Vehicle fault detection system based on Internet of Vehicles
By combining passive and active fault detection methods on the Internet of Vehicles platform and using back-end servers for data evaluation and arbitration, the problems of insufficient fault detection and insufficient robustness in the existing vehicle fault detection methods are solved, and efficient detection and intelligent early warning of vehicle faults are achieved.
Patent Information
- Application Number
- CN202411651080.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-19
- Publication Date
- 2025-02-25
AI Technical Summary
The existing vehicle fault detection methods have problems such as insufficient fault detection, insufficient robustness, and neglecting the vehicle's own fault detection results.
The vehicle fault detection system based on the Internet of Vehicles is adopted, combined with the passive fault signal acquisition unit and the active fault information acquisition unit, and data evaluation and arbitration are carried out through the back-end server to realize fault detection and intelligent early warning.
By integrating passive and active fault detection methods, the accuracy and robustness of vehicle fault detection are improved, and effective handling of potential vehicle faults and intelligent early warning are achieved.
Smart Images
Figure CN119512046A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vehicle fault detection using a back-end server, and in particular to a vehicle fault detection system based on an Internet of Vehicles. Background Art
[0002] Intelligent vehicle operation and maintenance has become an important direction for future development, among which fault detection and early warning during vehicle driving are of great significance to vehicle quality assurance. Existing methods usually rely on the vehicle's own diagnosis for fault detection, which cannot effectively deal with potential fault hazards. At the same time, excessive reliance on self-diagnosis will result in the inability to properly warn of faults when self-diagnosis fails.
[0003] The existing methods for automobile fault detection include: obtaining information of the module to be detected through a detection device and transmitting the detection information to an external detection device through a signal generator, and the external device determines whether it is working abnormally; or, judging vehicle fault information through brake light switch, position information, brake pedal position information, accelerator pedal position information, master cylinder pressure information, etc.; or, based on the improved SF method or isolation tree model, the real-time log data of the Internet of Vehicles platform is divided into blocks and feature extracted to obtain multiple data blocks and super features corresponding to each data block; based on the super features corresponding to each data block in the real-time log data and the predetermined features, the abnormal data is located. However, there are certain technical defects in the above, including: insufficient detection of potential faults and insufficient robustness; inability to effectively handle potential vehicle faults; fault detection based on the Internet of Vehicles platform fault detection method and the use of big data in the system, but ignoring the vehicle's own fault detection results, and insufficient interpretability. Summary of the invention
[0004] In order to solve at least one aspect of the above problems, the present invention provides a vehicle fault detection system based on the Internet of Vehicles, comprising: a passive fault signal acquisition unit, the passive fault signal acquisition unit is used to collect the vehicle's own diagnostic signal; an active fault information acquisition unit, the active fault information acquisition unit is used to collect vehicle operation information, the vehicle operation information includes speed, acceleration, position, torque; a back-end server, the back-end server receives the vehicle's own diagnostic signal by connecting to the passive fault signal acquisition unit, the back-end server evaluates the effective state of the passive fault signal acquisition unit, passively detects fault information and passively detects abnormal information through the vehicle's own diagnostic signal, and the back-end server collects the active fault information through the active fault information acquisition unit. The unit receives the vehicle operation information, the back-end server includes a preset fault rate assessment model, the preset fault rate assessment model outputs active fault detection information based on the vehicle operation information, the back-end server also includes an arbitration unit, when the passive detection fault information includes a fault signal, the arbitration unit outputs a warning instruction, when the active fault detection information includes a fault signal and the passive detection fault information is abnormal, the arbitration unit outputs a warning instruction, when the active fault detection information includes a fault signal and the passive detection fault fails, the arbitration unit outputs a warning instruction; a vehicle-side fault warning unit, the vehicle-side fault warning unit receives the warning instruction by connecting to the arbitration unit, and outputs warning information based on the warning instruction.
[0005] Preferably, the preset failure rate assessment model is obtained by training based on a vehicle fault training set using a machine learning algorithm, and the vehicle fault training set includes multiple vehicle fault categories and their corresponding vehicle operation information.
[0006] Preferably, the preset failure rate assessment model includes a feature value extraction layer, a fused feature value layer and a probability output layer, the feature value extraction layer includes multiple feature value models, multiple feature value models adopt an LSTM model, and multiple feature value models respectively extract single feature values based on vehicle operation information. The fused feature value layer adopts an LSTM model to extract fault feature values based on multiple single feature values, and the probability output layer adopts a Softmax activation function to output fault categories based on fault feature values.
[0007] Preferably, the LSTM model training uses cross entropy as a loss function, and the loss function L is: Among them, m is the number of samples, y i is the label of sample i, and p i The probability that sample i is detected as a fault.
[0008] Preferably, the back-end server further includes a data conversion unit, which performs Frenet coordinate system conversion based on the received vehicle operation information.
[0009] Preferably, the back-end server also includes an effective state evaluation unit, which is used to evaluate the effective state of the passive fault signal acquisition unit according to the vehicle's own diagnostic signal. The effective state evaluation unit includes a diagnostic threshold. When the vehicle's own diagnostic signal belongs to the interval corresponding to the diagnostic threshold, the effective state evaluation unit outputs that the passive fault detection unit is valid. When the vehicle's own diagnostic signal does not belong to the interval corresponding to the diagnostic threshold, the effective state evaluation unit outputs that the passive fault detection unit is invalid.
[0010] Preferably, the backend server is heartbeat connected to the passive fault signal collection unit.
[0011] Preferably, the backend server is heartbeat connected to the active fault information collection unit.
[0012] The vehicle fault detection system based on the Internet of Vehicles of the embodiment of the present invention has the following beneficial effects: vehicle fault detection and intelligent early warning are realized by the Internet of Vehicles using a method that integrates passive fault detection and active fault detection. The passive fault signal acquisition unit realizes the acquisition of the vehicle's own diagnostic signal and reports the operation status of its own module, and the active fault information acquisition unit collects information such as vehicle speed, vehicle acceleration, vehicle position information, vehicle torque, etc. and reports the operation status of its own module. Active fault detection realizes the fault detection function by monitoring the operation status of the whole vehicle. The back-end server collects passive fault signals and active fault information, completes active fault detection and completes the arbitration function. The arbitration unit assumes the arbitration function and runs on the back-end server, which can reduce the computing power load of the on-board processor; performs active fault detection and realizes early warning of faults; adopts the arbitration unit to realize active fault detection under the rules, prevents false alarms caused by active fault detection, and improves the availability of the early warning function; adopts the vehicle network to store data on the back-end server, which is convenient for the use of the vehicle intelligent operation and maintenance system. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] In order to better understand the above and other purposes, features, advantages and functions of the present invention, reference may be made to the embodiments shown in the accompanying drawings. The same reference numerals in the accompanying drawings refer to the same components. It should be understood by those skilled in the art that the accompanying drawings are intended to schematically illustrate the preferred embodiments of the present invention and have no limiting effect on the scope of the present invention, and the components in the drawings are not drawn to scale.
[0014] Figure 1 A structural block diagram of a vehicle fault detection system based on vehicle networking according to an embodiment of the present invention is shown;
[0015] Figure 2 A structural block diagram of a preset fault rate evaluation model of a vehicle fault detection system based on the Internet of Vehicles according to an embodiment of the present invention is shown;
[0016] Figure 3 A schematic diagram of the application process of a vehicle fault detection system based on the Internet of Vehicles according to an embodiment of the present invention is shown. DETAILED DESCRIPTION
[0017] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.
[0018] As used herein, the term "including" and its variations mean open inclusion, i.e., "including but not limited to". Unless otherwise stated, the term "or" means "and / or". The term "based on" means "based at least in part on". The terms "an example embodiment" and "an embodiment" mean "at least one example embodiment". The term "another embodiment" means "at least one additional embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.
[0019] In order to at least partially solve the above problems and one or more of other potential problems, the embodiments of the present disclosure propose a vehicle fault detection system based on the Internet of Vehicles, including: a passive fault signal acquisition unit, the passive fault signal acquisition unit is used to collect the vehicle's own diagnostic signal; an active fault information acquisition unit, the active fault information acquisition unit is used to collect vehicle operation information, the vehicle operation information includes speed, acceleration, position, torque; a back-end server, the back-end server receives the vehicle's own diagnostic signal by connecting to the passive fault signal acquisition unit, the back-end server evaluates the effective state of the passive fault signal acquisition unit, passive detection fault information and passive detection abnormal information through the vehicle's own diagnostic signal, and then The end server receives vehicle operation information through an active fault information collection unit, and the back-end server includes a preset fault rate assessment model, which outputs active fault detection information based on the vehicle operation information. The back-end server also includes an arbitration unit. When the passive detection fault information includes a dynamic fault signal, the arbitration unit outputs a warning instruction. When the active fault detection information includes a fault signal and the passive detection fault information is abnormal, the arbitration unit outputs a warning instruction. When the active fault detection information includes a fault signal and the passive detection fault information fails, the arbitration unit outputs a warning instruction; a vehicle-end fault warning unit, the vehicle-end fault warning unit receives the warning instruction by connecting to the arbitration unit, and outputs warning information based on the warning instruction.
[0020] Specifically, Figure 1 As shown, the passive fault collection unit is an on-board self-diagnosis device, which is used to collect the vehicle's own diagnostic signals. The vehicle's own diagnostic signals include various on-board sensor fault or abnormality information such as BMS fault and / or abnormality information, ESP fault and / or abnormality information, etc.
[0021] The active fault information collection unit obtains vehicle operation information, including vehicle speed, vehicle acceleration, vehicle position information and vehicle torque, by connecting to the on-board controller.
[0022] The back-end server is connected to the vehicle in communication, and the back-end server may be a cloud server. The back-end server is connected to the passive fault signal collection unit and the active fault information collection unit through the vehicle-mounted communication unit to receive the vehicle's own diagnostic signals and vehicle operation information.
[0023] The back-end server includes an evaluation of the effective state of the passive fault signal acquisition unit through the vehicle's own diagnostic signal. In some embodiments, the back-end server also includes an effective state evaluation unit, which is used to evaluate the effective state of the passive fault signal acquisition unit according to the vehicle's own diagnostic signal. The effective state evaluation unit includes a diagnostic threshold. When the vehicle's own diagnostic signal belongs to the interval corresponding to the diagnostic threshold, the effective state evaluation unit outputs that the passive fault detection unit is effective. When the vehicle's own diagnostic signal does not belong to the interval corresponding to the diagnostic threshold, the effective state evaluation unit outputs that the passive fault detection unit is invalid. Figure 3 As shown, the active state evaluation unit is activated in response to the active fault detection information including a fault signal.
[0024] When the vehicle's own diagnostic signal exceeds the preset normal range, the self-diagnosis system will immediately mark it as a potential problem. For example, the output voltage of the water temperature sensor should be between 0.30-4.70V (i.e., the diagnostic threshold). If the output voltage is less than 0.15V or greater than 4.85V, the system will think that its signal system may have a short circuit or open circuit fault, and the effective state evaluation unit outputs that the passive fault detection unit is invalid. In some embodiments, the diagnostic threshold is a time threshold, that is, when the input signal does not change as expected within the preset time threshold, the effective state evaluation unit evaluates that the passive fault detection unit is invalid. For example, after the engine reaches the normal operating temperature of the oxygen sensor, if the computer cannot detect that the output signal of the oxygen sensor exceeds a certain time or fluctuates within the range of 0.45V, the system will determine that there is an abnormality in the oxygen sensor signal system. In other embodiments, the effective state evaluation unit also includes an abnormal matching parameter. When the actuator instruction corresponding to the vehicle's own diagnostic signal is the same as the abnormal matching parameter fed back by the sensor, the system will confirm that the sensor is abnormal. For example, when the EGR system's EGR valve height sensor does not change according to preset parameters, the computer will detect other signals, such as the intake pressure sensor (MAP) output, to determine whether the EGR valve is working properly. The vehicle's own diagnostic signal determines the abnormal state of the EGR system based on the received intake pressure sensor signal.
[0025] The vehicle-side fault warning unit includes any one or more of a vehicle-mounted display, a vehicle-mounted player or a vehicle-mounted signal light, etc. The vehicle-side fault warning unit is connected to the back-end server through the vehicle-mounted controller to receive fault warning instructions, and outputs corresponding prompt audio or text and other prompt information based on the fault warning instructions.
[0026] like Figure 3As shown, during the operation of the vehicle, the passive fault signal collection unit and the active fault information collection unit are output to the back-end server in real time through the vehicle-side communication equipment. When the passive fault signal collection unit received by the back-end server contains a fault signal, the arbitration unit outputs a warning instruction; when the back-end server receives the vehicle operation information through the active fault information collection unit, and inputs the vehicle operation information into the preset fault rate evaluation model, when the active fault detection information output by the preset fault evaluation model includes fault information, the effective evaluation unit evaluates whether the passive fault collection unit is in a valid state according to the vehicle's own diagnostic signal, when the active fault detection information includes a fault signal and the passive detection fault information is abnormal, the arbitration unit outputs a warning instruction, and when the active fault detection information includes a fault signal and the passive detection fault fails, the arbitration unit outputs a warning instruction.
[0027] In some embodiments, the preset failure rate assessment model is obtained by training based on a vehicle fault training set using a machine learning algorithm, and the vehicle fault training set includes multiple vehicle fault categories and their corresponding vehicle operation information.
[0028] Specifically, a vehicle operation information set is collected, and corresponding fault tags are added according to the fault categories corresponding to each group of vehicle operation information in the vehicle operation information set to generate a vehicle fault training set. A classification machine learning algorithm is used to train the vehicle fault training set to obtain a preset fault rate assessment model. The classification machine learning algorithm includes one or a combination of convolutional neural networks (CNNs), recurrent neural networks (RNNs), etc.
[0029] In some embodiments, the preset failure rate assessment model includes a feature value extraction layer, a fused feature value layer and a probability output layer. The feature value extraction layer includes multiple feature value models. The multiple feature value models adopt an LSTM model. The multiple feature value models extract single feature values based on vehicle operation information respectively. The fused feature value layer adopts an LSTM model to extract fault feature values based on multiple single feature values. The probability output layer adopts a Softmax activation function to output fault categories based on the fault feature values.
[0030] Specifically, the active fault information collection unit collects information such as vehicle speed, vehicle acceleration, vehicle position information, vehicle torque, etc., and uploads it to the back-end server. The back-end server first cleans the received vehicle operation data, removes obvious outliers, and normalizes the data. Figure 2 As shown, the vehicle position change, vehicle speed, acceleration, and torque information are then input into a multi-channel long short-term memory network (feature value extraction layer LSTM) to extract feature values, the feature values are concatenated and input into the fusion feature value layer LSTM, and finally the classification result is output through softmax. In some embodiments, the LSTM model training uses cross entropy as the loss function, and the loss function L is: Among them, m is the number of samples, y i is the label of sample i, and p i The probability that sample i is detected as a fault.
[0031] In some embodiments, the back-end server further includes a data conversion unit, which performs Frenet coordinate system conversion based on the received vehicle operation information.
[0032] Specifically, the vehicle position information is converted into the Frenet coordinate system, and the vehicle position at time t can be expressed as x t (s t , d t ), s t is the coordinate along the center line at time t, d t is the vertical centerline coordinate. By transforming the coordinate system of the vehicle position information, when the vehicle moves forward and stays in the lane, the driving trajectory will be calculated as a straight line, which greatly simplifies the difficulty of trajectory planning, reduces the amount of calculation, and improves efficiency.
[0033] In some embodiments, the backend server is connected to the passive fault signal collection unit by heartbeat. The backend server is connected to the active fault information collection unit by heartbeat.
[0034] Specifically, the passive fault signal collection unit and the active fault information collection unit send a detection packet to the back-end server at intervals; and a timeout timer is started when the detection packet is sent; when the back-end server receives the detection packet, it should respond with a packet; if the passive fault signal collection unit and the active fault information collection unit receive a response packet from the back-end server, it means that the back-end server is normal and the timeout timer is deleted; if the timeout timers of the passive fault signal collection unit and the active fault information collection unit expire and no response packet is received, it means that the connection with the back-end server is disconnected.
[0035] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the present invention.
Claims
1. A vehicle fault detection system based on Internet of Vehicles, characterized in that: include: A passive fault signal acquisition unit, wherein the passive fault signal acquisition unit is used to acquire the vehicle's own diagnostic signal; An active fault information collection unit, the active fault information collection unit is used to collect vehicle operation information, the vehicle operation information includes speed, acceleration, position, and torque; A back-end server, wherein the back-end server receives the vehicle self-diagnosis signal by connecting to the passive fault signal acquisition unit, the back-end server evaluates the effective state of the passive fault signal acquisition unit, the passive detection fault information and the passive detection abnormality information through the vehicle self-diagnosis signal, the back-end server receives the vehicle operation information through the active fault information acquisition unit, the back-end server includes a preset fault rate evaluation model, the preset fault rate evaluation model outputs active fault detection information based on the vehicle operation information, the back-end server also includes an arbitration unit, when the passive detection fault information includes a fault signal, the arbitration unit outputs a warning instruction, when the active fault detection information includes a fault signal and the passive detection fault information is abnormal, the arbitration unit outputs a warning instruction, when the active fault detection information includes a fault signal and the passive detection fault information fails, the arbitration unit outputs a warning instruction; A vehicle-side fault warning unit receives a warning instruction by connecting to the arbitration unit, and outputs warning information based on the warning instruction.
2. The system according to claim 1, characterized in that The preset failure rate assessment model is obtained by training based on a vehicle failure training set using a machine learning algorithm, and the vehicle failure training set includes multiple vehicle failure categories and their corresponding vehicle operation information.
3. The system according to claim 1, characterized in that The preset failure rate assessment model includes a feature value extraction layer, a fused feature value layer and a probability output layer. The feature value extraction layer includes multiple feature value models, and the multiple feature value models adopt an LSTM model. The multiple feature value models extract single feature values based on vehicle operation information respectively. The fused feature value layer adopts an LSTM model to extract fault feature values based on multiple single feature values. The probability output layer adopts a Softmax activation function to output fault categories based on fault feature values.
4. The system according to claim 3, characterized in that The LSTM model training uses cross entropy as the loss function, and the loss function L is: Among them, m is the number of samples, y i is the label of sample i, and p i The probability that sample i is detected as a fault.
5. The system according to claim 4, characterized in that The back-end server also includes a data conversion unit, which performs Frenet coordinate system conversion based on the received vehicle operation information.
6. The system according to claim 1, characterized in that The back-end server also includes an effective state evaluation unit, which is used to evaluate the effective state of the passive fault signal acquisition unit according to the vehicle's own diagnostic signal. The effective state evaluation unit includes a diagnostic threshold. When the vehicle's own diagnostic signal belongs to the interval corresponding to the diagnostic threshold, the effective state evaluation unit outputs that the passive fault detection unit is effective. When the vehicle's own diagnostic signal does not belong to the interval corresponding to the diagnostic threshold, the effective state evaluation unit outputs that the passive fault detection unit is invalid.
7. The system according to claim 1, characterized in that The backend server is connected to the passive fault signal collection unit by heartbeat.
8. The system according to claim 1, characterized in that The backend server is connected to the active fault information collection unit by heartbeat.