Construction and application method for whole house control strategy output model and related device

Abstract

The invention relates to the technical field of whole house control, and discloses a construction method for a whole house control strategy output model, and the method comprises the steps: building a whole house control strategy initial model based on the whole house environment data and the equipment control data of whole house smart home equipment; based on a reinforcement learning algorithm, alternately training the whole house control strategy initial model by taking the environment comfort level and the equipment energy consumption value as rewards; and taking the whole house control strategy initial model with the reward signal maximization as a whole house control strategy output model. According to the method, the requirement of the user for comfort or energy consumption or the requirement for balance of the comfort and the energy consumption in the whole-house air system can be met, the comfort of the user in the whole-house air system can be improved, meanwhile, the energy consumption of the system can be reduced, and the intelligence of the system can be improved. The invention also discloses an application method for the whole house control strategy output model, a construction device for the whole house control strategy output model, and an electronic device.

Description

Technical Field

[0001] This application relates to the field of whole-house control technology, such as a method for constructing whole-house control strategy output, an application method and a construction device, and an electronic device. Background Technology

[0002] With the continuous advancement of technology, smart homes are becoming an increasingly important part of people's lives. Whole-house air systems, as a key element of smart homes, can provide users with a comfortable indoor environment.

[0003] The related technology discloses an air conditioning temperature control method based on deep reinforcement learning, including: establishing an air conditioning energy consumption optimization model, designing a reinforcement learning framework for air conditioning energy consumption, training a Q-value neural network based on the TD3 algorithm based on the state-action value function, historical temperature and humidity, wind speed sensor and outdoor temperature data; and deploying the trained Q-value neural network in the air conditioning control system to adjust the air conditioning temperature in real time.

[0004] In the process of implementing the embodiments of this disclosure, at least the following problems were found in the related art:

[0005] While the relevant technologies take into account the energy consumption cost of air conditioning throughout the day and can control the temperature more accurately and save energy while ensuring human comfort, they do not take into account the energy consumption of the entire whole-house air system or the user's comfort within the entire whole-house air system, thus reducing the system's intelligence.

[0006] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or describe the scope of protection of these embodiments, but rather as a prelude to the detailed description that follows.

[0008] This disclosure provides a method, application method, construction device, and electronic device for constructing whole-house control strategy output, which can improve user comfort in the whole-house air system while reducing system energy consumption and enhancing system intelligence.

[0009] In some embodiments, the method for constructing the whole-house control strategy output model includes: establishing an initial model of the whole-house control strategy based on whole-house environmental data and device control data of whole-house smart home devices; training the initial model of the whole-house control strategy alternately using environmental comfort and device energy consumption values ​​as rewards, respectively, based on a reinforcement learning algorithm; and using the initial model of the whole-house control strategy that maximizes the reward signal as the output model of the whole-house control strategy.

[0010] Optionally, based on whole-house environmental data and device control data of smart home devices, an initial model of whole-house control strategy is established, including: extracting time features and environmental features from the whole-house environmental data, and extracting state features and functional features from the device control data; integrating the time features, environmental features, state features and functional features to construct a feature dataset; and using a deep learning model to capture the feature relationships of the feature dataset to establish the initial model of whole-house control strategy.

[0011] Optionally, the initial model of the whole-house control strategy is trained alternately with environmental comfort and equipment energy consumption as rewards, including: training the initial model of the whole-house control strategy with environmental comfort as the reward for the first training; training the initial model of the whole-house control strategy with equipment energy consumption as the reward for the second training; wherein the first training and the second training are performed alternately.

[0012] Optionally, environmental comfort is determined as follows: the first whole-house control test strategy output by the initial model of the whole-house control strategy after the previous training is executed; after executing the first whole-house control test strategy, comfort environment data affecting human comfort is obtained; and the environmental comfort of the current environment is calculated based on the comfort environment data.

[0013] Optionally, the environmental comfort level of the current environment is calculated based on the comfort environment data, including: calculating the thermal balance parameters of the human body based on the comfort environment data; wherein the comfort environment data includes multiple parameters such as air temperature, air velocity, relative humidity, mean radiant temperature, human metabolic rate, and clothing thermal resistance; determining the human thermal comfort evaluation index based on the thermal balance parameters; using the human thermal comfort evaluation index as the environmental comfort level of the current environment; or, determining the predicted percentage of dissatisfaction, representing the user's degree of dissatisfaction with the current environment, based on the human thermal comfort evaluation index and the predicted percentage of dissatisfaction, and determining the environmental comfort level of the current environment based on the human thermal comfort evaluation index and the predicted percentage of dissatisfaction.

[0014] Optionally, the device energy consumption value is used as a reward to perform a second training on the initial model of the whole-house control strategy, including: determining the target whole-house control test strategy corresponding to the device energy consumption value; wherein the target whole-house control test strategy includes a second whole-house control test strategy or a third whole-house control test strategy; using the energy consumption value as a reward, adjusting the decision parameters and control strategy of the initial model of the whole-house control strategy according to the target whole-house control test strategy, so that the reward signal reaches the minimum energy consumption.

[0015] In some embodiments, the application method for the whole-house control strategy output model includes: obtaining user instructions; inputting the user instructions into the whole-house control strategy output model to obtain a whole-house control strategy; and controlling all smart home devices according to the whole-house control strategy.

[0016] In some embodiments, the apparatus for constructing a whole-house control strategy output model includes a processor and a memory. The memory stores program instructions, and the processor is configured to, when executing the program instructions, perform either the method for constructing the whole-house control strategy output model described above, or the method for applying the whole-house control strategy output model described above.

[0017] In some embodiments, the electronic device includes an electronic device body and the aforementioned construction device for the whole-house control strategy output model. The construction device for the whole-house control strategy output model is mounted on the electronic device body.

[0018] The method, application method, construction apparatus, and electronic device for constructing a whole-house control strategy output model provided in this disclosure can achieve the following technical effects:

[0019] Compared to related technologies, the embodiments disclosed herein can meet users' needs for comfort or energy consumption in whole-house air systems, or the need for a balance between comfort and energy consumption. They can improve users' comfort in the whole-house air system while reducing the system's energy consumption, and can also enhance the system's intelligence.

[0020] The above general description and the description below are exemplary and illustrative only and are not intended to limit this application. Attached Figure Description

[0021] One or more embodiments are illustrated by way of example with reference to the accompanying drawings. These illustrations and drawings do not constitute a limitation on the embodiments. Elements having the same reference numerals in the drawings are shown as similar elements. The drawings are not to be scaled. And wherein:

[0022] Figure 1 This is a schematic diagram of a system environment provided in an embodiment of this disclosure;

[0023] Figure 2This is a schematic diagram of a method for constructing a whole-house control strategy output model provided in an embodiment of this disclosure;

[0024] Figure 3 This is a schematic diagram of the HVAC system principle of a method for constructing a whole-house control strategy output model provided in an embodiment of this disclosure;

[0025] Figure 4 This is a schematic diagram illustrating the implementation principle of a reinforcement learning algorithm in a method for constructing a whole-house control strategy output model provided in this embodiment of the disclosure;

[0026] Figure 5 This is a schematic block diagram of a method for constructing a whole-house control strategy output model provided in an embodiment of this disclosure;

[0027] Figure 6 This is a schematic diagram of an application method for a whole-house control strategy output model provided in an embodiment of this disclosure;

[0028] Figure 7 This is a schematic diagram of a device for constructing a whole-house control strategy output model provided in an embodiment of this disclosure. Detailed Implementation

[0029] To provide a more detailed understanding of the features and technical content of the embodiments of this disclosure, the implementation of the embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. The accompanying drawings are for illustrative purposes only and are not intended to limit the embodiments of this disclosure. In the following technical description, for ease of explanation, several details are used to provide a full understanding of the disclosed embodiments. However, one or more embodiments may still be implemented without these details. In other cases, well-known structures and devices may be simplified in their depiction to simplify the drawings.

[0030] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure 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 disclosure described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.

[0031] Unless otherwise stated, the term "multiple" means two or more.

[0032] In this embodiment of the disclosure, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.

[0033] The term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.

[0034] The term "correspondence" can refer to an association or binding relationship. The correspondence between A and B means that there is an association or binding relationship between A and B.

[0035] In this embodiment of the disclosure, smart home devices refer to home appliances formed by introducing microprocessors, sensor technology, and network communication technology into home appliances. They have the characteristics of intelligent control, intelligent sensing, and intelligent application. The operation of smart home devices often relies on the application and processing of modern technologies such as the Internet of Things, the Internet, and electronic chips. For example, smart home devices can be connected to electronic devices to enable users to remotely control and manage smart home devices.

[0036] Combination Figure 1 As shown, the embodiments of this disclosure can be applied to, for example... Figure 1 The five-constant system shown is an integrated smart home system that uses intelligent control technology to maintain five constant states in the indoor environment: constant temperature, constant humidity, constant oxygen, constant cleanliness, and constant air quality. This provides users with a healthy, comfortable, and energy-efficient living environment. A five-constant system typically includes a heating / cooling source, an air handling unit, and an integrated control system that can adjust and precisely control temperature and humidity in response to changes in the indoor and outdoor environment. The method for constructing and applying the whole-house control strategy output model provided in this disclosure can be deployed in a smart control box, enabling edge control via a local area network or wired connection even when the user's home lacks a network connection. Alternatively, it can be deployed in the cloud. Since the cloud provides more computing power, cloud deployment improves efficiency and reduces operating costs while still meeting user renewal requirements. The edge control unit can be deployed in a smart control box, while the whole-house control strategy output model can be deployed in the cloud. In this way, the edge smart control box can respond to and handle urgent or time-sensitive tasks quickly in real time, reducing dependence on network latency. At the same time, if the cloud server has problems, the edge smart control box can continue to operate independently, improving the stability of the entire system.

[0037] It should be noted that the implementation environment of this disclosure embodiment includes, but is not limited to, a five-constant system, and this disclosure embodiment can be applied to various smart home systems. In these smart home systems, various smart home devices (such as floor-standing air conditioners, wall-mounted air conditioners, home air conditioners, ducted air conditioners, and fresh air systems), cloud servers, sensors, and control terminals (such as apps, wired controllers, and smart screens) are connected and interconnected through network communication technologies. These network communication technologies include, but are not limited to, Wi-Fi (Wireless Fidelity), Bluetooth, Zigbee, and PLC (Programmable Logic Controller).

[0038] Combination Figure 2 As shown in the embodiments of this disclosure, a method for constructing a whole-house control strategy output model is provided, including:

[0039] S201, the electronic device establishes an initial model of the whole-house control strategy based on whole-house environmental data and device control data of whole-house smart home devices.

[0040] S202, the electronic device uses a reinforcement learning algorithm to alternately train the initial model of the whole-house control strategy by using environmental comfort and device energy consumption as rewards respectively.

[0041] S203, the electronic device uses the initial model of the whole-house control strategy that maximizes the reward signal as the output model of the whole-house control strategy.

[0042] The method for constructing a whole-house control strategy output model provided in this disclosure can meet the user's comfort needs, energy consumption needs, or needs for a balance between comfort and energy consumption in the whole-house air system. It can improve the user's comfort in the whole-house air system while reducing the system's energy consumption, and it can also improve the system's intelligence.

[0043] For example, whole-house environmental data includes multiple parameters such as indoor ambient temperature, outdoor ambient temperature, indoor ambient humidity, outdoor ambient humidity, indoor air purification level, indoor oxygen concentration, and indoor space area. Equipment control data includes multiple parameters such as equipment operation commands, equipment status, equipment power, and equipment start-up and shutdown times.

[0044] For example, reinforcement learning algorithms may include deep Q-networks, deep deterministic policy gradients, or dual deep deterministic policy gradients. In practical applications, deep Q-networks can be used as approximate functions for reinforcement learning to learn the mapping relationship between states and actions. The system learns the optimal control policy through interaction with the environment, thereby maximizing the reward signal.

[0045] Optionally, based on whole-house environmental data and device control data of smart home devices, an initial model for whole-house control strategy is established. This includes: extracting temporal and environmental features from the whole-house environmental data, and extracting state and functional features from the device control data; integrating the temporal, environmental, state, and functional features to construct a feature dataset; and using a deep learning model to capture the feature relationships within the feature dataset to establish the initial model for whole-house control strategy. In this way, through data integration, the initial model for whole-house control strategy can be established more effectively, and the established model better understands users' lifestyles and preferences, thereby improving the user experience.

[0046] For example, data fusion methods can be used to integrate temporal features, environmental features, state features, and functional features. Data fusion methods include, but are not limited to, concatenation, addition, multiplication, attention mechanisms, pyramid pooling, deconvolution, FPN (Feature Pyramid Network), and ASPP (Atrous Spatial Pyramid Pooling). In practical applications, various data fusion methods can be used to integrate temporal features, environmental features, state features, and functional features, thereby providing richer feature representations.

[0047] For example, a deep learning model takes features from a feature dataset as input and learns the relationships between those features. Deep learning models can include convolutional neural networks, recurrent neural networks, and long short-term memory networks, with the choice made based on problem complexity and data characteristics in practical applications.

[0048] Optionally, the initial model of the whole-house control strategy can be trained alternately with environmental comfort and device energy consumption as rewards. This includes: first training the initial model with environmental comfort as the reward; and second training the initial model with device energy consumption as the reward; wherein the first and second training are performed alternately. In this way, the trained initial model of the whole-house control strategy can learn to find the optimal balance between improving comfort and reducing energy consumption, and can consider multiple objectives simultaneously, thus providing users with a more comprehensive control strategy. Furthermore, it can improve the robustness of the initial model of the whole-house control strategy under different environmental and device conditions.

[0049] Optionally, environmental comfort can be determined as follows: Execute the first whole-house control test strategy output by the initial model of the whole-house control strategy after the previous training; after executing the first whole-house control test strategy, acquire comfort environment data that affects human comfort; calculate the environmental comfort level of the current environment based on the comfort environment data. This allows for a quantitative assessment of environmental comfort, providing data for improving user experience. Furthermore, it facilitates adjustments to the initial model of the whole-house control strategy based on environmental comfort, further improving model performance.

[0050] Optionally, the environmental comfort level of the current environment can be calculated based on comfort environment data, including: calculating human thermal balance parameters based on the comfort environment data; wherein the comfort environment data includes multiple parameters such as air temperature, air velocity, relative humidity, mean radiant temperature, human metabolic rate, and clothing thermal resistance; determining human thermal comfort evaluation indicators based on the thermal balance parameters; and using the human thermal comfort evaluation indicators as the environmental comfort level of the current environment; or, based on the human thermal comfort evaluation indicators, determining the predicted percentage of dissatisfaction representing the user's degree of dissatisfaction with the current environment, and determining the environmental comfort level of the current environment based on the human thermal comfort evaluation indicators and the predicted percentage of dissatisfaction. In this way, considering multiple environmental data helps to more comprehensively assess the impact of the environment on the human body, thereby determining the environmental comfort level of the current environment more comprehensively and accurately. This helps to adjust the initial model of the whole-house strategy using environmental data, and helps to obtain a more accurate whole-house control strategy. Furthermore, determining the human thermal comfort evaluation indicators based on thermal balance parameters allows for a more objective and quantitative evaluation of environmental comfort.

[0051] For example, the human thermal comfort index PWV (Predicted Mean Vote) is calculated according to the following formula:

[0052] PMV = (0.303e -0.036M +A1){(MW)-3.05×10 -3 ×[5733-6.99(MW)-P a -0.42×[(MW)-58.15]-1.7×10 -5 M(5867-P a -0.0014M(34-t) a -3.96×10 -8 f c1 ×[(t c1 +273) 4 -(t r +273) 4 ]-f c1 h c (t c1 -t d)}+A2;

[0053] Where A1 is the first parameter and A2 is the second parameter, and A1 and A2 are determined based on the population group;

[0054] M represents the human metabolic rate, corresponding to the state of human activity, and is measured in W / m³. 2 ;

[0055] W represents the external mechanical work, which is usually taken as 0, and its unit is W / m. 2 ;

[0056] P a P is the partial pressure of water vapor in ambient air, expressed in kPa. a From air temperature t a The results were obtained from calculations based on relative humidity (RH).

[0057]

[0058] t a RH represents air temperature, in °C; RH represents relative humidity, in %;

[0059] f c1 For the surface area of ​​clothing, in I c1 Calculate f when ≤0.078. c1 =1.0+1.290I c1 ; in I c1 When f > 0.078, calculate f c1 =1.05+0.645I c1 ;I c1 Thermal resistance of clothing, measured in meters (m). 2 k / w corresponds to the human body's clothing pattern;

[0060] t c1 The surface temperature of the garment is expressed in °C.

[0061] t c1 =35.7-0.028(MW)-Ic1·[3.96×10 -8 f c1 ×{(t c1 +273) 4 -(t r +273) 4}+f c1 h c (t c1 -t a )]:

[0062] t r The average radiation temperature is expressed in °C.

[0063] hc This is a convective heat transfer system, with units of W / m³. 2 ×k;

[0064]

[0065] v represents the air velocity, measured in m / s.

[0066] More specifically, the corresponding PMV equation is automatically run based on the judgment of the population, and A1 and A2 are determined according to Table 1 below.

[0067] Table 1

[0068] category A1 A2 man +0.016 -0.050 woman +0.033 -0.227 elder +0.028 -0.107 child +0.043 -0.132

[0069] In practical applications, user categories can be determined using sensors such as cameras and voiceprints, or through manual input by the user.

[0070] More specifically, look up Table 2 to determine the value of M, and look up Table 3 to determine the value of I. c1 value.

[0071] Table 2

[0072]

[0073] Table 3

[0074]

[0075] In practical applications, the user's current activity status and home attire can be identified, and then the M and I values ​​can be determined by looking up a table. c1 The value is then substituted into the PMV formula to calculate the human thermal comfort index.

[0076] For example, PPD (Predicted Percentage of Dissatisfied) is calculated using the following formula:

[0077] PPD = 100 - 95exp[(0.03353PMV) 4 +0.21PMV 2 )];

[0078] The PMV and PPD values ​​exhibit a parabolic relationship according to the formula, with the lowest point of the parabola representing optimal comfort. Each user possesses a parabola characterizing their own features. Big data technology can be used to estimate the appropriate comfort level for users in the same location based on their parabolic range. Then, based on the user's comfort level estimated by the system and big data, the user's setting logic can be derived. This helps the whole-house control strategy output model to output whole-house control strategies more accurately.

[0079] More specifically, when the PDD value is in the range of 0 to 0.5, the current environment offers optimal comfort. When the PDD value is in the range of 0.5 to 1, the current environment offers a balanced level of comfort.

[0080] Optionally, the device energy consumption value can be determined as follows: obtain the second whole-house control test strategy output by the initial model of the whole-house control strategy after the previous training; determine the third whole-house control test strategy selected by the user based on the second whole-house control test strategy; obtain the second test energy consumption value for executing the second whole-house control test strategy and the third test energy consumption value for executing the third whole-house control test strategy; and take the minimum test energy consumption value between the second and third test energy consumption values ​​as the device energy consumption value. This allows the whole-house control strategy output model to output the whole-house control strategy with the lowest energy consumption according to user needs, providing strong data support for the whole-house control strategy, thereby making the whole-house control system more intelligent and more energy-efficient.

[0081] In practical applications, it can be applied to, for example Figure 3 The HVAC (Heating, Ventilation, Air Conditioning and Cooling) system shown. For example... Figure 3 As shown, the initial model of the whole-house control strategy predicts the energy consumption for the next time step (e.g., half an hour or one hour) and outputs a second whole-house control test strategy. The second whole-house control test strategy includes energy consumption optimization strategies or energy-reducing logic for air-related smart home devices, such as automatically opening blinds to reduce light intake; or turning off the air conditioner and turning on the fresh air system when the outdoor temperature is suitable. This can more effectively reduce the energy consumption of the entire house. Furthermore, the system has an optimization function that optimizes the initial model of the whole-house control strategy and outputs the parameter settings for energy optimization and implementation parameters. These parameter settings are then passed to the HVAC system so that the system can adjust its settings accordingly. After optimizing the initial model of the whole-house control strategy, the output model of the whole-house control strategy is obtained. This output model can output optimized strategies to control the system, forming an energy-saving control strategy that is then passed to the HVAC system. In addition, the HVAC system can predict future performance, thereby providing users with a more energy-efficient user experience.

[0082] More specifically, the whole-house control strategy output model predicts energy consumption for the next time step and prompts the user for confirmation. If the user chooses to execute, the model provides parameters and control logic. The user then executes the commands according to these parameters and logic, achieving energy-saving control of the entire system. The whole-house control strategy output model continuously trains its decision parameters and whole-house control strategy based on the predicted energy consumption and the user's execution results, ultimately achieving automatic decision-making and minimizing energy consumption.

[0083] Optionally, based on a reinforcement learning algorithm, the energy consumption value is used as a reward to perform a second training on the initial model of the whole-house control strategy. This includes: determining the target whole-house control test strategy corresponding to the device energy consumption value; wherein the target whole-house control test strategy includes a second whole-house control test strategy or a third whole-house control test strategy; using the energy consumption value as a reward, adjusting the decision parameters and control strategy of the initial model of the whole-house control strategy according to the target whole-house control test strategy, so that the reward signal reaches the minimum energy consumption. In this way, the energy consumption of the whole house can be reduced more effectively, thereby reducing energy consumption and significantly reducing costs for long-term operating whole-house systems.

[0084] For example, such as Figure 4 As shown, a reinforcement learning agent (DRL agent) can be constructed. During online applications, this agent can interact with device or vehicle models, learn their states, and perform simulation training based on state feedback from both the device / vehicle model and the control system. The agent then corrects the model parameters based on the simulation training results, and the control system can observe these parameter corrections. Finally, through multiple simulation training sessions, the reinforcement learning agent can better meet user needs.

[0085] Furthermore, the reinforcement learning algorithm and neural network are tuned and optimized based on the model performance during the training process. This improves the stability and performance of the whole-house control strategy output model.

[0086] In practical applications, such as Figure 5As shown in the schematic block diagram, this disclosure provides a method for constructing a whole-house control strategy output model. The method includes building an initial whole-house control strategy model based on environmental data and equipment control data; outputting the initial whole-house control strategy model; and training and outputting the whole-house control strategy based on a reinforcement learning algorithm. The initial strategy output by the initial whole-house control strategy model includes one or more of the following: optimal start-stop time, optimal equipment operation combination, dew point protection trend, optimal operating parameters, and changes in terminal heating and cooling loads. The output whole-house control strategy includes one or more of the following: temperature regulation, start-stop time regulation, terminal regulation, outlet water temperature regulation, fan speed regulation, predictive equipment maintenance, and operating parameter regulation. This approach reduces energy consumption while providing user comfort, making the system more intelligent.

[0087] Combination Figure 6 As shown in the embodiments of this disclosure, an application method for a whole-house control strategy output model is provided, including:

[0088] S601, Electronic device acquires user instructions.

[0089] S602, the electronic device inputs user commands into the whole-house control strategy output model to obtain the whole-house control strategy.

[0090] In this embodiment, the whole-house control strategy output model is constructed using the construction method for the whole-house control strategy output model described in the above embodiments.

[0091] S603, an electronic device controls all smart home devices in the house according to the whole-house control strategy.

[0092] The application method for the whole-house control strategy output model provided in this disclosure can meet the user's comfort needs, energy consumption needs, or the need for a balance between comfort and energy consumption in the whole-house air system. It can improve the user's comfort in the whole-house air system while reducing the system's energy consumption, and it can also improve the system's intelligence.

[0093] For example, user commands may include optimal comfort, minimum energy consumption, or a balance between comfort and energy consumption. User commands represent user needs, and the whole-house control strategy output model can output whole-house control strategies that meet user needs based on these commands. This enables the whole-house system to satisfy user needs, enhance system intelligence, and improve the user experience.

[0094] Optionally, inputting user commands into the whole-house control strategy output model includes: denoising the user commands and identifying the user intent; inputting the user commands and user intent into the whole-house control strategy output model to obtain the whole-house control strategy. This denoising of the user command data reduces erroneous operations caused by unclear or inaccurate commands, improving the accuracy of the whole-house control strategy output by the whole-house control strategy output model. Furthermore, identifying user commands and obtaining user intent allows the whole-house control strategy output model to output the whole-house control strategy more accurately based on the user commands and user intent, thus enhancing the user experience.

[0095] Combination Figure 7 As shown in the figure, this disclosure provides a device 70 for constructing a whole-house control strategy output model, including a processor 700 and a memory 701. Optionally, the device 70 may further include a communication interface 702 and a bus 703. The processor 700, communication interface 702, and memory 701 can communicate with each other via the bus 703. The communication interface 702 can be used for information transmission. The processor 700 can call logical instructions in the memory 701 to execute the method for constructing the whole-house control strategy output model in the above embodiment, or to execute the application method for the whole-house control strategy output model in the above embodiment.

[0096] Furthermore, the logic instructions in the aforementioned memory 701 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0097] The memory 701, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this disclosure. The processor 700 executes functional applications and data processing by running the program instructions / modules stored in the memory 701, thereby implementing the method for constructing the whole-house control strategy output model in the above embodiments, or the application method for the whole-house control strategy output model.

[0098] The memory 701 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory 701 may include high-speed random access memory and may also include non-volatile memory.

[0099] This disclosure provides an electronic device, including: an electronic device body and the aforementioned construction device 70 for a whole-house control strategy output model. The construction device 70 for the whole-house control strategy output model is mounted on the electronic device body. The mounting relationship described herein is not limited to placement within the electronic device body, but also includes mounting connections with other components of the electronic device, including but not limited to physical connections, electrical connections, or signal transmission connections. Those skilled in the art will understand that the construction device 70 for the whole-house control strategy output model can be adapted to feasible electronic device bodies to achieve other feasible embodiments.

[0100] In practical applications, electronic devices may include servers, or may include device terminals, or may include smart control boxes, or may include edge devices.

[0101] The technical solutions of this disclosure can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes one or more instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in this disclosure. The aforementioned storage medium can be a non-transitory storage medium, such as a USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc., and other media capable of storing program code.

[0102] The foregoing description and accompanying drawings fully illustrate embodiments of this disclosure to enable those skilled in the art to practice them. Other embodiments may include structural, logical, electrical, procedural, and other changes. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operation may vary. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. Moreover, the terminology used in this application is for describing embodiments only and is not intended to limit the claims. As used in the description of embodiments and claims, the singular forms “a,” “an,” and “the” are intended to equally include the plural forms unless the context clearly indicates otherwise. Similarly, the term “and / or” as used in this application means including one or more of the associated listed items and all possible combinations thereof. Additionally, when used in this application, the term "comprise" and its variations "comprises" and / or "comprising" refer to the presence of stated features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof. Without further limitations, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, or apparatus that includes said element. In this document, each embodiment may focus on the differences from other embodiments, and similar or identical parts between embodiments can be referred to mutually. For methods, products, etc., disclosed in the embodiments, if they correspond to the method section disclosed in the embodiments, the relevant parts can be referred to the description of the method section.

[0103] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of this disclosure. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0104] The methods and products (including but not limited to devices and equipment) disclosed in the embodiments herein can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units may be merely a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed units may be through some interfaces, and the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected to implement this embodiment according to actual needs. Furthermore, the functional units in the embodiments of this disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0105] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. In the descriptions corresponding to the flowcharts and block diagrams in the accompanying drawings, the operations or steps corresponding to different blocks may also occur in a different order than disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. Each block in a block diagram and / or flowchart, and combinations of blocks in a block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

Claims

1. A method for constructing a whole-house control strategy output model, characterized in that, include: Based on whole-house environmental data and device control data of whole-house smart home devices, an initial model of whole-house control strategy is established. Based on reinforcement learning algorithms, environmental comfort and device energy consumption values ​​are used as rewards to alternately train the initial model of the whole-house control strategy. The initial model of the whole-house control strategy that maximizes the reward signal is used as the output model of the whole-house control strategy.

2. The construction method according to claim 1, characterized in that, Based on whole-house environmental data and device control data of smart home devices, an initial model for whole-house control strategy is established, including: Extract time and environmental features from the whole-house environmental data, and extract status and functional features from the equipment control data; By integrating time features, environmental features, state features, and functional features, a feature dataset is constructed. By using deep learning models to capture the feature relationships in the feature dataset, an initial model for whole-house control strategy is established.

3. The method according to claim 1, characterized in that, The initial model of the whole-house control strategy was trained alternately using environmental comfort and equipment energy consumption as rewards, including: The initial model of the whole-house control strategy was trained by using environmental comfort as a reward. The energy consumption value of the equipment is used as a reward to conduct a second training on the initial model of the whole house control strategy; The first and second training sessions were conducted alternately.

4. The method according to claim 3, characterized in that, Environmental comfort is determined as follows: The first whole-house control test strategy output by the initial model of the whole-house control strategy after the previous training; After implementing the first whole-house control test strategy, obtain comfort environment data that affects human comfort; Calculate the environmental comfort level of the current environment based on comfort environment data.

5. The method according to claim 4, characterized in that, Based on environmental comfort data, calculate the environmental comfort level of the current environment, including: Based on comfort environment data, calculate the thermal balance parameters of the human body; among which, comfort environment data includes multiple parameters such as air temperature, air velocity, relative humidity, mean radiant temperature, human metabolic rate, and clothing thermal resistance. Based on the thermal balance parameters, determine the evaluation index of human thermal comfort; The human thermal comfort evaluation index can be used as the environmental comfort level of the current environment; or, based on the human thermal comfort evaluation index, the percentage of dissatisfaction predicted to represent the user's degree of dissatisfaction with the current environment can be determined, and the environmental comfort level of the current environment can be determined according to the human thermal comfort evaluation index and the percentage of dissatisfaction predicted.

6. The method according to claim 3, characterized in that, Determine the equipment energy consumption value as follows: Obtain the second whole-house control test strategy output by the initial model of the whole-house control strategy after the previous training; Determine the third whole-house control test strategy selected by the user based on the second whole-house control test strategy; The second test energy consumption value for implementing the second whole-house control test strategy and the third test energy consumption value for implementing the third whole-house control test strategy were obtained respectively. The minimum test energy consumption value between the second and third test energy consumption values ​​is taken as the device energy consumption value.

7. The method according to claim 6, characterized in that, The energy consumption value of the equipment is used as a reward to perform a second training on the initial model of the whole-house control strategy, including: Determine the target whole-house control test strategy corresponding to the equipment energy consumption value; wherein, the target whole-house control test strategy includes a second whole-house control test strategy or a third whole-house control test strategy; Using energy consumption as a reward, the decision parameters and control strategies of the initial model of the whole-house control strategy are adjusted according to the target whole-house control test strategy so that the reward signal achieves the lowest energy consumption.

8. An application method for a whole-house control strategy output model, characterized in that, include: Obtain user instructions; Input the user command into the whole-house control strategy output model as described in any one of claims 1 to 7 to obtain the whole-house control strategy; Control all smart home devices according to the whole-house control strategy.

9. A device for constructing a whole-house control strategy output model, comprising a processor and a memory storing program instructions, characterized in that, The processor is configured to, when running the program instructions, execute the method for constructing a whole-house control strategy output model as described in any one of claims 1 to 7, or the method for applying a whole-house control strategy output model as described in claim 8.

10. An electronic device, characterized in that, include: The electronic device itself; The device for constructing a whole-house control strategy output model as described in claim 9 is installed on the electronic device body.

Images