Automobile lamp fault intelligent operation and maintenance system and prediction method

Abstract

The invention relates to the technical field of automobile lamp fault prediction, and discloses an automobile lamp fault intelligent operation and maintenance system and a prediction method, and the system comprises a first data collection module which collects a first working parameter of an automobile lamp in a first preset time period T1; the data processing module is used for preprocessing to obtain a feature matrix; the second data acquisition module is used for acquiring second working parameters of the M vehicle lamps after a first preset time period T1; the fault scoring module is used for predicting and obtaining a fault score of the vehicle lamp through the trained prediction model; according to the method, the working parameters of the vehicle lamp in different working states are obtained through a simulation experiment, and correlation influence analysis on the working parameters in the time dimension and the space dimension is carried out on the working parameters through the prediction model, so that the nonlinear mapping relation between the working parameters of the vehicle lamp in the historical time and the working parameters of the vehicle lamp in the future time is established; compared with an existing linear model, the method can improve the prediction precision of the vehicle lamp fault.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile lamp fault prediction, and more specifically, to an automobile lamp fault intelligent operation and maintenance system and a prediction method. Background Art

[0002] With the rapid development of the automobile industry and the continuous advancement of intelligent technology, the safety and reliability of automobiles have gradually become the focus of consumer attention. As an important component of driving safety, car lights not only provide lighting at night and in bad weather conditions, but also bear important safety warning functions. The failure of car lights will not only bring driving safety risks, but also increase maintenance costs, seriously affecting the driver's driving experience. Traditional car light maintenance methods often rely on the driver's subjective judgment or regular inspections, and their maintenance efficiency is low, making it difficult to detect potential faults in a timely manner.

[0003] The existing method is to collect the humidity in the headlight and the current value of the headlight within a preset time period and at preset time intervals, and to calculate the mean value of the humidity, the mean value of the current, the mean square error, the maximum value and the minimum value, and then respectively predict the humidity and current parameters of the headlight in the future time period through linear fitting of the corresponding N-order polynomials, and to determine that when it reaches the corresponding fault warning threshold, a headlight fault warning is issued.

[0004] However, the working state of car lights is affected by many factors, such as temperature, humidity, current, etc. These factors often have complex interactions, and simple linear models may be difficult to fully capture the combined impact of these factors. In addition, the working parameters of car lights will also change significantly in different driving environments. For example, in a high temperature environment, the temperature and current values ​​of the car lights may increase. Therefore, static data models are difficult to adapt to dynamic environmental changes, resulting in low prediction accuracy of car light failures. Summary of the invention

[0005] The present invention provides an intelligent operation and maintenance system and a prediction method for automobile lamp failures, which solve the technical problems in the above-mentioned background technology.

[0006] The present invention provides an intelligent operation and maintenance system for automobile lamp failure, comprising: A first data acquisition module, which is used to collect the driving parameters of the vehicle and the first working parameters of the M headlights through a simulation experiment at a preset time interval t within a first preset time period T1; The driving parameters of the car include: ambient temperature, rainfall and vehicle speed; The first working parameters of the headlights include: temperature, humidity, current value, voltage value, light intensity and working time; A data processing module, which is used to pre-process the driving parameters of the vehicle and the first working parameters of the M headlights to obtain a feature matrix; The feature matrix includes M rows and N columns, and the element values ​​of the feature matrix are represented by a combination vector, and the combination vector of the mth row and nth column is obtained by preprocessing the first working parameter of the mth headlight and the driving parameter of the vehicle collected at the nth time point, wherein 1≤m≤M, 1≤n≤N, and N=T1 / t; A second data acquisition module, which is used to collect second operating parameters of the M vehicle lamps through a simulation experiment in a second preset time period T2 after the first preset time period T1; The second working parameters of the headlight include: temperature, humidity, current value and voltage value; A prediction model training module, which is used to train a prediction model according to the feature matrix and the second working parameters of the M vehicle lamps in a second preset time period T2; The input of the prediction model is the feature matrix, and the output is the second operating parameters of the M headlights in the second preset time period T2; A fault scoring module, which is used to predict the second operating parameters of the M headlights in the second preset time period T2 by using the trained prediction model, and calculate the fault scores of the M headlights; The fault warning module is used to generate corresponding fault warning information according to the fault scores of the M headlights and send them to the user of the car.

[0007] Furthermore, the first preset time period T1, the preset time interval t, the number of vehicle lights M, and the second preset time period T2 are all custom parameters.

[0008] Furthermore, preprocessing the first working parameter of the vehicle lamp and the driving parameter of the vehicle to obtain a combination vector includes the following steps: Step S201, taking the average value of the data of two adjacent time points of the missing values ​​to fill the missing values ​​in the first working parameter of the headlight and the driving parameter of the car; Step S202, calculating and obtaining characteristic parameters at each time point; The characteristic parameters include: ambient temperature change rate, rainfall change rate, vehicle speed change rate, temperature change rate, humidity change rate, current value change rate, voltage value change rate, light intensity change rate and working time ratio, where the working time ratio is equal to the working time divided by the total working time; Step S203, splicing the characteristic parameter at each time point with the first working parameter of the headlight and the driving parameter of the vehicle at the corresponding time point to obtain a splicing vector; Step S204, normalizing the concatenated vectors by using a Min-Max normalization method to obtain a combined vector.

[0009] Furthermore, the prediction model includes M hidden layers, each hidden layer includes N hidden units, the nth hidden unit of the mth hidden layer inputs the combination vector of the mth row and nth column of the feature matrix, and outputs an update vector; The update vector output by the Nth hidden unit of each hidden layer is input to the first classifier, the second classifier, the third classifier and the fourth classifier respectively, where the classification spaces of the first classifier, the second classifier, the third classifier and the fourth classifier respectively represent the temperature, humidity, current value and voltage value of the car light in the second preset time period T2.

[0010] Furthermore, the calculation formula of the nth hidden unit of the mth hidden layer of the prediction model includes: definition: Represents the combined vector of the mth row and nth column of the feature matrix input to the nth hidden unit of the mth hidden layer. concat represents the concatenation operation. represents point-wise multiplication, sigmoid represents the sigmoid activation function, Swish represents the Swish activation function, and tanh represents the hyperbolic tangent activation function; The forget gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in , and Respectively represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the forget gate of the nth hidden unit of the mth hidden layer; Input gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in , and Respectively represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the input gate of the nth hidden unit of the mth hidden layer; Candidate cell state value of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in and They respectively represent the weight parameter and bias parameter corresponding to the candidate cell state value of the nth hidden unit of the mth hidden layer; The cell state update value of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; Where M represents the number of hidden layers, N represents the number of hidden units in each hidden layer, represents the cell state update value of the nth hidden unit in the m-1th hidden layer, Assign a value of 0, represents the cell state update value of the n-1th hidden unit in the mth hidden layer, Assign a value of 0; Output gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in , and Respectively represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the output gate of the nth hidden unit of the mth hidden layer; The update vector for the output of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; The median vector of the n-1th hidden unit The calculation formula is as follows: ; Where FFNN represents a feed-forward neural network, represents the update vector output by the nth hidden unit in the m-1th hidden layer, Assign a value of 0, represents the update vector for the output of the nth hidden unit in the mth hidden layer, Assign a value of 0.

[0011] Furthermore, the steps in the first data acquisition module 101 to the data processing module 102 are repeatedly executed until U feature matrices are obtained as sample data of training samples for training the prediction model, and the steps in the second data acquisition module 103 are correspondingly executed until U second operating parameters of the M headlights in the second preset time period T2 are obtained as sample labels of training samples for training the prediction model.

[0012] Furthermore, the fault score of the mth headlight The calculation formula is as follows: ; Where 1≤m≤M, , , and Respectively represent the temperature, humidity, current value and voltage value of the mth headlight, , , and represent the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient respectively, , , and are custom parameters, and , , and The sum of is 1.

[0013] Further, when it is determined that the fault score of the vehicle light is greater than or equal to a first threshold or less than a second threshold, fault warning information of the corresponding vehicle light is generated, wherein the first threshold and the second threshold are both custom parameters.

[0014] Furthermore, after the fault warning SMS is generated, the type and number of the faulty headlights are sent to the car purchase store, which then finds a repair shop closest to the user for repair.

[0015] The present invention provides a method for intelligent operation and maintenance prediction of automobile lamp faults, comprising the following steps: Step S301, collecting driving parameters of the vehicle and first working parameters of M headlights through a simulation experiment at a preset time interval t within a first preset time period T1; Step S302, preprocessing the driving parameters of the vehicle and the first working parameters of the M headlights to obtain a feature matrix; Step S303, after the first preset time period T1, in the second preset time period T2, collecting the second working parameters of the M vehicle lamps through a simulation experiment; Step S304, training a prediction model according to the feature matrix and the second working parameters of the M vehicle lamps in the second preset time period T2; Step S305, predicting the second operating parameters of the M headlights in the second preset time period T2 by using the trained prediction model, and calculating the fault scores of the M headlights; Step S306: Generate corresponding fault warning information according to the fault scores of the M headlights and send them to the user of the car.

[0016] The beneficial effects of the present invention are as follows: the present invention obtains the working parameters of the headlights under different working conditions through simulation experiments, and analyzes the correlation influence in the time dimension and the space dimension through the prediction model, so as to establish a nonlinear mapping relationship between the working parameters of the headlights in the historical time and the working parameters of the headlights in the future time. Compared with the existing linear model, the prediction accuracy of the headlight failure can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of an intelligent operation and maintenance system for automobile headlight failure of the present invention; Figure 2 is a flow chart of obtaining a combination vector by preprocessing of the present invention; Figure 3 It is a flow chart of an intelligent operation and maintenance prediction method for automobile headlight failure of the present invention.

[0018] In the figure: a first data acquisition module 101, a data processing module 102, a second data acquisition module 103, a prediction model training module 104, a fault scoring module 105, and a fault warning module 106. DETAILED DESCRIPTION

[0019] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0020] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0021] like Figure 1 to Figure 3 As shown, an intelligent operation and maintenance system for automobile lamp failure includes: A first data acquisition module 101 is used to collect the driving parameters of the vehicle and the first working parameters of the M headlights through a simulation experiment at a preset time interval t within a first preset time period T1; The driving parameters of the car include: ambient temperature, rainfall and vehicle speed; The first working parameters of the headlights include: temperature, humidity, current value, voltage value, light intensity and working time; A data processing module 102, which is used to pre-process the driving parameters of the vehicle and the first working parameters of the M headlights to obtain a feature matrix; The feature matrix includes M rows and N columns, and the element values ​​of the feature matrix are represented by a combination vector, and the combination vector of the mth row and nth column is obtained by preprocessing the first working parameter of the mth headlight and the driving parameter of the vehicle collected at the nth time point, wherein 1≤m≤M, 1≤n≤N, and N=T1 / t; The second data collection module 103 is used to collect the second working parameters of the M vehicle lamps through a simulation experiment in a second preset time period T2 after the first preset time period T1; The second working parameters of the headlight include: temperature, humidity, current value and voltage value; A prediction model training module 104, which is used to train a prediction model according to the feature matrix and the second working parameters of the M vehicle lamps in the second preset time period T2; The input of the prediction model is the feature matrix, and the output is the second operating parameters of the M headlights in the second preset time period T2; A fault scoring module 105, which is used to predict the second operating parameters of the M headlights in the second preset time period T2 by using the trained prediction model, and calculate the fault scores of the M headlights; The fault warning module 106 is used to generate corresponding fault warning information according to the fault scores of the M headlights and send them to the user of the car.

[0022] It should be noted that the simulation experiment can directly simulate different rainfall amounts through the sprinkler system in the experimental environment, simulate different ambient temperatures through the heating equipment, and simulate the vehicle driving process through the fan. Although this simulation experiment method is flexible and changeable, the actual operating cost is relatively high. Therefore, various types of sensors can be directly installed on multiple vehicles to directly collect the driving parameters of the car and the first working parameters and the second working parameters of the headlights. In addition, different vehicle models or headlight types will also affect the prediction of headlight failures. The data collected from simulation experiments on different vehicle models or headlight types can be used to train the corresponding prediction model, thereby eliminating the differences between different vehicle models or headlight types.

[0023] In one embodiment of the present invention, the first preset time period T1, the preset time interval t, the number of lights M, and the second preset time period T2 are all custom parameters. Preferably, T1 is set to 5 minutes, t is set to 30 seconds, then N=T1 / t=10, T2 is set to 5 minutes, and the sizes of T1, t and T2 can be scaled to achieve the function of short-term or long-term prediction. M is set according to the actual number of lights. The default value of M is 4, i.e., left and right headlights, left and right rear taillights. The number of lights can also be refined, for example, including: daytime running lights, fog lights, reversing lights, etc., which can be set according to the actual number of light modules.

[0024] In one embodiment of the present invention, a temperature sensor is installed at the front end of the vehicle or the air intake grille to collect the ambient temperature, a rain sensor is installed on the windshield to collect the rainfall, a speed sensor is installed near the wheel to collect the vehicle speed, and a temperature sensor, a humidity sensor, a current sensor, a voltage sensor, and a light intensity sensor are installed inside the headlight to respectively collect the temperature, humidity, current value, voltage value and light intensity of the headlight, and the working time is obtained by collecting the data through the ECU (electronic control unit). The above data is converted into electrical signals, transmitted to the ECU through the vehicle bus, and then uploaded to the cloud server through the Internet of Vehicles module (such as 4G / 5G module, Wi-Fi module, etc.), and a prediction model is deployed on the cloud server to analyze the above data to generate fault warning information and return it to the user.

[0025] In one embodiment of the present invention, Figure 2 As shown, preprocessing the first working parameter of the headlight and the driving parameter of the vehicle to obtain a combined vector includes the following steps: Step S201, taking the average value of the data of two adjacent time points of the missing values ​​to fill the missing values ​​in the first working parameter of the headlight and the driving parameter of the car; Step S202, calculating and obtaining characteristic parameters at each time point; The characteristic parameters include: ambient temperature change rate, rainfall change rate, vehicle speed change rate, temperature change rate, humidity change rate, current value change rate, voltage value change rate, light intensity change rate and working time ratio, where the working time ratio is equal to the working time divided by the total working time; Step S203, splicing the characteristic parameter at each time point with the first working parameter of the headlight and the driving parameter of the vehicle at the corresponding time point to obtain a splicing vector; Step S204, normalizing the concatenated vectors by using a Min-Max normalization method to obtain a combined vector.

[0026] It should be noted that the ambient temperature change rate, rainfall change rate, vehicle speed change rate, temperature change rate, humidity change rate, current value change rate, voltage value change rate, light intensity change rate and working time ratio at the first time point are all assigned a value of 0, and the size of the combination vector is 1×18, which respectively represent the normalized ambient temperature, rainfall, vehicle speed, temperature, humidity, current value, voltage value, light intensity, working time, ambient temperature change rate, rainfall change rate, vehicle speed change rate, temperature change rate, humidity change rate, current value change rate, voltage value change rate, light intensity change rate and working time ratio, where the change rate at the current time point is obtained by subtracting the value at the previous time point from the value at the current time point and dividing it by the time interval between the two time points. For example, the ambient temperature at the previous time point is 29°C, the ambient temperature at the current time point is 28°C, and the time interval between the two time points is 30 seconds. Then, the ambient temperature change rate at the current time point = (28-29) / 30 = -0.033, which will not be elaborated here.

[0027] In one embodiment of the present invention, the prediction model includes M hidden layers, each hidden layer includes N hidden units, the nth hidden unit of the mth hidden layer inputs the combination vector of the mth row and nth column of the feature matrix, and outputs an update vector; The update vector output by the Nth hidden unit of each hidden layer is input to the first classifier, the second classifier, the third classifier and the fourth classifier respectively, where the classification spaces of the first classifier, the second classifier, the third classifier and the fourth classifier respectively represent the temperature, humidity, current value and voltage value of the car light in the second preset time period T2.

[0028] In one embodiment of the present invention, the calculation formula of the nth hidden unit of the mth hidden layer of the prediction model includes: definition: Represents the combined vector of the mth row and nth column of the feature matrix input to the nth hidden unit of the mth hidden layer. concat represents the concatenation operation. represents point-wise multiplication, sigmoid represents the sigmoid activation function, Swish represents the Swish activation function, and tanh represents the hyperbolic tangent activation function; The forget gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the forget gate of the nth hidden unit of the mth hidden layer, respectively. For example The size is 1×18, then It can be designed as a matrix of size 18×8, assuming is a 1×16 vector, then It needs to be designed as a 16×8 matrix, then is a vector of size 1×8; Input gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the input gate of the nth hidden unit of the mth hidden layer. Similarly, It can be designed as a matrix of size 18×8, assuming is a 1×16 vector, then If a 16×8 matrix is ​​required, Also a vector of size 1×8; Candidate cell state value of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in and They represent the weight parameters and bias parameters corresponding to the candidate cell state values ​​of the nth hidden unit in the mth hidden layer. According to the above content, and After concatenation, we get a vector of size 1×24. It can be designed as a 24×8 matrix. Also a vector of size 1×8; The cell state update value of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; Where M represents the number of hidden layers, N represents the number of hidden units in each hidden layer, represents the cell state update value of the nth hidden unit in the m-1th hidden layer, Assign a value of 0, represents the cell state update value of the n-1th hidden unit in the mth hidden layer, Assign a value of 0. According to the above, , and are all 1×8 vectors, then and The result of the addition must also be a vector of size 1×8, then is a vector of size 1×8; Output gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; in , and They respectively represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the output gate of the nth hidden unit of the mth hidden layer. Similarly, according to the above content, It can be designed into a matrix of size 18×8. It can be designed as a 16×8 matrix. Also a vector of size 1×8; The update vector for the output of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ; According to the above, Also a vector of size 1×8; The median vector of the n-1th hidden unit The calculation formula is as follows: ; Where FFNN represents a feed-forward neural network, represents the update vector output by the nth hidden unit in the m-1th hidden layer, Assign a value of 0, represents the update vector for the output of the nth hidden unit in the mth hidden layer, Assign a value of 0. According to the above content, and The concatenation is performed to obtain a 1×16 vector, which is calculated using a feedforward neural network model. and The attention weight between is represented by a real value, then the final is a 1×16 vector, which meets the above assumptions.

[0029] It should be noted that the update vector output by the Nth hidden unit of each hidden layer can be directly input into the fifth classifier. The classification space of the fifth classifier represents whether the headlights fail in the second preset time period T2. The second working parameters of the headlights after the second preset time period T2 can be used to pre-train the prediction model, thereby accelerating the convergence speed of the prediction model. In addition, the prediction model provided by the present invention can simultaneously analyze the correlation between multiple headlights in the time dimension and the space dimension. Compared with the linear model of the prior art, the prediction model establishes a nonlinear mapping relationship between the first working parameters of the headlights in the time and space dimensions and the second working parameters of the headlights in the future time period, thereby improving the prediction accuracy of headlight failures.

[0030] In one embodiment of the present invention, the steps in the first data acquisition module 101 to the data processing module 102 are repeatedly executed until U feature matrices are obtained as sample data of training samples for training the prediction model, and the steps in the second data acquisition module 103 are correspondingly executed until U second operating parameters of the M headlights in the second preset time period T2 are obtained as sample labels of training samples for training the prediction model.

[0031] It should be noted that the weight parameters and bias parameters in the prediction model are all learnable parameters. The mean square error is specified as the loss function, that is, the difference between the value output by the prediction model in the training phase and the sample label of the training sample. The weight and bias parameters of the prediction model are updated by the chain rule combined with the gradient descent algorithm. The gradient descent algorithm can be AdaGrad (adaptive learning rate), RMSProp (an improvement on AdaGrad, using exponentially weighted moving average to smooth the gradient square and avoid premature decrease in the learning rate), etc., which will not be elaborated here.

[0032] In one embodiment of the present invention, the fault score of the mth headlight The calculation formula is as follows: ; Where 1≤m≤M, , , and Respectively represent the temperature, humidity, current value and voltage value of the mth headlight, , , and represent the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient respectively, , , and are custom parameters, and , , and The sum of the values ​​is 1, preferably, Set to 0.1, Set to 0.3, Set to 0.3, Set to 0.3.

[0033] In one embodiment of the present invention, when it is determined that the fault score of a headlight is greater than or equal to a first threshold or less than a second threshold, fault warning information of the corresponding headlight is generated, wherein the first threshold and the second threshold are both custom parameters, that is, the first threshold and the second threshold respectively represent the upper limit and lower limit of the fault score of the headlight when it fails, and the fault warning information is sent to the user via SMS or sends a message to the user's car APP.

[0034] It should be noted that it is also possible to not calculate the fault score of the headlight, and directly use the second working parameter of the headlight to generate fault warning information, determine the upper and lower limits of the second working parameter of the headlight when it fails, and generate fault warning information when the second working parameter of the headlight obtained by the prediction model exceeds the upper or lower limit of the second working parameter.

[0035] In one embodiment of the present invention, after the fault warning SMS is generated, the type and number of faulty headlights are sent to the car purchase store, which then searches for a repair shop closest to the user for repairs.

[0036] In one embodiment of the present invention, Figure 3 As shown, the present invention provides a method for intelligent operation and maintenance prediction of automobile lamp failure, comprising the following steps: Step S301, collecting driving parameters of the vehicle and first working parameters of M headlights through a simulation experiment at a preset time interval t within a first preset time period T1; Step S302, preprocessing the driving parameters of the vehicle and the first working parameters of the M headlights to obtain a feature matrix; Step S303, after the first preset time period T1, in the second preset time period T2, collecting the second working parameters of the M vehicle lamps through a simulation experiment; Step S304, training a prediction model according to the feature matrix and the second working parameters of the M vehicle lamps in the second preset time period T2; Step S305, predicting the second operating parameters of the M headlights in the second preset time period T2 by using the trained prediction model, and calculating the fault scores of the M headlights; Step S306: Generate corresponding fault warning information according to the fault scores of the M headlights and send them to the user of the car.

[0037] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.

Claims

1. An intelligent operation and maintenance system for automobile lamp failure, characterized in that: include: A first data acquisition module, which is used to collect the driving parameters of the vehicle and the first working parameters of the M headlights through a simulation experiment at a preset time interval t within a first preset time period T1; The driving parameters of the car include: ambient temperature, rainfall and vehicle speed; The first working parameters of the headlights include: temperature, humidity, current value, voltage value, light intensity and working time; A data processing module, which is used to pre-process the driving parameters of the vehicle and the first working parameters of the M headlights to obtain a feature matrix; The feature matrix includes M rows and N columns, and the element values ​​of the feature matrix are represented by a combination vector, and the combination vector of the mth row and nth column is obtained by preprocessing the first working parameter of the mth headlight and the driving parameter of the vehicle collected at the nth time point, wherein 1≤m≤M, 1≤n≤N, and N=T1 / t; A second data acquisition module, which is used to collect second operating parameters of the M vehicle lamps through a simulation experiment in a second preset time period T2 after the first preset time period T1; The second working parameters of the headlight include: temperature, humidity, current value and voltage value; A prediction model training module, which is used to train a prediction model according to the feature matrix and the second working parameters of the M vehicle lamps in a second preset time period T2; The input of the prediction model is the feature matrix, and the output is the second operating parameters of the M headlights in the second preset time period T2; A fault scoring module, which is used to predict the second operating parameters of the M headlights in the second preset time period T2 by using the trained prediction model, and calculate the fault scores of the M headlights; The fault warning module is used to generate corresponding fault warning information according to the fault scores of the M headlights and send them to the user of the car.

2. The intelligent operation and maintenance system for automobile headlight failure according to claim 1 is characterized in that: The first preset time period T1, the preset time interval t, the number of vehicle lights M, and the second preset time period T2 are all custom parameters.

3. The intelligent operation and maintenance system for automobile headlight failure according to claim 1 is characterized in that: Preprocessing the first working parameter of the headlight and the driving parameter of the vehicle to obtain a combined vector includes the following steps: Step S201, taking the average value of the data of two adjacent time points of the missing values ​​to fill the missing values ​​in the first working parameter of the headlight and the driving parameter of the car; Step S202, calculating and obtaining characteristic parameters at each time point; The characteristic parameters include: ambient temperature change rate, rainfall change rate, vehicle speed change rate, temperature change rate, humidity change rate, current value change rate, voltage value change rate, light intensity change rate and working time ratio, where the working time ratio is equal to the working time divided by the total working time; Step S203, splicing the characteristic parameter at each time point with the first working parameter of the headlight and the driving parameter of the vehicle at the corresponding time point to obtain a splicing vector; Step S204, normalizing the concatenated vectors by using a Min-Max normalization method to obtain a combined vector.

4. The intelligent operation and maintenance system for automobile headlight failure according to claim 1 is characterized in that: The prediction model includes M hidden layers, each hidden layer includes N hidden units, the nth hidden unit of the mth hidden layer inputs the combination vector of the mth row and nth column of the feature matrix, and outputs an update vector; The update vector output by the Nth hidden unit of each hidden layer is input to the first classifier, the second classifier, the third classifier and the fourth classifier respectively, where the classification spaces of the first classifier, the second classifier, the third classifier and the fourth classifier respectively represent the temperature, humidity, current value and voltage value of the car light in the second preset time period T2.

5. The intelligent operation and maintenance system for automobile headlight failure according to claim 4 is characterized in that: The calculation formula for the nth hidden unit of the mth hidden layer of the prediction model includes: definition: Represents the combined vector of the mth row and nth column of the feature matrix input to the nth hidden unit of the mth hidden layer. concat represents the concatenation operation. represents point-wise multiplication, sigmoid represents the sigmoid activation function, Swish represents the Swish activation function, and tanh represents the hyperbolic tangent activation function; The forget gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ； in , and Respectively represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the forget gate of the nth hidden unit of the mth hidden layer; Input gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ； in , and Respectively represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the input gate of the nth hidden unit of the mth hidden layer; Candidate cell state value of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ； in and They respectively represent the weight parameter and bias parameter corresponding to the candidate cell state value of the nth hidden unit of the mth hidden layer; The cell state update value of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ； Where M represents the number of hidden layers, N represents the number of hidden units in each hidden layer, represents the cell state update value of the nth hidden unit in the m-1th hidden layer, Assign a value of 0, represents the cell state update value of the n-1th hidden unit in the mth hidden layer, Assign a value of 0; Output gate of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ； in , and Respectively represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the output gate of the nth hidden unit of the mth hidden layer; The update vector for the output of the nth hidden unit in the mth hidden layer The calculation formula is as follows: ； The median vector of the n-1th hidden unit The calculation formula is as follows: ； Where FFNN represents a feed-forward neural network, represents the update vector of the output of the nth hidden unit in the m-1th hidden layer, Assign a value of 0, represents the update vector for the output of the nth hidden unit in the mth hidden layer, Assign a value of 0.

6. The intelligent operation and maintenance system for automobile headlight failure according to claim 1 is characterized in that: Repeat the steps from the first data acquisition module 101 to the data processing module 102 until U feature matrices are obtained as sample data of training samples for training the prediction model, and correspondingly execute the steps in the second data acquisition module 103 until U second operating parameters of the M headlights in the second preset time period T2 are obtained as sample labels of training samples for training the prediction model.

7. The intelligent operation and maintenance system for automobile headlight failure according to claim 1 is characterized in that: Fault score of the mth headlight The calculation formula is as follows: ； Where 1≤m≤M, , , and Respectively represent the temperature, humidity, current value and voltage value of the mth headlight, , , and represent the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient respectively, , , and are custom parameters, and , , and The sum of is 1.

8. The intelligent operation and maintenance system for automobile headlight failure according to claim 1 is characterized in that: When it is determined that the fault score of the vehicle light is greater than or equal to a first threshold or less than a second threshold, fault warning information of the corresponding vehicle light is generated, wherein the first threshold and the second threshold are both custom parameters.

9. The intelligent operation and maintenance system for automobile headlight failure according to claim 1 is characterized in that: After the fault warning SMS is generated, the type and number of the faulty headlights are sent to the car purchase store, which then finds the nearest repair shop to the user for repair.

10. An intelligent operation and maintenance prediction method for automobile lamp failure, characterized in that: Executing an intelligent operation and maintenance system for automobile headlight failure as claimed in any one of claims 1 to 9 comprises the following steps: Step S301, collecting driving parameters of the vehicle and first working parameters of M headlights through a simulation experiment at a preset time interval t within a first preset time period T1; Step S302, preprocessing the driving parameters of the vehicle and the first working parameters of the M headlights to obtain a feature matrix; Step S303, after the first preset time period T1, in the second preset time period T2, collecting the second working parameters of the M vehicle lamps through a simulation experiment; Step S304, training a prediction model according to the feature matrix and the second working parameters of the M vehicle lamps in the second preset time period T2; Step S305, predicting the second operating parameters of the M headlights in the second preset time period T2 by using the trained prediction model, and calculating the fault scores of the M headlights; Step S306: Generate corresponding fault warning information according to the fault scores of the M headlights and send them to the user of the car.

Images