Steaming oven control method and device, steaming oven and storage medium

Abstract

The embodiment of the invention provides a steaming oven control method and device, a steaming oven and a storage medium. The steaming oven control method comprises the steps that operation parameters monitored by a sensor network are obtained; constructing a current state vector based on the operation parameters; determining a target action vector according to the current state vector; determining an actuator action sequence based on the target action vector; and controlling an execution component to operate according to the actuator action sequence. The embodiment of the invention can automatically adapt to different cooking environments, and the energy-saving effect is improved; the automatic optimization of the control strategy is realized, and the response speed and accuracy of the system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of kitchen appliance production, and in particular to a steam oven control method, a steam oven control device, a steam oven and a computer-readable storage medium. Background Art

[0002] As people's living standards continue to improve, steam ovens, as a multifunctional kitchen appliance, are becoming more and more popular among home users. However, traditional steam ovens often waste energy during operation, especially when cooking for a long time. How to achieve energy saving has become an urgent problem to be solved. The current energy-saving methods mainly rely on preset rules and manual adjustments, which are difficult to cope with complex and changing cooking environments. Summary of the invention

[0003] In view of the above problems, embodiments of the present invention are proposed to provide a method for controlling a steam oven that overcomes the above problems or at least partially solves the above problems, and involve a steam oven control device, a steam oven and a computer-readable storage medium.

[0004] In order to solve the above problems, in a first aspect of the present invention, an embodiment of the present invention discloses a method for controlling a steam oven, wherein a sensor network and an execution component are arranged inside the steam oven, and the method comprises:

[0005] Obtaining operating parameters monitored by the sensor network;

[0006] constructing a current state vector based on the operating parameters;

[0007] Determining a target action vector according to the current state vector;

[0008] determining an actuator motion sequence based on the target motion vector;

[0009] The execution component is controlled to run according to the actuator action sequence.

[0010] Optionally, the step of determining a target action vector according to the current state vector includes:

[0011] Inputting the current state vector into a preset neural network model, wherein the preset neural network model is used to output at least one execution action and its score value according to the state vector;

[0012] A target action vector is determined according to the executed action and its score value.

[0013] Optionally, the step of determining a target action vector according to the executed action and its score value comprises:

[0014] Sorting the execution actions according to the score values ​​to generate an execution order;

[0015] Based on the execution order, the execution actions are combined to determine a target action vector.

[0016] Optionally, the preset neural network model is trained in the following manner:

[0017] Build an initial deep neural network;

[0018] Obtain training sample data;

[0019] Construct reward function;

[0020] Based on the training sample data, fitting is performed according to the current state function and reward function of the initial deep neural network to generate the preset neural network model.

[0021] Optionally, the step of constructing a reward function includes:

[0022] Construct a reward function based on the target temperature and energy consumption of the steam oven.

[0023] Optionally, the reward function is [Rt=-(α\|T{target}-T{oven}|+β\Et)]; wherein T{target} is the target temperature, T{oven} is the current temperature, Et is the energy consumption, α is the temperature weight coefficient, and β is the energy consumption weight coefficient.

[0024] Optionally, the preset neural network model includes a Q network.

[0025] Optionally, the step of constructing a current state vector based on the operating parameters includes:

[0026] Preprocessing the operating parameters;

[0027] The preprocessed operating parameters are combined to obtain the current state vector.

[0028] Optionally, the method further comprises:

[0029] During the operation of the execution component, reading the real-time temperature values ​​at several moments;

[0030] Calculating temperature difference data according to the real-time temperature value;

[0031] After the execution component has finished running, determining the energy consumption;

[0032] The temperature difference data and the energy consumption are used to optimize the preset neural network model.

[0033] Optionally, the step of optimizing the preset neural network model using the temperature difference data and the energy consumption comprises:

[0034] The temperature difference data and the energy consumption are used to update the reward function of the preset neural network model; the preset neural network model is used to perform fitting based on the updated reward function.

[0035] In a second aspect of the present invention, an embodiment of the present invention discloses a steam oven control device, wherein a sensor network and an execution component are arranged inside the steam oven, and the device comprises:

[0036] An acquisition module, used to acquire the operating parameters monitored by the sensor network;

[0037] A construction module, configured to construct a current state vector based on the operating parameters;

[0038] A first determination module, used to determine a target action vector according to the current state vector;

[0039] A second determination module, configured to determine an actuator action sequence based on the target action vector;

[0040] The execution module is used to control the operation of the execution component according to the actuator action sequence.

[0041] Optionally, the first determining module includes:

[0042] A model processing module, used for inputting the current state vector into a preset neural network model, wherein the preset neural network model is used for outputting at least one execution action and its score value according to the state vector;

[0043] The target action vector determination submodule is used to determine the target action vector according to the execution action and its score value.

[0044] Optionally, the target motion vector determination submodule includes:

[0045] A sorting unit, used to sort the execution actions according to the score values ​​to generate an execution order;

[0046] The first combining unit is used to combine the execution actions based on the execution order to determine a target action vector.

[0047] Optionally, the preset neural network model is trained in the following manner:

[0048] Build an initial deep neural network;

[0049] Obtain training sample data;

[0050] Construct reward function;

[0051] Based on the training sample data, fitting is performed according to the current state function and reward function of the initial deep neural network to generate the preset neural network model.

[0052] Optionally, the step of constructing a reward function includes:

[0053] Construct a reward function based on the target temperature and energy consumption of the steam oven.

[0054] Optionally, the reward function is [Rt=-(α\|T{target}-T{oven}|+β\Et)]; wherein T{target} is the target temperature, Et is the energy consumption, α is the temperature weight coefficient, and β is the energy consumption weight coefficient.

[0055] Optionally, the preset neural network model includes a Q network.

[0056] Optionally, the building blocks include:

[0057] A preprocessing submodule, used for preprocessing the operating parameters;

[0058] The second combining submodule is used to combine the preprocessed operating parameters to obtain the current state vector.

[0059] Optionally, the device further comprises:

[0060] A reading module, used to read real-time temperature values ​​at several moments during the operation of the execution component;

[0061] A temperature difference calculation module, used to calculate temperature difference data according to the real-time temperature value;

[0062] A third determination module is used to determine the energy consumption after the execution component has finished running;

[0063] An optimization module is used to optimize the preset neural network model using the temperature difference data and the energy consumption.

[0064] Optionally, the optimization module includes:

[0065] An optimization submodule is used to update the reward function of the preset neural network model using the temperature difference data and the energy consumption; the preset neural network model is used to perform fitting based on the updated reward function.

[0066] In the third aspect of the present invention, an embodiment of the present invention discloses a water heater, including a processor, a memory, and a computer program stored in the memory and capable of running on the processor, and when the computer program is executed by the processor, the steps of the steam oven control method as described above are implemented.

[0067] In a fourth aspect of the present invention, an embodiment of the present invention discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the steam oven control method as described above are implemented.

[0068] The embodiments of the present invention include the following advantages:

[0069] The embodiment of the present invention obtains the operating parameters monitored by the sensor network; constructs the current state vector based on the operating parameters; determines the target action vector based on the current state vector; determines the actuator action sequence based on the target action vector; and controls the execution component to operate according to the actuator action sequence. By obtaining the operating parameters monitored by the sensor network, the current state of the steam oven is automatically identified, and the target action vector corresponding to the optimal energy-saving target is determined based on the current state, so as to control the corresponding control component to execute the corresponding action, thereby realizing the automation of the energy-saving control of the steam oven, being able to autonomously adapt to different cooking environments, and improving the energy-saving effect; realizing the automatic optimization of the control strategy, and improving the response speed and accuracy of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is a flow chart of steps of an embodiment of a steam oven control method of the present invention;

[0071] Figure 2 is a flowchart of another steam oven control method embodiment of the present invention;

[0072] Figure 3 is a flow chart of steps of an example of a steam oven control method of the present invention;

[0073] Figure 4 It is a structural block diagram of an embodiment of a steam oven control processing device of the present invention;

[0074] Figure 5 It is a structural block diagram of a steam oven provided by an embodiment of the present invention;

[0075] Figure 6 It is a structural block diagram of a storage medium provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0077] Reference Figure 1, shows a flow chart of steps of an embodiment of a steam oven control method of the present invention, wherein a sensor network and an execution component are arranged inside the steam oven. The sensor network is composed of a plurality of sensors, which can be used to monitor parameters such as temperature, humidity, current, voltage, etc. inside the steam oven in real time. The execution component is the execution parts such as various control elements, heating elements, fans, and steam generators inside the steam oven.

[0078] The steam oven control method may specifically include the following steps:

[0079] Step 101, obtaining the operating parameters monitored by the sensor network;

[0080] After the steam oven is started, when the user needs the steam oven to perform corresponding cooking, the operating parameters monitored by the sensor network can be obtained. These operating parameters are various parameters inside the steam oven when it is running. The operating parameters such as temperature, humidity, current, and voltage inside the steam oven can be obtained. The sensor network communicates with the controller through a communication link, and the controller obtains the operating parameters from the sensor network through the communication link.

[0081] Step 102, constructing a current state vector based on the operating parameters;

[0082] Based on the obtained various operating parameters, a current state vector is constructed, and the current operating status of the steam oven is represented by the current state vector.

[0083] Step 103, determining a target action vector according to the current state vector;

[0084] According to the current state vector, the operation status inside the steam oven can be determined, the optimal energy saving status corresponding to this operation status can be determined, the control strategy required to achieve the energy saving status can be determined, and the target action vector can be generated. The target action vector is used to represent the control strategy to be executed.

[0085] Step 104, determining an actuator action sequence based on the target action vector;

[0086] The target action vector is parsed to determine each action and the corresponding control parameters, and the actuator action sequence is generated. The actuator action sequence represents the control flow of the control execution component.

[0087] Step 105, controlling the execution component to run according to the execution device action sequence.

[0088] After the actuator action sequence is obtained, the corresponding actuator movement can be controlled according to each action in the actuator action sequence and the corresponding control parameters, thereby completing the overall control.

[0089] The embodiment of the present invention obtains the operating parameters monitored by the sensor network; constructs the current state vector based on the operating parameters; determines the target action vector based on the current state vector; determines the actuator action sequence based on the target action vector; and controls the execution component to operate according to the actuator action sequence. By obtaining the operating parameters monitored by the sensor network, the current state of the steam oven is automatically identified, and the target action vector corresponding to the optimal energy-saving target is determined based on the current state, so as to control the corresponding control component to execute the corresponding action, thereby realizing the automation of the energy-saving control of the steam oven, being able to autonomously adapt to different cooking environments, and improving the energy-saving effect; realizing the automatic optimization of the control strategy, and improving the response speed and accuracy of the system.

[0090] Reference Figure 2 , shows a flow chart of steps of another embodiment of a steam oven control method of the present invention, wherein a sensor network and an execution component are arranged inside the steam oven. The sensor network is composed of a plurality of sensors, which can be used to monitor parameters such as temperature, humidity, current, voltage, etc. inside the steam oven in real time. The execution component is various control elements inside the steam oven, such as heating elements, fans, steam generators, and other execution parts. Heating elements: The heating elements of the steam oven are usually made of high-temperature resistant materials such as nickel-chromium alloys, such as Cr20Ni80 nickel-chromium electric furnace wires. These heating elements are heated by electric current to provide the required heat energy for the steam oven. The heating elements are usually distributed on the inner wall or specific positions of the steam oven to ensure uniform heat distribution. Fan: The fan inside the steam oven (such as a brushless motor fan) is used for hot air circulation inside the machine to ensure uniform heat inside the machine cavity. The operation of the fan helps to accelerate the transfer and distribution of heat and improve cooking efficiency. The steam generator is a key component in the steam oven to realize the steam cooking function. It generates steam by heating and discharges the steam into the steam oven to provide the required steam environment for the food.

[0091] The steam oven control method may specifically include the following steps:

[0092] Step 201, obtaining the operating parameters monitored by the sensor network;

[0093] The embodiment of the present invention can be applied to a central control system in a steam oven. The central control system of the steam oven is connected to a sensor network and an execution component through relevant data communication lines and performs data transmission. The sensor network can monitor the operating parameters inside the steam oven in real time. Including but not limited to the operating parameters such as temperature, humidity, current, and voltage inside the steam oven. The central control system obtains the operating parameters monitored by the sensor network.

[0094] Step 202, constructing a current state vector based on the operating parameters;

[0095] The acquired operating parameters may be used to construct a current state vector according to a preset element format to determine the current operating status of the steam oven.

[0096] For example, the state space can be defined in advance: the state space includes all operating parameters monitored by the sensor network. The current state vector can be expressed as: [St = [T{oven}, H{oven}, I{oven}, V{oven}]], where (T{oven}) represents the internal temperature of the steam oven, (H{oven}) represents the internal humidity of the steam oven, (I{oven}) represents the current, and (V{oven}) represents the voltage.

[0097] The current state vector can be constructed by combining the acquired operating parameters in the format of the current state vector.

[0098] In an optional embodiment of the present invention, the step of constructing the current state vector based on the operating parameters includes: preprocessing the operating parameters; and combining the preprocessed operating parameters to obtain the current state vector.

[0099] After obtaining the operating parameters, the operating parameters can be preprocessed first, including but not limited to data cleaning, data integration and data transformation. Data cleaning includes: detecting and processing missing values, using statistical methods or machine learning algorithms to detect missing values ​​in the data, and using appropriate strategies to fill them, such as mean, median, mode filling, or using interpolation, model prediction and other methods. Processing outliers, identifying outliers in the data, which may be caused by data entry errors or measurement errors. For outliers, you can delete, replace them with missing values ​​or use statistical methods to correct them. Data smoothing, processing the data through certain algorithms to remove noise and fluctuations and make the data smoother. Commonly used data smoothing methods include binning (equal frequency or equal width binning), regression method, etc. Processing inconsistent data, for inconsistencies in the records, it is necessary to clean and correct them to ensure the accuracy and consistency of the data. Data integration includes entity recognition, matching real-world entities from multiple different information sources, and ensuring that different field names in different databases can point to the same entity. Data redundancy processing, detecting and processing redundant information in the data set to improve the quality and consistency of the data. Data value conflict detection and processing: when merging different data sources, there may be data value conflicts, which need to be detected and processed through data comparison, rule detection and other methods. Dimension reduction: reduce the dimension of data by deleting irrelevant attributes or reducing the number of attributes. Numerical reduction: select alternative data representations to reduce the amount of data, such as using parameter models (such as linear regression models) to replace actual data. Data compression: use data compression algorithms (such as principal component analysis PCA, singular value decomposition SVD, etc.) to map data from high-dimensional space to low-dimensional space to reduce data storage and computational burden. Data transformation includes data normalization, which scales the data according to a certain ratio so that it falls in a specific area (such as the [0,1] interval or the [-1,1] interval) to eliminate the impact of dimension on data analysis results. Data discretization: segmenting continuous data into discrete intervals helps convert continuous data into categorical data. Data sparsification: For discrete and nominal variables, when ordered LabelEncoder cannot be used, it is usually considered to sparsify the variables by making them 0,1 dummy variables.

[0100] Then the preprocessed operating parameters are combined according to the format of the current state vector to construct the current state vector.

[0101] Step 203, inputting the current state vector into a preset neural network model, wherein the preset neural network model is used to output at least one execution action and its score value according to the state vector;

[0102] After obtaining the current state vector, the current state vector can be input into a preset neural network model, which processes the current state vector and outputs at least one execution action and its score value that meet the energy-saving target. The higher the score value, the better the energy-saving target can be met by executing the execution action.

[0103] In an optional embodiment of the present invention, the preset neural network model is trained in the following manner: constructing an initial deep neural network; acquiring training sample data; constructing a reward function; based on the training sample data, fitting according to the current state function and reward function of the initial deep neural network to generate the preset neural network model.

[0104] An initial deep neural network can be constructed. The specific initial deep neural network can be determined according to the actual situation, such as a Q network. Then, training sample data is obtained and the initial deep neural network is trained using the training sample data. A reward function is constructed so that the model converges to the corresponding gradient. Finally, the current state function and reward function of the initial deep neural network are fitted to generate a preset neural network model.

[0105] Specifically, the step of constructing a reward function includes: constructing a reward function based on a target temperature and energy consumption of the steam oven.

[0106] In practical applications, the temperature stability and energy consumption of the steam oven can be used as optimization targets. The reward function can be designed based on the internal temperature and energy consumption of the steam oven, with the goal of maximizing temperature stability and minimizing energy consumption. Construct a reward function based on the target temperature and energy consumption of the steam oven.

[0107] Furthermore, the reward function is [Rt=-(α\|T{target}-T{oven}|+β\Et)]; wherein T{target} is the target temperature, Et is the energy consumption, α is the temperature weight coefficient, and β is the energy consumption weight coefficient.

[0108] In an example of the present invention, the preset neural network model includes a Q network.

[0109] Q network, which learns the approximation of the Q function through a neural network, takes the state as input and the action as output. Specifically, DQN (Deep Q Learning Network) uses a deep convolutional neural network (CNN) or a fully connected neural network (FCN) to process input data (such as game state) and output the Q value of each action. Then, the action is selected according to a greedy strategy or a random strategy under a certain probability. Its features include: Experience replay, DQN stores the agent's experience in the environment (state, action, reward, next state) in an experience pool, and then randomly extracts a batch of experience from it for training. This method can break the correlation between data and improve the training effect. Target network, DQN uses two neural networks, one is the main network (MainNetwork), which is used to select actions and calculate Q values; the other is the target network (Target Network), which is used to calculate the target Q value. The parameters of the target network are updated regularly to reduce the fluctuation of the objective function and improve the stability and convergence speed of training. Double DQN, uses the main network to select the optimal action, and uses the target network to calculate the Q value to reduce estimation bias and improve learning efficiency and stability.

[0110] In order to make those skilled in the art understand the model training process in the embodiment of the present invention, an example is used to illustrate the following:

[0111] 1. Initialize and interact with the environment

[0112] Initialize the Q-network: First, build a deep neural network, called the Q-network, to approximate the Q-function, which is the prediction of the reward for taking a specific action in a given state.

[0113] Initialize the experience replay buffer: Set up an experience replay mechanism to randomly extract previous interaction data to train the Q network to avoid excessive data correlation.

[0114] 2. Collect empirical data

[0115] Observe the state: first obtain the current state of the steam oven, that is, the state vector (St = [T{oven}, H{oven}, I{oven}, V{oven}]).

[0116] Select action: Based on the current state, the Q network predicts the action value and selects the action with the highest Q value or performs exploration (such as using the ε-greedy strategy).

[0117] Execute actions and receive feedback: Execute the selected action (At = [A{heating element}, A{fan}, A{steam generator}]) and observe the new state function (S{t+1}) and reward function (R{t+1}) of the steam oven.

[0118] Store experience: Store the interaction experience ((St,At,R{t+1},S{t+1})) into the experience playback buffer.

[0119] 3. Training the Q-Network

[0120] Randomly sample from the experience replay buffer: select a batch of previous experience data to train the Q network.

[0121] Calculate the target Q value: Based on the experience ((R{t+1},S{t+1})), use the target network (a regularly updated replica network) to predict the maximum Q value of the next state (R{t+1}+\gamma\max Q(s_{t+1},a)), where (\gamma) is a discount factor used to balance the relationship between immediate rewards and future rewards.

[0122] Update the Q network: Use the mean squared error loss function to compare the current Q network's prediction with the target Q value and optimize it to reduce the gap between the prediction and the actual value.

[0123] Step 204, determining a target action vector according to the executed action and its score value;

[0124] Then, according to the executed action and its score value, the target action to be executed and the corresponding target action vector are determined.

[0125] Specifically, the step of determining the target action vector based on the execution actions and their score values ​​includes: sorting the execution actions according to the score values ​​to generate an execution order; and combining the execution actions based on the execution order to determine the target action vector.

[0126] The execution actions can be sorted based on the size of the score value, that is, the larger the score value is, the higher the score value is, and the smaller the score value is, the lower the score value is, to generate the execution order. Then, the execution actions are combined according to the execution order to determine the target action vector. The target action vector definition includes the operation of all steam oven control elements, such as the switch state and power setting of the heating element, the speed of the fan, the working state of the steam generator, etc. The action vector is expressed as: [At = [A{heating element}, A{fan}, A{steam generator}]].

[0127] Step 205, determining an actuator action sequence based on the target action vector;

[0128] According to the target action vector analysis, the actuator action sequence corresponding to the actuator to be controlled is determined. The actuator action sequence is used to represent the action order of the actuator.

[0129] Step 206, controlling the execution component to run according to the execution action sequence;

[0130] According to the execution action of each execution component in the actuator action sequence, the execution component can be controlled to execute the corresponding execution action and then the heating amount can be adjusted; the corresponding cooking function can be achieved while reducing energy waste.

[0131] Step 207, reading real-time temperature values ​​at several moments during the operation of the execution component;

[0132] When the execution component performs the corresponding action, that is, during the operation of the execution component, the real-time temperature value in the steam oven can be continuously detected through the sensor network. The central control system can read the real-time temperature values ​​at multiple different times.

[0133] Step 208, calculating temperature difference data according to the real-time temperature value;

[0134] The temperature difference data between the operation periods is calculated based on the multiple real-time temperature values ​​read, and the temperature stability during the operation of the steam oven is determined based on the temperature difference data.

[0135] Step 209, after the execution component has finished running, determine the energy consumption;

[0136] After the execution components have finished running, the energy consumption can also be determined based on the data from the sensor network.

[0137] Step 210: Optimize the preset neural network model using the temperature difference data and the energy consumption.

[0138] The temperature difference data and energy consumption are used to update and optimize the preset neural network model to improve the subsequent processing capabilities of the preset neural network model.

[0139] In an optional embodiment of the present invention, the step of optimizing the preset neural network model using the temperature difference data and the energy consumption includes: updating the reward function of the preset neural network model using the temperature difference data and the energy consumption; the preset neural network model is used for fitting based on the updated reward function.

[0140] In an embodiment of the present invention, the temperature difference data and the energy consumption can be used to update the reward function of the preset neural network model. When the reward function is updated, the preset neural network model is fitted based on the updated reward function, thereby updating the relevant model parameters.

[0141] The update of the model can be to update the target network periodically. Continuing with the Q network as an example, the weights of the current Q network are regularly copied to the target network to ensure the stability of the training and reduce the fluctuation of the target value. Balance between exploration and utilization. According to the ε-greedy strategy, the probability of exploration is gradually reduced and the frequency of using the current strategy is increased. Evaluate and adjust the strategy. Over time, the optimization of the Q network will gradually improve the quality of the control strategy. By monitoring the performance of the steam oven in real time, the strategy is evaluated and tuned. Continuous learning and improvement: The model continues to learn in re-interaction and continuously updates the Q network until it reaches convergence or meets the preset performance standards. At this time, the Q network can provide an optimized strategy to control the steam oven, making the temperature stable and energy consumption low.

[0142] The embodiment of the present invention obtains the operating parameters monitored by the sensor network; constructs the current state vector based on the operating parameters; inputs the current state vector into a preset neural network model, and the preset neural network model is used to output at least one execution action and its score value according to the state vector; determines the target action vector according to the execution action and its score value; determines the actuator action sequence based on the target action vector; and controls the execution component to run according to the actuator action sequence. During the operation of the execution component, read the real-time temperature values ​​at several moments; calculate the temperature difference data according to the real-time temperature values; after the execution component is completed, determine the energy consumption; and optimize the preset neural network model using the temperature difference data and the energy consumption. The real-time operating parameters of the steam oven are obtained through sensors. The optimal control strategy is generated according to the current operating state by using the deep reinforcement learning algorithm. A reward function based on the internal temperature and energy consumption of the steam oven is constructed to achieve energy saving effect. The efficient deployment and iteration of the deep reinforcement learning algorithm is realized in the steam oven system. The deep reinforcement learning network technology is applied to improve the intelligence of the steam oven, so that it can adapt to different cooking environments autonomously and improve the energy saving effect. Through the application of the deep reinforcement learning algorithm, the automatic optimization of the control strategy is realized to reduce energy waste. Improve the system's response speed and accuracy, and be able to adjust control strategies in real time to meet different cooking needs. Through continuous learning and iteration, continuously improve its own performance and enhance user experience and satisfaction.

[0143] In order to make the implementation process of the embodiment of the present invention clear to those skilled in the art, an example is used to illustrate the following:

[0144] The energy-saving control system of the steam oven of the present invention mainly includes the following modules: The specific execution process may include the following modules: Sensor network: composed of multiple sensors, used to monitor the temperature, humidity, current, voltage and other parameters inside the steam oven in real time. Data acquisition and processing module: responsible for collecting data reported by the sensor and preprocessing the data, including data cleaning, standardization and other operations. Deep reinforcement learning module: the core part, using the deep reinforcement learning algorithm to generate the optimal control strategy according to the current operating status. Execution module: according to the strategy output by the deep reinforcement learning module, control the heating elements, fans, steam generators, etc. of the steam oven. Feedback module: based on the execution results and real-time operation data, generate a reward function to guide the strategy optimization of the deep reinforcement learning module.

[0145] Reference Figure 3 , the execution process is described:

[0146] Data acquisition and preprocessing: The sensor network monitors the temperature, humidity, current, voltage and other parameters inside the steam oven in real time and sends the data to the data acquisition and processing module. The data acquisition and processing module preprocesses the data and generates a standardized state vector.

[0147] Control strategy generation: The deep reinforcement learning module generates the optimal action vector based on the current state vector using the trained DQN model. The action vector includes the operating instructions for each control element of the steam oven.

[0148] Control strategy execution: The execution module controls the operation of the steam oven's heating elements, fans, steam generators, etc. based on the action vectors. The equipment status is adjusted in real time to optimize temperature stability and energy consumption.

[0149] Feedback and optimization: The feedback module calculates the reward function based on the execution results and new operation data. The deep reinforcement learning module uses the reward function to update the DQN model and optimize the control strategy. Over time, the system continues to learn and iterate to improve the effectiveness of the control strategy.

[0150] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0151] Reference Figure 4, shows a structural block diagram of an embodiment of a steam oven control device of the present invention, wherein a sensor network and an execution component are arranged inside the steam oven, and the steam oven control device may specifically include the following modules:

[0152] An acquisition module 401 is used to acquire the operating parameters monitored by the sensor network;

[0153] A construction module 402, configured to construct a current state vector based on the operating parameters;

[0154] A first determination module 403, configured to determine a target motion vector according to the current state vector;

[0155] A second determination module 404, configured to determine an actuator action sequence based on the target action vector;

[0156] The execution module 405 is used to control the execution component to run according to the actuator action sequence.

[0157] In an optional embodiment of the present invention, the first determining module 403 includes:

[0158] A model processing module, used for inputting the current state vector into a preset neural network model, wherein the preset neural network model is used for outputting at least one execution action and its score value according to the state vector;

[0159] The target action vector determination submodule is used to determine the target action vector according to the execution action and its score value.

[0160] In an optional embodiment of the present invention, the target motion vector determination submodule includes:

[0161] A sorting unit, used to sort the execution actions according to the score values ​​to generate an execution order;

[0162] The first combining unit is used to combine the execution actions based on the execution order to determine a target action vector.

[0163] In an optional embodiment of the present invention, the preset neural network model is trained in the following manner:

[0164] Build an initial deep neural network;

[0165] Obtain training sample data;

[0166] Construct reward function;

[0167] Based on the training sample data, fitting is performed according to the current state function and reward function of the initial deep neural network to generate the preset neural network model.

[0168] In an optional embodiment of the present invention, the step of constructing a reward function includes:

[0169] Construct a reward function based on the target temperature and energy consumption of the steam oven.

[0170] In an optional embodiment of the present invention, the reward function is [Rt=-(α\|T{target}-T{oven}|+β\Et)]; wherein T{target} is the target temperature, Et is the energy consumption, α is the temperature weight coefficient, and β is the energy consumption weight coefficient.

[0171] In an optional embodiment of the present invention, the preset neural network model includes a Q network.

[0172] In an optional embodiment of the present invention, the construction module 402 includes:

[0173] A preprocessing submodule, used for preprocessing the operating parameters;

[0174] The second combining submodule is used to combine the preprocessed operating parameters to obtain the current state vector.

[0175] In an optional embodiment of the present invention, the device further comprises:

[0176] A reading module, used to read real-time temperature values ​​at several moments during the operation of the execution component;

[0177] A temperature difference calculation module, used to calculate temperature difference data according to the real-time temperature value;

[0178] A third determination module is used to determine the energy consumption after the execution component has finished running;

[0179] An optimization module is used to optimize the preset neural network model using the temperature difference data and the energy consumption.

[0180] In an optional embodiment of the present invention, the optimization module includes:

[0181] An optimization submodule is used to update the reward function of the preset neural network model using the temperature difference data and the energy consumption; the preset neural network model is used to perform fitting based on the updated reward function.

[0182] The embodiment of the present invention obtains the operating parameters monitored by the sensor network; constructs the current state vector based on the operating parameters; inputs the current state vector into a preset neural network model, and the preset neural network model is used to output at least one execution action and its score value according to the state vector; determines the target action vector according to the execution action and its score value; determines the actuator action sequence based on the target action vector; and controls the execution component to run according to the actuator action sequence. During the operation of the execution component, read the real-time temperature values ​​at several moments; calculate the temperature difference data according to the real-time temperature values; after the execution component is completed, determine the energy consumption; and optimize the preset neural network model using the temperature difference data and the energy consumption. The real-time operating parameters of the steam oven are obtained through sensors. The optimal control strategy is generated according to the current operating state by using the deep reinforcement learning algorithm. A reward function based on the internal temperature and energy consumption of the steam oven is constructed to achieve energy saving effect. The efficient deployment and iteration of the deep reinforcement learning algorithm is realized in the steam oven system. The deep reinforcement learning network technology is applied to improve the intelligence of the steam oven, so that it can adapt to different cooking environments autonomously and improve the energy saving effect. Through the application of the deep reinforcement learning algorithm, the automatic optimization of the control strategy is realized to reduce energy waste. Improve the system's response speed and accuracy, and be able to adjust control strategies in real time to meet different cooking needs. Through continuous learning and iteration, continuously improve its own performance and enhance user experience and satisfaction.

[0183] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0184] Reference Figure 5 The embodiment of the present invention further provides a steam oven, comprising:

[0185] A processor 501 and a storage medium 502, wherein the storage medium 502 stores a computer program executable by the processor 501, and when the steam oven is controlled to run, the processor 501 executes the computer program to implement the steam oven control method as described in any one of the embodiments of the present invention.

[0186] A sensor network and an execution component are arranged inside the steam oven, and the steam oven control method includes:

[0187] Obtaining operating parameters monitored by the sensor network;

[0188] constructing a current state vector based on the operating parameters;

[0189] Determining a target action vector according to the current state vector;

[0190] determining an actuator motion sequence based on the target motion vector;

[0191] The execution component is controlled to run according to the actuator action sequence.

[0192] Optionally, the step of determining a target action vector according to the current state vector includes:

[0193] Inputting the current state vector into a preset neural network model, wherein the preset neural network model is used to output at least one execution action and its score value according to the state vector;

[0194] A target action vector is determined according to the executed action and its score value.

[0195] Optionally, the step of determining a target action vector according to the executed action and its score value comprises:

[0196] Sorting the execution actions according to the score values ​​to generate an execution order;

[0197] Based on the execution order, the execution actions are combined to determine a target action vector.

[0198] Optionally, the preset neural network model is trained in the following manner:

[0199] Build an initial deep neural network;

[0200] Obtain training sample data;

[0201] Construct reward function;

[0202] Based on the training sample data, fitting is performed according to the current state function and reward function of the initial deep neural network to generate the preset neural network model.

[0203] Optionally, the step of constructing a reward function includes:

[0204] Construct a reward function based on the target temperature and energy consumption of the steam oven.

[0205] Optionally, the reward function is [Rt=-(α\|T{target}-T{oven}|+β\Et)]; wherein T{target} is the target temperature, Et is the energy consumption, α is the temperature weight coefficient, and β is the energy consumption weight coefficient.

[0206] Optionally, the preset neural network model includes a Q network.

[0207] Optionally, the step of constructing a current state vector based on the operating parameters includes:

[0208] Preprocessing the operating parameters;

[0209] The preprocessed operating parameters are combined to obtain the current state vector.

[0210] Optionally, the method further comprises:

[0211] During the operation of the execution component, reading the real-time temperature values ​​at several moments;

[0212] Calculating temperature difference data according to the real-time temperature value;

[0213] After the execution component has finished running, determining the energy consumption;

[0214] The temperature difference data and the energy consumption are used to optimize the preset neural network model.

[0215] Optionally, the step of optimizing the preset neural network model using the temperature difference data and the energy consumption includes:

[0216] The temperature difference data and the energy consumption are used to update the reward function of the preset neural network model; the preset neural network model is used to perform fitting based on the updated reward function.

[0217] The embodiment of the present invention obtains the operating parameters monitored by the sensor network; constructs the current state vector based on the operating parameters; inputs the current state vector into a preset neural network model, and the preset neural network model is used to output at least one execution action and its score value according to the state vector; determines the target action vector according to the execution action and its score value; determines the actuator action sequence based on the target action vector; and controls the execution component to run according to the actuator action sequence. During the operation of the execution component, read the real-time temperature values ​​at several moments; calculate the temperature difference data according to the real-time temperature values; after the execution component is completed, determine the energy consumption; and optimize the preset neural network model using the temperature difference data and the energy consumption. The real-time operating parameters of the steam oven are obtained through sensors. The optimal control strategy is generated according to the current operating state by using the deep reinforcement learning algorithm. A reward function based on the internal temperature and energy consumption of the steam oven is constructed to achieve energy saving effect. The efficient deployment and iteration of the deep reinforcement learning algorithm is realized in the steam oven system. The deep reinforcement learning network technology is applied to improve the intelligence of the steam oven, so that it can adapt to different cooking environments autonomously and improve the energy saving effect. Through the application of the deep reinforcement learning algorithm, the automatic optimization of the control strategy is realized to reduce energy waste. Improve the system's response speed and accuracy, and be able to adjust control strategies in real time to meet different cooking needs. Through continuous learning and iteration, continuously improve its own performance and enhance user experience and satisfaction.

[0218] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the processor.

[0219] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0220] Reference Figure 6 The embodiment of the present invention further provides a computer-readable storage medium 601, on which a computer program is stored. When the computer program is executed by a processor, the steam oven control method as described in any one of the embodiments of the present invention is executed.

[0221] A sensor network and an execution component are arranged inside the steam oven, and the steam oven control method includes:

[0222] Obtaining operating parameters monitored by the sensor network;

[0223] constructing a current state vector based on the operating parameters;

[0224] Determining a target action vector according to the current state vector;

[0225] determining an actuator motion sequence based on the target motion vector;

[0226] The execution component is controlled to run according to the actuator action sequence.

[0227] Optionally, the step of determining a target action vector according to the current state vector includes:

[0228] Inputting the current state vector into a preset neural network model, wherein the preset neural network model is used to output at least one execution action and its score value according to the state vector;

[0229] A target action vector is determined according to the executed action and its score value.

[0230] Optionally, the step of determining a target action vector according to the executed action and its score value comprises:

[0231] Sorting the execution actions according to the score values ​​to generate an execution order;

[0232] Based on the execution order, the execution actions are combined to determine a target action vector.

[0233] Optionally, the preset neural network model is trained in the following manner:

[0234] Build an initial deep neural network;

[0235] Obtain training sample data;

[0236] Constructing reward function;

[0237] Based on the training sample data, fitting is performed according to the current state function and reward function of the initial deep neural network to generate the preset neural network model.

[0238] Optionally, the step of constructing a reward function includes:

[0239] Construct a reward function based on the target temperature and energy consumption of the steam oven.

[0240] Optionally, the reward function is [Rt=-(α\|T{target}-T{oven}|+β\Et)]; wherein T{target} is the target temperature, Et is the energy consumption, α is the temperature weight coefficient, and β is the energy consumption weight coefficient.

[0241] Optionally, the preset neural network model includes a Q network.

[0242] Optionally, the step of constructing a current state vector based on the operating parameters includes:

[0243] Preprocessing the operating parameters;

[0244] The preprocessed operating parameters are combined to obtain the current state vector.

[0245] Optionally, the method further comprises:

[0246] During the operation of the execution component, reading the real-time temperature values ​​at several moments;

[0247] Calculating temperature difference data according to the real-time temperature value;

[0248] After the execution component has finished running, determining the energy consumption;

[0249] The temperature difference data and the energy consumption are used to optimize the preset neural network model.

[0250] Optionally, the step of optimizing the preset neural network model using the temperature difference data and the energy consumption comprises:

[0251] The temperature difference data and the energy consumption are used to update the reward function of the preset neural network model; the preset neural network model is used to perform fitting based on the updated reward function.

[0252] The embodiment of the present invention obtains the operating parameters monitored by the sensor network; constructs the current state vector based on the operating parameters; inputs the current state vector into a preset neural network model, and the preset neural network model is used to output at least one execution action and its score value according to the state vector; determines the target action vector according to the execution action and its score value; determines the actuator action sequence based on the target action vector; and controls the execution component to run according to the actuator action sequence. During the operation of the execution component, read the real-time temperature values ​​at several moments; calculate the temperature difference data according to the real-time temperature values; after the execution component is completed, determine the energy consumption; and optimize the preset neural network model using the temperature difference data and the energy consumption. The real-time operating parameters of the steam oven are obtained through sensors. The optimal control strategy is generated according to the current operating state by using the deep reinforcement learning algorithm. A reward function based on the internal temperature and energy consumption of the steam oven is constructed to achieve energy saving effect. The efficient deployment and iteration of the deep reinforcement learning algorithm is realized in the steam oven system. The deep reinforcement learning network technology is applied to improve the intelligence of the steam oven, so that it can adapt to different cooking environments autonomously and improve the energy saving effect. Through the application of the deep reinforcement learning algorithm, the automatic optimization of the control strategy is realized to reduce energy waste. Improve the system's response speed and accuracy, and be able to adjust control strategies in real time to meet different cooking needs. Through continuous learning and iteration, continuously improve its own performance and enhance user experience and satisfaction.

[0253] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0254] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0255] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0256] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0257] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0258] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0259] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0260] The above is a detailed introduction to a steam oven control method provided by the present invention, which involves a steam oven control device, a steam oven and a computer-readable storage medium. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for a general technician in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A steam oven control method, characterized in that: A sensor network and an execution component are arranged inside the steam oven, and the method comprises: Obtaining operating parameters monitored by the sensor network; constructing a current state vector based on the operating parameters; Determining a target action vector according to the current state vector; determining an actuator motion sequence based on the target motion vector; The execution component is controlled to run according to the actuator action sequence.

2. The method according to claim 1, characterized in that The step of determining the target action vector according to the current state vector comprises: Inputting the current state vector into a preset neural network model, wherein the preset neural network model is used to output at least one execution action and its score value according to the state vector; A target action vector is determined according to the executed action and its score value.

3. The method according to claim 2, characterized in that The step of determining the target action vector according to the executed action and its score value comprises: Sorting the execution actions according to the score values ​​to generate an execution order; Based on the execution order, the execution actions are combined to determine a target action vector.

4. The method according to claim 2, characterized in that: The preset neural network model is trained in the following way: Build an initial deep neural network; Obtain training sample data; Constructing reward function; Based on the training sample data, fitting is performed according to the current state function and reward function of the initial deep neural network to generate the preset neural network model.

5. The method according to claim 4, characterized in that The steps of constructing the reward function include: Construct a reward function based on the target temperature and energy consumption of the steam oven.

6. The method according to claim 5, characterized in that The reward function is [Rt=-(α\|T{target}-T{oven}|+β\Et)]; wherein T{target} is the target temperature, T{oven} is the current temperature, Et is the energy consumption, α is the temperature weight coefficient, and β is the energy consumption weight coefficient.

7. The method according to any one of claims 2 to 6, characterized in that: The preset neural network model includes a Q network.

8. The method according to claim 1, characterized in that: The step of constructing the current state vector based on the operating parameters comprises: Preprocessing the operating parameters; The preprocessed operating parameters are combined to obtain the current state vector.

9. The method according to claim 2, characterized in that: The method further comprises: During the operation of the execution component, reading the real-time temperature values ​​at several moments; Calculating temperature difference data according to the real-time temperature value; After the execution component has finished running, determining the energy consumption; The temperature difference data and the energy consumption are used to optimize the preset neural network model.

10. The method according to claim 9, characterized in that The step of optimizing the preset neural network model by using the temperature difference data and the energy consumption comprises: The temperature difference data and the energy consumption are used to update the reward function of the preset neural network model; the preset neural network model is used to perform fitting based on the updated reward function.

11. A steam oven control device, characterized in that: A sensor network and an execution component are arranged inside the steam oven, and the device comprises: An acquisition module, used to acquire the operating parameters monitored by the sensor network; A construction module, configured to construct a current state vector based on the operating parameters; A first determination module, used to determine a target action vector according to the current state vector; A second determination module, configured to determine an actuator action sequence based on the target action vector; The execution module is used to control the operation of the execution component according to the actuator action sequence.

12. A steam oven, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the steam oven control method according to any one of claims 1 to 10 are implemented.

13. A computer-readable storage medium, characterized in that: The computer readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the steam oven control method according to any one of claims 1 to 10 are implemented.

Images