JP2009145298A - Failure pattern estimation method, failure pattern estimation device, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a failure pattern estimation method capable of estimating a failure pattern that is likely to occur in a vehicle, with high accuracy. <P>SOLUTION: The failure pattern estimation method includes a step of obtaining a diagnostic code from a failed vehicle; a step of estimating a failure pattern that occurs in the failed vehicle, with a high probability, on the basis of failure association table expressing the probability that each diagnostic code is recorded in the vehicle, at the occurrence of failure for each failure pattern and the diagnosis code obtained from the failed vehicle; and a step of outputting the estimated failure pattern. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a technique for accurately estimating a disconnection location of an in-vehicle LAN.

  When the vehicle breaks down, each ECU performs a failure diagnosis and stores the diagnosis result in a memory as a diagnosis code. The repair worker connects the failure diagnosis tool to the vehicle diagnosis connector and acquires the diagnosis code stored in the memory. Based on the acquired diagnosis code, the repair operator refers to a table that associates the diagnosis code with the failure location (hereinafter referred to as failure allocation table), identifies the failure location, and repairs the identified failure location. Do.

In the failure allocation table, what kind of diagnostic code each ECU outputs is recorded for each location where the in-vehicle LAN may be disconnected.
The failure allocation table is created by causing a failure in an actual vehicle on the test bench and acquiring a diagnostic code stored in each ECU. For this reason, when the size of the in-vehicle LAN increases, the number of failure modes increases, installation of large-scale evaluation equipment, placement of multiple workers with specialized knowledge, lengthening of processing time, and erroneous diag codes due to human error Several problems arise, such as acquisition.

  In recent years, with the improvement of computer-related technology, a system for testing a cooperative operation between a plurality of in-vehicle electrical components by a simulation using a computer has also been proposed (for example, see Patent Document 1).

JP 2005-181113 A

However, it is very difficult to specify a failure location as one location, and it is general to detect several locations where it can be determined that there is a possibility of failure.
For example, since a communication line such as an in-vehicle LAN is shared by a plurality of ECUs, even if a failure occurs at one location, the influence of the failure extends over a plurality of locations. For this reason, each ECU outputs various diagnostic codes, and it becomes difficult to specify the failure location.
Further, the disclosed technique of Patent Document 1 is a system for testing a cooperative operation between in-vehicle electrical components, and is not a system for estimating a fault location of the in-vehicle LAN.

  The present invention has been made in view of the above circumstances, and provides a failure pattern estimation method, a failure pattern estimation device, and a program capable of estimating with high accuracy a failure pattern that is highly likely to occur in an in-vehicle LAN. The purpose is to do.

In order to achieve such an object, the failure pattern estimation method of the present invention includes a step of acquiring a failure code from a failed vehicle, a management table that represents, for each failure pattern, a probability that each failure code is recorded in the vehicle when a failure occurs, and The method includes a step of estimating a failure pattern having a high probability of occurring in the failed vehicle based on a failure code acquired from the failed vehicle, and a step of outputting the estimated failure pattern.
As described above, according to the present invention, a failure pattern that is highly likely to occur in the in-vehicle LAN can be estimated with high accuracy using the management table.

In the failure pattern estimation method, the management table performs a computer simulation for each failure pattern, and the failure code detected by the control device when each failure pattern occurs and the probability that the failure code is recorded in the vehicle. It is characterized by being created by calculating for each of a plurality of mounted control devices.
Since the management table is created by computer simulation for each failure pattern, the management table can be created easily in a short time.

In the failure pattern estimation method, the software that performs the computer simulation constructs a virtual network that simulates the plurality of control devices mounted on a vehicle and a communication network that connects the plurality of control devices, and the virtual network A simulation of the communication state of the frame data within the virtual network, and a simulation of causing a failure in the virtual network and monitoring a failure code detected by the plurality of control devices. It is characterized by being performed at every data communication timing.
An accurate monitoring result can be obtained by simulating the communication state at every communication timing of the frame data.

In the failure pattern estimation method, as a communication network state actually identified by a plurality of control devices mounted on a vehicle, a normal first state, a second state in which an error is detected, an error And the third state in which the plurality of control devices are logically disconnected from the bus, but the virtual network constructed by the software omits the second state. It is a feature.
Therefore, it is possible to monitor only the state related to permission / stop of data transmission.

In the failure pattern estimation method, the step of estimating the failure pattern is characterized in that the order of failure patterns not including a failure code acquired from the failed vehicle is lowered.
Further, in the failure pattern estimation method, the step of estimating the failure pattern is characterized in that the order of failure patterns including a failure code that could not be acquired from the failed vehicle is lowered.
Therefore, the failure pattern can be estimated with high accuracy.

  The failure pattern estimation apparatus according to the present invention is generated for each failure pattern based on a management table representing a probability that each failure code is recorded in the vehicle when a failure occurs and a failure code acquired from the failure vehicle. In addition, the configuration includes an estimation unit that estimates a failure pattern with a high probability of being output, and an output unit that outputs the estimated failure pattern.

  The program of the present invention generates a computer in the failed vehicle based on a management table representing a probability that each failure code is recorded in the vehicle when a failure occurs and a failure code acquired from the failed vehicle for each failure pattern. It is characterized by functioning as a means for estimating a failure pattern having a high probability and a means for outputting the estimated failure pattern.

  According to the present invention, it is possible to estimate a failure pattern that is highly likely to occur in an in-vehicle LAN with high accuracy.

  Preferred embodiments of the present invention will be described with reference to the accompanying drawings.

  This embodiment has two steps. The first step is a failure assignment table (management table) in which a diagnosis code (fault code) recorded in a memory when a failure occurs is recorded for each failure pattern by a plurality of ECUs (Control Units) mounted on the vehicle. ). In the second step, a diagnosis code is acquired from the failed vehicle using a failure diagnosis tool or the like, and a failure pattern having a high probability of occurring in the failed vehicle is obtained based on the acquired diagnosis code and the failure allocation table. This is an estimation process. The estimated failure patterns are displayed on the display unit in descending order of occurrence probability.

First, the first step will be described in detail. Note that the first step and the second step are performed by software processing by the computer apparatus 1.
FIG. 1 shows the processing functions of the computer apparatus 1 realized by software control in blocks for each function. The virtual LAN simulation unit 20 and the result analysis unit 30 are realized by software control.

First, the virtual LAN simulation unit 20 is configured to construct a virtual network that simulates a plurality of ECUs mounted on a vehicle and an in-vehicle LAN that connects the plurality of ECUs.
Data necessary for constructing an in-vehicle LAN includes LAN routing, a format of communication data communicated with the in-vehicle LAN, a communication diagnosis specification, and the like. These data are input from the operation unit 11 of the interface unit 10.
The LAN routing is information indicating how a plurality of ECUs mounted on a vehicle are arranged on a network. The ECU has a self-diagnosis function for diagnosing what kind of failure has occurred when a failure (disconnection) occurs, but the communication diagnosis specification is a specification of the self-diagnosis function that the ECU has. Is defined.
In order to construct the virtual LAN 21, a virtual bus setting 22 and a virtual ECU setting 23 are required. The virtual bus setting 22 includes setting of a communication protocol used in the simulation (CAN (Controller Area Network) protocol is used in this embodiment), and the virtual ECU setting 23 includes self-diagnosis at the time of failure. The function includes a communication diagnosis and setting of CAN frame data transmission / reception function of each ECU.

When the above data is input and the virtual LAN is constructed in the virtual LAN simulation unit 20, the simulation parameters necessary for the actual simulation are input and the simulation conditions are set.
The simulation parameters include virtual LAN communication time T, number of simulations M, error occurrence probability X, time resolution S, and the like.
The virtual LAN communication time T is a time setting for performing a simulation by causing a failure on the virtual LAN. The number of simulations is the number of times of simulation. If the number of failure patterns occurring on the virtual LAN is N (N is an arbitrary natural number) and the number of simulations is M (M is an arbitrary natural number), the virtual LAN simulation unit 20 Repeats M × N simulations.
The error occurrence probability X is a probability of virtually generating an error (noise generated due to the effect of disconnection) on the virtual LAN. It is set as the failure model shown in FIG. 1, and an error is generated in the virtual LAN with an arbitrary probability. The time resolution S is a unit of time for performing simulation and monitoring the result. In this embodiment, the time resolution S is performed in accordance with the minimum time interval at which the ECU can continuously transmit frame data. The diag code recorded by the ECU is monitored.
Although the detection of the diag code is performed for each ECU, the diag code may be recorded for each ECU, or may be recorded collectively in a diag recorder or a master ECU.

There are two failure modes to be simulated by the virtual LAN simulation unit 20, the CAN communication line one-side disconnection mode and the CAN communication line complete disconnection mode.
In this embodiment, CAN is used as a communication protocol between ECUs, and the virtual LAN is connected by two communication lines. In the CAN communication line one-side disconnection mode in which one of the two communication lines is disconnected, the ECU using the communication line that has been disconnected on one side as a branch line cannot transmit / receive CAN frame data. In addition, the virtual LAN simulation unit 20 randomly changes a CAN frame transmitted on a communication line that is disconnected on one side to an error frame.
In the CAN communication line complete disconnection mode, the ECU that uses the communication line that has been completely disconnected as a branch line cannot transmit or receive CAN frame data. In addition, all ECUs connected to a completely disconnected communication line cannot be transmitted / received.

In the present embodiment, a device for simplifying the simulation and shortening the simulation time is provided. That is, the CAN protocol is not completely reproduced in the virtual LAN simulation unit 20, but only functions necessary for diagnostic evaluation are reproduced.
<Simplification of CAN frame>
FIG. 2A shows the configuration of a CAN frame that is actually transmitted and received by the CAN protocol, and FIG. 2B shows the configuration of a CAN frame used for simulation on the virtual LAN.
The CAN frame actually transmitted / received by the CAN protocol shown in FIG. 2A includes ID (Identification), DLC (Data Length Code), and actual data. However, in failure diagnosis, the presence / absence of CAN frame data transmission / reception is used as a determination condition, so the CAN frame length and actual data can be ignored.
Therefore, the CAN frame shown in FIG. 2B in which only the ID is recorded is used with the CAN frame length fixed.
<Simulation time concept>
By simplifying the CAN frame, it is not necessary to express data in bit units, so the time resolution in simulation is set low.
When the CAN frame was reproduced for each bit, the time resolution required 2 μs at a communication speed of 500 Kbps, but the shortest frame length was about 100 μs at a communication speed of 500 Kbps, so the time resolution was half that time. If it is set to 50 μs, there is no problem in operation.
That is, if the diagnostic code recorded by the ECU is monitored in accordance with the interval at which the ECU transmits data, a highly accurate monitoring result can be obtained.
<Simplification of error status>
In the diagnostic evaluation, in order to perform failure diagnosis based on whether or not a CAN frame is transmitted / received, only error active related to transmission permission, transmission stop, and bus off are defined as error statuses. The error active state is a state in which communication can be normally performed, and the bus off state is a state in which the ECU is completely disconnected from the communication LAN.
In addition to this, the CAN protocol that is actually performed in the vehicle has a state of error passive. The error-basis state is a state in which an abnormality has occurred in the LAN but communication is not stopped.

FIG. 3 shows an example of the failure allocation table.
In the failure allocation table, for each failure pattern, the occurrence probability of a diagnostic code stored in the ECU when the failure pattern occurs is recorded.
In the failure assignment table shown in FIG. 3, diagnostic codes output by each ECU are recorded. In addition, a failure pattern and a probability of occurrence of a diagnostic code stored in each ECU when a failure occurs are recorded.
For example, in the example shown in FIG. 3, when failure pattern 1 occurs, ECU_A stores diag code [U0100] with a probability of 80% and diag code [U0101] with a probability of 90%. Yes. Similarly, ECU_B indicates that when the failure pattern 1 occurs, the diagnosis code [U0200] is stored with a probability of 80%, and the diagnosis code [U0201] is stored with a probability of 100%. The ECU_C stores the diagnosis code [U0300] with a probability of 100% when the failure pattern 2 occurs, stores the diagnosis code [U0301] with a probability of 0%, and stores the diagnosis code [U0302] with a probability of 100%. It shows that

FIG. 4 shows an outline of creation of the failure allocation table. A fault is generated in the LAN for each failure pattern, and the diagnosis code recorded by each ECU at that time and the occurrence probability of the diagnosis code are recorded.
For example, assume that 100 simulations are performed for one failure pattern. Since the probability of occurrence in the failure pattern 1 of the diagnosis code [U0100] in the failure assignment table shown in FIG. 3 is 80%, the ECU_A indicates that 80 times [U0100] of the diagnosis code is recorded in 100 simulations. ing.
The log analysis function unit 31 shown in FIG. 2 acquires and analyzes the simulation result (log) from the virtual LAN simulation unit 20, and creates the failure assignment table shown in FIG. The created failure assignment table is displayed on the display unit 12 of the interface unit 10.

The procedure for creating the failure allocation table will be described with reference to the flowchart shown in FIG.
First, virtual LAN formation information, which is data for recognizing the configuration of a virtual LAN to be simulated, is input to the virtual LAN simulation unit 20 (step S1). The virtual LAN formation information includes LAN routing, communication data, communication diagnosis specifications, and the like. By inputting these pieces of information, a virtual LAN is formed in the virtual LAN simulation unit 20 (step S2).

  Next, the virtual LAN simulation unit 20 extracts a portion where a failure may occur in the formed virtual LAN, and calculates the number N of failure patterns from the extracted failure occurrence portion (step S3). The number N of failure patterns is a number N obtained by adding a case where a failure occurs in one place of the virtual LAN and a case where a plurality of places fail simultaneously.

  Next, the virtual LAN simulation unit 20 inputs simulation parameters (step S4). Simulation parameters include virtual LAN communication time T, number of simulations M, error occurrence probability X, time resolution S, and the like. These parameters are parameters set by the user.

Next, the virtual LAN simulation unit 20 performs simulation by causing a pseudo failure on the virtual LAN. If the number of failure patterns is N and the number of simulations is M, simulations are performed N × M times for all failure patterns (step S5).
When the virtual LAN simulation unit 20 executes N × M simulations (step S6 / YES), the virtual LAN simulation unit 20 analyzes the simulation log and creates a failure assignment table (step S7).

Next, a simulation procedure (details of step S5 described above) by the virtual LAN simulation unit 20 will be described with reference to the flowchart shown in FIG.
When the simulation by the virtual LAN simulation unit 20 is started, first, the timer t for measuring the virtual LAN communication time T set by the user is reset to 0 and measurement of the communication time is started (step S11).
Next, the virtual LAN simulation unit 20 sets variables so that an error occurs in the virtual LAN with the error occurrence probability X set by the user (step S12).

Next, the virtual LAN simulation unit 20 determines whether to determine the state of the virtual LAN. When the state of the virtual LAN is not an idle state in which data transmission is possible (step S13 / NO), the ECU that transmits the CAN frame stops the data transmission, and the ECU that receives the CAN frame receives the data Is stopped (step S14). When the virtual LAN state is an idle state in which data transmission is possible (step S13 / YES), the virtual LAN simulation unit 20 determines whether there is an ECU having a transmission request (step S15).
When there is no ECU with a transmission request (step S15 / NO), the virtual LAN simulation unit 20 proceeds to an error status determination process in step S23.
If there is an ECU having a transmission request (step S15 / YES) and there are a plurality of ECUs having a transmission request, the ECU that performs transmission is determined by arbitration (step S16). When there is one ECU with a transmission request, the ECU is set as the ECU that performs transmission as it is.

An ECU (hereinafter simply referred to as a transmission ECU) set in a transmission ECU that transmits a CAN frame transmits the CAN frame to another ECU (hereinafter referred to as a reception ECU) on the virtual LAN (step S17). The receiving ECU performs a CAN frame reception process of the CAN frame (step S18). The receiving ECU determines whether or not the reception of the CAN frame is completed (step S19). When the reception of the CAN frame is completed (step S19 / YES), an ACK (ACKnowledgement) indicating the completion of reception is returned to the transmission ECU.
The receiving ECU that has transmitted ACK clears the reception interval counter to zero (step S21). Further, when receiving the ACK from all the receiving ECUs, the transmitting ECU stops transmitting CAN frame data and clears the transmission interval counter to zero (step S22).

Next, the virtual LAN simulation unit 20 determines an error status (step S23). The error status determination process will be described separately using a flowchart. Further, the virtual LAN simulation unit 20 performs a communication diagnosis determination process (step S24). This communication diagnosis determination process will also be described using a separate flowchart.
When the above processing is completed, the virtual LAN simulation unit 20 refers to the count value of the timer t that measures the virtual LAN communication time T (step S25). If the count value of the timer t is not equal to the time T (t = T) set by the user (step S25 / NO), the process returns to step S11 and is repeated from the determination of the bus state.
If the count value of the timer t is the time T (t = T) set by the user (step S25 / YES), the simulation in the virtual LAN is terminated.

Next, the procedure of processing for determining the state of an error that has occurred in the virtual LAN (error status determination processing) will be described with reference to the flowchart shown in FIG.
The virtual LAN simulation unit 20 determines what state each ECU is in when an error occurs.
When the ECU state is error active (step S31 / YES), the transmission error counter is referred to (step S32). The error active state indicates a state where transmission / reception of a CAN frame is normally performed. The transmission error counter is a counter that is incremented every time a transmission error occurs. If the count value of the transmission error counter is 256 or more (step S32 / YES), the ECU error status is set to bus off (step S33). If the count value of the transmission error counter is smaller than 256 (step S32 / NO), the process is terminated as it is.

  In addition, when the error status of the ECU is not error active and is in the bus off state (step S31 / NO), the virtual LAN simulation unit 20 determines whether or not recovery from the bus off state is possible (step S31). S34). If it is possible to recover from the bus-off (step S34 / YES), the error status of the ECU is changed to error activation (step S35). If it is not possible to recover from the bus off (step S34 / NO), this process ends.

Next, the determination process of the diag code recorded in the ECU will be described with reference to the flowchart shown in FIG.
First, the virtual LAN simulation unit 20 determines whether or not the reception interval counter exceeds the threshold value (step S41). The reception interval counter is a counter for determining whether the CAN frame has been received. If a CAN frame is received, this reception interval counter is cleared to zero. If no CAN frame is received, the reception interval counter is incremented, and when the count value exceeds the threshold value, the virtual LAN simulation unit 20 records the corresponding diagnostic code.
If the count value of the reception interval counter exceeds the threshold value (step S41 / YES), the virtual LAN simulation unit 20 records a diagnosis code (reception interruption diagnosis code) indicating that reception has been interrupted (reception interruption diagnosis code). Step S42).

Next, the virtual LAN simulation unit 20 determines whether or not the transmission interval counter exceeds the threshold value (step S43). The transmission interval counter is a counter for determining transmission interruption of the CAN frame.
When the CAN frame transmission operation is started, the transmission interval counter is also counted up. When the transmission of the CAN frame data is completed, the transmission interval counter is cleared to zero. If ACK is returned to the receiving ECU and transmission of the CAN frame data is not completed, the transmission interval counter continues to count up.
When the count value of the transmission interval counter exceeds the threshold value (step S43 / YES), the virtual LAN simulation unit 20 records a diagnosis code (transmission interruption diagnosis code) indicating that transmission has been interrupted (transmission interruption diagnosis code) (step S43 / YES). Step S44).

  Next, the virtual LAN simulation unit 20 determines whether or not a bus off has occurred (step S45). When the status of the error status flag indicating the error status is determined and it is determined that a bus-off has occurred (step S45 / YES), a diagnosis code (bus-off determination diagnosis code) indicating the occurrence of the bus-off is recorded (step S46). .

  As described above, in this embodiment, the diagnostic evaluation work can be performed by a computer. For this reason, the time required for the diagnosis evaluation can be shortened. In addition, a large-scale evaluation facility such as an actual vehicle is not required, and an actual vehicle measurement vehicle is not required.

Next, the second step of estimating a fault location that has occurred in an actual vehicle using the created allocation table will be described.
In the second step, first, a diagnosis code is acquired from a vehicle in which a failure has occurred using a failure diagnosis tool. A diagnostic code is obtained from a vehicle that has been damaged and is stored in a store or the like by using a failure diagnosis tool.

Next, the diagnosis code acquired using the failure diagnosis tool is input to the computer apparatus 1, and the failure location is estimated using the diagnosis code acquired from the vehicle and the failure allocation table. In FIG. 9, the processing functions of the computer apparatus 1 realized by software control are represented by blocks for each function. The failure assignment logic unit 40 is built in the computer apparatus 1 by software control.
A pattern matching method is used to estimate the failure location of the failure allocation logic unit 40. In the pattern matching method, evaluation points are added for each failure pattern based on the diagnosis code acquired from the vehicle, and ranking is performed for each failure pattern. Evaluation points are added using a scoring table shown in FIG.
The scoring table shown in FIG. 10 is divided into six categories.
Classification A is a case where a diagnosis code can be acquired from an actual vehicle (hereinafter referred to as an actual vehicle) that identifies a failure location, and the probability of occurrence of the diagnosis code in the failure allocation table is 100%. [+1.0] is assigned as a value.
Class B is a case where a diagnosis code can be acquired from an actual vehicle, and the probability of occurrence of a diagnosis code in the failure assignment table is 00.1% to 99.9%. In this case, the score value is [+0.5 ] Is given.
Class C is a case where a diagnosis code cannot be obtained from an actual vehicle, and the diagnosis code generation probability in the failure assignment table is 0%. In this case, [0.0] is assigned as a scoring value.
Class D is a case where a diagnosis code cannot be obtained from an actual vehicle, and the probability of occurrence of a diagnosis code in the failure assignment table is 00.1% to 99.9%. In this case, the score value [0. 0].
Class E is a case where a diag code cannot be obtained from the actual vehicle and the diag code generation probability of the failure allocation table is 100%. In this case, [−0.5] is assigned as a scoring value.
Classification F is a case where a diagnosis code can be obtained from an actual vehicle, and the probability of occurrence of a diagnosis code in the failure assignment table is 0%. In this case, [−1.0] is assigned as a scoring value.

Next, a method for estimating a failure location using a diagnosis code acquired from a vehicle and a failure allocation table will be specifically described with reference to FIG.
The failure allocation logic unit 40 inputs the diagnosis code acquired from the failure diagnosis tool and the data of the failure allocation table, and estimates the failure location by the pattern matching method.
For each failure pattern, the location of the failure assignment table in which the diagnosis code acquired from the failure diagnosis tool is recorded is detected, and a score is assigned with reference to the scoring table.
The failure allocation logic unit 40 first searches the failure allocation table as shown in FIG. 11 and detects a diagnosis code (hereinafter referred to as a processing target diagnosis code) acquired from the failed vehicle. When the processing target diagnosis code is detected from the failure allocation table, the generation probability of the processing target diagnosis code is acquired for each failure pattern. Next, referring to the scoring table, the scoring value when the processing target diagnosis code is recorded in each failure pattern is obtained.
For example, when the failure pattern 1 occurs, the probability that the ECU_A stores the diagnosis code [U0100] is 80%, so [0.5] points are given from the scoring table.
In addition, when the failure pattern 2 occurs, the probability that the ECU_A stores the diagnosis code [U0100] is 100%, so [1.0] points are assigned from the scoring table.
It should be noted that a diagnosis code that could not be obtained from a vehicle by using a failure diagnosis tool and that has a probability of occurrence of 100% in the failure assignment table (ie, category E) has [−0.5] points. Is granted. Further, when the diag code acquired from the vehicle using the failure diagnosis tool has an occurrence probability of 0% in the failure assignment table (that is, classification F), [−1.0] points are assigned. .

When a score is assigned to each diag code, the failure assigning logic unit 40 adds the points assigned to the diag code for each failure pattern. The added total score is used as a failure pattern evaluation score.
When the failure location estimation software calculates the evaluation points for each failure pattern, the failure location estimation software arranges the failure patterns in descending order of the score, and causes the display unit 12 to display the estimation results.

  FIG. 12 shows an example of the display displayed on the display unit 12. In the display example shown in FIG. 12, the configuration of the virtual LAN and the top three candidates (failure location candidates) determined to have a possibility of failure are displayed.

The processing procedure of the failure assignment logic unit 40 will be described with reference to the flowchart shown in FIG.
The failure allocation logic unit 40 inputs, as input data, a diagnosis code acquired from a failed vehicle using a failure diagnosis tool and a failure allocation table (step S51).
Next, the failure allocation logic unit 40 estimates a failure pattern from the processing target diagnosis code acquired from the failed vehicle and the failure allocation table. First, as shown in FIG. 11, a failure allocation table is searched to detect a processing target diagnosis code acquired from a failed vehicle (step S52). When the processing target diagnosis code is detected from the failure allocation table, the generation probability of the processing target diagnosis code is acquired for each failure pattern (step S53). Next, referring to the scoring table, the scoring value when the processing target diagnosis code is recorded in each failure pattern is obtained (step S54).
It should be noted that a diagnosis code that could not be obtained from a vehicle by using a failure diagnosis tool and that has a probability of occurrence of 100% in the failure assignment table (ie, category E) has [−0.5] points. Is granted. Further, when the diag code acquired from the vehicle using the failure diagnosis tool has an occurrence probability of 0% in the failure assignment table (that is, classification F), [−1.0] points are assigned. .
The above process is repeated for all the diagnosis codes acquired from the failed vehicle (step S55). When the processing is completed for all diag codes acquired from the failed vehicle (step S55 / YES), the failure allocation logic unit 40 adds the points given to the diag codes for each failure pattern (step S56). The added total score is used as a failure pattern evaluation score. Then, the failure allocation logic unit 40 detects the top three failure patterns with the highest evaluation points (step S57). The detected failure pattern is displayed on the display unit 12.

As described above, according to the present embodiment, the failure location can be estimated using only the computer device. Since all processing is performed by a computer, the failure location can be estimated in a few hours.
In addition, since it is possible to narrow down the part where the failure has occurred by the computer device, it is possible to reduce the time for specifying the failure part.
Furthermore, since a failure part can be narrowed down accurately, generation | occurrence | production of problems, such as exchanging normal ECU, can be reduced.

The hardware configuration of the computer apparatus 1 will be described with reference to FIG.
The computer apparatus 1 includes a CPU 51, a ROM 52, a RAM 53, an NVRAM (Non Volatile RAM) 54, an input / output unit 55, and the like. The CPU 51 reads a program stored in the ROM 52 and performs a calculation according to this program. That is, when the CPU 51 reads the simulation program stored in the ROM 52, the virtual LAN simulation unit 20 and the result analysis unit 30 shown in FIG. 1 and the failure assignment logic unit 40 shown in FIG. 9 are configured. Further, the data of the calculation result is written in the RAM 33, and the NVRAM 34 is the data that has been written in the RAM 33, and the data that needs to be saved when the power is turned off.

  The embodiment described above is a preferred embodiment of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the scope of the present invention.

It is a figure which shows the functional block implement | achieved by the computer apparatus by software processing, and is a figure which shows the functional block at the time of failure allocation table preparation especially. (A) shows the configuration of a CAN frame that is actually transmitted and received by the CAN protocol, and (B) is a diagram showing the configuration of a CAN frame used for simulation on a virtual LAN. It is a figure which shows an example of a failure allocation table. It is a figure which shows the preparation outline of a failure allocation table. It is a flowchart which shows the procedure for failure allocation table | surface preparation. It is a flowchart which shows the procedure of the simulation in virtual LAN. It is a flowchart which shows the procedure of an error status determination process. It is a flowchart which shows the determination procedure of a diag code. It is a figure which shows the functional block implement | achieved by the computer apparatus by software processing, and is a figure which shows the functional block at the time of the estimation of a failure pattern especially. It is a figure which shows an example of a scoring table. It is a figure for demonstrating the estimation method of a failure pattern. It is a figure which shows the example of a display of the estimated failure pattern. It is a flowchart which shows the process sequence of a failure allocation logic part. It is a figure which shows the hardware constitutions of a computer apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Computer apparatus 10 Interface part 11 Operation part 12 Display part 20 Virtual LAN simulation part 30 Result analysis part 40 Fault allocation logic part 51 CPU
52 ROM
53 RAM
54 NVRAM
55 I / O section

Claims (8)

Obtaining a fault code from the faulty vehicle;
Based on the management table representing the probability that each failure code is recorded in the vehicle when a failure occurs for each failure pattern and the failure code acquired from the failed vehicle, a failure pattern having a high probability of occurring in the failed vehicle Estimating, and
Outputting an estimated failure pattern;
A failure pattern estimation method characterized by comprising:   The management table performs a computer simulation for each failure pattern, and includes a plurality of control devices mounted on the vehicle with a failure code detected by the control device when each failure pattern occurs and the probability that the failure code is recorded The fault pattern estimation method according to claim 1, wherein the fault pattern estimation method is created by calculating each time. The software for performing the computer simulation constructs a virtual network that simulates the plurality of control devices mounted on a vehicle and a communication network that connects the plurality of control devices,
While simulating the communication state of the frame data in the virtual network, causing the virtual network to simulate a failure, monitoring the failure code detected by the plurality of control devices,
The failure pattern estimation method according to claim 2, wherein the communication state simulation is performed for each communication timing of the frame data. As the state of the communication network that is actually identified by a plurality of control devices mounted on the vehicle, the first state that is normal, the second state that detects the occurrence of an error, and the error are detected continuously. A third state in which the plurality of control devices are logically disconnected from the bus,
4. The fault pattern estimation method according to claim 3, wherein the second state is omitted in the virtual network constructed by the software.   2. The failure pattern estimation method according to claim 1, wherein the step of estimating the failure pattern lowers the rank of failure patterns not including a failure code acquired from the failed vehicle.   2. The fault pattern estimation method according to claim 1, wherein the step of estimating the fault pattern lowers the rank of fault patterns including fault codes that could not be acquired from the faulty vehicle. For each failure pattern, based on a management table representing the probability that each failure code is recorded in the vehicle when a failure occurs and a failure code acquired from the failed vehicle, a failure pattern having a high probability of occurring in the failed vehicle Estimating means for estimating;
An output means for outputting the estimated failure pattern;
A failure pattern estimation apparatus comprising: Computer
For each failure pattern, based on a management table representing the probability that each failure code is recorded in the vehicle when a failure occurs and a failure code acquired from the failed vehicle, a failure pattern having a high probability of occurring in the failed vehicle Means to estimate;
A program that functions as means for outputting an estimated failure pattern.

Images