KR102262648B1 - Vehicle trouble diagnosis system - Google Patents

Abstract

The present invention relates to a vehicle disorder diagnosis system comprising: a diagnosis server which collects the data of status of parts, disorder history, and expected disorder of the parts from a large number of vehicles, stores the data provided from the plurality of vehicles as expected disorder reference data in accordance with the classification criteria including the type of vehicles, engine displacement, and an engine, transmission, suspension, and steering in accordance with the production year, and transmits the data for learning module for expected disorders meeting the driving conditions of the vehicles among the stored expected disorder reference data in accordance with a request of the vehicles; and a vehicle diagnosis unit installed on the vehicles to collect the data on the status of each part and the presence or absence of disorder, provide the data to the diagnosis server, and calculate the parts with expected disorder and the expected probability of disorder of the vehicles by using the expected disorder reference data provided from the diagnosis server. The vehicle diagnosis unit includes: a surrounding sound collection unit which measures a first to third operation sound measured in the vehicle, in the bonnet, and outside the vehicle; and a parts diagnosis unit which performs machine learning using the data for learning module for expected disorders provided from the diagnosis server with the first to third operation sound collected by the surrounding sound collection unit as an independent variable, and expects a disorder in an apparatus or part in the vehicles. The present invention aims to provide a vehicle disorder diagnosis system which is able to minimize errors in the diagnosis of a vehicle's disorder.

Description

Vehicle trouble diagnosis system

The present invention relates to a vehicle failure diagnosis system and a method for diagnosing the same, and more particularly, by machine learning the vehicle state and failure data collected from a plurality of vehicles to generate failure prediction reference data, and predict failure in each vehicle. It relates to a vehicle failure diagnosis system and a method for diagnosing a vehicle failure capable of predicting failure in advance using reference data.

Recently, electronic functions are increasing more and more in vehicles, and accordingly, various electronic devices are installed and used in the vehicle. These electronic devices sometimes replace functions existing in the vehicle, but are more often installed to implement new functions. Accordingly, as electronic devices are added to the vehicle along with the existing mechanical devices, the number of parts that may be damaged is greatly increased.

With the increase of these parts, electronic products for fault diagnosis that can automatically diagnose the failure of some functions of electronic devices or mechanical devices are being released and used. However, these electronic products are not only capable of diagnosing some parts, but also notify that a failure has occurred after the failure occurs. In addition, it is all about notifying the replacement cycle of vehicle consumables for which the replacement cycle is generally set, such as engine oil, transmission oil, spark plug, and timing belt.

Accordingly, since it is difficult to detect a sign of a vehicle failure until the vehicle breaks down, the vehicle may malfunction and stop on the road or cause an accident. Also, there were cases where consumables that were still usable were exchanged. Therefore, there is a need to develop a failure diagnosis system capable of predicting the occurrence of a failure in advance.

As an example of such a technology, Korean Patent Registration No. 10-1974347 "Vehicle Failure Diagnosis System and Diagnosis Method Thereof" can be cited.

The present invention provides an apparatus for diagnosing a vehicle failure that can reduce errors in vehicle diagnosis as much as possible by separating sounds related to driving of a vehicle from ambient noise.

The vehicle failure diagnosis system according to the present invention collects data on the state of parts and failure history and failure prediction data of the parts from a plurality of vehicles, and uses the data provided from the plurality of vehicles to determine the type of vehicle, the engine displacement, Failure prediction learning module that stores failure prediction reference data according to classification criteria including engine, transmission, suspension, and steering according to the year, and meets the operating conditions of the vehicle among the stored failure prediction reference data according to the request of the vehicle a diagnostic server that transmits data; And installed in the vehicle, the data on the state and failure of each part is collected and provided to the diagnosis server, and the parts expected to fail and the probability of failure of the vehicle using the failure prediction reference data provided from the diagnosis server A vehicle diagnosis unit that calculates

The vehicle diagnosis unit may include: an ambient sound collecting unit measuring first to third driving sounds measured inside the vehicle, inside the bonnet, and outside the vehicle; and performing machine learning using the failure prediction learning module data provided from the diagnosis server using the first to third driving sounds collected by the ambient sound collection unit as an independent variable to predict failure of devices or parts of the vehicle. It includes a parts diagnostic unit that does.

It also includes a communication information storage unit for receiving and storing data on the state and failure history of each component of the vehicle from the diagnosis server, wherein the component diagnosis unit includes a failure of each component of the vehicle stored in the communication information storage unit. Data on history can be added as an independent variable.

It also includes a CAN collecting unit that collects data from a plurality of electronic control units in the vehicle, and the parts diagnosis unit may add data collected from the CAN collecting unit as an independent variable.

In addition, the ambient sound collection unit may collect driving sounds by removing non-collection target sounds outside of a preset volume and frequency range that may be generated in the vehicle.

In addition, the parts diagnosis unit may perform a failure determination process through machine learning using the driving sound from which the non-collection target sound is removed.

In addition, the ambient sound collecting unit may learn a noise pattern according to a user environment of the vehicle by using a non-driving sound that is removed during a collection process among sounds generated during normal driving.

In addition, the learning of the noise pattern may be performed for each frequency band of a predetermined unit.

Also includes; an output sound collecting unit for collecting sound data transmitted to be output from the sound output unit of the vehicle, wherein the ambient sound collecting unit corresponds to the sound data collected by the output sound collecting unit from the collected driving sound sound can be removed.

According to the present invention, it is possible to improve the reliability of sound data used for machine learning by learning the usual ambient noise pattern, collecting the sound output from the peripheral device of the vehicle, removing it, and treating it as the vehicle driving sound, thereby improving the reliability of the sound data used in machine learning, such as diagnostic errors, etc. This occurrence can be minimized.

In addition, by using the noise-minimized car driving sound as the main data for vehicle failure diagnosis, and by using the age of vehicle parts and parts as auxiliary data for judging it, the diagnosis is made based on the actual failure rather than the prediction basis. With a high probability, it is possible to determine whether a car breakdown has occurred.

1 is a block diagram illustrating a connection relationship of a vehicle failure diagnosis system according to an embodiment of the present invention.
2 is a block diagram illustrating a vehicle diagnosis unit according to an exemplary embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. Unless there is a specific definition or reference, the terms indicating the direction used in this description are based on the state indicated in the drawings. Also, the same reference numerals refer to the same members throughout each embodiment. On the other hand, each component shown in the drawings may have an exaggerated thickness or dimension for convenience of description, and does not mean that it should actually be configured in a ratio between the corresponding dimensions or components.

A vehicle failure diagnosis system according to an exemplary embodiment will be described with reference to FIG. 1 . 1 is a block diagram illustrating a connection relationship of a vehicle failure diagnosis system according to an embodiment of the present invention.

The vehicle failure diagnosis system according to the present invention includes a diagnosis server 5 and a vehicle diagnosis unit provided in the vehicle 10 .

The diagnosis server 5 transmits and receives data to and from the vehicle 10 through the network 3 , and the vehicle 10 connects to the network 3 using an access point such as a wireless gateway 7 .

The diagnostic server 5 collects data on the state of parts and failure histories and failure prediction data of the parts from a plurality of vehicles 10, and collects the data provided from the plurality of vehicles 10, the type of vehicle, and the engine. According to the classification criteria including engine, transmission, suspension, and steering according to engine, transmission, suspension, and steering according to engine, transmission, suspension, and steering according to engine, engine, transmission, suspension, and steering according to engine displacement and year, it is stored as reference data, and thereafter, according to the request of the vehicle 10, among the stored failure prediction reference data, which meets the operating conditions of the vehicle. Transmit failure prediction learning module data.

The vehicle 10 includes a vehicle diagnosis unit 110 as shown in FIG. 2 . The vehicle diagnosis unit collects data on the status and failure of each component and provides it to the diagnosis server 5, and uses the failure prediction reference data provided from the diagnosis server 5 to predict the failure of parts and failure probability in the vehicle. to calculate Specific details of the vehicle diagnosis unit 110 will be described in detail below.

A vehicle diagnosis unit according to an exemplary embodiment will be described with reference to FIG. 2 . 2 is a block diagram illustrating a vehicle diagnosis unit according to an exemplary embodiment.

As described above, the vehicle diagnosis unit 110 is installed in the vehicle to collect data on the state and failure of each part of the vehicle and provide it to the diagnosis server to collect data on the health of the vehicle, On the other hand, using the failure prediction reference data calculated from the diagnosis server, parts and failure probability of the vehicle expected to fail are calculated.

Specifically, the ambient sound collection unit 210 , the output sound collection unit 220 , the CAN collection unit 230 , the navigation unit 240 , the parts diagnosis unit 250 , the communication information storage unit 260 and the communication unit 270 ) to provide

The ambient sound collecting unit 210 measures first to third driving sounds measured inside the vehicle, inside the bonnet, and outside the vehicle. That is, the first driving sound refers to sounds generated in the interior of the vehicle, and includes all of the ambient noise flowing into the vehicle, the noise generated inside the vehicle, the sound output through the vehicle's internal sound device, and the conversation occurring inside the vehicle. include The second driving sound is measured inside the bonnet and includes sounds related to mechanical operation and other noises generated in the engine room at the front or rear of the vehicle. The third driving sound refers to a sound outside the vehicle and includes various sounds generated outside the vehicle.

In this case, the ambient sound collection unit 210 may collect driving sounds by removing non-collection target sounds outside of a preset volume and frequency range that may be generated in the vehicle. In particular, in the case of measuring the second driving sound, it is possible to preset the volume and frequency band of the normal machine operation sound that may occur in the engine room and the machine operation sound in case of an abnormal failure, and remove the sound outside this range. .

As described above, sounds that are removed while the ambient sound collecting unit 210 collects sounds among sounds generated while the vehicle is driving are defined as non-driving sounds for convenience of explanation.

Such off-road sounds may be used as basic data for learning a noise pattern according to a user environment of the vehicle, and the learning of the noise pattern may be performed for each frequency band of a predetermined unit.

Standard noise is normally generated at each location, but when an abnormality starts to occur, a noise different from the normal noise starts to be generated. The noise collected through the ambient sound collecting unit 21 may be collected together with information on a location where the noise is recorded. In addition, the noise recorded through the ambient sound collection unit 210 may be collected by matching the driving distance of the vehicle. In this case, the noise when the vehicle is new may be separately collected and used as the reference noise in the diagnosis server. .

Also, the ambient sound collection unit 210 may remove a human voice range from the second driving sound.

Also, the ambient sound collecting unit 210 may remove sounds corresponding to the sound data collected by the output sound collecting unit from the collected first to third driving sounds.

The output sound collection unit 220 collects sound data transmitted to be output through a vehicle sound output unit, ie, a vehicle speaker. As described above, sounds not related to the driving of the vehicle may be removed by removing the sound corresponding to the output sound data collected by the output sound collecting unit 220 from the sound collected in the room.

The CAN collection unit 230 collects data from a plurality of electronic control units in the vehicle. The CAN collection unit 230 may collect data using a controller area network (CAN), which is a communication protocol for a vehicle. A number of electronic control units (ECUs) are installed in the vehicle, and the electronic control unit is a control device that controls and manages various electronic devices installed in the vehicle, and exchanges data using CAN. can do.

The CAN collection unit 230 may collect data from each electronic control unit using CAN communication, and the CAN collection unit 230 includes an engine electronic control unit, a brake electronic control unit, a steering electronic control unit, and a suspension electronic control unit. It is also possible to collect data directly from various electronic control units, such as, or collect information about each electronic control unit and various sensors through an on-board diagnosis (OBD) terminal.

For example, the CAN collection unit 230 may receive and collect the angle at which the driver turns the steering wheel and the speed at which the driver turns the steering from the electronic steering control unit, and the number of times the brake is applied from the electronic brake control unit, the characteristics of the driver applying the brake It can be collected and provided. In addition, the CAN collection unit 230 may collect information about the operation range or number of operations of the suspension from the suspension electronic control unit.

One of LIN (Local Interconnect Network) or Flexray may be used as the vehicle communication protocol, and in this case, data may be collected from each electronic control unit using LIN or Flexray.

The data collected from the CAN collection unit 230 is matched with the sound collected by the driving sound collection unit 210 on the timeline to determine whether it is a sound according to the driver's control action or automatic control of the vehicle. can be used as

The navigation unit 240 may collect information on the road on which the vehicle travels, and may provide information on whether the vehicle is traveling on an unpaved road or an uphill road. If the vehicle is mainly driven on unpaved roads or uphill roads, the probability of transmission failure may increase. On the other hand, the transmission probability may vary depending on the number of gear shifts or the driver's habit on gear shifting, and this information may also be collected and provided to the diagnosis server. As described above, the data collected from the navigation unit 240 may be used as data that is based on determination of a driving sound matched on the timeline by providing basic data on the driving environment of the vehicle.

The parts diagnosis unit 250 performs machine learning using the failure prediction learning module data provided from the diagnosis server using the first to third driving sounds collected by the ambient sound collecting unit as an independent variable to determine the device or part of the vehicle. Anticipate or predict failure.

In addition, the parts diagnosis unit 250 adds data on the state and failure history of each part of the vehicle transmitted from the diagnosis server and stored in the communication information storage unit 260 as an independent variable to cause failure of the device or component of the vehicle. can be predicted. In addition, the parts diagnosis unit 250 may add data related to the control and state of the vehicle collected from the CAN collection unit 230 as an independent variable. In this case, the parts diagnosis unit 250 may perform a failure determination process through machine learning using a non-collection target sound or a driving sound from which a non-driving sound is removed.

In the case of the existing vehicle failure prediction system, it is determined that there is a probability that a failure will occur if the noise data of the own vehicle collected through the surrounding sound collection unit and the noise data generated when a failure occurs as the failure prediction reference data are matched by a certain percentage or more. However, in this case, matching was difficult due to various noises, and in some cases, incorrect results were obtained. The parts diagnosis unit 250 according to the present embodiment removes as much as possible non-driving sounds that are not related to driving from the surrounding sound collection unit 210, so that the actual driving-related sound is used as a basis for judging the machine learning, thereby improving the efficiency of judgment. can improve

The communication information storage unit 260 receives and stores data on the status and failure history of each part of the vehicle from the diagnosis server. At this time, the data on the state of each part of the vehicle stored in the communication information storage unit 260 may include not only historical data transmitted from the vehicle to the diagnostic server in the past, but also general reference data related to the vehicle model and year of the vehicle. Similarly, the data on the failure history may include failure probability data for each part related to the model and year of the vehicle.

In the parts diagnosis unit 250, such a barrel? When data stored in the storage unit 260 is used, machine learning may be performed by using data of the corresponding vehicle and at the same time using the general state or failure probability data of the corresponding vehicle.

Meanwhile, the communication information storage unit 260 may further store data collected through the CAN collection unit 230 , data collected through the navigation unit 25 , and the like.

The communication unit 270 may support communication with the diagnostic server described above. In order to communicate with the diagnostic server, V2X (Vehicle to Everything) must be used. V2X is a technology including V2V (Vehicle to Vehicle), V2I (Vehicle to Infrastructure), and V2N (Vehicle to Nomadice device). It uses an access point such as a road-side unit (RSU) that is installed and communicates with the vehicle. The communication unit 270 may use a Dedicated Short Range Communication (DSRC) scheme of the 5.8 GHz band having high accuracy and high instantaneous communication speed.

Although the preferred embodiments of the present invention have been described above, the technical spirit of the present invention is not limited to the above-described preferred embodiments, and may be implemented in various ways within the scope not departing from the technical spirit of the present invention embodied in the claims. have.

110: vehicle diagnostic unit
210: ambient sound collection unit
220: output sound collection unit
230: CAN collection unit
240: navigation
250: parts diagnosis unit
260: communication information storage unit
270: communication department

Claims (8)

It collects data on the condition and failure history of parts and failure prediction data of the parts from a plurality of vehicles, and collects the data provided from the plurality of vehicles for the type of vehicle, engine displacement, engine, transmission, suspension, a diagnosis server that stores failure prediction reference data according to classification criteria including steering, and transmits failure prediction learning module data matching the driving conditions of the vehicle among the stored failure prediction reference data according to the request of the vehicle; and
It is installed in the vehicle and collects data on the state and failure of each part and provides it to the diagnosis server, and uses the failure prediction reference data provided from the diagnosis server to determine which parts are expected to fail and the probability of failure a vehicle diagnosis unit that calculates;
The vehicle diagnosis unit,
an ambient sound collecting unit measuring first to third driving sounds measured inside the vehicle, inside the bonnet, and outside the vehicle; and
Using the first to third driving sounds collected by the ambient sound collection unit as an independent variable, machine learning is performed using the failure prediction learning module data provided from the diagnosis server to predict the failure of the device or parts of the vehicle. Vehicle failure diagnosis system including; parts diagnosis unit.
The ambient sound collection unit collects driving sounds by removing non-collection target sounds outside of the volume and frequency range that can be generated in a preset vehicle,
The parts diagnosis unit performs a failure determination process through machine learning using the driving sound from which the non-collection target sound has been removed,
The ambient sound collecting unit learns a noise pattern according to a user environment of the vehicle by using a non-driving sound removed during a collection process among sounds generated during normal driving. According to claim 1,
The vehicle diagnosis unit includes a communication information storage unit for receiving and storing data on the status and failure history of each part of the vehicle from the diagnosis server,
The component diagnosis unit adds data on the failure history of each component of the vehicle stored in the communication information storage unit as an independent variable. According to claim 1,
The vehicle diagnosis unit includes a CAN collection unit that collects data from a plurality of electronic control units in the vehicle;
The component diagnosis unit adds the data collected from the CAN collection unit as an independent variable. delete delete delete According to claim 1,
The learning of the noise pattern is performed for each frequency band of a predetermined unit. According to claim 1,
The vehicle diagnosis unit includes an output sound collection unit that collects sound data transmitted to be output from a sound output unit of the vehicle;
The ambient sound collecting unit removes a sound corresponding to the sound data collected by the output sound collecting unit from the collected driving sound.

Images