Intelligent temperature control method and system for vehicle lamp based on adaptive learning and dynamic prediction

Abstract

The application relates to a vehicle lamp intelligent temperature control method and system based on adaptive learning and dynamic prediction, and belongs to the technical field of vehicle lamp temperature control. The method comprises the following steps: acquiring vehicle lamp temperature data in real time, and preprocessing the vehicle lamp temperature data; according to the preprocessed vehicle lamp temperature data, a preset temperature prediction algorithm is used to predict the vehicle lamp temperature change trend in a future period; according to the vehicle lamp temperature change trend, a temperature control time is determined, and a preset control algorithm is executed according to the temperature control time, so as to reduce the vehicle lamp temperature by controlling the working current of the vehicle lamp, and the starting time of the control algorithm is not later than the temperature control time; the control process is recorded, a control result is generated, and the temperature prediction algorithm and the control algorithm are optimized according to the control result. The application has the effects of dynamically adjusting the thermal management time, improving heat accumulation and prolonging the service life of the vehicle lamp.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of vehicle lamp temperature control, in particular to a vehicle lamp intelligent temperature control method and system based on adaptive learning and dynamic prediction. BACKGROUND

[0002] With the development of the automobile industry, the performance requirements of vehicle lighting systems, especially front headlights, are increasingly improved. New energy and fuel vehicles usually use LED light sources for front headlights, which generate a large amount of heat during operation. At present, the industry generally adopts passive heat dissipation methods, such as increasing or thickening the heat dissipation fins to enhance the heat dissipation efficiency. However, this method is prone to heat accumulation during long-term operation, which not only reduces the service life of the vehicle lamp, but also may directly damage the vehicle lamp, thereby affecting the normal driving of the vehicle and causing safety hazards. Therefore, there is an urgent need for a method that can efficiently dissipate the heat of the front headlight. SUMMARY

[0003] In order to dynamically adjust the heat management time and improve heat accumulation and prolong the service life of the vehicle lamp, the application provides a vehicle lamp intelligent temperature control method and system based on adaptive learning and dynamic prediction.

[0004] In a first aspect, the application provides a vehicle lamp intelligent temperature control method based on adaptive learning and dynamic prediction, comprising:

[0005] real-time acquisition of vehicle lamp temperature data, preprocessing of the vehicle lamp temperature data;

[0006] According to the preprocessed vehicle lamp temperature data, a temperature prediction algorithm is used to predict the vehicle lamp temperature trend in the future period;

[0007] According to the vehicle lamp temperature trend, the temperature control time is determined, and a preset control algorithm is executed according to the temperature control time to reduce the vehicle lamp temperature by controlling the working current of the vehicle lamp, and the start time of the control algorithm is not later than the temperature control time;

[0008] The control process is recorded to generate a control result, and the temperature prediction algorithm and the control algorithm are optimized according to the control result.

[0009] By adopting the above technical solution, the temperature prediction algorithm is used to predict the temperature rise trend in advance, so as to start the control before the temperature exceeds the standard (the default temperature control time is the time point when the temperature exceeds the standard), effectively reduce the heat accumulation, prolong the service life of the vehicle, and avoid the hysteresis of the traditional passive heat dissipation. Moreover, each control result will be fed back to the algorithm model (i.e. the temperature prediction algorithm and the control algorithm), so that the algorithm model is iteratively optimized by the control result, forming a closed-loop learning.

[0010] Optionally, the temperature control time is determined according to the temperature change trend of the vehicle lamp, and when the current time reaches the temperature control time, a preset control algorithm is executed to reduce the temperature of the vehicle lamp by regulating the working current of the vehicle lamp.

[0011] The temperature control time is determined according to the temperature change trend of the vehicle lamp, and the temperature control time refers to a time when the corresponding temperature data in the temperature change trend of the vehicle lamp exceeds a preset temperature threshold.

[0012] According to the real-time acquired temperature data of the vehicle lamp, it is analyzed whether the actual temperature data change trend deviates greatly from the temperature change trend of the vehicle lamp.

[0013] If so, the temperature change trend of the vehicle lamp is corrected based on a preset correction model, so that the corrected temperature change trend of the vehicle lamp is similar to the actual temperature data change trend; and the temperature control time is determined and updated according to the corrected temperature change trend of the vehicle lamp.

[0014] According to the real-time updated temperature control time, a preset control algorithm is executed to reduce the temperature of the vehicle lamp by regulating the working current of the vehicle lamp, and the starting time of the control algorithm is not later than the temperature control time.

[0015] By adopting the above technical solution, it is determined whether the actual temperature data change trend deviates from the temperature change trend of the vehicle lamp predicted by the temperature prediction algorithm according to the real-time detected actual temperature data of the vehicle lamp, so as to verify the prediction accuracy of the temperature prediction algorithm. If it deviates, the corresponding temperature change trend of the vehicle lamp is corrected in real time based on a preset correction model to ensure the prediction accuracy, so as to realize dynamic adjustment of the temperature control time according to the actual temperature change, and realize efficient emergency protection for sudden abnormal situations.

[0016] Optionally, the control algorithm at least includes a linear algorithm, a step algorithm, and an exponential algorithm; the linear algorithm is used to make the working current change linearly with time when regulating the working current; the step algorithm is used to make the working current change in a step form when regulating the working current; and the exponential algorithm is used to make the working current change in an exponential function form with time when regulating the working current.

[0017] The preset control algorithm is executed, including:

[0018] According to the real-time acquired temperature data of the vehicle lamp, the change rate of the temperature data with time is calculated to obtain a temperature slope kc.

[0019] A starting time is determined, and when the current time reaches the starting time, the control algorithm is selected according to the temperature slope kc and a preset first selection logic, and the selected control algorithm is executed; wherein the starting time is not later than the temperature control time.

[0020] The first selection logic is:

[0021] When |kc| < k1, the regulation algorithm is switched to a linear algorithm;

[0022] When k1≤|kc|≤k2, the regulation algorithm is switched to a step algorithm;

[0023] When |kc| > k2, the regulation algorithm is switched to an exponential algorithm.

[0024] By adopting the above technical solution, multiple regulation algorithms are set to dynamically adapt to different temperature rise scenarios. For example, the linear algorithm is suitable for a slow temperature rise scenario, which can achieve gentle control of the current, avoid brightness jump of the vehicle lamp, and improve user experience. The step algorithm is suitable for a temperature step rise scenario (such as intermittent failure of a cooling fan), which can quickly respond to temperature step changes while reducing regulation frequency and MCU load. The exponential algorithm is suitable for a temperature sudden rise scenario (such as complete failure of a radiator), which can achieve millisecond-level emergency suppression and prevent thermal runaway damage to the vehicle lamp.

[0025] Optionally, the method further comprises:

[0026] acquiring light-on data of the vehicle lamp in a historical period, wherein the light-on data at least includes a light-on period and a use scenario; and dividing the light-on period into a necessary light-on period and a non-necessary light-on period according to the use scenario;

[0027] learning the distribution of the necessary light-on period of the vehicle lamp in the historical period, and predicting the distribution of the necessary light-on period of the corresponding vehicle lamp in a future period;

[0028] fusing and comparing the vehicle lamp temperature change trend and the necessary light-on period distribution to determine whether there is a target necessary light-on period and output a determination result; the target necessary light-on period meets: according to the vehicle lamp temperature change trend, there is an over-temperature moment in the target necessary light-on period in which the temperature data exceeds a preset temperature threshold;

[0029] The determination of the start time, and when the current time reaches the start time, according to the temperature slope kc and the preset first selection logic, a regulation algorithm is selected, and the selected regulation algorithm is executed, including:

[0030] if the target necessary light-on period exists, then, according to preset third selection logic, a temperature difference between the temperature data at the over-temperature time point and a preset temperature threshold is calculated, a regulation algorithm is selected based on the temperature difference and a preset second correspondence table, and a corresponding regulation time period is determined; and when the current time reaches the regulation time period, the regulation algorithm selected is used to regulate the working current of the vehicle lamp within the regulation time period, so that the temperature data of the vehicle lamp does not exceed the preset temperature threshold within the target necessary light-on period at the current time;

[0031] if the target necessary light-on period does not exist, then, according to preset second selection logic, whether there is a regulation result similar to the current temperature slope is determined from the stored regulation results in the historical period, if there is, the regulation algorithm corresponding to the regulation result similar to the current temperature slope is selected as the currently selected regulation algorithm, the start time is determined, and when the current time reaches the start time, the regulation algorithm matched is executed;

[0032] if there is no regulation result similar to the current temperature slope, then, according to the first selection logic, a regulation algorithm is selected, a start time is determined, and when the current time reaches the start time, the regulation algorithm selected is executed.

[0033] By adopting the above technical solution, based on the light-on time distribution in the historical period and the vehicle use scenario, the necessary light-on period is divided, and the future necessary light-on period is predicted by fusion probability analysis, the necessary time (i.e. necessary light-on period) is strengthened, the driving behavior prediction-based vehicle lamp temperature pre-control method is realized, the driver driving habit, prediction algorithm and dynamic regulation are deeply coupled, the selection logic of the regulation algorithm is classified, and the predictive intervention (i.e. third selection logic) is used to prioritize the second selection logic and the first selection logic, so as to ensure that there is no over-temperature situation in the necessary light-on period.

[0034] Optionally, the method further comprises:

[0035] The aging degree of the vehicle lamp is updated in real time, and the preset temperature threshold is adjusted in real time according to the updated aging degree of the vehicle lamp.

[0036] By adopting the above technical solution, the preset temperature threshold is adaptively adjusted according to the aging degree of the vehicle lamp, so as to ensure that the determined temperature control time is more close to the actual use of the vehicle lamp, and the determination accuracy of the temperature control time is improved.

[0037] Optionally, the real-time updating of the aging degree of the vehicle lamp comprises:

[0038] Whenever a regulation result is generated, the historical regulation result is analyzed to determine the regulation frequency and heat dissipation efficiency of the vehicle lamp, the aging degree of the vehicle lamp is predicted based on the regulation frequency and heat dissipation efficiency, and the aging degree of the vehicle lamp is updated in real time.

[0039] By adopting the technical scheme, the aging degree of the car lamp is predicted and analyzed more accurately according to the real-time control frequency and the change in heat dissipation efficiency.

[0040] Optionally, the method further comprises:

[0041] The number of times of control of each control algorithm is periodically counted, and a maintenance cycle is output according to a statistical result and a preset maintenance analysis model, so that a maintenance personnel can know the maintenance cycle.

[0042] By adopting the technical scheme, the maintenance cycle is output according to the frequency of use of the control algorithm, so as to remind the maintenance personnel to perform the maintenance operation regularly.

[0043] In a second aspect, the present application provides a car lamp intelligent temperature control system based on adaptive learning and dynamic prediction, comprising:

[0044] A temperature data preprocessing module is configured to acquire car lamp temperature data in real time and preprocess the car lamp temperature data.

[0045] A car lamp temperature prediction module is configured to predict and output a car lamp temperature change trend in a future period of time according to the preprocessed car lamp temperature data and by using a preset temperature prediction algorithm.

[0046] A car lamp temperature control module is configured to determine a temperature control time according to the car lamp temperature change trend, and execute a preset control algorithm according to the temperature control time, so as to reduce the car lamp temperature by controlling the working current of the car lamp, and the start time of the control algorithm is not later than the temperature control time.

[0047] A control feedback optimization module is configured to record a control process, generate a control result, and optimize the temperature prediction algorithm and the control algorithm according to the control result.

[0048] In a third aspect, the present application provides a car lamp intelligent temperature control device based on adaptive learning and dynamic prediction, comprising a memory and a processor, and the memory stores a computer program capable of being loaded and executed by the processor and performing the method according to any one of the first aspect.

[0049] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program capable of being loaded and executed by a processor and performing the method according to any one of the first aspect.

[0050] In summary, the present application has at least one of the following beneficial technical effects:

[0051] In the present application, the temperature rising trend is determined in advance by the temperature prediction algorithm, so that the regulation and control can be started before the temperature exceeds the standard (the default temperature control time is the time point when the temperature exceeds the standard), effectively reducing the heat accumulation, prolonging the service life of the vehicle, avoiding the hysteresis of traditional passive heat dissipation, and the regulation and control result will be fed back to the algorithm model (i.e. temperature prediction algorithm and regulation and control algorithm) to drive the algorithm model to iterate and optimize, forming a closed loop learning;

[0052] Further, according to the actual temperature data of the vehicle lamp detected in real time, it is determined whether the actual temperature data change trend deviates from the vehicle lamp temperature change trend predicted by the temperature prediction algorithm, so as to verify the prediction accuracy of the temperature prediction algorithm, and if it deviates, the corresponding vehicle lamp temperature change trend is corrected based on the preset correction model in real time to ensure the prediction accuracy, and dynamic heat management is realized on the basis of the prediction result of the temperature prediction algorithm to reduce heat accumulation. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0054] Figure 1 is a schematic diagram of the intelligent temperature control method for vehicle lamp based on adaptive learning and dynamic prediction disclosed by the embodiments of the present application.

[0055] Figure 2 is a flowchart of the intelligent temperature control method for vehicle lamp based on adaptive learning and dynamic prediction disclosed by the embodiments of the present application.

[0056] Figure 3 is a structural block diagram of the intelligent temperature control system for vehicle lamp based on adaptive learning and dynamic prediction disclosed by the embodiments of the present application.

[0057] Explanation of reference numerals: 201, temperature data preprocessing module; 202, vehicle lamp temperature prediction module; 203, vehicle lamp temperature regulation and control module; 204, regulation and control feedback optimization module. DETAILED DESCRIPTION

[0058] The following will be described in combination with the Figures 1-3 The present application will be further described in detail.

[0059] The embodiments of the present application disclose an intelligent temperature control method for vehicle lamp based on adaptive learning and dynamic prediction (hereinafter referred to as intelligent temperature control method for vehicle lamp), and the execution subject is an intelligent temperature control system for vehicle lamp based on adaptive learning and dynamic prediction (hereinafter referred to as intelligent temperature control system for vehicle lamp).Figure 1 This section elaborates on the specific execution process of the vehicle lighting intelligent temperature control system.

[0060] S101 acquires real-time headlight temperature data and preprocesses the headlight temperature data.

[0061] S102, based on the preprocessed headlight temperature data, uses a preset temperature prediction algorithm to predict and output the headlight temperature change trend in the future period.

[0062] S103 determines the temperature control time based on the temperature change trend of the headlight and executes a preset control algorithm based on the temperature control time to reduce the headlight temperature by controlling the working current of the headlight, and the start time of the control algorithm is no later than the temperature control time.

[0063] S104 records the control process, generates control results, and optimizes the temperature prediction algorithm and control algorithm based on the control results.

[0064] S103 specifically includes the following sub-steps:

[0065] S1031, determine the temperature control time based on the temperature change trend of the headlights; the temperature control time refers to the time during which the corresponding temperature data in the headlight temperature change trend exceeds the preset temperature threshold.

[0066] S1032, based on the real-time acquired headlight temperature data, analyze whether the actual temperature data change trend deviates significantly from the headlight temperature change trend.

[0067] If S1033 is true, then based on the preset correction model, the trend of headlight temperature change is corrected so that the corrected headlight temperature change trend is similar to the actual temperature data change trend; based on the corrected headlight temperature change trend, the temperature control time is determined and updated.

[0068] S1034, based on the temperature control time determined by real-time updates, executes a preset control algorithm to reduce the temperature of the headlights by adjusting the working current of the headlights, and the start time of the control algorithm is no later than the temperature control time.

[0069] The control algorithm includes at least a linear algorithm, a step algorithm, and an exponential algorithm. The linear algorithm is used to control the operating current so that the operating current changes linearly with time. The step algorithm is used to control the operating current so that the operating current changes in a step manner. The exponential algorithm is used to control the operating current so that the operating current changes in an exponential function form with time.

[0070] Accordingly, the "execute the preset control algorithm" in S1034 specifically includes the following steps:

[0071] S1035, based on the real-time acquired headlight temperature data, calculates the rate of change of temperature data over time and obtains the temperature slope kc;

[0072] S1036, determine the start time, and when the start time is reached at the current time, select the control algorithm according to the temperature slope kc and the preset first selection logic, and execute the selected control algorithm; wherein, the start time is not later than the temperature control time;

[0073] The first choice logic is:

[0074] When |kc| < k1, the control algorithm is switched to a linear algorithm;

[0075] When k1≤|kc|≤k2, the control algorithm is switched to the step algorithm;

[0076] When |kc|>k2, the control algorithm is switched to an exponential algorithm.

[0077] In implementation, the intelligent temperature control system for vehicle lights uses preset temperature sensors to detect and obtain vehicle light temperature data (such as the ambient temperature of the vehicle lights). The acquired vehicle light temperature data is then preprocessed using an exponentially weighted moving average (EWMA) filter to eliminate sensor noise and highlight temperature change trends. The corresponding processing formula is: T t =α⋅Tt+(1−α)⋅T t−1 ',in, α For weighting coefficients (e.g., 0.3), T t 'This represents the filtered temperature.

[0078] Next, the intelligent temperature control system for vehicle lights uses a temperature prediction algorithm to predict the trend of vehicle light temperature changes over a future period. The prediction principle is as follows:

[0079] a. Temperature change slope calculation: Fit the instantaneous slope to the filtered temperature data sequence (e.g., using least squares or difference methods):

[0080] Temperature slope k= , where n is a positive integer.

[0081] b. Match similar scene instances from the historical model:

[0082] The historical model (EEPROM) stores several scenario instances. Each scenario instance includes a temperature array table (a data set consisting of temperature data that changes over time in historical periods) and corresponding temperature control parameters. Specifically, the temperature control parameters are: the temperature slope that changes over time, the corresponding control algorithm used, and the control results after control according to the corresponding control algorithm (such as temperature data after control, cooling rate, energy consumption, etc.).

[0083] In addition, each scene instance corresponds to a scene type. The distinction between different scene types can include: ambient temperature (i.e., the ambient temperature when the temperature control is started), LED working status (initial current I0, headlight switch status (on / off), cooling fan PWM duty cycle), heat dissipation conditions (cooling fan base speed), and stage (control stage, non-control stage). The scene instances can be pre-classified and assigned to the corresponding scene types based on the aforementioned distinction criteria.

[0084] The scene type of the current scene is determined based on the ambient temperature, temperature slope, LED working status, stage, and the corresponding control algorithm if it is in the control stage. Then, scene instances similar to the current scene (i.e., scene instances in the scene type that are in the same stage and whose ambient temperature and temperature slope are more similar than the preset similarity) are retrieved from the historical model as similar scene instances.

[0085] c. Use the temperature change trend contained in similar scenario examples as the headlight temperature change trend (which can be represented as a curve of temperature changing over time).

[0086] Ultimately, this allows for the prediction of trends in vehicle headlight temperature changes.

[0087] Then, the vehicle headlight intelligent temperature control system is used to determine whether there is a target time in the future based on the trend of vehicle headlight temperature change: starting from the target time, the temperature data corresponding to the target time is higher than the preset temperature threshold (such as 115℃), and the temperature data corresponding to all times before the target time is not higher than the preset temperature threshold. If this is the case, the target time is used as the temperature control time, and the start time is determined (the start time is not later than the temperature control time). When the start time is reached at the current time, the control operation is started (that is, the preset control algorithm is executed to control the working current of the vehicle headlight).

[0088] During this process, the temperature prediction algorithm will continuously adjust the predicted trend of the headlight temperature based on the actual temperature of the headlights detected in real time. Specifically:

[0089] The intelligent temperature control system for vehicle lights calculates the temperature slope k in real time based on temperature data detected by preset temperature sensors. c Real-time determination of the calculated temperature slope k cThe model checks whether the temperature slope is similar to the temperature slope at the time corresponding to the current moment in the headlight temperature change trend. This is used to determine whether the actual temperature trend deviates significantly from the predicted headlight temperature change trend. If the deviation (the difference between the current temperature slope and the temperature slope at the time corresponding to the headlight temperature change trend) is greater than a preset range, the preset dynamic response condition is considered met. In this case, the prediction model is updated using an exponential weighting method, that is, the temperature slope of the headlight temperature change trend is corrected, and the corrected temperature slope k' is calculated; where k' = β·k c +(1-β)·k h β is the dynamic weight (e.g., 0.7), k c The temperature slope is calculated based on the currently detected temperature data; k h Let k be the slope of the temperature change at the corresponding moment in the trend of vehicle headlight temperature change, and then replace k with the corrected temperature slope k'. h This allows for the correction and updating of the corresponding headlight temperature change trend, ensuring that the updated headlight temperature change trend closely matches the actual temperature changes.

[0090] If the temperature change trend of the vehicle headlights includes a temperature control period (i.e., an overheating situation will occur in the future), then a control algorithm needs to be selected and executed at the start time. Accordingly, since the control algorithms disclosed in this application include linear algorithms, ladder algorithms, and exponential algorithms, one of them needs to be selected as the control algorithm. The selection method is as follows:

[0091] The selection logic of the control algorithm includes a first selection logic and a second selection logic: the first selection logic is to set k... c The slope is compared with the preset slope thresholds (k1 and k2 shown below) corresponding to the current scene type, and the corresponding control algorithm is determined based on the comparison results and the preset correspondence table; that is:

[0092] |k c If | < k1, then choose the linear algorithm;

[0093] k1≤|k c If |≤k2, then choose the ladder algorithm;

[0094] |k c If |>k2, then choose the exponential algorithm.

[0095] It should be noted here that the preset slope thresholds for each scene type are pre-stored, and each scene type corresponds to a corresponding relationship table, which stores the control algorithms corresponding to different comparison results.

[0096] The second selection logic is based on the scene instances stored in the historical model. If there is a scene instance that matches the current temperature slope (i.e., the similar scene instance mentioned above), then the control algorithm corresponding to the similar scene instance is selected as the control algorithm for the current scene.

[0097] It should also be noted that the priority of the second selection logic is higher than that of the first selection logic. That is, if the control algorithm selected according to the second selection logic is inconsistent with the control algorithm selected according to the first selection logic, the control algorithm selected according to the second selection logic shall be selected as the control algorithm corresponding to the current temperature slope.

[0098] When the selected control algorithm is a linear algorithm, the formula is: I(t) = I0 - ε·t; where I0 refers to the initial current value (i.e., the LED operating current at the start of control), ε is a preset attenuation coefficient, and ε corresponding to different temperature slope ranges k is pre-stored. The range is determined by the temperature slope k, and then the specific value of ε is determined; t refers to the accumulated time from the start of the current control algorithm (t=0), and I(t) is the current value calculated corresponding to the start of the control algorithm. The operating current of the LED headlight is adjusted according to the real-time calculated I(t) to achieve linear and smooth control of the operating current. At this time, the cooling fan can be controlled to operate at a constant power (e.g., 50% PWM).

[0099] When the selected control algorithm is the stepped algorithm, the temperature range and its corresponding current adjustment ratio (e.g., X%) of the current temperature data are determined in real time based on several preset temperature ranges and the current adjustment ratio corresponding to each temperature range. The adjusted current value I is then calculated using the formula I = X%·Imax. Whenever the temperature range of the temperature data changes, the specific value of X% will also change, and the calculated current will also change accordingly. That is, the current will change in a stepped manner with temperature changes, thereby controlling the operating current of the LED headlight, where Imax is a preset value. At this time, the power of the cooling fan increases in a stepped manner with temperature (e.g., X%PWM).

[0100] When the selected control algorithm is an exponential function, according to the formula: I(t) = I0·e -λt Where λ is the preset decay rate, which is a constant; starting from the start time of the current control algorithm (t=0), the adjusted current value I(t) is calculated in real time, and the working current of the LED lamp is adjusted in real time according to the adjusted current value calculated in real time. At this time, the cooling fan is controlled to run at full speed (such as 100% PWM).

[0101] Finally, the start time is determined by the selected control algorithm. For example, when the control algorithm is a linear algorithm, the start time is the time corresponding to 10 minutes before the temperature control time. When the control algorithm is a linear algorithm, the start time is 10 seconds before the temperature control time. When the control algorithm is a stepped algorithm, the start time can be 5 minutes before the temperature control time. When the current time reaches the start time, the selected control algorithm is executed until the temperature data does not exceed the preset temperature threshold and then the control stops.

[0102] During the execution of the control algorithm, the following data will be recorded for feedback analysis: temperature change curve (a dataset consisting of real-time temperature data and the corresponding temperature slope over time), temperature control parameters (corresponding to the type of control algorithm used: linear algorithm / step algorithm / exponential algorithm), current adjustment amplitude ΔI, and performance indicators as control results. The historical model will be updated based on these control results. The performance indicators must include at least the cooling rate. Whether the target temperature has been reached (e.g., below 115℃) and whether it has caused side effects (e.g., LED flickering, increased noise). Specifically:

[0103] Based on the above recorded content, the slope thresholds (i.e., k1 and k2) and algorithm parameters (i.e., ε, temperature range, λ) stored in the historical model are dynamically optimized.

[0104] Analyze the records of successful temperature control (i.e., the temperature drops below the target temperature after control), and score the effectiveness of each successful control record. The score Q = preset cooling efficiency weight * cooling rate + preset stability weight * (1 - side effect level); where different side effects are preset with corresponding levels (levels are expressed as percentages). Then, for each scenario type, periodically select the temperature rate k corresponding to the record with the highest score Q from all successful control records for the corresponding scenario type within the corresponding period, as the optimal temperature rate k for that scenario type. p If the optimal temperature rate k p The corresponding score Q p If the score is greater than the preset slope threshold corresponding to the scene type, then the formula is used: k new =α·k old +(1-α)·k p The optimized slope threshold corresponding to the corresponding scene type is calculated, where k old This refers to the preset slope thresholds (i.e., k1 and k2) corresponding to the scene type, k newThis refers to the preset slope thresholds obtained after optimization for the corresponding scene type (i.e., the optimized k1 and k2). Specifically, the unoptimized k1 and k2 are substituted into the aforementioned formula to calculate the corresponding optimized k1 and k2. The optimized k1 and k2 are then used to update the corresponding preset slope thresholds, and the scores Q of the optimized k1 and k2 are used as well. p Update the score Q' for the corresponding scene type. Similarly, optimize the preset slope thresholds (i.e., k1 and k2) for each scene type to update the preset slope thresholds for each scene type stored in the historical model.

[0105] In addition, it is also used to optimize algorithm parameters based on the recorded content. For example, if the cooling is too slow (i.e., the cooling rate...) If the temperature range is less than the first preset value, then increase ε. If there is a temperature range with poor control effect (e.g., high probability of unsuccessful control or high level of side effects), then merge or split the corresponding temperature range. If the cooling is too fast and causes the LED to turn off, then decrease λ.

[0106] Furthermore, since the control algorithm actively alters the heat generation and dissipation behavior of the lamps after regulation, the temperature change trajectory deviates from the initial prediction. If the prediction is not updated, subsequent decisions may be based on erroneous data, leading to temperature control failure. To address this, a preset trigger scenario is proposed to activate the temperature prediction algorithm for re-prediction when the trigger scenario is met. Specifically, the trigger scenario is when the current stage changes (i.e., when transitioning from a non-regulation stage to a regulation stage, or from a regulation stage to a non-regulation stage). Based on the parameters at the time of the stage change (such as the initial current I', fan speed, heat generation power P=(I')2*R, ambient temperature, the control algorithm used, the real-time temperature after regulation activation, and the temperature slope), a similar scenario instance similar to the current scenario is retrieved from the historical model, and the temperature change trend corresponding to the similar scenario instance is used to update the headlight temperature change trend, thus updating the prediction results of the temperature prediction algorithm. In other embodiments, the parameters at the time of the stage change can also be input into a pre-learned and trained converged LSTM model to utilize the LSTM model to output the corresponding headlight temperature change trend.

[0107] Optionally, the intelligent temperature control method for vehicle lights also includes the following steps:

[0108] Acquire vehicle headlight usage data over historical periods, including at least the time periods during which lights are turned on and the usage scenarios. Based on the usage scenarios, divide the time periods during which lights are turned on into necessary and unnecessary times.

[0109] By studying the distribution of necessary lighting times for vehicle lights in historical periods, we can predict the distribution of necessary lighting times for the corresponding vehicle lights in future periods.

[0110] The trend of vehicle headlight temperature change and the distribution of necessary headlight-on periods are fused and compared to determine whether there is a target necessary headlight-on period and output the determination result; the target headlight-on period meets the following conditions: according to the trend of vehicle headlight temperature change, there are over-temperature moments in the target headlight-on period where the temperature data exceeds the preset temperature threshold.

[0111] S1036 further includes the following sub-steps:

[0112] If there is a target necessary lighting period, then according to the preset third selection logic, the temperature difference between the temperature data at the time of the overheating moment and the preset temperature threshold is calculated. Based on the temperature difference and the preset second correspondence table, the control algorithm and the corresponding control period are selected. When the current time reaches the control period, the selected control algorithm is used to control the working current of the headlights during the control period so that the headlight temperature data never exceeds the preset temperature threshold during the target necessary lighting period.

[0113] If there is no target necessary lighting period, then according to the preset second selection logic, determine whether there is a control result similar to the current temperature slope from the control results stored in the historical period. If there is, select the control algorithm corresponding to the control result similar to the current temperature slope as the current selected control algorithm, determine the start time, and execute the matched control algorithm when the start time is reached at the current time.

[0114] If no control result similar to the current temperature slope is found, the first selection logic is followed to select a control algorithm, determine the start time, and execute the selected control algorithm when the start time is reached at the current time.

[0115] In implementation, several usage scenario labels are preset, and usage scenarios are divided into two main categories based on the necessity of turning on the lights: necessary scenarios, such as nighttime driving, tunnels, and heavy rain (which can be confirmed through vehicle cameras / light sensors / weather APIs), and non-necessary scenarios, such as short-term daytime use and leaving the engine running while parked. Based on these two categories of usage scenarios, the time periods for turning on the lights are further divided into time periods corresponding to necessary scenarios (i.e., necessary lighting periods) and time periods corresponding to non-necessary scenarios (i.e., non-necessary lighting periods). Necessary lighting periods can be considered as the time periods during which the headlights must be on under the corresponding usage scenario. For each vehicle's headlights, a Gaussian Mixture Model (GMM) is used to cluster the necessary lighting periods of the corresponding headlights in historical periods, extracting typical time distribution patterns, thereby predicting and outputting the distribution of necessary lighting periods in future time periods for each vehicle's headlights; that is, based on the necessary lighting periods in historical periods, the time periods in future time periods that coincide with the necessary lighting periods in historical periods are considered necessary lighting periods.

[0116] Next, calculate the risk coefficient for overheating during each necessary lighting period in the future. Risk coefficient = ;in, Based on the predicted trend of headlight temperature change, the headlight temperature at time t within the future time period and the necessary headlight-on period is determined.

[0117] The necessary lighting periods can be represented as [ P(t) represents the probability that time t falls within the necessary lighting period. For example, If the risk coefficient is greater than the preset coefficient threshold, the corresponding necessary lighting period will be determined as the target necessary lighting period, and the time when the risk coefficient is greater than the preset coefficient threshold will be taken as the over-temperature moment.

[0118] Based on several preset risk levels (represented as risk coefficient ranges), and the corresponding control period and control algorithm for each risk level, the risk level corresponding to the necessary lighting period for the target is determined, along with the corresponding control period and control algorithm. For example, if the risk coefficient is [0.1, 0.3], the corresponding control period is [overheating moment - 30 min, overheating moment], and the corresponding control algorithm is a linear algorithm; if the risk coefficient is (0.3, 0.6), the corresponding control period is [overheating moment - 60 min, overheating moment], and the corresponding control algorithm is a stepped algorithm; if the risk coefficient is >0.6, the corresponding control period is [current moment, overheating moment], and the corresponding control algorithm is an exponential algorithm.

[0119] The specific selection method for the control algorithm based on the second or first logic, as well as the scheme for determining the start time, have been disclosed in the previous text and will not be repeated here.

[0120] Optionally, the intelligent temperature control method for vehicle lights also includes the following steps:

[0121] Whenever a control result is generated, historical control results are analyzed to determine the control frequency and heat dissipation efficiency of the headlights. Based on the control frequency and heat dissipation efficiency, the aging degree of the headlights is predicted, the aging degree of the headlights is updated in real time, and the preset temperature threshold is adjusted in real time according to the updated aging degree of the headlights.

[0122] In implementation, whenever a new control result is generated, the control frequency (such as the number of times the control is initiated within a specified time) and the heat dissipation efficiency are calculated as the temperature drop per unit time. The control frequency, heat dissipation efficiency, and cumulative start-up time since the vehicle headlight was manufactured are input into a preset aging degree prediction model so that the aging degree prediction model outputs an aging coefficient β. The prediction expression of the aging degree prediction model can be: β = w1 * control frequency + w2 * heat dissipation efficiency + w3 * cumulative start-up time; w1, w2, and w3 are preset weight values.

[0123] In addition, the system pre-stores the specific values ​​of the preset temperature threshold corresponding to the range of several vehicle headlight aging coefficients and the range of each vehicle headlight aging coefficient. The preset temperature threshold is updated by referring to the aforementioned correspondence.

[0124] Optionally, the intelligent temperature control method for vehicle lights also includes the following steps:

[0125] The number of adjustments for each control algorithm is periodically counted. Based on the statistical results and the preset maintenance analysis model, the maintenance cycle is output for maintenance personnel to know.

[0126] In implementation, the method for generating maintenance cycles using the maintenance analysis model is as follows:

[0127] First, calculate the control intensity S = (μ1 * N1 + μ2 * N2 + μ3 * N3) / (N1 + N2 + N3);

[0128] Where N1 is the number of times the linear algorithm is triggered within the period; N2 is the number of times the ladder algorithm is triggered within the period; N3 is the number of times the exponential algorithm is triggered within the period; and μ1, μ2, and μ3 are preset weights.

[0129] Then, based on the control intensity S, the following formula is used to calculate:

[0130] When S≥S';

[0131] When S is less than S'.

[0132] This application also discloses an intelligent vehicle headlight temperature control system based on adaptive learning and dynamic prediction. (Refer to...) Figure 3 The intelligent temperature control system for vehicle lights based on adaptive learning and dynamic prediction includes:

[0133] Temperature data preprocessing module 201 is used to acquire vehicle headlight temperature data in real time and preprocess the vehicle headlight temperature data.

[0134] The vehicle headlight temperature prediction module 202 is used to predict and output the trend of vehicle headlight temperature change in the future period based on the preprocessed vehicle headlight temperature data and a preset temperature prediction algorithm.

[0135] The headlight temperature control module 203 is used to determine the temperature control time based on the headlight temperature change trend, and execute a preset control algorithm based on the temperature control time to reduce the headlight temperature by controlling the headlight operating current, and the start time of the control algorithm is no later than the temperature control time.

[0136] The regulation feedback optimization module 204 is used to record the regulation process, generate regulation results, and optimize the temperature prediction algorithm and regulation algorithm based on the regulation results.

[0137] Optionally, the headlight temperature control module 203 is also used to determine the temperature control time based on the headlight temperature change trend; the temperature control time refers to the time during which the corresponding temperature data in the headlight temperature change trend exceeds a preset temperature threshold; based on the real-time acquired headlight temperature data, it analyzes whether the actual temperature data change trend deviates significantly from the headlight temperature change trend; if so, it corrects the headlight temperature change trend based on a preset correction model so that the corrected headlight temperature change trend is similar to the actual temperature data change trend; based on the corrected headlight temperature change trend, it determines and updates the temperature control time; it is also used to execute a preset control algorithm based on the real-time updated and determined temperature control time, so as to reduce the headlight temperature by controlling the working current of the headlight, and the start time of the control algorithm is not later than the temperature control time.

[0138] Optionally, the headlight temperature control module 203 is also used to calculate the rate of change of temperature data over time based on the real-time acquired headlight temperature data, and to obtain the temperature slope kc.

[0139] Determine the start time, and when the start time is reached at the current time, select a control algorithm based on the temperature slope kc and the preset first selection logic, and execute the selected control algorithm; wherein the start time is no later than the temperature control time;

[0140] The first choice logic is:

[0141] When |kc| < k1, the control algorithm is switched to a linear algorithm;

[0142] When k1≤|kc|≤k2, the control algorithm is switched to the step algorithm;

[0143] When |kc|>k2, the control algorithm is switched to an exponential algorithm.

[0144] Optionally, it also includes a necessary over-temperature determination module, used to acquire headlight operating data in historical periods, wherein the headlight operating data includes at least the headlight operating time and usage scenario; according to the usage scenario, the headlight operating time is divided into necessary headlight operating time and unnecessary headlight operating time; the distribution of necessary headlight operating time in historical periods is learned, and the distribution of necessary headlight operating time in the future is predicted; the headlight temperature change trend and the distribution of necessary headlight operating time are fused and compared to determine whether there is a target necessary headlight operating time and the determination result is output; the target headlight operating time meets the following criteria: according to the headlight temperature change trend, there is an over-temperature moment in the target headlight operating time where the temperature data exceeds a preset temperature threshold;

[0145] The headlight temperature control module 203 is further configured to, if a target necessary headlight-on period exists, calculate the temperature difference between the temperature data at the over-temperature moment and the preset temperature threshold according to a preset third selection logic, select a control algorithm and the corresponding control period based on the temperature difference and a preset second correspondence table; and when the current time reaches the control period, use the selected control algorithm to control the headlight operating current during the control period so that the headlight temperature data never exceeds the preset temperature threshold during the target necessary headlight-on period at the current time; it is also configured to, if no target necessary headlight-on period exists, determine whether there is a control result with a similar temperature slope from the control results stored in the historical period according to the preset second selection logic, and if so, select the control algorithm corresponding to the control result with a similar temperature slope as the currently selected control algorithm, determine the start time, and execute the matched control algorithm when the current time reaches the start time; and it is also configured to, if no control result with a similar temperature slope exists, select a control algorithm according to the first selection logic, determine the start time, and execute the selected control algorithm when the current time reaches the start time.

[0146] Optionally, it also includes a headlight aging assessment module, which is used to update the aging level of the headlights in real time and adjust the preset temperature threshold in real time according to the updated aging level of the headlights.

[0147] Optionally, the headlight aging assessment module is also used to analyze historical control results whenever control results are generated, determine the control frequency and heat dissipation efficiency of the headlight, predict the aging degree of the headlight based on the control frequency and heat dissipation efficiency, and update the aging degree of the headlight in real time.

[0148] Optionally, a dynamic maintenance cycle determination module is also included, which is used to periodically count the number of adjustments for each control algorithm, and output the maintenance cycle based on the statistical results and the preset maintenance analysis model, so that maintenance personnel can know it.

[0149] This application also discloses an intelligent temperature control device for vehicle lights based on adaptive learning and dynamic prediction. The intelligent temperature control device for vehicle lights based on adaptive learning and dynamic prediction includes a memory and a processor. The memory stores a computer program that can be loaded by the processor and executed as described above for the intelligent temperature control method for vehicle lights based on adaptive learning and dynamic prediction.

[0150] This application also discloses a computer-readable storage medium that stores a computer program that can be loaded by a processor and executed as described above for the intelligent temperature control method for vehicle lights based on adaptive learning and dynamic prediction. The computer-readable storage medium includes, for example, various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0151] It should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0152] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit the scope of protection of the application. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on these embodiments, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

Claims

1. An intelligent temperature control method for vehicle lamps based on adaptive learning and dynamic prediction, characterized in that, The method comprises the following steps: real-time acquisition of lamp temperature data, pre-processing of the lamp temperature data; According to the pre-processed lamp temperature data, the temperature prediction algorithm is used to predict the lamp temperature change trend in the future period; According to the lamp temperature change trend, the temperature control time is determined, and the preset control algorithm is executed according to the temperature control time to reduce the lamp temperature by adjusting the working current of the lamp, and the starting time of the control algorithm is not later than the temperature control time; Record the control process and generate the control result, and optimize the temperature prediction algorithm and the control algorithm according to the control result; The control algorithm at least includes linear algorithm, step algorithm, exponential algorithm; The linear algorithm is used to change the working current linearly with time when adjusting the working current; The step algorithm is used to change the working current in a stepwise manner when adjusting the working current; The exponential algorithm is used to change the working current in the form of exponential function with time when adjusting the working current; The execution of the preset control algorithm comprises: According to the real-time acquisition of lamp temperature data, the change rate of the temperature data with time is calculated, and the temperature slope kc is obtained; Determine the starting time, and when the current time reaches the starting time, select the control algorithm according to the temperature slope kc and the preset first selection logic, and execute the selected control algorithm; Wherein, the starting time is not later than the temperature control time; The first selection logic is: When |kc|<k1, the control algorithm is switched to linear algorithm; When k1≤|kc|≤k2, the control algorithm is switched to step algorithm; When |kc|>k2, the control algorithm is switched to exponential algorithm; The method further comprises: Obtain the lamp-on data of the lamp in the historical period, wherein the lamp-on data at least includes the lamp-on period and the use scene; According to the use scene, the lamp-on period is divided into necessary lamp-on period and unnecessary lamp-on period; Learn the distribution of necessary lamp-on period of the lamp in the historical period, and predict the distribution of necessary lamp-on period of the corresponding lamp in the future period; Fuse and compare the lamp temperature change trend and the necessary lamp-on period distribution to determine whether there is a target necessary lamp-on period and output the determination result; The target necessary lamp-on period meets: according to the lamp temperature change trend, there is an over-temperature moment in the target necessary lamp-on period, which exceeds the preset temperature threshold; The determination of the starting time, and when the current time reaches the starting time, the control algorithm is selected according to the temperature slope kc and the preset first selection logic, and the selected control algorithm is executed, comprising: If the target necessary light-on period exists, then according to the preset third selection logic, a temperature difference between the temperature data at the over-temperature time point and the preset temperature threshold is calculated, a control algorithm is selected based on the temperature difference and a preset second correspondence table, and a corresponding control period is determined; and when the current time reaches the control period, the selected control algorithm is used to control the working current of the vehicle lamp in the control period, so that the vehicle lamp temperature data does not exceed the preset temperature threshold in the target necessary light-on period; wherein, a risk coefficient of over-temperature in each necessary light-on period in a future period is calculated, if the risk coefficient is greater than a preset coefficient threshold, the corresponding necessary light-on period is determined as the target necessary light-on period, and the time when the risk coefficient is greater than the preset coefficient threshold is taken as the over-temperature time point of the target necessary light-on period; the risk level of the risk coefficient corresponding to the target necessary light-on period is determined, and the corresponding control period and control algorithm are determined; If the target necessary light-on period does not exist, then according to the preset second selection logic, it is determined from the stored control results in the historical period whether there is a control result similar to the current temperature slope, if there is, the control algorithm corresponding to the control result similar to the current temperature slope is selected as the currently selected control algorithm, the start time is determined, and when the current time reaches the start time, the control algorithm matched is executed; If there is no control result similar to the current temperature slope, then according to the first selection logic, the control algorithm is selected, the start time is determined, and when the current time reaches the start time, the selected control algorithm is executed.

2. The adaptive learning and dynamic prediction based intelligent temperature control method for vehicle lamp according to claim 1, wherein, The method further comprises: According to the vehicle lamp temperature change trend, the control time is determined, and the preset control algorithm is executed according to the control time to reduce the temperature of the vehicle lamp by controlling the working current of the vehicle lamp, comprising: According to the vehicle lamp temperature change trend, the control time is determined; the control time refers to the time corresponding to the temperature data exceeding the preset temperature threshold in the vehicle lamp temperature change trend; According to the real-time acquired vehicle lamp temperature data, it is analyzed whether the actual temperature data change trend deviates greatly from the vehicle lamp temperature change trend; If yes, the vehicle lamp temperature change trend is corrected based on a preset correction model, so that the corrected vehicle lamp temperature change trend is similar to the actual temperature data change trend; the control time is determined and updated according to the corrected vehicle lamp temperature change trend; 3.The adaptive learning and dynamic prediction based intelligent temperature control method of vehicle lamp according to claim 1, wherein, According to the real-time updated control time, the preset control algorithm is executed to reduce the temperature of the vehicle lamp by controlling the working current of the vehicle lamp, and the start time of the control algorithm is not later than the control time. The method further comprises:

4. The adaptive learning and dynamic prediction based intelligent temperature control method for vehicle lamp according to claim 3, characterized in that, The aging degree of the vehicle lamp is updated in real time, and the preset temperature threshold is adjusted in real time according to the updated aging degree of the vehicle lamp. The real-time updating of the aging degree of the vehicle lamp comprises:

5. The adaptive learning and dynamic prediction based intelligent temperature control method for vehicle lamp according to claim 1, wherein, Every time a control result is generated, the historical control results are analyzed to determine the control frequency and heat dissipation efficiency of the vehicle lamp, the aging degree of the vehicle lamp is predicted based on the control frequency and heat dissipation efficiency, and the aging degree of the vehicle lamp is updated in real time. The method further comprises: Periodically count the number of times of each regulation algorithm, and output a maintenance cycle according to the statistical result and a preset maintenance analysis model, so that a maintenance personnel can learn the maintenance cycle.

6. An intelligent temperature control system for vehicle lamps based on adaptive learning and dynamic prediction, characterized in that, The method comprises the following steps: a temperature data preprocessing module (201) is configured to acquire vehicle lamp temperature data in real time and preprocess the vehicle lamp temperature data; a vehicle lamp temperature prediction module (202) is configured to predict a vehicle lamp temperature change trend in a future period of time according to the preprocessed vehicle lamp temperature data and by using a preset temperature prediction algorithm; a vehicle lamp temperature regulation module (203) is configured to determine a temperature control time according to the vehicle lamp temperature change trend, and execute a preset regulation algorithm according to the temperature control time, so as to reduce the vehicle lamp temperature by regulating the working current of the vehicle lamp, and the starting time of the regulation algorithm is not later than the temperature control time; a regulation feedback optimization module (204) is configured to record a regulation process, generate a regulation result, and optimize the temperature prediction algorithm and the regulation algorithm according to the regulation result; the vehicle lamp temperature regulation module (203) is further configured to calculate a temperature slope kc by calculating a change rate of the temperature data with time according to the vehicle lamp temperature data acquired in real time; determine a starting time, and when the current time reaches the starting time, select a regulation algorithm according to the temperature slope kc and a preset first selection logic, and execute the selected regulation algorithm; wherein the starting time is not later than the temperature control time; the first selection logic is as follows: when |kc| < k1, the regulation algorithm is switched to a linear algorithm; when k1 ≤ |kc| ≤ k2, the regulation algorithm is switched to a step algorithm; when |kc| > k2, the regulation algorithm is switched to an exponential algorithm; a necessary period over-temperature determination module is configured to acquire vehicle lamp on-time data in a historical period, wherein the on-time data at least comprises an on-time period and a use scenario; divide the on-time period into a necessary on-time period and a non-necessary on-time period according to the use scenario; learn the distribution of the necessary on-time period in the historical period, and predict the distribution of the necessary on-time period in a future period of time; fuse and compare the vehicle lamp temperature change trend and the distribution of the necessary on-time period, determine whether there is a target necessary on-time period, and output a determination result; the target necessary on-time period satisfies: according to the vehicle lamp temperature change trend, there is an over-temperature time point in the target necessary on-time period, at which temperature data exceeds a preset temperature threshold. The vehicle lamp temperature regulation module (203) is further configured to: if there is a target necessary light-on period, calculate a temperature difference between the temperature data at the over-temperature time point and a preset temperature threshold according to a preset third selection logic, select a regulation algorithm and a corresponding regulation period based on the temperature difference and a preset second correspondence table; and when the current time reaches the regulation period, regulate the vehicle lamp working current in the regulation period using the selected regulation algorithm, so that the vehicle lamp temperature data does not exceed the preset temperature threshold during the target necessary light-on period; wherein, a risk coefficient of over-temperature in each necessary light-on period in a future period is calculated, if the risk coefficient is greater than a preset coefficient threshold, the corresponding necessary light-on period is determined as the target necessary light-on period, and the time when the risk coefficient is greater than the preset coefficient threshold is taken as the over-temperature time point of the target necessary light-on period; a risk level of the risk coefficient corresponding to the target necessary light-on period is determined, and the corresponding regulation period and regulation algorithm are determined; if there is no target necessary light-on period, whether there is a regulation result similar to the current temperature slope is determined from the stored regulation results of the historical period according to the preset second selection logic, if there is, the regulation algorithm corresponding to the regulation result similar to the current temperature slope is selected as the currently selected regulation algorithm, the start time is determined, and when the current time reaches the start time, the regulation algorithm obtained by matching is executed; if there is no regulation result similar to the current temperature slope, the regulation algorithm is selected according to the first selection logic, the start time is determined, and when the current time reaches the start time, the selected regulation algorithm is executed.

7. An intelligent temperature control device for vehicle lamps based on adaptive learning and dynamic prediction, characterized in that, A computer program is stored in a memory and loaded and executed by a processor, and the computer program comprises the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, A computer program is stored in a memory and loaded and executed by a processor, and the computer program comprises the method according to any one of claims 1 to 5.

Images