Microcontroller-based automotive cup holder temperature regulation control method

Abstract

The present application relates to the technical field of automobile cup holder temperature control method, and discloses a temperature regulation control method of automobile cup holder based on microcontroller. The method collects real-time temperature data and environmental parameters of the target container, generates an initial temperature parameter set containing a container surface temperature distribution curve and an environmental heat exchange coefficient sequence. Then, a first regulation model containing the mapping relationship between surface temperature change and refrigeration power is established according to the container surface temperature distribution curve, and a second regulation model containing the environmental heat interference and heating power compensation rule is established according to the environmental heat exchange coefficient sequence. During the cup holder temperature control, based on the real-time temperature difference fluctuation and the heat conduction characteristic data of the container material, the target model is activated to generate a dynamic power instruction, and finally the instruction is input into the semiconductor temperature control unit of the cup holder to adjust the current gradient of the refrigeration fin or the voltage distribution of the heating film. The present application can accurately and intelligently regulate the temperature of the automobile cup holder, and improve the user experience and the comfort of the automobile cabin.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile cup holder temperature control methods, in particular to an automobile cup holder temperature regulation control method based on a microcontroller. BACKGROUND

[0002] In the development process of modern automobiles, people pay more and more attention to the functionality and comfort of automobile interiors. As a common component in the automobile cabin, the function of the vehicle-mounted cup holder is also expanding. The traditional vehicle-mounted cup holder only has the basic function of placing water cups, water bottles and beverage containers, and can only provide a fixed container space for the driver and passengers, which is difficult to meet the diversified needs.

[0003] With the improvement of people's living quality, especially in long-distance driving, business travel and family outings, the demand for beverage temperature becomes diverse. For example, in winter, people expect the cup holder to heat the beverage to obtain a warm beverage experience; while in summer, they hope that the cup holder can cool down to provide a cool beverage. However, although some vehicle-mounted cup holders on the market have added simple heating and cooling functions, these functions have many shortcomings. These functions mostly rely on user-initiated operation to turn on and off, and cannot be automatically adjusted according to actual conditions. The existing cup holders lack precision and intelligence in temperature control. The traditional temperature control method is not up to the task when faced with complex and changing environmental factors during vehicle travel. The bumps during vehicle travel, the large fluctuations in external environmental temperature, and the differences in heat conduction of different materials all affect the temperature of the beverage in the cup holder, and the traditional cup holder is difficult to effectively cope with these disturbances, resulting in poor temperature regulation effect. Taking the cold and warm cup holder scheme as an example, its heating range of water temperature is limited, and the heating speed is slow. Generally, the water temperature can only be raised by about 3℃ within 10 minutes, and the maximum heating temperature can only reach about 40℃. Heating 350ml of water in a room temperature environment for 20 minutes, the water temperature increases very limitedly. Moreover, in terms of cooling, the traditional cup holder also has problems such as low cooling efficiency and unstable temperature control. In the hot summer, it is difficult to quickly cool the beverage to a suitable temperature, which cannot meet the urgent demand of users for cold drinks in a high-temperature environment. In addition, different materials of cups, such as glass cups, plastic cups, and thermos cups, have very different heat conduction characteristics, and the traditional cup holder cannot optimize the temperature according to the different materials of the container, further reducing the effect of temperature regulation and user experience. It can be seen that the existing vehicle-mounted cup holder temperature regulation method cannot meet the demand for convenient, efficient and intelligent temperature regulation, and an innovative control method is needed to solve these problems. SUMMARY

[0004] The purpose of the present application is to provide an automobile cup holder temperature regulation control method based on a microcontroller to solve the problems raised in the background.

[0005] To achieve the above object, the application provides a microcontroller-based temperature regulation control method for a car cup holder, which comprises the following steps:

[0006] collecting real-time temperature data and environmental parameters of a target container to generate an initial temperature parameter set, wherein the initial temperature parameter set comprises a container surface temperature distribution curve and an environmental heat exchange coefficient sequence;

[0007] establishing a first regulation model according to the container surface temperature distribution curve in the initial temperature parameter set, wherein the first regulation model comprises a mapping relationship between surface temperature variation and refrigeration power;

[0008] establishing a second regulation model according to the environmental heat exchange coefficient sequence in the initial temperature parameter set, wherein the second regulation model comprises a compensation rule between environmental heat interference and heating power;

[0009] activating a target model in the first regulation model or the second regulation model based on real-time temperature difference fluctuation data in a cup holder temperature control process and container material thermal conduction characteristic data, and generating a dynamic power instruction through the target model;

[0010] inputting the dynamic power instruction into a semiconductor temperature control unit of the cup holder to adjust a refrigeration fin current gradient or a heating film voltage distribution.

[0011] Preferably, the collecting of real-time temperature data and environmental parameters of a target container to generate an initial temperature parameter set comprises the following steps:

[0012] synchronously collecting temperature sampling values of each surface of the container through a multi-region temperature sensor, and analyzing original data streams of an environmental temperature and humidity sensor;

[0013] dividing the container surface into at least two thermal sections according to the spatial distribution characteristics of the temperature sampling values, and assigning each thermal section with a corresponding initial refrigeration reference value and a heating compensation coefficient;

[0014] extracting refrigeration response parameters and heating lag time constants matched with each thermal section from a historical parameter library;

[0015] performing time-domain normalization processing on the refrigeration response parameters to generate an optimized refrigeration gradient sequence corresponding to each thermal section;

[0016] performing frequency-domain decomposition processing on the heating lag time constants to generate an optimized heating response curve corresponding to each thermal section;

[0017] fusing the optimized refrigeration gradient sequence and the optimized heating response curve of each thermal section to generate the initial temperature parameter set.

[0018] Preferably, the first adjustment model and the second adjustment model are established, comprising:

[0019] The optimized refrigeration gradient sequence is input into a temperature prediction model, model weights are trained through historical temperature distribution data, and a refrigeration power mapping function in the first adjustment model is generated;

[0020] The optimized heating response curve is input into a power control model, a control threshold is adjusted through a heat conduction compensation parameter, and a heating compensation rule table in the second adjustment model is generated;

[0021] After the temperature prediction model and the power control model converge, key control vectors in the refrigeration power mapping function and the heating compensation rule table are extracted, respectively;

[0022] The key control vectors are matched with real-time collected container material heat conduction characteristic data for matching degree verification;

[0023] When the matching degree is lower than a preset threshold, the temperature prediction model and the power control model are re-iteratively trained until a convergence condition is met.

[0024] Preferably, based on real-time temperature difference fluctuation data and container material heat conduction characteristic data in the cup holder temperature control process, a target model in the first adjustment model or the second adjustment model is activated, and a dynamic power instruction is generated through the target model, comprising:

[0025] The cup holder internal temperature difference fluctuation amplitude and the container heat conduction delay time are monitored in real time;

[0026] When the temperature difference fluctuation amplitude exceeds a refrigeration activation threshold and the heat conduction delay time is in a stable interval, a refrigeration power mapping function in the first adjustment model is activated;

[0027] According to a gradient mapping rule in the refrigeration power mapping function, a dynamic current adjustment instruction of a refrigeration sheet is generated;

[0028] When the heat conduction delay time exceeds a heating activation threshold and the temperature difference fluctuation amplitude is in a stable interval, a heating compensation rule table in the second adjustment model is activated;

[0029] According to a compensation distribution rule in the heating compensation rule table, a dynamic voltage adjustment instruction of a heating film is generated.

[0030] Preferably, the dynamic power instruction is input into a semiconductor temperature control unit of the cup holder, comprising:

[0031] According to a gradient change amount in the dynamic current adjustment instruction, the driving current values of each region of the refrigeration sheet are adjusted in stages;

[0032] After each current adjustment, collect the container surface temperature feedback data and calculate the error with the predicted value of the refrigeration power mapping function;

[0033] According to the error change trend, dynamically correct the weight parameters of the refrigeration power mapping function.

[0034] Preferably, the method further comprises:

[0035] According to the voltage distribution parameters in the dynamic voltage adjustment instruction, adjust the output voltage values of the heating film in different temperature zones;

[0036] During the voltage adjustment process, the uniformity of the container temperature is detected in real time by an infrared sensor, and the compensation coefficients in the heating compensation rule table are updated according to the detection results.

[0037] Preferably, the method further comprises a feedback phase after the execution of the dynamic power instruction, comprising:

[0038] Collecting the final temperature distribution data of the container and the heating film energy consumption data;

[0039] Comparing the final temperature distribution data with the expected temperature range of the first adjustment model to generate a refrigeration model error signal;

[0040] Performing deviation analysis on the heating film energy consumption data and the reference energy consumption curve of the second adjustment model to generate a heating model error signal;

[0041] According to the system deviation component in the refrigeration model error signal, adjust the reference current gradient in the refrigeration power mapping function;

[0042] According to the energy consumption fluctuation component in the heating model error signal, optimize the voltage distribution parameters in the heating compensation rule table.

[0043] Preferably, adjusting the reference current gradient and optimizing the voltage distribution parameters comprises:

[0044] Identify the steady-state deviation in the refrigeration model error signal, and calculate the current compensation amount by a sliding mean algorithm;

[0045] Update the reference parameters of the refrigeration power mapping function according to the current compensation amount;

[0046] Identify the transient fluctuation component in the heating model error signal, and extract the effective voltage correction amount by a filtering algorithm;

[0047] Adjust the voltage distribution weight in the heating compensation rule table according to the effective voltage correction amount.

[0048] Preferably, the method further comprises, in the device initialization phase, performing:

[0049] Parsing the thermal capacity identifier in the material code of the target container and the thermal conductivity coefficient code segment, to generate a material thermal characteristic vector;

[0050] Inputting the material thermal characteristic vector into a preset material thermal characteristic database for similarity search, and screening candidate temperature control templates with a matching degree exceeding a threshold;

[0051] Selecting an optimal refrigeration reference template according to the historical control stability index of the candidate temperature control template;

[0052] Extracting a set of heating response curves associated with the optimal refrigeration reference template from the database;

[0053] Verifying the temperature-power synergy of the set of heating response curves, and eliminating abnormal curves with temperature mutation or power conflict;

[0054] Aligning the verified heating response curves with the optimal refrigeration reference template in time sequence, to generate a reference configuration of an initial temperature parameter set.

[0055] Preferably, the temperature-power synergy verification of the set of heating response curves comprises:

[0056] Extracting heating power distribution data of a single to-be-verified curve, and synchronously acquiring a refrigeration power sequence of a corresponding time node in the optimal refrigeration reference template;

[0057] According to the phase node of the refrigeration power sequence, labeling a synergy timestamp on the heating power distribution data;

[0058] Detecting whether the power change slope of the adjacent interval of the synergy timestamp exceeds a mutation threshold;

[0059] When a power conflict is detected, generating a replacement power smoothing segment based on a historical conflict correction record and updating the curve.

[0060] Compared with the prior art, the present application has the following beneficial effects:

[0061] The method can realize accurate temperature regulation. By collecting real-time temperature data of the target container, an initial temperature parameter set containing the container surface temperature distribution curve is generated, and then a first regulation model is established, and the mapping relationship between the surface temperature change and the refrigeration power is determined. This means that during the refrigeration process, the system can accurately adjust the refrigeration power according to the actual change of the container surface temperature. For example, when the container surface temperature drops too quickly and exceeds the preset appropriate temperature change range, the system can quickly reduce the refrigeration power to avoid the beverage temperature being too low. Conversely, if the temperature drops too slowly, the system will automatically increase the refrigeration power to speed up the refrigeration speed and ensure that the beverage temperature can stably approach the user's desired temperature value. Similarly, for the heating process, the second regulation model established according to the environmental heat exchange coefficient sequence contains the compensation rules of environmental heat interference and heating power. In cold winter, the low-temperature environment outside the vehicle during driving will have a greater heat interference on the beverage in the cup holder. At this time, the system can automatically increase the heating power according to the environmental heat exchange coefficient to compensate for the heat loss caused by environmental heat interference and maintain the stability of the beverage temperature.

[0062] The method has intelligent dynamic regulation capability. During the cup holder temperature control process, the system will monitor the temperature difference fluctuation data in real time and intelligently activate the target model in the first regulation model or the second regulation model in combination with the container material thermal conductivity characteristic data to generate a dynamic power instruction. Containers of different materials, such as glass, plastic, and stainless steel, have significant differences in thermal conductivity characteristics. Glass containers have fast thermal conductivity, plastic containers have relatively slow thermal conductivity, and stainless steel containers have unique thermal conductivity characteristics. When the user places containers of different materials in the cup holder, the system can automatically identify the container material and flexibly adjust the temperature regulation strategy according to the thermal conductivity characteristic data of the container material. For example, for glass containers with fast thermal conductivity, the system will appropriately increase the power regulation amplitude during heating or refrigeration to quickly reach and maintain the target temperature. For plastic containers with slow thermal conductivity, a relatively gentle power regulation method is adopted to avoid excessive or insufficient temperature regulation. At the same time, the monitoring of real-time temperature difference fluctuation data enables the system to respond to temperature changes in a timely manner. During vehicle driving, the ambient temperature around the cup holder may fluctuate due to factors such as direct sunlight and air conditioner outlet position. The system can quickly adjust the power instruction according to these real-time changes to ensure that the beverage temperature in the cup holder is not disturbed by external interference and always remains stable. This intelligent dynamic regulation capability enables the cup holder to adapt to various complex and variable actual use scenarios and provides users with a more convenient and intelligent temperature regulation experience.

[0063] The application of the method helps to improve the overall comfort and user experience of the automobile cabin. During long-distance driving, the user does not need to manually adjust the temperature of the cup holder frequently. The system can automatically adjust the temperature of the beverage intelligently and accurately according to various factors, so that the user can enjoy the beverage at the appropriate temperature at any time, reducing the distraction operation during driving and improving the driving safety. For business users, they can drink coffee or tea at the right temperature at any time in the car, which helps to improve the atmosphere and efficiency of business negotiations. For family outing scenarios, parents can conveniently provide children with beverages at the appropriate temperature, ensuring the health and comfort experience of children's diet. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 The working principle diagram of the microcontroller-based automobile cup holder temperature adjustment control method described in the application;

[0065] Figure 2 The flowchart of multi-region parameter acquisition and processing;

[0066] Figure 3 The flowchart of model training and verification;

[0067] Figure 4 The flowchart of feedback stage error signal processing. DETAILED DESCRIPTION

[0068] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0069] Please refer to Figure 1 The application provides a microcontroller-based automobile cup holder temperature adjustment control method, which comprises:

[0070] Precise temperature control of the container is achieved by integrating temperature sensing, environmental monitoring, and intelligent algorithms. The system first collects surface temperature distribution data of the target container using a multi-region temperature sensor array, simultaneously acquiring ambient temperature and humidity parameters to construct an initial temperature parameter set containing the container surface temperature distribution curve and the ambient heat exchange coefficient sequence. Based on this set, two independent control models are established: the first model generates a cooling control strategy by analyzing the mapping relationship between container surface temperature changes and cooling power; the second model forms a heating control strategy based on compensation rules for ambient thermal interference and heating power. The system monitors temperature fluctuations and heat conduction characteristics in real time, dynamically selecting and activating the corresponding control model, generating execution commands for the semiconductor temperature control unit, and achieving precise temperature control by adjusting the current gradient of the cooling chip or the voltage distribution of the heating film.

[0071] Example 1: See Figure 2 During the generation of the initial temperature parameter set, the system achieves accurate modeling of the container's temperature characteristics through multi-sensor collaborative acquisition and data processing. The temperature acquisition module employs six high-precision digital temperature sensors arranged in a ring, each mounted at a fixed spatial angle on the inner wall of the cup holder, forming a full-coverage monitoring network covering the container's sidewalls. The sensors are I2C-interface digital devices with built-in 16-bit ADC converters, achieving a temperature resolution of 0.015℃, and a sampling frequency of 10Hz to ensure real-time dynamic response. Each sampling point records not only the temperature value but also includes a three-dimensional spatial coordinate marker, forming a temperature data package with location information. Environmental parameter monitoring utilizes an integrated temperature and humidity composite sensor, with a humidity measurement range covering 5% to 95% RH and a temperature measurement accuracy of ±0.3℃. Hardware filtering circuitry eliminates signal noise caused by vehicle vibration.

[0072] The microcontroller polls data from each sensor via a multiplexer, employing a timestamp synchronization mechanism to ensure the temporal consistency of data across channels. After undergoing moving average filtering, the raw data enters the spatial temperature distribution reconstruction algorithm. This algorithm, based on the inverse distance weighted interpolation principle, calculates the estimated temperature values ​​of unmonitored areas on the container surface according to the spatial relationships of the sensors, generating a complete two-dimensional temperature field distribution map. The environmental parameter analysis module separates three key indicators from the raw data stream: absolute humidity, dew point temperature, and dry-bulb temperature, and calculates the current air heat exchange coefficient using a lookup table method.

[0073] The thermal zone division algorithm adopts a gradient-based edge detection technique to analyze the slope variation characteristics of each region in the temperature distribution map. When the temperature variation rate difference between adjacent regions exceeds the set threshold, the automatic division boundary line is generated. In typical applications, the container surface is divided into three independent control sections, top, middle and bottom, each with different thermodynamic characteristics. The top section is directly exposed to the air, and the environmental heat exchange effect is significant; the middle section is affected by the container wall thickness and liquid convection, and the thermal inertia is large; the bottom section is in contact with the cup holder base, and conduction heat dissipation dominates. The system assigns each section an independent control parameter set, including the reference cooling intensity, heating compensation coefficient and temperature response time constant.

[0074] The historical parameter library adopts a hierarchical storage structure, with the upper layer storing standard parameter templates for common container types and the lower layer recording adaptive adjustment data in actual operation. When retrieving data, first match the container material type and volume specification, and then filter the historical records under similar working conditions according to the current environmental conditions. The extraction process of refrigeration response parameters includes time domain feature analysis, the system identifies the inflection point position of the temperature drop curve in the historical data, and calculates the time constant and power variation rate of each stage. The analysis of heating lag characteristics focuses on the delay time and overshoot in the temperature rise process, and through curve fitting, the parameter estimation value of the first-order inertia link is obtained.

[0075] The time domain normalization processing module adopts the sliding window Fourier transform technique to convert the non-stationary refrigeration response signal into time-frequency joint domain features. In the processing process, a 256-point Hanning window function is set, and the signal segment is intercepted with a 50% overlap rate, and the energy distribution characteristics of each frequency band are extracted by fast Fourier transform. The generated optimized refrigeration gradient sequence contains a 32-dimensional feature vector, each feature corresponds to the energy integral value of a specific frequency band, and is arranged in time order to form a time spectrum diagram of refrigeration control. The frequency domain decomposition of the heating lag parameter uses the wavelet packet transform algorithm, and the db4 wavelet basis function is selected for 6-layer decomposition to accurately locate the feature frequency band in the thermal response process. The energy values of each node obtained by decomposition constitute the multi-scale description features of the heating response curve, which are weighted and fused with the thermal conductivity of the container material.

[0076] The improved Kalman filter algorithm is used in the parameter fusion stage to make optimal estimation of the time-frequency domain features and the real-time collected physical quantity measurement data. The state equation of the filter contains the physical model of temperature conduction, and the observation equation integrates the actual measurement values of each sensor. In each iteration process, the system dynamically adjusts the process noise covariance matrix to adapt to the characteristic changes of different thermal sections. The generated initial temperature parameter set is stored in a structured data format, including time series labels, spatial distribution matrices, and environmental correction coefficients. The time series labels record the time stamp and duration of each control period; the spatial distribution matrix stores the temperature gradient values and heat flux density of each thermal section; and the environmental correction coefficient contains the compensation values of the air convection coefficient and the radiation heat transfer coefficient.

[0077] The data verification module performs integrity checking after the parameter set is generated, ensuring that the data packet is error-free during transmission through the cyclic redundancy check. The abnormality detection algorithm analyzes the logical relationship between the parameters of each thermal section. When an abnormal combination that violates the laws of thermodynamics is found, the data re-sampling process is automatically triggered. The parameter set that passes the verification is sent to the double-buffer memory for real-time calling by the subsequent modeling module. The entire acquisition and processing process is completed within a 20 ms control period, meeting the strict real-time requirements of automotive electronic systems.

[0078] The system maintenance mechanism periodically performs self-diagnosis on the sensor network by injecting test signals to detect the response characteristics of each channel. The drift error of the temperature sensor is eliminated by the two-point calibration method, and the periodic calibration of the environmental sensor is automatically triggered based on the climate control parameters on the vehicle CAN bus. The update strategy of the historical database uses incremental learning, which only triggers database structure expansion when the feature distribution of newly collected data exceeds the original range. This design ensures the adaptive ability of the system while avoiding unlimited growth of storage space.

[0079] The hardware acceleration module plays an important role in the critical path of data processing. Temperature interpolation calculation is accelerated by the geometry processor unit, and Fourier transform and wavelet decomposition are completed by calling the dedicated digital signal processing coprocessor. This heterogeneous computing architecture reduces the load of the main control chip by about 40%, ensuring stable operation of the system under complex working conditions. The communication interface uses the automotive CANFD bus protocol, and the transmission bandwidth of the parameter set reaches 5 Mbps, supporting data sharing with other thermal management units such as the vehicle air conditioning system.

[0080] In terms of fault-tolerant mechanism design, the system automatically switches to a degraded mode based on the remaining sensors when a single sensor failure is detected. The hot section partitioning algorithm uses nearest neighbor interpolation to compensate for missing data at this time, and the abnormal state is prompted on the diagnostic interface with a red marker. When the environmental sensors fail, the system switches to a preset typical working condition parameter set, and the environmental heat exchange conditions are inversely calculated by monitoring the cup holder temperature change trend. These designs significantly improve the robustness and usability of the system.

[0081] The quality of the initial temperature parameter set directly affects the subsequent control effect, so the system sets up a multi-level verification mechanism. In the data acquisition stage, the sampling values of the analog channels need to pass the reasonableness test, and the abnormal values outside the physical possible range are removed; in the feature extraction stage, the intermediate results of each algorithm module are output to the debugging interface for offline analysis and verification; after the final parameter set is generated, it is cross-validated with the working condition data recorded by the on-board diagnostic system. These measures together constitute a complete quality assurance system, providing a reliable input basis for temperature regulation control.

[0082] Embodiment 2: see Figure 3 The first adjustment model is based on a deep neural network architecture, and its core is to establish a nonlinear mapping relationship between the temperature change of the container surface and the refrigeration power. The input layer of the network receives the optimized refrigeration gradient sequence generated in Embodiment 1, which contains 32-dimensional time-frequency domain feature vectors, and each feature dimension represents the energy integral value of a specific frequency band. The hidden layer is designed as a recursive structure containing 128 long short-term memory (LSTM) units, and the internal state update of the unit follows the principle of the gating mechanism. The forgetting gate controls the retention degree of historical information, the input gate adjusts the proportion of new information, and the output gate determines the external influence intensity of the current state. This structural characteristic enables the network to effectively capture the time sequence dependence of temperature changes, especially the thermal inertia effect and delay response characteristics in the refrigeration process.

[0083] The root mean square back propagation algorithm (RMSprop) is used to optimize the weight parameters in the network training stage, and the loss function is defined as the mean square error of the predicted temperature and the actual temperature. The training data set is composed of historical running records, each sample contains an input feature vector, a temperature distribution true value corresponding to the time stamp, and an environmental condition label. The input features are standardized in the data preprocessing link, so that their numerical values are distributed in the [-1, 1] interval. To prevent overfitting, the network introduces Dropout regularization technology, which randomly discards part of the neuron connections with a probability p dr = 0.2 during the training process, where p dr represents the probability of neuron being discarded. The training process is set with an early stopping mechanism, which terminates iteration when the validation set error does not decrease for 5 consecutive epochs, and the prediction error of the finally converged network is controlled within the engineering allowable range.

[0084] The implementation of the second regulation model relies on a fuzzy logic control framework, whose core is to transform the heating compensation problem into a multi-dimensional decision-making process. The input variables are defined as the heat conduction delay time and the temperature difference fluctuation amplitude, which are fuzzified into three linguistic variables of "low", "medium", and "high" through membership functions. The output variable is the heating film voltage adjustment, which is divided into 7 discrete levels. The rule base contains 27 control rules in the form of "if the delay time is high and the temperature difference is medium, then the voltage adjustment is level 4". The defuzzification process uses the gravity method to calculate the exact output value, generating the final heating compensation rule table. This table is stored in the form of a two-dimensional matrix, with the row index corresponding to the discretized heat conduction delay time, the column index representing the temperature difference fluctuation interval, and the matrix elements storing the recommended voltage adjustment level.

[0085] The model parameter adjustment phase introduces the heat conduction compensation coefficient α hc , which dynamically modifies the effect of the control rules according to the container material characteristics. hc The calculation formula of α

[0086]

[0087] where κ mat represents the measured thermal conductivity of the current container material, κ std is the reference thermal conductivity under standard conditions, ρ env is the measured value of the ambient air density, and ρ nom is the nominal air density. This coefficient is used to scale the amplitude of the voltage adjustment levels in the rule table, appropriately reducing the heating intensity when the material's thermal conductivity is better than the standard value, and increasing the compensation amount when the air density is higher to offset the effect of convective heat dissipation.

[0088] The model convergence verification uses a two-stage strategy. The first stage checks the monotonicity of the refrigeration power mapping function, requiring the output refrigeration power to strictly increase when the input temperature rises. This property is verified by traversing discrete test points within the function's domain. The second stage evaluates the completeness of the heating compensation rule table, ensuring that all possible input combinations can find corresponding entries with non-zero membership in the rule table. When boundary conditions are missing during the verification process, a rule interpolation mechanism is automatically triggered, generating new control entries based on the weighted average of adjacent rules.

[0089] The extraction process of the key control vectors focuses on the core decision features of the model. For the refrigeration power mapping function, the partial derivative matrix of its output with respect to the input features is extracted, and the top 3 feature dimensions with the largest variation rates are selected as the dominant control vectors. The heating compensation rule table then counts the frequency of occurrence of each voltage level, and rules with a usage frequency exceeding a threshold and a large adjustment amplitude are marked as key control points. When these vectors are matched with the real-time collected container material data, a distance measurement system for the feature space is established. The material data is obtained through an embedded RFID reader, decoded and converted into a 16-dimensional material feature vector, and the cosine similarity is calculated in the same measurement space with the model control vectors. When the similarity is lower than the preset threshold, the system records the current working condition as a new type of working mode and starts the model incremental learning process.

[0090] The model retraining mechanism uses the elastic weight consolidation algorithm to adapt to new working conditions while preserving existing knowledge. Before each retraining, the importance scores of the weights of each layer of the network are calculated, and important parameters are protected during subsequent training by adding a constraint term. The update of the heating compensation rule table uses a gradual adjustment strategy, modifying only rule entries that differ significantly from the current working condition. This design ensures the adaptability of the model while avoiding control instability caused by frequent adjustments. The retraining process is executed in the background thread, and after completion, the new model version is verified through at least 10 control cycles of shadow running, and only after confirming performance improvement is it switched to online control.

[0091] During real-time control, the model selection logic is based on dual judgment conditions. The temperature difference fluctuation monitoring module calculates the root mean square value of the temperature change of each thermal section within adjacent sampling periods, and when this value exceeds the refrigeration activation threshold, the first adjustment model is called preferentially. The heat conduction delay time is estimated by analyzing the initial slope of the temperature response after the heating instruction is issued, and when the delay exceeds the set threshold, the second adjustment model is enabled. The outputs of the two models are coordinated in the time and space dimensions, with refrigeration power adjustment mainly affecting the temperature rapid change stage and heating compensation focusing on maintaining steady-state accuracy. When the system detects a conflict between the control instructions of the two models, the arbitration algorithm is started, and the dominant control mode is selected based on the degree and direction of the current temperature deviation from the target value.

[0092] The model performance maintenance system continuously monitors the control effect, recording temperature tracking errors and energy consumption indicators. The error analysis module aligns the actual temperature trajectory with the model prediction curve using dynamic time warping, calculates the deviation at each time point to form an error distribution map. The energy consumption statistics unit accumulates the current and voltage integral values of the semiconductor temperature control unit to generate a power-time curve. These data are fed back to the model parameter optimization loop regularly, forming a closed-loop learning system. The anomaly detection algorithm scans for outliers in the error distribution, and when a persistent deviation pattern is found, a diagnostic report is automatically generated, indicating possible model degradation or hardware failure.

[0093] Hardware acceleration, neural network inference process is deployed on a dedicated AI accelerator, using matrix multiplication instruction set to realize parallel computing. The rule evaluation of fuzzy logic control is realized through lookup table, which discretizes the continuous input into 256 index to access the precomputed results directly. These optimizations control the joint inference time of the two models within 5ms, meeting the real-time requirements of automotive electronic systems. Memory management uses block caching strategy to save frequently accessed model parameters in cache area, reducing main memory access delay.

[0094] The safety mechanism design ensures that the model output is within a physically reasonable range. The refrigeration power instruction is saturated and limited within the safe working area of semiconductor devices. The heating film voltage adjustment amount is set to limit the rate to avoid thermal stress on the container caused by sudden temperature changes. All model output instructions are checked by an independent safety monitoring unit, and instructions that violate the constraints will be intercepted and trigger a system safety state transition. Model version management uses blockchain technology to record each update, supporting fast rollback to a stable version in case of failure.

[0095] Integration with the vehicle system is achieved through a standardized interface. The model receives environmental temperature, solar intensity and other auxiliary information from the vehicle bus, dynamically adjusting the control parameters. At the same time, the cup holder temperature state is uploaded to the vehicle infotainment system, supporting user interface display and voice prompt functions. The diagnostic interface supports engineers to read the internal state of the model through a special tool, including neuron activation patterns, rule trigger frequency and other deep information, providing data support for on-site troubleshooting.

[0096] Example 3: The temperature difference fluctuation monitoring module uses a differential signal processing architecture to detect abnormal changes by comparing the real-time readings of adjacent temperature sensors. The hardware circuit design includes six parallel instrumentation amplifiers, each connected to a pair of spatially adjacent sensors, amplifying the microvolt-level temperature difference signal to a processable range. The signal conditioning stage uses a second-order Butterworth filter with a cutoff frequency of 0.5Hz to suppress high-frequency noise while preserving the dynamic characteristics of real temperature changes. When the differential voltage of any channel exceeds the threshold V th , an interrupt signal is triggered to notify the microcontroller to start the refrigeration regulation process. The threshold voltage V th is dynamically calculated by the following formula:

[0097] V th = G amp · S sens · ΔT set

[0098] Where: G amp represents the gain coefficient of the instrumentation amplifier (typical value 500), S sens is the sensitivity of the temperature sensor (unit mV / ℃), and ΔT setThe threshold value is self-adaptive to the ambient temperature, which relaxes the fluctuation tolerance in high-temperature environment and improves the monitoring sensitivity in low-temperature condition.

[0099] The refrigeration film current control adopts a three-stage gradient regulation strategy, each stage corresponding to a different thermodynamic state. The initial stage increases the driving current at a preset slope, and the slope value is obtained from a table according to the heat capacity characteristics of the container material. During the current rising process, the microcontroller outputs an analog control signal through a 16-bit DAC, which drives the semiconductor refrigeration film through a power amplification circuit. The current monitoring uses a Hall effect sensor, and the sampled data is compared with the set value after ADC conversion to form a closed-loop control. In the stable stage, a fuzzy PID algorithm is introduced, and the proportional coefficient K p The integral time T i and the differential gain K d are nonlinearly adjusted according to the temperature difference error, and are dynamically optimized according to the temperature change rate. The final stage uses pulse width modulation technology, with a switching frequency set to 25kHz to avoid audible noise, and a duty cycle adjustment step of 0.5% to achieve fine temperature stabilization.

[0100] The temperature feedback verification system is started after each current adjustment, with a collection period set to 200ms to match the time constant of heat conduction. Multi-channel sensor data is captured through a synchronous sampling hold circuit to eliminate time deviation between channels. The data verification algorithm first eliminates outliers caused by electromagnetic interference, then calculates the temperature average and standard deviation of each thermal section. The prediction error evaluation uses the weighted residual sum of squares method, giving higher weight to the bottom section to reflect its key influence on the overall temperature. When the error exceeds the limit, the parameter correction process is triggered, and the weight matrix of the refrigeration power mapping function is updated iteratively through the conjugate gradient method, with each adjustment amplitude limited to within 5% to avoid system oscillation.

[0101] The heating film control uses spatial zoning voltage regulation technology, dividing the flexible heating film into eight independently controlled fan-shaped regions. Each zone is equipped with a dedicated drive circuit, including a MOSFET switch tube and a current detection resistor. The voltage distribution algorithm is based on the temperature field data input by the infrared thermal imager, calculating the compensation voltage value of each region. The Gaussian smoothing algorithm is used for boundary processing to avoid temperature unevenness caused by voltage sudden change in adjacent regions. The calculation of voltage regulation amount ΔV reg considers the local temperature difference ΔT local and the environmental heat flux density φ env :

[0102] ΔV reg = K v ·(ΔT local +C comp ·φ env )

[0103] where: K v is the voltage-temperature conversion coefficient (unit V / °C), C comp is the ambient compensation factor. This formula ensures automatic enhancement of heating power under strong convection conditions to offset additional heat loss.

[0104] The infrared temperature detection system uses a 32x32 pixel thermopile array with a spatial resolution of 1.5mm and a temperature refresh rate of 5Hz. The image processing algorithm first performs non-uniformity correction, and calibrates the response curve of each pixel point by referencing a blackbody. Temperature uniformity analysis extracts five characteristic parameters: the difference between the highest and lowest temperatures, the hotspot position offset, the radial temperature gradient, the circumferential asymmetry, and the overall standard deviation. These parameters are input into the query engine of the heating compensation rule table to generate targeted voltage adjustment instructions. After the instructions are executed, the system monitors the evolution trend of the temperature field, and if the uniformity improvement is insufficient, it triggers a secondary compensation cycle.

[0105] Hardware protection mechanisms run throughout the entire control process. The refrigeration sheet drive circuit includes an overcurrent protection chip that automatically cuts off the power supply when the current exceeds the safety threshold. Independent temperature fuses are set for each partition of the heating film to prevent local overheating from causing material aging. The power device heat dissipation uses a design combining heat pipes with fins to ensure that the junction temperature does not exceed the rated value under maximum load. All analog signal channels are configured with TVS diodes for protection, suppressing power fluctuations and electrostatic discharge interference.

[0106] The timing synchronization system ensures that all subsystems work in coordination. The main controller generates a reference clock signal, which is distributed to each functional module through the LVDS interface. Critical operations such as current adjustment and temperature sampling are strictly aligned to the time slot window to avoid resource conflicts caused by concurrent operations. The exception handling uses a priority interrupt mechanism, with the temperature difference overrun alarm having the highest priority, immediately suspending other task execution to perform the emergency temperature control program.

[0107] The dynamic parameter storage management system maintains three sets of operating parameters: factory default values, adaptive learning values, and temporary adjustment values. When the system detects a container change, it automatically loads the corresponding material's reference parameters; during stable operation, it gradually replaces them with learning-optimized values; and in the event of sudden working conditions, it uses temporary adjustment parameters for rapid response. Parameter version management uses the copy-on-write technology to ensure the atomicity and traceability of the parameter update process.

[0108] Interaction with the vehicle system is achieved through the CANFD bus. The cup holder controller sends temperature status messages regularly (period 100ms) and receives environmental parameter updates from the air conditioning system (period 1s). When the vehicle enters energy-saving mode, the temperature control system automatically switches to a low-power strategy, limiting the maximum operating current and relaxing the temperature control accuracy requirements. User interface integration is achieved through the AUTOSAR architecture, supporting the display of real-time temperature curves and energy consumption statistics on the central control screen.

[0109] The diagnostic function design contains both online self-check and offline analysis modes. The power-on self-check procedure verifies the health status of all sensor channels and drive circuits, and the ECC error correction capability of the memory is tested periodically during operation. Fault records are stored in a circular buffer, saving detailed data of the last 100 abnormal events, including temperature profile, control instruction sequence, and environmental snapshot at the time of occurrence. Engineers can inject test mode instructions through the diagnostic interface to simulate various boundary conditions and verify the system robustness.

[0110] The thermodynamic model online update mechanism allows the import of verified new control algorithms. The update package uses digital signature to ensure the source is trustworthy, and AES-256 encryption is used to protect the transmission process. The model switching uses a fade-in and fade-out strategy, with the new and old versions running in parallel for a period of time, gradually transferring control to avoid step disturbance. The version rollback function can quickly recover to a stable state when the new model performs poorly.

[0111] The manufacturing consistency guarantee measures include an end-to-end traceability system. Each cup holder module undergoes 72-hour aging test before leaving the factory, recording its baseline performance curve under different working conditions. Key parameters such as sensor sensitivity and drive circuit gain are written to one-time programmable memory to prevent later tampering. The batch management database associates material sources, production processes, and test results, supporting root cause analysis of quality problems.

[0112] The environmental adaptability design takes into account the harsh conditions of automotive applications. The circuit board is encapsulated with automotive-grade epoxy resin, capable of withstanding temperature cycles from -40°C to 125°C. The connector has an IP67 protection level to prevent car wash liquid from seeping in. All software algorithms are verified through -40°C low-temperature startup and 85°C high-temperature full-load operation tests. The electromagnetic compatibility design includes multi-layer board stacking, magnetic bead filtering, and shielded housing, meeting the CISPR25 Class 5 radiation standard.

[0113] The energy consumption optimization strategy is implemented throughout the system. The refrigeration sheet drive uses a zero-voltage switching topology to reduce switching loss; the heating film is powered in partitions and dynamically started and stopped according to demand; the microcontroller switches to low-power mode during idle periods. The energy recovery circuit stores the reverse thermoelectric conversion energy generated during braking in a super capacitor. The system calculates and accumulates energy consumption indicators in real time to provide decision-making basis for vehicle energy management.

[0114] The life cycle management includes performance degradation monitoring functions. The Seebeck coefficient of the refrigeration sheet is recorded over time to predict the remaining service life. The resistance value of the heating film is measured periodically, and maintenance recommendations are prompted when the resistance value changes by more than 10%. The equivalent series resistance (ESR) of key components such as electrolytic capacitors is continuously monitored, and an alarm is triggered for preventive maintenance when it abnormally rises. These data are uploaded to the cloud through the Internet of Vehicles to support predictive maintenance and supply chain optimization.

[0115] Example 4: Refer to Figure 4 The data processing of feedback stage starts from the acquisition of final temperature distribution of the container. The system initiates multi-channel synchronous sampling after the end of the refrigeration or heating cycle, and the six sensors record the stable values of the top, middle and bottom sections respectively. The temperature data is transmitted to the analysis module through the DMA channel, and the energy consumption monitoring circuit records the cumulative charge of the semiconductor temperature control unit with a resolution of 0.1 mAh.

[0116] The data preprocessing includes outlier rejection and unit unification. For the cooling scenario of aluminum beverage cans, assuming the target temperature is set to 5°C, the system records the final temperature distribution of a certain run, refer to Table 1.

[0117] Table 1: Final temperature distribution of a certain run.

[0118]

[0119]

[0120] The refrigeration model error analysis uses Gaussian blur processing technology. The system first establishes a tolerance interval for the target temperature, and constructs a Gaussian distribution curve with the target value as the center and the positive and negative tolerances. For the 5.8°C measured value of the top section, calculate the Mahalanobis distance outside the target interval [4.5°C, 5.5°C], quantify the deviation and generate a deviation signal. The deviation signals of the three sections are fused into the total error value of the refrigeration model, and are decomposed into two components of steady-state offset and dynamic fluctuation.

[0121] The energy consumption analysis module uses dynamic time warping algorithm. Take the current run's current-time series and the reference curve for nonlinear alignment, calculate the power spectral density difference after eliminating the time delay effect. When energy saving is detected in the middle section, extract the negative fluctuation component in the frequency range of 0.05-0.1 Hz, and mark it as an abnormal energy saving mode.

[0122] The reference current gradient adjustment uses adaptive filtering technology. The system identifies a +0.3°C steady-state offset in the top section, automatically calculates the compensation amount: the sliding window is set to 10 historical periods (200ms / period), and the offset mean of the last three periods is taken as the effective compensation value. In the case of aluminum can refrigeration, the compensation module increases the reference current intensity of the refrigeration mapping function by 5mA, and records the special tuning parameters of this material under the current environment simultaneously.

[0123] The voltage distribution optimization process transient fluctuation component. The bottom section of the energy consumption deviation Hilbert transform, extract the instantaneous amplitude characteristics at 0.15 Hz, determine the dominant mode of transient interference. Voltage weight adjustment using asynchronous zoning strategy: high temperature area compensation coefficient increases 0.15, low temperature area remains unchanged. New parameters are written to the backup sector of FLASH memory in real time, and hot update is realized through double buffering mechanism.

[0124] Data reliability safeguards include a triple check mechanism. Temperature data performs CRC16 check after transmission, and triggers re-sampling process when abnormal. Energy consumption data eliminates gain error through analog-to-digital converter self-calibration function, and the system injects test current regularly to verify the accuracy of the measurement channel. Parameter update process uses version transaction management, and automatically rolls back to the last stable version when any step fails.

[0125] The environmental disturbance compensation strategy is associated with vehicle state data. When the monitored solar intensity during this run reaches 800 W / m 2 2, the system automatically amplifies the compensation amount of the top temperature zone by 20%. If the vehicle is in a driving vibration environment, the bottom temperature zone increases the anti-interference coefficient by 0.1. These associated parameters form an environmental feature matrix, which is stored synchronously with the error signal.

[0126] The model parameter correlation update mechanism establishes a temperature-energy consumption double closed loop. After adjusting the slope of the mapping function of the top error, the refrigeration model is optimized simultaneously. The heating model updates the heat conduction delay parameter according to the bottom energy consumption fluctuation characteristics. The cross-correction module scans the physical constraint relationship between parameters regularly, such as the power conservation principle requires that the sum of the three section energy consumption correction amounts tends to zero.

[0127] The fault recovery system starts deep diagnosis when it detects continuous abnormalities. When the error of a certain temperature zone exceeds the set threshold for three consecutive times, the system automatically loads backup parameters and runs, and performs hardware self-detection sequence at the same time. The detection items include sensor impedance measurement, refrigeration plate junction temperature estimation and driving circuit conduction test. After generating the diagnosis report, the automatic optimization scheme is transmitted to the cloud analysis platform through OTA.

[0128] The industrial implementation example shows the whole process operation. During the heating process of the cylindrical stainless steel thermos, the system detects that the bottom temperature always lags behind the target value by 3℃. The analysis stage identifies the heat conduction delay characteristics of this material, and the error component shows that the steady-state deviation is 0.7℃. The parameter update link increases the voltage weight of the bottom heating zone by 22%, and prolongs the compensation gradient time by 300ms. The subsequent running data shows that the temperature tracking accuracy is improved, and the energy consumption deviation converges to the allowable range.

[0129] The historical data intelligent analysis engine automatically mines correction laws. The system screens successful correction cases of typical containers such as aluminum cans, glass cups, and vacuum cups, and establishes a material-environment-parameter three-dimensional knowledge base. When processing a new type of bamboo fiber material container, the system matches similar cases in the knowledge base according to the thermal conductivity coefficient 0.18 W / m·K, and initializes the optimization parameter to 85% of the ceramic material, avoiding starting from zero.

[0130] The operation monitoring system generates a visual report interface. Engineers can view the temperature-energy consumption hyperbolic curve with timestamps through the diagnostic device, where the benchmark curve is marked with a blue dashed line, and the actual trajectory is drawn with a red solid line. The parameter modification record is displayed below the curve, marking the specific value and time point of each adjustment. The historical operation log can be searched by date, container type, environmental conditions, and other dimensions.

[0131] The life cycle maintenance function records the performance degradation of key components. The Seebeck coefficient of the refrigeration sheet is automatically measured every 100 hours of operation, and the change curve is stored in the database. When the coefficient is detected to decrease by more than 8% of the initial value, the system automatically increases the driving current by 10% to compensate while maintaining the target temperature, and prompts a device aging alarm on the maintenance interface.

[0132] The system integration verification passes the vehicle vibration bench test. Under the condition of 5-500Hz random vibration, the feedback data processing module uses a time sequence latching mechanism to ensure data integrity. The electromagnetic compatibility test shows that after the variable frequency drive noise is eliminated by a band-stop filter, the temperature acquisition system has a measurement error of less than 0.1°C under a field intensity of 30V / m. These designs ensure reliable operation in harsh environments.

[0133] Example 5: The device initialization phase begins with the container identity recognition process. When the container is placed in the cup holder, the RFID sensing coil at the bottom of the cup holder activates the container tag and reads its data stream of material code. The code structure contains a 24-bit identifier, the first 8 bits are the material category code, the middle 8 bits store the heat capacity characteristic value, and the last 8 bits record the thermal conductivity coefficient range. The decoder converts the physical parameters into a feature vector, where the thermal conductivity coefficient code segment is processed by a floating point conversion engine to generate a standard format thermal conductivity characteristic value. The heat capacity identifier is mapped to a dimensionless thermal inertia index, and the two together form a 16-dimensional material thermal characteristic vector, each dimension representing the thermal response characteristics of the material in a specific temperature range.

[0134] The material thermal property database stores data using a graph neural network architecture, with nodes representing different material types and edge weights reflecting the similarity of thermal properties. The retrieval process first locates the current material vector in the feature space, and performs a nearest neighbor search using cosine similarity as the metric. The algorithm automatically ignores candidate records with a similarity of less than 90%, and ranks the candidate set in descending order of similarity. The retrieval mechanism adds an environmental compensation factor, which automatically relaxes the similarity threshold by 5 percentage points when the vehicle is in high-altitude areas, adapting to changes in thermal conductivity characteristics under low pressure conditions.

[0135] The selection of candidate temperature control templates is based on multi-dimensional stability indicators. The system calculates three core parameters for each candidate template in historical operation: temperature control standard deviation, power output fluctuation rate, and response time dispersion. Temperature standard deviation reflects long-term control accuracy, power fluctuation rate represents energy efficiency, and response time dispersion measures system robustness. The multi-objective decision-making process uses an improved entropy weight TOPSIS algorithm to normalize and weight the three indicators. The weight coefficient dynamic adjustment mechanism automatically updates according to the current environmental temperature change rate, giving more weight to the response time indicator when the temperature changes rapidly, and prioritizing control accuracy in stable environments. After selecting the optimal refrigeration reference template, the system automatically checks its compatibility with the current hardware version to exclude potential risks of firmware mismatch.

[0136] The extraction of heating response curve sets uses association indexing technology. In the database, heating curves are cross-table linked with refrigeration templates, and each heating curve record has its adapted refrigeration template number and effective temperature range. The system first filters the curve set that has successful pairing records with the optimal refrigeration reference template, and then performs working range cross-validation. The curve set contains multiple heating response curves under different environmental conditions, with each curve stored as a sequence of time-temperature-power three-dimensional coordinate points, with a uniform sampling interval of 500ms.

[0137] The temperature-power synergy verification establishes a dual-channel analysis model. The verification algorithm first performs time-domain alignment on the selected refrigeration reference template and the heating curve to be verified, using the inflection point of refrigeration power change as the synchronization marker. The synergy detection module scans the power output value of the heating curve at each time node, while detecting the phase characteristics of the refrigeration power curve. When the refrigeration power is in the rapid change stage, the system requires the heating power to remain relatively stable; in the refrigeration power platform period, it allows the heating power to adjust appropriately. The key verification point is set at the 300ms window period after the refrigeration power switches, and if the gradient change of the heating power exceeds the set threshold value at this time, it is marked as a power conflict event.

[0138] Anomaly curve processing adopts intelligent repair strategy. When power conflict is detected, the system first analyzes the power trend in the conflict period. If the conflict is caused by local mutation and the duration is less than 600ms, a smooth transition segment is generated by taking two normal reference points before and after the conflict point using cubic spline interpolation method. For large-scale persistent conflict curve, the system automatically retrieves successful curve samples of similar materials from the historical database, learns their power coordination mode through dynamic time warping algorithm, and synthesizes a new response curve that meets the requirements. Each curve repair operation generates a detailed correction log, including original data points, replacement point position and interpolation parameters.

[0139] The timing alignment stage implements a multi-scale phase synchronization mechanism. The main time axis takes the refrigeration power sequence as the reference, and identifies its key event marker points: including refrigeration start point, maximum power point, power switching point and steady state maintenance start point. The time axis of the heating response curve is mapped to the main time axis through a nonlinear stretching algorithm, and the temperature response characteristics of the heating curve are kept unchanged during the stretching process. The timing compensation module specially processes the initial delay characteristics of the heating response, and inserts a preset leading delay segment to ensure that the heating power and the refrigeration power are coordinated at the system response level. The aligned double curve generates a configuration file with collaborative control markers, and each control point contains a state code to identify the master-slave relationship between refrigeration and heating.

[0140] The initialization verification process performs closed-loop testing. After loading the baseline configuration, the system enters simulation mode and calls the built-in digital twin engine for virtual temperature control simulation. The simulation model includes cup holder structure heat transfer equation, container material thermal resistance network and environmental convection boundary conditions. The system injects standard temperature disturbance signals and monitors the temperature response data of the virtual container at each stage. The verification indicators focus on temperature overshoot, steady-state error and recovery time, and all indicators must meet the requirements before entering the actual control stage. If the verification fails, the system automatically triggers the configuration rollback mechanism and reinitializes the problem module according to the failure characteristics.

[0141] The hardware adaptation layer handles specific drive parameter conversion. The system converts the logical control parameters in the baseline configuration into hardware instruction sets, maps the refrigeration baseline template into PWM output frequency and duty cycle combinations, and converts the heating response curve into DAC output level sequences. The temperature sensor characteristic calibration value is added during the instruction conversion process to eliminate the influence of individual differences. After safety verification, the drive parameter set is stored in the FRAM memory with power failure protection function to ensure that the configuration data is complete after the vehicle is turned off.

[0142] The vehicle network cooperative initialization mechanism exchanges environmental parameters. Before the system startup is completed, the current environmental state snapshot is requested from the air conditioning control unit through the CAN bus, including the cabin temperature distribution map, the sunshine angle coefficient, and the air conditioning outlet speed. These parameters are used to fine-tune the environmental compensation factor in the baseline configuration, especially increasing the forced convection compensation coefficient for the cup holder position near the front row outlet. The system simultaneously registers the temperature control service to the cabin domain controller, establishing a priority communication channel, which can receive real-time control target updates when the user operates the center screen to switch the beverage type.

[0143] Configuration updates during operation use an incremental loading method. When the system detects that the container thermal characteristics continuously deviate from the initialization baseline, it starts a partial configuration update process in the background. The update process only replaces the module configurations with poor performance, such as updating only the refrigeration template or the local heating curve, maintaining the stability of the remaining parts of the system. The update package is verified for integrity through a dual-channel verification mechanism. The main controller and the security coprocessor independently calculate the hash value of the configuration data, and only after the comparison is consistent is the update applied.

[0144] The abnormality handling plan contains multiple response strategies. When the container material cannot be identified, the system switches to a general multi-layer configuration scheme: the bottom layer loads the stainless steel material baseline configuration, and the upper layer adds a heat conduction compensation coefficient estimated according to the container shape. When the database retrieval times out, the emergency mode is started, and the configuration parameters of the last successful run are used. All abnormal events generate a diagnostic snapshot with an environmental image, which is stored in an independent partition for subsequent analysis.

[0145] The firmware upgrade process includes configuration migration. When a new system firmware version is detected, the compatibility flag of the existing baseline configuration is automatically checked. In the compatible case, the configuration structure conversion is performed to migrate the old version parameters to the new format; in the incompatible case, the user is prompted to reset the configuration through a special tool. The migration process preserves historical running statistical information, and the new control algorithm can use these data to optimize the initialization accuracy.

[0146] The life cycle maintenance mechanism tracks the configuration validity. The system records the actual running effect indicators of each set of baseline configuration, establishes a configuration health degree evaluation model, and when the cumulative use time of a configuration reaches a certain threshold, the system automatically performs a re-initialization process before replacing the new configuration. The historical configurations with performance degradation are marked and archived for later fault analysis and backtracking comparison, forming a closed-loop learning system for configuration optimization.

[0147] The system integration test phase verifies the stability of the initialization process under extreme conditions such as voltage fluctuation, temperature shock, etc. Special attention is paid to the low-temperature cold start scenario, and the system ensures reliable reading of configuration data by preheating the FRAM memory at an environmental temperature of -30°C. In the startup scenario with serious electrical interference, the configuration data uses differential transmission and retransmission protocol to ensure complete loading of control parameters. The entire initialization process takes less than 2 seconds, and seamless switching of control strategies can be achieved when replacing the container during vehicle driving.

[0148] It should be noted that the relational terms herein such as first and second and the like are used solely to distinguish one from another entity or action without necessarily requiring or implying any actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0149] While embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, changes, and alterations can be made by those skilled in the art without departing from the principles and spirit of the present application, and it is intended that the present application be limited only by the scope of the claims hereinafter and their equivalents.

Claims

1. A microcontroller-based temperature regulation control method for a car cup holder, characterized by, The method comprises: Collecting real-time temperature data and environmental parameters of the target container to generate an initial temperature parameter set, the initial temperature parameter set including a container surface temperature distribution curve and an environmental heat exchange coefficient sequence; According to the container surface temperature distribution curve in the initial temperature parameter set, a first adjustment model is established, and the first adjustment model includes a mapping relationship between surface temperature change and refrigeration power; According to the environmental heat exchange coefficient sequence in the initial temperature parameter set, a second adjustment model is established, and the second adjustment model includes a compensation rule of environmental heat interference and heating power; Based on the real-time temperature difference fluctuation data in the cup holder temperature control process and the container material thermal conductivity characteristic data, the target model in the first adjustment model or the second adjustment model is activated, and a dynamic power instruction is generated through the target model; The dynamic power instruction is input to the semiconductor temperature control unit of the cup holder to adjust the refrigeration fin current gradient or the heating film voltage distribution; The collecting real-time temperature data and environmental parameters of the target container to generate an initial temperature parameter set comprises: Synchronously collecting temperature sampling values of each surface of the container through a multi-region temperature sensor, and analyzing original data streams of an environmental temperature and humidity sensor; According to the spatial distribution characteristics of the temperature sampling values, the container surface is divided into at least two thermal sections, and each thermal section is assigned a corresponding initial refrigeration reference value and a heating compensation coefficient; Refrigeration response parameters and heating lag time constants matched with each thermal section are extracted from a historical parameter library; The refrigeration response parameters are subjected to time domain normalization processing to generate an optimized refrigeration gradient sequence corresponding to each thermal section; The heating lag time constants are subjected to frequency domain decomposition processing to generate an optimized heating response curve corresponding to each thermal section; The initial temperature parameter set is fused and generated according to the optimized refrigeration gradient sequence and the optimized heating response curve of each thermal section; The first adjustment model and the second adjustment model are established, comprising: The optimized refrigeration gradient sequence is input to a temperature prediction model, the model weight is trained through historical temperature distribution data, and a refrigeration power mapping function in the first adjustment model is generated; The optimized heating response curve is input to a power control model, the control threshold is adjusted through a heat conduction compensation parameter, and a heating compensation rule table in the second adjustment model is generated; After the temperature prediction model and the power control model converge, key control vectors in the refrigeration power mapping function and the heating compensation rule table are extracted respectively; The key control vectors are matched with real-time collected container material thermal conductivity characteristic data for matching degree verification; When the matching degree is lower than a preset threshold, the temperature prediction model and the power control model are re-iteratively trained until the convergence condition is met.

2. The microcontroller-based automotive cup holder temperature regulation control method of claim 1, wherein, The based on the real-time temperature difference fluctuation data in the cup holder temperature control process and the container material thermal conductivity characteristic data, the target model in the first adjustment model or the second adjustment model is activated, and a dynamic power instruction is generated through the target model, comprising: Real-time monitoring of cup holder internal temperature difference fluctuation amplitude and container thermal conduction delay time; activating a refrigeration power mapping function in the first regulation model when the temperature difference fluctuation amplitude exceeds a refrigeration activation threshold and the heat conduction delay time is in a stable interval; generating a dynamic current adjustment instruction of a refrigeration sheet according to a gradient mapping rule in the refrigeration power mapping function; activating a heating compensation rule table in the second regulation model when the heat conduction delay time exceeds a heating activation threshold and the temperature difference fluctuation amplitude is in a stable interval; generating a dynamic voltage adjustment instruction of a heating film according to a compensation distribution rule in the heating compensation rule table.

3. The microcontroller-based automotive cup holder temperature regulation control method of claim 2, wherein, The input of the dynamic power instruction to the semiconductor temperature control unit of the cup holder comprises: adjusting the driving current value of each region of the refrigeration sheet in stages according to the gradient change amount in the dynamic current adjustment instruction; after each current adjustment, collecting container surface temperature feedback data and performing error calculation with the predicted value of the refrigeration power mapping function; dynamically correcting the weight parameter of the refrigeration power mapping function according to the error change trend.

4. The microcontroller-based automotive cup holder temperature regulation control method of claim 3, wherein, The method further comprises: adjusting the output voltage value of the heating film in different temperature zones according to the voltage distribution parameter in the dynamic voltage adjustment instruction; in the voltage adjustment process, the uniformity of the container temperature is detected in real time through an infrared sensor, and the compensation coefficient in the heating compensation rule table is updated according to the detection result.

5. The microcontroller-based automotive cup holder temperature regulation control method of claim 4, wherein, The method further comprises a feedback stage after the execution of the dynamic power instruction, comprising: collecting final temperature distribution data of the container and heating film energy consumption data; comparing the final temperature distribution data with the expected temperature range of the first regulation model to generate a refrigeration model error signal; performing deviation analysis on the heating film energy consumption data and the reference energy consumption curve of the second regulation model to generate a heating model error signal; adjusting the reference current gradient in the refrigeration power mapping function according to the system deviation component in the refrigeration model error signal; optimizing the voltage distribution parameter in the heating compensation rule table according to the energy consumption fluctuation component in the heating model error signal.

6. The microcontroller-based automotive cup holder temperature regulation control method of claim 5, wherein, Adjusting the reference current gradient and optimizing the voltage distribution parameter comprise: identifying the steady-state deviation amount in the refrigeration model error signal, and calculating the current compensation amount through a sliding mean algorithm; updating the reference parameter of the refrigeration power mapping function according to the current compensation amount; identifying the transient fluctuation component in the heating model error signal, and extracting the effective voltage correction amount through a filtering algorithm; adjusting the voltage distribution weight in the heating compensation rule table according to the effective voltage correction amount.

7. The microcontroller-based automotive cup holder temperature regulation control method of claim 1, wherein, The method further comprises executing in the device initialization stage: parsing the heat capacity identifier in the material quality code and the thermal conductivity coefficient code segment of the target container to generate a material thermal characteristic vector; inputting the material thermal characteristic vector into a pre-stored material thermal characteristic database for similarity retrieval, and screening candidate temperature control templates with a matching degree exceeding a threshold; selecting an optimal refrigeration reference template according to the historical control stability index of the candidate temperature control template; extracting a set of heating response curves associated with the optimal refrigeration reference template from the database; performing temperature-power cooperativity verification on the set of heating response curves to eliminate abnormal curves with temperature mutation or power conflict; The verified heating response curve is time-aligned with the optimal refrigeration benchmark template to generate a benchmark configuration of an initial temperature parameter set.

8. The microcontroller-based automotive cup holder temperature regulation control method of claim 7, wherein, The temperature-power synergy verification on the heating response curve set comprises: extracting heating power distribution data of a single to-be-verified curve, and synchronously acquiring a refrigeration power sequence of a corresponding time node in the optimal refrigeration benchmark template; annotating a synergy timestamp on the heating power distribution data according to a phase node of the refrigeration power sequence; detecting whether a power change slope of adjacent intervals of the synergy timestamp exceeds a sudden change threshold; when a power conflict is detected, generating a replacement power smoothing segment based on a historical conflict correction record and updating the curve.

Images