target is “3010000000”. The pattern of the number of output times of the fifth extraction target is “0001000010”. The pattern of the number of output times of the sixth extraction target is “0200010000”. The pattern of the number of output times of the seventh extraction target is “0001000011”. Since the appearance frequency is relatively low, each is selected as an extraction target.

  Note that, if the number of times of output of any log is 1 or a predetermined value or more, the log may be extracted.

  The unit time extracted above is handled as a time zone that may indicate a failure state. This is because the fact that the frequency of appearance of the pattern of the feature amount of the log in a certain time zone is low can be presumed to be highly likely to be in a state different from the normal state (normal state).

  Subsequently, the feature amount integration unit 13 integrates (sums) the output count of the extracted log of unit time and the output count of the log of the previous unit time (S104). As a result, the number of log outputs per minute is calculated.

  FIG. 13 is a diagram illustrating an example of a result of integration of the number of outputs. In the figure, “1” is set for the unit time to be integrated with respect to the extraction target. In addition, 261 to 267 include the result of integrating (summing) the log output counts of unit time to be integrated with respect to the output count of each log of the extracted unit time (that is, the log output count in 1 minute units). )It is shown. Hereinafter, a log output frequency pattern (combination) indicated by each of the integration results 261 to 267 is referred to as an “integration pattern”.

  Subsequently, the cause inference unit 14 searches for case data similar to the integrated pattern from the case data registered in the known case DB 17 for each integrated pattern (S104).

  FIG. 14 is a diagram illustrating a configuration example of a known case DB. As shown in the figure, in the known case DB 17, for each known case, the number of output of the state detection log in one minute unit and the cause of failure (failure location) are managed in association with each other. For example, the record of classification number 1 indicates that the cause of failure is the air conditioner unit 22 when the pattern of the number of times the state detection log is output per minute is “100,000”.

  Here, whether or not the integrated pattern and the case data are similar is determined by calculating the similarity between the two. A specific example of the similarity is a distance (Euclidean distance) between the integrated pattern and the case data. The Euclidean distance between one integrated pattern and one example data can be calculated by the following formula.

ｄは、ユークリッド距離である。ａ_ｉは、統合パターンの状態検知ログ出力回数におけるｉ番目の変数の値である。ｂ_ｉは、事例データのｉ番目の変数の値である。ｎは、変数の数（次元数）である。本実施の形態では、ｎは６である。また、ａ_１、ａ_２、ａ_３、ａ_４、ａ_５、ａ_６は、それぞれ統合パターンのＳ１、Ｓ２、Ｓ３、Ｓ４、Ｓ５、Ｓ６の値である。また、ｂ_１、ｂ_２、ｂ_３、ｂ_４、ｂ_５、ｂ_６は、それぞれ事例データのＳ１、Ｓ２、Ｓ３、Ｓ４、Ｓ５、Ｓ６の値である。

d is the Euclidean distance. a _i is the value of the i-th variable in the number of times of output of the integrated pattern state detection log. b _i is the value of the i-th variable of the case data. n is the number of variables (number of dimensions). In the present embodiment, n is 6. Further, a ₁ , a ₂ , a ₃ , a ₄ , a ₅ and a ₆ are values of S 1, S 2, S 3, S 4, S 5 and S ₆ of the integrated pattern, respectively. B ₁ , b ₂ , b ₃ , b ₄ , b ₅ , b ₆ are values of S 1, S 2, S 3, S 4, S 5, S ₆ of the case data, respectively.

  The cause inference unit 14 extracts the case data that is most similar for each integrated pattern (the Euclidean distance is short), and acquires the cause of failure associated with the case data.

  As a result, the cause of failure shown in FIG. 15 is estimated based on each integrated pattern. FIG. 15 is a diagram illustrating an example of a failure cause estimation result based on each integration pattern.

  According to the figure, it is estimated that the air conditioner unit 22 is a cause of failure based on the integrated pattern 261. Further, it is presumed that there is no cause of failure based on the integrated patterns 262 and 266. Further, it is presumed that the cause of failure is the DVD unit 23 based on the integrated patterns 263 and 264. Further, it is presumed that the cause of the failure is the TV unit 24 based on the integrated pattern 267.

  Subsequently, the estimation result comparison unit 15 determines the priority (priority order) of the estimation result by comparing the output counts of the logs of the plurality of integrated patterns (S106). In the present embodiment, the higher the switch log output count, the higher the priority is given. The switch log is output in response to a human operation. A person is likely to repeat the same or related operations if the behavior of the device is different than expected. For example, when the operation switch 212 is operated and the device does not operate as expected, the operation is performed again for about 30 seconds or the related switch is operated. Furthermore, only a person can always determine that a device failure is different from the expected operation. In other words, it is possible to determine the failure to some extent when it is in the state assumed in advance by the program, but it is not possible to determine the failure in the case of an unexpected state. Therefore, in the present embodiment, the highest priority is given to the integrated pattern having the largest number of switch log outputs.

  Therefore, in the example of FIG. 15, the integrated data 264 is selected. As a result, the cause estimating unit 14 determines that the most likely failure cause is the DVD log output count.

  Note that the priority of the integrated data or the cause of the failure may be determined based on other methods. For example, the failure cause having the largest number of estimation results may be given the highest priority. The priority may be determined based on the number of times other logs are output.

  By the way, in the present embodiment, the feature amount of the log data is calculated in units of 30 seconds. When the failure occurs, the related operation is repeated for about 30 seconds, and then gives up and recognizes the failure. Based. In addition, the reason for integrating the feature value in 30-second units into 1-minute units is that there is a high possibility that a state showing a sign of failure has occurred at a stage before a suspected failure is detected by a person. is there. That is, it can be said that the integrated data in 1 minute units includes information indicating a failure sign and information indicating the failure itself.

  Note that 30 seconds is a guide only. Therefore, for example, the unit time may be less than 30 seconds or may exceed 30 seconds. However, considering the operation while traveling, the unit time is preferably 1 minute or less. This is because it is considered that there are few cases where the operation is continued for more than one minute for safety of driving.

  Subsequently, the output unit 16 causes the display device 106 to display information indicating the cause of failure estimated by the cause estimating unit 14 (S107). As in the present embodiment, when a plurality of failure causes are estimated, the output unit 16 can identify the priorities determined by the estimation result comparison unit 15 (for example, in descending order of priority). Display information indicating the cause of failure. Note that the output unit 16 may output the start time and the end time of the unit time to which the integrated data that is the cause of the failure cause belongs. Moreover, you may make it display the location applicable to the said start time and end time in the log data 210-240. By doing so, log data useful for analyzing the cause of failure can be presented to the user from a large amount of log data.

  As described above, the failure cause inference apparatus 10 according to the present embodiment synchronizes each log data output individually with the output time, and calculates a feature value for each unit time considering human behavior. Also, by extracting (selecting) a unit time that is suspected of failure based on the feature amount, and integrating the extracted feature amount of the unit time with the feature amount of the previous unit time, a sign of failure can be obtained. An integrated pattern including information to be shown is generated. Further, the cause of the failure is estimated by comparing the integrated pattern with the case data.

  Therefore, it is possible to efficiently analyze log data output individually from a plurality of devices. As a result, the cause of the failure can be estimated in a short time.

  In addition to the Euclidean distance, an inner product, a cosine correlation value, or the like of the integrated pattern (a) and the case data (b) may be used as the similarity. The inner product may be calculated by the following equation.

また、コサイン相関値は、次の式によって算出すればよい。

The cosine correlation value may be calculated by the following equation.

各変数の意味は、ユークリッド距離の式において説明した通りである。

The meaning of each variable is as described in the Euclidean distance equation.

  Further, the present invention can be effectively applied to the analysis of the cause of failure of devices other than the in-vehicle device. It can also be effectively applied to the estimation of the cause of occurrence for events other than failures.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

Regarding the above description, the following items are further disclosed.
(Appendix 1)
A failure cause estimation method executed by a computer,
A feature amount calculation procedure for synchronizing log data individually output by a plurality of devices mounted on the same car according to an output time, and calculating a feature amount of each log data for each unit time,
A unit time extraction procedure for extracting a unit time to which a feature amount whose appearance frequency is lower than a threshold value belongs;
A feature amount integration procedure for integrating each extracted feature amount of the unit time and each feature amount of the unit time immediately before the unit time;
Case data extraction procedure for extracting case data most similar to the integrated feature amount pattern from among a plurality of case data recorded in the storage device in association with the cause of failure,
A failure cause estimation method comprising: an output procedure for outputting information indicating a failure cause associated with the extracted case data.
(Appendix 2)
The failure cause inference method according to attachment 1, wherein the feature amount is a log output count.
(Appendix 3)
When a plurality of case data is extracted in the case data extraction procedure, the priority determination procedure for determining the priority of the plurality of case data based on the feature amount corresponding to the case data,
The failure cause estimation method according to appendix 1 or 2, wherein the output procedure outputs information indicating the cause of the failure according to the priority.
(Appendix 4)
The failure cause inference method according to appendix 3, wherein the priority determination procedure includes determining a priority of the plurality of case data based on a feature amount of a log output by a human operation among the feature amounts corresponding to the case data. .
(Appendix 5)
A feature amount calculating means for synchronizing log data individually output by a plurality of devices mounted on the same car according to an output time, and calculating a feature amount of each log data per unit time;
A unit time extraction means for extracting a unit time to which a feature quantity whose appearance frequency is lower than a threshold value belongs;
Feature quantity integration means for integrating each extracted feature quantity of unit time and each feature quantity of unit time immediately before the unit time;
Case data extraction means for extracting the case data most similar to the integrated feature amount pattern from the plurality of case data recorded in the storage device in association with the cause of failure,
A failure cause inference device comprising: output means for outputting information indicating a cause of failure associated with extracted case data.
(Appendix 6)
The failure cause inference apparatus according to appendix 5, wherein the feature amount is a log output count.
(Appendix 7)
When a plurality of case data is extracted by the case data extracting means, the apparatus includes a priority determining means for determining the priority of the plurality of case data based on the feature amount corresponding to the case data.
The failure cause estimation device according to appendix 5 or 6, wherein the output means outputs information indicating the cause of the failure according to the priority.
(Appendix 8)
The failure cause inference device according to appendix 7, wherein the priority determination means determines the priority of the plurality of case data based on a feature amount of a log output by a human operation among feature amounts corresponding to the case data. .
(Appendix 9)
On the computer,
A feature amount calculation procedure for synchronizing log data individually output by a plurality of devices mounted on the same car according to an output time, and calculating a feature amount of each log data for each unit time,
A unit time extraction procedure for extracting a unit time to which a feature amount whose appearance frequency is lower than a threshold value belongs;
A feature quantity integration procedure for integrating each extracted feature quantity of the unit time and each feature quantity of the unit time immediately before the unit time;
Case data extraction procedure for extracting case data most similar to the integrated feature amount pattern from among a plurality of case data recorded in the storage device in association with the cause of failure,
The program for performing the output procedure which outputs the information which shows the failure cause matched with the extracted example data.
(Appendix 10)
The program according to appendix 9, wherein the feature amount is a log output count.
(Appendix 11)
When a plurality of case data is extracted in the case data extraction procedure, the priority determination procedure for determining the priority of the plurality of case data based on the feature amount corresponding to the case data,
The program according to appendix 9 or 10, wherein the output procedure outputs information indicating the cause of the failure according to the priority.
(Appendix 12)
The program according to appendix 11, wherein the priority determination procedure determines the priority of the plurality of case data based on a feature amount of a log output by a human operation among feature amounts corresponding to the case data.

It is a figure for demonstrating the outline | summary of the failure cause estimation apparatus in embodiment of this invention. It is a figure which shows the example of schematic structure of the vehicle-mounted system mounted in the motor vehicle of this Embodiment. It is a figure which shows the detailed structural example of the vehicle-mounted system of this Embodiment. It is a figure which shows the hardware structural example of the failure cause estimation apparatus in embodiment of this invention. It is a figure which shows the function structural example of the failure cause estimation apparatus in embodiment of this invention. It is a flowchart for demonstrating the process sequence of a failure cause estimation apparatus. It is a figure which shows the example of the log data of an operation panel. It is a figure which shows the example of the log data of an air-conditioner unit. It is a figure which shows the example of the log data of a DVD unit. It is a figure which shows the example of the log data of TV unit. It is a figure which shows the output frequency per unit time of each log data. It is a figure which shows the example of the unit time extracted by the unit time extraction part. It is a figure which shows the example of the integration result of the frequency | count of output. It is a figure which shows the structural example of known case DB. It is a figure which shows the example of the estimation result of the failure cause based on each integrated pattern.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Failure cause estimation apparatus 11 Feature quantity calculation part 12 Unit time extraction part 13 Feature quantity integration part 14 Cause estimation part 15 Prediction result comparison part 16 Output part 17 Known case DB
20 Car 21 Operation Panel 22 Air Conditioner Unit 23 DVD Unit 24 TV Unit 100 Drive Device 101 Recording Medium 102 Auxiliary Storage Device 103 Memory Device 104 CPU
105 Interface device 106 Display device 107 Input device 211 Liquid crystal display 212 Operation switches 213, 221, 231, 241 Clock 214, 222, 232, 242 Log recording devices 215, 224, 234, 244 Processing device 216 Display switches 223, 233, 243 Screen output part B bus

Claims (6)

A failure cause estimation method executed by a computer,
A feature amount calculation procedure for synchronizing log data individually output by a plurality of devices mounted on the same car according to an output time, and calculating a feature amount of each log data for each unit time,
A unit time extraction procedure for extracting a unit time to which a feature amount whose appearance frequency is lower than a threshold value belongs;
A feature amount integration procedure for integrating each extracted feature amount of the unit time and each feature amount of the unit time immediately before the unit time;
Case data extraction procedure for extracting case data most similar to the integrated feature amount pattern from among a plurality of case data recorded in the storage device in association with the cause of failure,
A failure cause estimation method comprising: an output procedure for outputting information indicating a failure cause associated with the extracted case data.   The failure cause estimation method according to claim 1, wherein the feature amount is a log output count. When a plurality of case data is extracted in the case data extraction procedure, the priority determination procedure for determining the priority of the plurality of case data based on the feature amount corresponding to the case data,
The failure cause inference method according to claim 1, wherein the output procedure outputs information indicating the cause of the failure according to the priority.   The failure cause inference according to claim 3, wherein the priority determination procedure determines the priority of the plurality of case data based on a feature value of a log output by a human operation among feature values corresponding to the case data. Method. A feature amount calculating means for synchronizing log data individually output by a plurality of devices mounted on the same car according to an output time, and calculating a feature amount of each log data per unit time;
A unit time extraction means for extracting a unit time to which a feature quantity whose appearance frequency is lower than a threshold value belongs;
Feature quantity integration means for integrating each extracted feature quantity of unit time and each feature quantity of unit time immediately before the unit time;
Case data extraction means for extracting the case data most similar to the integrated feature amount pattern from the plurality of case data recorded in the storage device in association with the cause of failure,
A failure cause inference device comprising: output means for outputting information indicating a cause of failure associated with extracted case data. On the computer,
A feature amount calculation procedure for synchronizing log data individually output by a plurality of devices mounted on the same car according to an output time, and calculating a feature amount of each log data for each unit time,
A unit time extraction procedure for extracting a unit time to which a feature amount whose appearance frequency is lower than a threshold value belongs;
A feature amount integration procedure for integrating each extracted feature amount of the unit time and each feature amount of the unit time immediately before the unit time;
Case data extraction procedure for extracting case data most similar to the integrated feature amount pattern from among a plurality of case data recorded in the storage device in association with the cause of failure,
The program for performing the output procedure which outputs the information which shows the failure cause matched with the extracted example data.

Images