Vehicle maintenance system using odb data and method performing thereof

Abstract

A vehicle maintenance system using OBD data according to the present invention comprises: a user terminal that collects and provides vehicle internal data from an ODB device that monitors the status of each of a plurality of parts of a vehicle and collects vehicle internal data; a vehicle maintenance brokerage server that collects and analyzes data for vehicle maintenance from the user terminal, an A / S center terminal, and an external server to predict the likelihood of failure and recommends an A / S center based on the prediction results; an A / S center entry / exit management device that recognizes the vehicle number of a customer entering the A / S center and receives and provides ODB data from the user terminal; and an A / S center terminal that receives vehicle entry information and ODB data from the A / S center entry / exit management device, and generates and provides a work order when a maintenance type is entered after a consultation between a mechanic and a customer is completed based on the ODB data.

Description

Technology Field

[0001] The present invention relates to a vehicle maintenance system using OBD data and a method for implementing the same. More specifically, it relates to a vehicle maintenance system using OBD data and a method for implementing the same that analyzes data collected through communication with an OBD device or an in-vehicle IoT sensor to predict the possibility of failure in advance and notifies the user and the service center of this, thereby enabling vehicle maintenance to be planned in advance. Background Technology

[0003] Vehicle technology has advanced dramatically over the past few decades, transforming from a simple means of transportation into a smart mobility platform that combines advanced electronics and software.

[0004] While these advancements have contributed to significantly improving vehicle performance and safety, they have simultaneously presented new challenges to vehicle maintenance and management methods. In particular, the introduction of technologies such as OBD-II (On-Board Diagnostics), IoT (Internet of Things), and AI (Artificial Intelligence) has laid the foundation for performing vehicle maintenance in a more systematic and predictive manner.

[0005] With the increasing adoption of advanced technologies, modern vehicles are providing the capability to monitor their condition in real time through various sensors. IoT sensors capable of collecting data such as engine status, oil change intervals, brake pad wear, and battery voltage are embedded in the vehicles. Furthermore, standardized interfaces that allow internal vehicle data to be transmitted to external devices via OBD-II devices are also widely utilized.

[0006] However, the utilization of this data remains limited. Generally, vehicle maintenance is carried out through reactive measures, such as periodic inspections or visiting a service center after a breakdown occurs. This can lead to the following problems.

[0007] First, because maintenance is performed only at fixed intervals without considering factors such as parts aging, usage environment, or driving patterns, unnecessary maintenance may occur, or conversely, serious vehicle damage may result from unexpected breakdowns.

[0008] Second, since vehicle owners typically visit service centers only after a breakdown occurs, there is a problem of long waiting times due to centers exceeding their processing capacity during emergencies. Additionally, repair schedules are sometimes delayed because necessary parts are not prepared in a timely manner.

[0009] Third, while vehicle manufacturers and repair shops possess some technology to utilize vehicle data, there is a lack of systematic platforms capable of analyzing this data in real time to predict maintenance timing or provide maintenance services optimized for users.

[0010] These issues not only cause economic and time losses for both vehicle owners and repair shops but can also have an adverse effect on vehicle safety. Accordingly, there is a growing need for technology that collects vehicle condition data in real time, predicts maintenance timing in advance based on this data, and recommends appropriate services. In particular, the advancement of big data analysis and artificial intelligence (AI) technologies is attracting attention as a key means to solve these problems. The problem to be solved

[0012] The present invention aims to provide a vehicle maintenance system using OBD data and a method for implementing the same, which analyzes data collected through communication with an OBD device or an in-vehicle IoT sensor to predict the possibility of failure in advance and notifies the user and the service center of this, thereby enabling vehicle maintenance to be planned in advance.

[0013] In addition, the present invention aims to provide a vehicle maintenance system using OBD data and a method for implementing the same, which enables a mechanic to check the vehicle status in real time by providing OBD data pre-stored in a user terminal or vehicle key to an A / S center terminal when visiting an A / S center.

[0014] In addition, the present invention aims to provide a vehicle maintenance system using OBD data and a method for implementing the same, which can calculate the timing for parts replacement and propose an optimized maintenance schedule by comprehensively analyzing the usage cycle of parts, vehicle driving habits, driving conditions, and external environment data based on OBD data. means of solving the problem

[0016] A vehicle maintenance system using OBD data to achieve this purpose includes: a user terminal that collects and provides internal vehicle data from an ODB device that monitors the status of each of multiple parts of the vehicle and collects internal vehicle data; a vehicle maintenance brokerage server that collects and analyzes data for vehicle maintenance from the user terminal, an A / S center terminal, and an external server to predict the likelihood of failure and recommends an A / S center based on the prediction results; an A / S center entry / exit management device that recognizes the vehicle number of a customer entering the A / S center and receives and provides ODB data from the user terminal; and an A / S center terminal that receives vehicle entry information and ODB data from the A / S center entry / exit management device, and generates and provides a work order when a maintenance type is entered after a consultation between a mechanic and a customer is completed based on the ODB data.

[0017] In one embodiment, the vehicle maintenance brokerage server can define variables that may affect failure prediction according to the characteristics of vehicle parts and the driving environment, and then input them into the failure learning model to learn a failure learning model that outputs a failure probability.

[0018] In one embodiment, the vehicle maintenance intermediary server inputs data for vehicle maintenance into the fault learning model to receive the current fault probability, and can predict the time when the fault probability reaches a threshold starting from the current fault probability and provide it to the user terminal.

[0019] In one embodiment, the vehicle maintenance brokerage server may provide information regarding the maintenance cycle to a user terminal in the form of a push notification or dashboard, and upon receiving a reservation request message in response thereto, extract a list of service centers and provide it to the user terminal.

[0020] In addition, a vehicle maintenance method using OBD data to achieve this purpose includes the steps of: a user terminal collecting internal vehicle data from an ODB device that collects internal vehicle data by monitoring the status of each of a plurality of vehicle parts; a vehicle maintenance intermediary server collecting and analyzing data for vehicle maintenance from the user terminal, an A / S center terminal, and an external server to predict the likelihood of failure, recommending an A / S center based on the prediction results, and then booking a specific A / S center; an A / S center entry / exit management device recognizing the vehicle number of a customer entering the A / S center, and if the customer is a booked customer, receiving ODB data from the user terminal; and an A / S center terminal receiving vehicle entry information and ODB data from the A / S center entry / exit management device, and, based on the ODB data, generating and providing a work order when a maintenance type is entered after a consultation between a mechanic and a customer is completed. Effects of the invention

[0022] According to the present invention as described above, there is an advantage in that data collected through communication with an OBD device or an in-vehicle IoT sensor is analyzed to predict the possibility of failure in advance, and vehicle maintenance can be planned in advance by notifying the user and the service center.

[0023] In addition, according to the present invention, there is an advantage that a mechanic can check the vehicle status in real time by providing ODB data, which is pre-stored in the user terminal or vehicle key, to the A / S center terminal when visiting the A / S center.

[0024] In addition, according to the present invention, there is an advantage in that the timing for replacing parts can be calculated and an optimized maintenance schedule proposed by comprehensively analyzing the usage cycle of parts, vehicle driving habits, driving conditions, and external environment data based on ODB data. Brief explanation of the drawing

[0026] FIG. 1 is a network configuration diagram for explaining a vehicle maintenance system using OBD data according to an embodiment of the present invention. FIG. 2 is a drawing for explaining an entry and exit management device for an after-sales service center according to an embodiment of the present invention. FIG. 3 is a block diagram illustrating the internal structure of a vehicle maintenance brokerage server according to one embodiment of the invention. FIG. 4 is a flowchart illustrating an embodiment of a vehicle maintenance method using OBD data according to the present invention. Specific details for implementing the invention

[0027] The aforementioned objectives, signatures, and advantages are described in detail below with reference to the attached drawings, thereby enabling those skilled in the art to easily implement the technical concept of the present invention. In describing the present invention, detailed descriptions of known technologies related to the present invention are omitted if it is determined that such descriptions would unnecessarily obscure the essence of the invention. Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the attached drawings. In the drawings, the same reference numerals are used to indicate the same or similar components.

[0029] FIG. 1 is a network configuration diagram for explaining a vehicle maintenance system using OBD data according to an embodiment of the present invention. FIG. 2 is a diagram for explaining an entry / exit management device for an after-sales service center according to an embodiment of the present invention.

[0030] Referring to FIGS. 1 and 2, a vehicle maintenance system using OBD data includes a vehicle (100), an OBD (On Board Dignosis) device (200), a user terminal (300), a vehicle maintenance intermediary server (400), an after-sales service center terminal (500), and an after-sales service center entry / exit management device (600).

[0031] An OBD (On Board Dignosis) device (200) is formed in the vehicle (100). Accordingly, the OBD device (200) refers to a self-diagnostic connector mounted inside the vehicle (100) to check the vehicle condition, and the OBD connector is provided on the dashboard side of the center fascia of the vehicle.

[0032] This OBD device (200) monitors various parts of the vehicle and collects internal vehicle data.

[0033] In one embodiment, the OBD device (200) monitors and collects engine status, battery voltage, brake pad wear status, oil status, fuel efficiency, temperature, etc. in real time. At this time, engine status includes engine temperature, speed, fuel efficiency, and exhaust gas status, battery voltage and status include battery remaining capacity and charge status, error codes (DTC, Diagnostic Trouble Codes) include error codes that occurred in the engine, brake system, electrical system, etc., and driving distance may include total driving distance, driving time, RPM (revolutions per minute), etc.

[0034] The user terminal (300) may be a smartphone, tablet PC, or other portable terminal owned by the driver or owner of the vehicle. The user terminal (300) may have an application installed that communicates with the vehicle maintenance brokerage server (400) and the OBD device (200) to collect internal vehicle data, convert and filter the data, and then transmit it to the vehicle maintenance brokerage server (400).

[0035] The user terminal (300) can collect vehicle data through communication with the OBD device (200) or an in-vehicle IoT sensor and provide it to the vehicle maintenance brokerage server (400). If the vehicle key communicates with the OBD device (200) to collect vehicle data, the user terminal (300) can collect vehicle data from the vehicle key and provide it to the vehicle maintenance brokerage server (400).

[0036] The vehicle maintenance brokerage server (400) collects and analyzes data for vehicle maintenance to predict the likelihood of failure and recommends an optimal after-sales service center. To this end, the vehicle maintenance brokerage server (400) can collect vehicle internal data and user driving data from a user terminal (300), collect vehicle maintenance data from an after-sales service center terminal (500), and collect external environment data from an external server (e.g., a weather agency server).

[0037] The aforementioned internal vehicle data may include engine temperature, battery voltage, brake pad wear, fuel efficiency, mileage, oil status, and fault codes (DTCs), while external environmental data may include weather, temperature, humidity, road conditions, driving route, and traffic flow. For example, temperature changes can be used to predict battery performance degradation when temperatures are low during winter, and road conditions can be used to predict the impact on tires and suspension systems when driving on frequently bumpy roads.

[0038] User driving data may include driving habits (frequency of sudden acceleration / sudden braking), driving patterns (long distance / short distance), etc. For example, the frequency of sudden acceleration can be used to predict the likelihood of failure of engine parts or fuel systems if sudden acceleration occurs frequently, the frequency of sudden braking can be used to predict the likelihood of wear of brake pads or tires if sudden braking occurs frequently, and driving style can be used to predict the likelihood of failure by analyzing the vehicle's average driving speed, driving distance, and driving patterns frequently used by the driver.

[0039] Maintenance data can include vehicle maintenance history, parts replacement records, maintenance costs, and time. For example, brake pad replacement history can be used to predict wear and failure probability based on replacement cycles or usage conditions, while battery replacement history can be used to predict the current battery's failure probability based on how many years ago it was replaced.

[0040] First, the vehicle maintenance brokerage server (400) defines variables that can affect failure prediction for each vehicle part.

[0041] In one embodiment, the vehicle maintenance brokerage server (400) can define variables that may affect failure prediction depending on the characteristics of the vehicle parts and the driving environment.

[0042] For example, if the vehicle maintenance brokerage server (400) is a brake pad, it can determine the input variables as brake pad wear, mileage, frequency of sudden braking, road condition, temperature, humidity, etc., and determine the target variable as the probability of brake pad failure.

[0043] In another example, if the vehicle maintenance brokerage server (400) is a battery, it can determine the input variables as battery voltage, temperature, and driving pattern, and determine the target variable as the probability of battery failure.

[0044] In another embodiment, the vehicle maintenance brokerage server (400) can analyze internal vehicle data and past failure history to extract relevant variables for each vehicle part. This is to define variables that affect the failure probability of each part.

[0045] In the above embodiment, the vehicle maintenance brokerage server (400) can determine the time of failure among past failure history as a variable that influences the probability of failure by identifying specific cycles or conditions in which specific parts frequently fail. For example, in the case of a battery, the probability of failure increases after a certain usage period.

[0046] After that, the vehicle maintenance brokerage server (400) defines variables that affect the prediction of failure of vehicle parts and inputs them to learn a failure learning model that outputs a failure probability.

[0047] To this end, the vehicle maintenance brokerage server (400) collects data on various variables that may affect the likelihood of failure for each vehicle part, preprocesses them, and trains a failure learning model.

[0048] In one embodiment, the vehicle maintenance brokerage server (400) learns to apply weights and biases to each input variable to perform a linear transformation and then output a failure probability through a non-linear activation function.

[0049] In the above embodiment, the vehicle maintenance brokerage server (400) may apply different weights and biases depending on the influence of the input variable on the failure probability.

[0050] For example, if the input variables are brake pad wear (x1), mileage (x2), and sudden braking frequency (x3), and the weights are the weight for brake pad wear (w1), the weight for mileage (w2), and the weight for sudden braking frequency (w3), the weight (w1) will be larger than that of the other variables because brake pad wear (x1) has a significant impact on the failure probability. Similarly, although mileage (x2) affects brake pad wear, its influence is not as great as that of brake pad wear, so the weight (w2) may be relatively small. Additionally, while sudden braking frequency (x3) is an important variable, its influence may be less than that of mileage or wear, so the weight (w3) may be smaller.

[0051] The vehicle maintenance brokerage server (400) can generate a linear output as shown in [Equation 1] by applying different weights and biases according to the influence of the input variable on the failure probability.

[0053] [Mathematical Formula 1]

[0054]

[0056] z: linear output,

[0057] x1, x2, x3: Input variables that can affect failure prediction for each vehicle part,

[0058] w1, w2, w3: weights for input variables,

[0059] b: bias,

[0061] In [Equation 1], the bias (b) helps the breakdown learning model generate an appropriate output even when the input value is 0. The bias (b) serves to adjust the output so that the breakdown learning model does not over-predict under certain conditions.

[0062] After that, the vehicle maintenance brokerage server (400) predicts when maintenance is needed for a part based on the failure probability of the part. For example, the vehicle maintenance brokerage server (400) predicts when the failure probability of a brake pad reaches 80% and determines this as the maintenance time.

[0063] First, the vehicle maintenance brokerage server (400) inputs input data (e.g., current state of a part, part usage history, past failure history, environmental data, etc.) into a failure probability model to receive a failure probability. For example, brake pad wear (50%), driving distance (10,000 km), and frequency of sudden braking (5 times / day) can be input into the failure probability model to receive a failure probability of 40%.

[0064] After that, the vehicle maintenance brokerage server (400) predicts when the failure probability reaches a threshold (e.g., 80%), starting from the current failure probability.

[0065] In one embodiment, the vehicle maintenance brokerage server (400) uses a failure probability model to predict how the failure probability changes over time (or usage).

[0066] For example, the vehicle maintenance intermediary server (400) can predict that if the probability of failure increases by 10% for every 1,000 km of driving, the current probability of failure is 40%, and the remaining driving distance to reach a probability of failure of 80% is 4,000 km (i.e., (80%?40%)÷10%=4(in units of 1,000 km)).

[0067] As another example, if the vehicle maintenance brokerage server (400) is not constant in the rate of increase of failure probability (e.g., the rate of increase increases as the wear level increases), it can use a machine learning model (e.g., regression model, curve fitting) to calculate the point at which the failure probability reaches 80% by simulating through the machine learning model starting from the current failure probability of 40%.

[0068] As another example, the vehicle maintenance brokerage server (400) calculates the difference remaining from the current state to the failure threshold by dividing it by the usage rate as in [Equation 2].

[0070] [Mathematical Formula 2]

[0071]

[0072] C current : Current status indicators (i.e., current wear, current usage, etc.),

[0073] C threshold : Threshold at which a component reaches failure or limit state,

[0074] R usage : Rate of increase in usage (e.g., increase in wear per hour)

[0075] T remaining : Time remaining until failure threshold or usage distance

[0077] C in [Mathematical Formula 2] threshold - Ccurrent It measures how much of the component remains until it reaches the critical state by calculating the difference between the current state and the critical threshold. For example, if the current wear is 70% and the failure threshold is 90%, then C current It can be measured as 20%.

[0078] Then, in [Equation 2], the remaining usage is converted into time or distance by considering the rate of increase in usage. For example, if the wear rate per hour is 2%, it is 2% / hour.

[0079] As another example, the vehicle maintenance brokerage server (400) calculates using a differential equation if the usage speed is not constant and is dynamic (e.g., driving conditions, environmental changes) as in [Equation 3].

[0081] [Mathematical Formula 3]

[0082]

[0084] C current: Current status indicators (i.e., current wear, current usage, etc.),

[0085] C threshold : Threshold at which a component reaches failure or limit state,

[0086] R usage : Rate of increase in usage (e.g., increase in wear per hour)

[0087] T remaining : Time remaining until failure threshold or usage distance

[0088] R usage (t): Rate of increase in usage changing over time,

[0090] As described above, the vehicle maintenance brokerage server (400) calculates the remaining usage in the current state and calculates the maintenance cycle. For example, if the current driving distance is 10,000 km, the expected failure probability is 80%, the achievable driving distance is 14,000 km, and the remaining usage distance is 4,000 km, the maintenance cycle can be predicted to be approximately 4,000 km.

[0091] As described above, the vehicle maintenance brokerage server (400) calculates the remaining usage in the current state to determine the maintenance cycle, and then provides information to the user terminal (300) in the form of a push notification or dashboard.

[0092] For example, the vehicle maintenance brokerage server (400) provides information to the user terminal (300) in the form of a push notification or dashboard, stating, "Brake pad replacement is required. The recommended maintenance date is January 15, 2024." Thus, the user can check the current status of parts, the expected probability of failure, the recommended maintenance date, etc., through the notification.

[0093] The vehicle maintenance brokerage server (400) provides maintenance cycle information to the user terminal (300) in the form of a push notification or dashboard, and upon receiving a reservation request message in response thereto, extracts a list of service centers and provides it to the user terminal (300).

[0094] In one embodiment, the vehicle maintenance brokerage server (400) extracts a list of service centers based on the availability of necessary parts for parts with a high probability of failure, the distance between the user's location and the service center, the availability of reservations at the service center, and the service center rating, and provides this list to the user terminal (300).

[0095] In the above embodiment, when a specific service center and visit time are selected from the user terminal (300), the vehicle maintenance brokerage server (400) generates a reservation request message and provides it to the corresponding service center terminal (500).

[0096] The A / S center terminal (500) receives vehicle entry information and ODB data from the A / S center entry / exit management device (200), and generates a work order when a maintenance type (light maintenance, heavy maintenance, accident vehicle, etc.) is entered after a consultation between a mechanic and a customer is completed based on the ODB data. At this time, the work order may include vehicle information, maintenance type, estimated time required, worker assignment, etc.

[0097] After that, the A / S center terminal (500) issues a work order and automatically assigns an appropriate work bay based on the type of work and the characteristics of the work bay. For example, a specific work bay may be specialized for light maintenance, while another work bay may be suitable for heavy maintenance or accident vehicles.

[0098] The A / S center entry / exit management device (600) recognizes the vehicle number of a customer entering or exiting the vehicle A / S center and provides vehicle entry information or vehicle exit information to the vehicle A / S center terminal (300). The A / S center entry / exit management device (100) is located at the entrance and exit of the A / S center and includes, respectively, a vehicle number recognition device (20-1, 20-2), an automatic barrier (22-1, 22-2), a lobby phone (24-1, 24-2), a surveillance camera (26-1, 26-2), and an OBD data receiving device (28).

[0099] In one embodiment, the A / S center entry / exit management device (600) recognizes the vehicle number of a customer entering the A / S center and provides vehicle entry information. It also verifies whether the customer is a customer with a reservation at the A / S center and provides service details, and after the work is completed, recognizes the customer's vehicle number and provides vehicle exit information.

[0100] A vehicle number recognition device (20-1, 20-2) according to one embodiment is configured as an integrated display board and displays vehicle entry / exit information and service details together on the screen.

[0101] The OBD data receiving device (28) has a mounting module formed therein in which a user terminal (300) or a vehicle key can be mounted, and can receive OBD data by communicating with the user terminal (300) or the vehicle key and provide it to the A / S center terminal (500).

[0103] FIG. 2 is a block diagram illustrating the internal structure of a vehicle maintenance brokerage server according to one embodiment of the present invention.

[0104] Referring to FIG. 2, the vehicle maintenance brokerage server (400) includes a data collection unit (410), a fault learning model generation unit (420a), a fault timing prediction unit (430), and an after-sales service center reservation brokerage unit (440).

[0105] The data collection unit (410) collects data for vehicle maintenance. To this end, the data collection unit (410) can collect vehicle internal data and user driving data from the user terminal (300), collect vehicle maintenance data from the A / S center terminal (500), and collect external environment data from an external server (e.g., a weather agency server).

[0106] The aforementioned internal vehicle data may include engine temperature, battery voltage, brake pad wear, fuel efficiency, mileage, oil status, and fault codes (DTCs), while external environmental data may include weather, temperature, humidity, road conditions, driving route, and traffic flow. For example, temperature changes can be used to predict battery performance degradation when temperatures are low during winter, and road conditions can be used to predict the impact on tires and suspension systems when driving on frequently bumpy roads.

[0107] User driving data may include driving habits (frequency of sudden acceleration / sudden braking), driving patterns (long distance / short distance), etc. For example, the frequency of sudden acceleration can be used to predict the likelihood of failure of engine parts or fuel systems if sudden acceleration occurs frequently, the frequency of sudden braking can be used to predict the likelihood of wear of brake pads or tires if sudden braking occurs frequently, and driving style can be used to predict the likelihood of failure by analyzing the vehicle's average driving speed, driving distance, and driving patterns frequently used by the driver.

[0108] Maintenance data can include vehicle maintenance history, parts replacement records, maintenance costs, and time. For example, brake pad replacement history can be used to predict wear and failure probability based on replacement cycles or usage conditions, while battery replacement history can be used to predict the current battery's failure probability based on how many years ago it was replaced.

[0109] The fault learning model generation unit (420a) defines variables that can affect fault prediction for each vehicle part.

[0110] In one embodiment, the fault learning model generation unit (420a) can define variables that may affect fault prediction depending on the characteristics of the vehicle parts and the driving environment.

[0111] For example, when the vehicle part is a brake pad, the failure learning model generation unit (420a) can determine the input variables as brake pad wear, mileage, frequency of sudden braking, road condition, temperature, humidity, etc., and determine the target variable as the brake pad failure probability.

[0112] In another example, when the vehicle component is a battery, the failure learning model generation unit (420a) can determine the input variables as battery voltage, temperature, and driving pattern, and determine the target variable as the battery failure probability.

[0113] In another embodiment, the failure learning model generation unit (420a) can extract relevant variables for each vehicle part by analyzing vehicle internal data and past failure history. This is to define variables that affect the failure probability of each part.

[0114] In the above embodiment, the failure learning model generation unit (420a) can determine the time of failure among past failure history as a variable that influences the probability of failure by identifying specific cycles or conditions where a specific part frequently fails. For example, in the case of a battery, the probability of failure increases after a certain usage period.

[0115] After that, the fault learning model generation unit (420a) defines variables that affect the prediction of failure of vehicle parts and inputs them to learn a fault learning model (420b) that outputs a failure probability.

[0116] To this end, the fault learning model generation unit (420a) collects data on various variables that may affect the failure probability of each vehicle part, preprocesses them, and trains the fault learning model (420b).

[0117] In one embodiment, the failure learning model generation unit (420a) applies weights and biases to each input variable to perform a linear transformation, and then learns to output a failure probability through a non-linear activation function.

[0118] In the above embodiment, the failure learning model generation unit (420a) can apply different weights and biases depending on the influence of the input variables on the failure probability.

[0119] For example, if the input variables are brake pad wear (x1), mileage (x2), and sudden braking frequency (x3), and the weights are the weight for brake pad wear (w1), the weight for mileage (w2), and the weight for sudden braking frequency (w3), the weight (w1) will be larger than that of the other variables because brake pad wear (x1) has a significant impact on the failure probability. Similarly, although mileage (x2) affects brake pad wear, its influence is not as great as that of brake pad wear, so the weight (w2) may be relatively small. Additionally, while sudden braking frequency (x3) is an important variable, its influence may be less than that of mileage or wear, so the weight (w3) may be smaller.

[0120] The failure learning model generation unit (420a) can generate a linear output as shown in [Equation 1] by applying different weights and biases according to the influence of the input variable on the failure probability.

[0121] The failure time prediction unit (430) receives the current failure probability by inputting data for vehicle maintenance into the failure learning model (420b) generated by the failure learning model generation unit (420a), and predicts the time when the failure probability reaches a threshold starting from the current failure probability. For example, the failure time prediction unit (430) predicts the time when the failure probability of the brake pad reaches 80% and determines this as the maintenance time.

[0122] First, the failure time prediction unit (430) inputs input data (e.g., current state of a part, part usage history, past failure history, environmental data, etc.) into a failure probability model to receive a failure probability. For example, brake pad wear (50%), driving distance (10,000 km), and frequency of sudden braking (5 times / day) are input into the failure probability model, and a failure probability of 40% can be received.

[0123] After that, the failure time prediction unit (430) predicts the time when the failure probability reaches a threshold (e.g., 80%), starting from the current failure probability.

[0124] In one embodiment, the failure time prediction unit (430) uses a failure probability model to predict how the failure probability changes over time (or usage).

[0125] For example, the failure time prediction unit (430) can predict that if the failure probability increases by 10% for every 1,000 km of driving, the current failure probability is 40%, and the remaining driving distance to reach a failure probability of 80% is 4,000 km (i.e., (80%?40%)÷10%=4(units of 1,000 km)).

[0126] As another example, the failure time prediction unit (430) can calculate the time when the failure probability reaches 80% by simulating through a machine learning model (e.g., regression model, curve fitting) starting from the current failure probability of 40% when the failure probability increase rate is not constant (e.g., the rate of increase increases as the wear rate increases) using a machine learning model.

[0127] As another example, the failure time prediction unit (430) calculates the difference remaining from the current state to the failure threshold by dividing it by the usage rate as in [Equation 2].

[0128] As another example, the failure time prediction unit (430) calculates using a differential equation if the usage speed is not constant and is dynamic (e.g., driving conditions, environmental changes) as in [Equation 3].

[0129] As described above, the failure time prediction unit (430) calculates the remaining usage in the current state to determine the maintenance cycle and then provides information to the user terminal (300) in the form of a push notification or dashboard.

[0130] For example, the failure time prediction unit (430) provides information to the user terminal (300) in the form of a push notification or dashboard, stating, "Brake pad replacement is required. The recommended maintenance date is January 15, 2024." Thus, the user can check the current part status, expected failure probability, recommended maintenance date, etc. through the notification.

[0131] The A / S center reservation intermediary (440) provides maintenance cycle information to the user terminal (300) in the form of a push notification or dashboard, and upon receiving a reservation request message in response thereto, extracts a list of A / S centers and provides it to the user terminal (300).

[0132] In one embodiment, the service center reservation broker (440) extracts a list of service centers based on the availability of necessary parts for parts with a high probability of failure, the distance between the user's location and the service center, the availability of reservations at the service center, and the repair shop rating, and provides this list to the user terminal (300).

[0133] In the above embodiment, when a specific A / S center and a visit time are selected from the user terminal (300), the A / S center reservation intermediary (440) generates a reservation request message and provides it to the corresponding A / S center terminal (500).

[0135] FIG. 4 is a flowchart illustrating an embodiment of a vehicle maintenance method using OBD data according to the present invention.

[0136] Referring to FIG. 4, the user terminal (300) collects vehicle internal data from an ODB device that collects vehicle internal data by monitoring the status of each of a plurality of parts of the vehicle (step S410).

[0137] The vehicle maintenance brokerage server (400) collects data for vehicle maintenance from the user terminal, the service center terminal, and the external server, analyzes it, and predicts the possibility of failure (step S420).

[0138] The vehicle maintenance brokerage server (400) recommends an after-sales service center based on the above prediction results and then reserves a specific after-sales service center (step S430).

[0139] The A / S center entry / exit management device (600) recognizes the vehicle number of a customer entering the A / S center, and if the customer is a reservation customer, receives ODB data from the user terminal (step S440).

[0140] The A / S center terminal (500) receives vehicle entry information and ODB data from the A / S center entry / exit management device, and based on the ODB data, when a consultation between a mechanic and a customer is completed and a maintenance type is entered, it generates and provides a work order (step S450).

[0142] Although the present invention has been described by the embodiments and drawings described above, the present invention is not limited to the above embodiments, and various modifications and variations are possible from this description by those skilled in the art to which the present invention pertains. Accordingly, the concept of the present invention should be understood only by the claims set forth below, and all equivalent or analogous variations thereof shall be considered to fall within the scope of the concept of the present invention. Explanation of the symbols

[0144] 100: Vehicle, 200: OBD (On Board Dignosis) device, 300: User terminal, 400: Vehicle maintenance brokerage server, 410: Data Collection Department, 420a: Failure learning model generation unit, 420b: Failure learning model, 430: Failure time prediction unit, 440: Service Center Reservation Brokerage, 500: Service Center Terminal, 600: Service Center Entry / Exit Control Device

Claims

Claim 1 A user terminal that collects and provides vehicle internal data from an ODB device that collects vehicle internal data by monitoring the status of each of multiple vehicle parts; a vehicle maintenance brokerage server that defines input variables that may affect failure prediction according to the characteristics of the vehicle parts and the driving environment for each vehicle part received from the user terminal, the A / S center terminal, and an external server, and trains a model to output a failure probability by generating a linear output such as [Equation 1] by applying different weights and biases according to the influence of the input variables on the failure probability, inputs vehicle internal data received from the user terminal, vehicle maintenance data received from the A / S center terminal, and external environment data received from the external server into the failure probability model to output a failure probability, predicts the time when the failure probability reaches a threshold value and determines that time as the maintenance time, collects and analyzes data for vehicle maintenance to predict the probability of failure, and recommends a specific A / S center based on the prediction result; and an A / S center entry / exit management device that recognizes the vehicle number of a customer entering the A / S center and receives and provides ODB data from the user terminal. and includes an A / S center terminal that receives vehicle entry information and ODB data from the A / S center entry / exit management device, and generates and provides a work order when a maintenance type is entered after a consultation between a mechanic and a customer is completed based on the ODB data, [Mathematical Formula 1] A vehicle maintenance system using OBD data, characterized in that z represents a linear output, x1, x2, and x3 represent input variables that can influence failure prediction for each vehicle part, w1, w2, and w3 represent weights for the input variables, and b represents a bias that adjusts the output so that the failure learning model does not over-predict specific conditions. Claim 2 delete Claim 3 A vehicle maintenance system using OBD data according to claim 1, wherein the vehicle maintenance brokerage server inputs data for vehicle maintenance into the fault learning model to receive a current fault probability, and predicts the time when the fault probability reaches a threshold starting from the current fault probability and provides it to a user terminal. Claim 4 A vehicle maintenance system using OBD data according to paragraph 3, wherein the vehicle maintenance brokerage server provides information regarding the maintenance cycle to a user terminal in the form of a push notification or dashboard, and upon receiving a reservation request message in response thereto, extracts a list of service centers and provides it to the user terminal. Claim 5 A step in which a user terminal collects vehicle internal data from an ODB device that collects vehicle internal data by monitoring the status of each of multiple parts of the vehicle; a step in which a vehicle maintenance brokerage server defines input variables affecting the failure of each vehicle part, and trains a model to output a failure probability by generating a linear output such as [Equation 1] by applying different weights and biases according to the influence of said input variables on the failure probability; a step in which the vehicle maintenance brokerage server inputs vehicle internal data received from the user terminal, vehicle maintenance data received from the A / S center terminal, and external environment data received from an external server into the failure probability model to output a failure probability; a step in which the time when said failure probability reaches a threshold is predicted, the time is determined as the maintenance time, and a specific A / S center is booked; a step in which an A / S center entry / exit management device recognizes the vehicle number of a customer entering the A / S center, and if the customer is a reserved customer, receives ODB data from the user terminal; a step in which an A / S center terminal receives vehicle entry information and ODB data from the A / S center entry / exit management device, and based on said ODB data, a consultation between a mechanic and a customer Includes a step of generating and providing a work order when the maintenance type is entered after completion, [Equation 1] A vehicle maintenance method using OBD data, characterized in that z represents a linear output, x1, x2, and x3 represent input variables that can influence failure prediction for each vehicle part, w1, w2, and w3 represent weights for the input variables, and b represents a bias that adjusts the output so that the failure learning model does not over-predict specific conditions.

Images