Intelligent vehicle maintenance management system and method based on big data analysis

Abstract

The invention discloses an intelligent vehicle maintenance management system and method based on big data analysis, and relates to the technical field of intelligent vehicle maintenance, and the method comprises the following steps: collecting various kinds of vehicle driving data in real time in the vehicle driving process through a data collection device, and providing a reliable basis for subsequent analysis and decision making. Through real-time analysis of vehicle driving data and combination of a machine learning model, the system can dynamically adjust the detection frequency and the maintenance strategy, and it is ensured that maintenance better fits the actual use environment. For a rugged road surface, the system accurately recognizes a high-risk state according to the impact load, the vehicle body rolling acceleration and other characteristics, early warning is conducted in time, the maintenance frequency is optimized, excessive maintenance is avoided, meanwhile, the safety is improved, and the maintenance cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent vehicle maintenance, and specifically to an intelligent vehicle maintenance management system and method based on big data analysis. Background Art

[0002] Intelligent vehicle maintenance management based on big data analysis is to collect and analyze a large amount of data generated during the use of vehicles, such as vehicle speed, fuel consumption, temperature, driving mileage, driving behavior, fault records, etc. By using advanced machine learning and data mining technologies, it can intelligently predict and manage the operating status, potential faults and maintenance requirements of vehicles. By monitoring and analyzing data, the system can identify possible fault points or maintenance opportunities in advance, optimize the maintenance plan, improve the operating efficiency of vehicles, extend their service life, reduce the failure rate, and reduce unnecessary maintenance costs. In addition, maintenance strategies can be adjusted according to different driving modes and usage environments, making vehicle maintenance more personalized and intelligent, so as to achieve more efficient resource allocation and cost control.

[0003] The prior art has the following deficiencies:

[0004] Existing vehicle maintenance management technologies usually set fixed inspection periods based on the driving history data of vehicles. In this way, by collecting and analyzing the driving data of vehicles, regular evaluations are carried out to determine whether the vehicle needs maintenance. This method works well on urban paved roads because the urban roads are flat, the road surface conditions are good, and the driving conditions of vehicles are relatively stable. However, when the vehicle enters a large number of unpaved roads such as the Sichuan-Tibet Line, the harsh road environment poses a great challenge to the stability and reliability of the vehicle. The driving conditions on unpaved roads are more complex, and the vehicle will encounter frequent vibrations, impacts and frictions in this environment, which significantly increase the risk of failures. For example, key components such as the suspension system, tires, braking system, and transmission system are more likely to wear and damage in such an environment. If the vehicle still relies on a fixed inspection period for maintenance, potential hidden dangers may have occurred before the next inspection, resulting in mechanical failures not being detected in time. If the failures are not handled in time, it may cause the vehicle to be unable to effectively respond to sudden problems in extreme road conditions, such as brake failure, suspension breakage or transmission system failure, thus triggering serious traffic accidents and even endangering life safety.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, so it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide an intelligent vehicle maintenance management system and method based on big data analysis. By collecting and analyzing vehicle driving data in real time and combining with a machine learning model for intelligent evaluation, the detection frequency and maintenance plan can be dynamically adjusted to ensure that the maintenance is more suitable for the actual use environment. Under rough road conditions, the system accurately identifies high-risk states based on key features such as the impact load index and the body roll acceleration index, issues early warnings in a timely manner and conducts maintenance. At the same time, by classifying the driving state into normal roads and rough roads through intelligent evaluation, the maintenance strategy is optimized to avoid over-maintenance and increase the detection frequency, thereby reducing the maintenance cost and improving the vehicle safety and reliability, so as to solve the problems in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: An intelligent vehicle maintenance management method based on big data analysis, including the following steps:

[0008] Through a data collection device, various types of vehicle driving data are collected in real time during the vehicle driving process, providing a reliable basis for subsequent analysis and decision-making;

[0009] Sort and preprocess the collected real-time driving data, establish a data analysis set, and extract key features reflecting the vehicle driving state from the analysis set;

[0010] Within the detection window, analyze the extracted key features, input the analyzed key features into a pre-trained machine learning model, and intelligently evaluate the current driving state of the vehicle through the machine learning model;

[0011] Based on the evaluation results of the machine learning model, classify the driving state of the vehicle into two categories: normal road driving and rough road driving;

[0012] For normal road driving, continue to evaluate and maintain according to the preset fixed detection cycle to ensure the reasonable use of resources and cost control;

[0013] For rough road driving, since the vehicle faces higher failure risks under harsh road conditions, based on the evaluation results of the machine learning model, increase the detection frequency and implement more stringent maintenance measures.

[0014] Preferably, key features reflecting the driving state of the vehicle are extracted from the analysis set. The extracted key features include the impact intensity suffered by each component of the vehicle during driving and the lateral stability of the vehicle during driving. Under the detection window, the impact intensity suffered by each component of the vehicle during driving and the lateral stability of the vehicle during driving are analyzed, and an impact load index and a body roll acceleration index are generated respectively. The impact load index quantifies the impact intensity borne by each component of the vehicle during driving; the body roll acceleration index quantifies the lateral stability of the vehicle during driving.

[0015] Preferably, after obtaining the impact load index and the body roll acceleration index generated after analyzing the extracted key features, the impact load index and the body roll acceleration index are input into a pre-trained machine learning model, and a stable driving coefficient is generated through the machine learning model, and the current vehicle driving state is intelligently evaluated through the stable driving coefficient.

[0016] Preferably, the stable driving coefficient generated by comprehensively analyzing the obtained key features through the machine learning model is compared and analyzed with a pre-set stable driving coefficient reference threshold to divide the vehicle driving state. The specific division steps are as follows:

[0017] If the stable driving coefficient is greater than or equal to the pre-set stable driving coefficient reference threshold, the vehicle driving state is classified as driving on a rough road;

[0018] If the stable driving coefficient is less than the pre-set stable driving coefficient reference threshold, the vehicle driving state is classified as driving on a regular road.

[0019] Preferably, for driving on a rough road, since the vehicle faces a higher failure risk under harsh road conditions, based on the evaluation result of the machine learning model, the detection frequency is increased, and the specific steps for implementing more stringent maintenance measures are as follows:

[0020] Compare the generated stable driving coefficient SDC with the preset stable driving coefficient reference threshold SDC Th to determine the vehicle driving state. The calculation expression is as follows:

[0021] ΔSDC = SSDC - SDC Th

[0022] , where ΔSDC is the difference between the stable driving coefficient SDC and the preset stable driving coefficient reference threshold SDC Th ;

[0023] When it is determined that the vehicle is driving on a rough road and the vehicle driving state is unstable, the detection frequency is adjusted according to the degree of instability. The calculation expression is as follows:

[0024]

[0025] , where DFAR is the detection frequency adjustment ratio for adjusting the detection frequency, A and B are both adjustment coefficients. A controls the intensity of the influence of the difference ΔSDC on the detection frequency adjustment ratio DFAR, B controls the power response of the difference ΔSDC, e is the natural base, C defines the gain of the Sigmoid function, and D is the decay factor of the Sigmoid function;

[0026] Based on a preset fixed detection period and the detection frequency adjustment ratio DFAR, a new detection frequency is calculated, and the calculation expression is as follows:

[0027]

[0028] , where NDF is the new detection frequency, that is, the adjusted detection period, FDC is the preset fixed detection period, Q is an adjustment coefficient used to further adjust the non-linear influence of the ΔSDC difference on the detection frequency, and P is an exponential coefficient used to adjust the non-linear influence of the ΔSDC difference.

[0029] Preferably, under the detection window, the specific steps for analyzing the impact intensity suffered by each vehicle component during driving to generate an impact load index are as follows:

[0030] Under the detection window, the impact data suffered by each vehicle component during driving is collected in real time through a data acquisition device, and then the acquired data is subjected to noise filtering to remove high-frequency noise and ensure the timeliness and accuracy of the data;

[0031] By performing real-time measurement on the preprocessed data, the instantaneous impact intensity of each component at each detection moment is calculated, and the calculation expression is as follows:

[0032]

[0033] , where I i (t) is the instantaneous impact intensity, that is, the instantaneous impact intensity of vehicle component i at time point t, k n is the weight coefficient of acceleration and force, A i (t) is the instantaneous acceleration, that is, the instantaneous acceleration of vehicle component i at time point t, α n is the non-linear exponent of acceleration on the impact intensity, F i (t) is the instantaneous force, that is, the instantaneous force borne by vehicle component i at time point t, β n is the non-linear exponent of force on the impact intensity, n is the specific item of the nth acceleration and force on the impact intensity, and N is the total number of specific items of acceleration and force on the impact intensity;

[0034] After calculating the instantaneous impact strength I of each component i (t), these data are summarized to obtain the impact load of the whole vehicle. The calculation expression is as follows:

[0035]

[0036] , where I total (t) is the impact load, that is, the impact load of the vehicle at time point t. M is the total number of vehicle components, and W i is the weight factor of the component, that is, the weight factor contributed by each component to the total impact load. γ i is the non-linear exponent of component i, and δ is the overall adjustment coefficient of the impact load;

[0037] After obtaining the impact load I total (t), further considering the cumulative effect of continuous impacts during vehicle driving, the impact load index is dynamically adjusted to generate the final impact load index. The calculation expression is as follows:

[0038]

[0039] , where I adjusted is the impact load index, I total (τ) is the impact load at a certain time point τ. e -λ(t-τ) is the time decay function, which is used to describe the influence degree of the impact load between a certain past moment τ and time point t on the current state. e is the natural base, Δt is the time interval, and λ is the decay factor, which is used to control the decay rate of the influence of the impact load on the current moment.

[0040] Preferably, under the detection window, the lateral stability of the vehicle during driving is analyzed, and the specific steps for generating the body roll acceleration index are as follows:

[0041] Under the detection window, collect the lateral acceleration data of the vehicle and preprocess it. Calculate the lateral acceleration of the vehicle based on the preprocessed acceleration data. The calculation expression is as follows:

[0042] a filtered (t) = Filter(a raw (t))

[0043] , where a raw (t) is the original lateral acceleration, which represents the lateral acceleration value of the vehicle at time point t. Filter is the filter, and a filtered (t) is the filtered lateral acceleration;

[0044] By analyzing the sudden change of the lateral acceleration during vehicle driving, the instantaneous change rate of the lateral acceleration is extracted. By calculating the instantaneous change rate of the acceleration, the moment of sharp change is identified, and the change amplitude is further analyzed. The calculation expression is as follows:

[0045] Δa(t)=|a filtered (t)-a filtered (t-Δt)|

[0046] , where a filtered (t-Δt) is the lateral acceleration value at the previous moment with a time interval of Δt from the time point t, and Δa(t) is the instantaneous change rate of the lateral acceleration;

[0047] According to the fluctuation characteristics of the lateral acceleration during vehicle driving, a dynamic model is established to quantify the lateral stability of the vehicle. The vehicle speed and the road surface friction factor are introduced, and an acceleration model is constructed to describe the dynamic change of the vehicle lateral stability. The calculation expression is as follows:

[0048]

[0049] , where v(t) is the current vehicle speed, that is, the vehicle speed at the time point t, μ road is the road surface friction factor, which refers to the friction coefficient between the vehicle and the current road surface, m is the vehicle weight, and a dynamic (t) is the dynamic lateral acceleration of the vehicle;

[0050]

[0051] , where RAI is the vehicle body roll acceleration index, ω is the speed weight factor, is the road surface friction factor weight, θ is the index adjustment factor, s is the total number of sampling points, and j is the sampling point.

[0052] An intelligent vehicle maintenance and management system based on big data analysis includes a data collection module, a data sorting and preprocessing module, a feature analysis and model input module, a driving state evaluation and classification module, a conventional road maintenance management module, and a rough road maintenance management module:

[0053] The data collection module, through the data collection device, collects various vehicle driving data in real time during vehicle driving, providing a reliable basis for subsequent analysis and decision-making;

[0054] The data sorting and preprocessing module sorts and preprocesses the collected real-time driving data, establishes a data analysis set, and extracts the key features reflecting the vehicle driving state from the analysis set;

[0055] Feature analysis and model input module: within the detection window, analyze the extracted key features and input the analyzed key features into a pre-trained machine learning model to intelligently evaluate the current driving state of the vehicle through the machine learning model;

[0056] Driving state evaluation and classification module: based on the evaluation results of the machine learning model, classify the driving state of the vehicle into two categories: normal road driving and rough road driving;

[0057] Normal road maintenance management module: for normal road driving, continue to evaluate and maintain according to the preset fixed detection cycle to ensure the rational use of resources and cost control;

[0058] Rough road maintenance management module: for rough road driving, since the vehicle faces higher failure risks under harsh road conditions, based on the evaluation results of the machine learning model, increase the detection frequency and implement more stringent maintenance measures.

[0059] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0060] By collecting and analyzing the driving data of the vehicle in real time and combining with the machine learning model to intelligently evaluate the driving state of the vehicle, the present invention can dynamically adjust the detection frequency and maintenance plan. Under rough road conditions, the system can accurately identify whether the vehicle is in a high-risk state according to the changes in key features such as the impact load index and the body roll acceleration index, issue early warnings in a timely manner and perform maintenance. This intelligent, data-driven method can ensure that vehicle maintenance is more in line with the actual use environment, avoid potential hidden dangers that may be missed by traditional methods, and significantly improve the timeliness and accuracy of maintenance.

[0061] Through the intelligent evaluation based on the machine learning model, the driving state of the vehicle can be accurately classified into two categories: normal road driving and rough road driving, and then the detection cycle and maintenance strategy can be adjusted according to the driving state. For normal road driving, the system continues to adopt the traditional fixed detection cycle, which can effectively avoid unnecessary over-maintenance and optimize resource utilization; for rough road driving, the system will increase the detection frequency and take more stringent maintenance measures, thus avoiding equipment damage or safety hazards caused by untimely detection and maintenance. This can not only ensure the safety of the vehicle, but also effectively reduce unnecessary maintenance costs and achieve the best balance between resources and costs. Description of the Drawings

[0062] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for use in the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0063] Figure 1 This is the method flow chart of the intelligent vehicle maintenance and management method based on big data analysis of the present invention.

[0064] Figure 2 This is the module schematic diagram of the intelligent vehicle maintenance and management system based on big data analysis of the present invention. Detailed implementation manners

[0065] Now, the exemplary embodiments will be described more comprehensively with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that the present disclosure will be more thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0066] The present invention provides an intelligent vehicle maintenance and management method based on big data analysis as shown in Figure 1 and includes the following steps:

[0067] During the vehicle's driving process, various driving data of the vehicle are collected in real time through a data collection device, providing a reliable basis for subsequent analysis and decision-making.

[0068] By collecting various driving data in real time through a data collection device during the vehicle's driving process, a reliable basis can be provided for subsequent analysis and decision-making. These data are usually obtained through in-vehicle sensors and devices, including but not limited to vehicle speed, acceleration, braking state, engine temperature, fuel consumption, suspension system vibration, tire pressure, braking distance, etc. The frequency and accuracy of data collection should be optimized according to specific requirements. By integrating Internet of Things (IoT) technology, the data can be transmitted to the cloud or a local server in real time for storage and processing, ensuring the high availability and timeliness of the data. In addition, the data collection device should have high durability and anti-interference ability to ensure stable operation in harsh environments. In this way, through the systematic monitoring of these real-time data, accurate basis can be provided for subsequent vehicle state analysis, fault prediction, and maintenance decision-making, which helps to improve the safety, reliability, and maintenance efficiency of vehicle operation.

[0069] Sort and preprocess the collected real-time driving data, establish a data analysis set, and extract key features reflecting the vehicle's driving state from the analysis set.

[0070] Organizing and preprocessing the collected real-time driving data is the crucial first step to ensure the accuracy and reliability of subsequent data analysis. First of all, it is necessary to clean the original data, removing noise and outliers, such as missing sensor data or incorrect data caused by environmental interference. Common cleaning methods include interpolating to fill in missing values, removing data points outside the reasonable range, or using filters to remove high-frequency noise. At the same time, it is also necessary to unify the formats of data from different sources, such as converting the data units of different sensors and synchronizing timestamps. Ensuring the consistency, integrity, and accuracy of the data provides a stable foundation for subsequent analysis.

[0071] After completing the data cleaning, it is necessary to integrate the organized data into a unified analysis set for subsequent in-depth analysis and modeling. This step involves data normalization and standardization processing to ensure that the data ranges of different features are consistent and to prevent specific variables from dominating the analysis due to differences in dimensions. For example, converting features with different dimensions such as vehicle speed, acceleration, and fuel consumption into dimensionless standard scores to bring them into the same scale range. Next, through feature selection algorithms, key features closely related to the vehicle's driving state are extracted. These features may include vibration frequency, tire pressure changes, brake system pressure, etc. Finally, by constructing a clear and structured data analysis set, it will lay a solid foundation for subsequent machine learning-based modeling and intelligent evaluation.

[0072] Within the detection window, analyze the extracted key features and input the analyzed key features into a pre-trained machine learning model to intelligently evaluate the current driving state of the vehicle through the machine learning model;

[0073] Extract key features reflecting the vehicle's driving state from the analysis set. The extracted key features include the impact intensity suffered by each vehicle component during driving and the lateral stability of the vehicle during driving. Under the detection window, analyze the impact intensity suffered by each vehicle component during driving and the lateral stability of the vehicle during driving, and generate an impact load index and a body roll acceleration index respectively. The impact load index quantifies the impact intensity borne by each vehicle component during driving; the body roll acceleration index quantifies the lateral stability of the vehicle during driving, especially on unpaved roads.

[0074] After obtaining the impact load index and the body roll acceleration index generated by analyzing the extracted key features, input the impact load index and the body roll acceleration index into a pre-learned machine learning model, generate a stable driving coefficient through the machine learning model, and intelligently evaluate the current driving state of the vehicle through the stable driving coefficient.

[0075] "Pre-trained machine learning model" refers to a model with a certain generalization ability that is pre-trained through the training and optimization of historical data in the vehicle maintenance management system. Based on a large amount of vehicle driving data, especially historical data containing features such as vehicle status, impact force, and acceleration under different road conditions, the model is trained through machine learning algorithms (such as supervised learning, deep learning, etc.) to learn the dynamic behavior patterns of vehicles under different road conditions. Through such a learning process, the model can identify key factors affecting vehicle stability and performance, such as impact load, vehicle body roll, and tire wear, and make predictions and evaluations in various driving environments. Therefore, the pre-trained model can not only identify abnormal situations during vehicle driving but also provide an intelligent assessment of the vehicle driving state based on current sensor data and learned rules during real-time driving.

[0076] The key to this pre-trained model lies in its ability to make accurate predictions in different driving scenarios and environments through a large amount of labeled training data and continuous optimization. During the training phase of the model, data scientists usually use a large amount of historical driving data to extract key features such as vehicle acceleration, speed, impact frequency, and lateral stability, and use these data to establish a mapping relationship between input and output. Through iterative optimization, the model continuously adjusts its parameters and finally achieves an optimal prediction ability. When the model is trained, it can quickly generate a stable driving coefficient (for example, by judging the current driving state through changes in the impact load index and vehicle body roll acceleration index) after the input of new, real-time collected vehicle data. This process does not depend on a pre-set detection cycle or human intervention but automatically makes an assessment through real-time data and machine learning algorithms, thereby improving the accuracy and timeliness of the assessment and ensuring that the maintenance management system can more flexibly adapt to the needs of different driving conditions.

[0077] During the driving process, various components of the vehicle are continuously subjected to relatively large impacts, which usually indicates that the vehicle has been driving on rough roads for a long time. Rough roads include unpaved dirt roads, gravel roads, or areas with uneven road surfaces. These road surfaces usually have numerous potholes, obstacles, and irregular surface morphologies, resulting in the vehicle constantly encountering frequent and intense vibrations and impacts while driving. When the vehicle drives on such roads for a long time, the impact forces borne by components such as the suspension system, tires, and chassis will increase significantly, manifested as the vehicle's vibration intensity, swing amplitude, and driving stability being affected. These long-term impacts not only increase the wear and fatigue of various vehicle components but may also lead to early component failures, such as tire blowouts and suspension system failures. Therefore, if it is found through the monitoring data of the impact load index that the vehicle has been subjected to relatively large impacts for a long time, this indicates that the vehicle is in a state of driving on rough roads for a long time, and this state may have a negative impact on the long-term stability and reliability of the vehicle.

[0078] Under the detection window, the specific steps for analyzing the impact intensity suffered by various vehicle components during driving to generate the impact load index are as follows:

[0079] Under the detection window, use a data acquisition device to collect in real-time the impact data suffered by various vehicle components during driving, and then perform noise filtering on the obtained data to remove high-frequency noise and ensure the timeliness and accuracy of the data;

[0080] When collecting in real-time the impact data suffered by various vehicle components, it is mainly necessary to pay attention to the impact data of the following parts. The impact conditions of these parts directly affect the vehicle's stability, reliability, and failure risk. Through precise impact data collection and noise filtering, potential mechanical failures can be effectively monitored and prevented, especially when driving on rough roads. The following are the parts that need to be focused on and their related impact data:

[0081] Suspension system

[0082] The suspension system is a key connecting component between the vehicle and the road surface, responsible for absorbing the impacts from the uneven ground and maintaining the vehicle body's stability. When driving on rough roads, the suspension system bears relatively large impacts, especially the shock absorbers and springs. It is necessary to focus on the acceleration of the suspension system (such as wheel acceleration, suspension component acceleration) and the vibration frequency of the suspension system. These parameters reflect the suspension's response to impacts and its ability to effectively mitigate impacts.

[0083] Tire system

[0084] When a tire comes into contact with the road surface, especially on unpaved roads, it bears a large impact force. It is necessary to monitor the acceleration data of the tire. Especially when the vehicle passes through obstacles or rough roads, the instantaneous force changes on the tire can reveal the health status of the tire. Key attention should be paid to tire pressure fluctuations and instantaneous torque, as these data reflect whether the tire is damaged due to impacts during driving.

[0085] Brake system

[0086] The brake system often faces large impacts on rough roads, especially when the driver brakes suddenly or encounters an uneven road surface. It is necessary to monitor the pressure changes of the brake pedal and the temperature of the brake pads. At the same time, key attention should be paid to the vibration data of the brake disc and the changes in braking force in the brake system. These data can provide clues as to whether the brake system can respond in a timely manner and whether there is excessive wear or failure.

[0087] Body and chassis

[0088] The body and chassis are the main structural parts of the vehicle, and the impact effects they bear are directly related to the stability of the vehicle. On rough roads, the body is subjected to large vibrations and impact forces. It is necessary to pay attention to data such as vehicle body acceleration (such as acceleration in the front-back and left-right directions), longitudinal, lateral, and vertical vibrations of the vehicle body. In addition, changes in the vehicle body roll angle and the vibration amplitude of the chassis should also be the focus of monitoring, especially when passing through large potholes or slopes.

[0089] Drivetrain

[0090] The drivetrain includes important components such as the engine, transmission, and differential. When driving on rough roads, it often bears large vibrations and is prone to overload or damage. Special attention should be paid to the torque fluctuations of the drivetrain and the vibrations of the bearings. Monitoring these parameters can identify changes in the load of the drivetrain and whether it has been damaged due to impacts.

[0091] Steering system

[0092] The steering system is directly related to the controllability and driving stability of the vehicle. Especially on rough roads, the steering system bears large impacts, which may lead to poor control or damage. It is necessary to focus on the data of the steering force sensor, especially the fluctuations of the steering force. The steering angle change rate and steering response time of the steering system are also important indicators to measure its stability.

[0093] In-vehicle vibration sensor

[0094] Inside the vehicle, especially in the driver's cab, the vibrations of the vehicle may also affect the driver's operating experience and the comfort of the vehicle. By installing vibration sensors inside the vehicle, it is possible to monitor the vibration acceleration data inside the vehicle and analyze the vibration intensity and frequency when the vehicle passes through rough roads to determine whether there is excessive vibration that causes driving difficulties or affects the driver's health.

[0095] Collision avoidance system

[0096] The collision avoidance system is usually triggered during a collision or a strong impact. Therefore, it is also important to monitor the response of the collision avoidance system. Attention should be paid to the collision acceleration sensor and the collision time response data, which can help determine whether the protection mechanism of the collision avoidance system has been triggered due to external impacts during vehicle driving, and then warn of potential collision risks.

[0097] Fuel system and battery system

[0098] When driving on rough roads, the fuel system and the battery system may be vibrated, resulting in fuel leakage or battery damage, especially for the battery pack. In this case, the pressure fluctuation of the fuel pump and the changes in battery voltage / current are the key indicators to focus on. The change in the internal resistance of the battery and the battery temperature fluctuation can also reflect whether the system is damaged or there are potential faults.

[0099] Noise filtering

[0100] Since sensor data is easily affected by high-frequency noise, environmental interference, and sensor itself errors, noise filtering is crucial. The Kalman filter or low-pass filter can be used to effectively filter out high-frequency noise, ensuring that only the signals related to the impact intensity are retained. In addition, for vibration data, methods such as wavelet transform can be used to further extract the effective signals in the relevant frequency bands.

[0101] Through the above steps of impact data acquisition and noise filtering, the stress state of each vehicle component can be monitored in real time, and potential faults can be detected in a timely manner, especially when driving on rough roads, providing a reliable basis for intelligent maintenance management.

[0102] By performing real-time measurements on the preprocessed data, calculate the instantaneous impact intensity of each component at each detection moment. The calculation expression is as follows:

[0103]

[0104] , where I i (t) is the instantaneous impact intensity, that is, the instantaneous impact intensity of vehicle component i at time point t, k n is the weight coefficient of acceleration and acting force, A i (t) is the instantaneous acceleration, that is, the instantaneous acceleration of vehicle component i at time point t, α n is the non-linear exponent of acceleration on the impact intensity, used to adjust the influence degree of acceleration and acting force on the impact intensity, and is adjusted according to the actual characteristics (such as rigidity, mass, etc.) of different components, F i (t) is the instantaneous acting force, that is, the instantaneous acting force borne by vehicle component i at time point t, βn is the non - linear exponent of the acting force on the impact strength, reflecting the non - linear relationship between the acting force and the impact strength. n is the specific item of the nth acceleration and the influencing factor of the acting force on the impact strength, and N is the total number of specific items of the influencing factors of the acceleration and the acting force on the impact strength;

[0105] Calculating the impact strength of each component at each detection moment can accurately evaluate and quantify the load conditions borne by each component at a specific moment, thereby identifying potential risks that may occur during vehicle driving. Different components such as the suspension system, tires, brakes, transmission system, etc., their impact strengths are affected by driving environment, road type, and vehicle motion state. By calculating the impact strength of each component at each detection moment, potential damage or abnormalities can be detected in a timely manner. Especially in complex environments such as rough roads, some components may be damaged prematurely due to excessive force. Such calculations not only contribute to real - time monitoring and diagnosis but also provide data support for subsequent maintenance decisions to ensure the reliability and safety of each vehicle component under different road conditions.

[0106] After calculating the instantaneous impact strength I i (t) of each component, these data are summarized to obtain the impact load of the whole vehicle. The calculation expression is as follows:

[0107]

[0108] , where I total (t) is the impact load, that is, the impact load of the vehicle at time point t. M is the total number of vehicle components, W i is the weight factor of the component, that is, the weight factor contributed by each component to the total impact load. γ i is the non - linear exponent of component i, reflecting the non - linear relationship between the impact strength and the load of the component. δ is the overall adjustment coefficient of the impact load;

[0109] The overall adjustment coefficient δ of the impact load is a global parameter used to adjust the relative weights and influencing degrees of the impact loads borne by each vehicle component when calculating the total impact load. The role of this coefficient is to make the impact load I total (t) more in line with the actual vehicle usage requirements, especially in complex road conditions. Specifically, a larger δ value means that the loads of each component of the vehicle under extreme vibration and impact conditions will be more prominently reflected in the total load index, thereby enhancing the sensitivity to potential failures; while a smaller δ value makes the change of the impact load index relatively stable, reducing the impact of over - reaction of some components on the overall assessment. Therefore, the adjustment coefficient δ is an important tool to ensure that the model can accurately reflect the vehicle's force condition and identify potential risks in different environments.

[0110] Obtain the impact load I total After (t), further considering the cumulative effect of continuous impacts during vehicle driving, dynamically adjust the impact load index to generate the final impact load index. The calculation expression is as follows:

[0111]

[0112] , where I adjusted is the impact load index, I total (τ) is the impact load at a certain time point τ, e -λ(t-τ) is the time decay function, used to describe the influence degree of the impact load between a certain past moment τ and the time point t on the current state. e is the natural base, Δt is the time interval, and λ is the decay factor, used to control the decay rate of the influence of the impact load on the current moment.

[0113] Under the detection window, the larger the value of the impact load index performance generated after analyzing the impact intensity suffered by each vehicle component during driving usually means that the current vehicle has been driving on a rough road surface for a long time. Rough road surfaces include unpaved roads, gravel roads, or potholed roads, and these road surfaces will cause the vehicle to frequently encounter large impacts and vibrations during driving. Therefore, a higher value of the impact load index under the monitoring window reflects the greater impact intensity suffered by the vehicle on such uneven road surfaces. When the vehicle drives on a rough road surface for a long time, components such as the suspension system, tires, and chassis will bear more vibrations and impacts, thus causing the value of the impact load index to continuously increase. On the contrary, if the impact load index is low, it indicates that the road surface on which the vehicle is driving is relatively flat, lacking significant vibrations and impacts, usually an urban paved road or a relatively smooth road surface.

[0114] The reduction of the lateral stability of the vehicle during driving usually indicates that the vehicle has been driving on a rough road for a long time, because the complex terrain on the rough road will have a direct impact on the lateral stability of the vehicle. When the vehicle is driving on an uneven unpaved road, obstacles such as uneven road surface, stones, potholes, etc. will frequently change the driving direction and posture of the vehicle, generating more lateral impacts and severe lateral forces. These forces cause the body of the vehicle to swing violently from side to side, increasing the risk of reduced lateral stability. In particular, when the vehicle is driving on such a road for a long time, the design of the suspension system, wheels and body, and the overall balance of the vehicle will be greatly challenged and will not be able to effectively suppress lateral shaking as on a flat road. Over time, the lateral stability of the vehicle gradually decreases, which may be manifested as the driver's control of the vehicle's direction becomes more difficult during the operation, the body roll increases, and even a large shaking or slipping phenomenon occurs. Therefore, the reduction of lateral stability can usually be used as an important indicator of the vehicle's long-term driving on a rough road, reflecting the gradual impact of the driving environment on the stability of the vehicle.

[0115] In the detection window, the lateral stability of the vehicle during driving is analyzed and the vehicle body rolling acceleration index is generated. The specific steps are as follows:

[0116] In the detection window, the lateral acceleration data of the vehicle is collected and preprocessed, and the lateral acceleration of the vehicle is calculated based on the preprocessed acceleration data. The calculation expression is as follows:

[0117] a filtered (t)=Filter(a raw (t))

[0118] , where a raw (t) is the original lateral acceleration, which indicates the lateral acceleration value of the vehicle at time point t. Filter is a filter, which is an algorithm used to process raw data. In the field of signal processing, it is used to remove noise and unnecessary signal components. filtered (t) is the filtered lateral acceleration;

[0119] Lateral acceleration refers to the acceleration generated by a vehicle along an axis perpendicular to the direction of travel (i.e., the lateral axis) during driving. Specifically, it reflects the change in acceleration along the left and right directions of the vehicle when the vehicle turns, brakes suddenly, or encounters external forces. Lateral acceleration is usually measured by the vehicle's acceleration sensor (such as an inertial measurement unit) in meters per second squared (m / s 2) This kind of acceleration is very important during vehicle driving. Especially when turning, changing lanes, or encountering crosswinds and rough roads, the vehicle will be subjected to lateral forces of varying degrees, resulting in lateral displacement or tilt of the vehicle body. Therefore, lateral acceleration can effectively reflect the lateral stability and handling performance of the vehicle, and is of great significance for analyzing vehicle safety, driving experience, and fault warning.

[0120] By analyzing the sudden changes in lateral acceleration during vehicle driving and extracting the instantaneous change rate of lateral acceleration, it is an important feature reflecting vehicle driving stability. When the vehicle is driving on a rough road, due to frequent vibrations and impacts, the lateral acceleration usually fluctuates significantly. By calculating the instantaneous change rate of acceleration, identifying the moments of sharp changes, and further analyzing the change amplitude, it helps to judge the lateral instability of the vehicle when driving on an unpaved road. The calculation expression is as follows:

[0121] Δa(t) = |a filtered (t) - a filtered (t - Δt)|

[0122] , where a filtered (t - Δt) is the lateral acceleration value at the previous moment with a time interval of Δt from the time point t, and Δa(t) is the instantaneous change rate of lateral acceleration;

[0123] According to the fluctuation characteristics of lateral acceleration during vehicle driving, establish a dynamic model to quantify the lateral stability of the vehicle. Introduce vehicle speed and road surface friction factor, and construct an acceleration model to describe the dynamic changes of vehicle lateral stability. The calculation expression is as follows:

[0124]

[0125] , where v(t) is the current vehicle speed, that is, the vehicle speed at the time point t, μ road is the road surface friction factor, which refers to the friction coefficient between the vehicle and the current road surface, m is the vehicle weight, and a dynamic (t) is the dynamic lateral acceleration of the vehicle, used to describe the lateral stability of the vehicle during driving, especially the dynamic behavior shown on an unpaved road (rough road);

[0126]

[0127] , where RAI is the vehicle body roll acceleration index, ω is the speed weight factor, used to adjust the influence degree of vehicle speed v(t) on the vehicle body roll acceleration index, is the road surface friction factor weight, which is the friction factor μ roadThe weight factor controls the magnitude of the effect of the friction factor on the vehicle body rolling acceleration index. θ is the exponential adjustment factor used to control the final weight of the vehicle body rolling acceleration index. s is the total number of sampling points, and j is the sampling point.

[0128] The summation symbol ∑ indicates that this formula accumulates the calculated values among multiple data points. For each sampling point j, the in the formula is calculated, and the result is raised to the power of θ. Then these values are accumulated to finally obtain the vehicle body rolling acceleration index RAI.

[0129] Under the detection window, the larger the value of the vehicle body rolling acceleration index performance generated by analyzing the lateral stability of the vehicle during driving, generally indicates that the current vehicle has been driving on a rough road surface for a long time. The vehicle body rolling acceleration index quantifies the degree of roll and sway of the vehicle by analyzing the changes in the lateral stability of the vehicle during driving. When the vehicle is driving on a rough road surface, due to the uneven ground and road surface obstacles, the vehicle body will frequently encounter large lateral impacts and vibrations, which leads to an increase in the rolling acceleration of the vehicle body. Being in such a road condition for a long time, the rolling acceleration index of the vehicle will continue to increase, reflecting the severe shaking and instability of the vehicle body under the action of lateral forces. On the contrary, when the vehicle is driving on a flat paved road surface, there are fewer lateral impacts and vibrations, and the vehicle body rolling acceleration index will be lower.

[0130] The machine learning model is not limited here. Any machine learning model that can achieve comprehensive analysis of the impact load index I adjusted and the vehicle body rolling acceleration index RAI to generate the stable driving coefficient SDC is acceptable. To implement the technical solution of the present invention, the present invention provides a specific implementation method:

[0131] The formula for generating the stable driving coefficient SDC is as follows:

[0132]

[0133] , where k 1 , k 2 are respectively the preset proportionality coefficients of the impact load index I adjusted and the vehicle body rolling acceleration index RAI, and k 1 , k 2 are both greater than 0.

[0134] As can be seen from the calculation expression of the stable driving coefficient, under the detection window, the greater the performance value of the impact load index generated after analyzing the impact intensity suffered by each vehicle component during driving, and the greater the performance value of the body roll acceleration index generated after analyzing the lateral stability of the vehicle during driving, the greater the performance value of the stable driving coefficient generated by comprehensively analyzing the obtained key features through the machine learning model, indicating that the current vehicle has been driving on a rough road for a long time, and vice versa, indicating that the current vehicle is driving on a flat and good road surface.

[0135] In the above formula, the preset proportionality coefficients k 1 and k 2 are parameters used to weight the impact load index I adjusted and the body roll acceleration index RAI, with the aim of adjusting according to the relative importance of each index's contribution to the stable driving coefficient SDC. These proportionality coefficients k 1 and k 2 are used to adjust the influence degree of the two key features during the comprehensive analysis process, so as to better reflect the actual driving state of the vehicle. For example, if the impact load of the vehicle has a greater impact on driving stability, then k 1 can be set larger, and vice versa. By presetting these proportionality coefficients, the model can optimize the contributions of the impact load index and the body roll acceleration index to the stability assessment according to the actual situation, making the generated stable driving coefficient SDC more accurate and reasonable. These proportionality coefficients k 1 and k 2 are usually set according to historical data, expert experience or experimental results to ensure that the model can exert its maximum efficiency under different working conditions.

[0136] Based on the evaluation results of the machine learning model, the driving state of the vehicle is divided into two categories: driving on a regular road and driving on a rough road;

[0137] The stable driving coefficient generated by comprehensively analyzing the obtained key features through the machine learning model is compared and analyzed with the pre-set stable driving coefficient reference threshold to divide the driving state of the vehicle. The specific division steps are as follows:

[0138] If the stable driving coefficient is greater than or equal to the pre-set stable driving coefficient reference threshold, the driving state of the vehicle is divided into driving on a rough road;

[0139] If the stable driving coefficient is less than the pre-set stable driving coefficient reference threshold, the driving state of the vehicle is divided into driving on a regular road;

[0140] Normal road driving refers to the state where a vehicle travels on a flat and well-paved road. Such roads usually include urban roads, highways, and other well-maintained roads; rough road driving refers to the state where a vehicle travels on an unpaved or poor-condition road. Such roads usually include rural dirt roads, mountain roads, gravel roads, desert sections, and other rough road surface environments.

[0141] For normal road driving, continue to evaluate and maintain according to the preset fixed inspection cycle to ensure the rational use of resources and cost control;

[0142] For normal road driving, the purpose of continuing to evaluate and maintain according to the preset fixed inspection cycle is mainly to ensure that the vehicle maintenance work can be carried out efficiently and orderly in a stable and relatively ideal driving environment. Under normal road conditions such as urban roads, the vibration and impact on the vehicle are relatively small, and the frequency of faults is relatively low. Therefore, there is no need to frequently adjust the inspection cycle. By following the fixed cycle for routine inspections, such as regular evaluations of key components such as the engine, braking system, and tires, the safety and performance stability of the vehicle during long-term use can be ensured, while avoiding waste of resources caused by excessive inspections. This can not only reasonably allocate and use limited maintenance resources, but also help effectively control costs, ensure that the operating efficiency and operating costs of the vehicle are within a controllable range, and thus bring economic benefits to the transportation company or vehicle owner.

[0143] For rough road driving, since the vehicle faces a higher risk of failure under harsh road conditions, based on the evaluation results of the machine learning model, increase the inspection frequency and implement more stringent maintenance measures;

[0144] For rough road driving, since the vehicle faces a higher risk of failure under harsh road conditions, based on the evaluation results of the machine learning model, the specific steps to increase the inspection frequency and implement more stringent maintenance measures are as follows:

[0145] Compare the generated stable driving coefficient SDC with the preset stable driving coefficient reference threshold SDC Th to determine the vehicle driving state. The calculation expression is as follows:

[0146] ΔSDC = SSDC - SDC Th

[0147] , where ΔSDC is the difference between the stable driving coefficient SDC and the preset stable driving coefficient reference threshold SDC Th ;

[0148] If ΔSDC < 0, it indicates that the current vehicle is driving on a normal road and the vehicle driving state is stable; if ΔSDC ≥ 0, it indicates that the vehicle is driving on a rough road and the vehicle driving state is unstable, and the inspection frequency needs to be increased.

[0149] When it is determined that the vehicle is driving on a rough road surface and the driving state of the vehicle is unstable, the detection frequency is adjusted according to the degree of instability, and the calculation expression is as follows:

[0150]

[0151] , where DFAR is the detection frequency adjustment ratio for adjusting the detection frequency, A and B are both adjustment coefficients. A controls the intensity of the influence of the difference ΔSDC on the detection frequency adjustment ratio DFAR. Increasing A will make the driving state differences of the vehicle more significantly reflected in the frequency adjustment. B controls the power response of the difference ΔSDC. When increasing, the response of the detection frequency adjustment ratio DFAR to the ΔSDC difference will be more sensitive, and even a small change can greatly affect the detection frequency. e is the natural base, C defines the gain of the Sigmoid function, controls the intensity of the detection frequency adjustment, and D is the attenuation factor of the Sigmoid function, which determines the change range of the detection frequency when the difference ΔSDC is small;

[0152] By introducing the Sigmoid function, a smooth transition of the detection frequency near the threshold can be achieved, thus avoiding overly drastic frequency fluctuations and increasing the robustness of the system.

[0153] Based on the preset fixed detection period and the detection frequency adjustment ratio DFAR, a new detection frequency is calculated, and the calculation expression is as follows:

[0154]

[0155] , where NDF is the new detection frequency, that is, the adjusted detection period, FDC is the preset fixed detection period, Q is the adjustment coefficient used to further adjust the non-linear influence of the ΔSDC difference on the detection frequency, and P is the exponential coefficient used to adjust the non-linear influence of the ΔSDC difference;

[0156] Through this formula, the detection period can be dynamically shortened or extended according to the driving state of the vehicle, the detection frequency can be increased, and it can be ensured that the vehicle can be detected more frequently when the vehicle is driving on a rough road surface and the driving state of the vehicle is unstable.

[0157] Under the driving conditions of rough road surfaces, the vehicle faces a more complex and extreme driving environment than on conventional roads, such as frequent vibrations, shocks, and frictions, which all significantly increase the probability of vehicle failures. Traditional fixed-cycle-based detection methods are difficult to cope with this dynamic and unpredictable risk. Especially under extreme road conditions, the vehicle may have been damaged without being detected in time. Therefore, it is of great importance to increase the detection frequency based on the evaluation results of the machine learning model and implement more stringent maintenance measures.

[0158] First, through the intelligent evaluation of the machine learning model, the system can real-time identify the changes in key features such as the impact intensity, vibration frequency, and vehicle body stability suffered by various vehicle components (such as the suspension system, tires, braking system, etc.) when the vehicle is driving on rough roads. These changes can reflect the risks faced by the vehicle during driving, and then predict potential faults. Through this intelligent evaluation, the system can timely detect those latent faults or impending faults that may be caused by poor road conditions.

[0159] Secondly, based on the evaluation results, increasing the detection frequency means that when key components are subjected to greater impacts or increased instability, the system can automatically remind the operator to conduct more frequent inspections and maintenance on these components. This flexible detection mechanism, combined with real-time evaluation, can significantly improve the early warning ability of faults, thereby reducing the probability of sudden faults occurring. Finally, implementing more stringent maintenance measures can ensure that vehicles on rough roads can be repaired and adjusted in a timely manner, reduce the fault risk, extend the service life of the vehicle, and improve safety.

[0160] The present invention can dynamically adjust the detection frequency and maintenance plan by collecting and analyzing the driving data of the vehicle in real time and combining the machine learning model to conduct intelligent evaluation on the driving state of the vehicle. Under rough road conditions, the system can accurately identify whether the vehicle is in a high-risk state according to the changes in key features such as the impact load index and the vehicle body rolling acceleration index, issue early warnings in a timely manner, and conduct maintenance. This intelligent and data-driven method can ensure that vehicle maintenance is more in line with the actual use environment, avoid potential hidden dangers that may be missed by traditional methods, and significantly improve the timeliness and accuracy of maintenance.

[0161] Through the intelligent evaluation based on the machine learning model, the driving state of the vehicle can be accurately classified into two categories: driving on conventional roads and driving on rough roads. Then, the detection cycle and maintenance strategy are adjusted according to the driving state. For driving on conventional roads, the system continues to adopt the traditional fixed detection cycle, which can effectively avoid unnecessary over-maintenance and optimize resource utilization; while for driving on rough roads, the system will increase the detection frequency and adopt more stringent maintenance measures, thereby avoiding equipment damage or safety hazards caused by untimely detection and maintenance. This can not only ensure the safety of the vehicle, but also effectively reduce unnecessary maintenance costs, achieving the best balance between resources and costs.

[0162] The present invention provides an intelligent vehicle maintenance management system based on big data analysis as shown in the figure, including a data collection module, a data sorting and preprocessing module, a feature analysis and model input module, a driving state evaluation and classification module, a conventional road maintenance management module, and a rough road maintenance management module:

[0163] A data acquisition module, through a data acquisition device, collects various types of vehicle driving data in real time during vehicle driving, providing a reliable basis for subsequent analysis and decision-making;

[0164] A data sorting and preprocessing module sorts and preprocesses the collected real-time driving data, establishes a data analysis set, and extracts key features reflecting the vehicle driving state from the analysis set;

[0165] A feature analysis and model input module analyzes the extracted key features within a detection window, inputs the analyzed key features into a pre-trained machine learning model, and intelligently evaluates the current driving state of the vehicle through the machine learning model;

[0166] A driving state evaluation and classification module divides the driving state of the vehicle into two categories: normal road driving and rough road driving based on the evaluation results of the machine learning model;

[0167] A normal road maintenance management module continues to evaluate and maintain according to a preset fixed detection cycle for normal road driving to ensure the reasonable use of resources and cost control;

[0168] A rough road maintenance management module, for rough road driving, since the vehicle faces higher failure risks under harsh road conditions, based on the evaluation results of the machine learning model, increases the detection frequency and implements more stringent maintenance measures.

[0169] The intelligent vehicle maintenance management method based on big data analysis provided by the embodiments of the present invention is implemented through the above-mentioned intelligent vehicle maintenance management system based on big data analysis. The specific methods and processes of the intelligent vehicle maintenance management system based on big data analysis are detailed in the embodiments of the above-mentioned intelligent vehicle maintenance management method based on big data analysis, and will not be elaborated here.

[0170] Only some exemplary embodiments of the present invention have been described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, without departing from the spirit and scope of the present invention, the described embodiments can be modified in various different ways. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. An intelligent vehicle maintenance management method based on big data analysis, characterized in that: The following steps are involved: Through the data acquisition device, various types of vehicle driving data are collected in real time during the vehicle driving process, providing a reliable basis for subsequent analysis and decision-making; Organize and preprocess the collected real-time driving data, establish a data analysis set, and extract key features reflecting the vehicle driving status from the analysis set; Within the detection window, the extracted key features are analyzed and input into a pre-trained machine learning model, which is used to intelligently evaluate the current driving status of the vehicle. Based on the evaluation results of the machine learning model, the vehicle's driving status is divided into two categories: regular road driving and rough road driving; For regular road driving, assessment and maintenance will continue to be carried out according to the preset fixed inspection cycle to ensure the rational use of resources and cost control; For driving on rough roads, since vehicles face a higher risk of failure under harsh road conditions, the inspection frequency is increased and more stringent maintenance measures are implemented based on the evaluation results of the machine learning model.

2. The intelligent vehicle maintenance management method based on big data analysis according to claim 1 is characterized by: Key features reflecting the driving state of the vehicle are extracted from the analysis set. The extracted key features include the impact strength suffered by various components of the vehicle during driving and the lateral stability of the vehicle during driving. Under the detection window, the impact strength suffered by various components of the vehicle during driving and the lateral stability of the vehicle during driving are analyzed to generate an impact load index and a body rolling acceleration index respectively. The impact load index quantifies the impact strength suffered by various components of the vehicle during driving; the body rolling acceleration index quantifies the lateral stability of the vehicle during driving.

3. The intelligent vehicle maintenance management method based on big data analysis according to claim 2 is characterized by: After obtaining the impact load index and the body rolling acceleration index generated by analyzing the extracted key features, the impact load index and the body rolling acceleration index are input into a pre-learned machine learning model, and a stable driving coefficient is generated through the machine learning model. The stable driving coefficient is used to perform an intelligent evaluation of the current vehicle driving state.

4. The intelligent vehicle maintenance management method based on big data analysis according to claim 3 is characterized by: The stable driving coefficient generated by comprehensive analysis of the key features obtained by the machine learning model is compared with the preset stable driving coefficient reference threshold to divide the vehicle driving state. The specific division steps are as follows: If the stable driving coefficient is greater than or equal to a preset stable driving coefficient reference threshold, the vehicle driving state is classified as driving on a rough road; If the stable driving coefficient is less than a preset stable driving coefficient reference threshold, the vehicle driving state is classified as normal road driving.

5. The intelligent vehicle maintenance management method based on big data analysis according to claim 4 is characterized by: For driving on rough roads, since vehicles face a higher risk of failure under bad road conditions, the specific steps to increase the inspection frequency and implement more stringent maintenance measures based on the evaluation results of the machine learning model are as follows: The generated stable driving coefficient SDC is compared with the preset stable driving coefficient reference threshold SDC Th Compare and determine the vehicle driving state. The calculation expression is as follows: ΔSDC=SSDC-SDC Th Where ΔSDC is the difference between the stable driving coefficient SDC and the preset stable driving coefficient reference threshold SDC. Th The difference between When it is judged that the vehicle is driving on a rough road and the driving state of the vehicle is unstable, the detection frequency is adjusted according to the degree of instability. The calculation expression is as follows: Wherein, DFAR is the detection frequency adjustment ratio, which is used to adjust the detection frequency, A and B are adjustment coefficients, A controls the intensity of the difference ΔSDC on the detection frequency adjustment ratio DFAR, B controls the power response of the difference ΔSDC, e is the natural base, C defines the gain of the Sigmoid function, and D is the attenuation factor of the Sigmoid function; The new detection frequency is calculated based on the preset fixed detection period and the detection frequency adjustment ratio DFAR. The calculation expression is as follows: Where NDF is the new detection frequency, i.e., the adjusted detection period, FDC is the preset fixed detection period, Q is the adjustment coefficient, which is used to further adjust the nonlinear effect of the ΔSDC difference on the detection frequency, and P is the exponential coefficient used to adjust the nonlinear effect of the ΔSDC difference.

6. The intelligent vehicle maintenance management method based on big data analysis according to claim 2 is characterized in that: Under the detection window, the impact strength of each vehicle component during driving is analyzed, and the specific steps for generating the impact load index are as follows: In the detection window, the impact data of various components of the vehicle during driving is collected in real time through the data acquisition device, and then the acquired data is subjected to noise filtering to remove high-frequency noise and ensure the timing and accuracy of the data; By measuring the preprocessed data in real time, the instantaneous impact strength of each component at each detection moment is calculated. The calculation expression is as follows: In the formula, I i (t) is the instantaneous impact strength, i.e., the instantaneous impact strength of vehicle component i at time point t, k n is the weight coefficient of acceleration and force, A i (t) is the instantaneous acceleration, i.e., the instantaneous acceleration of vehicle component i at time point t, α n is the nonlinear index of acceleration on impact strength, F i (t) is the instantaneous force, i.e., the instantaneous force on vehicle component i at time point t, β n is the nonlinear index of the effect of force on impact strength, n is the specific item of the nth acceleration and force affecting impact strength, and N is the total number of specific items of the acceleration and force affecting impact strength; After calculating the instantaneous impact strength I of each component i (t), these data are summarized to obtain the impact load of the entire vehicle. The calculation expression is as follows: In the formula, I total (t) is the impact load, i.e., the impact load of the vehicle at time point t, M is the total number of vehicle components, W i is the weight factor of the component, that is, the weight factor of each component's contribution to the total impact load, γ i is the nonlinear index of component i, δ is the overall adjustment coefficient of impact load; Get the impact load I total (t), the cumulative effect of continuous impacts during vehicle driving is further considered, and the impact load index is dynamically adjusted to generate the final impact load index. The calculation expression is as follows: In the formula, I adjusted is the shock load index, I total (τ) is the impact load at a certain time point τ, e -λ(t-τ) It is a time decay function, which is used to describe the influence of the impact load from a certain time τ in the past to the time point t on the current state. e is the natural base, Δt is the time interval, and λ is the decay factor, which is used to control the decay rate of the impact load on the current moment.

7. The intelligent vehicle maintenance management method based on big data analysis according to claim 2 is characterized in that: In the detection window, the lateral stability of the vehicle during driving is analyzed and the vehicle body rolling acceleration index is generated. The specific steps are as follows: In the detection window, the lateral acceleration data of the vehicle is collected and preprocessed, and the lateral acceleration of the vehicle is calculated based on the preprocessed acceleration data. The calculation expression is as follows: a filtered (t)=Filter(a raw (t)) In the formula, a raw (t) is the original lateral acceleration, which represents the lateral acceleration value of the vehicle at time point t, Filter is the filter, a filtered (t) is the filtered lateral acceleration; By analyzing the sudden changes in lateral acceleration during vehicle driving, the instantaneous rate of change of lateral acceleration is extracted. By calculating the instantaneous rate of change of acceleration, the moment of sharp change is identified, and the change amplitude is further analyzed. The calculation expression is as follows: Δa(t)=|a filtered (t)-a filtered (t-Δt)| In the formula, a filtered (t-Δt) is the lateral acceleration value at the previous moment with a time interval of Δt from the time point t, and Δa(t) is the instantaneous rate of change of lateral acceleration; According to the fluctuation characteristics of the lateral acceleration during vehicle driving, a dynamic model is established to quantify the lateral stability of the vehicle. The vehicle speed and road friction factors are introduced to construct an acceleration model to describe the dynamic changes of the vehicle's lateral stability. The calculation expression is as follows: Where v(t) is the current speed of the vehicle, i.e. the speed at time t, μ road is the road friction factor, which refers to the friction coefficient between the vehicle and the current road surface, m is the vehicle weight, a dynamic (t) is the dynamic lateral acceleration of the vehicle; Where RAI is the vehicle rolling acceleration index, ω is the speed weight factor, is the road friction factor weight, θ is the exponential adjustment factor, s is the total number of sampling points, and j is the sampling point.

8. An intelligent vehicle maintenance and management system based on big data analysis, used to implement the intelligent vehicle maintenance and management method based on big data analysis as described in any one of claims 1 to 7 above, characterized in that: It includes data acquisition module, data sorting and preprocessing module, feature analysis and model input module, driving status evaluation and classification module, conventional road maintenance management module and rough road maintenance management module: The data acquisition module collects various types of vehicle driving data in real time through the data acquisition device during the vehicle driving process, providing a reliable basis for subsequent analysis and decision-making; The data sorting and preprocessing module sorts and preprocesses the collected real-time driving data, establishes a data analysis set, and extracts key features reflecting the vehicle driving status from the analysis set; The feature analysis and model input module analyzes the extracted key features within the detection window, inputs the analyzed key features into the pre-trained machine learning model, and intelligently evaluates the current driving status of the vehicle through the machine learning model; The driving status assessment and classification module classifies the vehicle's driving status into two categories: regular road driving and rough road driving based on the evaluation results of the machine learning model; The regular road maintenance management module continues to evaluate and maintain regular road driving according to the preset fixed inspection cycle to ensure the rational use of resources and cost control; The rough road maintenance management module increases the inspection frequency and implements more stringent maintenance measures based on the evaluation results of the machine learning model for driving on rough roads, as vehicles face a higher risk of failure under harsh road conditions.

Images