A wireless power transmitting chip and its method

Abstract

This application relates to the field of intelligent control. Specifically, it discloses a wireless power transmission chip and its method. By obtaining the time queue of the voltage level of the input power supply and the time queue of the real-time temperature of the chip, and using deep learning data analysis and processing techniques to perform temporal correlation implicit encoding of the voltage level and the real-time temperature of the chip, the switching frequency value at the next time point is automatically recommended based on the conditional significant interaction fusion representation between the temporal correlation implicit features of the voltage level and the temporal correlation implicit features of the real-time temperature of the chip. In this way, the switching frequency can be dynamically adjusted according to the real-time monitored data to ensure the best performance under various working conditions. This not only significantly improves the thermal management ability of the chip but also greatly enhances the degree of intelligent control.

Description

Technical Field

[0001] This application relates to the field of intelligent control, and more specifically, to a wireless power transmission chip and its method. Background Art

[0002] Wireless power transmission technology has become an indispensable part of modern electronic devices, especially widely used in portable devices and Internet of Things (IoT) devices. Wireless power transmission technology enables power transmission without physical connection, improving the convenience and safety of use. And the wireless power transmission chip is the core component in wireless power transmission technology and an important support for the realization of wireless charging technology.

[0003] In the design of wireless power transmission chips, thermal management is a crucial technical consideration. Since the power consumption varies under different voltages, heat dissipation issues need to be taken into account, and the corresponding output should be adjusted in a timely manner to avoid overheating. And the switching frequency has a direct impact on efficiency and the generated heat. Therefore, the efficiency can be improved by changing the switching frequency, thereby reducing unnecessary heat generation.

[0004] In traditional wireless switched-mode power supply designs, a fixed switching frequency is usually set. However, at a fixed switching frequency, when the load changes, the efficiency of the system may be significantly affected. For example, under light load conditions, a higher switching frequency will result in higher switching losses, thus reducing the overall efficiency. In addition, fixed-frequency control cannot be flexibly adjusted according to environmental changes (such as temperature, input voltage fluctuations, etc.), resulting in poor performance in a dynamic environment. Therefore, under different operating conditions, it is difficult for traditional control methods to find the optimal switching frequency, making the system performance and energy efficiency ratio unable to reach the best state, thus limiting the adaptability and flexibility and making it difficult to cope with complex and changing practical application scenarios.

[0005] Therefore, an optimized wireless power transmission chip is desired. Summary of the Invention

[0006] To solve the above technical problems, this application is proposed. Embodiments of this application provide a wireless power transmission chip and its method. By obtaining a time queue of the voltage level of the input power supply and a time queue of the real-time temperature of the chip, and using deep learning data analysis and processing techniques to perform sequential correlation implicit coding on the voltage level and the real-time temperature of the chip, a recommended value of the switching frequency at the next time point is automatically recommended based on the conditional significant interaction fusion representation between the voltage level sequential correlation implicit feature and the chip real-time temperature sequential correlation implicit feature. In this way, whether under different load conditions or in changing ambient temperatures, the switching frequency can be dynamically adjusted according to the real-time monitored data to ensure optimal performance under various working conditions. This not only significantly improves the thermal management ability of the chip but also greatly enhances the degree of intelligent control.

[0007] According to one aspect of this application, a wireless power transmission chip is provided, which includes:

[0008] A voltage identification module for identifying the voltage level of the input power supply;

[0009] A temperature detection module for detecting the real-time temperature of the chip through a temperature sensor;

[0010] An adaptive control module for adaptively adjusting the switching frequency based on the voltage level of the input power supply and the real-time temperature of the chip;

[0011] Among them, the adaptive control module includes:

[0012] A voltage level time sequence acquisition unit for acquiring a time queue of the voltage level of the input power supply;

[0013] A real-time temperature time sequence acquisition unit for acquiring a time queue of the real-time temperature of the chip;

[0014] A voltage level - real-time temperature time sequence correlation unit for performing time series coding on the time queue of the voltage level and the time queue of the real-time temperature of the chip to obtain a voltage level time sequence correlation implicit feature vector and a chip real-time temperature time sequence correlation implicit feature vector;

[0015] A voltage level - chip temperature time sequence significant interaction unit for performing conditional significant interaction fusion between the voltage level time sequence correlation implicit feature vector and the chip real-time temperature time sequence correlation implicit feature vector to obtain a voltage level - chip temperature time sequence significant interaction fusion representation;

[0016] An optimization instruction generation unit for obtaining an optimization instruction based on the voltage level - chip temperature time sequence significant interaction fusion representation, where the optimization instruction includes a recommended value of the switching frequency at the next time point.

[0017] In the above wireless power transmission chip, the voltage level real-time temperature time series correlation unit is configured to: input the time queue of the voltage level and the time queue of the real-time temperature of the chip into a time series encoder based on a forward LSTM model to obtain the voltage level time series correlation hidden feature vector and the chip real-time temperature time series correlation hidden feature vector.

[0018] In the above wireless power transmission chip, the voltage level-chip temperature time series significant interaction unit includes: a voltage level-chip temperature hidden feature joint subunit configured to perform a shared hidden feature joint on the voltage level time series correlation hidden feature vector and the chip real-time temperature time series correlation hidden feature vector to obtain a voltage level-chip temperature conditional feature vector; a voltage level-chip temperature conditional feature interaction subunit configured to perform a conditional information-based feature weighted interaction on the voltage level time series correlation hidden feature vector and the chip real-time temperature time series correlation hidden feature vector based on the voltage level-chip temperature conditional feature vector to obtain the voltage level-chip temperature time series significant interaction fusion representation.

[0019] In the above wireless power transmission chip, the voltage level-chip temperature hidden feature joint subunit includes: a voltage level-chip temperature time series joint secondary subunit configured to input the voltage level time series correlation hidden feature vector and the chip real-time temperature time series correlation hidden feature vector into a joint implicit feature capture network to obtain a voltage level-chip temperature time series joint implicit feature vector; a voltage level-chip temperature conditional feature generation secondary subunit configured to perform a feature activation on the voltage level-chip temperature time series joint implicit feature vector based on the Sigmoid function to obtain the voltage level-chip temperature conditional feature vector.

[0020] In the above wireless power transmission chip, the voltage level-chip temperature time series joint secondary subunit is configured to: perform a position-wise addition on the voltage level time series correlation hidden feature vector and the chip real-time temperature time series correlation hidden feature vector, multiply the obtained voltage level-chip temperature time series sum vector by a weight matrix, and then perform a position-wise addition with a bias vector to obtain a voltage level-chip temperature time series joint interaction vector; process the voltage level-chip temperature time series joint interaction vector using the tanh function to obtain the voltage level-chip temperature time series joint implicit feature vector.

[0021] In the above wireless power transmission chip, the voltage level-chip temperature condition feature interaction sub-unit includes: a voltage level semantic contribution degree calculation secondary sub-unit, configured to calculate the voltage level semantic contribution degree of the voltage level time-series associated implicit feature vector relative to the voltage level-chip temperature condition feature vector; a chip temperature semantic contribution degree calculation secondary sub-unit, configured to calculate the chip temperature semantic contribution degree of the chip real-time temperature time-series associated implicit feature vector relative to the voltage level-chip temperature condition feature vector; a voltage level-chip temperature feature modulation secondary sub-unit, configured to perform normalization processing on the voltage level semantic contribution degree and the chip temperature semantic contribution degree, and use the normalized voltage level semantic contribution degree and the normalized chip temperature semantic contribution degree to perform weighted modulation on the voltage level time-series associated implicit feature vector and the chip real-time temperature time-series associated implicit feature vector to obtain a modulated voltage level time-series associated implicit feature vector and a modulated chip real-time temperature time-series associated implicit feature vector; a voltage level-chip temperature interaction fusion secondary sub-unit, configured to use the modulated voltage level time-series associated implicit feature vector as a query vector, the modulated chip real-time temperature time-series associated implicit feature vector as a key vector, and the voltage level-chip temperature condition feature vector as a value vector, and input the modulated voltage level time-series associated implicit feature vector, the modulated chip real-time temperature time-series associated implicit feature vector, and the voltage level-chip temperature condition feature vector into a feature-intermediate significant guidance interaction module based on a converter structure to obtain a voltage level-chip temperature time-series significant interaction fusion representation vector as the voltage level-chip temperature time-series significant interaction fusion representation.

[0022] In the above wireless power transmission chip, the voltage level semantic contribution degree calculation secondary sub-unit is configured to: calculate the element-wise division of the voltage level time-series associated implicit feature vector and the voltage level-chip temperature condition feature vector to obtain a voltage level semantic contribution vector; calculate the base-2 logarithmic function value of the absolute value of each eigenvalue of the voltage level semantic contribution vector to obtain a voltage level semantic contribution logarithmic vector; calculate the element-wise dot product of the voltage level time-series associated implicit feature vector and the voltage level semantic contribution logarithmic vector, and perform element-wise summation on the obtained dot product vector to obtain a voltage level semantic contribution value; calculate the exponential function with the natural constant e as the base and the voltage level semantic contribution value as the exponent to obtain the voltage level semantic contribution degree.

[0023] In the above wireless power transmission chip, the voltage level chip temperature feature modulation secondary subunit is configured to: calculate the sum value of the voltage level semantic contribution degree and the chip temperature semantic contribution degree to obtain the voltage level-chip temperature semantic contribution sum value; divide the voltage level semantic contribution degree and the chip temperature semantic contribution degree by the voltage level-chip temperature semantic contribution sum value respectively to obtain the normalized voltage level semantic contribution degree and the normalized chip temperature semantic contribution degree; perform element-wise multiplication of the voltage level time-series associated implicit feature vector and the normalized voltage level semantic contribution degree to obtain the modulated voltage level time-series associated implicit feature vector; perform element-wise multiplication of the chip real-time temperature time-series associated implicit feature vector and the normalized chip temperature semantic contribution degree to obtain the modulated chip real-time temperature time-series associated implicit feature vector.

[0024] In the above wireless power transmission chip, the optimization instruction generation unit is configured to: input the voltage level-chip temperature time-series significant interaction fusion representation vector into an optimization controller based on a decoder to obtain the optimization instruction, where the optimization instruction includes a recommended value of the switching frequency at the next time point.

[0025] According to another aspect of the present application, a method for a wireless power transmission chip is provided, which includes:

[0026] Obtain a time queue of the voltage level of the input power supply;

[0027] Obtain a time queue of the chip real-time temperature;

[0028] Perform time series encoding on the time queue of the voltage level and the time queue of the chip real-time temperature to obtain a voltage level time-series associated implicit feature vector and a chip real-time temperature time-series associated implicit feature vector;

[0029] Perform conditional significant interaction fusion between the voltage level time-series associated implicit feature vector and the chip real-time temperature time-series associated implicit feature vector to obtain a voltage level-chip temperature time-series significant interaction fusion representation;

[0030] Based on the voltage level-chip temperature time-series significant interaction fusion representation, obtain an optimization instruction, where the optimization instruction includes a recommended value of the switching frequency at the next time point.

[0031] Compared with the prior art, a wireless power transmission chip and its method provided by the present application obtain a time queue of the voltage level of the input power supply and a time queue of the real-time temperature of the chip, and use deep learning data analysis and processing techniques to perform sequential correlation implicit coding on the voltage level and the real-time temperature of the chip, so as to automatically recommend the switching frequency value at the next time point based on the conditional significant interaction fusion representation between the sequential correlation implicit features of the voltage level and the sequential correlation implicit features of the real-time temperature of the chip. In this way, whether under different load conditions or in changing ambient temperatures, the switching frequency can be dynamically adjusted according to the real-time monitored data to ensure the best performance under various working conditions, which not only significantly improves the thermal management ability of the chip but also greatly enhances the degree of intelligent control. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other objects, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0033] Figure 1 FIG. is a block diagram of a wireless power transmission chip according to an embodiment of the present application.

[0034] Figure 2 FIG. is a block diagram of an adaptive control module in a wireless power transmission chip according to an embodiment of the present application.

[0035] Figure 3 FIG. is a schematic diagram of data flow in an adaptive control module in a wireless power transmission chip according to an embodiment of the present application.

[0036] Figure 4 FIG. is a block diagram of a voltage level-chip temperature sequential significant interaction unit in a wireless power transmission chip according to an embodiment of the present application.

[0037] Figure 5 FIG. is a circuit design diagram of a wireless power transmission chip according to an embodiment of the present application.

[0038] Figure 6 FIG. is a flowchart of a method for a wireless power transmission chip according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0040] It should be understood that the various steps recited in the method embodiments of the present disclosure can be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0041] In the description of the embodiments of the present disclosure, the term "comprising" and its like should be understood as an open inclusion, that is, "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "an embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. There may also be other explicit and implicit definitions hereinafter.

[0042] It should be noted that the modifications of "one" and "plural" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise clearly stated in the context, it should be understood as "one or more".

[0043] Wireless power transmission technology has become an indispensable part of modern electronic devices, especially widely used in portable devices and Internet of Things (IoT) devices. Wireless power transmission technology can achieve power transmission without physical connection, improving the convenience and safety of use. And the wireless power transmission chip is the core component in wireless power transmission technology and an important support for the realization of wireless charging technology.

[0044] In the design of wireless power transmission chips, thermal management is a crucial technical consideration. Since the power consumption under different voltages varies, it is necessary to consider the heat dissipation problem and adjust the corresponding output in a timely manner to avoid overheating. And the switching frequency has a direct impact on the efficiency and the heat generated. Therefore, the efficiency can be improved by changing the switching frequency, thereby reducing unnecessary heat generation.

[0045] In traditional wireless switched-mode power supply designs, a fixed switching frequency is usually set. However, at a fixed switching frequency, when the load changes, the efficiency of the system may be significantly affected. For example, under light load conditions, a higher switching frequency will result in higher switching losses, thereby reducing the overall efficiency. In addition, fixed-frequency control cannot be flexibly adjusted according to environmental changes (such as temperature, input voltage fluctuations, etc.), resulting in poor performance in a dynamic environment. Therefore, under different operating conditions, it is difficult for traditional control methods to find the optimal switching frequency, making it impossible to achieve the best system performance and energy efficiency ratio, thus limiting the adaptive ability and flexibility and making it difficult to cope with complex and changing practical application scenarios.

[0046] Based on this, Figure 1 FIG. is a block diagram of a wireless power transmitting chip according to an embodiment of the present application. As Figure 1 shown, the wireless power transmitting chip 100 according to an embodiment of the present application includes: a voltage identification module 110 for identifying the voltage level of the input power supply; a temperature detection module 120 for detecting the real-time temperature of the chip through a temperature sensor; and an adaptive control module 130 for adaptively adjusting the switching frequency based on the voltage level of the input power supply and the real-time temperature of the chip.

[0047] Specifically, in the embodiment of the present application, the voltage identification module 110 is used to identify the voltage level of the input power supply. It should be understood that the voltage level of the input power supply may change due to factors such as power grid fluctuations and load changes. A stable input voltage is a prerequisite for ensuring the normal operation of the system. Different voltage levels will affect the efficiency of the system. For example, too high a voltage may result in excessive current, thereby increasing losses; too low a voltage may cause the system to malfunction. Therefore, by identifying the voltage level of the input power supply, corresponding protection measures can be taken to avoid chip damage.

[0048] Specifically, in the embodiment of the present application, the temperature detection module 120 is used to detect the real-time temperature of the chip through a temperature sensor. Correspondingly, too high a temperature will affect the performance of the chip and even cause system failures or damage. By monitoring the temperature in real time, measures can be taken in a timely manner, such as reducing the output power or adjusting the switching frequency, to control the temperature and better perform thermal management.

[0049] Accordingly, in the adaptive control module, the technical concept of the present application is to obtain the time queue of the voltage level of the input power supply and the time queue of the real-time temperature of the chip, and use the data analysis and processing technology of deep learning to perform the time-series correlation implicit coding of the voltage level and the real-time temperature of the chip, so as to automatically recommend the switching frequency value at the next time point based on the conditional significant interaction fusion representation between the voltage level time-series correlation implicit feature and the chip real-time temperature time-series correlation implicit feature. In this way, whether under different load conditions or in changing ambient temperatures, the switching frequency can be dynamically adjusted according to the real-time monitored data to ensure the best performance under various working conditions, which not only significantly improves the thermal management ability of the chip but also greatly enhances the degree of intelligent control.

[0050] Figure 2 FIG. is a block diagram of an adaptive control module in a wireless power transmission chip according to an embodiment of the present application. Figure 3 FIG. is a schematic diagram of data flow in an adaptive control module in a wireless power transmission chip according to an embodiment of the present application. As Figure 2 and Figure 3 shown, the adaptive control module 130 includes: a voltage level time series acquisition unit 131 for acquiring the time queue of the voltage level of the input power supply; a real-time temperature time series acquisition unit 132 for acquiring the time queue of the real-time temperature of the chip; a voltage level-real-time temperature time series correlation unit 133 for performing time series coding on the time queue of the voltage level and the time queue of the real-time temperature of the chip to obtain a voltage level time series correlation implicit feature vector and a chip real-time temperature time series correlation implicit feature vector; a voltage level-chip temperature time series significant interaction unit 134 for performing conditional significant interaction fusion between the voltage level time series correlation implicit feature vector and the chip real-time temperature time series correlation implicit feature vector to obtain a voltage level-chip temperature time series significant interaction fusion representation; and an optimization instruction generation unit 135 for obtaining an optimization instruction based on the voltage level-chip temperature time series significant interaction fusion representation, where the optimization instruction includes a recommended value of the switching frequency at the next time point.

[0051] In an embodiment of the present application, the voltage level time sequence acquisition unit 131 is configured to acquire a time queue of the voltage level of the input power supply. The real-time temperature time sequence acquisition unit 132 is configured to acquire a time queue of the real-time temperature of the chip. It should be understood that the voltage level of the input power supply refers to the voltage value provided by the power supply, and different voltage levels will affect the power consumption of the chip. The real-time temperature of the chip refers to the actual temperature of the wireless power transmission chip during operation. If the temperature is too high, it will affect the performance of the chip and even cause system failure or damage. Based on this, in order to more accurately dynamically adjust the switching frequency according to the current working state and ensure the best performance under different load conditions and environmental temperatures, in the technical solution of the present application, a time queue of the voltage level of the input power supply is acquired, and a time queue of the real-time temperature of the chip is acquired.

[0052] In an embodiment of the present application, the voltage level-real-time temperature time sequence association unit 133 is configured to perform time series encoding on the time queue of the voltage level and the time queue of the real-time temperature of the chip to obtain a voltage level time sequence association hidden feature vector and a real-time temperature time sequence association hidden feature vector of the chip. Specifically, in an embodiment of the present application, the voltage level-real-time temperature time sequence association unit is configured to: input the time queue of the voltage level and the time queue of the real-time temperature of the chip into a time series encoder based on a forward LSTM model to obtain the voltage level time sequence association hidden feature vector and the real-time temperature time sequence association hidden feature vector of the chip. Correspondingly, considering that there are different degrees of time sequence feature associations between the voltage level and the real-time temperature of the chip within a specific time period, and in order to capture the relationship between early information and current information, so as to better understand the influence of time sequence associations between different time spans, in the technical solution of the present application, the time queue of the voltage level and the time queue of the real-time temperature of the chip are input into a time series encoder based on a forward LSTM model to obtain a voltage level time sequence association hidden feature vector and a real-time temperature time sequence association hidden feature vector of the chip. That is, the forward LSTM model can capture long-term dependencies in the data. In particular, the model uses a gating mechanism to control the inflow and outflow of information, so as to effectively manage long-term dependencies at different specific time scales to better understand the time sequence association relationship between the voltage level and the real-time temperature.

[0053] In an embodiment of the present application, the voltage level-chip temperature time sequence significant interaction unit 134 is configured to perform conditional significant interaction fusion between the voltage level time sequence association hidden feature vector and the real-time temperature time sequence association hidden feature vector of the chip to obtain a voltage level-chip temperature time sequence significant interaction fusion representation. Figure 4Block diagram of the voltage level-chip temperature timing significant interaction unit in the wireless power transmission chip according to an embodiment of the present application. Specifically, in the embodiment of the present application, as Figure 4 shown, the voltage level-chip temperature timing significant interaction unit 134 includes: a voltage level-chip temperature implicit feature joint sub-unit 1341, configured to perform shared implicit feature joint on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector to obtain a voltage level-chip temperature conditional feature vector; a voltage level-chip temperature conditional feature interaction sub-unit 1342, configured to perform conditional information-based feature weighted interaction on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector based on the voltage level-chip temperature conditional feature vector to obtain the voltage level-chip temperature timing significant interaction fusion representation. Further, considering that although the voltage level and the chip temperature are two independent measurement indicators, there may be an internal connection between them. For example, input voltage fluctuations may cause changes in the chip temperature, and vice versa. In order to capture and integrate the complex timing interaction relationship between the two sequence features, and extract deeper correlations and patterns to obtain a more comprehensive and useful representation, in the technical solution of the present application, conditional significant interaction fusion is performed on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector to obtain a voltage level-chip temperature timing significant interaction fusion representation vector as the voltage level-chip temperature timing significant interaction fusion representation.

[0054] Specifically, first, shared implicit feature joint is performed on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector to obtain a voltage level-chip temperature conditional feature vector. In this way, the information of the voltage level and the chip temperature can be integrated together, while capturing and mining the shared implicit features between the two to form a higher-level feature representation. Specifically, the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector are input into a joint implicit feature capture network to generate a more representative voltage level-chip temperature timing joint implicit feature vector by learning the internal association pattern between the voltage level and the chip temperature. Then, the Sigmoid function is used to process it to further capture and extract the most critical feature part for comprehensively describing the condition, and suppress the relatively less important features to obtain the voltage level-chip temperature conditional feature vector.

[0055] Then, based on the voltage level-chip temperature condition feature vector, perform conditional information-based feature weighted interaction on the voltage level time series associated implicit feature vector and the chip real-time temperature time series associated implicit feature vector to emphasize those features that have a significant impact on switching frequency control, and obtain the voltage level-chip temperature time series significant interaction fusion representation. Specifically, calculate the semantic contribution degrees of the voltage level time series associated implicit feature vector and the chip real-time temperature time series associated implicit feature vector with respect to the voltage level-chip temperature condition feature vector respectively to determine the importance of each feature for the conditional comprehensive representation, and obtain the voltage level semantic contribution degree and the chip temperature semantic contribution degree. After that, in order to ensure that they are compared on the same scale and there is no bias due to the difference in numerical magnitudes during the subsequent feature weighted modulation interaction process, normalize the voltage level semantic contribution degree and the chip temperature semantic contribution degree, and use the normalized voltage level semantic contribution degree and the normalized chip temperature semantic contribution degree to perform weighted modulation on the voltage level time series associated implicit feature vector and the chip real-time temperature time series associated implicit feature vector to highlight and characterize the more important parts for the conditional features, thereby generating the modulated voltage level time series associated implicit feature vector and the modulated chip real-time temperature time series associated implicit feature vector. Finally, use the modulated voltage level time series associated implicit feature vector as the query vector, the modulated chip real-time temperature time series associated implicit feature vector as the key vector, and the voltage level-chip temperature condition feature vector as the value vector, and input them into the feature-interaction module with significant guidance based on the transformer structure to utilize the transformer structure mechanism to better capture the non-linear relationship and the complex time series dependence of the significant key information, and obtain the voltage level-chip temperature time series significant interaction fusion representation vector.

[0056] Specifically, in the embodiment of the present application, the voltage level-chip temperature implicit feature joint subunit 1341 includes: a voltage level-chip temperature time series joint secondary subunit, configured to input the voltage level time series associated implicit feature vector and the chip real-time temperature time series associated implicit feature vector into the joint implicit feature capture network to obtain a voltage level-chip temperature time series joint implicit feature vector; a voltage level-chip temperature condition feature generation secondary subunit, configured to perform feature activation on the voltage level-chip temperature time series joint implicit feature vector based on the Sigmoid function to obtain the voltage level-chip temperature condition feature vector.

[0057] More specifically, in the embodiments of the present application, the voltage level-chip temperature timing joint secondary subunit is configured to: perform element-wise addition on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector, multiply the obtained voltage level-chip temperature timing summation vector by a weight matrix, and then perform element-wise addition with a bias vector to obtain a voltage level-chip temperature timing joint interaction vector; use the tanh function to process the voltage level-chip temperature timing joint interaction vector to obtain the voltage level-chip temperature timing joint implicit feature vector.

[0058] Specifically, in the embodiments of the present application, the voltage level-chip temperature conditional feature interaction subunit 1342 includes: a voltage level semantic contribution degree calculation secondary subunit, configured to calculate the voltage level semantic contribution degree of the voltage level timing associated implicit feature vector relative to the voltage level-chip temperature conditional feature vector; a chip temperature semantic contribution degree calculation secondary subunit, configured to calculate the chip temperature semantic contribution degree of the chip real-time temperature timing associated implicit feature vector relative to the voltage level-chip temperature conditional feature vector; a voltage level-chip temperature feature modulation secondary subunit, configured to perform normalization processing on the voltage level semantic contribution degree and the chip temperature semantic contribution degree, and use the normalized voltage level semantic contribution degree and the normalized chip temperature semantic contribution degree to perform weighted modulation on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector to obtain a modulated voltage level timing associated implicit feature vector and a modulated chip real-time temperature timing associated implicit feature vector; a voltage level-chip temperature interaction fusion secondary subunit, configured to use the modulated voltage level timing associated implicit feature vector as a query vector, the modulated chip real-time temperature timing associated implicit feature vector as a key vector, and the voltage level-chip temperature conditional feature vector as a value vector, and input the modulated voltage level timing associated implicit feature vector, the modulated chip real-time temperature timing associated implicit feature vector, and the voltage level-chip temperature conditional feature vector into a feature-wise significant guidance interaction module based on a transformer structure to obtain a voltage level-chip temperature timing significant interaction fusion representation vector as the voltage level-chip temperature timing significant interaction fusion representation.

[0059] More specifically, in the embodiments of the present application, the voltage level semantic contribution degree calculation secondary subunit is configured to: calculate the position-wise division of the voltage level time-series associated implicit feature vector and the voltage level-chip temperature conditional feature vector to obtain a voltage level semantic contribution vector; calculate the base-2 logarithmic function value of the absolute value of each eigenvalue of the voltage level semantic contribution vector to obtain a voltage level semantic contribution logarithmic vector; calculate the position-wise dot product of the voltage level time-series associated implicit feature vector and the voltage level semantic contribution logarithmic vector, and perform element-wise addition on the obtained dot product vector to obtain a voltage level semantic contribution value; calculate the exponential function with the natural constant e as the base and the voltage level semantic contribution value as the exponent to obtain the voltage level semantic contribution degree. Similarly, the calculation process of the chip temperature semantic contribution degree is the same as that of the voltage level semantic contribution degree.

[0060] More specifically, in the embodiments of the present application, the voltage level-chip temperature feature modulation secondary subunit is configured to: calculate the sum of the voltage level semantic contribution degree and the chip temperature semantic contribution degree to obtain a voltage level-chip temperature semantic contribution sum value; divide the voltage level semantic contribution degree and the chip temperature semantic contribution degree by the voltage level-chip temperature semantic contribution sum value respectively to obtain the normalized voltage level semantic contribution degree and the normalized chip temperature semantic contribution degree; perform a position-wise dot product of the voltage level time-series associated implicit feature vector and the normalized voltage level semantic contribution degree to obtain the modulated voltage level time-series associated implicit feature vector; perform a position-wise dot product of the chip real-time temperature time-series associated implicit feature vector and the normalized chip temperature semantic contribution degree to obtain the modulated chip real-time temperature time-series associated implicit feature vector.

[0061] More specifically, the voltage level-chip temperature interaction fusion secondary subunit is configured to: perform a vector multiplication of the modulated voltage level time-series associated implicit feature vector and the transposed vector of the modulated chip real-time temperature time-series associated implicit feature vector, and perform a position-wise division of the obtained modulated voltage level-chip temperature time-series associated matrix by the square root of the length of the modulated chip real-time temperature time-series associated implicit feature vector to obtain a modulated voltage level-chip temperature time-series associated scaling matrix; input the modulated voltage level-chip temperature time-series associated scaling matrix into the softmax function, and perform a matrix-vector multiplication of the obtained modulated voltage level-chip temperature time-series associated scaling activation matrix and the voltage level-chip temperature conditional feature vector to obtain the voltage level-chip temperature time-series significant interaction fusion representation vector.

[0062] In the embodiments of the present application, specifically, a conditional significant interaction fusion between the voltage level timing-related implicit feature vector and the chip real-time temperature timing-related implicit feature vector can be expressed by the formula:

[0063] ;

[0064] ;

[0065] ;

[0066] ;

[0067] ;

[0068] ;

[0069] ;

[0070] ;

[0071] ;

[0072] Among them, V 1 and V 2 are the voltage level timing-related implicit feature vector and the chip real-time temperature timing-related implicit feature vector respectively, is addition by position points, W VT and b VT are the weight matrix and the bias vector respectively, is function, V J is the voltage level-chip temperature timing joint implicit feature vector, is function, represents the voltage level-chip temperature conditional feature vector, is the feature value at each position in the voltage level timing-related implicit feature vector, is the feature value at each position in the chip real-time temperature timing-related implicit feature vector, is the feature value at each position in the voltage level-chip temperature conditional feature vector, L is the number of feature values in the voltage level timing-related implicit feature vector and the chip real-time temperature timing-related implicit feature vector, represents the logarithmic function value with base 2, represents the exponential function with base e, the natural constant, is the voltage level semantic contribution degree, is the chip temperature semantic contribution degree, and are the normalized semantic contribution degrees of the voltage level and the normalized semantic contribution degree of the chip temperature, respectively, and are the time-sequence correlation implicit feature vectors of the modulated voltage level and the time-sequence correlation implicit feature vector of the real-time chip temperature after modulation, respectively, is the transposed vector of, d is the length of, represents the vector multiplication, is the function, is the voltage level-chip temperature time-sequence significant interaction fusion representation vector described above.

[0073] In the embodiment of the present application, the optimization instruction generation unit 135 is configured to obtain an optimization instruction based on the voltage level-chip temperature time-sequence significant interaction fusion representation, and the optimization instruction includes a recommended value of the switching frequency at the next time point. Specifically, in the embodiment of the present application, the optimization instruction generation unit is configured to: input the voltage level-chip temperature time-sequence significant interaction fusion representation vector into an optimization controller based on a decoder to obtain the optimization instruction, and the optimization instruction includes a recommended value of the switching frequency at the next time point. That is, decoding processing is performed using the voltage level-chip temperature time-sequence significant interaction fusion representation obtained by the significant interaction fusion of the voltage level time-sequence correlation implicit feature vector and the chip real-time temperature time-sequence correlation implicit feature vector, so as to automatically recommend the switching frequency value at the next time point. In this way, whether under different load conditions or in a changing ambient temperature, the switching frequency can be dynamically adjusted according to the real-time monitored data to ensure the best performance under various working conditions, which not only significantly improves the thermal management ability of the chip but also greatly enhances the degree of intelligent control.

[0074] It should be understood that in the technical solution of this application, the voltage level time - series associated implicit feature vector and the chip real - time temperature time - series associated implicit feature vector respectively represent the long - range - short - range time - series associated features of the voltage level of the input power supply of the wireless power transmission chip and the long - range - short - range time - series associated features of the chip real - time temperature. When the voltage level time - series associated implicit feature vector and the chip real - time temperature time - series associated implicit feature vector are subjected to conditional significant interaction fusion between features, there is a relatively significant fine - grained deviation in the feature distribution between the voltage level time - series associated implicit feature vector and the chip real - time temperature time - series associated implicit feature vector due to the source data modality difference and source data dimension difference. This will cause fine - grained expression distortion in the extraction of joint implicit features, resulting in sparsity and uneven distribution of the feature manifold of the voltage level - chip temperature time - series significant interaction fusion representation vector obtained by significant guidance interaction between features of conditional information in the high - dimensional feature space, and further affecting the accuracy of the optimization instruction obtained through the optimization controller based on the decoder.

[0075] Based on this, in a preferred embodiment, inputting the voltage level - chip temperature time - series significant interaction fusion representation vector into the optimization controller based on the decoder to obtain an optimization instruction includes: performing L random samplings on the voltage level - chip temperature time - series significant interaction fusion representation vector to obtain L chip state random perturbation eigenvalues; arranging the L chip state random perturbation eigenvalues in an orderly manner from large to small to obtain an ordered vector of chip state random perturbations; calculating the self - inner product of the voltage level - chip temperature time - series significant interaction fusion representation vector to obtain a chip state energy modulation eigenvalue; scaling the semantic intensity of the ordered vector of chip state random perturbations based on the chip state energy modulation eigenvalue to obtain an ordered vector of semantically enhanced chip state random perturbations; calculating the product between the ordered vector of semantically enhanced chip state random perturbations and the transposed vector of the voltage level - chip temperature time - series significant interaction fusion representation vector to obtain a chip state optimal example hint orthogonal modulation matrix; calculating the matrix product between the chip state optimal example hint orthogonal modulation matrix and the voltage level - chip temperature time - series significant interaction fusion representation vector to obtain an optimized voltage level - chip temperature time - series significant interaction fusion representation vector; inputting the optimized voltage level - chip temperature time - series significant interaction fusion representation vector into the optimization controller based on the decoder to obtain the optimization instruction.

[0076] The above - mentioned optimization process is expressed by the formula as follows:

[0077] ;

[0078] ;

[0079] ;

[0080] ;

[0081] ;

[0082] ;

[0083] ;

[0084] ;

[0085] Among them, and represent the th random sampling of the significantly interactive fusion representation vector of the voltage level-chip temperature timing to obtain the th chip state random perturbation eigenvalue, represents L times of random sampling, represents the Gaussian density probability function, represents the mean of the significantly interactive fusion representation vector of the voltage level-chip temperature timing, represents the significantly interactive fusion representation vector of the voltage level-chip temperature timing, represents calculating the mean of the vector, represents the variance of the significantly interactive fusion representation vector of the voltage level-chip temperature timing, is the eigenvalue at the th position in the significantly interactive fusion representation vector of the voltage level-chip temperature timing, is the length of the significantly interactive fusion representation vector of the voltage level-chip temperature timing, represents being sorted in descending order, represents the semantic enhanced chip state random perturbation ordered vector, represents the square of the two-norm of the vector, represents the chip state energy modulation eigenvalue, represents element-wise multiplication by position, represents matrix multiplication, represents the transpose vector of the significantly interactive fusion representation vector of the voltage level-chip temperature timing, represents the chip state optimal example hint orthogonal modulation matrix, represents the optimized significantly interactive fusion representation vector of the voltage level-chip temperature timing.

[0086] Accordingly, by performing \(L\) random samplings on the significantly interactive fusion representation vector of voltage level - chip temperature time series to obtain a set of new eigenvalue sets based on the original features but with slight variations, randomness is introduced to explore different aspects of the feature space, and by simulating the noise or variations in the actual data, it helps the model better generalize to unseen data. Then, these randomly perturbed eigenvalues of chip states are sorted in descending order to form a randomly perturbed ordered vector of chip states, to identify which perturbation directions might be the most important for the current problem. Such sorting not only helps the algorithm focus on the feature differences that can best distinguish different classes, but also helps reveal hidden data structures or patterns. Subsequently, the self - inner product of the significantly interactive fusion representation vector of voltage level - chip temperature time series is calculated to obtain the eigenvalue of chip state energy modulation, which reflects the overall strength or "energy" of the original feature vector, providing a benchmark for measuring the importance of features. Using this eigenvalue of chip state energy modulation as a reference point, semantic intensity scaling is performed on the previously formed randomly perturbed ordered vector of chip states to construct a semantically enhanced randomly perturbed ordered vector of chip states. In this way, those perturbations that appear more important in the original feature space are amplified, enhancing the model's sensitivity to key feature changes, while still retaining relatively small but potentially useful perturbation information.

[0087] Furthermore, the product between the semantically enhanced randomly perturbed ordered vector of chip states and the transpose of the significantly interactive fusion representation vector of voltage level - chip temperature time series is calculated to obtain the optimal exemplar - hint orthogonal modulation matrix of chip states, whose purpose is to find the way of interaction between features from two different perspectives. The formed matrix can be regarded as a transformation tool for capturing and representing complex linear relationships between features. Finally, by applying the above - constructed optimal exemplar - hint orthogonal modulation matrix of chip states to the original significantly interactive fusion representation vector of voltage level - chip temperature time series, the final feature structure modulation is completed, making the features more in line with the requirements of the objective function, reducing redundant information, and at the same time maximizing the retention of useful information to improve learning efficiency and prediction performance.

[0088] In summary, the optimization process of the feature manifold of the voltage level-chip temperature timing significant interaction fusion representation vector is elucidated. It extracts the best prompt examples based on fine-grained object retrieval with random perturbations, and then guides adversarial robustness learning based on the best prompt examples to complete the structure in the feature space of the high-dimensional manifold of the feature vector, so as to improve the detail and structure of the feature expression. In this way, the accuracy of the optimization instruction obtained through the decoder-based optimization controller is improved. In this way, whether under different load conditions or in changing ambient temperatures, the switching frequency can be dynamically adjusted according to the real-time monitored data, ensuring the best performance under various working conditions. This not only significantly improves the thermal management ability of the chip, but also greatly enhances the degree of intelligent control.

[0089] Specifically, in another embodiment of the present application, a circuit design diagram of a wireless power transmission chip is also provided. Figure 5 It is a circuit design diagram of a wireless power transmission chip according to an embodiment of the present application.

[0090] In summary, the wireless power transmission chip 100 based on the embodiment of the present application is elucidated. It obtains the time queue of the voltage level of the input power supply and the time queue of the real-time temperature of the chip, and uses deep learning data analysis and processing techniques to perform timing correlation implicit coding of the voltage level and the real-time temperature of the chip, so as to automatically recommend the switching frequency value at the next time point based on the conditional significant interaction fusion representation between the voltage level timing correlation implicit feature and the chip real-time temperature timing correlation implicit feature. In this way, whether under different load conditions or in changing ambient temperatures, the switching frequency can be dynamically adjusted according to the real-time monitored data, ensuring the best performance under various working conditions. This not only significantly improves the thermal management ability of the chip, but also greatly enhances the degree of intelligent control.

[0091] As described above, the wireless power transmission chip 100 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a wireless power transmission chip algorithm. In a possible implementation manner, the wireless power transmission chip 100 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the wireless power transmission chip 100 can be a software module in the operating system of the wireless terminal, or can be an application program developed for the wireless terminal; of course, the wireless power transmission chip 100 can also be one of the many hardware modules of the wireless terminal.

[0092] Alternatively, in another example, the wireless power transmitting chip 100 and the wireless terminal may also be separate devices, and the wireless power transmitting chip 100 can be connected to the wireless terminal through a wired and / or wireless network and transmit interaction information in accordance with a predefined data format.

[0093] Figure 6 FIG. is a flowchart of a method for a wireless power transmitting chip according to an embodiment of the present application. As Figure 6 shown, the method for a wireless power transmitting chip according to an embodiment of the present application includes: S110, obtaining a time queue of voltage levels of the input power supply; S120, obtaining a time queue of the real-time temperature of the chip; S130, performing time series encoding on the time queue of the voltage levels and the time queue of the real-time temperature of the chip to obtain a voltage level time series correlation implicit feature vector and a chip real-time temperature time series correlation implicit feature vector; S140, performing conditional significant interaction fusion between the voltage level time series correlation implicit feature vector and the chip real-time temperature time series correlation implicit feature vector to obtain a voltage level-chip temperature time series significant interaction fusion representation; S150, based on the voltage level-chip temperature time series significant interaction fusion representation, obtaining an optimization instruction, where the optimization instruction includes a recommended value of the switching frequency at the next time point.

[0094] Here, those skilled in the art can understand that the specific operations of each step in the above method of the wireless power transmitting chip have been described in detail in the description of the wireless power transmitting chip above with reference to Figures 1 to 4 and therefore, the repeated description thereof will be omitted.

[0095] The various implementations of the present disclosure have been described above. The above description is exemplary and not exhaustive. And it is not limited to the disclosed implementations. Many modifications and variations are obvious to those of ordinary skill in the art in the technical field without departing from the scope and spirit of the described implementations. The selection of the terms used herein is intended to best explain the principles of the implementations, practical applications, or improvements to the technology in the market, or to enable other ordinary skilled in the art in the technical field to understand the various implementation manners disclosed herein.

Claims

1. A wireless power transmission chip, characterized in that: include: A voltage identification module, used to identify the voltage level of the input power supply; Temperature detection module, used to detect the real-time temperature of the chip through a temperature sensor; An adaptive control module, configured to adaptively adjust the switching frequency based on the voltage level of the input power supply and the real-time temperature of the chip; Wherein, the adaptive control module comprises: A voltage level timing acquisition unit, used to acquire a time sequence of a voltage level of the input power supply; A real-time temperature timing acquisition unit, used to acquire a time sequence of the real-time temperature of the chip; A voltage level real-time temperature timing association unit, used for performing time series encoding on the time queue of the voltage level and the time queue of the chip real-time temperature to obtain a voltage level timing association implicit feature vector and a chip real-time temperature timing association implicit feature vector; A voltage level-chip temperature timing significant interaction unit, used for performing conditional feature significant interaction fusion on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector to obtain a voltage level-chip temperature timing significant interaction fusion representation; The optimization instruction generating unit is used to obtain the optimization instruction based on the voltage level-chip temperature timing significant interaction fusion representation, and the optimization instruction includes a recommended value of the switching frequency at the next time point.

2. The wireless power transmitting chip according to claim 1, characterized in that: The voltage level real-time temperature timing association unit is used to: input the time queue of the voltage level and the time queue of the chip real-time temperature into a time series encoder based on a forward LSTM model to obtain the voltage level timing association implicit feature vector and the chip real-time temperature timing association implicit feature vector.

3. The wireless power transmitting chip according to claim 2, characterized in that: The voltage level-chip temperature timing significant interaction unit includes: A voltage level-chip temperature implicit feature combining subunit, used for combining the voltage level timing-related implicit feature vector and the chip real-time temperature timing-related implicit feature vector by sharing implicit features to obtain a voltage level-chip temperature condition feature vector; The voltage level-chip temperature condition feature interaction subunit is used to perform feature weighted interaction based on conditional information on the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector based on the voltage level-chip temperature condition feature vector to obtain the voltage level-chip temperature timing significant interaction fusion representation.

4. The wireless power transmitter chip according to claim 3, characterized in that: The voltage level-chip temperature implicit feature combined subunit includes: A voltage level-chip temperature timing joint secondary subunit, used for inputting the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector into a joint implicit feature capture network to obtain a voltage level-chip temperature timing joint implicit feature vector; The voltage level-chip temperature condition feature generation secondary subunit is used to perform feature activation based on the Sigmoid function on the voltage level-chip temperature timing joint implicit feature vector to obtain the voltage level-chip temperature condition feature vector.

5. The wireless power transmitting chip according to claim 4, characterized in that: The voltage level-chip temperature timing is combined with a secondary subunit for: After adding the voltage level timing associated implicit feature vector and the chip real-time temperature timing associated implicit feature vector by position, the obtained voltage level-chip temperature timing sum vector is multiplied by the weight matrix and then added by position with the bias vector to obtain the voltage level-chip temperature timing joint interaction vector; The voltage level-chip temperature timing joint interaction vector is processed using a tanh function to obtain the voltage level-chip temperature timing joint implicit feature vector.

6. The wireless power transmitting chip according to claim 5, characterized in that: The voltage level-chip temperature condition characteristic interaction subunit includes: A voltage level semantic contribution calculation secondary subunit is used to calculate the voltage level semantic contribution of the voltage level timing associated implicit feature vector relative to the voltage level-chip temperature condition feature vector; A chip temperature semantic contribution calculation secondary subunit is used to calculate the chip temperature semantic contribution of the chip real-time temperature timing associated implicit feature vector relative to the voltage level-chip temperature condition feature vector; The voltage level chip temperature feature modulation secondary subunit is used to normalize the voltage level semantic contribution and the chip temperature semantic contribution, and use the normalized voltage level semantic contribution and the normalized chip temperature semantic contribution to weighted modulate the voltage level timing-associated implicit feature vector and the chip real-time temperature timing-associated implicit feature vector to obtain the modulated voltage level timing-associated implicit feature vector and the modulated chip real-time temperature timing-associated implicit feature vector; The voltage level-chip temperature interaction fusion secondary sub-unit is used to use the implicit feature vector associated with the modulation voltage level timing as the query vector, the implicit feature vector associated with the modulation chip real-time temperature timing as the key vector and the voltage level-chip temperature condition feature vector as the value vector, and input the implicit feature vector associated with the modulation voltage level timing, the implicit feature vector associated with the modulation chip real-time temperature timing and the voltage level-chip temperature condition feature vector into the feature-significant guided interaction module based on the converter structure to obtain the voltage level-chip temperature timing significant interaction fusion representation vector as the voltage level-chip temperature timing significant interaction fusion representation.

7. The wireless power transmitting chip according to claim 6, characterized in that: The voltage level semantic contribution calculation secondary subunit is used to: Calculate the voltage level semantic contribution vector by dividing the voltage level timing association implicit feature vector and the voltage level-chip temperature condition feature vector by the position point; Calculating a logarithmic function value with base 2 of the absolute value of each eigenvalue of the voltage level semantic contribution vector to obtain a voltage level semantic contribution logarithmic vector; Calculate the positional dot product of the voltage level time series associated implicit feature vector and the voltage level semantic contribution logarithmic vector, and perform positional multiplication on the obtained dot product vector to obtain the voltage level semantic contribution value; An exponential function with the natural constant e as the base and the voltage level semantic contribution value as the exponent is calculated to obtain the voltage level semantic contribution degree.

8. The wireless power transmitting chip according to claim 7, characterized in that: The voltage level chip temperature characteristic modulation secondary subunit is used to: Calculating a sum of the voltage level semantic contribution and the chip temperature semantic contribution to obtain a voltage level-chip temperature semantic contribution sum; Respectively dividing the voltage level semantic contribution and the chip temperature semantic contribution by the voltage level-chip temperature semantic contribution sum to obtain the normalized voltage level semantic contribution and the normalized chip temperature semantic contribution; Multiplying the voltage level time series associated implicit feature vector and the normalized voltage level semantic contribution by position points to obtain the modulated voltage level time series associated implicit feature vector; The chip real-time temperature timing-associated implicit feature vector is multiplied by the normalized chip temperature semantic contribution by position point to obtain the modulated chip real-time temperature timing-associated implicit feature vector.

9. The wireless power transmitting chip according to claim 8, characterized in that: The optimization instruction generating unit is used to: input the voltage level-chip temperature timing significant interaction fusion representation vector into a decoder-based optimization controller to obtain the optimization instruction, and the optimization instruction includes a recommended value of the switching frequency at the next time point.

10. A method for transmitting a wireless power supply chip according to any one of claims 1 to 9, characterized in that: include: Obtaining a time queue of a voltage level of the input power supply; A time queue for obtaining the real-time temperature of the chip; Performing time series encoding on the time queue of the voltage level and the time queue of the chip real-time temperature to obtain a voltage level timing-related implicit feature vector and a chip real-time temperature timing-related implicit feature vector; Performing conditional feature-inter-significant interaction fusion on the voltage level timing-related implicit feature vector and the chip real-time temperature timing-related implicit feature vector to obtain a voltage level-chip temperature timing-significant interaction fusion representation; Based on the voltage level-chip temperature timing significant interaction fusion representation, an optimization instruction is obtained, and the optimization instruction includes a recommended value of the switching frequency at the next time point.

Images