Air conditioning control method, device and storage medium

Abstract

The present invention discloses an air-conditioning control method, device, equipment and storage medium, and relates to the field of vehicle technology. The present invention includes obtaining the current environmental sensor data of the vehicle air-conditioning system, then performing data processing on the environmental sensor data to obtain target sensor data, and then inputting the target sensor data into a multi-output classification model for processing to determine the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensor data, wherein the multi-output classification model is trained based on user behavior data and environmental sensor data associated with the user behavior data. Finally, the vehicle air-conditioning system is adjusted according to the system operating parameters. Therefore, the dynamic self-adjustment of the air-conditioning system based on the system operating parameters preferred by the vehicle user obtained by model prediction can reduce the user's adjustment complexity and ensure that the in-vehicle environment is always in the comfort range preferred by the vehicle user.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioning, and in particular to an air conditioning control method, device and storage medium. Background Art

[0002] Currently, the vehicle air conditioning system is a key component in the car that directly interacts with the user. Usually, when it detects that the user has adjusted the operating mode of the air conditioning, it adjusts the air conditioning based on the operating mode set by the user.

[0003] When the external environment changes, the user needs to adjust the operating mode of the air conditioner multiple times to achieve the indoor air-conditioning environment desired by the current user, which increases the complexity of adjustment and reduces the user experience. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a vehicle air conditioning control method, device and storage medium that overcome the above problems or at least partially solve the above problems.

[0005] According to a first aspect of the present invention, an air conditioning control method is provided, which is applied to a vehicle air conditioning system. The method includes:

[0006] Obtain the current environmental sensor data of the vehicle air conditioning system;

[0007] performing data processing on the environmental sensing data to obtain target sensing data;

[0008] inputting the target sensor data into a multi-output classification model for processing to determine the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensor data, wherein the multi-output classification model is trained based on the user behavior data and environmental sensor data associated with the user behavior data;

[0009] The vehicle air conditioning system is adjusted according to the system operating parameters.

[0010] According to the second aspect of the present invention, there is further provided an electronic device, comprising:

[0011] one or more processors;

[0012] Memory;

[0013] One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to perform any of the above methods.

[0014] Based on the third aspect of the present invention, a computer-readable storage medium is further provided, which stores a computer program used in conjunction with an electronic device, and the computer program can be executed by a processor to implement any of the above methods.

[0015] Compared with the prior art, the present invention includes obtaining the current environmental sensor data of the vehicle air-conditioning system, then performing data processing on the environmental sensor data to obtain target sensor data, and then inputting the target sensor data into a multi-output classification model for processing to determine the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensor data, wherein the multi-output classification model is trained based on user behavior data and environmental sensor data associated with the user behavior data. Finally, the vehicle air-conditioning system is adjusted according to the system operating parameters. Therefore, the dynamic self-adjustment of the air-conditioning system based on the system operating parameters preferred by the vehicle user obtained by model prediction can reduce the user's adjustment complexity and ensure that the in-vehicle environment is always in the comfort range preferred by the vehicle user.

[0016] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be construed as limiting the present invention. The same reference numerals are used throughout the accompanying drawings to denote the same components.

[0018] In the attached figure:

[0019] Figure 1 This is a schematic flow chart of the steps of an air conditioning control method provided by an embodiment of the present invention;

[0020] Figure 2 1 is a schematic flow chart of another air conditioning control method provided by an embodiment of the present invention;

[0021] Figure 3 Schematic diagram of a multi-output classification model provided by an embodiment of the present invention;

[0022] Figure 4 Schematic diagram of the module structure of another multi-output classification model provided by an embodiment of the present invention;

[0023] Figure 51 is a flow chart of the training steps of a multi-output classification model provided by an embodiment of the present invention;

[0024] Figure 6 It is a structural schematic diagram of an air conditioning control device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0026] Reference Figure 1 , shows an air conditioning control method provided by an embodiment of the present invention, which is applied to a vehicle control system. The method may include:

[0027] S101: Acquire current environmental sensor data of the vehicle air-conditioning system.

[0028] In an embodiment of the present invention, the vehicle air conditioning system may include a vehicle terminal and various sensor devices. The air conditioning control method is applied to the vehicle terminal. When the vehicle terminal detects that the vehicle air conditioning system is started, the sensor data corresponding to each sensor device can be obtained.

[0029] For example, the sensing devices include but are not limited to temperature sensing devices, light sensing devices, humidity sensing devices, wind speed sensing devices, and carbon dioxide sensing devices. Correspondingly, there can be multiple sensing devices of each type. For example, if the temperature sensing device is installed outside the vehicle, the corresponding detection is the outside temperature of the vehicle. If the temperature sensing device is installed inside the vehicle, the corresponding detection is the inside temperature of the vehicle. And so on. Those skilled in the art can arrange the corresponding sensing devices according to the actual system requirements. And the sensing data detected by all sensing devices are input into the vehicle-mounted terminal as environmental sensing data. In one example, the environmental sensing data includes at least one of the following sensing types: inside vehicle temperature, vehicle speed, outside vehicle temperature, outside vehicle humidity, and outside vehicle light intensity.

[0030] S102: Process the environmental sensing data to obtain target sensing data.

[0031] In the embodiment of the present invention, data processing may include, but is not limited to, data cleaning, data interpolation, and data arrangement, thereby processing the environmental sensor data to obtain target sensor data. Compared with the environmental sensor data, the target sensor data includes the same sensor type.

[0032] S103. Input the target sensor data into a multi-output classification model for processing to determine the system operating parameters preferred by the vehicle user in the air-conditioned environment corresponding to the environmental sensor data, wherein the multi-output classification model is trained based on the user behavior data and the environmental sensor data associated with the user behavior data.

[0033] S104: Adjust the vehicle air conditioning system according to the system operating parameters.

[0034] In an embodiment of the present invention, the multi-output classification model can be understood as a model that outputs multiple prediction results. For example, for a vehicle-mounted air-conditioning system, its system operating parameters may include air circulation mode, air outlet mode, temperature value, and wind speed level. For example, the air circulation mode may include an internal circulation mode and an external circulation mode, and the air outlet mode may include a blow-off mode, a blow-off mode plus a lower air outlet mode, a separate lower air outlet mode, a lower air outlet mode, a front air outlet mode, and a separate front air outlet mode. Thus, after the target sensor data is input into the multi-output classification model for processing, the prediction results corresponding to the above-mentioned system operating parameters can be obtained. For example, the prediction results may be: external circulation mode; front air outlet mode; temperature 24.5°C; wind speed level 7.

[0035] Furthermore, user behavior data can be used to determine the system operating parameters of the vehicle air-conditioning system preferred by the user based on the driving environment under the environmental sensor data. Since the above-mentioned multi-output classification model is trained based on user behavior data and corresponding environmental sensor data, the system operating parameters predicted by the multi-output classification model are the system operating parameters preferred by the vehicle user in the current environmental sensor data. By adjusting the vehicle air-conditioning system according to these system operating parameters, the in-vehicle environment can be quickly adjusted to the comfort range desired by the vehicle user. This improves the driving experience of the vehicle user and reduces the complexity of the air-conditioning system adjustment.

[0036] Reference Figure 2 , shows another air conditioning control method provided by an embodiment of the present invention, the method may include:

[0037] S201: Acquire current environmental sensor data of the vehicle air-conditioning system.

[0038] In the embodiment of the present invention, the description of step S201 refers to the description of step S101. The target sensor data includes at least one of the following sensor types: vehicle interior temperature, vehicle speed, vehicle exterior temperature, vehicle exterior humidity, and vehicle exterior light intensity.

[0039] In another example, for some vehicle air-conditioning systems that can adjust the humidity inside the vehicle, the target sensing data may also include the following sensing type: humidity inside the vehicle.

[0040] S202: Process the environmental sensing data to obtain target sensing data.

[0041] In the embodiment of the present invention, data processing may include, but is not limited to, data cleaning, data interpolation, and data arrangement, thereby processing the environmental sensor data to obtain target sensor data. Compared with the environmental sensor data, the target sensor data includes the same sensor type.

[0042] In one example, environmental sensor data can be cleaned. Data cleaning can be understood as removing multiple environmental sensor data corresponding to the same user sensor type. For example, if a target sensor data item contains two or more vehicle interior temperature data items, one of the vehicle interior temperature data items is deleted. This process is repeated in this way to obtain the target sensor data.

[0043] In another example, data interpolation can be performed on the environmental sensor data after data cleaning. Data interpolation can be understood as adding environmental sensor data that the user lacks. For example, if the data on the temperature inside the car is missing in a target data, the current temperature inside the car can be re-obtained and added to this target data to complete the data interpolation.

[0044] In another example, the environmental sensor data that has completed data interpolation may be arranged in order according to different sensor types to determine target sensor data corresponding to the environmental sensor data.

[0045] The multi-output classification model can be understood as a model that outputs multiple prediction results. For example, for a vehicle air conditioning system, its system operating parameters may include air circulation mode, air outlet mode, temperature value, wind speed level, etc. For example, the air circulation mode may include internal circulation mode and external circulation mode, and the air outlet mode may include blow-off mode, blow-off mode with lower air outlet mode, separate lower air outlet mode, lower air outlet mode, front air outlet mode, and separate front air outlet mode.

[0046] Therefore, in the process of inputting the target sensor data into the multi-output classification model for processing, it is necessary to extract a large number of features and learn the correlation between the features through deep learning. Figure 3 As shown, the multi-output classification model includes a feature embedding module, a deep learning module and a prediction output module, wherein step S103 may further include the following steps: S203-S206.

[0047] S203: Input the target sensor data into the feature embedding module for feature mapping, and determine embedded feature data corresponding to the target sensor data.

[0048] In an embodiment of the present invention, the feature embedding module is responsible for converting the data features of the target sensor data into a feature embedding vector, and using the feature data carried by the feature embedding vector as embedded feature data, so that the subsequent deep learning module can perform the next step of feature cross-talk on the embedded feature data.

[0049] S204: Input the embedded feature data into the deep learning module for feature cross-pollination, and determine deep cross-pollination data corresponding to the embedded feature data.

[0050] In the embodiment of the present invention, the feature cross refers to the cross combination of features corresponding to different types or dimensions. The depth cross data refers to feature data obtained through multiple feature crosses.

[0051] In an optional embodiment of the invention, referring to Figure 4 As shown, the deep learning module includes a deep cross network (DCN) and a feedforward neural network (DNN). Therefore, step S204 may further include the following sub-steps:

[0052] The embedded feature data is input into the deep cross network to perform crossover of explicit features to obtain initial crossover data.

[0053] The initial cross-data is input into the feedforward neural network to perform cross-talk of implicit features, and deep cross-data corresponding to the initial cross-data is determined.

[0054] In the embodiment of the present invention, the display feature can be understood as a feature with practical significance and can be described and explained in words. The deep cross network performs data cross-talk on embedded feature data of different orders (whose feature type is display feature) to obtain initial cross-talk data corresponding to the embedded feature data. The initial cross-talk data integrates the features of the embedded feature data at different orders, which can improve the accuracy of model prediction.

[0055] For example, assuming that the feature vector corresponding to the embedded feature data is e, the calculation of the deep cross network at the L+1 layer can be as follows:

[0056]

[0057] In the above formula (1), W L+1 and b L+1is the learnable parameter of the L+1th layer.

[0058] The implicit features can be understood as having no practical meaning. Typically, the displayed features are matrix-decomposed based on different dimensions to obtain the implicit features corresponding to each displayed feature. For example, the initial cross-data is input into the feedforward neural network to perform cross-linking of implicit features to obtain deep cross-data. The deep cross-data output by the feedforward neural network can accurately express the semantic information corresponding to the target sensor data, thereby further improving the accuracy of model prediction.

[0059] For example, assuming that the feature vector corresponding to the initial cross data is Z, the cross process of the implicit features of the feedforward neural network can be as follows:

[0060] Z (1) =σ(W (1) Z+b (1) ) Formula (2) ......

[0062] Z (L) =σ(W (L) Z (L-1) +b (L) ) Formula (3)

[0063] When L is greater than or equal to 2 layers, the above formula (3) is applied for calculation. (L) is the output feature vector of the Lth layer of the feedforward neural network; W (L) is the weight matrix of the Lth layer of the feedforward neural network; b (L) is the bias value of the Lth layer of the feedforward neural network, and σ is the activation function. Those skilled in the art can determine the number of network layers of the deep cross network and the number of network layers of the feedforward neural network according to actual conditions, and no further restrictions are made here.

[0064] S205: Input the deep cross-data into the prediction output module for feature classification, and predict the system operating parameters preferred by the vehicle user under each sensing type.

[0065] In this embodiment of the present invention, the multi-output classification model can be understood as a model that outputs multiple prediction results. Thus, after the target sensor data is input into the multi-output classification model for processing, prediction results corresponding to the aforementioned system operating parameters can be obtained. For example, the prediction results may be: external circulation mode; front air outlet type; temperature 24.5°C; wind speed level 7.

[0066] Since the existing model usually only trains a single prediction result and only outputs one prediction result, considering that the deep cross data can accurately express the semantic information corresponding to the target sensor data, in an optional embodiment of the invention, referring to Figure 4 As shown, the prediction output module includes a binary classification network, a multi-classification network, and a regression network. Furthermore, the binary classification network, the multi-classification network, and the regression network are connected to the feedforward neural network respectively. This enables the sharing of semantic information of sensor data obtained through deep learning. The step S205 may also include the following sub-steps:

[0067] The deep cross-data is respectively input into the binary classification network, the multi-classification network and the regression network for feature classification, and the system operating parameters preferred by the vehicle user under each sensing type are predicted.

[0068] In an embodiment of the present invention, the deep cross data is respectively input into the binary classification network, the multi-classification network and the regression network for feature classification. Among them, those skilled in the art can determine the number of classifications output by the multi-classification network according to the actual business scenario, and no excessive restrictions are made here. For example, when the air outlet mode can include six modes such as blowing type, blowing type plus lower air outlet type, separated lower air outlet type, lower air outlet type, front air outlet type and separated front air outlet type, the corresponding number of multi-classifications is 6, that is, the multi-classification network outputs a 6-dimensional vector. Similarly, the temperature value and wind speed level are outputted by two regression networks respectively, and the output is a specific value. In the case where the air circulation mode can include an internal circulation mode and an external circulation mode, a 2-dimensional vector can be outputted by the binary classification network.

[0069] The binary classification network and the multi-classification network usually output characteristic probability data corresponding to different system operating parameters (i.e., the output vector of the corresponding network). For example, the air outlet mode corresponding to the maximum value in the above-mentioned 6-dimensional vector is selected as the system operating parameter predicted by the multi-classification network. The air circulation mode corresponding to the maximum value in the two-dimensional vector is selected as the system operating parameter predicted by the binary classification network. The output of the regression network is the system operating parameter predicted under a single mode. For example, predictions involving temperature values ​​and wind speed levels. For example, the system operating parameters obtained by the prediction output module can be as follows: external circulation mode; front air outlet type; temperature 24.5°C; wind speed level 7.

[0070] User behavior data can be used to determine the user's preferred system operating parameters for the vehicle air conditioning system based on the driving environment under the environmental sensor data. Since the multi-output classification model is trained based on the user behavior data and the corresponding environmental sensor data, the system operating parameters predicted by the multi-output classification model are the system operating parameters preferred by the vehicle user based on the current environmental sensor data.

[0071] S206: Adjust the vehicle air conditioning system according to the system operating parameters.

[0072] In this embodiment of the present invention, by adjusting the vehicle air conditioning system according to the system operating parameters, the vehicle interior environment can be quickly adjusted to the comfort range desired by the vehicle user, thereby improving the driving experience of the vehicle user and reducing the complexity of air conditioning system adjustment.

[0073] In an optional embodiment of the invention, the method may further include a training step of the multi-output classification model, wherein the training step includes:

[0074] S501. Acquire multiple target sensor data corresponding to the vehicle air-conditioning system and user behavior data corresponding to different vehicle users, wherein the user behavior data is system operating parameters obtained based on control adjustment operations of different vehicle users on environmental sensor data.

[0075] S502: Input each target sensor data into a multi-output classification model for processing, and predict the system operating parameters preferred by the corresponding vehicle user in the air-conditioning environment corresponding to the environmental sensor data.

[0076] S503 : Determine a model loss function of the multi-output classification model based on the predicted system operating parameters and the user behavior data of the corresponding vehicle user.

[0077] S504: Adjust the model parameters in the multi-output classification model according to the model loss function to determine a trained multi-output classification model.

[0078] In the embodiment of the present invention, the description of steps S501-S503 refers to the description of the above steps S203-S205, and it can also be understood that the model training process and the model application process are partially consistent.

[0079] The difference lies in determining the model loss function of the multi-output classification model based on the predicted system operating parameters and the system operating parameters in the user behavior data of the corresponding vehicle user. Since the prediction output module includes a binary classification network, a multi-classification network, and a regression network, the model loss function can be calculated for each of the binary classification network, the multi-classification network, and the regression network. The network parameters of the corresponding network are adjusted based on each model loss function.

[0080] In one example, for a binary classification network, the Sigmoid function can be used to calculate the final prediction result. The formula can be as follows:

[0081]

[0082] The estimated result obtained according to formula (4) is a two-dimensional vector, and the two eigenvalues ​​in the vector are both numbers between 0 and 1. In addition, a binary cross entropy function can be used as a loss function. The binary cross entropy loss function value JX1(Zy,Hy) can be calculated by the following formula:

[0083]

[0084] In the above formula (5), Hy is the prediction result, Zy is the system operation parameter in the user behavior data (also known as the true label of the sample), and N is the total number of samples. Therefore, the function value of the binary cross entropy function can be used as the loss function value of the binary classification network, and the network parameters of the binary classification network can be adjusted based on this loss function value.

[0085] In another example, for a multi-classification network, the Softmax function can be used to calculate the final result. The formula can be shown as follows:

[0086]

[0087] In the above formula (6), Z is the multidimensional vector output by the multi-classification network. M is the total number of categories corresponding to the multi-classification network. Therefore, the cross entropy function can be used as the loss function. For example, the calculation formula of the cross entropy loss function value JX2(Zy,Hy) can be as follows:

[0088]

[0089] In the above formula (7), Hyic is the probability value of each predicted category corresponding to the i-th sample; Zyic is a sign function in the range of 0-1. When the true category corresponding to the i-th sample is equal to Zyic, c takes 1, otherwise it takes 0. Therefore, the network parameters of the multi-classification network can be adjusted according to the cross entropy loss function value.

[0090] In another example, for the regression network, the prediction results obtained by predicting multiple target sensor data of different in-vehicle users can be obtained by using the mean square error loss as the loss function. The existing technology usually does not take into account that what users need is often a more comfortable value within a certain range, rather than a single value in the traditional regression problem. For example, for a car air conditioner with a wind speed level of 0-8, when the user expects the wind speed level of the 7th gear, often the 6th gear and the 8th gear can meet the user's needs. Therefore, the regression problem that the regression network needs to solve can be further transformed into a classification problem. Correspondingly, the regression network can include a multi-classification subnetwork. For example, the temperature value is 23.5℃. The number of output classifications of the multi-classification subnetwork, that is, multiple temperature levels, can be determined based on the number of temperature grades in the existing system. The predicted M-dimensional vectors are weighted respectively and used as the final prediction result, wherein the prediction result Hy of the regression network can be calculated according to the following formula:

[0091]

[0092] In the above formula (8), M represents the number of categories, pi represents the probability of the i-th category taking a value in the M-dimensional vector, and ci represents the numerical value of the i-th category.

[0093] The loss function of the above regression network can use the mean square error loss as the loss function. The mean square error loss function MSE(Zy,Hy) can be calculated using the following formula (9):

[0094]

[0095] In the above formula (9), Zy is the system operation parameter corresponding to the user behavior data, and N is the total number of target sensor data (which can also be understood as the total number of samples).

[0096] In the above process, a distance function loss can also be introduced. The distance loss function DIS(Zy,Hy) can be calculated by the following formula (10):

[0097]

[0098] In summary, the corresponding comprehensive loss function can be determined based on the mean square error loss function and the distance loss function. The comprehensive loss function L(Zy,Hy) can be calculated by the following formula (11):

[0099] L(Zy,Hy)=λMSE(Zy,Hy)+(1-λ)DIS(Zy,Hy) Formula (11)

[0100] The value of λ in the above formula (11) is between 0 and 1, which can be understood as a hyperparameter used to adjust the ratio of the two parts in the loss function, wherein the value of λ can be preset. Therefore, the network parameters of the regression network can be adjusted based on the above comprehensive loss function. It focuses on the processing of the regression problem of some system operating parameters (such as temperature value, humidity value, wind speed gear, etc.), and uses the processing method of converting the regression problem into a classification problem together with the additional distance loss to realize the processing of the user's comfort zone. Thereby, the range corresponding to the user's preferred comfort environment is expanded, and the in-car environment can quickly reach the user's preferred comfort environment.

[0101] In summary, an embodiment of the present invention discloses an air-conditioning control method, which may include obtaining the current environmental sensor data of a vehicle-mounted air-conditioning system, then performing data processing on the environmental sensor data to obtain target sensor data, and then inputting the target sensor data into a multi-output classification model for processing to determine the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensor data, wherein the multi-output classification model is trained based on user behavior data and environmental sensor data associated with the user behavior data. Finally, the vehicle-mounted air-conditioning system is adjusted according to the system operating parameters. Thus, the dynamic self-adjustment of the air-conditioning system based on the system operating parameters preferred by the vehicle user obtained by model prediction can reduce the user's adjustment complexity and ensure that the in-vehicle environment is always in the comfort range preferred by the vehicle user.

[0102] It should be noted that for the method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present application are not limited by the order of the actions described, because according to the embodiments of the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present application.

[0103] Reference Figure 6 , shows an air conditioning control device provided by an embodiment of the present invention, the device may include:

[0104] The data acquisition module 601 is used to acquire the current environmental sensor data of the vehicle air-conditioning system.

[0105] The data processing module 602 is used to process the environmental sensing data to obtain target sensing data.

[0106] The model processing module 603 is used to input the target sensor data into a multi-output classification model for processing, and determine the system operating parameters preferred by the vehicle user in the air-conditioned environment corresponding to the environmental sensor data, wherein the multi-output classification model is trained based on user behavior data and environmental sensor data associated with the user behavior data.

[0107] The system adjustment module 604 is used to adjust the vehicle air conditioning system according to the system operating parameters.

[0108] In an optional embodiment of the invention, the target sensor data includes at least one of the following sensor types: vehicle interior temperature, vehicle speed, vehicle exterior temperature, vehicle exterior humidity, and vehicle exterior light intensity.

[0109] In an optional embodiment of the invention, the multi-output classification model includes a feature embedding module, a deep learning module, and a prediction output module, and the model processing module 603 may include:

[0110] The embedding feature submodule is used to input the target sensor data into the feature embedding module for feature mapping, and determine the embedded feature data corresponding to the target sensor data.

[0111] The feature crossover submodule is used to input the embedded feature data into the deep learning module for feature crossover, and determine the deep crossover data corresponding to the embedded feature data.

[0112] The feature classification submodule is used to input the deep cross data into the prediction output module for feature classification, and predict the system operating parameters preferred by the vehicle user under each sensing type.

[0113] The operating parameter determination submodule is used to combine the system operating parameters corresponding to each sensing type preferred by the vehicle user to obtain the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensing data.

[0114] In an optional embodiment of the invention, the deep learning module includes a deep cross network and a feedforward neural network, and the feature cross submodule may include:

[0115] The first cross unit is used to input the embedded feature data into the deep cross network to perform crossover of explicit features to obtain initial crossover data.

[0116] The second crossover unit is configured to input the initial crossover data into the feedforward neural network to perform crossover of implicit features, and determine deep crossover data corresponding to the initial crossover data.

[0117] In an optional embodiment of the invention, the prediction output module includes a binary classification network, a multi-classification network, and a regression network, and the feature classification submodule may include:

[0118] The feature classification unit is used to input the deep cross data into the binary classification network, the multi-classification network and the regression network respectively for feature classification, and predict the system operating parameters preferred by the vehicle user under each sensing type.

[0119] In an optional embodiment of the invention, the data processing module 602 may include:

[0120] The data cleaning submodule is used to clean the environmental sensor data to obtain cleaned sensor data.

[0121] The data interpolation submodule is used to match the cleaned sensor data, and when the matched cleaned sensor data is missing, interpolate the cleaned sensor data to obtain target sensor data.

[0122] In an optional embodiment of the invention, the apparatus may further include a training module for the multi-output classification model, wherein the training module is configured to:

[0123] Acquire multiple target sensor data corresponding to the vehicle air conditioning system and user behavior data corresponding to different vehicle users, wherein the user behavior data is system operating parameters obtained based on control adjustment operations of different vehicle users for environmental sensor data.

[0124] Each target sensor data is input into a multi-output classification model for processing, and the system operating parameters preferred by the corresponding vehicle user in the air-conditioning environment corresponding to the environmental sensor data are predicted.

[0125] A model loss function of the multi-output classification model is determined based on the predicted system operating parameters and the user behavior data of the corresponding vehicle user.

[0126] According to the model loss function, the model parameters in the multi-output classification model are adjusted to determine a trained multi-output classification model.

[0127] In an optional embodiment of the invention, the multi-output classification model includes a prediction output module, the prediction output module includes a regression network, and the apparatus may further include a loss determination module, the loss determination module may be used to:

[0128] Based on the mean square error loss and the distance loss, a network loss function corresponding to the regression network is determined.

[0129] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0130] It is easy for those skilled in the art to think that any combination of the above embodiments is feasible, so any combination of the above embodiments is an implementation scheme of the present invention. However, due to space limitations, this specification will not describe them in detail here.

[0131] In the description provided herein, numerous specific details are described. However, it is understood that embodiments of the present invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques are not shown in detail so as not to obscure the understanding of this description.

[0132] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the claims, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Accordingly, the claims that follow the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of the present invention.

[0133] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition may be divided into multiple submodules or subunits or subcomponents. All features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed herein may be combined in any combination, except that at least some of such features and / or processes or units are mutually exclusive. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0134] An electronic device, comprising:

[0135] one or more processors;

[0136] Memory;

[0137] One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the method described in the above embodiment.

[0138] A computer-readable storage medium stores a computer program for use in conjunction with an electronic device, wherein the computer program can be executed by a processor to implement the method described in the above embodiment.

[0139] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, embodiments of the present invention may take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining software and hardware. Furthermore, embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0140] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the process in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0141] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0142] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1A step that specifies a function in one or more boxes.

[0143] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0144] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or terminal device that includes the element.

[0145] The above is a detailed introduction to the air-conditioning control method provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. An air conditioning control method, characterized in that: Applied to a vehicle air conditioning system, the method includes: Obtain the current environmental sensor data of the vehicle air conditioning system; performing data processing on the environmental sensing data to obtain target sensing data; inputting the target sensor data into a multi-output classification model for processing to determine the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensor data, wherein the multi-output classification model is trained based on the user behavior data and environmental sensor data associated with the user behavior data; Adjusting the vehicle air conditioning system according to the system operating parameters; The multi-output classification model includes a feature embedding module, a deep learning module, and a prediction output module. The target sensor data is input into the multi-output classification model for processing to determine the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensor data, including: Inputting the target sensor data into the feature embedding module for feature mapping to determine embedded feature data corresponding to the target sensor data; Inputting the embedded feature data into the deep learning module for feature cross-pollination, and determining deep cross-pollination data corresponding to the embedded feature data; Inputting the deep cross-data into the prediction output module for feature classification, and predicting the system operating parameters preferred by the vehicle user under each sensing type; combining the system operating parameters corresponding to the various sensing types preferred by the vehicle user to obtain the system operating parameters preferred by the vehicle user in the air-conditioning environment corresponding to the environmental sensing data; The deep learning module includes a deep cross network and a feedforward neural network, and the embedded feature data is input into the deep learning module for feature cross-pollination, and the deep cross-pollination data corresponding to the embedded feature data is determined, including: Inputting the embedded feature data into the deep cross network to perform crossover of explicit features to obtain initial crossover data; The initial cross-data is input into the feedforward neural network to perform cross-talk of implicit features, and deep cross-data corresponding to the initial cross-data is determined.

2. The air conditioning control method according to claim 1, characterized in that: The target sensor data includes at least one of the following sensor types: vehicle interior temperature, vehicle speed, vehicle exterior temperature, vehicle exterior humidity, and vehicle exterior light intensity.

3. The air conditioning control method according to claim 1, wherein: The prediction output module includes a binary classification network, a multi-classification network, and a regression network. The deep cross-data is input into the prediction output module for feature classification to predict the system operating parameters preferred by the vehicle user under each sensing type, including: The deep cross-data is respectively input into the binary classification network, the multi-classification network and the regression network for feature classification, and the system operating parameters preferred by the vehicle user under each sensing type are predicted.

4. The air conditioning control method according to claim 1, wherein: The processing of the environmental sensor data to obtain target sensor data includes: performing data cleaning on the environmental sensor data to obtain cleaned sensor data; The cleaned sensor data are matched, and when it is found that the cleaned sensor data are missing, the cleaned sensor data are interpolated to obtain target sensor data.

5. The air conditioning control method according to claim 1, characterized in that: The method further comprises a step of training the multi-output classification model, the training step comprising: Acquire multiple pieces of target sensor data corresponding to the vehicle air conditioning system and user behavior data corresponding to different vehicle users, wherein the user behavior data is system operating parameters obtained based on control adjustment operations of different vehicle users in response to environmental sensor data; Input each target sensor data into a multi-output classification model for processing, and predict the system operating parameters preferred by the corresponding vehicle user in the air-conditioning environment corresponding to the environmental sensor data; Determining a model loss function of the multi-output classification model based on the predicted system operating parameters and user behavior data of the corresponding vehicle user; According to the model loss function, the model parameters in the multi-output classification model are adjusted to determine a trained multi-output classification model.

6. The air conditioning control method according to claim 5, characterized in that: The multi-output classification model includes a prediction output module, the prediction output module includes a regression network, and the method further includes: Based on the mean square error loss and the distance loss, a network loss function corresponding to the regression network is determined.

7. An electronic device comprising: one or more processors; Memory; One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the method according to any one of claims 1 to 6.

8. A computer-readable storage medium storing a computer program for use in conjunction with an electronic device, wherein the computer program can be executed by a processor to implement the method according to any one of claims 1 to 6.

Images