JP7789588B2 - Vehicle inspection system, vehicle inspection method and program - Google Patents

Description

The present invention relates to an effective technology for conducting vehicle inspections.

In recent years, technology related to daily vehicle inspections has been attracting attention.
For example, Patent Document 1 proposes a technology that determines whether the equipment of a vehicle is in a normal state, a faulty state, or an abnormal state, and controls wasteful inspection and repair of equipment in an abnormal state that can be restored to a normal state.
Furthermore, Patent Document 2 proposes a technology that takes into account information about the driver's driving behavior when diagnosing the deterioration or consumption of a vehicle's consumables, and associates the degree of deterioration or consumption of the consumables with the driving behavior and displays the diagnosis results on a display screen.

Japanese Patent Application Laid-Open No. 2015-3570 Japanese Patent Application Laid-Open No. 2005-227141

In the logistics industry, drivers of trucks and other vehicles used for delivery are sometimes required to repeatedly inspect them over short intervals to ensure safe operation. For example, laws and regulations require drivers to inspect their vehicles daily to ensure proper operation of lights and brakes, and some even require drivers to inspect the vehicle once a day before getting into the vehicle. Delivery companies may also voluntarily order drivers to inspect their vehicles during and after work. These so-called daily inspections are a burden on delivery drivers. Therefore, automated vehicle inspections are needed to reduce the burden on drivers.
However, the techniques described in Patent Documents 1 and 2 do not allow for automatic inspection of vehicles.

The inventors recognized the need for a vehicle inspection system, vehicle inspection method, and program that can automatically perform vehicle inspections.

Therefore, the present invention aims to provide a vehicle inspection system, vehicle inspection method, and program that can automatically perform vehicle inspections.

The present invention provides a vehicle inspection system for inspecting a vehicle, comprising:
an inspection result acquisition unit that acquires inspection result data regarding inspection results for each inspection item of the vehicle;
an inspection data acquisition unit that acquires inspection sensor data measured by a sensor provided in the vehicle;
an inspection model generation unit that learns the correlation between the acquired inspection sensor data and the inspection result data and generates an inspection model;
an inspection implementation unit that applies the acquired inspection sensor data to the inspection model to inspect the vehicle for each inspection item;
A vehicle inspection system is provided.

According to the present invention, a vehicle inspection system that inspects vehicles acquires inspection sensor data measured by sensors installed in the vehicle, and inspects the vehicle for each inspection item based on the acquired inspection sensor data and accumulated past inspection sensor data.

This invention makes it possible to automatically perform vehicle inspections.

1 is a diagram illustrating an overview of a vehicle inspection system 1. FIG. FIG. 1 is a diagram illustrating a functional configuration of a vehicle inspection system 1. FIG. 2 is a diagram showing an inspection model generation process executed by the vehicle inspection system 1. FIG. 2 is a diagram showing an inspection implementation process executed by the vehicle inspection system 1. FIG. 2 is a diagram illustrating an example of an inspection database. FIG. 2 is a diagram schematically illustrating an example of operation data. FIG. 10 is a diagram illustrating an example of an inspection result notification. FIG. 2 is a diagram showing a fault model generation process executed by the vehicle inspection system 1. FIG. 3 is a diagram schematically illustrating an example of driving tendency data. FIG. 2 is a diagram schematically illustrating an example of weather data. FIG. 2 is a diagram showing a fault diagnosis process executed by the vehicle inspection system 1. FIG. 10 is a diagram illustrating an example of a failure database. FIG. 10 is a diagram illustrating an example of a failure notification.

Hereinafter, a detailed description of an embodiment of the present invention (hereinafter, "embodiment") will be given with reference to the accompanying drawings. In the following drawings, the same elements are designated by the same numbers or symbols throughout the description of the embodiment.

[Basic concept/basic configuration]
1 is a diagram for explaining an overview of a vehicle inspection system 1. The vehicle inspection system 1 is a system for inspecting a vehicle 5, which includes at least a computer 10 that automatically inspects predetermined inspection items of the vehicle 5.

This embodiment assumes that various sensors, described below, are attached to the vehicle 5, and these sensors detect predetermined sensor data.

The processing steps when the vehicle inspection system 1 inspects the vehicle 5 will be described with reference to FIG.
First, the computer 10 acquires inspection sensor data measured by various sensors provided on the vehicle 5 (step S1).
The vehicle 5 is equipped with various sensors that measure the status of items corresponding to inspection items. The inspection items include, for example, daily inspections of loose V-belts, tire conditions (air pressure, cracks, grooves), the remaining amount of brake fluid, the operation status of the parking brake, the remaining amount of coolant, and the lighting status of lights. The sensors include, for example, a vibration sensor installed near the V-belt, a tire air pressure sensor, a fluid level sensor, a parking brake sensor, a coolant level sensor that measures the water level in a reservoir tank, and an illuminance sensor installed inside the lights that measures light intensity. The sensor data includes, for example, vibration sensor data measured by the vibration sensor, air pressure sensor data measured by the air pressure sensor, fluid level sensor data measured by the fluid level sensor, parking brake sensor data measured by the parking brake sensor, coolant level sensor data measured by the coolant level sensor, and illuminance sensor data measured by the illuminance sensor.
The vehicle 5 transmits sensor data measured by sensors as inspection sensor data to the computer 10 via its own ECU (Electronic Control Unit). The computer 10 receives the inspection sensor data and thereby acquires the inspection sensor data.

The computer 10 inspects the vehicle 5 for each inspection item based on the acquired inspection sensor data and the accumulated past inspection sensor data (step S2).
The computer 10 stores in advance past inspection sensor data previously acquired from the vehicle 5. The computer 10 compares the currently acquired inspection sensor data with the past inspection sensor data, and performs an inspection for the presence or absence of an abnormality or a malfunction for each inspection item. For example, if the currently acquired inspection sensor data matches or is similar to past inspection sensor data showing a state without an abnormality or malfunction, it determines that there is no abnormality or malfunction. Additionally, the computer 10 learns past inspection sensor data and inspection results, and performs an inspection of the vehicle 5 for each inspection item based on the learning results and the currently acquired inspection sensor data.

With this vehicle inspection system 1, vehicle inspections can be carried out automatically.

[Functional configuration]
The functional configuration of the vehicle inspection system 1 will be described with reference to FIG.
The vehicle inspection system 1 is a system that includes a computer 10 that is connected to an ECU 7 equipped in a vehicle 5, an ECU 8 equipped in another vehicle 6, a driver terminal carried by the driver of the vehicle 5, an operations manager terminal carried by the manager of the vehicle 5, etc., so as to enable data communication via a network 3 such as a public line network.
The vehicle inspection system 1 may include the above-mentioned ECUs 7 and 8, a driver terminal, an administrator terminal, other terminals and devices, etc. In this case, the vehicle inspection system 1 executes the processes described below using one or a combination of the computer 10, other terminals and devices, etc.

The computer 10 is a computer or personal computer having a server function for inspecting the vehicle 5 .
The computer 10 may be realized, for example, by a single computer, or by multiple computers, such as a cloud computer. The cloud computer in this specification may refer to either a computer that uses any computer in a scalable manner to perform a specific function, or a computer that includes multiple functional modules to realize a system and uses the functions in any combination.

The computer 10 includes, as a control unit, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), etc., and, as a communication unit, includes a device that enables communication with other terminals and devices, etc., and an inspection data acquisition unit 11 that acquires inspection sensor data measured by a sensor provided on the vehicle 5, etc.
The computer 10 also includes a data storage unit such as a hard disk, semiconductor memory, storage medium, or memory card as a memory unit.
In addition, the computer 10 includes, as a processing unit, various devices that perform various processes, an inspection implementation unit 12 that performs vehicle inspections for each inspection item based on inspection sensor data and accumulated past inspection sensor data, and the like.

In the computer 10, the control unit loads a predetermined program and works in cooperation with the communication unit to realize an inspection sensor data acquisition module, an inspection result data acquisition module, an ignition data acquisition module, an inspection result notification module, a sensor data acquisition module in the event of a failure, and a failure notification module.
In addition, in the computer 10, the control unit reads a predetermined program, and in cooperation with the storage unit, realizes an inspection model storage module, an inspection result recording module, a fault determination model storage module, and a fault diagnosis result recording module.
In addition, in computer 10, the control unit reads a specified program and works in cooperation with the processing unit to realize an inspection model generation module, a condition judgment module, a time acquisition module, an inspection completion judgment module, an inspection execution module, a problem judgment module, an inspection operation data acquisition module, a malfunction operation data acquisition module, another vehicle driver driving tendency data acquisition module, another vehicle operation day weather data acquisition module, a fault judgment model generation module, a vehicle driver driving tendency data acquisition module, a vehicle operation day weather data acquisition module, and a fault judgment module.

[Inspection model generation process executed by computer 10]
The inspection model generation process executed by the computer 10 will be described with reference to Fig. 3. Fig. 3 is a diagram showing a flowchart of the inspection model generation process executed by the computer 10. The processes executed by each of the above-mentioned modules will be described together with this process.

First, the inspection sensor data acquisition module acquires inspection sensor data from the vehicle 5 (step S10).
The vehicle 5 transmits sensor data measured by various sensors equipped on the vehicle 5 that measure the status of items corresponding to inspection items, etc., to the computer 10 via the ECU 7 as inspection sensor data. As described above, the inspection items are items that are inspected daily, such as loose V-belts, tire conditions (air pressure, cracks, grooves), remaining brake fluid, parking brake operation status, remaining coolant level, and lighting status of lights. As described above, the sensors include vibration sensors, air pressure sensors, fluid level sensors, parking brake sensors, coolant level sensors, and illuminance sensors. As described above, the sensor data include vibration sensor data, air pressure sensor data, fluid level sensor data, parking brake sensor data, coolant level sensor data, and illuminance sensor data.
The inspection sensor data acquisition module receives the inspection data transmitted by the ECU 7. The inspection sensor data acquisition module receives the inspection data and thereby acquires inspection sensor data from the vehicle 5.

The inspection result data acquisition module acquires inspection result data relating to the inspection results for each inspection item of the vehicle 5 (step S11).
The driver terminal or the operations manager terminal accepts input of the inspection results for each inspection item of the vehicle 5. The inspection results include whether or not the V-belt is loose, whether or not the condition of the tires is normal, whether or not the remaining amount of brake fluid is appropriate, whether or not the operation state of the parking brake is normal, whether or not the remaining amount of coolant is appropriate, whether or not the lighting state of the lights is normal, etc.
The driver terminal or the operations manager terminal transmits the inspection results that it has received as input to the computer 10 as inspection result data. The inspection result data acquisition module receives this inspection result data. By receiving this inspection result data, the inspection result data acquisition module acquires the inspection result data.

The inspection model generation module learns the correlation between the inspection sensor data and the inspection result data, and generates an inspection model (step S12).
The learning method executed by the inspection model generation module includes machine learning using supervised learning, unsupervised learning, reinforcement learning, etc., and deep learning using convolutional neural networks, recurrent neural networks, long-short-term memory, etc. In this embodiment, machine learning using supervised learning will be described as an example.
The inspection model generation module performs machine learning using supervised learning to determine the correlation between each acquired inspection sensor data and the inspection result data for the inspection items measured by each inspection sensor data.
The inspection model generation module learns the correlation between vibration sensor data and the inspection results for V-belt slack. The inspection model generation module also learns the correlation between air pressure sensor data and the inspection results for tire condition. The inspection model generation module also learns the correlation between fluid level sensor data and the inspection results for the remaining amount of brake fluid. The inspection model generation module also learns the correlation between parking brake sensor data and the inspection results for the parking brake operation status. The inspection model generation module also learns the correlation between coolant level sensor data and the inspection results for the remaining amount of coolant. The inspection model generation module also learns the correlation between illuminance sensor data and the inspection results for the lighting status of the lights.
The inspection model generation module generates an inspection model for each inspection item based on the learning results.

The inspection model storage module stores the inspection model (step S13). The inspection model storage module stores the generated inspection model for each inspection item.

The inspection model generation process is completed as described above.
As a result of the inspection model generation process described above, the computer 10 accumulates past inspection sensor data and past inspection results.

[Inspection Implementation Process of Computer 10]
The inspection implementation process executed by the computer 10 will be described with reference to Fig. 4. Fig. 4 is a diagram showing a flowchart of the inspection implementation process executed by the computer 10. The processes executed by each of the above-mentioned modules will be described together with this process. This inspection implementation process is a detailed description of the above-mentioned inspection data acquisition process (step S1) and the process of implementing inspections on the vehicle 5 for each inspection item (step S2). Note that detailed descriptions of processes similar to those described above will be omitted.

First, the inspection sensor data acquisition module acquires inspection sensor data from the vehicle 5 (step S20). The processing in step S20 is similar to the processing in step S10 described above.

The condition determination module determines whether a predetermined condition is met (step S21).
The predetermined conditions include whether ignition data indicating that the ignition switch of the vehicle 5 has been turned on or off has been acquired, whether the current time is a preset time, and so on.
A case where the predetermined condition is whether or not ignition data has been acquired will be described.
When the ignition switch is turned on or off, the ECU 7 transmits ignition data indicating this to the computer 10. The ignition data acquisition module receives this ignition data.
The condition determination module executes the process of step S21 depending on whether or not the ignition data has been acquired. If the condition determination module determines that the ignition data has not been acquired (NO in step S21), the inspection execution process ends.
On the other hand, if the condition determination module determines that the ignition data has been acquired (YES in step S21), the computer 10 executes an inspection completion determination process (step S22) to be described later.
A case where the predetermined condition is whether the current time is a preset time or not will be described.
The time acquisition module acquires the current time from a timer or the like provided in itself.
The condition determination module executes the process of step S21 based on whether the current time is a preset time. The preset time is, for example, a time when a daily inspection is generally performed, such as the start time of the vehicle 5 or a time a predetermined time earlier than the start time of the vehicle 5. If the condition determination module determines that the current time is not the preset time (NO in step S21), it ends the inspection execution process.
On the other hand, when the condition determination module determines that the current time is the preset time (YES in step S21), it executes an inspection completion determination process (step S22) to be described later.

If the condition determination module determines that the above-described predetermined conditions are met, the inspection completion determination module determines whether today's inspection has been completed (step S22). The inspection completion determination module references an inspection database (hereinafter referred to as inspection DB) that records the inspection date and time and the inspection results, and determines whether today's inspection has been completed.
If the inspection results are recorded in the inspection DB at the date and time of today's inspection, the inspection completion determination module determines that today's inspection has been completed (step S22 YES), and ends this inspection implementation process.
By the process of step S22, the inspection is performed only once a day. If the inspection is performed only once a day, there is no need to wait for the inspection to finish.

On the other hand, if the inspection results are not recorded in the inspection DB for today's inspection date and time, the inspection completion determination module determines that today's inspection has not been completed (step S22 NO), and the inspection execution module applies the inspection sensor data to the inspection model and inspects the vehicle 5 for each inspection item (step S23).
The inspection execution module inspects the inspection items for the currently acquired inspection sensor data using the inspection model generated by the processing of step S12 described above. The inspection execution module estimates the inspection results for the currently acquired inspection sensor data that match or are similar to the inspection sensor data in the inspection model as the inspection results for the inspection items.
If the inspection item is V-belt slack, the inspection execution module uses a V-belt inspection model to estimate V-belt slack based on the amplitude and frequency of the acquired vibration sensor data. If the inspection item is tire condition, the inspection execution module uses the tire inspection model to estimate tire condition based on the acquired air pressure sensor data. Similarly, for other inspection items, the inspection execution module uses the inspection model for the inspection item to estimate the inspection result for the inspection item based on the acquired inspection sensor data.
The inspection execution module determines that there is no problem with the inspection item if the estimated inspection result is no problem, and determines that there is a problem with the inspection item if the estimated inspection result is problematic.The inspection execution module determines that there is a problem with the inspection item if the estimated inspection result is neither.
The inspection execution module determines the nature of the problem for an inspection item that is determined to have a problem from the inspection results.If the tire condition estimated from the air pressure sensor data indicates that the tire is not air-pressure-rich, the inspection execution module determines that the problem is insufficient air pressure.The inspection execution module also determines the nature of the problem for other inspection items that are determined to have a problem from the inspection results.

The inspection execution module can also be configured to perform inspections without using an inspection model.
In this case, the inspection execution module compares previously acquired inspection sensor data stored in advance with the currently acquired inspection sensor data. If the currently acquired inspection sensor data matches or is similar to the past inspection sensor data that was determined to be problem-free in the inspection result data, the inspection execution module determines that there is no problem with the inspection item. If the currently acquired inspection sensor data matches or is similar to the past inspection sensor data that was determined to be problematic in the inspection result data, the inspection execution module determines that there is a problem with the inspection item. If the currently acquired inspection sensor data is neither of these, the inspection execution module determines that there is a problem with the inspection item. The computer 10 may execute the processing described below according to these determination results.

The inspection result recording module records the inspection date and time and the inspection results in the inspection DB (step S24). The inspection result recording module records the identifier of this vehicle 5 (vehicle ID, control number, driver ID, etc.), the inspection date and time, and the inspection results for each inspection item in association with each other in the inspection DB.
The inspection result recording module records the sensor data for each inspection item, linking it to the inspection results for each inspection item.

[Inspection DB]
The inspection DB will be described with reference to Fig. 5. Fig. 5 is a diagram schematically illustrating an example of the inspection DB recorded by the inspection result recording module.
5, the inspection DB records the vehicle ID of the vehicle 5 that was inspected, the inspection date and time, the inspection items, and the inspection results for each inspection item, all in association with each other. The inspection result recording module records "no problem" in the inspection DB if the inspection results show no problems with the inspection items, and records the fact that there is a problem with the inspection items and the details of the problem if there is a problem with the inspection items. The inspection sensor data acquired this time is linked to each inspection item.
The computer 10 uses this inspection DB in the inspection model update process (step S28) described below.

Returning to FIG. 4, the rest of the inspection execution process will be described.
The problem determination module determines whether or not there is a problem with the inspection item (step S25). The problem determination module refers to the inspection DB and determines whether or not the inspection result this time is recorded as having a problem. Note that the problem determination module may also determine whether or not there is a problem as a result of the inspection performed by the processing of step S23 described above.
If the problem determination module determines that there is no problem with the inspection item (step S25: NO), the computer 10 executes an inspection model update process (step S28) to be described later.

On the other hand, if the problem determination module determines that there is a problem with the inspection item (YES in step S25), the inspection operation data acquisition module acquires inspection operation data (step S26).
The inspection operation data is data indicating operation data such as the operation date and time when the vehicle 5 delivers the delivery when the inspection is carried out, the vehicle ID, the delivery origin, the delivery destination, the delivery route (delivery distance, whether or not there is high ground), the product weight, etc. The operation data is stored in advance in the storage unit.
The inspection operation data acquisition module acquires, as inspection operation data, operation data of the vehicle 5 that has been determined to have a problem with the current inspection item based on the date and time of the inspection and the vehicle ID.
The inspection operation data acquisition module may acquire operation data stored in the driver's terminal or the operation manager's terminal as inspection operation data.

[Operation data]
The operation data will be described with reference to Fig. 6. Fig. 6 is a diagram showing an example of the operation data stored in the storage unit.
In FIG. 6, the operation data is stored in association with the vehicle operation date and time, vehicle ID, delivery origin, delivery destination, delivery route (delivery distance, presence or absence of high altitude), and product weight.
If the vehicle ID of the vehicle 5 inspected this time is "100001", the inspection operation data acquisition module acquires the operation data of the delivery origin, delivery destination, delivery route, and product weight associated with the date and time the inspection was performed from the operation dates and times associated with this vehicle ID as inspection operation data.

Returning to FIG. 4, the rest of the inspection execution process will be described.
The inspection result notification module notifies the driver that there is a problem and the inspection operation data (step S27). The inspection result notification module generates an inspection result notification that includes details of the inspection item that was determined to have a problem, details of the problem, and the acquired inspection operation data. The inspection result notification module sends the generated inspection result notification to the driver's terminal or the operations manager's terminal.
The driver terminal or the operations manager terminal receives this inspection result notification and displays it on its own display unit, etc. The inspection result notification module outputs this inspection result notification to the driver terminal or the operations manager terminal, thereby notifying the driver or the operations manager that there is a problem and the operation data at the time of inspection.

[Inspection result notification]
The inspection result notification will be described with reference to Fig. 7. Fig. 7 is a diagram showing an example of an inspection result notification output to the operation manager terminal.
In FIG. 7, an inspection item that is determined to have a problem is shown as a "problem item." The details of the problem are shown as "problem details." If a tire air pressure problem is determined to have a problem, the inspection result notification will show the tire as the problem item and insufficient air pressure as the problem details. The inspection operation data is shown as "operation data." This operation data is the inspection operation data acquired by the processing of step S26 described above.
By viewing this inspection result notification, the operations manager can ascertain the inspection item that has a problem, the details of the problem, and the operation data of the vehicle 5 that has the problem. When the inspection result notification is output to the driver's terminal, the same inspection result notification as shown in Fig. 7 is also output to the driver's terminal. By viewing this inspection result notification, the driver can ascertain the inspection item that has a problem, the details of the problem, and the operation data of the vehicle 5 that has the problem.

Returning to FIG. 4, the rest of the inspection execution process will be described.
The inspection model generation module updates the inspection model based on the current inspection result (step S28).
The inspection model generation module performs machine learning using supervised learning to determine the correlation between the acquired inspection sensor data and the inspection results based on the current inspection results. The inspection model generation module updates each inspection model generated by the processing in step S12 based on the learning results.
The inspection model storage module stores each updated inspection model.
In the next and subsequent inspection execution processes, the computer 10 executes the process of step S23 using the inspection model updated by the process of step S28.

This completes the inspection process.

[Fault Model Generation Process Executed by the Computer 10]
The fault model generation process executed by the computer 10 will be described with reference to Fig. 8. Fig. 8 is a diagram showing a flowchart of the fault model generation process executed by the computer 10. The processes executed by each of the above-mentioned modules will be described together with this process.

First, the failure sensor data acquisition module acquires failure sensor data from the other vehicle 6 (step S40).
The other vehicle 6 is a vehicle related to the vehicle 5 and is a vehicle of the same model as the vehicle 5.
When another vehicle 6 has previously experienced a breakdown or is determined to have a breakdown, the other vehicle 6 transmits sensor data measured by various sensors that measure the status of items corresponding to its own inspection items to the computer 10 via the ECU 8 as failure sensor data. The inspection items are the same as the inspection items of the vehicle 5 described above, such as looseness of the V-belt, tire condition (air pressure, cracks, grooves), remaining amount of brake fluid, parking brake operation status, remaining amount of coolant, and lighting status of lights. The sensors are the same as the sensors of the vehicle 5 described above, such as a vibration sensor, air pressure sensor, fluid level sensor, parking brake sensor, coolant level sensor, and illuminance sensor. The sensor data are the same as the sensor data described above, such as vibration sensor data measured by a vibration sensor, air pressure sensor data measured by an air pressure sensor, fluid level sensor data measured by a fluid level sensor, parking brake sensor data measured by a parking brake sensor, coolant level sensor data measured by a coolant level sensor, and illuminance sensor data measured by an illuminance sensor.
The failure sensor data acquisition module receives the failure sensor data transmitted by the ECU 8. The failure sensor data acquisition module receives the failure sensor data, thereby acquiring failure sensor data from the other vehicle 6.

The breakdown operation data acquisition module acquires breakdown operation data (step S31).
The operation data in the event of a breakdown is data that indicates the operation data of the other vehicle 6 when the other vehicle 6 breaks down. As described above, the operation data includes the operation date and time when delivering the delivery, the vehicle ID, the delivery origin, the delivery destination, the delivery route (delivery distance, whether or not there is high ground), the product weight, etc. The operation data is stored in advance in the storage unit.
The breakdown operation data acquisition module acquires operation data of the vehicle 6 at the date and time when the breakdown occurred as breakdown operation data based on the date and time when the breakdown occurred and the vehicle ID.

The other vehicle driver driving tendency data acquisition module acquires other vehicle driver driving tendency data (step S32).
The other vehicle driver's driving tendency data is driving tendency data that indicates the driving tendency of the driver when the other vehicle 6 breaks down. The driving tendency data is data that is stored in advance in a storage unit or in another computer.
The other vehicle driver driving tendency data acquisition module acquires the driving tendency data of the driver who will drive the other vehicle 6 at this time as the other vehicle driver driving tendency data based on the driver ID of the driver of the other vehicle 6 when the other vehicle 6 breaks down.

[Driving tendency data]
The driving tendency data will be described with reference to Fig. 9. Fig. 9 is a diagram schematically showing an example of driving tendency data stored in the storage unit.
In FIG. 9, the driving tendency data is stored in association with a driver ID, whether or not sudden braking has occurred, whether or not sudden acceleration has occurred, whether or not a shift change has occurred early or late, and whether or not sudden steering has occurred.
If the driver ID of the driver of the broken-down other vehicle 6 this time is "200001", the other vehicle driver driving tendency data acquisition module acquires the driving tendency data of the other vehicle driver associated with this driver ID, including whether or not the driver braked suddenly, accelerated suddenly, shifted gears early or late, and steered suddenly.

The other vehicle operation day weather data acquisition module acquires other vehicle operation day weather data (step S33).
The weather data for the other vehicle operation day is weather data that indicates the weather on the operation day when the other vehicle 6 breaks down. The weather data is stored in advance in a storage unit or in another computer.
The other vehicle operation day weather data acquisition module acquires weather data of the date and place where the other vehicle 6 moves as the other vehicle operation day weather data based on the operation date and time, delivery origin and delivery destination of the other vehicle 6 when the other vehicle 6 breaks down.

[Weather Data]
The weather data will be described with reference to Fig. 10. Fig. 10 is a diagram showing an example of weather data stored in the storage unit.
10, the weather data is stored in such a manner that the date and time, the area, the presence or absence of heavy rain, the presence or absence of strong winds, the presence or absence of high waves, and the presence or absence of high tides are associated with each other. The weather data is stored for each date.
In this case, if the date and time when the other vehicle 6 broke down was "2020/02/05 8:00", and the delivery origin in the operation data at the time of the breakdown was "Southern Kanto" and the delivery destination was "West Kanto", the other vehicle operation day weather data acquisition module will acquire the presence or absence of heavy rain, strong winds, high waves, and high tides as weather data for the other vehicle operation day, associated with this date and time and the regions of the delivery origin and destination.

The failure determination model generation module learns the correlations between the sensor data at the time of failure, the operation data at the time of failure, the driving tendency data of the driver of the other vehicle, and the weather data on the day the other vehicle is operated, and generates a failure determination model (step S34).
Similar to the learning method executed by the inspection model generation module described above, the learning method executed by the fault determination model generation module includes machine learning using supervised learning, unsupervised learning, reinforcement learning, etc., and deep learning using convolutional neural networks, recurrent neural networks, long-short-term memory, etc. In this embodiment, machine learning using supervised learning will be described as an example.
The failure judgment model generation module performs machine learning using supervised learning to determine the correlation between the acquired sensor data at the time of failure, operation data at the time of failure, driving tendency data of drivers of other vehicles, and weather data on the day other vehicles are operated.
The failure determination model generation module learns the correlations between vibration sensor data, operation data at the time of a failure, driving tendency data of the driver of another vehicle, and weather data on the day the other vehicle is operated. The failure determination model generation module also learns the correlations between tire pressure sensor data, operation data at the time of a failure, driving tendency data of the driver of another vehicle, and weather data on the day the other vehicle is operated. The failure determination model generation module also learns the correlations between flute level sensor data, operation data at the time of a failure, driving tendency data of the driver of another vehicle, and weather data on the day the other vehicle is operated. The failure determination model generation module also learns the correlations between parking brake sensor data, operation data at the time of a failure, driving tendency data of the driver of another vehicle, and weather data on the day the other vehicle is operated. The failure determination model generation module also learns the correlations between coolant level sensor data, operation data at the time of a failure, driving tendency data of the driver of another vehicle, and weather data on the day the other vehicle is operated. The failure determination model generation module also learns the correlation between the illuminance sensor data, the operation data at the time of the failure, the driving tendency data of the driver of the other vehicle, and the weather data on the day the other vehicle is operated.
The failure determination model generation module generates a failure determination model for each inspection item based on the learning results.
The failure determination model generation module may generate a failure determination model for each model year of the other vehicle 6. The model year may also be acquired from the ECU 8.

The fault determination model storage module stores the fault determination model (step S35). The fault determination model storage module stores the generated fault determination model for each inspection item.

The above is the fault determination model generation process.
As a result of the above-described fault determination model generation process, the computer 10 generates a fault determination model based on sensor data measured when the other vehicle 6 previously experienced a fault.

[Fault diagnosis process of computer 10]
The fault diagnosis process executed by the computer 10 will be described with reference to Fig. 11. Fig. 11 is a diagram showing a flowchart of the fault diagnosis process executed by the computer 10. The processes executed by each of the modules described above will be described together with this process. Note that detailed descriptions of processes similar to those described above will be omitted.

First, the inspection sensor data acquisition module acquires inspection sensor data from the vehicle 5 (step S40). The processing in step S40 is similar to the processing in step S20 described above.

The condition determination module determines whether or not a predetermined condition is met (step S41). The processing in step S41 is the same as the processing in step S21 described above. If the condition determination module determines that the predetermined condition is not met (step S41 NO), the fault diagnosis process ends.

On the other hand, if the condition determination module determines that the above-mentioned predetermined condition is met (YES in step S41), the inspection completion determination module determines whether or not today's inspection has been completed (step S42). The inspection completion determination module determines whether or not today's inspection has been completed by referring to a fault database (hereinafter referred to as the fault DB) described below in which the inspection date and time and the fault diagnosis results are recorded.
If the fault diagnosis result is recorded in the fault DB at the date and time of today's inspection, the inspection completion determination module determines that today's inspection has been completed (step S42 YES), and ends this fault diagnosis process.
By the process of step S42, the inspection is performed only once a day. If the inspection is performed only once a day, there is no need to wait for the inspection to finish.

On the other hand, if the fault diagnosis result is not recorded in the fault DB for today's inspection date and time, the inspection completion determination module determines that today's inspection has not been performed (step S42: NO), and the inspection operation data acquisition module acquires inspection operation data (step S43). The processing of step S43 is the same as the processing of step S26 described above.

The vehicle driver driving tendency data acquisition module acquires the vehicle driver driving tendency data (step S44).
The vehicle driver driving tendency data is driving tendency data that indicates the driving tendency of the driver who drives the vehicle 5 to be inspected. As described above, the driving tendency data is data that has been stored in advance in a storage unit or in another computer. The driving tendency data is the same as that described in connection with the processing of step S32 above, and is shown in FIG. 9.
The vehicle driver driving tendency data acquisition module acquires, based on the driver ID of the driver of the vehicle 5, driving tendency data of the driver currently driving the vehicle 5 as vehicle driver driving tendency data.
If the driver ID of the driver of vehicle 5 undergoing inspection this time is "200001", the vehicle driver driving tendency data acquisition module acquires the vehicle driver driving tendency data associated with this driver ID, including whether or not the driver braked suddenly, accelerated suddenly, shifted gears early or late, and made sudden turns.

The vehicle operation day weather data acquisition module acquires vehicle operation day weather data (step S45).
The vehicle operation day weather data is weather data that indicates the weather on the operation day on which the inspection is to be performed. As described above, the weather data is stored in advance in the storage unit or in another computer. The weather data is the same as that described in the processing of step S33 above, and is the same as that shown in FIG. 10.
The vehicle operation day weather data acquisition module acquires weather data for the date and location where the vehicle 5 travels as vehicle operation day weather data based on the operation date and time when the inspection is performed, the delivery origin, and the delivery destination.
If the date and time for inspection of vehicle 5 this time is "2021/02/05 8:00", and the origin of delivery in the inspection operation data is "Southern Kanto" and the destination of delivery is "West Kanto", the vehicle operation day weather data acquisition module will acquire the presence or absence of heavy rain, strong winds, high waves, and high tides as weather data for other vehicles operation days, associated with this date and time and the regions of the origin and destination of delivery.

The inspection execution module applies the inspection sensor data, the operation data at the time of inspection, the vehicle driver's driving tendency data, and the weather data on the day the vehicle is operated to a fault determination model, and performs an inspection to determine whether or not there is a fault in the vehicle 5 for each inspection item (step S46).
The inspection execution module uses the failure determination model generated by the processing of step S34 described above to predict the presence or absence of a failure for each inspection item if the currently acquired inspection sensor data, inspection operation data, vehicle driver driving tendency data, and vehicle operation day weather data match or approximate each other. Here, predicting the presence or absence of a failure includes not only predicting the presence or absence of a current failure, but also predicting the possibility of a failure during operation or at the end of operation.
When the inspection item is the condition of the tires, the inspection execution module uses a tire failure determination model to predict the presence or absence of a tire failure (e.g., a burst) based on the air pressure sensor data, the delivery distance in the inspection operation data, the presence or absence of sudden acceleration or sudden braking in the vehicle driver's driving tendency data, and the presence or absence of rain in the operation day weather data. Similarly, for other inspection items, the inspection execution module uses the inspection item failure determination model to predict the presence or absence of a failure in the inspection item based on the acquired inspection sensor data, inspection operation data, vehicle driver's driving tendency data, and operation day weather data.
The inspection execution module determines that there is no fault in the inspection results if the inferred inspection result shows no fault, and determines that there is a fault in the inspection results if the inferred inspection result shows there is a fault.The inspection execution module determines that there is no fault in the inspection results if the inferred inspection result shows neither.
The inspection execution module determines the nature of the failure from the inspection results for inspection items that are determined to have a malfunction, and determines how to deal with the malfunction.If the tire condition estimated from air pressure sensor data, delivery distance, sudden acceleration, sudden braking, and rain is determined to be a possible tire burst, the module determines the nature of the malfunction as a tire burst and determines that the appropriate remedy is to replace the tire.The inspection execution module also determines the nature of the malfunction and how to deal with other inspection items that are determined to have a malfunction.

The fault diagnosis result recording module records the inspection date and time and the inspection result in the fault DB (step S47). The fault diagnosis result recording module records the identifier of the vehicle 5, the inspection date and time, and the fault diagnosis result for each inspection item in association with each other in the fault DB.
The inspection result recording module records the sensor data for each inspection item, linking it to the fault diagnosis results for each inspection item.

[Failure DB]
The fault DB will be described with reference to Fig. 12. Fig. 12 is a diagram showing an example of the fault DB recorded by the fault diagnosis result recording module.
12, the fault DB records the vehicle ID of the vehicle 5 that was inspected, the inspection date and time, the inspection items, and the inspection results for each inspection item, all in association with each other. The fault diagnosis result recording module records "no fault" in the inspection DB if the inspection results show no fault in the inspection item, and records the fact that a fault exists and the remedy (fault present and tire replacement) if a fault is found in the inspection item. The inspection sensor data acquired this time is linked to each inspection item.
The computer 10 uses this fault DB in the process of updating the fault determination model (step S50) described later.

Returning to FIG. 11, the rest of the fault diagnosis process will be described.
The fault determination module determines whether or not there is a fault in the inspection item (step S48). The fault determination module refers to the fault DB and determines whether or not there is a record of a fault in the current inspection result. Note that the fault determination module may determine whether or not there is a fault based on the inspection result by the processing of step S46 described above.
If the fault determination module determines that there is no fault in the inspection item (NO in step S48), the computer 10 ends this fault diagnosis process.

On the other hand, if the malfunction determination module determines that there is a malfunction in the inspection item (step S48 YES), the malfunction notification module notifies the driver that there is a malfunction and the inspection operation data (step S49). The malfunction notification module generates a malfunction notification including the details of the inspection item determined to have a malfunction, a countermeasure for the malfunction, and the acquired inspection operation data. The malfunction notification module transmits the generated malfunction notification to the driver's terminal or the operation manager's terminal.
The driver terminal or the operations manager terminal receives this failure notification and displays it on its own display unit, etc. The failure notification module causes the driver terminal or the operations manager terminal to output this failure notification, thereby notifying the driver or the operations manager that there is a failure, how to deal with it, and the inspection operation data.

[Failure notification]
The malfunction notification will be described with reference to Fig. 13. Fig. 13 is a diagram showing an example of a malfunction notification output to the operation manager terminal.
In FIG. 13, an inspection item that is determined to have a malfunction is shown as a "malfunction item." Furthermore, the remedy is shown as a "remedy." If a tire burst is determined to be the malfunction, the malfunction notification will indicate tire burst as the malfunction item and the remedy that tire replacement is required. Furthermore, the inspection operation data is shown as "operation data." This operation data is the inspection operation data acquired by the processing of step S43 described above.
By viewing this malfunction notification, the operations manager can ascertain the inspection item with the malfunction, how to deal with it, and the operation data of the vehicle 5 with the malfunction. When the malfunction notification is output to the driver's terminal, a malfunction notification similar to that shown in Fig. 13 is also output to the driver's terminal. By viewing this malfunction notification, the driver can ascertain the inspection item with the malfunction, how to deal with it, and the operation data of the vehicle 5 with the malfunction.

Returning to FIG. 11, the rest of the fault diagnosis process will be described.
The fault determination model generation module updates the fault determination model based on the current inspection result (step S50).
Based on the inspection results, the fault determination model generation module performs machine learning using supervised learning to determine correlations between the acquired inspection sensor data, inspection operation data, vehicle driver driving tendency data, vehicle operation day weather data, and the inspection results.The fault determination model generation module updates the fault determination model generated by the processing of step S34 described above based on the learning results.At this time, only the fault determination models for inspection items determined to have a fault are updated.
The fault determination model storage module stores each updated fault determination model.
In the next and subsequent fault diagnosis processes, the computer 10 executes the process of step S46 using the fault determination model updated by the process of step S50.

This completes the fault diagnosis process.

The fault diagnosis process described above is for inspecting whether or not there is a fault in the vehicle 5, but the computer 10 can also inspect for signs of a fault for each inspection item in the vehicle 5. This case will be described.
The computer 10 refers to the inspection DB to acquire the inspection sensor data of the other vehicle 6 at a predetermined timing (e.g., several days, one day, or one week) before the timing at which the sensor data at the time of failure was acquired. The computer 10 also acquires the inspection operation data, the vehicle driver's driving tendency data, and the weather data on the vehicle operation day of the other vehicle 6 at this timing.
The computer 10 learns the correlations between the acquired inspection sensor data, inspection operation data, vehicle driver driving tendency data, and vehicle operation day weather data, and generates a sign determination model.
When the computer 10 checks for the presence or absence of a fault in the above-mentioned fault diagnosis process (step S46), it simply applies the generated precursor determination model and checks for the presence or absence of precursors of a fault for each inspection item.
If the computer 10 determines that there is a sign of a malfunction, it notifies the user of the presence of the sign of a malfunction, suggests whether maintenance is necessary, and provides the operational data at the time of inspection, in the same manner as in the process of step S49.
By executing such processing, the computer 10 is able to check for signs of failure.

The above-mentioned means and functions are realized by a computer (including a CPU, information processing device, and various terminals) loading and executing a specified program. The program may be provided, for example, from the computer via a network (Software as a Service (SaaS)) or as a cloud service. The program may also be provided in a form recorded on a computer-readable recording medium. In this case, the computer reads the program from the recording medium, transfers it to an internal or external recording device, records it, and executes it. The program may also be pre-recorded on a recording device (recording medium) and provided from the recording device to the computer via a communication line.

Although the embodiments of the present invention have been described above, the present invention is not limited to these above-mentioned embodiments. Furthermore, the effects described in the embodiments of the present invention are merely a list of the most favorable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments of the present invention.

(1) A vehicle inspection system that performs an inspection of a vehicle (e.g., vehicle 5),
an inspection data acquisition unit (e.g., an inspection data acquisition unit 11, an inspection sensor data acquisition module) that acquires inspection sensor data (e.g., vibration sensor data, air pressure sensor data, fluid level sensor data, parking brake sensor data, coolant level sensor data, illuminance sensor data) measured by sensors (e.g., vibration sensor, air pressure sensor, fluid level sensor, parking brake sensor, coolant level sensor, illuminance sensor) provided in the vehicle;
an inspection execution unit (e.g., inspection execution unit 12, inspection execution module) that inspects the vehicle for each inspection item (e.g., looseness of a V-belt, tire condition (air pressure, cracks, grooves), remaining amount of brake fluid, operating state of a parking brake, remaining amount of coolant, lighting state of lights) based on the acquired inspection sensor data and the accumulated past inspection sensor data;
A vehicle inspection system equipped with:

The invention (1) makes it possible to perform inspections automatically, reducing the burden on the driver.

(2) A failure data acquisition unit (e.g., a failure sensor data acquisition module, a failure operation data acquisition module) that acquires failure sensor data (e.g., vibration sensor data, air pressure sensor data, fluid level sensor data, parking brake sensor data, coolant level sensor data, illuminance sensor data) measured when another vehicle related to the vehicle (e.g., another vehicle 6) was previously determined to be in a failure state, and failure operation data related to the other vehicle's failure (e.g., operation date and time, vehicle ID, delivery origin, delivery destination, delivery route (delivery distance, presence or absence of high altitude), product weight, etc.);
a model generation unit (e.g., a failure determination model generation module) that learns a correlation between the acquired failure-time sensor data and the failure-time operation data and generates a failure determination model;
Further provided with
The inspection data acquisition unit (e.g., an inspection operation data acquisition module) acquires inspection operation data related to operation data at the time of inspection of the vehicle,
The inspection execution unit applies the failure determination model to the acquired inspection sensor data and the inspection operation data to inspect whether or not there is a failure in the vehicle.
A vehicle inspection system as described in (1).

According to the invention (2), faults can be learned and automatically inspected.

(3) an ignition data acquisition unit (e.g., an ignition data acquisition module) that acquires ignition data indicating whether an ignition switch of the vehicle has been turned on or off;
Further provided with
The inspection execution unit executes an inspection when acquiring ignition data.
A vehicle inspection system according to (1) or (2).

The invention (3) makes it easier to perform inspections reliably.

(4) The inspection execution unit executes the inspection at a preset time.
A vehicle inspection system according to (1) or (2).

The invention (4) makes it easier to perform inspections reliably.

(5) The inspection data acquisition unit (e.g., an inspection operation data acquisition module) acquires inspection operation data related to operation data at the time of inspection of the vehicle,
a notification unit (for example, an inspection result notification module) that notifies the user that a problem has been detected in the inspection item as a result of the inspection and the operation data at the time of the inspection;
The vehicle inspection system according to (1) further comprises:

The invention (5) makes it easier to identify vehicles with problems.

(6) a notification unit (for example, a fault notification module) that notifies the user of the fault and the vehicle's operation data at the time of inspection if a fault is found in the inspection item as a result of the inspection;
The vehicle inspection system according to (2) further comprises:

The invention (6) makes it easier to identify vehicles with malfunctions.

(7) An other vehicle driver driving tendency data acquisition unit (e.g., other vehicle driver driving tendency data acquisition module) that acquires other vehicle driver driving tendency data regarding the driving tendency of the driver of the other vehicle (e.g., whether or not there is sudden braking, whether or not there is sudden acceleration, whether or not there is early or late shift change, whether or not there is sudden steering),
an other vehicle operation day weather data acquisition unit (e.g., other vehicle operation day weather data acquisition module) that acquires other vehicle operation day weather data related to the weather on the operation day of the other vehicle (e.g., date and time, region, presence or absence of heavy rain, presence or absence of storms, presence or absence of waves, presence or absence of high tides);
Further provided with
the model generation unit learns correlations between the acquired sensor data at the time of failure, the operation data at the time of failure, the driving tendency data of the driver of the other vehicle, and the weather data on the day the other vehicle is operated, and generates the failure determination model.
(2) A vehicle inspection system according to (2).

The invention (7) makes it possible to respond to driving habits and bad weather before long-distance travel, and to prevent breakdowns.

(8) A vehicle driver driving tendency data acquisition unit (for example, a vehicle driver driving tendency data acquisition module) that acquires vehicle driver driving tendency data related to the driving tendency of the driver of the vehicle;
a vehicle operation day weather data acquisition unit (e.g., a vehicle operation day weather data acquisition module) that acquires vehicle operation day weather data relating to the weather on the vehicle operation day;
Further provided with
The inspection implementation unit inspects the vehicle for the presence or absence of a malfunction by applying the malfunction determination model to the acquired inspection sensor data, the inspection operation data, the vehicle driver's driving tendency data, and the vehicle operation day weather data.
(7) A vehicle inspection system according to (7).

The invention (8) makes it possible to respond to driving habits and bad weather before long-distance travel, and to prevent breakdowns.

(9) A vehicle inspection method executed by a computer that performs vehicle inspection, comprising:
A step (e.g., step S20) of acquiring inspection sensor data measured by a sensor provided in the vehicle;
A step (e.g., step S23) of inspecting the vehicle for each inspection item based on the acquired inspection sensor data and the accumulated past inspection sensor data;
A vehicle inspection method comprising:

(10) A computer that performs vehicle inspections,
A step of acquiring inspection sensor data measured by a sensor provided in the vehicle (e.g., step S20);
A step of inspecting the vehicle for each inspection item based on the acquired inspection sensor data and the accumulated past inspection sensor data (e.g., step S23);
A computer-readable program for executing the program.

1 Vehicle inspection system 3 Network 5 Vehicle 6 Other vehicles 7, 8 ECU
10 Computer 10
11 Inspection data acquisition unit 12 Inspection execution unit

Claims (10)

A vehicle inspection system for inspecting a vehicle,
an inspection result acquisition unit that acquires inspection result data regarding inspection results for each inspection item of the vehicle;
an inspection data acquisition unit that acquires inspection sensor data measured by a sensor provided in the vehicle;
an inspection model generation unit that learns the correlation between the acquired inspection sensor data and the inspection result data and generates an inspection model;
an inspection implementation unit that applies the acquired inspection sensor data to the inspection model to inspect the vehicle for each inspection item;
A vehicle inspection system equipped with: a failure data acquisition unit that acquires failure sensor data measured when another vehicle related to the vehicle is determined to be in a failure state and failure operation data related to operation data of the other vehicle when the other vehicle is in a failure state;
a model generation unit that learns a correlation between the acquired sensor data at the time of failure and the operational data at the time of failure and generates a failure determination model;
Further provided with
the inspection data acquisition unit acquires inspection operation data relating to operation data at the time of inspection of the vehicle;
The inspection execution unit applies the failure determination model to the acquired inspection sensor data and the inspection operation data to inspect whether or not there is a failure in the vehicle.
The vehicle inspection system according to claim 1 . an ignition data acquisition unit that acquires ignition data indicating whether an ignition switch of the vehicle has been turned on or off;
Further provided with
The inspection execution unit executes an inspection when acquiring ignition data.
3. A vehicle inspection system according to claim 1 or 2. The inspection execution unit executes the inspection at a preset time.
3. A vehicle inspection system according to claim 1 or 2. the inspection data acquisition unit acquires inspection operation data relating to operation data at the time of inspection of the vehicle;
a notification unit that notifies the driver that a problem has been detected in the inspection item as a result of the inspection and the inspection-time operation data when the problem has been detected;
The vehicle inspection system of claim 1 further comprising: a notification unit that notifies the driver of the presence of a malfunction in the inspection item as a result of the inspection and the vehicle's operation data at the time of inspection;
The vehicle inspection system of claim 2 further comprising: an other vehicle driver driving tendency data acquisition unit that acquires other vehicle driver driving tendency data relating to the driving tendency of the driver of the other vehicle;
an other vehicle operation day weather data acquisition unit that acquires other vehicle operation day weather data relating to the weather on the operation day of the other vehicle;
Further provided with
the model generation unit learns correlations between the acquired sensor data at the time of failure, the operation data at the time of failure, the driving tendency data of the driver of the other vehicle, and the weather data on the day the other vehicle is operated, and generates the failure determination model.
The vehicle inspection system according to claim 2 . a vehicle driver driving tendency data acquisition unit that acquires vehicle driver driving tendency data relating to the driving tendency of the driver of the vehicle;
a vehicle operation day weather data acquisition unit that acquires vehicle operation day weather data relating to the weather on the operation day of the vehicle;
Further provided with
The inspection implementation unit inspects the vehicle for the presence or absence of a malfunction by applying the malfunction determination model to the acquired inspection sensor data, the inspection operation data, the vehicle driver's driving tendency data, and the vehicle operation day weather data.
The vehicle inspection system according to claim 7. A computer-implemented vehicle inspection method for inspecting a vehicle, comprising:
acquiring inspection result data relating to inspection results for each inspection item of the vehicle;
acquiring inspection sensor data measured by a sensor provided in the vehicle;
A step of learning a correlation between the acquired inspection sensor data and the inspection result data and generating an inspection model;
applying the acquired inspection sensor data to the inspection model to inspect the vehicle for each inspection item;
A vehicle inspection method comprising: The computer that performs the vehicle inspection
acquiring inspection result data relating to inspection results for each inspection item of the vehicle;
acquiring inspection sensor data measured by a sensor provided in the vehicle;
A step of learning a correlation between the acquired inspection sensor data and the inspection result data and generating an inspection model;
applying the acquired inspection sensor data to the inspection model to inspect the vehicle for each inspection item;
A computer-readable program for executing the program.