Vehicle lamp light-emitting module state real-time monitoring method based on edge calculation

Abstract

The invention discloses a vehicle lamp light-emitting module state real-time monitoring method based on edge calculation, and relates to the technical field of intelligent vehicle electronics, and the method comprises the steps: collecting vehicle lamp light-emitting module operation data according to an operation environment object, carrying out the filtering and alignment preprocessing, and generating a preprocessing data window; based on the diagnosis input object, a diagnosis model is selected from the edge calculation nodes of the vehicle lamp light-emitting module and reasoning is carried out, and an advanced diagnosis result object is generated; according to the state category, the health degree, the remaining life interval and the alarm level in the vehicle lamp light-emitting module state object, generating a vehicle lamp light-emitting module state real-time monitoring result; according to the method, efficient organization and configuration management of edge computing node resources are realized by constructing a structured operation environment object, a mixed diagnosis strategy of data driving and rule constraint is formed, the rationality of health degree evaluation and the credibility of residual life prediction are improved, and the reliability of the prediction is improved. And predictive maintenance and functional safety guarantee of the intelligent automobile on key lighting components are effectively supported.

Description

Technical Field

[0001] This invention relates to the field of intelligent automotive electronics technology, and in particular to a method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing. Background Technology

[0002] Against the backdrop of the rapid development of intelligent vehicles and vehicle networking technology, the reliability and intelligence level of the vehicle lighting system, as an important component of vehicle active safety and human-machine interaction, are receiving increasing attention. With the popularization of LED light sources, adaptive lighting control and vehicle communication technology, the vehicle lighting module has evolved from a traditional lighting device into an intelligent terminal integrating perception, control and communication. Remote status monitoring solutions based on vehicle networks have gradually become a research hotspot. Some solutions adopt a centralized architecture, which uploads the vehicle lighting operation data to the central computing unit (such as a domain controller) via CAN bus or Ethernet, and performs unified processing and fault diagnosis by the cloud or vehicle main control platform.

[0003] Centralized monitoring architectures have inherent limitations in terms of real-time performance and local response capabilities. Due to the transient and highly dynamic nature of the operating status of vehicle headlight modules (such as luminous flux attenuation, abnormal drive current, and thermal management failure), if remote diagnostic paths are relied upon, factors such as data transmission delays, network congestion, or fluctuations in the load of the master control node may prevent critical fault information from being identified and responded to in a timely manner, thereby affecting driving safety and maintenance efficiency. Frequent uploading of raw sensor data will also increase the load and energy consumption of the vehicle network, which is not conducive to the overall energy efficiency optimization of the system. In contrast, the edge computing paradigm emphasizes completing the perception, processing, and decision-making closed loop near the data source, which can significantly reduce communication overhead and improve response speed. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a real-time monitoring method for the status of vehicle headlight modules based on edge computing to solve the problem of delayed response in vehicle headlight status diagnosis caused by the reliance on remote processing in existing centralized monitoring architectures.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a real-time monitoring method for the status of a vehicle headlight module based on edge computing. The method includes: after the edge computing node of the vehicle headlight module is powered on, loading configuration parameters and initializing data structures to generate a runtime environment object; collecting runtime data of the vehicle headlight module based on the runtime environment object, and performing filtering and alignment preprocessing to generate a preprocessed data window; calculating diagnostic indicators based on the preprocessed data window and comparing the diagnostic indicators with diagnostic thresholds for the vehicle headlight module in the runtime environment object to generate a diagnostic input object; selecting a diagnostic model in the edge computing node of the vehicle headlight module based on the diagnostic input object and performing inference to generate a high-level diagnostic result object; making a comprehensive judgment based on the high-level diagnostic result object and referring to the indicators in the diagnostic input object to generate a vehicle headlight module status object; and generating a real-time monitoring result for the vehicle headlight module status based on the status category, health level, remaining lifespan range, and alarm level in the vehicle headlight module status object.

[0008] As a preferred embodiment of the edge computing-based real-time monitoring method for vehicle headlight illumination module status according to the present invention, the specific steps for generating the runtime environment object are as follows:

[0009] After the edge computing node of the vehicle headlight illumination module is powered on, the clock, watchdog and memory control are started, and storage space is reserved in the random access memory;

[0010] Read the original content of the monitoring configuration parameters and the diagnostic threshold parameters of the headlight illumination module, and perform format parsing and verification to obtain the monitoring configuration parameters and the diagnostic threshold of the headlight illumination module;

[0011] Within the reserved storage space, establish queue data structures for the task scheduling queue and the exception event queue, and set initial values ​​for read / write pointers and counters; allocate contiguous storage space for the historical data circular buffer according to the buffer size in the monitoring configuration parameters, and set the starting address and index of the historical data circular buffer;

[0012] The monitoring configuration parameters, vehicle headlight illumination module diagnostic thresholds, task scheduling queue status information, abnormal event queue status information, and historical data ring buffer addresses and indexes are encapsulated into runtime environment objects.

[0013] As a preferred embodiment of the edge computing-based real-time monitoring method for vehicle headlight illumination module status according to the present invention, the specific steps for generating the preprocessed data window are as follows:

[0014] Based on the sampling period and sensor channel configuration in the runtime environment object, start the sampling timing task in the edge computing node of the vehicle light emission module;

[0015] Within each sampling period, the vehicle headlight module operation data is read sequentially from the sensors specified by the operating environment object, and a timestamp in a uniform format is added to each group of vehicle headlight module operation data. The vehicle headlight module operation data with the timestamp is then written to the buffer area.

[0016] Digital filtering is performed on the operating data of the vehicle headlight illumination module in the same sensing channel within the buffer area, and interpolation is performed on the operating data of the vehicle headlight illumination module in different sensing channels using timestamps to generate a vehicle headlight illumination module operating dataset.

[0017] Based on the vehicle headlight module operation dataset, a continuous segment of vehicle headlight module operation data is extracted from the cache according to the time window length in the operation environment object and spliced ​​together to generate a preprocessed data window.

[0018] As a preferred embodiment of the edge computing-based real-time monitoring method for the status of vehicle headlight modules described in this invention, the step of calculating diagnostic indicators based on a preprocessed data window is as follows:

[0019] Based on the operating data of the vehicle lamp light-emitting module in the current sensing channel of the preprocessed data window, the current standard deviation is calculated within the time range covered by the preprocessed data window, and the current fluctuation amplitude is obtained from the current standard deviation.

[0020] Based on the operating data of the vehicle headlight emitting module in the temperature sensing channel of the preprocessed data window, the temperature rise slope is calculated and the temperature peak is extracted by linear fitting of the temperature difference and time difference between adjacent sampling points.

[0021] Based on the operating data of the vehicle lamp light-emitting module corresponding to the optical sensing channel in the preprocessed data window, the luminous flux change rate is calculated by the ratio of the current luminous flux to the reference luminous flux, and the mean and standard deviation of the luminous flux in the preprocessed data window are calculated.

[0022] Based on the operating data of the vehicle lamp light-emitting module corresponding to the voltage sensing channel in the preprocessed data window, the voltage ripple amplitude and voltage ripple coefficient are calculated by separating the DC component and the AC component. The average current, current fluctuation amplitude, temperature rise slope, temperature peak, luminous flux change rate, luminous flux standard deviation, voltage ripple amplitude and voltage ripple coefficient are combined to form diagnostic indicators.

[0023] As a preferred embodiment of the edge computing-based real-time monitoring method for the status of vehicle headlight emission modules described in this invention, the specific steps for generating the diagnostic input object are as follows:

[0024] Read the diagnostic threshold of the vehicle headlight module corresponding to each diagnostic indicator in the diagnostic indicators from the runtime environment object, compare each diagnostic indicator in the diagnostic indicators with the corresponding diagnostic threshold of the vehicle headlight module, and obtain a set of comparison results.

[0025] The number and type of out-of-limit diagnostic indicators are counted based on the comparison result set, and the vehicle headlight illumination module operation data, diagnostic indicators and comparison result set in the preprocessed data window are encapsulated into a diagnostic input object.

[0026] As a preferred embodiment of the real-time monitoring method for the status of vehicle headlight emitting modules based on edge computing described in this invention, the diagnostic model is a neural network model, which simultaneously outputs the status category of the emitting module and the estimated remaining lifespan of the emitting module during a forward inference process.

[0027] As a preferred embodiment of the edge computing-based real-time monitoring method for the status of vehicle headlight emission modules described in this invention, the specific steps for generating the advanced diagnostic result object are as follows:

[0028] The edge computing node based on the vehicle headlight illumination module reads model library information from the runtime environment object and obtains a list of diagnostic model identifiers, while reading diagnostic indicators and comparison result sets from the diagnostic input object;

[0029] Based on the exceedance situation reflected by the diagnostic indicators and the comparison result set, and combined with the monitoring configuration parameters in the operating environment object, the target diagnostic model identifier is selected from the diagnostic model identifier list, and the diagnostic model corresponding to the target diagnostic model identifier is loaded in the edge computing node of the vehicle headlight emitting module.

[0030] Diagnostic indicators are extracted from the diagnostic input object to construct the diagnostic model input vector. The diagnostic model input vector is then input into the diagnostic model to perform inference operations, resulting in a high-level diagnostic result object.

[0031] As a preferred embodiment of the edge computing-based real-time monitoring method for vehicle headlight illumination module status described in this invention, the specific steps for generating the vehicle headlight illumination module status object are as follows:

[0032] Read the luminescent module status category, luminescent module health level, and luminescent module remaining lifespan estimate from the advanced diagnostic results object; and read the diagnostic indicators and comparison result set from the diagnostic input object.

[0033] The consistency of the status category of the light-emitting module with the types of out-of-limit diagnostic indicators in the comparison result set is compared, and the consistency correction rule is applied to correct the status category of the light-emitting module. The corrected status category of the light-emitting module is then output.

[0034] Based on the estimated remaining lifespan of the light-emitting module and the number of out-of-limit diagnostic indicators, a health level and alarm level of the light-emitting module are generated. The corrected light-emitting module status category, light-emitting module health level, estimated remaining lifespan of the light-emitting module, and corresponding vehicle light-emitting module identifier are encapsulated into a vehicle light-emitting module status object.

[0035] As a preferred embodiment of the edge computing-based real-time monitoring method for vehicle headlight emitting module status described in this invention, the consistency correction rule refers to a set of pre-defined rules based on real vehicle operation history data, maintenance records, and other relevant data.

[0036] As a preferred embodiment of the edge computing-based real-time monitoring method for vehicle headlight illumination module status described in this invention, the specific steps for generating the real-time monitoring results of the vehicle headlight illumination module status are as follows:

[0037] Read the status category, health level, remaining lifespan, alarm level, and corresponding headlight module identifier and monitoring time stamp from the headlight module status object, and convert them into monitoring level and text prompt information;

[0038] According to the storage format, the vehicle headlight module identifier, monitoring time stamp, status category, health status, remaining lifespan range, alarm level, monitoring level, and text prompt information are assembled into structured data and output as the real-time monitoring result of the vehicle headlight module status.

[0039] The beneficial effects of this invention are as follows: By constructing a structured operating environment object, efficient organization and configuration management of edge computing node resources are achieved. At the same time, by integrating multi-source diagnostic indicators, dynamically selecting neural network models, and introducing a consistency correction mechanism based on historical experience, a hybrid diagnostic strategy of "data-driven + rule-constrained" is formed, which improves the accuracy of vehicle headlight illumination module status recognition, the rationality of health assessment, and the credibility of remaining life prediction, effectively supporting the predictive maintenance and functional safety assurance of key lighting components in intelligent vehicles. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart of a method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing.

[0042] Figure 2 A flowchart for generating runtime environment objects.

[0043] Figure 3 A flowchart for generating a preprocessed data window.

[0044] Figure 4 A flowchart for generating diagnostic input objects. Detailed Implementation

[0045] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0046] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0048] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for real-time monitoring of the status of a vehicle headlight module based on edge computing, including the following steps:

[0049] S1. After the edge computing node of the vehicle headlight illumination module is powered on, it loads the configuration parameters and initializes the data structure to generate the runtime environment object.

[0050] S1.1. After the edge computing node of the vehicle headlight illumination module is powered on, the clock, watchdog and memory control are started, and storage space is reserved in the random access memory.

[0051] Specifically, after the edge computing node of the headlight illumination module is connected to the vehicle power supply and powered on, it performs a self-check on its own power supply voltage and reset status. After confirming stable power supply, it activates the internal clock circuit to configure the clock signal oscillation, frequency multiplication, and frequency division, providing a unified time reference for subsequent instruction execution. The edge computing node then starts a watchdog timer, setting the watchdog counting period and timeout reload mode to ensure that the watchdog timer responds to program errors. When there is no response for a long time, the watchdog counter can be reset to keep the edge computing node of the headlight module in an available state. The edge computing node of the headlight module starts memory control, initializes the access timing, address mapping method and data width of the random access memory, and performs a zeroing operation on the address range of the random access memory. The edge computing node of the headlight module divides several continuous or segmented address intervals in the random access memory and marks the address intervals as reserved storage space for storing task scheduling information and headlight module operation data.

[0052] S1.2. Read the original content of the monitoring configuration parameters and the diagnostic threshold parameters of the vehicle headlight module, and perform format parsing and verification to obtain the monitoring configuration parameters and the diagnostic threshold of the vehicle headlight module.

[0053] Specifically, the edge computing node of the headlight module sequentially reads the original content of the monitoring configuration parameters and the diagnostic threshold parameters of the headlight module according to the preset address and identifier through the non-volatile memory interface. The original content of the monitoring configuration parameters is temporarily stored in the monitoring configuration parameter buffer in the reserved storage space, and the original content of the headlight module diagnostic threshold is temporarily stored in the headlight module diagnostic threshold buffer in the reserved storage space. At the same time, the edge computing node of the headlight module parses the format of the original content of the headlight module diagnostic threshold, and parses out the headlight module diagnostic threshold corresponding to each diagnostic indicator. The node performs integrity verification and range verification on each headlight module diagnostic threshold. When the headlight module diagnostic threshold passes the verification, it is directly written into the headlight module diagnostic threshold working area. When the headlight module diagnostic threshold fails the verification, the pre-stored default value of the headlight module diagnostic threshold is written into the headlight module diagnostic threshold working area to obtain the monitoring configuration parameters and the headlight module diagnostic threshold.

[0054] S1.3. Establish queue data structures for the task scheduling queue and the exception event queue within the reserved storage space, and set the initial values ​​of the read / write pointers and counters; allocate contiguous storage space for the historical data circular buffer according to the buffer size in the monitoring configuration parameters, and set the starting address and index of the historical data circular buffer.

[0055] Specifically, within the reserved storage space, a contiguous address region is first allocated for the task scheduling queue to sequentially store the queue element information. The initial values ​​of the task scheduling queue read pointer, write pointer, and counter (used to count the current number of elements) are recorded at the starting address of the task scheduling queue. In the remaining area of ​​the reserved storage space, another contiguous address region is allocated for the exception event queue. The initial values ​​of the exception event queue read pointer, write pointer, and counter are recorded at the starting address of the exception event queue. Queue data structures are then established for both the task scheduling queue and the exception event queue within the reserved storage space, and the initial values ​​of the read / write pointers and counters are set.

[0056] The edge computing node of the vehicle headlight module calculates the size of the continuous storage space required for the historical data ring buffer based on the buffer size in the monitoring configuration parameters. It allocates a continuous storage space of the same size as required for the historical data ring buffer in the random access memory, records the starting address of the continuous storage space as the starting address of the historical data ring buffer, and records the initial value of the index used to locate the current position of the historical data ring buffer as the historical data ring buffer index.

[0057] S1.4. Encapsulate the monitoring configuration parameters, vehicle headlight illumination module diagnostic thresholds, task scheduling queue status information, abnormal event queue status information, and historical data ring buffer address and index into a runtime environment object.

[0058] Specifically, after loading the diagnostic thresholds for the headlight illumination module and initializing the task scheduling queue, abnormal event queue, and historical data ring buffer, the edge computing node of the headlight illumination module selects a continuous storage area in the reserved storage space of the random access memory as the storage area for the runtime environment object. Within the runtime environment object storage area, it first writes the monitoring configuration parameters, such as the sampling period, sensor channel identifier, buffer size, and types of diagnostic indicators (e.g., current-related diagnostic indicators, temperature-related diagnostic indicators, luminous flux-related diagnostic indicators, and voltage-related diagnostic indicators). Then, it writes the headlight illumination module diagnostic threshold content corresponding to each diagnostic indicator in the headlight illumination module diagnostic threshold working area. Next, it writes the current values ​​of the task scheduling queue read pointer, task scheduling queue write pointer, and task scheduling queue counter in the task scheduling queue status information, and the current values ​​of the abnormal event queue read pointer, abnormal event queue write pointer, and abnormal event queue counter in the abnormal event queue status information. Finally, it writes the historical data ring buffer address and historical data ring buffer index sequentially at the end of the runtime environment object storage area, forming the runtime environment object.

[0059] S2. Collect the operating data of the vehicle headlight module based on the operating environment object, and perform filtering and alignment preprocessing to generate a preprocessed data window.

[0060] S2.1. Start the sampling timing task in the edge computing node of the vehicle headlight illumination module according to the sampling period and sensor channel configuration in the running environment object.

[0061] Specifically, the sampling period value (example value: 10 milliseconds) and sensor channel configuration information are read from the runtime environment object. The sampling period value specifies the time interval for data acquisition, and the sensor channel configuration information defines the identifiers and parameters of the current sensor, voltage sensor, temperature sensor, and optical sensor that need to be accessed. The edge computing node of the vehicle headlight illumination module uses a built-in timer hardware to set a periodic interrupt, with the interrupt period consistent with the sampling period value. A lightweight real-time monitoring task descriptor is created in the task scheduling queue, and the trigger condition of the lightweight real-time monitoring task descriptor is configured as a timer interrupt event. At the same time, the sensor channel configuration information is written to the associated data area of ​​the lightweight real-time monitoring task descriptor to ensure that the specified sensor can be correctly addressed and read during task execution. The lightweight real-time monitoring task is activated, enters the ready state of the task scheduling queue, and waits for the timer to trigger execution.

[0062] S2.2. In each sampling period, read the vehicle light-emitting module operation data from the sensors specified by the operating environment object in sequence, and attach a timestamp of uniform format to each group of vehicle light-emitting module operation data, and write the vehicle light-emitting module operation data with the attached timestamp into the buffer area.

[0063] Specifically, within each sampling period, when the lightweight real-time monitoring task is triggered by a timer interrupt, the edge computing node of the vehicle headlight module reads the vehicle headlight module's operating data sequentially from the current sensor, voltage sensor, temperature sensor, and optical sensor according to the order defined by the sensor channel configuration in the operating environment object. During the reading process, the edge computing node of the vehicle headlight module synchronously obtains the current time value from the internal high-precision clock and adds a timestamp to each group of vehicle headlight module operating data in a unified format. After the timestamp is added, the edge computing node of the vehicle headlight module writes the timestamped vehicle headlight module operating data into the pre-allocated buffer in the monitoring status object. The address and size of the buffer are determined by the buffer configuration in the operating environment object.

[0064] S2.3. Perform digital filtering on the operating data of the vehicle headlight module in the same sensor channel within the buffer area, and perform interpolation on the operating data of the vehicle headlight module in different sensor channels using timestamps to generate a vehicle headlight module operating dataset.

[0065] Specifically, the data is grouped according to sensor channel identifiers. For the operating data of the vehicle headlight module from multiple consecutive sampling points of the same sensor channel, a moving average filtering algorithm is used for digital filtering. An example window size for the moving average filtering algorithm is 5 sampling points. After filtering, using the earliest acquired timestamp sequence of the sensor channel as the baseline time axis, the operating data of the vehicle headlight module from other sensor channels are calculated using a linear interpolation algorithm to obtain the corresponding values ​​at each time point on the baseline time axis, ensuring that the operating data of the vehicle headlight module from all sensor channels have a unified timestamp sequence. The digitally filtered and interpolated operating data of the vehicle headlight module from each sensor channel are then combined in chronological order to form a complete vehicle headlight module operating dataset.

[0066] S2.4. Based on the vehicle headlight illumination module operation dataset, extract a continuous segment of vehicle headlight illumination module operation data from the cache according to the time window length in the operation environment object and splice it to generate a preprocessed data window.

[0067] Specifically, the edge computing node of the vehicle headlight module reads the time window length parameter from the runtime environment object based on the vehicle headlight module's runtime dataset. The example value of the time window length parameter is 100 milliseconds. According to the time window length parameter, it locates and extracts a continuous time span of vehicle headlight module runtime data from the cache. The extraction process selects data points corresponding to the time window length parameter backward from the latest timestamp to ensure that the timestamps of all data points are continuous. Then, the extracted vehicle headlight module runtime data is spliced ​​in chronological order to form a continuous sequence covering the complete time window, generating a preprocessed data window.

[0068] S3. Based on the preprocessed data window, calculate the diagnostic indicators and compare the diagnostic indicators with the diagnostic threshold of the vehicle headlight emission module in the operating environment object to generate a diagnostic input object.

[0069] S3.1. Based on the operating data of the vehicle lamp light-emitting module in the current sensing channel of the preprocessed data window, calculate the current standard deviation within the time range covered by the preprocessed data window and obtain the current fluctuation amplitude from the current standard deviation.

[0070] Specifically, the arithmetic mean of the operating data sequence of the vehicle headlight emitting module in the current sensing channel is calculated to obtain the average current value. The maximum and minimum current values ​​are obtained by traversing the operating data sequence of the vehicle headlight emitting module in the current sensing channel and finding the maximum and minimum sample values ​​respectively. The standard deviation of the current is calculated based on the sample standard deviation of the operating data sequence of the vehicle headlight emitting module in the current sensing channel. The calculated standard deviation of the current is directly used as the quantification value of the current fluctuation amplitude.

[0071] S3.2. Based on the operating data of the vehicle headlight illumination module in the temperature sensing channel of the preprocessed data window, the temperature rise slope is calculated by linear fitting of the temperature difference and time difference between adjacent sampling points and the temperature peak is extracted.

[0072] Specifically, the temperature difference and corresponding time difference are calculated for each adjacent sampling point in the operating data sequence of the vehicle headlight emitting module in the temperature sensing channel to form a temperature change rate sequence; the least squares method is used to perform linear fitting on the temperature change rate sequence, and the slope of the fitted line is used as the temperature rise slope; at the same time, all sampling points in the operating data sequence of the vehicle headlight emitting module in the temperature sensing channel are traversed to find the maximum temperature sampling value as the temperature peak value.

[0073] It should be noted that the expression for the linear fit of the temperature rise slope is:

[0074] ;

[0075] in, Indicates the slope of the temperature rise. This indicates the total number of sampling points for the operating data of the vehicle headlight illumination module in the temperature sensing channel within the preprocessing data window. Indicates the sampling point number. Indicates the first The time value corresponding to each sampling point Indicates the first The temperature values ​​of the vehicle headlight illumination module operating data corresponding to each sampling point in the temperature sensing channel. This represents the arithmetic mean of all temperature values. This represents the arithmetic mean of all time values.

[0076] S3.3. Based on the operating data of the vehicle lamp light-emitting module corresponding to the optical sensing channel in the preprocessing data window, calculate the rate of change of luminous flux by the ratio of the current luminous flux to the reference luminous flux, and calculate the mean and standard deviation of luminous flux in the preprocessing data window.

[0077] Specifically, the operating data sequence of the vehicle headlight emitting module corresponding to the optical sensing channel is located from the preprocessed data window. Each sampling point in the operating data sequence of the vehicle headlight emitting module contains the current luminous flux value and the corresponding time stamp. The reference luminous flux matching the optical sensing channel is read from the operating environment object. The ratio of each current luminous flux to the reference luminous flux is formed and continuously arranged to form a luminous flux change rate sequence. The edge computing node of the vehicle headlight emitting module summarizes and calculates all the current luminous flux in the operating data sequence of the vehicle headlight emitting module of the optical sensing channel and determines the average luminous flux representing the overall brightness level by combining the number of sampling points. Then, the deviation of each current luminous flux from the average luminous flux is calculated based on the average luminous flux, and the square of the deviation is averaged to obtain the standard deviation of luminous flux used to characterize the degree of brightness fluctuation.

[0078] S3.4. Based on the operating data of the vehicle lamp light-emitting module corresponding to the voltage sensing channel in the preprocessed data window, the voltage ripple amplitude and voltage ripple coefficient are calculated by separating the DC component and AC component, and the average current, current fluctuation amplitude, temperature rise slope, temperature peak, luminous flux change rate, luminous flux standard deviation, voltage ripple amplitude and voltage ripple coefficient are combined to form diagnostic indicators.

[0079] Specifically, by applying smoothing filtering to the operating data of the vehicle headlight module in the voltage sensing channel, a voltage DC component sequence that changes slowly over time is obtained. Then, the operating data of the vehicle headlight module in the voltage sensing channel is compared point by point with the voltage DC component sequence to obtain a voltage AC component sequence that fluctuates around the voltage DC component. By finding the maximum and minimum values ​​of the voltage AC component in the voltage AC component sequence and calculating the amplitude range between them, the voltage ripple amplitude is obtained. At the same time, the voltage ripple coefficient is obtained by comparing the representative value of the voltage DC component sequence in the preprocessed data window with the voltage ripple amplitude. The voltage ripple coefficient is used to characterize the relative fluctuation degree of the voltage AC component with the voltage DC component, reflecting the stability of the power supply voltage of the vehicle headlight module and the electrical stress level of the vehicle headlight module caused by voltage fluctuation. The average current, current fluctuation amplitude, temperature rise slope, temperature peak, luminous flux change rate, luminous flux standard deviation, voltage ripple amplitude, and voltage ripple coefficient are combined to output diagnostic indicators.

[0080] S3.5. Read the diagnostic threshold of the vehicle headlight module corresponding to each diagnostic indicator in the diagnostic indicators from the runtime environment object, compare each diagnostic indicator in the diagnostic indicators with the corresponding diagnostic threshold of the vehicle headlight module, and obtain a set of comparison results.

[0081] Specifically, the diagnostic threshold of the headlight illumination module, corresponding to each diagnostic indicator name, is read from the runtime environment object. The diagnostic indicators are then paired with the headlight illumination module diagnostic thresholds one by one in a preset order. For each pair of diagnostic indicators and headlight illumination module diagnostic thresholds, a size relationship judgment is performed. The degree of exceeding the limit is determined by combining the difference between the diagnostic indicator and the headlight illumination module diagnostic threshold. The degree of exceeding the limit is used to characterize the deviation of each diagnostic indicator value from the corresponding headlight illumination module diagnostic threshold in terms of the magnitude and direction of deviation. The size relationship status and degree of exceeding the limit corresponding to each diagnostic indicator are combined in the same order as the diagnostic indicators to form a set of comparison results.

[0082] S3.6. Based on the comparison result set, count the number and type of out-of-limit diagnostic indicators, and encapsulate the vehicle headlight illumination module operation data, diagnostic indicators and comparison result set in the preprocessed data window into a diagnostic input object.

[0083] Specifically, the comparison result set is traversed according to the order of the diagnostic indicators. For each diagnostic indicator, it is determined whether the comparison result is in the state of exceeding the diagnostic threshold of the headlight module. When the size relationship of the comparison result set records indicates that the diagnostic indicator is greater than the upper limit of the corresponding headlight module diagnostic threshold, or the diagnostic indicator is less than the lower limit of the corresponding headlight module diagnostic threshold, or the diagnostic indicator exceeds the single-sided headlight module diagnostic threshold when the headlight module diagnostic threshold is a single-sided limit, the diagnostic indicator is determined to be in the state of exceeding the diagnostic threshold of the headlight module. The number of exceeding diagnostic indicators is accumulated according to the current, temperature, luminous flux, and voltage diagnostic indicators to form the distribution of the number and type of exceeding diagnostic indicators. The headlight module operation data, diagnostic indicators, and comparison result set in the preprocessed data window are written into the continuous storage area in the order of the predetermined fields. The continuous storage area is marked as a diagnostic input object to form a diagnostic input object.

[0084] S4. Based on the diagnostic input object, select the diagnostic model in the edge computing node of the headlight illumination module and perform inference to generate an advanced diagnostic result object.

[0085] S4.1. The diagnostic model is a neural network model. During one forward inference process, the neural network model simultaneously outputs the state category of the light-emitting module and the remaining lifetime estimate of the light-emitting module.

[0086] Specifically, the neural network model receives the diagnostic indicator sequence from the diagnostic input object as input at the edge computing node of the vehicle headlight module. It extracts the operating features of the vehicle headlight module through multi-layer nonlinear transformation. After the shared feature extraction layer, a classification output layer for outputting the state category of the headlight module and a regression output layer for outputting the remaining life estimate of the headlight module are set. This allows the neural network model to generate the headlight module state category result and the headlight module remaining life estimate result simultaneously in one forward inference process. The headlight module state category result and the headlight module remaining life estimate result are then used as the headlight module state category field and the headlight module remaining life estimate field in the high-level diagnostic result object.

[0087] It should be noted that the neural network model training process involves collecting a sample set containing diagnostic indicators, actual light-emitting module state categories, and actual remaining lifespan during the mass production stage or the test calibration stage with example values ​​for the vehicle headlight emitting module. The diagnostic indicators are used as input to the neural network model, while the actual light-emitting module state categories and actual remaining lifespan are used as supervision signals. The parameters of the classification output layer and the regression output layer in the neural network model are adjusted simultaneously through supervised learning. After the neural network model achieves the expected accuracy requirements in predicting the light-emitting module state categories and estimating the remaining lifespan of the light-emitting module, the neural network model parameters are stored in the edge computing nodes of the vehicle headlight emitting module for online inference.

[0088] S4.2. Based on the edge computing node of the vehicle headlight illumination module, read the model library information from the runtime environment object and obtain the list of diagnostic model identifiers, and at the same time read the diagnostic indicators and comparison result set from the diagnostic input object.

[0089] Specifically, the system accesses the pre-stored model library information fields in the runtime environment object. These fields include attributes such as the diagnostic model identifier, storage address, complexity level, applicable operating condition category of the vehicle lighting module, number of dimensions of the input diagnostic indicators, and output content type of multiple diagnostic models. The system then reads all diagnostic model identifiers belonging to the real-time monitoring of the vehicle lighting module status from the model library information fields in a predetermined order and organizes them into a diagnostic model identifier list, storing it in the working storage area of ​​the vehicle lighting module edge computing node. The vehicle lighting module edge computing node accesses the diagnostic input object, sequentially reading the diagnostic indicator field and comparison result set field. The diagnostic indicators are used as the basis for constructing the input vector of the subsequent diagnostic model, and the comparison result set is used as reference information reflecting the current over-limit status of the vehicle lighting module. Within the vehicle lighting module edge computing node, a current diagnostic preparation state is formed, simultaneously containing the diagnostic model identifier list, diagnostic indicators, and comparison result set.

[0090] S4.3. Based on the diagnostic indicators and the out-of-limit situation reflected by the comparison result set, and combined with the monitoring configuration parameters in the operating environment object, select the target diagnostic model identifier from the diagnostic model identifier list, and load the diagnostic model corresponding to the target diagnostic model identifier into the edge computing node of the vehicle headlight emitting module.

[0091] Specifically, after reading model library information and obtaining the list of diagnostic model identifiers from the runtime environment object, and reading diagnostic indicators and comparison result sets from the diagnostic input object, the edge computing node of the vehicle headlight module classifies the severity of the current operating status of the vehicle headlight module based on the number and type of out-of-limit diagnostic indicators reflected in the comparison result set (for example, when the number of out-of-limit diagnostic indicators is 1 and only the luminous flux diagnostic indicator is slightly out of limit, the luminous flux change rate is slightly higher than the corresponding vehicle headlight module diagnostic threshold, and the temperature, current, and voltage diagnostic indicators are all within the normal range, the severity of the vehicle headlight module operating status is classified as slight). This is combined with the monitoring configuration parameters related to the sampling period, diagnostic indicator types, and vehicle headlight module operating conditions. The content determines the diagnostic model complexity level suitable for the current operating state of the vehicle lighting module. From the diagnostic model identifier list, a target diagnostic model identifier matching the diagnostic model complexity level is selected. The diagnostic model identifier uniquely identifies a diagnostic model within the edge computing node of the vehicle lighting module. It is a one-to-one coded information corresponding to the diagnostic model, typically in the form of a string, numerical number, or a combination of both. It is used to distinguish different diagnostic models in the model library information field of the runtime environment object. The diagnostic model corresponding to the target diagnostic model identifier is loaded from the non-volatile memory of the vehicle lighting module edge computing node into the working storage area of ​​the vehicle lighting module edge computing node, and the target diagnostic model is placed in a ready state to receive diagnostic indicators from diagnostic input objects as inference input.

[0092] S4.4. Extract diagnostic indicators from the diagnostic input object to construct the diagnostic model input vector, input the diagnostic model input vector into the diagnostic model to perform inference operations, and obtain the advanced diagnostic result object.

[0093] Specifically, after the target diagnostic model is loaded and ready for use, the edge computing node of the headlight module accesses the diagnostic input object. According to the pre-defined input order of the target diagnostic model, it sequentially extracts diagnostic indicators such as average current, current fluctuation amplitude, temperature rise slope, temperature peak, luminous flux change rate, luminous flux standard deviation, voltage ripple amplitude, and voltage ripple coefficient from the diagnostic input object. The diagnostic indicators are scaled according to a unified dimension and numerical range, and the scaled diagnostic indicators are combined in a fixed arrangement to form the diagnostic model input vector. The edge computing node of the headlight module inputs the diagnostic model input vector into the diagnostic model, and the neural network model completes the feature extraction and output operations to obtain the diagnostic model output content, which includes the headlight module state category result and the headlight module remaining lifetime estimation result. The edge computing node of the headlight module then encapsulates the diagnostic model output content into a high-level diagnostic result object.

[0094] S5. Based on the advanced diagnostic result object and referring to the indicators in the diagnostic input object, a comprehensive judgment is made to generate the vehicle headlight illumination module status object.

[0095] S5.1. Read the luminescent module status category, luminescent module health level, and luminescent module remaining lifespan estimate from the advanced diagnostic results object, and read the diagnostic indicators and comparison result set from the diagnostic input object.

[0096] Specifically, the system reads the character data of the light-emitting module status category, the floating-point value of the light-emitting module health status, and the integer value of the estimated remaining lifespan of the light-emitting module according to the predefined field structure of the advanced diagnostic result object. At the same time, it accesses the memory storage area of ​​the diagnostic input object and extracts the diagnostic indicators and comparison results stored in the diagnostic input object in sequence. The diagnostic indicators include a series of floating-point values ​​such as average current, current fluctuation amplitude, temperature rise slope, temperature peak, luminous flux change rate, luminous flux standard deviation, voltage ripple amplitude, and voltage ripple coefficient. The comparison results record the over-limit status of each diagnostic indicator and the corresponding vehicle light-emitting module diagnostic threshold in binary form. The read light-emitting module status category, light-emitting module health status, light-emitting module remaining lifespan estimate, diagnostic indicator array, and comparison result set bitmap are temporarily stored in the register group of the edge computing node of the vehicle light-emitting module.

[0097] S5.2. Perform a consistency comparison between the status category of the light-emitting module and the types of out-of-limit diagnostic indicators in the comparison result set, and correct them according to the consistency correction rules, and output the corrected status category of the light-emitting module.

[0098] Specifically, a correspondence table is established between the status categories of the light-emitting module and the types of diagnostic indicators. For example, the "early aging" status corresponds to excessive luminous flux change rate and excessive temperature rise slope, while the "driving abnormality" status corresponds to excessive current fluctuation amplitude and excessive voltage ripple amplitude. All out-of-limit diagnostic indicator types in the comparison result set are iterated through, and the number of out-of-limit diagnostic indicators matching the current light-emitting module status category is counted as the matching degree. Simultaneously, the number of out-of-limit diagnostic indicators that do not match the current light-emitting module status category is counted as the conflict degree. According to the consistency correction rule, when the matching degree is greater than the conflict degree and the matching degree exceeds the matching degree judgment threshold (e.g., value 2), the original light-emitting module status category remains unchanged. When the conflict degree is greater than the matching degree or the matching degree does not reach the matching degree judgment threshold, the most matching light-emitting module status category is selected as the correction result according to the main distribution characteristics of the out-of-limit diagnostic indicator types. The corrected light-emitting module status category is then output.

[0099] It should be noted that the process involves collecting operational samples of vehicle headlight modules that include diagnostic indicators, actual headlight module status categories, and headlight module diagnostic threshold exceedances. First, the matching degree and conflict degree of each operational sample of headlight modules are calculated according to the consistency correction rule. Then, several candidate matching degree judgment thresholds with values ​​of 1, 2, and 3 are tried. For each candidate matching degree judgment threshold, the consistency ratio between the corrected headlight module status category and the actual headlight module status category, as well as the number of misjudgments and omissions, are statistically analyzed. The candidate matching degree judgment threshold that performs best under the condition of a high consistency ratio and a low combined index of misjudgments and omissions is selected as the target matching degree judgment threshold.

[0100] S5.3. Based on the estimated remaining lifespan of the light-emitting module and the number of out-of-limit diagnostic indicators, generate the health level and alarm level of the light-emitting module, and encapsulate the corrected light-emitting module status category, light-emitting module health level, estimated remaining lifespan of the light-emitting module, and corresponding vehicle light-emitting module identifier into a vehicle light-emitting module status object.

[0101] Specifically, the remaining lifespan estimate of the light-emitting module is compared with a preset lifespan threshold, for example, 500 hours. Simultaneously, the total number of diagnostic indicators marked as exceeding the limit in the comparison result set is counted. Based on the percentage of the remaining lifespan estimate of the light-emitting module relative to the preset lifespan threshold and the number of exceeding diagnostic indicators, the health level of the light-emitting module is determined through a light-emitting module health level mapping table, for example, the health level is divided into four levels: excellent, good, average, and poor. At the same time, combining the type distribution and severity of the exceeding diagnostic indicators, an alarm level is generated according to a rule base, for example, the alarm level is divided into four levels: none, prompt, warning, and severe. The corrected light-emitting module status category character, the light-emitting module health level value, the light-emitting module remaining lifespan estimate value, and the vehicle light-emitting module identifier are written into a contiguous storage area in a fixed field order to form a vehicle light-emitting module status object.

[0102] It should be noted that, for multiple vehicle light-emitting modules, under rated working conditions or example values ​​of accelerated aging conditions, the cumulative operating time of each vehicle light-emitting module from initial lighting to the occurrence of a drop in luminous flux to the target ratio (e.g., 70% of the initial luminous flux in the example value), abnormal temperature rise, or failure to work is recorded. At the same time, combined with long-term real vehicle operation statistics, the failure probability of vehicle light-emitting modules in different lifespan ranges is calculated. The lifespan range with a failure probability lower than the target threshold and matching the real vehicle maintenance cycle is determined as the reference lifespan, and the operating time constant corresponding to the reference lifespan is fixed as the preset lifespan threshold.

[0103] Operating samples of vehicle headlight light modules were collected, including estimated remaining lifespan of the light modules, preset lifespan thresholds, the number of out-of-limit diagnostic indicators, and the health levels of the light modules marked in the maintenance records. The ratio of the estimated remaining lifespan of the light modules to the preset lifespan thresholds was divided into several lifespan ratio intervals, and the number of out-of-limit diagnostic indicators was divided into several quantity intervals. For each combination of lifespan ratio intervals and out-of-limit diagnostic indicator quantity intervals, the frequency of occurrence of the corresponding light module health level was counted. The light module health level with the most occurrences was selected as the target light module health level corresponding to the combination of lifespan ratio intervals and out-of-limit diagnostic indicator quantity intervals. For combinations of lifespan ratio intervals and out-of-limit diagnostic indicator quantity intervals with small sample sizes or scattered statistical results, the adjustments were made based on the actual severity of the vehicle headlight light module faults, forming a light module health level mapping table that maps the combination of lifespan ratio intervals and out-of-limit diagnostic indicator quantity intervals to the light module health level.

[0104] S5.4. Consistency correction rules refer to a set of rules based on real vehicle operation history data, maintenance records, and pre-defined rules.

[0105] Specifically, fleet operation data, including diagnostic indicators, actual fault phenomena, and final repair measures, is continuously collected during vehicle use to form a real-vehicle operation history database. Simultaneously, the actual status of the headlight modules and replacement records confirmed during each maintenance are recorded to establish a maintenance record database. Statistical analysis of the real-vehicle operation history database and the maintenance record database identifies the correspondence between specific combinations of diagnostic indicators and actual fault states. For example, when both the luminous flux change rate and temperature rise slope exceed limits simultaneously, there is an 85% probability of early aging. These statistical patterns are then transformed into specific if-then rules, forming a pre-defined set of rules. This set of rules is periodically updated to the edge computing nodes of the headlight modules via OTA (Over-The-Air) updates to provide a basis for correction when contradictory state judgments occur.

[0106] S6. Generate real-time monitoring results of the vehicle headlight module status based on the status category, health level, remaining lifespan, and alarm level in the headlight module status object.

[0107] S6.1. Read the status category, health level, remaining lifespan, alarm level, and corresponding vehicle headlight illumination module identifier and monitoring time stamp from the headlight illumination module status object, and convert them into monitoring level and text prompt information.

[0108] Specifically, the edge computing node of the headlight illumination module accesses the memory storage area of ​​the headlight illumination module status object, and sequentially reads the character data of the illumination module status category, health value, remaining lifespan interval value, alarm level code, headlight illumination module identifier, and monitoring time stamp. According to the correspondence between the alarm level code and the health level mapping table of the illumination module, the alarm level code is converted into a monitoring level identifier. For example, the alarm level "Severe" corresponds to the monitoring level "Level 1". At the same time, the headlight illumination module status category, health value, and remaining lifespan interval value are combined to generate text prompt information. For example, the status category "early aging", "health 65", and "remaining lifespan 600 hours" are combined into the prompt text "Left front low beam headlight: early aging, health 65%, remaining lifespan approximately 600 hours". The monitoring level identifier and text prompt information are organized together with the headlight illumination module identifier and monitoring time stamp into a standardized output data block.

[0109] S6.2. Assemble the vehicle headlight module identifier, monitoring time stamp, status category, health status, remaining lifespan range, alarm level, monitoring level, and text prompt information into structured data according to the storage format, and output it as the real-time monitoring result of the vehicle headlight module status.

[0110] Specifically, the edge computing node of the headlight module sequentially writes the headlight module identifier, monitoring time stamp, status category, health value, remaining lifespan range value, alarm level code, monitoring level identifier, and text prompt information into a continuous storage area according to the storage format. The storage format adopts a structure with a fixed field length and delimiter combination. After the data is assembled, the edge computing node of the headlight module sends the structured data to the body controller through the vehicle communication interface to form the real-time monitoring result of the headlight module status.

[0111] In summary, this invention achieves efficient organization and configuration management of edge computing node resources by constructing a structured runtime environment object. Simultaneously, by integrating multi-source diagnostic indicators, dynamically selecting neural network models, and introducing a consistency correction mechanism based on historical experience, a hybrid diagnostic strategy of "data-driven + rule-constrained" is formed. This improves the accuracy of vehicle headlight illumination module status recognition, the rationality of health assessment, and the reliability of remaining lifespan prediction, effectively supporting predictive maintenance and functional safety assurance of key lighting components in intelligent vehicles.

[0112] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing, characterized in that: include, After the edge computing node of the vehicle headlight illumination module is powered on, it loads the configuration parameters, initializes the data structure, and generates the runtime environment object. The operating data of the vehicle headlight illumination module is collected based on the operating environment object, and filtered and aligned preprocessing is performed to generate a preprocessed data window; Based on the preprocessed data window, diagnostic indicators are calculated, and the diagnostic indicators are compared with the diagnostic threshold of the vehicle light emission module in the operating environment object to generate a diagnostic input object. Based on the diagnostic input object, a diagnostic model is selected and reasoned in the edge computing node of the vehicle headlight module to generate an advanced diagnostic result object. Based on the advanced diagnostic result object and with reference to the indicators in the diagnostic input object, a comprehensive judgment is made to generate the vehicle headlight illumination module status object; Based on the status category, health level, remaining lifespan, and alarm level of the vehicle headlight module status object, generate real-time monitoring results of the vehicle headlight module status.

2. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 1, characterized in that: The specific steps for generating the runtime environment object are as follows: After the edge computing node of the vehicle headlight illumination module is powered on, the clock, watchdog and memory control are started, and storage space is reserved in the random access memory; Read the original content of the monitoring configuration parameters and the diagnostic threshold parameters of the headlight illumination module, and perform format parsing and verification to obtain the monitoring configuration parameters and the diagnostic threshold of the headlight illumination module; Within the reserved storage space, establish queue data structures for the task scheduling queue and the exception event queue, and set initial values ​​for read / write pointers and counters; allocate contiguous storage space for the historical data circular buffer according to the buffer size in the monitoring configuration parameters, and set the starting address and index of the historical data circular buffer; The monitoring configuration parameters, vehicle headlight illumination module diagnostic thresholds, task scheduling queue status information, abnormal event queue status information, and historical data ring buffer addresses and indexes are encapsulated into runtime environment objects.

3. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 1, characterized in that: The specific steps for generating the preprocessed data window are as follows. Based on the sampling period and sensor channel configuration in the runtime environment object, start the sampling timing task in the edge computing node of the vehicle light emission module; Within each sampling period, the vehicle headlight module operation data is read sequentially from the sensors specified by the operating environment object, and a timestamp in a uniform format is added to each group of vehicle headlight module operation data. The vehicle headlight module operation data with the timestamp is then written to the buffer area. Digital filtering is performed on the operating data of the vehicle headlight illumination module in the same sensing channel within the buffer area, and interpolation is performed on the operating data of the vehicle headlight illumination module in different sensing channels using timestamps to generate a vehicle headlight illumination module operating dataset. Based on the vehicle headlight module operation dataset, a continuous segment of vehicle headlight module operation data is extracted from the cache according to the time window length in the operation environment object and spliced ​​together to generate a preprocessed data window.

4. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 1, characterized in that: The calculation of diagnostic indicators based on the preprocessed data window follows these steps. Based on the operating data of the vehicle lamp light-emitting module in the current sensing channel of the preprocessed data window, the current standard deviation is calculated within the time range covered by the preprocessed data window, and the current fluctuation amplitude is obtained from the current standard deviation. Based on the operating data of the vehicle headlight module in the temperature sensing channel of the preprocessed data window, the temperature rise slope is calculated and the temperature peak is extracted by linear fitting of the temperature difference and time difference between adjacent sampling points. Based on the operating data of the vehicle lamp light-emitting module corresponding to the optical sensing channel in the preprocessed data window, the luminous flux change rate is calculated by the ratio of the current luminous flux to the reference luminous flux, and the mean and standard deviation of the luminous flux in the preprocessed data window are calculated. Based on the operating data of the vehicle lamp light-emitting module corresponding to the voltage sensing channel in the preprocessed data window, the voltage ripple amplitude and voltage ripple coefficient are calculated by separating the DC component and the AC component. The average current, current fluctuation amplitude, temperature rise slope, temperature peak, luminous flux change rate, luminous flux standard deviation, voltage ripple amplitude and voltage ripple coefficient are combined to form diagnostic indicators.

5. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 1, characterized in that: The specific steps for generating the diagnostic input object are as follows: Read the diagnostic threshold of the vehicle headlight module corresponding to each diagnostic indicator in the diagnostic indicators from the runtime environment object, compare each diagnostic indicator in the diagnostic indicators with the corresponding diagnostic threshold of the vehicle headlight module, and obtain a set of comparison results. The number and type of out-of-limit diagnostic indicators are counted based on the comparison result set, and the vehicle headlight illumination module operation data, diagnostic indicators and comparison result set in the preprocessed data window are encapsulated into a diagnostic input object.

6. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 5, characterized in that: The diagnostic model is a neural network model, which simultaneously outputs the state category of the light-emitting module and the estimated remaining lifetime of the light-emitting module during a single forward inference process.

7. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 1, characterized in that: The specific steps for generating the advanced diagnostic results object are as follows: The edge computing node based on the vehicle headlight illumination module reads model library information from the runtime environment object and obtains a list of diagnostic model identifiers, while reading diagnostic indicators and comparison result sets from the diagnostic input object; Based on the exceedance situation reflected by the diagnostic indicators and the comparison result set, and combined with the monitoring configuration parameters in the operating environment object, the target diagnostic model identifier is selected from the diagnostic model identifier list, and the diagnostic model corresponding to the target diagnostic model identifier is loaded in the edge computing node of the vehicle headlight emitting module. Diagnostic indicators are extracted from the diagnostic input object to construct the diagnostic model input vector. The diagnostic model input vector is then input into the diagnostic model to perform inference operations, resulting in a high-level diagnostic result object.

8. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 1, characterized in that: The specific steps for generating the vehicle headlight illumination module state object are as follows: Read the luminescent module status category, luminescent module health level, and luminescent module remaining lifespan estimate from the advanced diagnostic results object; and read the diagnostic indicators and comparison result set from the diagnostic input object. The consistency of the status category of the light-emitting module with the types of out-of-limit diagnostic indicators in the comparison result set is compared, and the consistency correction rule is applied to correct the status category of the light-emitting module. The corrected status category of the light-emitting module is then output. Based on the estimated remaining lifespan of the light-emitting module and the number of out-of-limit diagnostic indicators, a health level and alarm level of the light-emitting module are generated. The corrected light-emitting module status category, light-emitting module health level, estimated remaining lifespan of the light-emitting module, and corresponding vehicle light-emitting module identifier are encapsulated into a vehicle light-emitting module status object.

9. The method for real-time monitoring of the status of vehicle headlight emitting modules based on edge computing as described in claim 8, characterized in that: The consistency correction rule refers to a set of rules based on real vehicle operation history data, maintenance records, and pre-defined rules.

10. The method for real-time monitoring of the status of a vehicle headlight emitting module based on edge computing as described in claim 1, characterized in that: The specific steps for generating real-time monitoring results of the vehicle headlight illumination module status are as follows. Read the status category, health level, remaining lifespan, alarm level, and corresponding headlight module identifier and monitoring time stamp from the headlight module status object, and convert them into monitoring level and text prompt information; According to the storage format, the vehicle headlight module identifier, monitoring time stamp, status category, health status, remaining lifespan range, alarm level, monitoring level, and text prompt information are assembled into structured data and output as the real-time monitoring result of the vehicle headlight module status.

Images