Air purifier control method, system and device and computer equipment

Abstract

The invention relates to an air purifier control method, system and device, computer equipment, a storage medium and a computer program product. The method comprises the steps that in the environment where the air purifier is located, actual parameter information collected for a plurality of environment parameters and expected parameter information of a user for a target parameter in the environment parameters are obtained; performing correlation analysis on the environmental parameters based on the actual parameter information, and determining an environmental characteristic relation of the environment; the environment characteristic relation comprises an incidence relation between the environment parameters and / or an incidence relation between parameter information corresponding to different spatial positions of the same environment parameter in the environment; and the control mode of the air purifier is determined according to the expected parameter information of the target parameters and the environment characteristic relation. By adopting the method, the diversity of control modes can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of household appliances, and in particular to an air purifier control method, system, device, computer equipment, storage medium and computer program product. Background Art

[0002] With the development of household appliance technology, air purifiers have emerged. Air purifiers can drive the fan blades through the motor, thereby driving the air flow through the filter or electric purification module, absorbing the polluted gases, particles, and harmful bacteria and viruses, and then filtering the air.

[0003] In traditional technology, the motor power of the air purifier is adjusted according to the concentration of pollutants in the environment where the air purifier is located to achieve air purification. The control mode is single and cannot meet the increasingly diversified user needs. Summary of the invention

[0004] Based on this, it is necessary to provide an air purifier control method, system, device, computer equipment, storage medium and computer program product that can reduce the complexity of operation in order to address the technical problem of the single control mode of traditional air purifiers.

[0005] In a first aspect, the present application provides an air purifier control method. The method comprises:

[0006] In the environment where the air purifier is located, obtaining actual parameter information collected for multiple environmental parameters respectively, and the user's expected parameter information for the target parameter in each of the environmental parameters;

[0007] Based on the actual parameter information, each environmental parameter is subjected to correlation analysis to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each environmental parameter and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment;

[0008] The control mode of the air purifier is determined according to the expected parameter information of the target parameter and the environmental characteristic relationship.

[0009] In one embodiment, determining the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship includes:

[0010] Determining the expected parameter information of each of the environmental parameters associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter;

[0011] Based on each of the expected parameter information, a control mode of the air purifier is determined.

[0012] In one embodiment, each of the environmental parameters corresponds to multiple parameter information at multiple spatial locations; the correlation analysis of each of the environmental parameters based on each of the actual parameter information to determine the environmental characteristic relationship of the environment includes:

[0013] For each of the spatial positions, integrating the parameter information of each of the environmental parameters in the spatial position to obtain a multi-dimensional information group with the spatial position and the environmental parameters as dimensional features;

[0014] Determine, from the multidimensional information group, a plurality of multidimensional sub-information groups matching the convolution size of the convolution kernel;

[0015] For each of the multidimensional sub-information groups, performing a dot product operation on the multidimensional sub-information group based on the convolution kernel to obtain an output value corresponding to the multidimensional sub-information group;

[0016] Based on the output values ​​of each of the multi-dimensional sub-information groups, an environmental characteristic map of the environment is obtained; the environmental characteristic map is used to characterize the environmental characteristic relationship of the environment.

[0017] In one embodiment, determining the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship includes:

[0018] Taking the environment feature map and the expected parameter information of the target parameter as inputs of a forgetting function, and obtaining a weight matrix with the same dimension as the environment feature map;

[0019] Updating the environment feature map based on the weight matrix to obtain an updated feature map of the environment;

[0020] Based on the updated characteristic map, a control mode of the air purifier is determined.

[0021] In one embodiment, the update characteristic graph includes a plurality of graphs; each of the update characteristic graphs corresponds to a parameter collection moment; and determining the control mode of the air purifier based on the update characteristic graph includes:

[0022] For each of the updated feature maps, flatten the updated feature map to obtain a feature vector corresponding to the updated feature map;

[0023] According to the time sequence between the parameter collection moments, each of the feature vectors is analyzed in turn based on the long short-term memory network model to determine the final update state of the memory cell of the feature vector in the long short-term memory network model;

[0024] According to the final updated state, a control mode corresponding to the air purifier is determined.

[0025] In one embodiment, the method further comprises:

[0026] Determining, based on an input gate of the long short-term memory network model, newly added information corresponding to the feature vector in the memory cell;

[0027] Determining, based on a forget gate of the long short-term memory network model, discarding information corresponding to the feature vector in the memory cell;

[0028] Based on the newly added information and the discarded information, the state of the memory cell is updated to obtain an updated state.

[0029] In a second aspect, the present application also provides an air purifier control system, the system comprising:

[0030] An information collection device corresponding to each of the multiple environmental parameters; the information collection device is used to collect corresponding actual parameter information for each of the environmental parameters in the environment where the air purifier is located;

[0031] A control device, used for obtaining each of the actual parameter information and the user's expected parameter information for the target parameter in each of the environmental parameters; performing correlation analysis on each of the environmental parameters based on the actual parameter information to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each of the environmental parameters and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment; determining the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship.

[0032] In a third aspect, the present application further provides an air purifier control device, the device comprising:

[0033] A perception module, used to obtain actual parameter information collected for multiple environmental parameters and user's expected parameter information for target parameters in each of the environmental parameters in the environment where the air purifier is located;

[0034] A first intelligent module is used to perform correlation analysis on each of the environmental parameters based on each of the actual parameter information to determine an environmental characteristic relationship of the environment; the environmental characteristic relationship includes a correlation relationship between each of the environmental parameters and / or a correlation relationship between parameter information corresponding to different spatial positions of the same environmental parameter in the environment;

[0035] The second intelligent module is used to determine the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship.

[0036] In a fourth aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above method are implemented.

[0037] In a fifth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and the computer program implements the steps of the above method when executed by a processor.

[0038] In a sixth aspect, the present application further provides a computer program product, wherein the computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0039] The above-mentioned air purifier control method, device, computer equipment, storage medium and computer program product can obtain actual parameter information collected for multiple environmental parameters respectively, and the user's expected parameter information for the target parameters in each environmental parameter in the environment where the air purifier is located, so as to determine the actual environmental conditions of the current environment and the user's preferences and needs. Then, based on the actual parameter information, each environmental parameter is correlated and analyzed to determine the environmental characteristic relationship of the environment. It can comprehensively and accurately determine the correlation relationship between the environmental parameters in the current environment, and / or the correlation relationship between the corresponding parameter information of the same environmental parameter in different spatial positions in the environment. Then, according to the expected parameter information of the target parameter and the environmental characteristic relationship, the control mode of the air purifier is determined. On the premise of meeting the diversified needs of users, the corresponding control mode of the air purifier can be determined in combination with the needs of the user, thereby improving the diversity of the control mode. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 An application environment diagram of an air purifier control method in an embodiment;

[0041] Figure 2 A schematic flow chart of an air purifier control method in one embodiment;

[0042] Figure 3 A schematic diagram of a flow chart of an environmental characteristic map determination step in one embodiment;

[0043] Figure 4 A schematic flow chart of a control mode determination step in one embodiment;

[0044] Figure 5A schematic flow chart of an environmental characteristic map determination step in another embodiment;

[0045] Figure 6 A schematic diagram of a flow chart of an update status determination step in one embodiment;

[0046] Figure 7 A schematic flow chart of a control mode determination step in yet another embodiment;

[0047] Figure 8 A flowchart of a neural network model training step in one embodiment;

[0048] Fig. 9 A schematic diagram of a process of training a neural network model in another embodiment;

[0049] Fig.10 A logic diagram of a smart module control process in one embodiment;

[0050] Fig.11 A module connection diagram of an air purifier control system in one embodiment;

[0051] Fig.12 A structural block diagram of an intelligent module in an air purifier control system in an embodiment;

[0052] Fig.13 A decision diagram of a neural network model in an air purifier control system in one embodiment;

[0053] Fig.14 A product usage flow chart of an air purifier control system in one embodiment;

[0054] Fig.15 A schematic flow chart of an air purifier control method in another embodiment;

[0055] Fig.16 is a structural block diagram of an air purifier control device in one embodiment;

[0056] Fig.17 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0058] The air purifier control method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the control device 102 communicates with the information collection device 104 corresponding to multiple environmental parameters through the network. Among them, the control device 102 refers to a device used to monitor and adjust a specific object or system, and implements the control strategy by receiving and processing signals, outputs control signals, and monitors and protects the controlled object in real time, thereby realizing precise control of the entire system. The information collection device 104 is a tool used to collect different types of data, information and signals, also known as a data acquisition device, a signal acquisition device, a sensor or a data recorder, etc., which can collect various data information, signals and events in real time and transmit them to a computer, a controller, a storage device or other processing device for processing and analysis. For example, the information collection device 104 can be a temperature sensor, a humidity sensor, an infrared sensor, etc., and when working, it will automatically identify the surrounding environment, monitor the temperature, humidity, biological position and other different information in the environment, and use this as a basis for decision-making. Specifically, during the process of controlling the air purifier, the control device 102 obtains actual parameter information collected for multiple environmental parameters from the information collection device 104 in the environment where the air purifier is located, and obtains the user's expected parameter information for the target parameters in each environmental parameter; based on the actual parameter information, correlation analysis is performed on each environmental parameter to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between the environmental parameters, and / or the correlation relationship between the corresponding parameter information of the same environmental parameter in different spatial positions in the environment; the control mode of the air purifier is determined according to the expected parameter information of the target parameter and the environmental characteristic relationship.

[0059] In one embodiment, Figure 2 As shown, a method for controlling an air purifier is provided, and the method is applied to Figure 1 The control device in the example is used to illustrate, including the following steps:

[0060] Step S202, in the environment where the air purifier is located, actual parameter information collected for multiple environmental parameters and the user's expected parameter information for the target parameter in each environmental parameter are obtained.

[0061] Among them, an air purifier is an electrical appliance that purifies the air by filtering, electric purification or a combination of the two. The basic principle is to drive the fan blades through the motor, thereby driving the airflow through the filter or electric purification module, adsorbing the pollutants, particles, and harmful bacteria and viruses, and then filtering the air. Now the development of air purifiers is basically mature, and its performance and technology have been greatly improved compared to the initial ones, including air purification volume, catalytic formaldehyde removal technology, etc. Environmental parameters refer to various physical, chemical and biological indicators that can reflect environmental conditions. These parameters can be classified according to different environmental factors, such as temperature environmental parameters, humidity environmental parameters, wind speed environmental parameters, etc. Actual parameter information refers to the parameter information actually collected for each environmental parameter, which can be used to characterize the actual environmental conditions of the environment in which the air purifier is located. The target parameter refers to the environmental parameter involved in the user's preference requirements, and the expected parameter information is the user's expected parameter value for the target parameter. For example, if the user's preference requirements are: like 25 degrees room temperature and 30% humidity, then the target parameters are temperature and humidity, and the expected parameter information is temperature 25 degrees and humidity 30%.

[0062] Specifically, in order to control the air purifier, it is necessary to first clarify the actual environmental conditions in the environment where the air purifier is located and the user's preferences, that is, it is necessary to obtain the actual parameter information collected for multiple environmental parameters in the environment where the air purifier is located, as well as the user's expected parameter information for the target parameters in each environmental parameter. In some specific embodiments, multiple types of information collection devices can collect corresponding actual parameter information for multiple environmental parameters respectively, and then the control device obtains the above-mentioned actual parameter information from the information collection device. Exemplarily, the step of obtaining the actual parameter information can be active acquisition or passive reception.

[0063] Step S204: performing correlation analysis on each environmental parameter based on each actual parameter information to determine the environmental characteristic relationship of the environment.

[0064] The environmental characteristic relationship includes the correlation between various environmental parameters, and / or the correlation between the corresponding parameter information of the same environmental parameter at different spatial locations in the environment. For example, the correlation between various environmental parameters may refer to the correlation between temperature and humidity, the correlation between temperature and biological position, etc. The correlation between the corresponding parameter information of the same environmental parameter at different spatial locations in the environment may refer to the correlation between temperature values ​​at different spatial locations.

[0065] Specifically, after obtaining the actual parameter information collected for multiple environmental parameters, the environmental parameters can be correlated and analyzed based on the actual parameter information to determine the environmental characteristic relationship in the environment. In some specific embodiments, the control device can integrate the parameter information of each environmental parameter in the spatial position for each spatial position to obtain a multidimensional information group with spatial position and environmental parameters as dimensional features. From the multidimensional information group, multiple multidimensional sub-information groups that match the convolution size of the convolution kernel are determined. For each multidimensional sub-information group, a dot product operation is performed on the multidimensional sub-information group based on the convolution kernel to obtain the output value corresponding to the multidimensional sub-information group. Based on the output values ​​of each multidimensional sub-information group, an environmental characteristic map of the environment is obtained, which is used to characterize the environmental characteristic relationship of the environment. In other specific embodiments, the control device can also directly perform correlation analysis on each environmental parameter based on the actual parameter information to determine the environmental characteristic relationship in the environment. For example, when the actual parameter information of temperature is 25 degrees and the humidity is 30%, if the temperature is relatively high and the humidity is relatively high, a correlation analysis can be performed on the temperature and humidity based on these actual parameter information to determine that the correlation between the temperature and humidity is weak; if the temperature is relatively low and the humidity is relatively high, a correlation analysis can be performed on the temperature and humidity based on these actual parameter information to determine that the correlation between the temperature and humidity is strong.

[0066] Step S206, determining the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship.

[0067] The control mode of the air purifier refers to the set of operating parameters of each component of the air purifier during operation, which is designed to meet the air purification needs in different scenarios. The components may include fans, humidifiers, heaters, etc. The operating parameters may include the wind speed of the fan, the humidification power of the humidifier, the heating power of the heater, etc.

[0068] Specifically, after determining the environmental characteristic relationship in the environment, in order to meet the needs of the user, it is also necessary to combine the expected parameter information of the target parameter and the environmental characteristic relationship to finally determine the control mode of the air purifier. Optionally, the control device can determine the associated parameters that have an associated relationship with the target parameter from the environmental characteristic relationship, and determine the operating parameters of the components corresponding to each associated parameter according to the expected parameter information of the target parameter, and then determine the control mode adapted to each operating parameter. Exemplarily, when the target parameter is temperature, the associated parameters may include temperature and humidity, and the operating parameters may include the humidification power of the humidifier and the heating power of the heater.

[0069] In some specific embodiments, the environmental characteristic relationship can be represented by an environmental characteristic graph. Therefore, the environmental characteristic graph can be updated and adjusted according to the expected parameter information of the target parameter, and the control mode of the air purifier can be determined based on the obtained updated characteristic graph.

[0070] The above-mentioned air purifier control method obtains actual parameter information collected for multiple environmental parameters respectively, and the user's expected parameter information for the target parameters in each environmental parameter in the environment where the air purifier is located, so as to determine the actual environmental conditions of the current environment and the user's preferences and needs. Then, based on the actual parameter information, each environmental parameter is correlated and analyzed to determine the environmental characteristic relationship of the environment. It can comprehensively and accurately determine the correlation between the environmental parameters in the current environment, and / or the correlation between the corresponding parameter information of the same environmental parameter in different spatial positions in the environment. Then, according to the expected parameter information of the target parameter and the environmental characteristic relationship, the control mode of the air purifier is determined. On the premise of meeting the diversified needs of users, the corresponding control mode of the air purifier can be determined in combination with the needs of the user, thereby improving the diversity of the control mode.

[0071] In one embodiment, each environmental parameter corresponds to multiple parameter information at multiple spatial locations. In the case of this embodiment, Figure 3 As shown, based on the actual parameter information, each environmental parameter is correlated and analyzed to determine the environmental characteristic relationship of the environment, including:

[0072] Step S302: for each spatial position, the parameter information of each environmental parameter in the spatial position is integrated to obtain a multi-dimensional information group with spatial position and environmental parameter as dimensional features.

[0073] Among them, spatial position is a term that describes the relative position of an object or point in three-dimensional or two-dimensional space. It can be determined by various coordinate systems, which can be Cartesian coordinates, polar coordinates, or other more complex systems. Multidimensional information group refers to the combination of information in multiple dimensions.

[0074] Specifically, it is necessary to clarify multiple spatial locations in the environment and integrate the sorted environmental parameter information with the corresponding spatial location information. This usually involves associating or matching the environmental parameter data with the spatial location data in the database, and the integrated information forms a multidimensional information group, in which each spatial location has a corresponding environmental parameter data set. The multidimensional information group is characterized by spatial location and environmental parameters. Each dimension can be regarded as a variable or attribute used to describe the spatial location and environmental state. Among them, these locations can be specific geographic coordinates (such as longitude and latitude) or relative positions (such as different rooms in a building). The determination of spatial location can be achieved through technical means such as geographic information systems and global positioning systems.

[0075] Step S304: determine, from the multidimensional information group, a plurality of multidimensional sub-information groups that match the convolution size of the convolution kernel.

[0076] Step S306: for each multidimensional sub-information group, perform a dot product operation on the multidimensional sub-information group based on a convolution kernel to obtain an output value corresponding to the multidimensional sub-information group.

[0077] Step S308, obtaining an environmental characteristic map of the environment based on the output values ​​of each multi-dimensional sub-information group.

[0078] Among them, the convolution kernel, also known as the filter, is a coefficient matrix used for feature extraction during image processing. In a convolutional neural network, given an input image, the convolution kernel will slide on the input image and perform a convolution operation with a small area in the image, that is, a weighted sum of pixels at each position is performed to obtain each corresponding pixel in the output image. The convolution size of the convolution kernel is an important parameter in the convolution operation, which defines the size of the convolution kernel. The environmental feature map is used to characterize the environmental feature relationship of the environment.

[0079] Specifically, in order to facilitate subsequent convolution sliding, we can first determine multiple multidimensional sub-information groups that match the convolution size of the convolution kernel from the multidimensional information group, and perform dot product operations on the convolution kernel and each multidimensional sub-information group to obtain the output values ​​corresponding to each multidimensional sub-information group. By combining the output values ​​of each multidimensional sub-information group, we can obtain the environmental characteristic map of the environment.

[0080] In this embodiment, each parameter information is first integrated to obtain a multidimensional information group, and then the multidimensional information group is operated through a convolution kernel, and finally an environmental characteristic map that can characterize the environmental characteristic relationship of the environment is obtained, which can improve the accuracy of determining the environmental characteristic relationship.

[0081] In one embodiment, Figure 4 As shown, according to the expected parameter information of the target parameter and the environmental characteristic relationship, the control mode of the air purifier is determined, including:

[0082] Step S402: Taking the environment feature map and the expected parameter information of the target parameter as inputs of the forgetting function, a weight matrix having the same dimension as the environment feature map is obtained.

[0083] Step S404, updating the environment feature map based on the weight matrix to obtain an updated feature map of the environment.

[0084] Among them, the forgetting function determines which features should be retained and which should be forgotten based on the user's expected information.

[0085] Specifically, the forgetting function can be a fully connected layer, whose input is the user's expected information and the output feature map of the current convolutional layer, and whose output is a weight matrix with the same size dimension as the feature map. The weight matrix is ​​multiplied by the feature map at the element level to achieve selective forgetting of the features and obtain an updated feature map of the environment.

[0086] Step S406, determining the control mode of the air purifier based on the updated characteristic map.

[0087] Specifically, after the updated characteristic graph is determined, since the updated characteristic graph has combined the user's expectations and the actual parameter information of the environment, the control mode of the air purifier can be determined according to the updated characteristic graph.

[0088] In this embodiment, the expected parameter information and the environmental characteristic map are input into the forgetting function to obtain a weight matrix for updating the environmental characteristic map. Finally, the control mode of the air purifier is determined according to the updated characteristic map, which effectively reduces the user's operation complexity.

[0089] In one embodiment, the update characteristic graph includes a plurality of update characteristic graphs; each update characteristic graph corresponds to a parameter collection moment. In the case of this embodiment, Figure 5 As shown, based on the updated characteristic map, the control mode of the air purifier is determined, including:

[0090] Step S502: for each updated feature map, flatten the updated feature map to obtain a feature vector corresponding to the updated feature map.

[0091] Specifically, the update feature map is a two-dimensional or three-dimensional array output by the convolution layer, which contains the features extracted from the input data after the convolution operation. As the network goes deeper, the update feature map is usually further processed, such as reducing the size through the pooling layer, or classifying through the fully connected layer. The update feature map is converted into a one-dimensional feature vector for subsequent processing.

[0092] Step S504, analyzing each feature vector based on the long short-term memory network model in turn according to the time sequence between the parameter collection moments, and determining the final update state of the memory cell of the feature vector in the long short-term memory network model.

[0093] Step S506, determining the control mode corresponding to the air purifier according to the final update state.

[0094] Among them, the long short-term memory network model is a special recurrent neural network (RNN) architecture that aims to solve the long-term dependency problem encountered by the standard RNN when processing long sequence data.

[0095] Specifically, in the Long Short-Term Memory (LSTM) network model, the state of the memory cell is updated at each time step based on the input feature vector and the previous cell state. The final updated state is obtained after processing the entire feature vector sequence. Therefore, the control device needs to analyze each feature vector based on the Long Short-Term Memory network model in the time sequence between the parameter acquisition moments, determine the final updated state of the feature vector in the memory cell of the Long Short-Term Memory network model, and determine the corresponding control mode of the air purifier based on the final updated state.

[0096] In this embodiment, each feature vector is analyzed once based on the long short-term memory network model, and finally the control mode corresponding to the air purifier is determined according to the final update state of the memory cell, which can ensure the accuracy of the control mode determination.

[0097] In one embodiment, Figure 6 As shown, the air purifier control method also includes:

[0098] Step S602, based on the input gate of the long short-term memory network model, determine the new information corresponding to the feature vector in the memory cell.

[0099] The function of the input gate is to determine what new information should be added to the memory cell in the current time step. The new information refers to the information that the memory cell needs to include during the update process.

[0100] Specifically, first, the input gate receives the input feature vector x_t of the current time step and the previous hidden state h_{t-1} through a sigmoid function. The output of the sigmoid function is a value between 0 and 1. For each unit of the memory cell, this value indicates the degree to which the new information of the unit is allowed to pass. If the sigmoid output is close to 1, most of the new information will be retained; if it is close to 0, most of the new information will be discarded. At the same time, another tanh layer also receives x_t and h_{t-1} as input and generates a new candidate value vector C_tilde_t. This vector contains the new information that may be added to the memory cell.

[0101] Step S604, based on the forget gate of the long short-term memory network model, determine the discard information corresponding to the feature vector in the memory cell.

[0102] Among them, the role of the forget gate is to decide which old information in the memory cell should be discarded in the current time step. Discarded information refers to the information that needs to be excluded during the update process of the memory cell.

[0103] Specifically, the forget gate also receives x_t and h_{t-1} as input through a sigmoid function and outputs a value between 0 and 1. For each unit of the memory cell, this value indicates the extent to which the old information of the unit is allowed to be retained. If the sigmoid output is close to 0, most of the old information will be discarded; if it is close to 1, most of the old information will be retained.

[0104] Step S606: based on the newly added information and the discarded information, the state of the memory cell is updated to obtain an updated state.

[0105] Specifically, first, the output of the forget gate is multiplied by the current state of the memory cell C_{t-1}. This multiplication operation actually applies the output value of the forget gate element by element to determine which old information is discarded. Then, the output of the input gate (which determines the extent to which new information is allowed to pass) is multiplied by the new candidate value vector C_tilde_t. This multiplication operation also applies the output value of the input gate element by element to determine which new information is added to the memory cell. Finally, the results of the above two steps are added to obtain the new state of the memory cell C_t. This new state contains both the filtered old information and the newly added useful information.

[0106] In this embodiment, the information corresponding to the feature vector in the memory cell is determined according to the input gate and the forgetting gate of the long short-term memory network model, and then the state of the memory cell is updated to obtain the updated state, which can ensure the accuracy of the updated state determination and further determine the accuracy of the control mode determination.

[0107] In one embodiment, Figure 7 As shown, according to the expected parameter information of the target parameter and the environmental characteristic relationship, the control mode of the air purifier is determined, including:

[0108] Step S702: Determine the expected parameter information of each environmental parameter associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter.

[0109] Step S704: Determine the control mode of the air purifier based on the expected parameter information.

[0110] The expected parameter information refers to the expected parameter information required to achieve the target parameter, and is the expected parameter information of each environmental parameter that is associated with the target parameter.

[0111] Specifically, after determining the expected parameter information and actual parameter information of the target parameter, since the environmental characteristic relationship can be used to characterize the correlation between the environmental parameters, the environmental parameters associated with the target parameter can be determined first, and then the expected parameter information of each environmental parameter associated with the target parameter can be determined based on the difference between the expected parameter information and the actual parameter information of the target parameter. After that, the control mode of the air purifier is determined based on each expected parameter information. For example, the target parameter is temperature, the expected parameter information is 25 degrees, and the environmental parameter associated with the temperature is determined to be humidity. Then, assuming that the expected parameter information of the humidity is 40%, the temperature can reach the expected value, then the control mode corresponding to the air purifier can be determined based on the expected parameter information of the humidity being 40%.

[0112] In this embodiment, according to the difference between the expected parameter information and the actual parameter information of the target parameter, the expected parameter information of each environmental parameter that is associated with the target parameter is determined, and based on the expected parameter information, the control mode of the air purifier is determined, which can improve the accuracy of the control mode determination.

[0113] In one embodiment, an air purifier control system is also provided, the system comprising:

[0114] An information collection device corresponding to each of the multiple environmental parameters; the information collection device is used to collect corresponding actual parameter information for each environmental parameter in the environment where the air purifier is located;

[0115] A control device is used to obtain actual parameter information and user's expected parameter information for target parameters in each environmental parameter; perform correlation analysis on each environmental parameter based on the actual parameter information to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between the environmental parameters and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment; determine the control mode of the air purifier based on the expected parameter information of the target parameter and the environmental characteristic relationship.

[0116] In this embodiment, the air purifier control system is implemented in the same manner as the above-mentioned air purifier control method, which will not be repeated here.

[0117] In one embodiment, the process of this embodiment is as follows Figure 8 As shown, based on the background of big data and deep learning methods, an intelligent module is trained through the steps of constructing a data set, building and training a deep learning model, etc. This module can be used in an air purifier and make decisions.

[0118] In terms of constructing the data set, it is necessary to collect environmental information of the air purifier in actual operation or simulated actual operation, including using corresponding sensors to sense the temperature, light, airflow, biological position, humidity, and other information that can reflect and judge the environmental conditions. Then, a microcontroller is used to write a data acquisition program to collect a large amount of different environmental information and record the different decision labels corresponding to different environmental information. The overall process is as follows: Figure 8 shown.

[0119] This is repeated to obtain a large amount of data for model training, which is divided into three data sets: training set, test set and validation set, with a ratio of about 6:3:1. The training set is used to repeatedly train the model and continuously optimize the decision weight of the model. The role of the test set is to evaluate the degree of fit of the model at the end of each training and give feedback to judge the model training status and help the model correct the weight. The validation set will be used to verify the degree of fit of the trained model to make decisions in unfamiliar environments.

[0120] CNN (Convolutional Neural Network) can effectively extract local features of multi-variable inputs. For example, environmental data such as temperature, humidity, airflow, biological position and lighting can be regarded as multi-channel inputs, similar to the RGB channels of an image. CNN can capture the spatial relationship between these data through convolution operations. Through convolution operations, CNN can reduce the number of parameters and improve the computational efficiency and generalization ability of the model. LSTM (Long Short-Term Memory Network) can process time series data and capture time dependencies. This is very useful for processing changes in environmental data, because environmental parameters usually change over time. LSTM can effectively retain long-term dependencies through a gating mechanism, avoiding the gradient vanishing problem in traditional RNNs.

[0121] CNN has a strong ability to extract data features, while LSTM has memory and can process time series data, so a composite neural network of CNN and LSTM is constructed, and then the neural network is trained based on the pre-made data set. The training process logic is as follows Fig. 9 As the training progresses, the model weights will be continuously updated and the degree of fit will become higher and higher. When the degree of fit of the model no longer increases with continued training, the training is completed. The model is tested on the validation set, and the decision results given by the model are compared with the decision results in the validation set. If the results are basically consistent and the difference is within the allowable range, the training is completed and a mature model is obtained.

[0122] Among them, choosing CNN+LSTM is not the only option. In the subsequent development, there will be more and more neural networks suitable for this work. Here we only select a more suitable solution at present. If there is a better solution that can meet the requirements in the future, you can selectively replace it.

[0123] The model is stored in the built-in chip of the air purifier, and the model will be read from the built-in chip when it is used. During use, for different environments, independent decisions are made based on factors such as temperature, light, airflow, biological location, humidity, etc., and the operating status of the air purifier is adjusted, such as wind speed, silent mode, sleep mode, strong mode, etc. In addition, the humidification mode can be adjusted according to humidity, the warm air can be turned on according to temperature, and the warm air temperature and wind speed can be adjusted according to the distance of the organism, etc.

[0124] A working example, its overall operation logic is as follows Fig.10 As shown:

[0125] ① User personalized preset: If a user lives in the north, he likes a fresh environment with an air humidity of 60% and a room temperature of 25 degrees Celsius, and likes the wind blowing towards him, and uses this as a personalized preset.

[0126] ②Environmental conditions: The current ambient humidity is 40%, the indoor temperature is 22 degrees Celsius, and the user is located close to the air purifier.

[0127] ③ Real-time monitoring of the environment by intelligent modules: The intelligent modules are equipped with a variety of different sensors, such as temperature sensors, humidity sensors, infrared sensors, etc. When working, they will automatically identify the surrounding environment, monitor various different information in the environment such as temperature, humidity, biological location, etc., and use this as the basis for decision-making.

[0128] ④ Intelligent module decision-making: The intelligent module monitors the environment in real time, and based on a large amount of previous training, the intelligent module will analyze the environmental conditions and make specific decisions, output the decision basis and take effect in the execution module.

[0129] ⑤ After receiving the decision, the execution module executes the instructions to make the air purifier increase the air blowing in the direction of people and turn on the humidification mode. At the same time, it will generate heat to play a heating role and turn on the negative ion generator to make the air fresher.

[0130] To sum up, this embodiment applies the intelligent module of deep learning to the air purifier, which greatly improves the intelligence of the air purifier, improves the adaptability of the air purifier to different environments, improves the competitiveness of the product, can help users make decisions, solves the problem of customers' insufficient understanding of product applications, and improves customer loyalty. It can also further understand and learn customer needs during use, optimize and improve customer personalized settings, and provide better quality personalized services.

[0131] In a specific embodiment, in order to solve the problems that the current air purifier functions are relatively fixed and modular, users do not understand their own needs, product functions and product operations well, the air purifier is complex to operate, lacks intelligence, and lacks personalized services, this embodiment is based on big data conditions and deep learning technology to build an air purifier control system suitable for air purifiers based on deep learning, so that the air purifier can adapt to a variety of complex working environments, monitor the temperature, humidity, biology, airflow, light and other environmental information in real time, and make decisions based on a large number of trained neural networks to determine and execute the working mode of the air purifier, including adjusting the wind speed and direction, controlling humidification and heating, etc. The user only needs to make simple personalized settings, and the air purifier can intelligently serve the user, and record the user's usage mode and continue to train the neural network based on this in the intelligent cloud, so that the decision of the neural network is more in line with the user's needs, so that even if the user's air environment needs are vague and the air purifier control is unfamiliar, the air purifier can actively cater to the user's preferences and work in line with the user's own needs.

[0132] The air purifier control system includes a main controller, an intelligent module (including a first intelligent module and a second intelligent module), an intelligent cloud, a human-computer interaction system, a safe operation system, a sensing module, a data storage system and a driving system.

[0133] The main controller includes a data processing program for collecting data from the perception module and user operations and managing the data storage system; it also includes a communication module for data transmission with the smart cloud.

[0134] The sensing module includes a temperature sensor for sensing the ambient temperature, a humidity sensor for sensing the ambient humidity, a light sensor for sensing the ambient light intensity, a biological position sensor for sensing the position of people in the space, and an airflow sensor for sensing the dynamics of the ambient airflow.

[0135] The data storage system is equipped with a storage medium for storing data output by the main controller.

[0136] The driving system includes the end effector of the air purifier, such as the motor, humidification system, electric heating system, fan blades, etc.

[0137] The safe operation system includes the safety monitoring module and some driving functions in the main controller. It has the highest priority and responds first when the product is abnormal.

[0138] The human-computer interaction system is used to set preset temperature, wind speed, humidity and other personalized information, or directly control the air purifier. It is the medium of human-computer interaction.

[0139] The intelligent module includes a data processing function to process the data from the perception module; an intelligent decision-making function to identify the features of the data and make decisions to drive the air purifier to work.

[0140] The smart cloud includes a cloud database that stores neural network training data; cloud model backup will store the neural network weights of each air purifier; and cloud training will continue to train the neural network.

[0141] The main issues addressed by this solution include:

[0142] 1. Users have insufficient knowledge of air purifier products and are vague about their own needs.

[0143] 2. The functions of the air purifier itself are relatively modular and fixed, and it often requires complex operations to meet users' needs for different aspects of the air environment.

[0144] 3. The functions of air purifier products are standardized and cannot accurately meet user needs, and there is a lack of personalized services for users.

[0145] The main technical means adopted in this solution include:

[0146] 1. Big data background and deep learning technology.

[0147] 2. Powerful communication function, which can ensure long-term and effective data transmission between local and cloud.

[0148] 3. The perfect deployment of cloud systems can allow local neural networks to continue training in the cloud environment.

[0149] The beneficial effects are mainly reflected in:

[0150] 1. Improve the intelligence of air purifiers so that they can make autonomous decisions based on the environment and needs without the need for complex controls.

[0151] 2. It can provide personalized services to users, so that the air purifier can be used according to the user's own needs and the user only needs to provide a simple and general idea at the beginning.

[0152] 3. Users do not need to understand their needs and products very well, they only need to have a general idea, and the air purifier can serve users well, thus lowering the threshold for use.

[0153] 4. The air purifier can record and adjust personalized settings and neural network weights based on user usage, actively catering to user preferences and making it more convenient for users to use.

[0154] In this embodiment, the relationship between the modules is as follows Fig.11 As shown, intelligent control of the air purifier is achieved through the interaction between the main controller, intelligent module, intelligent cloud, human-computer interaction system, safe operation system, perception module, data storage system and drive system.

[0155] The overall block diagram of the air purifier control system is as follows Fig.12 As shown, the main controller includes a data processing program. When the user operates the air purifier, the user operation data and the environmental recognition data of the perception module at this time will be fed back to the data processing program of the main controller. The data processing program will perform certain preprocessing on these data, including marking the data time, etc., and then import the data into the data storage system and manage it; it also includes a communication module to ensure data transmission between the main controller and the smart cloud, including the transmission of environmental data, operation data and neural network data between the data processing program and the smart cloud, and the management of the cloud database and model backup by the main controller.

[0156] The perception module will capture different environmental data through different sensors, including temperature sensors for sensing the ambient temperature, humidity sensors for sensing the ambient humidity, light sensors for sensing the ambient light intensity, biological position sensors for sensing the position of people in the space, and airflow sensors for sensing the dynamics of ambient airflow. The perception module will provide real-time feedback to the safety operation system and the intelligent module, and will feed back the data at the current time point to the main controller when the user has control behavior on the air purifier.

[0157] The data storage system is equipped with a storage medium to store the user's operating information on the air purifier and the corresponding environmental information. After this information is marked by the data processing program of the main controller, it will be stably stored in the storage medium. The main controller can communicate with the cloud through the communication module and upload the data in the storage medium to build a cloud database.

[0158] The drive system includes the end effectors of the air purifier, such as the motor, humidification system, electric heating system, fan blades, etc. Specifically, the drive system generally receives instructions from the drive module, or receives instructions from the safety detection module when the air purifier is working abnormally, and takes corresponding actions, such as adjusting the motor speed to adjust the wind speed, controlling the humidification system to adjust the air humidity, controlling the electric heating system to adjust the air temperature, changing the air direction by adjusting the fan blade direction, etc.

[0159] The safe operation system includes the safety monitoring module and some driving functions in the main controller, and the safe operation system maintains the highest command level during the operation of the air purifier. The information of the sensing module and the driving system will be fed back to the safety monitoring module in real time. When the surrounding environment is abnormal or the drive fails, the decision of the main controller driving module and the intelligent module will fail, and the safe operation system will control the driving system and enter the safe mode until the abnormality is resolved and restarted. In the safe mode, the alarm will continue to sound, the driving system will stop humidification and electric heating, and the fan blades will be fully opened. The motor will run at the lowest speed. If the sensing module feedback is blocked by foreign objects, the motor will stop running.

[0160] The human-computer interaction system is used to set preset temperature, wind speed, humidity and other personalized information in order to provide execution standards for the intelligent module; it can also be used to directly control the operation of the air purifier and execute the control instructions through the operation module and the drive module.

[0161] The intelligent module includes a data processing function that normalizes the information fed back by the perception module, so that it can be better recognized by the neural network; it also includes an intelligent decision-making function provided by a large number of trained neural networks. The neural network can efficiently identify the feedback information of the perception module after processing, and quickly identify its characteristics. After understanding the environmental information, it will make decisions based on the user's preset or personalized preferences, thereby driving the air purifier to work. Among them, the intelligent module includes a first intelligent module and a second intelligent module. The first intelligent module is specifically used to perform correlation analysis on each environmental parameter based on each actual parameter information to determine the environmental characteristic relationship of the environment. The second intelligent module is specifically used to determine the control mode of the air purifier based on the expected parameter information of the target parameter and the environmental characteristic relationship.

[0162] The smart cloud can build a rich software system, the most basic of which include cloud database, cloud model backup and cloud training. In addition, it also provides a neural network training environment. The cloud database is used to store the training data of the neural network. The main controller data processing program calls the data storage system to send local data to the cloud and store it in the cloud database for subsequent calls by the cloud; the cloud model backup will store the neural network weights of the air purifier when it leaves the factory and the neural network weights of each update, and users can overwrite the local neural network version by themselves; cloud training will continue to train the neural network based on the existing neural network, cloud database, cloud framework and cloud deep learning environment. After training, the neural network weights will be stored in the cloud model backup, which can be transmitted to the main controller through the communication module, and can overwrite the local neural network weights.

[0163] Neural network decision logic Fig.13 As shown in the figure, the personalized settings set by the user, that is, the user's preferences, will be used as the basis for the neural network decision. The neural network receives the temperature, humidity, airflow, biology, light and other information of the perception module in real time, and makes corresponding decisions based on the user's personalized settings. Figure 2 Block diagram of the intelligent system. The decision information of the neural network is ultimately executed by the drive system.

[0164] The process of users using air purifiers Fig.14 As shown, combined Fig.12 From the working logic of the system: ① The products are mass-produced, and the attached intelligent systems are initialized uniformly. The weights of the neural networks are the same and they can make decisions based on environmental information and preset personalized requirements. ② After purchasing the air purifier product, the user will first set the personalized settings and preset their own needs for the air environment. The decision of the neural network is based on this. However, it should be pointed out that the user's settings at this time are relatively vague, and do not comprehensively consider the surrounding environment and various needs. It just sets a rough starting point for the neural network.

[0165] ③When the machine is turned on, the perception module continuously monitors the temperature, humidity, airflow, biology, light and other information of the surrounding environment, and continuously feeds back to the safe operation system and the intelligent module. The intelligent module normalizes the information and performs other preprocessing. The neural network receives the processed information and efficiently identifies and extracts features, and then issues instructions to the drive system to make corresponding actions according to these features and the personalized needs of users. It should be emphasized that the decision-making here is not a modular work. The perception module can feedback information of different indicators, which can reflect a variety of information such as temperature, humidity, spatial biological position in the air, etc., and make corresponding detailed decisions based on different detailed information. In the whole process, users basically do not need to do anything except the preset when purchasing the product at the beginning and the switch when using it. For example, if the air purifier is placed by the window, the window is open, and there is wind blowing in, the air purifier will not open or close all the blades at the same time, but only close the blades facing the window. For another example, if the air purifier is placed next to a person, the electric heating system is working at this time. If the warm air blows on the person, the person will feel stuffy and dry. The decision-making system will turn off the warm air in the direction of the person and change it to normal temperature wind, while continuing to blow warm air normally in other places.

[0166] ④ But generally speaking, when users buy air purifiers, their understanding of product functions is limited to the instructions, and they have no experience in using them. In addition, it is usually difficult to plan the air environment in a short period of time with the human brain alone. Therefore, it can be basically concluded that the initial personalized settings are very vague, and at most they can only stay at settings such as preferring a humid environment or preferring wind blowing on people. Therefore, in the subsequent use process, users will often manually adjust the various functions of the air purifier to make its working effect meet their expectations.

[0167] ⑤ The user's operation of the air purifier will be recorded. At the same time, every time the user operates the air purifier, the perception module will collect the current environmental information and feed it back to the main controller. After preliminary data processing in the main controller, it will be input and stored in the data storage system. These operation data and environmental data correspond one by one and the data processing program will mark the time.

[0168] ⑥ The main controller manages the data of the data storage system in real time. If it detects that there are more data flows in a certain period of time, it means that the user controls frequently, which means that there is a gap between the system preset and the actual needs of the user. The main controller is equipped with a communication module, which can stably transmit data with the intelligent cloud. When connected to the network, the main controller will send the data of the data storage system to the cloud database. The cloud system is equipped with a deep learning environment and a neural network framework. It can be based on the existing neural network and use the data in the cloud database for training, so as to obtain new neural network weights that can better meet user demand decisions.

[0169] ⑦ Users can choose whether to overwrite the neural network weights trained in the cloud to the local area. If overwritten, the decision-making system of the air purifier will make decisions from an angle that is more in line with user needs. In this way, the air purifier can serve users completely independently, and users do not need to perform too many operations. It also solves the problem of users' lack of understanding of the product and unclear about their own needs. Even if users do not know their detailed needs, the intelligent system can provide users with high-quality services.

[0170] In summary, the air purifier control system enables users to use the air purifier well even if they do not understand the product and their own needs are unclear. The air purifier itself will make more detailed decisions to serve users, and users do not need to perform too many complicated operations. More importantly, although the product is produced in a standardized manner, the functions implemented can meet the personalized needs of users. Moreover, as the number of uses increases, the logic of the air purifier's personalized service will continue to improve, making a standardized product a personalized tool tailored for each user.

[0171] In a specific embodiment, Fig.15 As shown, a method for controlling an air purifier is also provided, comprising:

[0172] Step S1501, in the environment where the air purifier is located, obtaining actual parameter information collected for multiple environmental parameters respectively, and the user's expected parameter information for the target parameter in each environmental parameter;

[0173] Wherein, each environmental parameter corresponds to multiple parameter information at multiple spatial locations;

[0174] Step S1502, for each spatial position, integrating the parameter information of each environmental parameter in the spatial position to obtain a multi-dimensional information group with spatial position and environmental parameter as dimensional features;

[0175] Step S1503, determining a plurality of multidimensional sub-information groups matching the convolution size of the convolution kernel from the multidimensional information group;

[0176] Step S1504, for each multidimensional sub-information group, performing a dot product operation on the multidimensional sub-information group based on a convolution kernel to obtain an output value corresponding to the multidimensional sub-information group;

[0177] Step S1505, based on the output values ​​of each multi-dimensional sub-information group, an environmental characteristic map of the environment is obtained; the environmental characteristic map is used to characterize the environmental characteristic relationship of the environment;

[0178] The environmental characteristic relationship includes the correlation between environmental parameters and / or the correlation between the corresponding parameter information of the same environmental parameter at different spatial locations in the environment;

[0179] Step S1506, using the environment feature map and the expected parameter information of the target parameter as inputs of the forgetting function to obtain a weight matrix with the same dimension as the environment feature map;

[0180] Step S1507, updating the environment feature map based on the weight matrix to obtain an updated feature map of the environment, and flattening each updated feature map to obtain a feature vector corresponding to the updated feature map;

[0181] Wherein, the updating characteristic graph includes a plurality of updating characteristic graphs; each updating characteristic graph corresponds to a parameter collection time;

[0182] Step S1508, determining the newly added information corresponding to the feature vector in the memory cell based on the input gate of the long short-term memory network model in sequence according to the time sequence between the acquisition moments of each parameter;

[0183] Step S1509, based on the forget gate of the long short-term memory network model, determining the discarded information corresponding to the feature vector in the memory cell;

[0184] Step S1510, based on the newly added information and the discarded information, the state of the memory cell is updated to obtain a final updated state, and the control mode corresponding to the air purifier is determined according to the final updated state;

[0185] Step S1511, determining the expected parameter information of each environmental parameter associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter;

[0186] Step S1512, determining the control mode of the air purifier based on the expected parameter information.

[0187] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0188] Based on the same inventive concept, the embodiment of the present application also provides an air purifier control device for implementing the air purifier control method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more air purifier control device embodiments provided below can refer to the limitations of the air purifier control method above, and will not be repeated here.

[0189] In one embodiment, Fig.16 As shown, an air purifier control device 1600 is provided, including: a sensing module 1602, a first intelligent module 1604 and a second intelligent module 1606, wherein:

[0190] The perception module 1602 is used to obtain actual parameter information collected for multiple environmental parameters and the user's expected parameter information for the target parameter in each environmental parameter in the environment where the air purifier is located;

[0191] The first intelligent module 1604 is used to perform correlation analysis on each environmental parameter based on each actual parameter information to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each environmental parameter and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment;

[0192] The second intelligent module 1606 is used to determine the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship.

[0193] In one embodiment, each environmental parameter corresponds to multiple parameter information at multiple spatial locations. In this embodiment, the first intelligent module 1604 is specifically used to:

[0194] For each spatial position, the parameter information of each environmental parameter in the spatial position is integrated to obtain a multi-dimensional information group with spatial position and environmental parameters as dimensional features;

[0195] Determine, from the multidimensional information group, a plurality of multidimensional sub-information groups matching the convolution size of the convolution kernel;

[0196] For each multidimensional sub-information group, a dot product operation is performed on the multidimensional sub-information group based on a convolution kernel to obtain an output value corresponding to the multidimensional sub-information group;

[0197] Based on the output values ​​of each multidimensional sub-information group, an environmental characteristic map of the environment is obtained; the environmental characteristic map is used to characterize the environmental characteristic relationship of the environment.

[0198] In one embodiment, the second smart module 1606 includes:

[0199] A weight matrix determination unit, used to use the expected parameter information of the environment feature map and the target parameter as inputs of the forgetting function to obtain a weight matrix with the same dimension as the environment feature map;

[0200] An updated feature map determining unit, used to update the environment feature map based on the weight matrix to obtain an updated feature map of the environment;

[0201] The control mode determination unit is used to determine the control mode of the air purifier based on the updated characteristic map.

[0202] In one embodiment, the updating characteristic graph includes a plurality of updating characteristic graphs; each updating characteristic graph corresponds to a parameter collection moment. In the case of this embodiment, the control mode determination unit is specifically used to:

[0203] For each updated feature map, flatten the updated feature map to obtain a feature vector corresponding to the updated feature map;

[0204] According to the time sequence between the acquisition moments of each parameter, each feature vector is analyzed in turn based on the long short-term memory network model to determine the final update state of the memory cell of the feature vector in the long short-term memory network model;

[0205] According to the final update status, the corresponding control mode of the air purifier is determined.

[0206] In one embodiment, the air purifier control device 1600 further includes an update status determination module, which is specifically used to:

[0207] Based on the input gate of the long short-term memory network model, the newly added information corresponding to the feature vector in the memory cell is determined;

[0208] Based on the forget gate of the long short-term memory network model, the discarded information corresponding to the feature vector in the memory cell is determined;

[0209] Based on the newly added information and the discarded information, the state of the memory cell is updated to obtain an updated state.

[0210] In one embodiment, the second intelligent module 1606 is further configured to:

[0211] Determine the expected parameter information of each environmental parameter associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter;

[0212] Based on the expected parameter information, the control mode of the air purifier is determined.

[0213] Each module in the above-mentioned air purifier control device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0214] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.17 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, an air purifier control method is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0215] Those skilled in the art will understand that Fig.17 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0216] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0217] In the environment where the air purifier is located, actual parameter information collected for multiple environmental parameters and expected parameter information of the user for the target parameter in each environmental parameter are obtained;

[0218] Performing correlation analysis on each environmental parameter based on each actual parameter information to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each environmental parameter and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment;

[0219] According to the expected parameter information of the target parameter and the environmental characteristic relationship, the control mode of the air purifier is determined. In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0220] In one embodiment, each environmental parameter corresponds to multiple parameter information at multiple spatial locations. In the case of this embodiment, the processor also implements the following steps when executing the computer program:

[0221] For each spatial position, the parameter information of each environmental parameter in the spatial position is integrated to obtain a multi-dimensional information group with spatial position and environmental parameters as dimensional features;

[0222] Determine, from the multidimensional information group, a plurality of multidimensional sub-information groups matching the convolution size of the convolution kernel;

[0223] For each multidimensional sub-information group, a dot product operation is performed on the multidimensional sub-information group based on a convolution kernel to obtain an output value corresponding to the multidimensional sub-information group;

[0224] Based on the output values ​​of each multidimensional sub-information group, an environmental characteristic map of the environment is obtained; the environmental characteristic map is used to characterize the environmental characteristic relationship of the environment.

[0225] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0226] The expected parameter information of the environment feature map and the target parameter is used as the input of the forgetting function to obtain a weight matrix with the same dimension as the environment feature map;

[0227] The environment feature map is updated based on the weight matrix to obtain an updated feature map of the environment;

[0228] Based on the updated characteristic map, a control mode of the air purifier is determined.

[0229] In one embodiment, the updating characteristic graph includes a plurality of updating characteristic graphs; each updating characteristic graph corresponds to a parameter collection moment. In the case of this embodiment, when the processor executes the computer program, the following steps are also implemented:

[0230] For each updated feature map, flatten the updated feature map to obtain a feature vector corresponding to the updated feature map;

[0231] According to the time sequence between the acquisition moments of each parameter, each feature vector is analyzed in turn based on the long short-term memory network model to determine the final update state of the memory cell of the feature vector in the long short-term memory network model;

[0232] According to the final update status, the corresponding control mode of the air purifier is determined.

[0233] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0234] Based on the input gate of the long short-term memory network model, the newly added information corresponding to the feature vector in the memory cell is determined;

[0235] Based on the forget gate of the long short-term memory network model, the discarded information corresponding to the feature vector in the memory cell is determined;

[0236] Based on the newly added information and the discarded information, the state of the memory cell is updated to obtain an updated state.

[0237] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0238] Determine the expected parameter information of each environmental parameter associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter;

[0239] Based on the expected parameter information, the control mode of the air purifier is determined.

[0240] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0241] In the environment where the air purifier is located, actual parameter information collected for multiple environmental parameters and expected parameter information of the user for the target parameter in each environmental parameter are obtained;

[0242] Performing correlation analysis on each environmental parameter based on each actual parameter information to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each environmental parameter and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment;

[0243] The control mode of the air purifier is determined based on the expected parameter information of the target parameter and the environmental characteristic relationship.

[0244] In one embodiment, each environmental parameter corresponds to multiple parameter information at multiple spatial locations. In the case of this embodiment, when the computer program is executed by the processor, the following steps are also implemented:

[0245] For each spatial position, the parameter information of each environmental parameter in the spatial position is integrated to obtain a multi-dimensional information group with spatial position and environmental parameters as dimensional features;

[0246] Determine, from the multidimensional information group, a plurality of multidimensional sub-information groups matching the convolution size of the convolution kernel;

[0247] For each multidimensional sub-information group, a dot product operation is performed on the multidimensional sub-information group based on a convolution kernel to obtain an output value corresponding to the multidimensional sub-information group;

[0248] Based on the output values ​​of each multidimensional sub-information group, an environmental characteristic map of the environment is obtained; the environmental characteristic map is used to characterize the environmental characteristic relationship of the environment.

[0249] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0250] The expected parameter information of the environment feature map and the target parameter is used as the input of the forgetting function to obtain a weight matrix with the same dimension as the environment feature map;

[0251] The environment feature map is updated based on the weight matrix to obtain an updated feature map of the environment;

[0252] Based on the updated characteristic map, a control mode of the air purifier is determined.

[0253] In one embodiment, the updating characteristic graph includes a plurality of updating characteristic graphs; each updating characteristic graph corresponds to a parameter collection moment. In the case of this embodiment, when the computer program is executed by the processor, the following steps are also implemented:

[0254] For each updated feature map, flatten the updated feature map to obtain a feature vector corresponding to the updated feature map;

[0255] According to the time sequence between the acquisition moments of each parameter, each feature vector is analyzed in turn based on the long short-term memory network model to determine the final update state of the memory cell of the feature vector in the long short-term memory network model;

[0256] According to the final update status, the corresponding control mode of the air purifier is determined.

[0257] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0258] Based on the input gate of the long short-term memory network model, the newly added information corresponding to the feature vector in the memory cell is determined;

[0259] Based on the forget gate of the long short-term memory network model, the discarded information corresponding to the feature vector in the memory cell is determined;

[0260] Based on the newly added information and the discarded information, the state of the memory cell is updated to obtain an updated state.

[0261] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0262] Determine the expected parameter information of each environmental parameter associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter;

[0263] Based on the expected parameter information, the control mode of the air purifier is determined.

[0264] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0265] In the environment where the air purifier is located, actual parameter information collected for multiple environmental parameters and expected parameter information of the user for the target parameter in each environmental parameter are obtained;

[0266] Performing correlation analysis on each environmental parameter based on each actual parameter information to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each environmental parameter and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment;

[0267] The control mode of the air purifier is determined based on the expected parameter information of the target parameter and the environmental characteristic relationship.

[0268] In one embodiment, each environmental parameter corresponds to multiple parameter information at multiple spatial locations. In the case of this embodiment, when the computer program is executed by the processor, the following steps are also implemented:

[0269] For each spatial position, the parameter information of each environmental parameter in the spatial position is integrated to obtain a multi-dimensional information group with spatial position and environmental parameters as dimensional features;

[0270] Determine, from the multidimensional information group, a plurality of multidimensional sub-information groups matching the convolution size of the convolution kernel;

[0271] For each multidimensional sub-information group, a dot product operation is performed on the multidimensional sub-information group based on a convolution kernel to obtain an output value corresponding to the multidimensional sub-information group;

[0272] Based on the output values ​​of each multidimensional sub-information group, an environmental characteristic map of the environment is obtained; the environmental characteristic map is used to characterize the environmental characteristic relationship of the environment.

[0273] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0274] The expected parameter information of the environment feature map and the target parameter is used as the input of the forgetting function to obtain a weight matrix with the same dimension as the environment feature map;

[0275] The environment feature map is updated based on the weight matrix to obtain an updated feature map of the environment;

[0276] Based on the updated characteristic map, a control mode of the air purifier is determined.

[0277] In one embodiment, the updating characteristic graph includes a plurality of updating characteristic graphs; each updating characteristic graph corresponds to a parameter collection moment. In the case of this embodiment, when the computer program is executed by the processor, the following steps are also implemented:

[0278] For each updated feature map, flatten the updated feature map to obtain a feature vector corresponding to the updated feature map;

[0279] According to the time sequence between the acquisition moments of each parameter, each feature vector is analyzed in turn based on the long short-term memory network model to determine the final update state of the memory cell of the feature vector in the long short-term memory network model;

[0280] According to the final update status, the corresponding control mode of the air purifier is determined.

[0281] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0282] Based on the input gate of the long short-term memory network model, the newly added information corresponding to the feature vector in the memory cell is determined;

[0283] Based on the forget gate of the long short-term memory network model, the discarded information corresponding to the feature vector in the memory cell is determined;

[0284] Based on the newly added information and the discarded information, the state of the memory cell is updated to obtain an updated state.

[0285] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0286] Determine the expected parameter information of each environmental parameter associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter;

[0287] Based on the expected parameter information, the control mode of the air purifier is determined.

[0288] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0289] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0290] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0291] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An air purifier control method, characterized in that: The method comprises: In the environment where the air purifier is located, obtaining actual parameter information collected for multiple environmental parameters respectively, and the user's expected parameter information for the target parameter in each of the environmental parameters; Based on the actual parameter information, each environmental parameter is subjected to correlation analysis to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each environmental parameter and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment; The control mode of the air purifier is determined according to the expected parameter information of the target parameter and the environmental characteristic relationship.

2. The method according to claim 1, characterized in that Each of the environmental parameters corresponds to multiple parameter information at multiple spatial positions; the correlation analysis of each of the environmental parameters based on each of the actual parameter information to determine the environmental characteristic relationship of the environment includes: For each of the spatial positions, integrating the parameter information of each of the environmental parameters in the spatial position to obtain a multi-dimensional information group with the spatial position and the environmental parameters as dimensional features; Determine, from the multidimensional information group, a plurality of multidimensional sub-information groups matching the convolution size of the convolution kernel; For each of the multidimensional sub-information groups, performing a dot product operation on the multidimensional sub-information group based on the convolution kernel to obtain an output value corresponding to the multidimensional sub-information group; Based on the output values ​​of each of the multi-dimensional sub-information groups, an environmental characteristic map of the environment is obtained; the environmental characteristic map is used to characterize the environmental characteristic relationship of the environment.

3. The method according to claim 2, characterized in that The step of determining the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship includes: Taking the environment feature map and the expected parameter information of the target parameter as inputs of a forgetting function, and obtaining a weight matrix with the same dimension as the environment feature map; Updating the environment feature map based on the weight matrix to obtain an updated feature map of the environment; Based on the updated characteristic map, a control mode of the air purifier is determined.

4. The method according to claim 3, characterized in that: The update characteristic graph includes a plurality of graphs; each of the update characteristic graphs corresponds to a parameter collection moment; and the control mode of the air purifier is determined based on the update characteristic graph, including: For each of the updated feature maps, flatten the updated feature map to obtain a feature vector corresponding to the updated feature map; According to the time sequence between the parameter collection moments, each of the feature vectors is analyzed in turn based on the long short-term memory network model to determine the final update state of the memory cell of the feature vector in the long short-term memory network model; According to the final updated state, a control mode corresponding to the air purifier is determined.

5. The method according to claim 4, characterized in that The method further comprises: Determining, based on an input gate of the long short-term memory network model, newly added information corresponding to the feature vector in the memory cell; Determining, based on a forget gate of the long short-term memory network model, discarding information corresponding to the feature vector in the memory cell; Based on the newly added information and the discarded information, the state of the memory cell is updated to obtain an updated state.

6. The method according to claim 1, characterized in that The step of determining the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship includes: Determining the expected parameter information of each of the environmental parameters associated with the target parameter according to the difference between the expected parameter information and the actual parameter information of the target parameter; Based on each of the expected parameter information, a control mode of the air purifier is determined.

7. An air purifier control system, characterized in that: The system comprises: An information collection device corresponding to each of the plurality of environmental parameters; the information collection device is used to collect corresponding actual parameter information for each of the environmental parameters in the environment where the air purifier is located; A control device, used for obtaining each of the actual parameter information and the user's expected parameter information for the target parameter in each of the environmental parameters; performing correlation analysis on each of the environmental parameters based on the actual parameter information to determine the environmental characteristic relationship of the environment; the environmental characteristic relationship includes the correlation relationship between each of the environmental parameters and / or the correlation relationship between the corresponding parameter information of the same environmental parameter at different spatial positions in the environment; determining the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship.

8. An air purifier control device, characterized in that: The device comprises: A perception module, used to obtain actual parameter information collected for multiple environmental parameters and user's expected parameter information for target parameters in each of the environmental parameters in the environment where the air purifier is located; A first intelligent module is used to perform correlation analysis on each of the environmental parameters based on each of the actual parameter information to determine an environmental characteristic relationship of the environment; the environmental characteristic relationship includes a correlation relationship between each of the environmental parameters and / or a correlation relationship between parameter information corresponding to different spatial positions of the same environmental parameter in the environment; The second intelligent module is used to determine the control mode of the air purifier according to the expected parameter information of the target parameter and the environmental characteristic relationship.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images