Vehicle repair store vehicle diagnosis method and device based on AI and VR technologies

Abstract

The invention provides a vehicle repair store vehicle diagnosis method and device based on AI and VR technologies, and relates to the field of data processing. The method comprises the steps that vehicle operation data of a target vehicle in a vehicle repair shop are acquired, and the vehicle operation data comprise temperature data, vibration data and sound data; determining an operation feature group according to the temperature data, the vibration data and the sound data; inputting the operation feature group into a preset calculation model, and calculating the fault probability of the target vehicle; mapping the fault probability and the fault position corresponding to the fault probability to a digital twin model of the target vehicle to obtain a to-be-displayed picture; and displaying the to-be-displayed picture to the user through the VR equipment. By implementing the technical scheme provided by the invention, the accuracy of vehicle diagnosis of the vehicle repair shop is improved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and particularly to a vehicle diagnosis method and device for auto repair shops based on AI and VR technologies. Background Art

[0002] With the rapid development of the automotive industry, the design of modern vehicles has become increasingly complex, covering a variety of advanced technologies such as electronic control systems, sensor networks, intelligent powertrains, and autonomous driving assistance functions. This makes vehicle fault diagnosis no longer just a mechanical structure-level inspection, but also involves the comprehensive analysis of electronics, software, and sensor data.

[0003] Currently, in traditional auto repair shops, vehicle fault diagnosis mainly relies on the experience of technicians and manual inspections. Technicians usually judge the health status of vehicles through visual inspections, auscultation, tactile tests, and fault code readings. However, this method relying on manual experience has limitations. Due to differences in the professional levels and experiences of technicians, different technicians may draw different conclusions for the same fault, making it difficult to guarantee the accuracy of the diagnosis results.

[0004] Therefore, there is an urgent need for a vehicle diagnosis method and device for auto repair shops based on AI and VR technologies. Summary of the Invention

[0005] This application provides a vehicle diagnosis method and device for auto repair shops based on AI and VR technologies, which is convenient for improving the accuracy of vehicle diagnosis in auto repair shops.

[0006] In the first aspect of this application, a vehicle diagnosis method for auto repair shops based on AI and VR technologies is provided. The method includes: obtaining vehicle operation data of a target vehicle in an auto repair shop, where the vehicle operation data includes temperature data, vibration data, and sound data; determining an operation feature group according to the temperature data, the vibration data, and the sound data; inputting the operation feature group into a pre-designed calculation model to calculate the fault probability of the target vehicle; mapping the fault probability and the fault location corresponding to the fault probability to the digital twin model of the target vehicle to obtain a to-be-displayed picture; and displaying the to-be-displayed picture to a user through a VR device.

[0007] By adopting the above technical solutions, combining temperature, vibration, and sound data, the vehicle operation status is analyzed from multiple angles, avoiding the limitations of a single sensor, and improving the accuracy of fault identification. By means of a pre-designed calculation model, the fault probability is automatically calculated, reducing the time for technicians to manually check and accelerating the fault diagnosis process, thereby improving the maintenance efficiency. Through a digital twin model, the fault location is mapped to the vehicle model, making the fault point visually presented, helping technicians quickly locate the problem and improving the maintenance accuracy. By displaying the fault scenario through VR devices, technicians or users can more intuitively understand the fault situation, enhancing interactivity, reducing communication errors, and improving the quality of maintenance services. By collecting and analyzing operation data, an intelligent diagnosis system is established, which can be used for predictive maintenance, reducing sudden failures and enhancing the vehicle operation stability. Therefore, it is convenient to improve the accuracy of vehicle diagnosis in auto repair shops.

[0008] Optionally, the obtaining of the vehicle operation data of the target vehicle in the auto repair shop specifically includes: receiving the original measurement data of the target vehicle sent by the sensor device group; performing data processing on the original measurement data to obtain the vehicle operation data, and the data processing includes denoising, filtering, and normalization processing.

[0009] By adopting the above technical solutions, the vehicle operation data is automatically collected by the sensor device group, avoiding the errors of manual detection and improving the real-time performance and accuracy of data acquisition. Through denoising, filtering, and normalization processing, interference signals are eliminated and the data quality is optimized, enabling the subsequent calculation model to more accurately analyze the vehicle status. The preprocessed data is more standardized, which can enhance the adaptability of the AI calculation model and enable it to maintain a high fault diagnosis accuracy rate in different vehicles and different environments. Through effective data cleaning, the influence of sensor noise or abnormal data on the analysis results is avoided, improving the accuracy of fault detection and reducing the false positive rate. The data after normalization processing is more standardized, which helps to reduce the calculation complexity, improve the system processing speed, and optimize the utilization of storage resources.

[0010] Optionally, the determining of the operation feature group according to the temperature data, the vibration data, and the sound data specifically includes: performing smoothing processing on the temperature data to obtain a temperature feature, where the temperature feature is used to locate the specific part with abnormal temperature; performing spectral analysis on the vibration data to extract vibration features; performing noise reduction and feature extraction on the sound data to obtain voiceprint features; and generating the operation feature group based on the temperature feature, the vibration feature, and the voiceprint feature.

[0011] By adopting the above technical solution, combining three types of data: temperature, vibration, and sound, comprehensively analyzing the vehicle operating state, avoiding the limitations of a single data source, improving the comprehensiveness and accuracy of fault identification. Reducing data fluctuations, accurately positioning the abnormal temperature part, and helping to identify overheating or cooling system faults. Extracting key frequency features to identify possible mechanical faults, such as bearing wear, component looseness, etc. Removing environmental noise, obtaining the key voiceprint features of vehicle operation, and helping to detect abnormal noise sources. By comprehensively analyzing temperature features, vibration features, and voiceprint features, a more comprehensive operation feature group is formed, reducing the risk of misjudgment of a single data source and improving the stability of the diagnostic result. Through the combination of multiple types of features, the calculation model can adapt to different vehicle models and different working conditions, improving the recognition ability of multiple fault modes. Through reasonable noise reduction, smoothing, spectrum analysis and other processing methods, data redundancy is reduced, the efficiency of the subsequent calculation model is improved, and the fault diagnosis process is accelerated.

[0012] Optionally, inputting the operation feature group into a pre-designed calculation model to calculate the fault probability of the target vehicle specifically includes: for the same position of the target vehicle, inputting the temperature data into the first sub-model to obtain the first fault probability; inputting the vibration data into the second sub-model to obtain the second fault probability; inputting the sound data into the third sub-model to obtain the third fault probability, and the pre-designed calculation model includes the first sub-model, the second sub-model, and the third sub-model; performing compensation probability fusion on the first fault probability, the second fault probability, and the third fault probability to obtain the fault probability of the target vehicle.

[0013] By adopting the above technical solution, the first, second, and third sub-models are respectively constructed for temperature, vibration, and sound data, ensuring that each data type can be optimally analyzed, and improving the fault detection accuracy of a single data dimension. Each sub-model processes different physical characteristics respectively, avoiding mutual interference between data, making the fault diagnosis more accurate and objective, and reducing the impact of noise or abnormal data on the overall judgment. By fusing multiple fault probabilities, the misjudgment that may occur in a single model is compensated, ensuring that the final fault probability is more stable and reliable, and reducing the possibility of misdiagnosis or missed diagnosis. Adopting the method of sub-model + fusion model enables the system to be flexibly adjusted according to different vehicle models and working conditions, such as adding sub-models for current, voltage, exhaust gas analysis, etc., to enhance the system adaptability. Combining with AI algorithms, each sub-model can be continuously optimized based on historical data, improving the long-term fault prediction ability, and gradually constructing a more accurate intelligent diagnosis system.

[0014] Optionally, performing compensation probability fusion on the first fault probability, the second fault probability, and the third fault probability to obtain the fault probability of the target vehicle is specifically calculated using the following calculation formula: ; Among them, P 目标 is the failure probability of the target vehicle, i is the i-th sensor device, n is the number of sensor devices, and P i is the failure probability calculated by the i-th sensor device.

[0015] By adopting the above technical solution, probability fusion is carried out through the data of multiple sensors, avoiding misjudgment caused by a single data source and improving the reliability of the diagnosis result. Using a mathematical model to calculate the data of different sensors is more standardized than manual experience judgment, reducing errors caused by human factors and making the calculation of failure probability more scientific and reproducible. By means of compensating probability fusion, the weights of data from different sensors can be balanced, reducing the interference of abnormal data of individual sensors on the overall judgment and improving the stability of fault detection. The formula supports the input of different types or quantities of sensors, enabling this method to adapt to different vehicle models and various fault modes and improving the versatility of the system. This method can calculate the failure probability on real-time data streams, support online monitoring and real-time fault warning, and improve the maintenance efficiency and service quality of auto repair shops.

[0016] Optionally, mapping the failure probability and the failure location corresponding to the failure probability to the digital twin model of the target vehicle to obtain a display screen to be shown specifically includes: obtaining a three-dimensional model of the target vehicle through 3D scanning technology; performing model layering processing on the target vehicle to obtain the digital twin model; marking the failure probability and the failure location corresponding to the failure probability in the digital twin model in a preset manner to obtain the display screen to be shown, and the preset manner includes color highlighting and dynamic animation.

[0017] By adopting the above technical solution, a digital twin model of the vehicle is constructed through 3D scanning technology, presenting the fault diagnosis result in a three-dimensional visualization manner to help technicians quickly understand the fault location and severity. Through model layering processing, the fault area can be viewed step by step to accurately identify the faults of specific components, improving the pertinence of maintenance work and reducing the disassembly time. Using preset methods such as color highlighting and dynamic animation makes the fault points stand out significantly, enabling maintenance personnel to quickly focus on the problem and improving the efficiency and accuracy of maintenance decision-making. Combining the calculation result of the failure probability can not only display the current fault point but also predict potential fault risks based on historical data, optimizing the predictive maintenance ability. The digital twin model obtained by 3D scanning can adapt to different brands and models of vehicles. By adjusting the model layering structure and annotation method, the versatility and scalability of the solution are ensured. The digital twin model can be combined with VR devices to support remote maintenance guidance or technician training, reducing on-site misjudgment and improving the efficiency of maintenance collaboration.

[0018] Optionally, the method further includes: obtaining training information; inputting the training information into an adaptive feature fusion network for training to obtain a first training result, where the adaptive feature fusion network includes LSTM, CNN, and DNN; performing superposition and normalization processing on the first training result and the training information to obtain a second training result; inputting the second training result into the adaptive feature fusion network for processing to obtain a third training result; performing superposition and normalization processing on the third training result and the second training result until the training information similarity matrix is output, and the training information similarity matrix satisfies a preset logistic regression condition.

[0019] By adopting the above technical solution, LSTM is suitable for processing time series data, such as the temperature, vibration, and sound signals of vehicle sensors, and can learn historical dependence relationships. CNN is good at extracting features from local data patterns, such as the spectral features of sound signals and vibration waveforms, etc., to improve the ability to extract fault features. DNN has a powerful feature learning ability, which can further optimize the fusion and classification of multi-modal data and improve the accuracy of fault recognition. Through collaborative training of multiple models, different data features are adaptively fused to ensure a comprehensive understanding of complex vehicle states and reduce misjudgments caused by the limitations of a single model. With multi-stage training, the result of each round of training will be superimposed and normalized with the original data, continuously optimizing the feature representation and making the model gradually converge to the optimal solution. By measuring the relationship between different fault modes through the similarity matrix, the model can automatically learn the correlation between fault features, improving the accuracy and generalization ability of classification. The training process is continuously optimized until the training information similarity matrix satisfies the preset logistic regression condition, ensuring that the output of the final model is stable and reliable and avoiding overfitting or underfitting problems.

[0020] In a second aspect of the present application, a vehicle diagnosis device for an auto repair shop based on AI and VR technologies is provided. The vehicle diagnosis device for the auto repair shop includes an acquisition module and a processing module. Among them, the acquisition module is used to acquire vehicle operation data of a target vehicle in the auto repair shop, and the vehicle operation data includes temperature data, vibration data, and sound data; the processing module is used to determine an operation feature group according to the temperature data, the vibration data, and the sound data; the processing module is further used to input the operation feature group into a pre-designed calculation model to calculate the fault probability of the target vehicle; the processing module is further used to map the fault probability and the fault location corresponding to the fault probability to the digital twin model of the target vehicle to obtain a to-be-displayed picture; the processing module is further used to display the to-be-displayed picture to the user through a VR device.

[0021] In a third aspect of the present application, an electronic device is provided. The electronic device includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions. Both the user interface and the network interface are used for communicating with other devices. The processor is used to execute the instructions stored in the memory so that the electronic device executes the method described above.

[0022] In a fourth aspect of the present application, a computer-readable storage medium is provided. The computer-readable storage medium stores instructions that, when executed, perform the method described above.

[0023] In summary, one or more technical solutions provided in the present application have at least the following technical effects or advantages: Combining temperature, vibration, and sound data to analyze the vehicle operation status from multiple angles, avoiding the limitations of a single sensor, and improving the accuracy of fault identification. Automatically calculating the fault probability through a pre-designed calculation model, reducing the time for technicians to manually check, accelerating the fault diagnosis process, and improving the maintenance efficiency. Through the digital twin model, mapping the fault location to the vehicle model to visually present the fault point, helping technicians quickly locate the problem and improving the maintenance accuracy. Displaying the fault picture through VR devices, enabling technicians or users to more intuitively understand the fault situation, enhancing interactivity, reducing communication errors, and improving the quality of maintenance services. By collecting and analyzing operation data, establishing an intelligent diagnosis system, which can be used for predictive maintenance, reducing sudden failures, and enhancing the vehicle operation stability. Therefore, it is convenient to improve the accuracy of vehicle diagnosis in auto repair shops. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic flowchart of a method for diagnosing vehicles in auto repair shops based on AI and VR technologies provided by an embodiment of the present application; Figure 2 It is another schematic flowchart of a method for diagnosing vehicles in auto repair shops based on AI and VR technologies provided by an embodiment of the present application; Figure 3 It is a schematic block diagram of a device for diagnosing vehicles in auto repair shops based on AI and VR technologies provided by an embodiment of the present application; Figure 4 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

[0025] Description of the reference numerals: 31, acquisition module; 32, processing module; 41, processor; 42, communication bus; 43, user interface; 44, network interface; 45, memory. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments.

[0027] In the description of the embodiments of this application, words such as "for example" or "for instance" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "for example" or "for instance" is intended to present relevant concepts in a specific manner.

[0028] In the description of the embodiments of this application, the meaning of the term "plural" refers to two or more. For example, a plurality of systems refers to two or more systems, and a plurality of screen terminals refers to two or more screen terminals. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The terms "include", "comprise", "have" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0029] With the rapid development of the automotive industry, the design of modern cars has become increasingly complex, integrating advanced technologies such as electronic control systems, sensor networks, intelligent powertrains, and autonomous driving assistance functions. The diagnosis of vehicle faults is no longer limited to the detection at the mechanical structure level, but has evolved into a comprehensive analysis process involving electronic systems, software algorithms, and multi-source sensing data.

[0030] However, in current traditional auto repair shops, fault diagnosis still mainly relies on the experience of technicians and manual inspections. Usually, technicians judge the health status of vehicles by means of visual inspections, auscultation, tactile tests, and reading fault codes. However, this method based on manual experience has significant limitations. On the one hand, the professional levels and experiences of technicians vary, resulting in different or even contradictory conclusions for the same fault by different technicians, reducing the accuracy of diagnosis. On the other hand, the involvement of complex electronic systems and sensor data makes it difficult to comprehensively and accurately identify faults solely by traditional methods, affecting the repair efficiency and user experience.

[0031] To solve the above technical problems, this application provides a vehicle diagnosis method for auto repair shops based on AI and VR technologies, referring to Figure 1 , Figure 1Schematic flowchart of a vehicle diagnosis method for an auto repair shop integrating AI and VR technologies provided by an embodiment of this application. This method is applied to a server and includes steps S110 to S150 as follows: S110. Obtain the vehicle operation data of the target vehicle in the auto repair shop, where the vehicle operation data includes temperature data, vibration data, and sound data.

[0032] Specifically, the server obtains the vehicle operation data of the target vehicle in the auto repair shop, which means the whole process of data collection, transmission, and processing. The vehicle operation data here includes temperature data, vibration data, and sound data. These data come from sensors installed in different parts of the vehicle and are used to monitor the vehicle operation status in real time. The temperature data comes from engine temperature sensors, cooling system sensors, brake temperature sensors, transmission temperature sensors, etc., to monitor the temperature of various vehicle components and identify problems such as overheating, insufficient cooling, or abnormal friction. The vibration data comes from acceleration sensors or gyroscopes to detect mechanical faults such as engine jitter, abnormal suspension systems, and wheel imbalance. The sound data is captured by microphones or special acoustic sensors to detect abnormal noises during vehicle operation, such as abnormal engine noises, brake pad wear sounds, and bearing fault noises.

[0033] The data collected by these sensors will be transmitted to the server of the auto repair shop through in-vehicle networks (CAN bus, Ethernet) or wireless communication (Wi-Fi, 5G, Bluetooth) for real-time monitoring or storage and analysis. The server will perform preprocessing such as denoising, filtering, and normalization on the obtained raw data for subsequent analysis. These data can be automatically analyzed in combination with AI models and provide information such as fault prediction, trend analysis, and health scores to help technicians more accurately judge the vehicle status.

[0034] In a possible implementation, obtaining the vehicle operation data of the target vehicle in the auto repair shop specifically includes: receiving the raw measurement data of the target vehicle sent by the sensor device group; performing data processing on the raw measurement data to obtain the vehicle operation data, and the data processing includes denoising, filtering, and normalization processing.

[0035] Specifically, multiple groups of sensor devices on the vehicle will monitor parameters such as temperature, vibration, sound, pressure, current, and voltage in real time and send the data to the server. These raw data may be disturbed by factors such as environmental noise, sampling errors, and sensor drift, so further processing is required. The engine temperature sensor detects 90°C, but the high surrounding temperature may cause errors and need to be corrected. The tire vibration sensor records fluctuations within a certain range, but some may be caused by uneven road surfaces and need to filter out interference signals. The microphone picks up the engine sound, but it may also contain wind noise or background noise and needs noise reduction processing. The data collected by the sensors will be transmitted to the server or the diagnostic equipment in the auto repair shop through in-vehicle buses (such as CAN bus), wireless networks (such as Wi-Fi, 5G, Bluetooth), or local storage. After receiving the raw data, the server will store it and enter the data processing stage. When the vehicle enters the shop, the technician reads the real-time data of the vehicle through the OBD-II interface and transmits it to the server through Wi-Fi. In the remote diagnosis scenario, the in-vehicle terminal directly sends the data to the cloud server for processing through 4G / 5G. After receiving the raw data from the sensors, the server needs to perform a series of data cleaning and preprocessing to improve the accuracy and usability of the data.

[0036] Since the raw data may contain environmental noise, electromagnetic interference, or sensor errors, filtering algorithms or signal processing methods need to be used to eliminate invalid data. Engine noise may be mixed with wind noise, road noise, or background voices, and wavelet noise reduction can be used to extract the sound features related to faults. The vibration caused by uneven road surfaces can be removed by high-pass filtering, and only abnormal vibration signals are retained. There may be fluctuations in the sensor sampling data, and smoothing processing is required to extract the real signal. The engine temperature may have short-term fluctuations due to measurement errors of the sensor, and moving average filtering can be used to smooth the temperature curve and avoid misjudging as overheating. The vibration signal of the tire is affected by road conditions, and low-pass filtering can be used to remove high-frequency interference and only retain low-frequency abnormal vibration information.

[0037] Since the data dimensions of different sensors are different (such as temperature in °C, vibration in m / s², and sound in dB), direct processing may lead to problems of numerical scale mismatch. Normalization can unify the data range and improve the stability of model training and calculation. For example: Min-Max normalization, Z-score standardization, etc. The engine temperature ranges from 0 - 150°C and is normalized to the [0, 1] interval for unified calculation with vibration data (0 - 5 m / s²) and sound data (0 - 100 dB). The vibration data may have different units (g-force, m / s²), and after normalization, the contribution weights of different data sources are ensured to be consistent.

[0038] S120. Determine the operation characteristic group according to the temperature data, vibration data, and sound data.

[0039] Specifically, the server extracts various operation characteristics based on the temperature data, vibration data, and sound data to form an operation characteristic group. The core of this process is signal processing, feature extraction, and data fusion for subsequent fault diagnosis and analysis. The operation characteristic group refers to representative characteristic parameters extracted from the original data, which can describe the operation state of the vehicle and be used to determine whether there are abnormal conditions. For example, the temperature change trend of the engine, the vibration frequency pattern of the braking system, the noise characteristics of the transmission, etc. The temperature data comes from temperature sensors that monitor the temperature of components such as the engine, brake discs, cooling system, and exhaust pipes. The vibration data comes from acceleration sensors and gyroscopes that monitor the vibration of components such as the engine, transmission, suspension system, and tires. The sound data comes from microphones that monitor abnormal engine noises, abnormal brake noises, transmission noises, etc.

[0040] From the temperature curve recorded by the engine temperature sensor, if it rises sharply within a short period, it may indicate a cooling system failure. If the transmission vibration sensor detects a high-frequency vibration signal, it may be due to gear wear or jamming. If the sound sensor captures a sharp friction sound, it may be due to severe wear of the brake pads. To ensure the reliability of the data, the server needs to perform preprocessing on the original data, such as denoising, smoothing, filtering, and normalization. Use moving average filtering to remove the sampling noise of the sensor and obtain a smooth temperature curve. Calculate the temperature rise rate (the amount of temperature change per unit time) to determine whether there is an abnormal temperature rise trend. Use high-pass filtering to remove the low-frequency interference caused by road bumps and only retain the vibration signals related to mechanical failures. Calculate the vibration amplitude and frequency distribution to determine whether there is abnormal vibration. Use wavelet denoising to remove background noise (such as wind noise, human voices, etc.) and retain the mechanical sounds of the vehicle. Perform FFT (Fourier transform) to extract the main sound characteristic frequencies and analyze whether there is abnormal noise.

[0041] The server combines the above various characteristic data to form a complete operation characteristic group for fault diagnosis. If the operation data of the target vehicle is as follows: the engine temperature rises abnormally (110°C), the temperature rise rate is greater than 10°C / min, the transmission vibration spectrum shows abnormal high-frequency components, and there is a sharp noise of 80 Hz in the sound data, then the server will combine these characteristics into an operation characteristic group and input it into the AI diagnosis model to calculate the fault probability.

[0042] In a possible implementation, an operation feature group is determined based on temperature data, vibration data, and sound data, specifically including: performing smoothing processing on the temperature data to obtain temperature features, which are used to locate the specific parts with temperature anomalies; performing spectral analysis on the vibration data to extract vibration features; performing noise reduction and feature extraction on the sound data to obtain voiceprint features; and generating an operation feature group based on the temperature features, vibration features, and voiceprint features.

[0043] Specifically, through the smoothing processing of temperature data, sudden noises are eliminated, the data is ensured to be stable, and the specific parts with temperature anomalies are located, such as the engine, brake disc, or cooling system. The sliding mean filter smooths the temperature change and reduces data jitter. The temperature rise rate calculation is used to judge whether there is an abnormal temperature rise trend. The heat distribution analysis combines multiple temperature sensors to locate the overheated area. For example, if the engine temperature rises from 85°C to 110°C within 5 minutes and the temperature rise rate exceeds the normal value (10°C / min), it may indicate coolant leakage or water pump failure. The brake disc temperature reaches 400°C, which is much higher than the normal range (100 - 250°C), and it may be brake sticking or severe wear of the brake pads.

[0044] Perform spectral analysis on the vibration data to extract features for judging whether there are abnormal vibration modes, such as engine jitter, transmission gear wear, suspension system looseness, etc. FFT (Fast Fourier Transform) converts the vibration signal from the time domain to the frequency domain to extract the main frequency components of the vibration. The root mean square value (RMS) of the vibration is used to measure the vibration intensity, and if it exceeds the threshold, there may be a fault. Peak analysis identifies high-frequency vibrations to judge abnormal mechanical faults. For example, if a peak at 2.5 kHz appears in the vibration spectrum of the transmission, corresponding to the gear wear feature, it may be poor gear meshing. The RMS value of the suspension system vibration is 50% higher than normal, which may be due to shock absorber aging or looseness.

[0045] Perform noise reduction and feature extraction on the sound data to identify abnormal noises, such as engine abnormal noises, brake abnormal noises, bearing wear noises, etc. Wavelet noise reduction removes environmental noises, such as wind noise and human voices, and only retains mechanical sounds. MFCC (Mel Frequency Cepstral Coefficients) feature extraction simulates the auditory characteristics of the human ear to extract the key features of the sound. Sound spectrum analysis identifies abnormal frequency components, such as sharp friction sounds and impact sounds. Engine sound analysis finds a "clicking" impact sound (the main components are in the range of 1.2 kHz - 1.8 kHz), which may be a valve fault. A sharp friction sound (>3 kHz) appears in the brake sound, indicating severe wear of the brake pads and the need for replacement. Finally, the server combines the temperature features, vibration features, and voiceprint features to form an operation feature group for subsequent AI analysis.

[0046] S130: Input the operation feature group into a pre-designed calculation model to calculate the failure probability of the target vehicle.

[0047] Specifically, the server obtains the operation feature groups of the vehicle (temperature features, vibration features, voiceprint features, etc.) and inputs them into a pre-designed calculation model. The calculation model consists of multiple sub-models, and each sub-model processes different sensor data. Finally, the failure probability of the vehicle is calculated through compensated probability fusion. Further, the feature data such as temperature, vibration, and sound received by the server are respectively input into the corresponding sub-models. The first sub-model (temperature analysis model) identifies the temperature abnormal area and determines whether it exceeds the warning threshold. For example, if the temperature of the brake disc rises abnormally (higher than 450 °C), it may indicate a problem with the braking system. The failure probability P1 caused by the temperature abnormality is output. The second sub-model (vibration analysis model) detects whether there is an abnormal vibration mode through spectrum analysis. For example, if an abnormal peak appears in the 2.5 kHz frequency band of the transmission, it may be gear wear. The failure probability P2 caused by the vibration abnormality is output. The third sub-model (voiceprint analysis model) detects abnormal noises through voiceprint analysis, such as "clicking" abnormal noises or friction sounds. For example, if a sharp impact sound (1.2 Hz) appears in the engine noise spectrum, it may be a valve failure. The failure probability P3 caused by the sound abnormality is output.

[0048] Based on the failure probabilities (P1, P2, P3) calculated by each sub-model, the server performs fusion calculation through the compensated joint probability method. This can avoid misjudgment caused by a single data source (such as misjudging as an increase in ambient temperature when relying solely on temperature). The multi-sensor complementarity improves the diagnostic accuracy (such as increased temperature + abnormal vibration + abnormal noise, which can clarify the transmission failure).

[0049] In a possible implementation manner, inputting the operation feature group into the pre-designed calculation model to calculate the failure probability of the target vehicle specifically includes: for the same position of the target vehicle, inputting the temperature data into the first sub-model to obtain the first failure probability; inputting the vibration data into the second sub-model to obtain the second failure probability; inputting the sound data into the third sub-model to obtain the third failure probability. The pre-designed calculation model includes the first sub-model, the second sub-model, and the third sub-model; performing compensated probability fusion on the first failure probability, the second failure probability, and the third failure probability to obtain the failure probability of the target vehicle.

[0050] Specifically, for example, Case 1: Transmission gear failure, abnormal temperature (P1 = 60%): The temperature of the transmission rises and exceeds the normal value, abnormal vibration (P2 = 70%): High-frequency vibration is detected, which may be gear wear, abnormal sound (P3 = 80%): "Clicking" noise appears, which conforms to the characteristics of gear damage. After fusion calculation, P (transmission failure) = 90%. The transmission gear may be severely worn or damaged, and it is recommended to repair immediately.

[0051] Case 2: Poor engine heat dissipation and abnormal temperature (P1 = 75%): The engine temperature is too high (110°C), the vibration is normal (P2 = 15%): There is no abnormal vibration, and the sound is normal (P3 = 10%): There is no abnormal noise. After fusion calculation, P(engine failure) = 50%. There may be a problem with the engine heat dissipation, but no mechanical damage is found. It is recommended to check the coolant or the status of the water pump.

[0052] Therefore, the server synthesizes temperature, vibration, and sound data to reduce false positives and false negatives, avoiding human errors. The server calculates the probability of failure based on a mathematical model, which is more accurate than traditional manual diagnosis. With real-time monitoring, it improves the maintenance efficiency: It can quickly determine the location of the failure, reduce the vehicle downtime, and improve the maintenance efficiency. Compared with the traditional technician's experience-based judgment, this method combines AI + multi-sensor analysis, making the vehicle fault diagnosis in auto repair shops more efficient and intelligent.

[0053] In a possible implementation, a compensated probability fusion is performed on the first failure probability, the second failure probability, and the third failure probability to obtain the failure probability of the target vehicle. The specific calculation is carried out using the following formula: ; where, P 目标 is the failure probability of the target vehicle, i is the i-th sensor device, n is the number of sensor devices, and P i is the failure probability calculated by the i-th sensor device.

[0054] Specifically, this section describes a method of compensated probability fusion, that is, how to use the independent failure probabilities of multiple sensors to calculate the final failure probability of the target vehicle. During the vehicle diagnosis process, the failure probability provided by a single sensor may be affected by various factors, such as environmental noise, signal interference, measurement errors, etc. Therefore, different sensors may yield different failure probabilities. To improve the accuracy, these failure probabilities need to be fused and calculated to obtain a more reliable final result. The final failure probability of the target vehicle is the final conclusion obtained by comprehensively considering the data of multiple sensors. The failure probability calculated by the i-th sensor, for example, the failure probabilities given by the temperature sensor, vibration sensor, and sound sensor respectively. The total number of sensors, for example, 3 sensors (temperature, vibration, sound) may be used in a system. The goal of compensated probability fusion is to synthesize multiple independent failure probabilities to obtain the most reasonable final failure probability, using the following logic: The measurement data of different sensors may be inconsistent (for example, the temperature is abnormal but the vibration is normal). Some sensors are more reliable than others, so different weights can be assigned. If multiple sensors detect abnormalities simultaneously, the probability of failure should be higher; otherwise, it may just be a false alarm.

[0055] For example, if the temperature sensor shows a 70% probability of failure, it means that the engine temperature is relatively high and there may be a heat dissipation problem. The vibration sensor only shows a 30% probability of failure, indicating that the internal mechanical vibration of the engine is basically normal. The sound sensor shows a 20% probability of failure, indicating that there is no abnormal noise in the engine. After fusion calculation, due to the abnormal temperature (70%), but the vibration and sound are basically normal (30%, 20%), the system may judge that the engine has poor heat dissipation, but there is no mechanical failure. Through compensated probability fusion, the final calculated P 目标 = 50%, indicating that this problem requires further inspection, but it does not belong to an emergency failure. It is recommended to check the coolant or the radiator fan, but there is no need to immediately replace engine components.

[0056] S140. Map the probability of failure and the corresponding failure location to the digital twin model of the target vehicle to obtain the picture to be displayed.

[0057] Specifically, the server has calculated the probability of failure of each component through multiple sensors (temperature, vibration, sound, etc.) and determined the possible location of the failure. The server obtains the three-dimensional structure model of the vehicle through 3D scanning or a CAD database. It performs hierarchical modeling, refining to different subsystems such as the engine, transmission, and braking system. The server matches the probability of failure value with the specific failure location. It uses color highlighting, animation, or other visual methods to display the health status of the failed component. A visual three-dimensional vehicle diagnosis picture is formed and displayed to the maintenance personnel through VR devices, smart terminals, and maintenance systems. It is intuitive and visual, improving the maintenance efficiency. Technicians no longer need to rely on abstract numerical values or fault codes, but can directly understand the vehicle problems through 3D, saving the diagnosis time. Traditional maintenance relies on experience judgment and may result in misdiagnosis. The combination of digital twin and AI analysis can provide more accurate maintenance guidance. Maintenance personnel can remotely view the failure picture, and experts at the headquarters can remotely guide novice technicians. It can also be used for training new employees to improve their maintenance capabilities. The system can record historical failure data, form a failure mode library, optimize future diagnosis models, and improve the diagnosis accuracy.

[0058] In a possible implementation, mapping the probability of failure and the corresponding failure location to the digital twin model of the target vehicle to obtain the picture to be displayed specifically includes: obtaining the three-dimensional model of the target vehicle through 3D scanning technology; performing model hierarchical processing on the target vehicle to obtain the digital twin model; and marking the probability of failure and the corresponding failure location in the digital twin model in a preset manner to obtain the picture to be displayed, where the preset manner includes color highlighting and dynamic animation.

[0059] Specifically, through 3D scanning technology, using methods such as laser scanning or photogrammetry, three-dimensional geometric data of the target vehicle is obtained, thereby generating a three-dimensional model of the vehicle. This model can accurately restore the vehicle's appearance and internal structure. The obtained three-dimensional model is subjected to layer processing, that is, each system or component of the vehicle (such as the engine, transmission, braking system, chassis, etc.) is separated to form a structured digital twin model. The purpose of this is to facilitate subsequent fault location and visualization annotation of specific components. In the digital twin model, through a preset annotation method, the calculated fault probability and the corresponding fault location are mapped onto the model. For example, the system marks each component according to the fault probability. Components with a high fault probability are displayed in a prominent color (such as red or orange), while components with a low risk are displayed in a more ordinary color. Dynamic animations can set flashing, pulsing, or other animation effects for high-risk areas to further alert the technician to the possible faults in these components. After integrating the above steps, the server generates a three-dimensional fault diagnosis screen containing all annotation information for the maintenance technician to view and interact with in real time through a display device (such as a VR device, tablet, or computer screen).

[0060] For example, Example 1: Engine cooling system failure. First, use 3D scanning technology to obtain a three-dimensional model of the engine compartment of a certain vehicle model. The server divides the interior of the engine compartment into layers such as the engine body, cooling system, and exhaust system. The system detects that the fault probability of the water pump component in the engine cooling system is 75%, while the fault probability of the coolant pipeline is 40%. In the digital twin model, the location of the water pump is highlighted in red and accompanied by a flashing animation to indicate a high risk; while the coolant pipeline is highlighted in orange to indicate a certain potential hazard. The maintenance technician sees a three-dimensional model through the VR device, and the water pump and pipeline in the engine cooling system are clearly marked. After the technician clicks on the water pump, detailed fault information and a recommended repair plan can be seen.

[0061] Example 2: Abnormal transmission gears. Obtain the internal three-dimensional model of the vehicle's transmission through 3D scanning. The server decomposes the transmission into different components such as gears, bearings, and transmission mechanisms. The system detects that the fault probability of the second gear in the transmission reaches 85%, while the fault probabilities of other components are relatively low. In the digital twin model, the area of the second gear is highlighted in red and accompanied by a dynamic pulsing animation to attract the attention of the technician; other areas remain normally displayed. The technician views the model through the touch screen or VR interface and immediately focuses on the fault area of the second gear. The system simultaneously displays the detection data, fault probability, and repair suggestions.

[0062] S150. Display the screen to be shown to the user through the VR device.

[0063] Specifically, in the previous steps, the server has mapped the failure probabilities of each vehicle component and their corresponding locations onto the digital twin model of the vehicle, generating a three-dimensional fault diagnosis screen with annotation information such as color highlighting and dynamic animations. This screen visually displays the status of each vehicle component and can clearly identify the areas where faults may exist. To enable maintenance technicians or users to view this screen more intuitively, the server transmits the screen to a virtual reality-supported display device (such as a VR headset, VR glasses, or a monitor compatible with the VR system). After wearing the VR device, the user can observe the digital twin model of the vehicle from a first-person perspective in a three-dimensional virtual space. The user can interact with the screen through a handle or body sensing operations, such as rotating the model, zooming in on details, and clicking to view detailed diagnostic information of specific components. This display method not only makes the fault location and probability visually visible but also highlights high-risk areas using dynamic animations and color annotations, helping users quickly identify problems and guiding subsequent maintenance decisions.

[0064] For example, Example 1: Engine fault display. Background: During the maintenance of a vehicle, a relatively high failure probability of the water pump in the engine cooling system was detected. Generated screen: The server used 3D scanning technology to construct a digital twin model of the engine compartment and marked the failure probability of the water pump in the cooling system area as 75%, highlighted in red, and accompanied by a flashing animation. VR display: The maintenance technician wears a VR headset and enters the virtual environment, seeing a three-dimensional engine compartment model. The position of the water pump in the model flashes red light. The technician can view detailed fault data (such as abnormal temperature, failure probability, etc.) by rotating and zooming in on the model. The system may also provide text descriptions or voice prompts to explain the currently detected problems and recommended maintenance measures.

[0065] Example 2: Transmission fault display. Background: The failure probability of the second gear in the transmission was detected to be as high as 85%, shown in red and accompanied by a dynamic pulse animation. Generated screen: The server has mapped this information into the digital twin model of the transmission, and the fault location is clearly marked. VR display: Through the VR device, the technician can "enter" the interior of the transmission and observe all gear components in a 360° panoramic view. The high-risk area (the second gear highlighted in red and with a pulse animation) is particularly conspicuous. After clicking, the technician can see a detailed fault description, such as "The gear is severely worn, and disassembly and inspection are recommended." The technician can use the tools in the virtual environment to simulate disassembly and assembly operations to pre-enact the maintenance process in advance and reduce the on-site operation difficulty.

[0066] Therefore, by displaying the screen to be shown through the VR device, the system transforms traditional two-dimensional fault information into an immersive three-dimensional visualization effect, enabling technicians and users to visually observe and understand the vehicle's fault condition. This not only improves the accuracy of diagnosis but also significantly enhances the maintenance efficiency and user experience, providing strong technical support for intelligent vehicle repair.

[0067] In a possible implementation manner, referring to Figure 2 , Figure 2 is another process schematic diagram of a vehicle diagnosis method for an auto repair shop based on AI and VR technologies provided by an embodiment of the present application. It includes steps S210 to S250, and the above steps are as follows: S210, obtain training information; S220, input the training information into an adaptive feature fusion network for training to obtain a first training result, and the adaptive feature fusion network includes LSTM, CNN, and DNN; S230, after superimposing and normalizing the first training result and the training information, obtain a second training result; S240, input the second training result into the adaptive feature fusion network for processing to obtain a third training result; S250, superimpose and normalize the third training result and the second training result until a training information similarity matrix is output, and the training information similarity matrix satisfies a preset logistic regression condition.

[0068] Specifically, the server first collects and obtains a batch of training data (i.e., "training information"), which can be multimodal data of vehicles (such as temperature, vibration, sound data) or other relevant feature data. These data are used as the initial input for subsequent model training. The first feature fusion training inputs the collected training information into an "adaptive feature fusion network", which integrates three different neural network architectures: LSTM is good at capturing dependencies in time series data and is suitable for processing continuous data (such as temperature changes in time series). CNN is used to extract local spatial features, such as local patterns in images or signals. DNN extracts higher-level abstract features through a multi-layer fully connected network. The network processes the training information and outputs a first training result, which is usually a preliminary extracted feature vector or feature representation. The first training result is "superimposed" (i.e., the two are fused and added) and normalized with the original training information. The purpose of this step is to fuse the information in the original data with the features extracted by the network to obtain a more complete and balanced feature representation. This new result is called the second training result. The second training result is input into the same adaptive feature fusion network again for further processing, and a third training result is output. This process is equivalent to further optimizing and abstracting the already fused features to extract deeper-level feature information. The third training result and the second training result are superimposed and normalized again, and this process is repeated continuously until a training information similarity matrix is finally output. This similarity matrix reflects the feature similarity degree between each training sample and can be regarded as a description of the entire training data space. When the similarity matrix satisfies the preset logistic regression condition, it indicates that after multiple iterations, the feature expression of the data is stable enough and has strong discriminability, which is suitable for subsequent classification tasks or fault probability prediction.

[0069] For example, assume there are three vehicles A, B, and C. After the above iterative training and feature fusion processing, the similarity matrix output by the system is as follows: the similarity between A and B is 0.85 (indicating that they are very similar in terms of fault features and are likely to belong to the same fault category), the similarity between A and C is 0.40 (indicating that their features are quite different and the possible fault causes may be different), and the similarity between B and C is 0.45. When this matrix meets the requirements of the logistic regression model, logistic regression can be used to classify these vehicles to determine which vehicle has a higher fault risk, thus assisting in the maintenance decision-making.

[0070] This application also provides a vehicle diagnostic device for an auto repair shop based on AI and VR technologies. Refer to Figure 3 , Figure 3 which is a schematic diagram of the modules of a vehicle diagnostic device for an auto repair shop based on AI and VR technologies provided by an embodiment of this application. The device is a server, and the server includes an acquisition module 31 and a processing module 32. Among them, the acquisition module 31 acquires the vehicle operation data of the target vehicle in the auto repair shop, and the vehicle operation data includes temperature data, vibration data, and sound data; the processing module 32 determines an operation feature group according to the temperature data, vibration data, and sound data; the processing module 32 inputs the operation feature group into a pre-designed calculation model to calculate the fault probability of the target vehicle; the processing module 32 maps the fault probability and the fault location corresponding to the fault probability to the digital twin model of the target vehicle to obtain a to-be-displayed picture; the processing module 32 displays the to-be-displayed picture to the user through a VR device.

[0071] In a possible implementation manner, the acquisition module 31 acquires the vehicle operation data of the target vehicle in the auto repair shop, specifically including: the acquisition module 31 receives the original measurement data sent by the sensor device group for the target vehicle; the processing module 32 performs data processing on the original measurement data to obtain the vehicle operation data, and the data processing includes denoising, filtering, and normalization processing.

[0072] In a possible implementation manner, the processing module 32 determines an operation feature group according to the temperature data, vibration data, and sound data, specifically including: the processing module 32 performs smoothing processing on the temperature data to obtain a temperature feature, and the temperature feature is used to locate the specific part with abnormal temperature; the processing module 32 performs spectral analysis on the vibration data to extract vibration features; the processing module 32 performs noise reduction and feature extraction on the sound data to obtain voiceprint features; the processing module 32 generates an operation feature group based on the temperature feature, vibration feature, and voiceprint feature.

[0073] In a possible implementation, the processing module 32 inputs the operation feature group into a pre-designed calculation model to calculate the failure probability of the target vehicle, which specifically includes: the processing module 32 inputs the temperature data into the first sub-model for the same position of the target vehicle to obtain the first failure probability; the processing module 32 inputs the vibration data into the second sub-model to obtain the second failure probability; the processing module 32 inputs the sound data into the third sub-model to obtain the third failure probability. The pre-designed calculation model includes the first sub-model, the second sub-model, and the third sub-model; the processing module 32 performs compensated probability fusion on the first failure probability, the second failure probability, and the third failure probability to obtain the failure probability of the target vehicle.

[0074] In a possible implementation, the processing module 32 performs compensated probability fusion on the first failure probability, the second failure probability, and the third failure probability to obtain the failure probability of the target vehicle, and specifically calculates using the following calculation formula: ; where P 目标 is the failure probability of the target vehicle, i is the i-th sensor device, n is the number of sensor devices, and P i is the failure probability calculated by the i-th sensor device.

[0075] In a possible implementation, the processing module 32 maps the failure probability and the failure location corresponding to the failure probability to the digital twin model of the target vehicle to obtain a display-ready screen, which specifically includes: the processing module 32 obtains the three-dimensional model of the target vehicle through 3D scanning technology; the processing module 32 performs model layering on the target vehicle to obtain the digital twin model; the processing module 32 marks the failure probability and the failure location corresponding to the failure probability in the digital twin model in a preset manner to obtain the display-ready screen, and the preset manner includes color highlighting and dynamic animation.

[0076] In a possible implementation, the acquisition module 31 acquires training information; the processing module 32 inputs the training information into an adaptive feature fusion network for training to obtain a first training result. The adaptive feature fusion network includes LSTM, CNN, and DNN; the processing module 32 performs superposition and normalization processing on the first training result and the training information to obtain a second training result; the processing module 32 inputs the second training result into the adaptive feature fusion network for processing to obtain a third training result; the processing module 32 performs superposition and normalization processing on the third training result and the second training result until a training information similarity matrix is output, and the training information similarity matrix meets the preset logistic regression condition.

[0077] It should be noted that: when the device provided in the above embodiment realizes its functions, only the division of the above-mentioned functional modules is used for illustration. In practical applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept. For the specific implementation process, please refer to the method embodiment, which will not be elaborated here.

[0078] This application also provides an electronic device. Referring to Figure 4 , Figure 4 which is a schematic structural diagram of an electronic device provided by an embodiment of this application. The electronic device may include: at least one processor 41, at least one network interface 44, a user interface 43, a memory 45, and at least one communication bus 42.

[0079] Among them, the communication bus 42 is used to realize the connection and communication between these components.

[0080] Among them, the user interface 43 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 43 may further include a standard wired interface and a wireless interface.

[0081] Among them, the network interface 44 may optionally include a standard wired interface and a wireless interface (such as a Wi-Fi interface).

[0082] Among them, the processor 41 may include one or more processing cores. The processor 41 connects various parts within the entire server through various interfaces and lines, and executes various functions of the server and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 45, and by calling data stored in the memory 45. Optionally, the processor 41 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 41 may integrate one or a combination of several of a central processing unit (CPU), a graphics processing unit (GPU), and a modem, etc. Among them, the CPU mainly processes the operating system, user interface, and application programs, etc.; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 41 and may be implemented separately by a single chip.

[0083] Among them, the memory 45 may include a Random Access Memory (RAM), or may include a Read-Only Memory. Optionally, the memory 45 includes a non-transitory computer-readable storage medium. The memory 45 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 45 may include a program storage area and a data storage area. Among them, the program storage area can store instructions for implementing the operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-mentioned method embodiments, etc.; the data storage area can store the data involved in the above-mentioned method embodiments. Optionally, the memory 45 may also be at least one storage device located far from the aforementioned processor 41. As Figure 4 shown, in the memory 45 as a computer storage medium, there may be included an operating system, a network communication module, a user interface module, and an application program of a vehicle diagnosis method for an auto repair shop based on AI and VR technologies.

[0084] In Figure 4 the electronic device shown, the user interface 43 is mainly used to provide an input interface for the user to obtain the data input by the user; and the processor 41 can be used to call the application program of a vehicle diagnosis method for an auto repair shop based on AI and VR technologies stored in the memory 45. When executed by one or more processors, the electronic device is caused to execute the method as described in one or more of the above embodiments.

[0085] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0086] This application also provides a computer-readable storage medium, and the computer-readable storage medium stores instructions. When executed by one or more processors, the electronic device is caused to execute the method as described in one or more of the above embodiments.

[0087] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0088] In several embodiments provided in this application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling, direct coupling, or communication connection between each other can be through some service interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0089] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] In addition, each functional unit in various embodiments of this application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0091] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of this application. And the aforementioned memory includes: various media such as USB flash drives, mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0092] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure still fall within the scope covered by the present disclosure. Those skilled in the art will easily think of other implementation schemes of the present disclosure after considering the specification and the disclosure of the practical truth. This application aims to cover any variations, uses, or adaptive changes of the present disclosure, and these variations, uses, or adaptive changes follow the general principles of the present disclosure and include common general knowledge or conventional technical means in the technical field not recorded in the present disclosure. The specification and the embodiments are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A vehicle diagnosis method for auto repair shops based on AI and VR technology, characterized in that: The method comprises: Acquire vehicle operation data of the target vehicle at the auto repair shop, wherein the vehicle operation data includes temperature data, vibration data, and sound data; determining an operation characteristic group according to the temperature data, the vibration data, and the sound data; Inputting the operation feature group into a preset calculation model to calculate the failure probability of the target vehicle; Mapping the fault probability and the fault location corresponding to the fault probability to the digital twin model of the target vehicle to obtain a picture to be displayed; The image to be displayed is displayed to the user through the VR device.

2. The vehicle diagnosis method for auto repair shops based on AI and VR technology according to claim 1 is characterized in that: The obtaining of vehicle operation data of the target vehicle in the auto repair shop specifically includes: Receiving raw measurement data for the target vehicle sent by the sensor device group; The original measurement data is processed to obtain the vehicle operation data, wherein the data processing includes denoising, filtering and normalization processing.

3. The vehicle diagnosis method for auto repair shops based on AI and VR technology according to claim 1 is characterized in that: The determining of the operation feature group according to the temperature data, the vibration data and the sound data specifically includes: Smoothing the temperature data to obtain temperature features, where the temperature features are used to locate specific locations of temperature anomalies; Performing spectrum analysis on the vibration data to extract vibration features; Performing noise reduction and feature extraction on the sound data to obtain voiceprint features; The operation feature group is generated based on the temperature feature, the vibration feature, and the voiceprint feature.

4. The vehicle diagnosis method for auto repair shops based on AI and VR technology according to claim 1 is characterized in that: The step of inputting the operation feature group into a preset calculation model to calculate the failure probability of the target vehicle specifically includes: For the same position of the target vehicle, input the temperature data into a first sub-model to obtain a first failure probability; Inputting the vibration data into a second sub-model to obtain a second failure probability; Inputting the sound data into a third sub-model to obtain a third fault probability, wherein the preset calculation model includes the first sub-model, the second sub-model and the third sub-model; The first failure probability, the second failure probability and the third failure probability are fused with compensation probabilities to obtain the failure probability of the target vehicle.

5. The vehicle diagnosis method for auto repair shops based on AI and VR technology according to claim 4 is characterized in that: The first failure probability, the second failure probability and the third failure probability are subjected to compensation probability fusion to obtain the failure probability of the target vehicle, which is specifically calculated using the following calculation formula: ； Among them, P 目标 is the failure probability of the target vehicle, i is the i-th sensor device, n is the number of sensor devices, P i The failure probability calculated for the i-th sensor device.

6. The vehicle diagnosis method for auto repair shops based on AI and VR technology according to claim 1 is characterized in that: Mapping the fault probability and the fault position corresponding to the fault probability to the digital twin model of the target vehicle to obtain a picture to be displayed specifically includes: Acquire a three-dimensional model of the target vehicle by using 3D scanning technology; Performing model layering processing on the target vehicle to obtain the digital twin model; The fault probability and the fault position corresponding to the fault probability are marked in the digital twin model in a preset manner to obtain the picture to be displayed, and the preset manner includes color highlighting and dynamic animation.

7. The vehicle diagnosis method for auto repair shops based on AI and VR technology according to claim 1 is characterized in that: The method further comprises: Get training information; Inputting the training information into an adaptive feature fusion network for training to obtain a first training result, wherein the adaptive feature fusion network includes LSTM, CNN and DNN; After superimposing and normalizing the first training result and the training information, a second training result is obtained; Inputting the second training result into the adaptive feature fusion network for processing to obtain a third training result; The third training result and the second training result are superimposed and standardized until the training information similarity matrix is ​​output, and the training information similarity matrix satisfies a preset logistic regression condition.

8. A vehicle diagnostic device for auto repair shops based on AI and VR technology, characterized in that: The vehicle diagnostic device for a car repair shop comprises an acquisition module (31) and a processing module (32), wherein: The acquisition module (31) is used to acquire vehicle operation data of the target vehicle in the auto repair shop, wherein the vehicle operation data includes temperature data, vibration data and sound data; The processing module (32) is used to determine an operation feature group based on the temperature data, the vibration data and the sound data; The processing module (32) is further used to input the operation feature group into a preset calculation model to calculate the failure probability of the target vehicle; The processing module (32) is further used to map the fault probability and the fault position corresponding to the fault probability to the digital twin model of the target vehicle to obtain a picture to be displayed; The processing module (32) is also used to display the image to be displayed to the user through a VR device.

9. An electronic device, characterized in that: The electronic device comprises a processor (41), a memory (45), a user interface (43) and a network interface (44), wherein the memory (45) is used to store instructions, the user interface (43) and the network interface (44) are both used to communicate with other devices, and the processor (41) is used to execute the instructions stored in the memory (45) so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is performed.

Images