Household intelligent environment regulation and control system

Abstract

The invention relates to the technical field, in particular to a home intelligent environment regulation and control system which comprises an acquisition processing unit, a prediction model unit and a result output unit, the acquisition processing unit is composed of a humidity sensor, a temperature sensor and a carbon dioxide concentration sensor and is used for acquiring indoor environment data and performing normalization processing, and the prediction model unit adopts an LSTM network structure. Normalized data is subjected to embedding processing and then input into a network, a dynamic prediction model is established through a forgetting door, an input door and an output door, an Adam optimizer is used for optimization, a result output unit is combined with an indoor humidity value and outdoor weather for comprehensive analysis, an intelligent decision algorithm is introduced, and a prediction result is obtained. According to the system, intelligent and accurate regulation and control of the home environment are achieved, the comfort degree is improved, energy is saved, and a healthy and comfortable living environment is created for a user.

Description

Technical Field

[0001] The present invention relates to the field of smart home technology, and in particular to a smart home environment control system. Background Art

[0002] Smart home is an important technology. As people's requirements for quality of life continue to increase, the comfort and intelligent control of the home environment have become increasingly important. In the field of smart home, accurately predicting and controlling indoor humidity is one of the key links to improve the living experience.

[0003] Traditional humidity prediction methods have many shortcomings when processing time series data. On the one hand, it is difficult for traditional methods to effectively handle long-term dependencies in time series data and cannot accurately capture the complex patterns of indoor humidity changes over time. For example, humidity data for a certain period of time in the past may become unrepresentative due to special events (such as a window suddenly opened), but it is difficult for traditional methods to automatically decide whether to forget these data. On the other hand, traditional methods also have problems in model training and optimization. The commonly used loss function may not accurately measure the difference between the model prediction value and the actual value, resulting in unsatisfactory training results and unable to improve the accuracy and stability of the model to create a healthy and comfortable living environment for users. In order to solve this technical problem, we provide a home intelligent environment control system. Summary of the invention

[0004] The purpose of the present invention is to provide a home intelligent environment control system to solve the problems raised in the above background technology.

[0005] To achieve the above-mentioned purpose, a home intelligent environment control system is provided, comprising a collection and processing unit, a prediction model unit, and a result output unit;

[0006] The acquisition and processing unit includes a humidity sensor, a temperature sensor, and a carbon dioxide concentration sensor, which are used to acquire indoor environmental data and process the indoor environmental data;

[0007] The prediction model unit uses the long short-term memory network algorithm in deep learning to predict indoor humidity, adopts an LSTM network structure with a hidden layer, inputs indoor environmental data, and enters the LSTM network after being processed by the embedding layer. In the network, the long-term dependency in the time series data is processed by the forget gate, input gate and output gate to establish a dynamic prediction model, and the loss function is used to measure the difference between the predicted value and the actual value of the dynamic prediction model. Finally, the dynamic prediction model is optimized by updating the weight and bias of the network through the Adam optimizer to obtain the humidity value;

[0008] The result output unit conducts a comprehensive analysis of the indoor humidity value and the outdoor weather, and introduces an intelligent decision-making algorithm to make a decision.

[0009] As a further improvement of the technical solution, the acquisition and processing unit preprocesses the indoor environment data by normalization, and the specific method is as follows:

[0010] The expression for humidity value processing is: Where x is the original humidity value, x min and x max is the minimum and maximum humidity value, H gy is the normalized humidity value, which is mapped to the interval [0, 1];

[0011] For the temperature value, according to the range of indoor temperature [T min , T max ], the above calculation method is used to process and map the temperature value to the interval [0, 1];

[0012] For the carbon dioxide concentration value, according to its measurement range [C min , C max ], and the above calculation method is used to process and map the normalized carbon dioxide concentration value to the [0, 1] interval.

[0013] As a further improvement of the technical solution, the processing method of the input layer of the prediction model unit LSTM network is as follows:

[0014] The input layer receives the normalized humidity, temperature, and carbon dioxide concentration data and embeds them respectively;

[0015] For humidity, temperature, and carbon dioxide concentration, use a h Dimensions, E t Dimensions and E c The embedding layer of the 3D embedding vector maps the humidity value, temperature value and carbon dioxide concentration value into a high-dimensional space;

[0016] The humidity, temperature and carbon dioxide concentration data after embedding are concatenated at the input layer to form a comprehensive input vector, which is then input into the LSTM network.

[0017] As a further improvement of the technical solution, the LSTM network is a long short-term memory network with N hidden layers, each hidden layer contains M neurons, and the input data is X={x1, x2, ..., x t}, where x t The data vector at time t is processed by the embedding layer and then enters the LSTM network. In the network, the data vector is processed by the forget gate f t, input gate i t And the output gate O t The calculation formula is as follows:

[0018] Forget gate: f t =σ(W f ×[h t-1 , x t ]+b f )

[0019] Input gate: i t =σ(W i ×[h t-1 , x t ]+b i )

[0020] Output gate: O t =σ(W O ×[h t-1 , x t ]+b O )

[0021] Among them, σ is the sigmoid function, W f , W i and W O is the weight matrix, b f 、b i and b O is the bias term, h t-1 For the hidden state of the previous moment, a dynamic prediction model is established by learning the long-term dependencies in time series data through the synergy of the forget gate, input gate and output gate.

[0022] As a further improvement of the technical solution, the method of optimizing the dynamic prediction model by updating the weights and biases of the network through the Adam optimizer specifically includes:

[0023] The expression of the loss function is:

[0024]

[0025] Where α is the loss function, n is the amount of vector data, H i is the true value of humidity at the ith level;

[0026] Let the model parameter be θ, the learning rate be α, and the first-order moment be m t , the second-order moment is v t , the time step is t;

[0027] First, calculate the gradient and update the first moment and the second moment. Calculate the corrected first moment and the second moment. Finally, update the model parameters and continuously adjust the model parameters until the difference range between the humidity value of the model on the validation dataset and the actual humidity value is within ±1°, and then output the humidity value.

[0028] As a further improvement of this technical solution, the way in which the result output unit comprehensively analyzes the indoor humidity value and the outdoor weather specifically includes:

[0029] Let H in be the current indoor humidity value, and ΔH in be the indoor temperature change rate, which is obtained from the ratio of the current indoor humidity value to the predicted indoor humidity value. Let H out be the humidity value in the outdoor weather forecast, P be the precipitation probability, and T out be the outdoor temperature;

[0030] Define the regulation index I, and the mathematical expression of the regulation index I is where ω1, ω2, ω3, ω4 are the weight coefficients of each factor. Set the thresholds T1 and T2. When I < T1, start strong humidification. When T1 < I < T2, start weak humidification. When T2 < I < 0, keep it unchanged. When 0 < I < -T2, start weak dehumidification. When I = T2, start strong dehumidification.

[0031] As a further improvement of this technical solution, the way in which the result output unit introduces an intelligent decision-making algorithm to make decisions specifically includes:

[0032] Obtain the carbon dioxide concentration values of each room from the carbon dioxide concentration sensors in the environmental perception unit. Judge whether there is human activity and ventilation in the room according to the change trend of the carbon dioxide concentration values of each room. When I < T1 and the carbon dioxide concentration is higher than the preset standard carbon dioxide concentration, give priority to starting the ventilation system, and start strong humidification while reducing the carbon dioxide concentration. When 0 < I < -T2 and the carbon dioxide concentration is higher than the preset standard carbon dioxide concentration, start strong dehumidification. Real-time monitor the indoor humidity value through the humidity sensor. When the preset standard humidity value is reached, stop humidification or dehumidification, and real-time feedback the humidity values of each room in the indoor to the central processing module.

[0033] Compared with the prior art, the beneficial effects of the present invention:

[0034] In a home intelligent environment control system, the input layer of the prediction model unit LSTM network embeds humidity values, temperature values ​​and carbon dioxide concentration values ​​respectively, and can map low-dimensional and sparse original data to high-dimensional space, enhance data feature expression, make data have richer expression ability, and be better learned and understood by the model. The LSTM network processes the long-term dependency in time series data through the operation of forget gate, input gate and output gate. The forget gate can determine which information needs to be forgotten according to the historical information of the data, remove irrelevant or outdated information, the input gate determines the storage of new information, so that the model can learn new features to adapt to environmental changes, and the output gate determines the output information to ensure that the prediction result is based on relevant information. This synergy can better capture the long-term dependency and complex change pattern of indoor humidity data, more accurately capture humidity changes, and provide a reliable basis for environmental control decisions. The Adam optimizer is used to update network weights and biases, combining the advantages of adaptive learning rate and momentum. The learning rate can be automatically adjusted to adapt to different parameters and data characteristics, solve the problems of difficult learning rate selection and slow convergence speed of traditional optimization algorithms, and improve training efficiency and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is an overall block diagram of the present invention.

[0036] The meaning of each number in the figure is:

[0037] 1. Acquisition and processing unit; 2. Prediction model unit; 3. Result output unit. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] Example 1

[0040] The present invention provides a home intelligent environment control system, please refer to Figure 1 As shown, it includes a collection and processing unit 1, a prediction model unit 2, and a result output unit 3;

[0041] The acquisition and processing unit 1 includes a humidity sensor, a temperature sensor, and a carbon dioxide concentration sensor, which are used to acquire indoor environmental data and process the indoor environmental data;

[0042] The raw data of humidity, temperature, and carbon dioxide concentration values ​​have different scales and units. If they are not normalized, when these data are input into the deep learning model, some features will dominate due to the huge difference in the numerical range, affecting the training effect and accuracy of the model;

[0043] The acquisition and processing unit 1 uses normalization to preprocess the indoor environment data. The specific method is as follows:

[0044] The expression for humidity value processing is: Where x is the original humidity value, x min and x max is the minimum and maximum humidity value, H gy is the normalized humidity value, which is mapped to the interval [0, 1];

[0045] For the temperature value, according to the range of indoor temperature [T min , T max ], the above calculation method is used to process and map the temperature value to the interval [0, 1];

[0046] For the carbon dioxide concentration value, according to its measurement range [C min , C max ], the above calculation method is used for processing, and the normalized carbon dioxide concentration value is mapped to the [0, 1] interval. Through normalization, the model can better adapt to different data sets and environments, and improve the generalization ability of the model. No matter how the specific values ​​of the input data change, the role of the normalized features in the model is relatively stable, thereby improving the accuracy of indoor humidity prediction.

[0047] Prediction model unit 2 uses the long short-term memory network algorithm in deep learning to predict indoor humidity. It uses an LSTM network structure with a hidden layer, inputs indoor environmental data, and enters the LSTM network after being processed by the embedding layer. In the network, the long-term dependency in the time series data is processed through the forget gate, input gate, and output gate to establish a dynamic prediction model, and the loss function is used to measure the difference between the predicted value and the actual value of the dynamic prediction model. Finally, the Adam optimizer is used to update the weights and biases of the network to optimize the dynamic prediction model and obtain the humidity value;

[0048] Raw data such as humidity, temperature, and carbon dioxide concentration values ​​are low-dimensional and sparse. Directly inputting them into a deep learning model cannot fully express the characteristics and relationships of the data. Through embedding processing, these low-dimensional data can be mapped to a high-dimensional space, making the data more expressive and better learned and understood by the model.

[0049] The input layer of the prediction model unit 2 LSTM network is processed as follows:

[0050] The input layer receives the normalized humidity, temperature, and carbon dioxide concentration data and embeds them respectively. The embedding layer can learn the potential features of the data, convert the original data into a more expressive high-dimensional vector, and improve the model's ability to understand the data.

[0051] For humidity, temperature, and carbon dioxide concentration, use a h Dimensions, E t Dimensions and E c The embedding layer of the 3D embedding vector maps the humidity value, temperature value and carbon dioxide concentration value into a high-dimensional space;

[0052] The humidity, temperature and carbon dioxide concentration data after embedding processing are spliced ​​at the input layer to form a comprehensive input vector, which is input into the LSTM network. Through embedding processing and splicing input vectors, the model can better learn the complex relationship between multiple factors such as humidity, temperature and carbon dioxide concentration and indoor humidity, thereby improving the accuracy of indoor humidity prediction.

[0053] Indoor humidity data is a typical time series data, and its changes are affected by many factors in the long term. Traditional prediction methods are difficult to effectively handle the long-term dependencies in time series data. The LSTM network structure is specially designed to process time series data. Its forget gate, input gate, and output gate can dynamically determine which information needs to be retained, updated, and output based on the historical information of the data, thereby better capturing the long-term dependencies and complex change patterns in indoor humidity data.

[0054] Compared with other deep learning models, LSTM has higher accuracy and stability in processing time series data and can better adapt to different indoor environments and seasonal changes;

[0055] The LSTM network is a long short-term memory network with N hidden layers. Each hidden layer contains M neurons. The input data is X = {x1, x2, ..., x t}, where x t The data vector at time t is processed by the embedding layer and then enters the LSTM network. In the network, the data vector is processed by the forget gate f t , input gate i t And the output gate O t The calculation formula is as follows:

[0056] Forget gate: f t =σ(W f ×[h t-1 , x t ]+b f )

[0057] This gate can be used to determine which information needs to be forgotten based on historical data information, which helps the model remove irrelevant or outdated information and improve the efficiency and accuracy of the model;

[0058] For example, in indoor humidity prediction, humidity data from a certain period of time in the past may become unrepresentative due to special events (such as a window suddenly opening). The forget gate can automatically decide whether to forget this data based on the current input data and the hidden state of the previous moment.

[0059] Input gate: i t =σ(W i ×[h t-1 , x t ]+b i )

[0060] The input gate determines what new information needs to be stored, which enables the model to continuously learn new data features and adapt to changes in the indoor environment;

[0061] For example, when the indoor temperature suddenly rises, the input gate can decide whether to store this new temperature information into the model based on the current input data and the hidden state at the previous moment, so as to better predict the change in humidity;

[0062] Output gate: O t =σ(W O ×[h t-1 , x t ]+b O )

[0063] The output gate determines which information needs to be output, which ensures that the prediction results output by the model are based on the most relevant information at the moment;

[0064] For example, when predicting indoor humidity, the output gate can select appropriate information for output based on the current input data and the hidden state at the previous moment, thereby improving the accuracy of the prediction;

[0065] Through the synergistic effect of the forget gate, input gate and output gate, the long-term dependencies in time series data can be effectively processed and the accuracy of indoor humidity prediction can be improved. Compared with traditional prediction methods, it can more accurately capture sudden changes and seasonal changes in humidity, providing a more reliable basis for subsequent environmental control decisions.

[0066] Among them, σ is the sigmoid function, W f , W i and W O is the weight matrix, b f , b i and b O is the bias term, ht-1 For the hidden state of the previous moment, a dynamic prediction model is established by learning the long-term dependencies in time series data through the synergy of the forget gate, input gate and output gate.

[0067] Traditional optimization algorithms such as stochastic gradient descent may encounter problems such as difficulty in selecting learning rates and slow convergence when training deep learning models. The Adam optimizer combines the advantages of adaptive learning rate and momentum, and can automatically adjust the learning rate to adapt to different parameters and data characteristics, thereby improving training efficiency and stability.

[0068] The method of optimizing the dynamic prediction model by updating the weights and biases of the network through the Adam optimizer includes:

[0069] The expression of the loss function is:

[0070]

[0071] Where α is the loss function, n is the amount of vector data, H i is the true value of humidity at the ith level;

[0072] Let the model parameter be θ, the learning rate be α, and the first-order moment be m t , the second-order moment is v t , the time step is t;

[0073] First, calculate the gradient and update the first-order moment and second-order moment, calculate the corrected first-order moment and second-order moment, and finally update the model parameters;

[0074] The expression for updating the first-order moment is: β1 is a hyperparameter that controls the exponential decay rate of the first-order moment. This allows the model to take historical gradient information into account when updating parameters to avoid falling into a local optimal solution.

[0075] The expression for updating the second-order moment is: Among them, β2 is a hyperparameter that controls the exponential decay rate of the second-order moment. The second-order moment can help the model adaptively adjust the learning rate and improve training efficiency;

[0076] Automatically adjust the learning rate according to the first-order and second-order moments of the gradient, so that the model can converge quickly in different parameter spaces and data distributions;

[0077] Compute the corrected first and second moments: To eliminate the influence of the initial value;

[0078] Update model parameters: where ε is a constant to prevent the denominator from being zero, and continuously adjust the model parameters until the difference range between the humidity value of the model on the validation dataset and the actual humidity value is within ±1, and then output the humidity value.

[0079] The result output unit 3 comprehensively analyzes the indoor humidity value and the outdoor weather, and introduces an intelligent decision-making algorithm to make decisions.

[0080] The way that the result output unit 3 comprehensively analyzes the indoor humidity value and the outdoor weather specifically includes:

[0081] Let H in be the current indoor humidity value, ΔH in be the indoor temperature change rate, which is obtained by the ratio of the current indoor humidity value and the predicted indoor humidity value, H out be the humidity value in the outdoor weather forecast, P be the precipitation probability, T out be the outdoor temperature;

[0082] Define the regulation index I, and the mathematical expression of the regulation index I is where ω1, ω2, ω3, ω4 are the weight coefficients of each factor, set the thresholds T1 and T2. When I < T1, start strong humidification; when T1 < I < T2, start weak humidification; when T2 < I < 0, keep unchanged; when 0 < I < -T2, start weak dehumidification; when I = T2, start strong dehumidification.

[0083] The way that the result output unit 3 introduces an intelligent decision-making algorithm to make decisions specifically includes:

[0084] Obtain the carbon dioxide concentration values of each room from the carbon dioxide concentration sensor in the environment perception unit 1, and judge whether there is human activity and ventilation in the room according to the change trend of the carbon dioxide concentration values of each room. When I < T1 and the carbon dioxide concentration is higher than the preset standard carbon dioxide concentration, give priority to starting the ventilation system, and start strong humidification while reducing the carbon dioxide concentration. When 0 < I < -T2 and the carbon dioxide concentration is higher than the preset standard carbon dioxide concentration, start strong dehumidification. Real-time monitor the indoor humidity value through the humidity sensor. When the preset standard humidity value is reached, stop humidification or dehumidification, and real-time feedback the humidity values of each room in the indoor to the central processing module.

[0085] In the present invention, the collection and processing unit is composed of humidity, temperature and carbon dioxide concentration sensors, which obtain indoor environmental data and perform normalization processing. The prediction model unit adopts an LSTM network structure, embeds the normalized data and inputs it into the network, establishes a dynamic prediction model through a forget gate, an input gate and an output gate, and optimizes it using an Adam optimizer. The result output unit performs a comprehensive analysis in combination with the indoor humidity value and the outdoor weather, introduces an intelligent decision-making algorithm, and judges the room situation according to the carbon dioxide concentration value for corresponding regulation. The system realizes intelligent and precise regulation of the home environment, improves comfort, saves energy, and creates a healthy and comfortable living environment for users.

[0086] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A home intelligent environment control system, characterized in that: It includes a data acquisition and processing unit (1), a prediction model unit (2), and a result output unit (3); The data acquisition and processing unit (1) includes a humidity sensor, a temperature sensor, and a carbon dioxide concentration sensor, which are used to obtain indoor environmental data and process the indoor environmental data; The prediction model unit (2) uses the long short-term memory network algorithm in deep learning to predict indoor humidity. It adopts an LSTM network structure with [number of hidden layers] hidden layers. The indoor environmental data is input, and after being processed by the embedding layer, it enters the LSTM network. In the network, the long-term dependencies in the time series data are processed through the forget gate, input gate, and output gate to establish a dynamic prediction model. The loss function is used to measure the difference between the predicted value and the actual value of the dynamic prediction model. Finally, the weights and biases of the network are updated through the Adam optimizer to optimize the dynamic prediction model, and the humidity value is obtained; The result output unit (3) comprehensively analyzes the indoor humidity value and the outdoor weather, and introduces an intelligent decision-making algorithm to make decisions.

2. The home intelligent environment control system according to claim 1, characterized in that: The data acquisition and processing unit (1) preprocesses the indoor environmental data using normalization. The specific method is as follows: The expression for humidity value processing is: Where x is the original humidity value, x min and x max is the minimum and maximum humidity value, H gy is the normalized humidity value, which is mapped to the interval [0, 1]; For the temperature value, according to the range of indoor temperature [T min , T max ], the above calculation method is used to process and map the temperature value to the interval [0, 1]; For the carbon dioxide concentration value, according to its measurement range [C min , C max ], and the above calculation method is used to process and map the normalized carbon dioxide concentration value to the [0, 1] interval.

3. The home intelligent environment control system according to claim 2, characterized in that: The processing method of the input layer of the LSTM network in the prediction model unit (2) is as follows: The input layer receives the humidity value, temperature value, and carbon dioxide concentration value data after normalization processing, and performs embedding processing on them respectively; For humidity, temperature, and carbon dioxide concentration, use a h Dimensions, E t Dimensions and E c The embedding layer of the 3D embedding vector maps the humidity value, temperature value and carbon dioxide concentration value into a high-dimensional space; The humidity, temperature, and carbon dioxide concentration data after embedding processing are concatenated in the input layer to form a comprehensive input vector, which is input into the LSTM network.

4. The home intelligent environment control system according to claim 3, characterized in that: The LSTM network is a long short-term memory network with N hidden layers, each hidden layer contains M neurons, and the input data is X={x1, x2, ..., x t }, where x t The data vector at time t is processed by the embedding layer and then enters the LSTM network. In the network, the data vector is processed by the forget gate f t , input gate i t And the output gate O t The calculation formula is as follows: Forget gate: f t =σ(W f ×[h t-1 , x t ]+b f ) Input gate: i t =σ(W i ×[h t-1 , x t ]+b i ) Output gate: O t =σ(W O ×[h t-1 , x t ]+b O ) Among them, σ is the sigmoid function, W f , W i and W O is the weight matrix, b f , b i and b O is the bias term, h t-1 For the hidden state of the previous moment, a dynamic prediction model is established by learning the long-term dependencies in time series data through the synergy of the forget gate, input gate and output gate.

5. The home intelligent environment control system according to claim 1, characterized in that: The method of optimizing the dynamic prediction model by updating the weights and biases of the network through the Adam optimizer specifically includes: The expression of the loss function is: Where α is the loss function, n is the amount of vector data, H i is the true value of humidity at the ith level; Let the model parameter be θ, the learning rate be α, and the first-order moment be m t , the second-order moment is v t , the time step is t; First, calculate the gradient and update the first moment and second moment, calculate the corrected first moment and second moment, and finally update the model parameters. Continuously adjust the model parameters until the difference range between the humidity value of the model on the validation dataset and the actual humidity value is within ±1°, and then output the humidity value.

6. The home intelligent environment control system according to claim 5, characterized in that: The method of the result output unit (3) comprehensively analyzing the indoor humidity value and the outdoor weather specifically includes: Let H in is the current indoor humidity value, ΔH in is the indoor temperature change rate, which is obtained by the ratio of the current indoor humidity value to the predicted indoor humidity value. out is the humidity value in the outdoor weather forecast, P is the probability of precipitation, T out is the outdoor temperature; Define the regulation index I, and the mathematical expression of the regulation index I is where ω1, ω2, ω3, ω4 are the weight coefficients of each factor, set the thresholds T1 and T2. When I < T1, start strong humidification; when T1 < I < T2, start weak humidification; when T2 < I < 0, keep unchanged; when 0 < I < -T2, start weak dehumidification; when I = T2, start strong dehumidification.

7. The home intelligent environment control system according to claim 1, characterized in that: The method of the result output unit (3) introducing an intelligent decision-making algorithm to make decisions specifically includes: Obtain the carbon dioxide concentration values of each room from the carbon dioxide concentration sensor in the environmental perception unit (1). According to the change trend of the carbon dioxide concentration values of each room, judge whether there is human activity and ventilation in the room. When I < T1 and the carbon dioxide concentration is higher than the preset standard carbon dioxide concentration, give priority to starting the ventilation system, and start strong humidification while reducing the carbon dioxide concentration. When 0 < I < -T2 and the carbon dioxide concentration is higher than the preset standard carbon dioxide concentration, start strong dehumidification. Real-time monitor the indoor humidity value through the humidity sensor. When the preset standard humidity value is reached, stop humidification or dehumidification, and feedback the humidity values of each room in the indoor to the central processing module in real time.

Images