A vehicle precise fault diagnosis method and device

Abstract

The present invention discloses a method and device for accurate vehicle fault diagnosis. The method includes: obtaining the remaining power of the vehicle battery; determining whether the remaining power is less than a first preset value; the in-vehicle central control obtaining the actual unlocking times of the vehicle within a preset time period; determining whether the difference between the number of received unlocking commands and the actual unlocking times within the preset time period is greater than a second preset value; obtaining vehicle sample data and determining a vehicle profile, and inputting the vehicle profile into a trained model to output a fault type. The present invention performs functional detection on various parts of the vehicle based on multiple sensors of the vehicle itself, and cooperates with other basic data and usage data, and inputs them into a pre-trained fault detection model for fault detection to confirm whether the vehicle is a faulty vehicle. In addition, the present invention uses a Bluetooth module to assist the communication module for information transmission, which can effectively and timely and accurately reflect fault information and improve the user experience.

Description

Technical Field

[0001] The present invention belongs to the technical field of power-assisted vehicles, and in particular, relates to a method and device for accurate vehicle fault diagnosis. Background Art

[0002] With the continuous development of the shared transportation industry such as shared bicycles and power-assisted vehicles, shared bicycles and power-assisted vehicles are spread all over the country. In solving the problem of short-distance travel, travelers are increasingly relying on shared transportation modes, and the supply of shared bicycles / power-assisted vehicles is also increasing. However, when the number of vehicles reaches a certain level, with the increase in the number of vehicle uses and time, the failure rate begins to rise. Vehicle failures directly affect the riding experience of users, and there are even potential safety hazards. At the same time, it also increases the workload of vehicle operation and maintenance personnel. In particular, shared power-assisted vehicles are equipped with power components such as storage batteries and auxiliary motors, increasing the types of failures.

[0003] Existing shared transportation tools usually report faults passively after a fault occurs, or operation and maintenance personnel conduct on-site inspections to determine whether a vehicle is faulty. Vehicle active fault reporting can usually only report some systematic problems. For example, the vehicle cannot be unlocked, the vehicle battery level is low, etc. Moreover, many reported faults are suspected faults rather than real faults, resulting in a relatively low accuracy rate of fault reporting. On the other hand, on-site inspection of vehicle faults by operation and maintenance personnel requires covering a large number of vehicles, consuming a large amount of labor costs, and it is difficult to ensure the timeliness of fault detection. In addition, when the vehicle positioning module and / or communication module does not work, the vehicle cannot actively report fault information, resulting in insufficient available vehicles at the station, affecting users' borrowing and travel.

[0004] In view of the above defects, the technical problem to be solved by the present invention is that the vehicle fault reporting rate in the prior art is relatively low, or the fault feedback is not timely, inaccurate, or even unable to be reported. Summary of the Invention

[0005] In order to overcome the above technical defects, the present invention provides a method and device for accurate vehicle fault diagnosis to solve the problems involved in the background art.

[0006] The present invention provides a method and device for accurate vehicle fault diagnosis, including:

[0007] The in-vehicle central control obtains the remaining power of the vehicle battery; determines whether the remaining power is less than a first preset value; if so, outputs the fault type as a low-battery vehicle and notifies the operation and maintenance personnel; if not, proceeds to the next step;

[0008] The vehicle-mounted central control obtains the actual number of unlocking times of the vehicle within a preset time period; determines whether the difference between the number of received unlocking commands and the actual number of unlocking times within the preset time period is greater than a second preset value; if so, outputs a fault type of unlocking timeout fault vehicle and notifies the operation and maintenance personnel; if not, proceeds to the next step;

[0009] The vehicle-mounted central control obtains vehicle sample data and transmits it to the background server; the sample data includes multiple-dimensional information related to the vehicle riding state and a fault label for marking a fault vehicle;

[0010] The background server determines the vehicle portrait of the vehicle based on the sample data, and determines whether the assist ratio of the vehicle exceeds a predetermined range within a predetermined time period, where the assist ratio = motor output torque / pedal torque;

[0011] And inputs the vehicle portrait into a training model to output a fault type; the training model is a fault recognition model obtained by weighting the vehicle portrait processed by a machine learning model and its corresponding fault label, the training model includes a first training model and a second training model, the first training model is trained based on the vehicle portrait and a first fault type; the second training model is trained based on the vehicle portrait and a second fault type; if it is higher than the maximum value of the predetermined range, the vehicle portrait is input into the first training model, and if it is lower than the minimum value of the predetermined range, the vehicle portrait is input into the second training model.

[0012] Preferably or optionally, the dimensional information includes: detection data of the operating states of various vehicle parts.

[0013] Preferably or optionally, the detection data of the operating states of various vehicle parts includes: vehicle riding speed, pedal output torque, motor output torque, battery output voltage, battery temperature, braking pressure, braking distance.

[0014] Preferably or optionally, the dimensional information includes: detection data of the operating states of various vehicle parts, vehicle basic attribute data, historical fault and maintenance data, vehicle historical riding data.

[0015] Preferably or optionally, the detection data of the operating states of various vehicle parts includes: vehicle riding speed, pedal output torque, motor output torque, battery output voltage, battery temperature, braking pressure, braking distance;

[0016] The vehicle basic attribute data includes: the production batch of the vehicle, the operating time of the vehicle, the operating area of the vehicle, the order situation and the order cancellation volume within a preset time;

[0017] The vehicle historical fault and maintenance data includes: the historical fault situation and maintenance situation of the vehicle within a preset time period;

[0018] The vehicle historical riding data includes: historical average driving duration, historical average driving speed, historical average driving distance, and historical operation records.

[0019] Preferably or optionally, before inputting the vehicle portrait into the training model, the diagnosis method further includes the following steps:

[0020] Judge whether the assist ratio of the vehicle exceeds a predetermined range multiple times within a predetermined time period; the assist ratio = motor output torque / pedal torque;

[0021] If it is higher than the maximum value of the predetermined range, input the vehicle portrait into the first training model; if it is lower than the minimum value of the predetermined range, input the vehicle portrait into the second training model; the first training model is trained based on the vehicle portrait and the first fault type; the second training model is trained based on the vehicle portrait and the second fault type.

[0022] Preferably or optionally, the first fault type includes: power supply fault, motor fault, controller fault, control device fault;

[0023] The second fault type includes: foot pedal fault, transmission device fault, braking device fault, wheel and accessory fault, handle fault, vehicle body shaking or abnormal noise.

[0024] Preferably or optionally, the method for the in-vehicle central control to transmit sample data to the background server includes:

[0025] Judge whether the communication module is connected to the background server; if so, transmit the sample data to the background server through the communication module; if not, execute the next step;

[0026] Broadcast a request for assistance information through the Bluetooth module until the Bluetooth module on the neighboring vehicle or parking device responds and communicates with the Bluetooth transmitter of the vehicle to be detected;

[0027] Transmit the sample data to the Bluetooth module on the neighboring vehicle or parking device through the Bluetooth module;

[0028] Transmit the sample data of the vehicle to be detected to the background server through the communication module on the neighboring vehicle or parking device.

[0029] In a third aspect, the present invention also provides a system based on the vehicle precise fault diagnosis method, including:

[0030] A first acquisition unit configured to acquire the remaining power of the vehicle battery;

[0031] The first judgment unit is configured to judge whether the remaining power is less than a first preset value; if so, it outputs that the fault type is a vehicle with low power and notifies the operation and maintenance personnel; if not, it proceeds to the next step;

[0032] The second acquisition unit is configured to the in-vehicle central control to acquire the actual unlocking times of the vehicle within a preset time period;

[0033] The second judgment unit is configured to judge whether the difference between the number of received unlocking commands and the actual unlocking times within the preset time period is greater than a second preset value; if so, it outputs that the fault type is a vehicle with an unlocking timeout fault and notifies the operation and maintenance personnel; if not, it proceeds to the next step;

[0034] The third acquisition unit is configured to acquire sample data of the vehicle and transmit it to the background server; the sample data includes multiple dimension information related to the riding state of the vehicle and a fault label for marking a faulty vehicle;

[0035] The first processing unit is configured to determine the vehicle portrait of the vehicle based on the sample data and input the vehicle portrait into a training model to output a fault type; the training model is a fault recognition model obtained by weighting the vehicle portrait processed by a machine learning model and its corresponding fault label.

[0036] In a third aspect, the present invention also provides a server for accurate vehicle fault diagnosis, which is characterized by including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the method when executing the program.

[0037] The present invention relates to a method and device for accurate vehicle fault diagnosis, and compared with the prior art, has the following beneficial effects:

[0038] 1. The present invention performs functional detection on various parts of the vehicle based on multiple sensors of the vehicle itself, and cooperates with other basic data and usage data, and inputs them into a pre-trained fault detection model for fault detection to confirm whether the vehicle is a faulty vehicle, and can even effectively judge which vehicles have a tendency to have faults, avoiding the problems of low vehicle fault reporting rate, untimely or inaccurate fault feedback in the prior art.

[0039] 2. All kinds of data obtained in the present invention can directly use the previous detection functions of the equipment without adding other sensors and detection devices, reducing the operation cost of the entire diagnosis method.

[0040] 3. The present invention designs a suitable pre-classification method according to the types of defects, improves the detection efficiency while ensuring the detection accuracy, and can solve the detection of more than 95% of faulty vehicles according to experiments.

[0041] 4. The present invention uses the method of temporarily storing data in a memory to record sample data, and then conducts data exchange regularly, which ensures the reliability of data transmission, avoids occupying a large amount of bandwidth, and also reduces the operation cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a schematic flowchart of a vehicle precise fault diagnosis method in Embodiment 1 of the present invention.

[0043] Figure 2 It is a vehicle precise fault diagnosis device in Embodiment 2 of the present invention.

[0044] Figure 3 It is a schematic structural diagram of an exemplary electronic device in Embodiment 3 of the present invention.

[0045] Description of reference numerals: First acquisition unit 11, First judgment unit 12, Second acquisition unit 13, Second judgment unit 14, Third acquisition unit 15, First processing unit 16, Bus 300, Receiver 301, Processor 302, Transmitter 303, Memory 304, Bus interface 305. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] In the following description, a large number of specific details are given to provide a more thorough understanding of the present invention. However, it is obvious to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, some technical features well known to the art are not described to avoid confusion with the present invention.

[0047] Embodiment 1

[0048] Refer to the attached Figure 1 , a vehicle precise fault diagnosis method, including:

[0049] S100. The in-vehicle central control obtains the remaining power of the vehicle battery; determines whether the remaining power is less than a first preset value; if so, outputs the fault type as low-power vehicle and notifies the operation and maintenance personnel; if not, proceeds to the next step;

[0050] Specifically, the vehicle mainly refers to a shared power-assisted vehicle, and the remaining battery power can be obtained through a battery sensor provided on the shared power-assisted vehicle. Since the battery sensor is connected to the central control panel of the shared power-assisted vehicle, during implementation, the remaining battery power information can be directly obtained through the central control panel. The first preset power is the lowest value of the remaining battery power set to ensure battery safety, and the first preset power is determined by the characteristics of the rechargeable battery on the shared power-assisted vehicle, and is generally determined according to the relevant parameters of the rechargeable battery.

[0051] Of course, it can be understood that the battery can be a rechargeable battery, but is not limited to a rechargeable battery. For those skilled in the art, the battery can also be a fuel cell. By detecting the remaining hydrogen content in the hydrogen storage device, the remaining ionization of the fuel cell can be confirmed. For the sake of convenience of description, the rechargeable battery will still be used as an example in the following text for introduction and explanation.

[0052] S200. The vehicle-mounted central control obtains the actual unlocking times of the vehicle within a preset time period; determines whether the difference between the number of received unlocking commands and the actual unlocking times within the preset time period is greater than a second preset value; if so, outputs the fault type as an unlocking timeout fault vehicle and notifies the operation and maintenance personnel; if not, proceeds to the next step;

[0053] Specifically, the user unlocks the vehicle by scanning the QR code on the vehicle body. The system issues an unlocking instruction, and the lock of the shared power-assisted vehicle executes the unlocking command and gives feedback to the system after unlocking is completed, which is a complete unlocking action. If the system does not receive feedback information within a certain time, it will execute the unlocking action again. If the lock feedback is not received within the predetermined time, and if the number of consecutive unlocking timeouts of the vehicle exceeds the preset threshold within the preset time, the threshold is generally set to be greater than or equal to 3 times continuously, and it can be considered that the unlocking fails, and the vehicle can be marked as a vehicle with a lock failure.

[0054] S300. The vehicle-mounted central control obtains sample data of the vehicle and transmits it to the background server; the sample data includes multiple dimension information related to the riding state of the vehicle and a fault label for marking a fault vehicle.

[0055] Specifically, find the pattern in the various characteristic parameter data of each part of the vehicle, and find the relevant data that has a relatively large correlation with the fault vehicle and is easy to judge the fault vehicle. In other words, there will be obvious characteristic differences in the dimension information of each dimension of the relevant latitude between the fault vehicle and the non-fault vehicle. By excavating the differences between these characteristic data, it can be reversely confirmed whether the shared power-assisted vehicle is a fault vehicle, and even effectively judge which shared power-assisted vehicles have a tendency to have faults, make timely maintenance and avoid further deterioration of the fault area, reducing the operating cost.

[0056] In a preferred embodiment, the sample data includes: detection data of the operating states of each part of the vehicle and a fault label for marking a fault vehicle.

[0057] Among them, the detection data of the operating states of various vehicle parts include: vehicle riding speed, pedal output torque, motor output torque, battery output voltage, battery temperature, braking pressure, braking distance and other parameters. The vehicle riding speed is detected by a ranging sensor set on the shared power-assisted vehicle. Additionally, it can also directly rely on the positioning system built into the shared power-assisted vehicle. Although there is a certain deviation, it is within the allowable range, generally about 1 to 2 kilometers per hour, which can be ignored for the entire calculation method. The pedal pressure is measured by the array pedal output torque set on both sides of the pedal, and the motor output torque is measured by the rotational speed sensors set at the pedal and the motor; the battery output voltage and battery temperature are obtained through the voltage detection device and temperature sensor set on the battery; the braking pressure is detected by a braking sensor that detects the distance of the piston movement or the pressure change between the piston and the wheel hub inside the drum brake, and the braking distance is measured by the cooperation of the ranging sensor and the braking sensor. The acquisition of the above various parameters can directly use the sensors and detection devices on the existing shared power-assisted vehicles. During the implementation process, only the relevant integration and summary of the above various data are required, without adding new sensors and detection devices, reducing the operating cost of the entire diagnostic method.

[0058] The fault labels used to mark faulty vehicles include: battery fault, motor fault, controller fault, control transmission line fault, foot pedal fault, transmission device fault, braking device fault, wheel and appendage fault, handle fault, vehicle body shaking or abnormal noise.

[0059] By integrating the detection data of the operating states of various vehicle parts and the fault labels used to mark faulty vehicles, for example, when the battery fails while the riding speed remains basically the same, it generally causes a decrease in the battery output voltage, an increase in the battery temperature, a decrease in the motor output torque, and an increase in the pedal torque; when the transmission device fails, it will lead to an increase in the pedal output torque and an increase in the motor output torque; when the braking device fails, during braking, compared with the standard braking pressure and standard motor output torque, the braking pressure becomes smaller and the motor output torque becomes larger. When different types of faults occur, the detection data of the operating states of one or more vehicle parts will also fluctuate accordingly. By identifying the differences between these characteristic data, it can be reversely confirmed whether the shared power-assisted vehicle is a faulty vehicle, and even effectively determine which shared power-assisted vehicles have a tendency to malfunction.

[0060] In another preferred embodiment, the sample data includes: the detection data of the operating states of various vehicle parts, vehicle basic attribute data, historical fault and maintenance data, vehicle historical riding data, and the fault labels used to mark faulty vehicles.

[0061] The detection data of the operating states of the various parts of the vehicle and the fault labels for marking fault vehicles are the same as those in the above preferred embodiment, and will not be elaborated here.

[0062] The vehicle basic attribute data includes: the production batch of the vehicle, the operating time of the vehicle, the operating area of the vehicle, the order situation and the order cancellation volume in a recent period. The specific similarity of the parts produced in the same batch also shows a certain similarity after being integrated into a whole vehicle. In addition, fault vehicles usually appear in clusters, that is to say, the road conditions in the use area also have a certain impact on the occurrence of faults. Therefore, extracting the operating area of the vehicle can also provide corresponding information for fault identification, and data mining has also proved this point. The higher the order volume in a recent period, the relatively higher the wear on the vehicle, and the probability of failure also increases accordingly. A high order cancellation volume in a recent period indicates that the users are not satisfied with the riding performance of the shared power-assisted vehicle, which indirectly indicates that there is a fault or a tendency to have a fault in the shared power-assisted vehicle. The recent period can be determined by comprehensively considering factors such as weather, operating area, and pedestrian flow. Therefore, extracting the production batch of the vehicle, the operating time of the vehicle, the operating area of the vehicle, the order situation within a predetermined time, and the order cancellation volume can also provide corresponding information for fault identification, and data mining has also proved this point.

[0063] The vehicle historical fault and maintenance data includes: the historical fault situation and maintenance situation of the vehicle within a preset time period; generally speaking, the faults of the vehicle itself have a certain periodicity within a certain range. At the same time, due to the running-in problem between parts, the vehicle is more likely to have a fault again after a fault occurs, especially for some vehicles with high-frequency faults. Therefore, the portrait of the health dimension of the vehicle can be established by using relevant data such as the vehicle's historical faults, repair times, and historical fault locations in advance.

[0064] The vehicle historical riding data includes: the historical average driving duration, the historical average driving speed, the historical average driving distance, and the historical operation records. It can be understood that the usage mode and usage habits of the vehicle are closely related to the possibility of the vehicle having a fault. Therefore, in this embodiment, the possibility of the vehicle having a fault is inferred by recording the historical operation records of the vehicle. The historical operation records can obtain information on the usage conditions of the user when using the vehicle, such as the braking frequency, steering curvature, braking of the front braking device and / or the rear braking device, and whether the pedal is overused, through parameters such as the rotation sensor, braking pressure, and pedal output torque set on the handle. In addition, by expanding the historical average driving duration, the historical average driving speed, the historical average driving distance, etc., and extracting relevant riding information as data features, the features of the vehicle in the riding dimension are supplemented, making the portrait of the vehicle in this dimension richer.

[0065] S400. Obtain the assist ratio of the vehicle within a predetermined time period; and determine whether the vehicle has a fault and the fault type of the faulty vehicle.

[0066] To improve the recognition efficiency and accuracy of the entire training model, pre-classify the fault types before inputting the determined portrait into the training model. Divide according to the area where the fault occurs. The shared power-assisted bicycle includes two parts: an electric power-assisted drive system and a bicycle system. Among them, the common faults of the electric power-assisted drive system correspond to the first fault type, specifically including: power supply fault, motor fault, controller fault, control device fault. The common faults of the bicycle system correspond to the second fault type, specifically including: pedal fault, transmission device fault, braking device fault, wheel and accessories fault, handlebar fault, vehicle body shaking or abnormal noise.

[0067] An electric shared power-assisted bicycle generally estimates the pedaling torque when the user is riding, and then controls the output torque of the assist motor according to this estimated torque to achieve the entire assist process; during normal driving, the motor torque is generally proportional to the pedal torque. The higher the pedaling speed, the greater the output torque provided by the motor. That is to say, under uniform driving conditions, the ratio between the motor output torque and the pedal torque of an electric shared power-assisted bicycle should be a fixed value, that is, the assist ratio is a constant value. Define the assist ratio as = motor output torque / pedal torque. Therefore, by comparing the resistance ratio at different speeds, if the assist ratio is too large, it means that the fault of the shared power-assisted bicycle is probably in the bicycle system, and the weight of the first fault type can be appropriately increased, or the first training model trained based on the first fault type can be used; if the assist ratio is too low, it means that the fault of the power-assisted bicycle is probably in the electric power-assisted drive system, and the weight of the second fault type can be appropriately increased, or the second training model trained based on the second fault type can be used.

[0068] S500. The background server determines the vehicle portrait of the vehicle based on the sample data, and inputs the vehicle portrait into the training model to output the fault type; the training model is a fault recognition model obtained by weighting the vehicle portrait processed by the machine learning model and its corresponding fault labels.

[0069] The training model can be a model including a neural network, which is obtained by training with a large amount of sample data. Specifically, the fault detection model can include a neural network composed of an input layer, a convolution layer, a pooling layer, an output layer, etc., mainly setting the number of convolution kernels according to the input feature type or the feature type to be analyzed, constructing a basic network structure, and then using the training samples in the sample set for training and learning, and using the verification samples to verify the prediction accuracy, and completing the final training process when a certain convergence is reached, thereby obtaining the fault detection model. It should be noted that the network structure in the fault detection model can be adjusted according to the input and output parameters. Based on the known existing neural network structure, the technical personnel in this field can set the basic model structure based on the input and output parameters given in this embodiment. This embodiment mainly combines the detection data of the operating status of each part of the vehicle, the basic attribute data of the vehicle, the historical fault and maintenance data, and the historical riding data of the vehicle as input to analyze and predict whether the vehicle has a fault, and the corresponding specific model structure is not limited.

[0070] In a further embodiment, corresponding to step S400, the training model includes a first training model and a second training model, the first training model is obtained by training based on the vehicle portrait and the first fault type; the second training model is obtained by training based on the vehicle portrait and the second fault type. When obtaining the vehicle portrait, some interference data in the detection data of the operating status of each part of the vehicle can be eliminated according to the fault type to improve the processing rate of the entire training model. When further judging the first fault type, parameters such as braking pressure and braking distance can be eliminated; when further judging the second fault type, parameters such as battery temperature can be eliminated. In this way, a suitable pre-classification method is designed according to the type of defect, while ensuring the detection accuracy, the detection efficiency is improved. According to experiments, the detection of more than 95% of faulty vehicles can be solved.

[0071] It should be noted that since the fault occurrence area (the shared power-assisted vehicle with a fault) and the fault analysis area (the back-end server) are separated, how to conduct effective information exchange has become the focus of the entire fault detection method. When the vehicle positioning module and / or the communication module fails to work, the vehicle cannot actively report fault information, resulting in insufficient available vehicles at the station, which affects users' borrowing and travel. Therefore, the applicant proposes a data exchange method, including the following steps: The vehicle-mounted central control periodically polls the assist ratio of the vehicle; the assist ratio = motor output torque / pedal torque; determine whether the assist ratio of the vehicle exceeds a predetermined range multiple times within a predetermined time period; if so, obtain the detection data of the operating states of various parts of the vehicle within this period in the memory to form sample data; the vehicle-mounted central control of the vehicle to be detected determines whether the vehicle-mounted communication module is connected to the back-end server; if so, transmit the sample data to the back-end server through the communication module, and the communication module mainly refers to a wireless transmission device; if not, it indicates that there may be a fault in the vehicle-mounted wireless module, and proceed to the next step; broadcast a request for assistance information through the Bluetooth module until the Bluetooth module on the neighboring vehicle or parking equipment responds and is connected to the Bluetooth transmitter of the vehicle to be detected; the Bluetooth module can be a Bluetooth transmitter set on the vehicle or a Bluetooth repeater set on the parking post; transmit the sample data to the Bluetooth module on the neighboring vehicle or parking equipment through the Bluetooth module; transmit the sample data of the vehicle to be detected to the back-end server through the communication module on the neighboring vehicle or parking post. This embodiment uses the Bluetooth module to assist the communication module in information transmission, which can effectively and accurately reflect fault information in a timely manner and improve the user experience. Among them, the Bluetooth transmitter is a part of the vehicle-mounted intelligent lock, which will not increase the hardware cost, but can further improve the stability of information transmission and the accuracy of vehicle fault diagnosis.

[0072] Embodiment 2

[0073] Based on the same inventive concept as a vehicle precise fault diagnosis method in the foregoing Embodiment 1, the present invention also provides a vehicle precise fault diagnosis device, as Figure 2 shown, the device includes:

[0074] A first acquisition unit 11, configured to acquire the remaining power of the vehicle battery;

[0075] A first judgment unit 12, configured to judge whether the remaining power is less than a first preset value; if so, output the fault type as a low-power vehicle and notify the operation and maintenance personnel; if not, proceed to the next step;

[0076] A second acquisition unit 13, configured to the vehicle-mounted central control acquire the actual unlocking times of the vehicle within a preset time period;

[0077] The second judgment unit 14 is configured to judge whether the difference between the number of received unlocking commands and the actual unlocking times within a preset time period is greater than a second preset value; if so, output the fault type as the unlocking timeout fault vehicle and notify the operation and maintenance personnel; if not, proceed to the next step;

[0078] The third acquisition unit 15 is configured to acquire sample data of the vehicle and determine a vehicle portrait, where the sample data includes multiple dimensional information related to the riding state of the vehicle and a fault label for marking a fault vehicle;

[0079] The first processing unit 16 is configured to input the vehicle portrait into a training model and output a fault type; the training model is a fault recognition model obtained by weighting the vehicle portrait processed by a machine learning model and its corresponding fault label.

[0080] All the various change methods and specific examples of the vehicle precise fault diagnosis method in the foregoing Embodiment 1 are equally applicable to the vehicle precise fault diagnosis device in this embodiment. Through the foregoing detailed description of the vehicle precise fault diagnosis method, those skilled in the art can clearly know the implementation method of the vehicle precise fault diagnosis device in this embodiment. Therefore, for the sake of brevity of the specification, it will not be elaborated herein.

[0081] Embodiment 3

[0082] Based on the same inventive concept as the vehicle precise fault diagnosis method in the foregoing embodiment, the present invention also provides a server for vehicle precise fault diagnosis, as Figure 3 shown Figure 3 The exemplary electronic device in Embodiment 3 includes a memory 304, a processor 302, and a computer program stored on the memory 304 and executable on the processor 302. When the processor 302 executes the program, it implements the steps of any of the vehicle precise fault diagnosis methods described above.

[0083] Among them, in Figure 3 the bus architecture (represented by the bus 300), the bus 300 may include any number of interconnected buses and bridges. The bus 300 links various circuits including one or more processors represented by the processor 302 and the memory represented by the memory 304 together. The bus 300 may also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art. Therefore, they will not be further described herein. The bus interface 305 provides an interface between the bus 300 and the receiver 301 and the transmitter 303. The receiver 301 and the transmitter 303 may be the same element, that is, a transceiver, which provides a unit for communicating with various other devices on the transmission medium.

[0084] The processor 302 is responsible for managing the bus 300 and general processing, while the memory 304 can be used to store data used by the processor 302 when performing operations.

[0085] In addition, it should be noted that, in the above specific embodiments, the various specific technical features described can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the present invention will not separately describe various possible combination methods.

Claims

1. A vehicle precise fault diagnosis method, characterized in that, Including: The in-vehicle central control obtains the remaining power of the vehicle battery; determines whether the remaining power is less than a first preset value; If so, it outputs that the fault type is a low-battery vehicle and notifies the operation and maintenance personnel; if not, it proceeds to the next step; The in-vehicle central control obtains the actual unlocking times of the vehicle within a preset time period; determines whether the difference between the number of received unlocking commands and the actual unlocking times within the preset time period is greater than a second preset value; if so, it outputs that the fault type is an unlocking timeout fault vehicle and notifies the operation and maintenance personnel; if not, it proceeds to the next step; The in-vehicle central control obtains vehicle sample data and transmits it to the background server; the sample data includes multiple-dimensional information related to the vehicle riding state and a fault label for marking a fault vehicle; The background server determines the vehicle portrait of the vehicle based on the sample data, and determines whether the assist ratio of the vehicle exceeds a predetermined range and reaches a predetermined number of times within a predetermined time period, where the assist ratio = motor output torque / pedal torque; And inputs the vehicle portrait into a training model to output the fault type; The training model is a fault recognition model obtained by weighting the vehicle portrait processed by the machine learning model and its corresponding fault label. The training model includes a first training model and a second training model. The first training model is trained based on the vehicle portrait and the first fault type; The second training model is trained based on the vehicle portrait and the second fault type; if the assist ratio is higher than the maximum value of the predetermined range, the vehicle portrait is input into the first training model; if the assist ratio is lower than the minimum value of the predetermined range, the vehicle portrait is input into the second training model.

2. The vehicle precise fault diagnosis method according to claim 1, characterized in that The dimensional information includes: detection data on the operating status of each vehicle part.

3. The vehicle precise fault diagnosis method according to claim 2, wherein The detection data on the operating status of each vehicle part includes: vehicle riding speed, pedal output torque, motor output torque, battery output voltage, battery temperature, braking pressure, braking distance.

4. The vehicle precise fault diagnosis method according to claim 1, wherein The dimensional information includes: detection data on the operating status of each vehicle part, vehicle basic attribute data, historical fault and maintenance data, vehicle historical riding data.

5. The vehicle precise fault diagnosis method according to claim 4, wherein The detection data on the operating status of each vehicle part includes: vehicle riding speed, pedal output torque, motor output torque, battery output voltage, battery temperature, braking pressure, braking distance; The vehicle basic attribute data includes: vehicle production batch, vehicle operation time, vehicle operation area, order situation and order cancellation volume within a preset time; The vehicle historical fault and maintenance data includes: historical fault situation and repair situation of the vehicle within a preset time period; The vehicle historical riding data includes: historical average driving duration, historical average driving speed, historical average driving distance, historical operation records.

6. The vehicle precise fault diagnosis method according to claim 1, characterized in that, The first fault type includes: power supply fault, motor fault, controller fault, control device fault; The second fault type includes: foot pedal fault, transmission device fault, braking device fault, wheel and appendage fault, handle fault, vehicle body shaking or abnormal noise.

7. The vehicle precise fault diagnosis method according to claim 1, characterized in that The method for the in-vehicle central control to transmit the sample data to the background server includes: Determine whether the communication module is connected to the background server; if so, transmit the sample data to the background server through the communication module; if not, proceed to the next step; Broadcast a request for assistance information through the Bluetooth module until the Bluetooth module on a neighboring vehicle or parking device responds; Transmit the sample data to the Bluetooth module on a neighboring vehicle or parking device through the Bluetooth module; Transmit the sample data of the vehicle to be detected to the background server through the communication module on a neighboring vehicle or parking device.

8. A diagnostic device for the vehicle precise fault diagnosis method according to any one of claims 1 to 7, characterized in that, Includes: A first acquisition unit configured to acquire the remaining power of the vehicle battery; A first judgment unit configured to judge whether the remaining power is less than a first preset value; If so, output the fault type as a low-battery vehicle and notify the operation and maintenance personnel; if not, proceed to the next step; A second acquisition unit configured to acquire the actual unlocking times of the vehicle by the in-vehicle central control within a preset time period; A second judgment unit configured to judge whether the difference between the number of received unlocking commands and the actual unlocking times within the preset time period is greater than a second preset value; if so, output the fault type as an unlocking timeout fault vehicle and notify the operation and maintenance personnel; if not, proceed to the next step; A third acquisition unit configured to acquire vehicle sample data and transmit it to the background server; the sample data includes multiple-dimensional information related to the vehicle riding state and a fault label for marking a fault vehicle; A first processing unit configured to determine the vehicle portrait of the vehicle based on the sample data and judge whether the assist ratio of the vehicle exceeds a predetermined range and reaches a predetermined number of times within a predetermined time period; The assist ratio is = motor output torque / pedal torque; And input the vehicle portrait into the training model to output the fault type; The training model is a fault recognition model obtained by weighting the vehicle portrait processed by the machine learning model and its corresponding fault label. The training model includes a first training model and a second training model. The first training model is trained based on the vehicle portrait and the first fault type; The second training model is trained based on the vehicle portrait and the second fault type; if the assist ratio is higher than the maximum value of the predetermined range, input the vehicle portrait into the first training model; if the assist ratio is lower than the minimum value of the predetermined range, input the vehicle portrait into the second training model.

9. A server for precise vehicle fault diagnosis, characterized in that, Includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 7.

Images