Method, device, processor and electronic device for predicting environmental comfort

Abstract

The application discloses a kind of environmental comfort degree prediction method, device, processor and electronic equipment.It relates to artificial intelligence field, the method comprises: determining the M sampling points of indoor space, and the data of N environmental influence indicators is collected from each sampling point, obtain M group of environmental index data, wherein M, N are positive integers;Each group of environmental index data in M group of environmental index data is input into neural network model respectively, and the initial comfort level of M sampling points is output;The spatial coordinates of M sampling points and the initial comfort level are jointly input into Kriging model, and the comfort information of indoor space is obtained.By the present application, the problem of low accuracy in predicting indoor space comfort in related art is solved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, and more specifically, to a method, apparatus, processor, and electronic device for predicting environmental comfort. Background Technology

[0002] With the continuous development of technology and the maturation of the Internet of Things (IoT), more and more intelligent devices are appearing in all aspects of people's lives and work, providing convenient services. People are also increasingly demanding higher quality living environments. As the most critical system for indoor environmental regulation, the air conditioning system has evolved from simply meeting basic cooling and heating needs to providing a more comfortable environment and achieving environmentally friendly operation. This has driven the development of air conditioning systems towards higher efficiency, energy saving, and comfort. The intelligentization of air conditioning systems makes it possible to predict and regulate indoor comfort, contributing to a higher quality of life.

[0003] Although regulating and predicting environmental comfort presents many challenges, experts and scholars both domestically and internationally have achieved certain results in comfort research and air conditioning control. They have utilized the Predicted Mean Vote (PMV), a predictive mean value that reflects human thermal comfort, to predict environmental comfort. This parameter provides a foundation for predicting and regulating environmental comfort, and under truly PMV-based control requirements, it also reduces much unnecessary energy consumption.

[0004] However, predicting indoor environmental comfort based on PMV values ​​requires adjusting various environmental variables involved in the index. This index uses a non-linear calculation model, which not only requires multiple iterations and involves a complex calculation process, but also makes it difficult to measure some of the parameters involved. Furthermore, prediction methods such as trend analysis and regression analysis are insufficient to accurately predict comfort results. Existing backpropagation (BP) neural networks require a large amount of test data and cannot reflect the distribution and changes in comfort within three-dimensional space; moreover, the lengthy model training process makes timely system adjustments difficult. In addition, if the indoor space is too large or has a complex and varied layout, and there is high traffic, environmental conditions can change rapidly, and the number of usable measurement points is limited, making it difficult to rationally arrange the measurement points.

[0005] There is currently no effective solution to the problem of low accuracy in predicting indoor space comfort in related technologies. Summary of the Invention

[0006] The main objective of this application is to provide a method, apparatus, processor, and electronic device for predicting environmental comfort, in order to solve the problem of low accuracy in predicting indoor space comfort in related technologies.

[0007] To achieve the above objectives, according to one aspect of this application, a method for predicting environmental comfort is provided. The method includes: determining M sampling points in an indoor space, and collecting data on N environmental impact indicators from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers; inputting each set of environmental indicator data into a neural network model to output the initial comfort level of the M sampling points; and inputting the spatial coordinates of the M sampling points and the initial comfort level into a Kriging model to obtain the comfort information of the indoor space.

[0008] Optionally, determining the M sampling points of the indoor space includes: dividing the indoor space into dimensions, dividing the indoor space of each dimension into non-overlapping intervals to obtain M sampling spaces; and randomly selecting a sampling point from each of the M sampling spaces according to the Latin hypercube sampling algorithm to obtain M sampling points.

[0009] Optionally, the neural network model includes an input layer, a hidden layer, a summing layer, and an output layer. M sets of environmental indicator data are input into the neural network model, and the output of the initial comfort levels for M sampling points includes: for each sampling point, inputting N environmental indicator data of the sampling point into the input layer for processing, inputting the processed data into the hidden layer for processing to obtain initial comfort data; inputting the initial comfort data into the summing layer for processing, and inputting the processed data into the output layer for processing to obtain the initial comfort level of the sampling point.

[0010] Optionally, for each sampling point, the N environmental index data of the sampling point are input into the input layer for processing, and the processed data are input into the hidden layer for processing to obtain the initial comfort data. This includes: determining the first weight of the N environmental index data of the sampling point based on the spatial coordinates of the sampling point, and determining the first activation function; performing weighted calculation on each environmental index data with the matched first weight to obtain N weighted environmental index data, and inputting the N weighted environmental index data into the first activation function to calculate the initial comfort data of the sampling point.

[0011] Optionally, the initial comfort data is input into a summation layer for processing, and the processed data is input into an output layer for processing to obtain the initial comfort level of the sampling point. This includes: determining the second weight of the N initial comfort data of the sampling point based on the spatial coordinates of the sampling point, and calculating the weighted average of each initial comfort data with the comfort data of the corresponding type based on the second weight to obtain N comfort data; selecting the comfort data with the highest probability density from the N comfort data, and determining the comfort level associated with the comfort data with the highest probability density as the initial comfort level of the sampling point.

[0012] Optionally, the spatial coordinates of M sampling points and the initial comfort level are input into the Kriging model to obtain the comfort information of the indoor space, including: calculating the distance between every two sampling points based on the spatial coordinates of the sampling points to obtain Z sampling distances, where Z is a positive integer; determining the regression model and the second activation function of the Kriging model; standardizing the Z sampling distances to obtain Z standardized sampling distances, and standardizing the comfort values ​​associated with the M initial comfort levels to obtain M standardized initial comfort data; fitting the Z standardized sampling distances and M standardized initial comfort data using the regression model, and inputting the fitted data into the second activation function to calculate the comfort information.

[0013] Optionally, after obtaining the comfort information of the indoor space, the method further includes: determining a comfort adjustment strategy for the indoor space based on the comfort information, and triggering a start command to execute the comfort adjustment strategy.

[0014] To achieve the above objectives, according to another aspect of this application, an environmental comfort prediction device is provided. The device includes: a determining unit, configured to determine M sampling points in an indoor space and collect data from N environmental impact indicators from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers; a first input unit, configured to input each set of environmental indicator data from the M sets into a neural network model and output the initial comfort level of the M sampling points; and a second input unit, configured to input the spatial coordinates of the M sampling points and the initial comfort level into a Kriging model to obtain comfort information of the indoor space.

[0015] According to another aspect of the present invention, a processor is also provided, which is used to run a program, wherein the program controls a device containing a non-volatile storage medium to perform an environmental comfort prediction method during runtime.

[0016] According to another aspect of the present invention, an electronic device is also provided, comprising one or more processors and a memory; the memory stores computer-readable instructions, and the processor is configured to execute the computer-readable instructions, wherein the computer-readable instructions, when executed, perform a method for predicting environmental comfort.

[0017] This application employs the following steps: M sampling points are determined within the indoor space, and N environmental impact indicators are collected from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers; each set of environmental indicator data is input into a neural network model, outputting the initial comfort level of the M sampling points; the spatial coordinates of the M sampling points and the initial comfort level are input into a Kriging model to obtain the comfort information of the indoor space. This solves the problem of low accuracy in predicting indoor space comfort in related technologies. By determining the sampling points within the indoor space and sampling environmental indicator data at each point, inputting the obtained data into a neural network model, and then inputting the output of the neural network model into a Kriging model, the comfort information of the indoor space is finally output, thus achieving the effect of accurately predicting the comfort of the indoor space. Attached Figure Description

[0018] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a flowchart of a method for predicting environmental comfort according to an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of a neural network model provided according to an embodiment of this application;

[0021] Figure 3 This is a schematic diagram of a kriging model provided according to an embodiment of this application;

[0022] Figure 4 This is a schematic diagram of an optional method for predicting environmental comfort provided according to an embodiment of this application;

[0023] Figure 5 This is a schematic diagram of an environmental comfort prediction device provided according to an embodiment of this application;

[0024] Figure 6 This is a schematic diagram of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0025] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0028] It should be noted that all information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) involved in this disclosure are information and data authorized by the user or fully authorized by all parties.

[0029] The present invention will now be described in conjunction with preferred implementation steps. Figure 1 This is a flowchart of a method for predicting environmental comfort according to embodiments of this application, such as... Figure 1 As shown, the method includes the following steps:

[0030] Step S101: Determine M sampling points in the indoor space, and collect data on N environmental impact indicators from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers.

[0031] Specifically, the indoor space can be the branch of a financial institution. In order to provide users with a good indoor environment in the branch, the air conditioning system in the branch needs to be intelligently adjusted. The intelligent adjustment of the air conditioning system first requires predicting the environmental comfort of the indoor scene. After obtaining the prediction results, adjustment instructions are generated, and the air conditioning system is adjusted through the adjustment instructions.

[0032] Comfort levels can be predicted using predicted average evaluation indices from different locations within an indoor space. These indices are then used to calculate comfort levels, which are subsequently used to adjust the air conditioning system. It's important to note that the predicted average evaluation index, also known as the Predicted Mean Vote (PMV), is an evaluation indicator used to characterize human thermal response. The PMV value is a numerical value derived from environmental impact indicators and represents the average temperature sensation experienced by most people in the same environment. Environmental impact indicators can include both environmental factors and human factors. Environmental factors can include dry-bulb temperature, relative humidity, mean radiant temperature, and air velocity, while human factors can include human metabolic rate and clothing thermal resistance.

[0033] Furthermore, if the indoor setting is a large space, it is difficult to maintain consistency in the predicted average evaluation index at different locations within the space, and the distribution of the predicted average evaluation index is also difficult to predict through simulation. If the same predicted average evaluation index is used to predict comfort, the accuracy of the calculated prediction results will be low, making it difficult to achieve precise control of the air conditioning system. Therefore, it is necessary to sample environmental impact indicators at different locations within the indoor space to obtain multiple sets of environmental indicator data. By using multiple sets of environmental indicator data to predict comfort, the accuracy of the prediction results can be improved.

[0034] Step S102: Input each set of environmental index data from the M sets of environmental index data into the neural network model, and output the initial comfort level of the M sampling points.

[0035] Specifically, after sampling multiple sets of environmental indicator data from different locations within the indoor space, these data are input into a neural network model. The model then calculates the predicted average evaluation index for each sampling point, i.e., the initial comfort level for each point. For example, the neural network model can be a PNN (Pulse Neutron Neutron) model. A PNN is a feedforward neural network model based on Bayesian decision theory, offering advantages such as easy training and fast convergence, making it suitable for real-time processing. Using a PNN model, the initial comfort level for each sampling point can be accurately predicted.

[0036] Step S103: Input the spatial coordinates of the M sampling points and the initial comfort level into the Kriging model to obtain the comfort information of the indoor space.

[0037] Specifically, after obtaining the initial comfort level of each sampling point, the spatial coordinates of all sampling points and the initial comfort level are input into the Kriging model. The Kriging model then outputs the overall comfort information of the indoor space. The Kriging model is a model based on the Kriging interpolation method, also known as a Gaussian process. It consists of a regression model and a nonparametric stochastic model. The parametric model is used to calculate and predict the overall value of the variable or object, while the nonparametric stochastic model is used to establish the correlation between the predicted point and the surrounding sample points, thereby achieving local correction of the predicted point and estimating the comfort level of the sampling location. This results in the output of accurate comfort information for the indoor space.

[0038] The environmental comfort prediction method provided in this application determines M sampling points in an indoor space and collects N environmental impact indicators from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers. Each set of environmental indicator data is input into a neural network model, outputting the initial comfort level of the M sampling points. The spatial coordinates of the M sampling points and the initial comfort level are then input into a Kriging model to obtain the comfort information of the indoor space. This method solves the problem of low accuracy in predicting indoor space comfort in related technologies. By determining sampling points in the indoor space and sampling environmental indicator data at each point, inputting the obtained data into a neural network model, and then inputting the output of the neural network model into a Kriging model, the method finally outputs the comfort information of the indoor space, thereby achieving the effect of accurately predicting indoor space comfort.

[0039] To obtain comprehensive environmental impact indicators of various types in an indoor space, it is necessary to reasonably determine the sampling points within the indoor space. Optionally, in the environmental comfort prediction method provided in this application embodiment, determining the M sampling points of the indoor space includes: dividing the indoor space into dimensions, dividing the indoor space of each dimension into non-overlapping intervals to obtain M sampling spaces; and randomly selecting a sampling point from each of the M sampling spaces according to the Latin hypercube sampling algorithm to obtain M sampling points.

[0040] Specifically, the first step is to determine the number of sampling points to be preset, and then determine the dimensional data of the indoor space to be divided based on the number of sampling points. For example, if 27 sampling points need to be set, the space can be divided into three dimensions based on the area of ​​the indoor space and the number of sampling points.

[0041] Furthermore, based on the dimensionality, the interior space in each dimension is divided into multiple non-overlapping intervals according to the building layout, thus obtaining multiple non-overlapping sampling spaces for the entire interior space. Further, the Latin Hypercube Sampling (LHS) algorithm is used to randomly select points in each sampling space, resulting in multiple sampling points. The LHS algorithm is a stratified sampling method that approximates random sampling from a multivariate parameter distribution, thereby achieving equal probability sampling of the entire space. By using the LHS method, which maintains statistical significance with small-scale samples, equal probability sampling of the entire space can be achieved, laying the foundation for calculating interior space comfort information.

[0042] The initial comfort level of each sampling point is calculated by processing each set of environmental index data through the various layers of the neural network model. Optionally, in the environmental comfort prediction method provided in this application embodiment, the neural network model includes an input layer, a hidden layer, a summing layer, and an output layer. The initial comfort level of M sampling points is output by inputting M sets of environmental index data into the neural network model respectively and outputting the initial comfort level of M sampling points. For each sampling point, the N environmental index data of the sampling point are input into the input layer for processing, and the processed data is input into the hidden layer for processing to obtain the initial comfort data. The initial comfort data is input into the summing layer for processing, and the processed data is input into the output layer for processing to obtain the initial comfort level of the sampling point.

[0043] Specifically, after obtaining the spatial coordinates of the sampling points and collecting data on multiple environmental impact indicators from each point, the collected data needs to undergo data processing and normalization. For example, after obtaining data on metabolic rate and clothing thermal resistance, parameters need to be converted and calculated based on the movement and working state of the monitored objects in the indoor space and the clothing they are wearing. The resulting data is then used as network input. Similarly, duplicate and redundant sampling information needs to be removed from the data to ensure its uniqueness. Furthermore, missing or ambiguous data acquired during the acquisition process needs to be processed, using methods such as averaging and median imputation to ensure that each sampling point can collect data on all impact indicators. Through these methods, the data input into the neural network model has appropriate units and orders of magnitude, effectively reducing data variability and conflict.

[0044] Furthermore, the processed environmental indicator data is input into a neural network model. The neural network model performs weighted calculations on the input data, and the results of the weighted calculations are processed by an activation function to obtain the processed data. The processed data is then summed with data of the same comfort level. Based on the summation result, the probability that the comfort level of the sampling point belongs to each comfort level is determined, thus obtaining the initial comfort level of the sampling point.

[0045] This embodiment uses a neural network model to calculate the initial comfort level of each sampling point, laying a data foundation for accurately calculating the current environmental comfort and timely adjusting and controlling the air conditioning system.

[0046] The hidden layer of the neural network model contains a first activation function. Optionally, in the environmental comfort prediction method provided in this application embodiment, for each sampling point, the N environmental index data of the sampling point are input into the input layer for processing, and the processed data is input into the hidden layer for processing to obtain initial comfort data. This includes: determining the first weight of the N environmental index data of the sampling point based on the spatial coordinates of the sampling point, and determining the first activation function; performing weighted calculation on each environmental index data with the matched first weight to obtain N weighted environmental index data, and inputting the N weighted environmental index data into the first activation function to calculate the initial comfort data of the sampling point.

[0047] Specifically, the environmental index data input into the neural network model can be represented by Xi (i = 1, 2, 3, 4), where X1, X2, X3, and X14 represent dry-bulb temperature, relative humidity, radiation temperature, and air velocity, respectively. Yi (i = 1, 2) represents metabolic rate and clothing thermal resistance. The initial comfort data can be represented by T(x), which can be expressed as: T(x) = F(X1, X2, X3, X4, Y1, Y2).

[0048] Figure 2 This is a schematic diagram of a neural network model provided according to an embodiment of this application, such as... Figure 2 As shown, the neural network model includes an input layer, a hidden layer, a summing layer, and an output layer. After receiving the input environmental index data, the input layer determines the corresponding first weight for each environmental index data based on the spatial location of each sampling point. For example, if the sampling point is located on the ceiling, the first weight indicating the metabolic rate in the environmental index data is set to a smaller value, and the first weight indicating the relative humidity in the environmental index data is set to a larger value; if the sampling point is located at the service window in the indoor space, the first weight of the environmental index data indicating the dry-bulb temperature in the environmental index data is set to a larger value.

[0049] Furthermore, the input layer inputs the environmental indicator data into the hidden layer, where each environmental indicator data is weighted and calculated with its matching first weight, and the calculated weighted environmental indicator data is input into the selected first activation function.

[0050] It should be noted that activation functions are used to introduce nonlinearity, enabling neural network models to better solve more complex problems. These functions can include Gaussian functions, exponential functions, and linear functions, as shown in Table 1.

[0051] Table 1

[0052]

[0053] Furthermore, substituting the value of the first activation function into the following formula, the initial comfort data of the sampling points are calculated:

[0054]

[0055] Where, φ ij (x) represents the initial comfort data, x ij Let σ be the first weight for each environmental indicator data point, d be the smoothing factor, d be the distance between each sampling point and the center of the indoor space, and X be the environmental indicator data (.). T This indicates the transpose operation.

[0056] Optionally, in the environmental comfort prediction method provided in this application embodiment, the initial comfort data is input into a summation layer for processing, and the processed data is input into an output layer for processing to obtain the initial comfort level of the sampling point. This includes: determining the second weight of N initial comfort data of the sampling point based on the spatial coordinates of the sampling point, and calculating a weighted average of each initial comfort data and the comfort data of the corresponding type based on the second weight to obtain N comfort data; selecting the comfort data with the highest probability density from the N comfort data, and determining the comfort level associated with the comfort data with the highest probability density as the initial comfort level of the sampling point.

[0057] Specifically, the summation layer in the neural network model serves as an intermediate layer, establishing the connection between the hidden layers and the output layer. Each "neuron" in the summation layer represents a comfort level, which includes comfortable, cold, cool, slightly cold, slightly warm, warm, and hot. The relationship between each comfort level and the corresponding comfort data is shown in Table 2.

[0058] Table 2

[0059]

[0060] The summation layer connects neurons in the hidden layer that belong to the same comfort level, and determines the second weight of the initial comfort data for each sampling point based on the spatial coordinates of the sampling points. The weighted average of the initial comfort data calculated by the hidden layer and the second weight matched to each initial comfort data point is then calculated, and the output result can be obtained using the following formula:

[0061]

[0062] Where vi refers to the weighted average of a certain comfort level, i.e., the comfort data; L is the number of initial comfort data points for that comfort level; j represents the j-th comfort level; and φ ij (x) represents the initial comfort data.

[0063] Furthermore, the summing layer passes the calculated result to the output layer and selects the comfort data with the highest probability density, i.e., T(x) = argmax(vi), where the probability density is determined by the first weight and the second weight. After obtaining the comfort data with the highest probability density, the data is rounded to obtain the closest PMV value, and this PMV value is used as the initial comfort level for that sampling point.

[0064] It should be noted that before predicting the comfort level of sampling points using a neural network model, the model needs to be trained and validated. After training using environmental index data from multiple historical sampling points and the actual comfort level of each sampling point, the accuracy of the PMV index is used as the evaluation metric for the model, and the error probability is calculated. The simulation test results of the environmental comfort model are compared with the pre-collected object comfort levels. Through cross-validation, the accuracy and model are recorded after each training iteration. The training samples and model parameters are continuously adjusted to achieve the highest accuracy, and this model is then used as the final neural network model for comfort prediction.

[0065] The comfort information of the indoor space is calculated by the Kriging model. Optionally, in the environmental comfort prediction method provided in this application embodiment, the spatial coordinates of M sampling points and the initial comfort level are input into the Kriging model to obtain the comfort information of the indoor space, including: calculating the distance between every two sampling points based on the spatial coordinates of the sampling points to obtain Z sampling distances, where Z is a positive integer; determining the regression model and the second activation function of the Kriging model; standardizing the Z sampling distances to obtain Z standardized sampling distances, and standardizing the comfort values ​​associated with the M initial comfort levels to obtain M standardized initial comfort data; fitting the Z standardized sampling distances and the M standardized initial comfort data using the regression model, and inputting the fitted data into the second activation function to calculate the comfort information.

[0066] Figure 3 This is a schematic diagram of a kriging model provided according to an embodiment of this application, such as... Figure 3 As shown, the dataset X and the result dataset Y, i.e., the spatial coordinates of the sampling points and the initial comfort level of each sampling point, are input into the Kriging model. The distance between every two sampling points is calculated, and the initialization parameter β and activation function of the Kriging model are determined. The data is then standardized using the z-score standardization method to obtain the standardized sampling distance and standardized initial comfort data. The transformation function is: x * = (x-μ) / σ; where μ is the mean of the data and σ is the standard deviation of the data.

[0067] A regression model was used to fit the standardized sampling distance and standardized initial comfort data, and the Levenberg-Marquardt method was called to solve the model. This established the correlation between the sampling point and the surrounding sampling points, enabling the determination of the comfort level of the sampling location and thus obtaining the comfort information of the indoor space.

[0068] It should be noted that after obtaining the comfort level of each sampling point, in order to establish the relationship between each sampling point and adjacent sampling points, and to correct the sampling points and adjust the comfort level of the sampling location, a Kriging model is needed for smoothing. The Kriging model can be expressed as y(x)=F(β,x)+z(x)=f T (x)β+z(x).

[0069] Here, F(β,x) is a polynomial form of linear combination, which can be decomposed into a basis function vector f(x) and a linear regression coefficient vector β. The linear regression coefficient vector β can be determined by the regression model. It should be noted that, due to the high degree of nonlinearity between the spatial location of the sampling points and the PMV value, a linear regression model can be used.

[0070]

[0071] Here, z(x) is a random distribution process, which is essentially a random variable dependent on x. The expected value of the sampled points in the global scope satisfies E[z(x)]=0. The correlation between any two points in space is represented by the covariance. Generally speaking, the closer the sample points are, the stronger the correlation.

[0072] Specifically, the comfort information of the indoor space is calculated in the following ways:

[0073] First, calculate the predicted response values ​​for the sampling points: The deviation between the predicted value and the actual value is: The standard deviation is: Where c is the estimated parameter to be solved, and R is the activation function.

[0074] Furthermore, by introducing the Lagrange multiplier: L(c,λ)=σ 2 (1+c T Rc-2c T r)-λ T (F T cf(x)); the mean square error equation with minimum deviation can be obtained: L(c,λ)=σ 2 (1+c T Rc-2c T r)-λ T (F T Let cf(x)); let the derivative L(c,λ) ′ c =2σ 2 From (Rc-r)-Fλ=0, we can obtain: In the formula, When f T With (x) a fixed form and p specified, the predicted value can be estimated using the maximum likelihood estimation method with the sampled points: β * =(F T R -1 F) -1 F T R -1 Y.

[0075] It should be noted that before using the Kriging model to determine the comfort information of indoor spaces, the model needs to be trained, and the comfort level of actual test subjects should be compared and analyzed with the prediction results of three-dimensional space. The training samples and correction factors of the model should be gradually adjusted to achieve accurate prediction of the model.

[0076] In this embodiment, after determining the parameters β and activation function of the Kriging model, the Kriging model is derived by introducing Lagrange multipliers, thereby obtaining the predicted value of the sampling point, which is the comfort information of the sampling point.

[0077] After obtaining comfort information, an adjustment strategy needs to be generated. Optionally, in the environmental comfort prediction method provided in this application embodiment, after obtaining the comfort information of the indoor space, the method further includes: determining the comfort adjustment strategy of the indoor space based on the comfort information, and triggering a start command to execute the comfort adjustment strategy.

[0078] Specifically, after using a comfort model obtained by combining a neural network model and a Kriging model to predict the comfort information of an indoor space, a comfort adjustment strategy for that space is generated from the comfort information. For example, if the comfort information indicates that the indoor space temperature is high and the relative humidity is low, the generated comfort adjustment strategy could be to lower the air conditioning temperature and turn on the humidifier.

[0079] Furthermore, a start command is generated using a comfort adjustment strategy. After the air conditioning system or environmental system receives the start command of the comfort adjustment strategy, it executes the command to adjust the overall indoor environment.

[0080] This application also provides an optional method for predicting environmental comfort. Figure 4 This is a schematic diagram of an optional environmental comfort prediction method provided according to an embodiment of this application, such as... Figure 4 As shown:

[0081] First, determine the number of sampling points for the indoor environment. Then, based on the number of sampling points, determine the dimensional data of the indoor space to be divided. Next, use a data acquisition system to randomly select points in each sampling space to obtain data on multiple environmental impact indicators from multiple sampling points. Further, the collected data needs to be processed and normalized. For example, missing or ambiguous data from the data acquisition process can be processed by using methods such as averaging or median imputation to ensure that data on all impact indicators are collected at each sampling point. The processed data is then stored in a database.

[0082] Furthermore, the data stored in the database is input into the neural network model of the comfort prediction system. The neural network model is used to perform weighted calculations on the input data, and after processing through activation functions, the processed data is obtained. The processed data is then summed with data of the same comfort level to determine the probability of each comfort level, thereby obtaining the initial comfort level of the sampling point.

[0083] Furthermore, the spatial coordinates of the sampling points and the initial comfort level of each sampling point are input into the Kriging model of the comfort prediction system. The distance between every two sampling points is calculated, and the initialization parameter β and activation function of the Kriging model are determined. The data is standardized using the z-score standardization method to obtain standardized sampling distances and standardized initial comfort data. The standardized sampling distances and standardized initial comfort data are fitted using a regression model, and the Levenberg-Marquardt method is called to solve the model. This establishes the correlation between the sampling points and surrounding sampling points, enabling the determination of the comfort level of the sampling location of the sampling points, and thus obtaining the comfort information of the indoor space.

[0084] Furthermore, a comfort adjustment strategy for the space is generated from the comfort information and input into the control system. The control system then sends instructions to the intelligent air conditioning system. Upon receiving the instructions, the intelligent air conditioning system adjusts the environment and stores the adjusted data in the database.

[0085] This embodiment determines sampling points in the indoor space and samples environmental index data for each sampling point. The obtained data is then input into a neural network model, and the output of the neural network model is input into a Kriging model. Finally, the comfort information of the indoor space is output, thereby achieving the effect of accurately predicting the comfort of the indoor space.

[0086] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0087] This application also provides an environmental comfort prediction device. It should be noted that the environmental comfort prediction device of this application can be used to execute the environmental comfort prediction method provided in this application. The environmental comfort prediction device provided in this application will be described below.

[0088] Figure 5 This is a schematic diagram of an environmental comfort prediction device provided according to an embodiment of this application, such as... Figure 5 As shown, the device includes: a determining unit 50, a first input unit 51, and a second input unit 52.

[0089] Unit 50 is used to determine M sampling points in the indoor space and collect data of N environmental impact indicators from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers;

[0090] The first input unit 51 is used to input each set of environmental index data from the M sets of environmental index data into the neural network model and output the initial comfort level of the M sampling points.

[0091] The second input unit 52 is used to input the spatial coordinates of M sampling points and the initial comfort level into the Kriging model to obtain the comfort information of the indoor space.

[0092] Optionally, in the environmental comfort prediction device provided in the embodiments of this application, the determining unit 50 includes: a division module, used to divide the indoor space into dimensions, dividing the indoor space of each dimension into non-overlapping intervals to obtain M sampling spaces; and a first selection module, used to randomly select a sampling point from the M sampling spaces according to the Latin hypercube sampling algorithm to obtain M sampling points.

[0093] Optionally, in the environmental comfort prediction device provided in this application embodiment, the first input unit 51 includes: a first input module, used for processing N environmental index data of each sampling point by inputting them into an input layer and processing the processed data into a hidden layer to obtain initial comfort data; and a second input module, used for processing the initial comfort data by inputting them into a summation layer and processing the processed data into an output layer to obtain the initial comfort level of the sampling point.

[0094] Optionally, in the environmental comfort prediction device provided in the embodiments of this application, the first input unit 51 includes: a first determining module, used to determine the first weights of N environmental index data of the sampling point according to the spatial coordinates of the sampling point, and determine a first activation function; and a first calculation module, used to perform weighted calculation on each environmental index data with the matched first weight to obtain N weighted environmental index data, and input the N weighted environmental index data into the first activation function to calculate the initial comfort data of the sampling point.

[0095] Optionally, in the environmental comfort prediction device provided in this application embodiment, the first input unit 51 includes: a second determining module, used to determine the second weight of N initial comfort data of the sampling point according to the spatial coordinates of the sampling point, and calculate the weighted average of each initial comfort data and the comfort data of the corresponding type according to the second weight to obtain N comfort data; and a second selecting module, used to select the comfort data with the highest probability density from the N comfort data, and determine the comfort level associated with the comfort data with the highest probability density as the initial comfort level of the sampling point.

[0096] Optionally, in the environmental comfort prediction device provided in this application embodiment, the second input unit 52 includes: a second calculation module, used to calculate the distance between every two sampling points according to the spatial coordinates of the sampling points, to obtain Z sampling distances, where Z is a positive integer; a third determination module, used to determine the regression model of the Kriging model and the second activation function; a processing module, used to standardize the Z sampling distances to obtain Z standardized sampling distances, and to standardize the comfort values ​​associated with M initial comfort levels to obtain M standardized initial comfort data; and a fitting module, used to fit the Z standardized sampling distances and the M standardized initial comfort data using the regression model, and input the fitted data into the second activation function to calculate comfort information.

[0097] Optionally, in the environmental comfort prediction device provided in the embodiments of this application, the device further includes: an adjustment unit, configured to determine an indoor space comfort adjustment strategy based on the comfort information after obtaining the comfort information of the indoor space, and trigger a start command to execute the comfort adjustment strategy.

[0098] The environmental comfort prediction device provided in this application embodiment uses a determining unit 50 to determine M sampling points in an indoor space and collect data on N environmental impact indicators from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers. A first input unit 51 inputs each set of environmental indicator data into a neural network model and outputs the initial comfort level of the M sampling points. A second input unit 52 inputs the spatial coordinates of the M sampling points and the initial comfort level into a Kriging model to obtain the comfort information of the indoor space. This solves the problem of low accuracy in predicting indoor space comfort in related technologies. By determining the sampling points in the indoor space and sampling environmental indicator data at each sampling point, inputting the obtained data into a neural network model, and then inputting the output of the neural network model into a Kriging model, the device finally outputs the comfort information of the indoor space, thereby achieving the effect of accurately predicting the comfort of the indoor space.

[0099] The environmental comfort prediction device includes a processor and a memory. The aforementioned determination unit 50, first input unit 51, second input unit 52, etc., are all stored in the memory as program units. The processor executes the aforementioned program units stored in the memory to realize the corresponding functions.

[0100] The processor contains a kernel, which retrieves the corresponding program unit from memory. One or more kernels can be configured, and adjusting kernel parameters can address the low accuracy issue in predicting indoor space comfort in related technologies.

[0101] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0102] This invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements a method for predicting environmental comfort.

[0103] This invention provides a processor for running a program, wherein the program executes a method for predicting environmental comfort during runtime.

[0104] Figure 6 This is a schematic diagram of an electronic device provided according to an embodiment of this application, such as... Figure 6 As shown, this embodiment of the invention provides an electronic device 60, which includes a processor, a memory, and a program stored in the memory and executable on the processor. The processor is used to execute computer-readable instructions, wherein the computer-readable instructions execute a method for predicting environmental comfort. The device described herein may be a server, PC, PAD, mobile phone, etc.

[0105] This application also provides a computer program product that, when executed on a data processing device, is suitable for performing a method for predicting environmental comfort.

[0106] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0107] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0108] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0111] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0112] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0113] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0114] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0115] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for predicting environmental comfort, characterized in that, include: M sampling points are determined in the indoor space, and data of N environmental impact indicators are collected from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers; Each of the M sets of environmental index data is input into the neural network model, and the initial comfort level of M sampling points is output. The spatial coordinates of the M sampling points and the initial comfort level are input into the Kriging model to obtain the comfort information of the indoor space. The neural network model includes an input layer, a hidden layer, a summing layer, and an output layer. The process involves inputting the M sets of environmental indicator data into the neural network model and outputting the initial comfort levels for M sampling points. For each sampling point, the process involves inputting N environmental indicator data from the sampling point into the input layer for processing, then inputting the processed data into the hidden layer for further processing to obtain initial comfort data. Finally, the initial comfort data is input into the summing layer for processing, and then input into the output layer for further processing to obtain the initial comfort level for that sampling point. Specifically, for each sampling point, N environmental index data of the sampling point are input into the input layer for processing, and the processed data is input into the hidden layer for processing to obtain initial comfort data. This includes: determining the first weight of the N environmental index data of the sampling point based on the spatial coordinates of the sampling point, and determining a first activation function; performing a weighted calculation on each environmental index data with the matched first weight to obtain N weighted environmental index data, and inputting the N weighted environmental index data into the first activation function to calculate the initial comfort data of the sampling point. The step of determining the first weight of N environmental indicator data of the sampling point based on the spatial coordinates of the sampling point includes: when the spatial coordinates of the sampling point indicate that the sampling point is located at the ceiling in the indoor space, setting the first weight of the metabolic rate of the sampling point to a smaller value and setting the first weight of the relative humidity of the sampling point to a larger value; when the spatial coordinates of the sampling point indicate that the sampling point is located at a service window in the indoor space, setting the first weight of the dry bulb temperature of the sampling point to a larger value, wherein the N environmental impact indicators include the metabolic rate, the relative humidity, and the dry bulb temperature; The step of inputting the spatial coordinates of the M sampling points and the initial comfort level into the Kriging model to obtain the comfort information of the indoor space includes: calculating the distance between every two sampling points based on the spatial coordinates of the sampling points to obtain Z sampling distances, where Z is a positive integer; determining the regression model and the second activation function of the Kriging model; standardizing the Z sampling distances to obtain Z standardized sampling distances, and standardizing the comfort values ​​associated with the M initial comfort levels to obtain M standardized initial comfort data; fitting the Z standardized sampling distances and the M standardized initial comfort data using the regression model, and inputting the fitted data into the second activation function to calculate the comfort information.

2. The method according to claim 1, characterized in that, The M sampling points for determining the indoor space include: The indoor space is divided into dimensions, and each dimension of the indoor space is divided into non-overlapping intervals to obtain M sampling spaces; The M sampling points are obtained by randomly selecting one sampling point from each of the M sampling spaces according to the Latin hypercube sampling algorithm.

3. The method according to claim 1, characterized in that, The initial comfort data is input into the summing layer for processing, and the processed data is input into the output layer for further processing to obtain the initial comfort level of the sampling point, including: The second weight of the N initial comfort data of the sampling point is determined based on the spatial coordinates of the sampling point, and the weighted average of each initial comfort data and the comfort data of the corresponding type is calculated based on the second weight to obtain N comfort data. Select the comfort data with the highest probability density from the N comfort data, and determine the comfort level associated with the comfort data with the highest probability density as the initial comfort level of the sampling point.

4. The method according to claim 1, characterized in that, After obtaining the comfort information of the indoor space, the method further includes: Based on the comfort information, a comfort adjustment strategy for the indoor space is determined, and a start command for executing the comfort adjustment strategy is triggered.

5. A device for predicting environmental comfort, characterized in that, include: The determination unit is used to determine M sampling points in the indoor space and collect data on N environmental impact indicators from each sampling point to obtain M sets of environmental indicator data, where M and N are positive integers; The first input unit is used to input each set of environmental index data from the M sets of environmental index data into the neural network model and output the initial comfort level of M sampling points. The second input unit is used to input the spatial coordinates of the M sampling points and the initial comfort level into the Kriging model to obtain the comfort information of the indoor space. The neural network model includes an input layer, a hidden layer, a summing layer, and an output layer. The first input unit includes: a first input module, used to input N environmental index data of each sampling point into the input layer for processing, and input the processed data into the hidden layer for processing to obtain initial comfort data; and a second input module, used to input the initial comfort data into the summing layer for processing, and input the processed data into the output layer for processing to obtain the initial comfort level of the sampling point. The first input unit further includes: a first determining module, used to determine the first weights of N environmental index data of the sampling point according to the spatial coordinates of the sampling point, and to determine a first activation function; and a first calculation module, used to perform weighted calculation on each environmental index data with the matched first weight to obtain N weighted environmental index data, and input the N weighted environmental index data into the first activation function to calculate the initial comfort data of the sampling point. The first determining module is further configured to: when the spatial coordinates of the sampling point indicate that the sampling point is located at the ceiling in the indoor space, set a first weight of the metabolic rate of the sampling point to a smaller value and set a first weight of the relative humidity of the sampling point to a larger value; when the spatial coordinates of the sampling point indicate that the sampling point is located at a service window in the indoor space, set a first weight of the dry-bulb temperature of the sampling point to a larger value, wherein the N environmental impact indicators include the metabolic rate, the relative humidity, and the dry-bulb temperature; The second input unit includes: a second calculation module for calculating the distance between every two sampling points based on their spatial coordinates, resulting in Z sampling distances, where Z is a positive integer; a third determination module for determining the regression model of the Kriging model and the second activation function; a processing module for standardizing the Z sampling distances to obtain Z standardized sampling distances, and standardizing the comfort values ​​associated with M initial comfort levels to obtain M standardized initial comfort data; and a fitting module for fitting the Z standardized sampling distances and M standardized initial comfort data using the regression model, and inputting the fitted data into the second activation function to calculate comfort information.

6. A processor, characterized in that, The processor is used to run a program, wherein the program executes the environmental comfort prediction method according to any one of claims 1 to 4.

7. An electronic device, characterized in that, Includes one or more processors and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the environmental comfort prediction method according to any one of claims 1 to 4.

Images