Electrical component abnormality determination method and device for vehicle

Abstract

To diagnose abnormality of an AI model 3 itself by an electrical component abnormality determination device that uses the AI model 3 having learned a reference current waveform to determine whether respective electrical components are abnormal from the power current of a battery 1.SOLUTION: An electrical component abnormality determination device determines whether respective electrical components are abnormal from the power current of a battery 1 through an AI model 3. A speaker 4 is selected as a reference electrical component for abnormality diagnosis by the AI model 3, and a test signal is input. A sound that a microphone 5 in a cabin detects is compared by a first comparison part 6 with the test signal to confirm the normal operation of the speaker 4. A second comparison part 7 compares a determination result of the first comparison part 6 with a determination result of the AI model 3. If the first comparison part 6 makes a normality determination and the AI model 3 makes an abnormality determination, it is diagnosed that the AI model 3 itself is abnormal.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to electrical equipment that receives power from a battery, and to a vehicle electrical equipment abnormality determination technology that detects abnormalities in each electrical equipment based on the current waveform of the power supply current using an AI model that has learned the reference current waveform when each electrical equipment is operating normally, and in particular to an electrical equipment abnormality determination that checks whether the AI ​​model is abnormal prior to determining whether the electrical equipment is abnormal. [Background technology]

[0002] Patent Document 1, a previous application filed by the present applicant, discloses a technology for determining abnormalities in various electrical components of a vehicle based on the battery's power supply current. Specifically, for multiple electrical components powered by the vehicle's battery, the power supply current flowing from the battery is acquired, and an AI model that has learned the reference current waveform when each electrical component is operating normally is used to detect abnormalities in each electrical component based on the current waveform. This technology makes it relatively easy to determine abnormalities in each electrical component while the electrical components are installed in the vehicle. While Patent Document 1 does not use the term "AI model," the part that identifies electrical components from the current waveform and determines their abnormality corresponds to a so-called AI model. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-100889 Summary of the Invention [Problem to be solved by the invention]

[0004] Naturally, such abnormality determination of electrical equipment is premised on the AI ​​model functioning normally, including whether the amount of learning data is sufficient, etc. However, Patent Document 1 does not mention how to determine whether the AI ​​model is normal or abnormal. [Means for solving the problem]

[0005] The present invention relates to a method for determining abnormality in a vehicle's electrical equipment, which relates to a plurality of electrical equipment supplied with power from a vehicle battery, and which acquires a power supply current flowing from the battery and detects abnormality in each electrical equipment based on the current waveform using an AI model that has learned a reference current waveform when each electrical equipment is operating normally, determining one or more reference electrical components included in the electrical components for the purpose of diagnosing abnormalities of the AI ​​model, and providing a second normal / abnormal determination means for the reference electrical components that is independent of the AI ​​model; The reference electrical equipment is operated, and an abnormality determination is made using the AI ​​model based on the current waveform at that time, and an abnormality determination is made using the second normality / abnormality determination means, If the determination result by the second normality / abnormality determination means is normal and the determination result using the AI ​​model is abnormal, the AI ​​model is deemed to be abnormal.

[0006] The second normal / abnormal determination means may be of any type, but in a vehicle with many electrical components, it is possible to confirm that a certain electrical component, such as a microphone for a speaker, is operating by using other electrical components or sensors that the vehicle originally has. If such a second normal / abnormal determination means determines that the reference electrical component is operating normally, and the determination result using the AI ​​model is abnormal, then the AI ​​model itself can be suspected to be abnormal. [Effects of the Invention]

[0007] According to this invention, if the AI ​​model that determines an abnormality in an electrical component is itself abnormal, this can be easily detected, thereby improving the reliability of abnormality determination in an electrical component based on power supply current. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a block diagram showing a first embodiment of an electrical equipment abnormality determination device according to the present invention; [Figure 2] 4 is a flowchart showing the flow of processing in the first embodiment. [Figure 3] FIG. 4 is a block diagram showing a second embodiment of the electrical equipment abnormality determination device. DETAILED DESCRIPTION OF THE INVENTION

[0009] An embodiment of the present invention will be described in detail below with reference to the drawings. Fig. 1 is a block diagram showing a first embodiment of an electrical equipment abnormality determination device according to the present invention, and in particular, shows only the essential parts of the present invention. As disclosed in Patent Document 1, the basic electrical equipment abnormality determination device is configured to, for a plurality of electrical equipment supplied with power from a vehicle battery 1, acquire the power supply current flowing from the battery 1 using a current sensor 2, and detect an abnormality in each electrical equipment based on the current waveform using an AI model 3 that has learned the reference current waveform when each electrical equipment is operating normally.

[0010] This invention determines whether the AI ​​model 3 itself is abnormal. In a preferred embodiment, the speaker 4 in the vehicle audio system is defined as the reference electrical component, and a receiver, i.e., a microphone 5, that detects the sound of the speaker 4 is preselected as the operation detector that detects the operation of the speaker 4. The microphone 5 is a microphone installed in the vehicle cabin for voice input to give some kind of command. It may also be a microphone installed in the vehicle cabin that constitutes a hands-free phone. The first comparator 6 determines whether the speaker 4 is operating by comparing the signal input to the speaker 4 with the signal obtained from the microphone 5.

[0011] In this embodiment, a second normality / abnormality discrimination means for discriminating between normality and abnormality of the speaker 4, which is the reference electrical equipment, is constituted by a microphone 5 that detects the sound when the speaker 4 is operating, and a first comparison unit 6 that compares the input signal with the microphone 5, which detects the sound when the speaker 4 is operating, without relying on the AI ​​model 3.

[0012] In other words, the reference electrical component is determined based on whether the normal operation of the electrical component can be determined using devices (electrical components, sensors, etc.) that are generally provided in a vehicle. In other words, an electrical component that can be used to establish a second normal / abnormal determination means using devices that are generally provided in a vehicle can be the reference electrical component.

[0013] More specifically, if speaker 4 is the reference electrical component, a test signal (test current) is input to speaker 4 to output a sound of a specific frequency (preferably an inaudible sound). If speaker 4 is operating normally, a sound of a frequency corresponding to the test signal is output from speaker 4, and microphone 5, located in the same vehicle cabin, detects this sound and outputs a corresponding signal. First comparator 6 compares the frequencies of the input signal and the output signal, and if the two match or are similar, outputs a signal indicating that speaker 4 is normal. If the two do not match, this means that the operation of speaker 4, the reference electrical component, cannot be confirmed. Therefore, as described below, it is desirable not to perform an abnormality diagnosis of AI model 3 and to change the reference electrical component as necessary. The output of first comparator 6, i.e., the normality determination result by the second normality / abnormality determination means, is input to second comparator 7.

[0014] On the other hand, the speaker 4, which is the reference electrical component, is subjected to an abnormality determination by the AI ​​model 3 according to the method disclosed in Patent Document 1. Specifically, in the case of the speaker 4, an appropriate test current is input to the speaker 4, similar to the determination by the second normal / abnormal determination means. At this time, the power supply current flowing from the battery 1 is detected by the current sensor 2 and input to the AI ​​model 3. In one embodiment, the input section of the AI ​​model 3 decomposes the power supply current waveform into current waveforms corresponding to each of the multiple electrical components, including the speaker 4. The AI ​​model 3 has learned the reference current waveform when the speaker 4 is operating normally, and determines whether the speaker 4 is operating normally or abnormal based on the input current waveform. Note that in this embodiment, the process of decomposing the current waveforms into those for each electrical component is considered to be included in the AI ​​model 3, but the process of decomposing the current waveforms into those for each electrical component may be considered to be performed separately from the AI ​​model 3.

[0015] The output of the AI ​​model 3, i.e., the normal / abnormal judgment result by the AI ​​model 3, is input to the second comparison unit 7 and compared with the judgment result by the second normal / abnormal judgment means (output of the first comparison unit 6). If both judgment results indicate that the speaker 4, which is the reference electrical component, is normal, the AI ​​model 3 diagnoses that there is no abnormality. In this case, the process proceeds to abnormality diagnosis of each electrical component using the AI ​​model 3 as disclosed in Patent Document 1.

[0016] On the other hand, if the determination result by the second normality / abnormality determination means determines that the speaker 4 is normal and the determination result by the AI ​​model 3 indicates that the speaker 4 is abnormal, the AI ​​model 3 is deemed to be abnormal, and a signal indicating this is output by the second comparison unit 7. When the AI ​​model 3 diagnoses an abnormality in this way, for example, a warning light is turned on, and abnormality diagnosis of each electrical component using the AI ​​model 3 is prohibited.

[0017] 2 is a flowchart showing the flow of processing in the first embodiment, which will be described below. This processing is executed, for example, immediately after the vehicle's main switch (the so-called ignition switch) is turned on (before the AI ​​model 3 starts diagnosing electrical equipment). Alternatively, it may be executed repeatedly at appropriate times during the trip.

[0018] In the first step, step 1, the operation of the reference electrical component is started. In the case of speaker 4, as described above, a test signal is input to operate speaker 4. In step 2, it is determined whether the reference electrical component is operating. In the case of speaker 4, as described above, operation is determined when microphone 5 detects sound from speaker 4. If normal operation cannot be confirmed (shown as NG in the flowchart), the process returns to step 1 after determining the number of retries in step 3, and the operation of the reference electrical component and the process of confirming its operation are repeated.

[0019] If the number of retries reaches a predetermined number N without confirming the normal operation of the reference electrical equipment, the process proceeds from step 3 to step 4 to determine whether the reference electrical equipment has already been changed. If the reference electrical equipment has already been changed, some kind of abnormality (e.g., a malfunction of the microphone 5 or other abnormality) is suspected, so the process proceeds to the system maintenance process shown as step 15 without diagnosing the AI ​​model 3. If the reference electrical equipment has not yet been changed, the process proceeds from step 4 to step 5 to change the reference electrical equipment. Then, steps 1 to 3 are performed again. Changing the reference electrical equipment means changing the electrical equipment used as the reference electrical equipment to another electrical equipment. For example, if there are multiple speakers, a second normal / abnormal determination means can be constructed by using the same microphone 5 or a different microphone for another speaker as the reference electrical equipment. The combination of "electrical equipment and operation confirmation means" may be changed to a different combination from the "speaker and microphone" combination.

[0020] Once the operation of the reference electrical component is confirmed in step 2 (indicated as OK in the flowchart), the operation of the reference electrical component is stopped, and the diagnosis of the AI ​​model 3 begins (steps 6 and 7). Specifically, in step 8, the operation of the reference electrical component is started again for the diagnosis of the AI ​​model 3. In the case of the speaker 4, a test signal is input to operate the speaker 4 as described above. In step 9, the power supply current of the battery 1 is acquired, and an abnormality determination of the speaker 4, which is the reference electrical component, by the AI ​​model 3 begins. After the operation of the reference electrical component is stopped in step 10, in step 11, it is determined whether the determination result by the AI ​​model 3 was normal (OK) or abnormal (NG). If the determination result by the AI ​​model 3 is normal (OK), it matches the result of the operation check of the speaker 4 via the microphone 5, so the AI ​​model 3 can be considered normal, and the diagnosis of the AI ​​model 3 ends (step 12). Then, the process proceeds to the diagnostic process of normal electrical components using the AI ​​model 3, which is comprehensively shown as step 16.

[0021] If the judgment result by AI model 3 in step 11 is abnormal (NG), the process returns to step 7 after judging the number of diagnoses in step 13, and repeats the abnormality judgment based on the power supply current using AI model 3. If the judgment of abnormality (NG) in step 11 is repeated a predetermined N times, the process proceeds from step 13 to step 14, and it is judged that AI model 3 itself is abnormal. Then, the process proceeds to system maintenance processing in step 15.

[0022] In this way, in the first embodiment described above, a simple process using the existing speaker 4 and microphone 5 that are generally provided in many vehicles can diagnose abnormalities in the AI ​​model 3, which is the premise for diagnosing abnormalities in electrical equipment using power supply current.

[0023] Next, Fig. 3 shows a second embodiment of the present invention. In this second embodiment, in order to increase the reliability of abnormality diagnosis of the AI ​​model 3, when an abnormality in the AI ​​model 3 is suspected by the processing of the first embodiment, a diagnosis of the AI ​​model 3 is performed using a different reference electrical component. In other words, at least two diagnoses using different reference electrical components are performed before it is finally determined that the AI ​​model 3 is abnormal.

[0024] 3A is a block diagram of the first AI model diagnostic system, and FIG. 3B is a block diagram of the second AI model diagnostic system. Both are basically configured in the same manner as the electrical equipment abnormality determination device of the first embodiment shown in FIG. 1, and include a battery 1, a current sensor 2 that detects the power supply current flowing from the battery 1, an AI model 3, a reference electrical equipment that constitutes the second normality / abnormality determination means, its operation detection unit, a first comparison unit 6, and a second comparison unit 7 that determines whether the AI ​​model 3 is abnormal by comparing the two determination results.

[0025] Here, in the first AI model diagnostic system, the reference electrical component is the first speaker 4A, and its operation detection unit is the first microphone 5A. Using the first speaker 4A and the first microphone 5A, the AI ​​model 3 is diagnosed in the same manner as in the first embodiment described above. If the first AI model diagnostic system diagnoses that the AI ​​model 3 is abnormal, a further diagnosis is performed using the second AI model diagnostic system.

[0026] In the second AI model diagnostic system, the reference electrical component is the second speaker 4B, and its operation detection unit is the second microphone 5B. The first microphone 5A of the first AI model diagnostic system may be used as the operation detection unit. The diagnostic method itself is the same as that of the first AI model diagnostic system. If the second AI model diagnostic system also diagnoses that the AI ​​model itself is abnormal, this is considered to be the final, confirmed diagnosis.

[0027] In a preferred embodiment, the first speaker 4A is a relatively small-diameter tweeter that can output inaudible high-pitched sounds in response to a test signal, and the second speaker 4B is a relatively large-diameter woofer that consumes more current than the tweeter. In other words, a tweeter and a woofer that are placed in appropriate positions in the vehicle cabin are selected as the reference electrical components for the vehicle audio system.

[0028] Thus, in the second embodiment, when it is determined that the AI ​​model 3 itself is abnormal, at least two diagnoses are performed using different electrical components as the reference electrical component, thereby increasing reliability. Furthermore, when diagnosing an abnormality in an electrical component based on the power supply current using the AI ​​model 3, if the electrical component consumes little power, the AI ​​model 3 may not be able to fully analyze the current waveform. In the above embodiment, in the second diagnosis, a reference electrical component (e.g., a woofer) that consumes more power than the reference electrical component (e.g., a tweeter) used in the first diagnosis is used, thereby making it less likely that an erroneous diagnosis will occur.

[0029] Although one embodiment of the present invention has been described in detail above, the present invention is not limited to the above embodiment and various modifications are possible. In particular, the reference electrical component is not limited to the speaker described above, and may be any electrical component that can be diagnosed as abnormal using the power supply current by the AI ​​model 3 and can be determined as abnormal by the second normal / abnormal determination means.

[0030] For example, the reference electrical component is an electrical component that outputs sound, vibration, radio waves, or light during operation, and is configured to detect the sound, vibration, radio waves, or light output by the reference electrical component with an appropriate receiver. Alternatively, the reference electrical component is an electrical component that generates heat during operation, and is configured to detect the heat output by the reference electrical component with a temperature sensor. Alternatively, the reference electrical component may be an electrical component that generates mechanical movement during operation, and the mechanical movement of the reference electrical component may be detected with a sensor.

[0031] Specific examples include the following: A map lamp in the vehicle cabin is set as the reference electrical component, and its light is detected by an illuminance sensor installed in the vehicle cabin; A power window is set as the reference electrical component, and the sound it makes when operating is detected by a microphone installed in the vehicle cabin; A radiator fan is set as the reference electrical component, and the mechanical vibrations that accompany its operation are detected by an acceleration sensor installed in the vehicle's anti-skid system, etc. [Explanation of symbols]

[0032] 1. Battery 2...Current sensor 3. AI model 4...Speaker 5...Microphone 6...First comparison section 7...Second comparison section

Claims

1. A method for determining an abnormality in a vehicle's electrical equipment, which is related to a plurality of electrical equipment supplied with power from the vehicle's battery, acquires a power supply current flowing from the battery, and detects an abnormality in each electrical equipment based on the current waveform using an AI model that has learned a reference current waveform when each electrical equipment is operating normally, determining one or more reference electrical components included in the electrical components for the purpose of diagnosing abnormalities of the AI ​​model, and providing a second normality / abnormality determination means for the reference electrical components that is independent of the AI ​​model; The reference electrical equipment is operated, and an abnormality determination is made using the AI ​​model based on the current waveform at that time, and an abnormality determination is made by the second normality / abnormality determination means, If the determination result by the second normality / abnormality determination means is normal and the determination result using the AI ​​model is abnormal, the AI ​​model is deemed to be abnormal. A method for determining abnormalities in vehicle electrical equipment.

2. The standard electrical equipment is an electrical equipment that outputs any of sound, vibration, radio wave, and light when in operation. The second normality / abnormality determination means detects sound, vibration, radio wave or light output from the reference electrical component via a receiver and compares it with an input to the reference electrical component to determine whether it is normal or abnormal. The method for determining an abnormality in an electrical component of a vehicle according to claim 1.

3. The standard electrical equipment is an electrical equipment that generates heat during operation, The second normality / abnormality determination means detects the heat output from the reference electrical component via a temperature sensor and compares it with the input to the reference electrical component to determine whether the reference electrical component is normal or abnormal. The method for determining an abnormality in an electrical component of a vehicle according to claim 1.

4. The reference electrical equipment is an electrical equipment that generates mechanical movement when in operation, The second normality / abnormality determination means detects mechanical movement of the reference electrical component via a sensor and compares the detected movement with an input to the reference electrical component to determine whether the component is normal or abnormal. The method for determining an abnormality in an electrical component of a vehicle according to claim 1.

5. The standard electrical equipment includes a plurality of standard electrical equipment, A first reference electrical component is operated for the abnormality diagnosis of the AI ​​model, and if the judgment result by the second normality / abnormality judgment means is normal and the judgment result using the AI ​​model is abnormal, a second reference electrical component is operated, and an abnormality determination is made using the AI ​​model based on the current waveform at that time, and in parallel, an abnormality determination is made using the second normality / abnormality determination means; If the determination result by the second normality / abnormality determination means is normal and the determination result using the AI ​​model is abnormal, the AI ​​model is deemed to be abnormal. The method for determining an abnormality in an electrical component of a vehicle according to claim 1.

6. selecting an electrical component that consumes more power during operation than the first reference electrical component as the second reference electrical component; The method for determining an abnormality in an electrical component of a vehicle according to claim 5.

7. The vehicle audio system includes a first speaker with a relatively low power consumption and a second speaker with a relatively high power consumption, The first reference electrical component is the first speaker, and the second reference electrical component is the second speaker. The method for determining an abnormality in an electrical component of a vehicle according to claim 6.

8. The reference electrical component is a speaker of a vehicle audio system, and the receiver is a microphone for audio input provided in the vehicle cabin. The method for determining an abnormality in an electrical component of a vehicle according to claim 2.

9. The reference electrical component is operated to determine whether or not there is an abnormality by the second normality / abnormality determining means; If it is determined to be normal, the same reference electrical equipment is operated and an abnormality determination is performed using the AI ​​model. If the judgment result using the AI ​​model is abnormal, the AI ​​model is deemed to be abnormal. The method for determining an abnormality in an electrical component of a vehicle according to claim 1.

10. The power supply current flowing from the battery is decomposed into current waveforms for each electrical component, and an abnormality in each electrical component is detected based on the individual current waveforms. The method for determining an abnormality in an electrical component of a vehicle according to claim 1.

11. An abnormality determination device for a vehicle's electrical equipment, which acquires a power supply current flowing from a vehicle battery for a plurality of electrical equipment to which power is supplied from the battery, and detects abnormalities in each electrical equipment based on the current waveform using an AI model that has learned a reference current waveform when each electrical equipment is operating normally, This electrical equipment abnormality determination device One or more reference electrical components preselected among the electrical components for abnormality diagnosis of the AI ​​model; an operation detection unit that detects that the reference electrical component has been operated; Furthermore, The reference electrical equipment is operated, and whether the reference electrical equipment is normal or abnormal is determined based on a response from the operation detection unit. The reference electrical equipment is operated, and an abnormality is determined using the AI ​​model based on the current waveform at that time. If the determination result based on the response of the motion detection unit is normal and the determination result using the AI ​​model is abnormal, the AI ​​model is deemed to be abnormal. A device for determining abnormalities in vehicle electrical equipment.

Images