Control instruction generation method and device, storage medium and electronic device

Abstract

The present application discloses a method and device for generating a control instruction, a storage medium and an electronic device, and relates to the field of smart home technology. The method for generating a control instruction comprises: identifying the fuzzy intention of a target object through an intention recognition model to obtain an intention result corresponding to the fuzzy intention, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device; determining a second adjustment parameter for controlling the control device based on the first operating status information of the control device, the historical operation data of the control device and the environmental information of the area where the control device is located; generating a control instruction for controlling the control device based on the intention result and the second adjustment parameter; adopting the above technical solution solves the problem in the prior art that the adjustment parameters of the device are set according to characteristics such as time and region, and cannot meet the different preferences of different users for the device.

Description

Technical Field

[0001] The present application relates to the field of smart home technology, and more specifically, to a method and device for generating control instructions, a storage medium, and an electronic device. Background Art

[0002] With the continuous development and popularization of smart homes, the number of smart homes in users' homes is increasing. Although smart homes can now be controlled using intelligent voice, in most cases users are required to specify the control device, such as "turn on the air conditioner in the bedroom" or "close the curtains"... For some vague user intentions, such as "a bit cold" or "too dark", the smart home cannot make corresponding controls because it cannot recognize the user's intention.

[0003] In the existing technology, a fixed model is usually used to complete the user's intelligent control of smart home devices. However, the use of such a fixed model has the following disadvantages: 1. Insufficient protection for user personalization: After the existing model selects a device, it directly controls the device for adjustment. The parameters of the device adjustment are set according to conditions such as time and geographic location. Therefore, most users, such as users in the same community, may have the same adjustment parameters. Therefore, there is a lack of personalization and it is impossible to meet the different preferences of different users for devices. 2. Insufficient number of features used: The features used by the existing model are basically external features (such as weather, wind direction, wind level, air pressure, sunrise time, sunset time, outdoor temperature, etc.), but device features (such as indoor temperature, voiceprint category for voice control, device status, etc.) are not taken into account. 3. Huge cost: If you try to build a model for each user to predict a single user's usage preference, although it is theoretically feasible, building and training a separate model for each user will consume a huge amount of data and computing power.

[0004] Therefore, the prior art does not propose a method that can predict a user's use of smart home devices based on the user's preferences.

[0005] In the existing technology, the adjustment parameters of the device are set according to characteristics such as time and region, which cannot meet the different preferences of different users for the device. No effective solution has been proposed yet. Summary of the Invention

[0006] Embodiments of the present invention provide a method and device for generating control instructions, a storage medium, and an electronic device to at least solve the problem in the prior art that the adjustment parameters of the device are set according to characteristics such as time and region, and cannot meet the different preferences of different users for the device.

[0007] According to one embodiment of the present invention, a method for generating a control instruction is provided, comprising: identifying the fuzzy intention of a target object through an intention recognition model to obtain an intention result corresponding to the fuzzy intention, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device; determining a second adjustment parameter for controlling the control device based on the first operating status information of the control device, the historical operation data of the control device and the environmental information of the area where the control device is located; and generating a control instruction for controlling the control device based on the intention result and the second adjustment parameter.

[0008] In an exemplary embodiment, the fuzzy intention of the target object is identified through an intention recognition model, including: obtaining the fuzzy intention input by the target object through an intelligent interactive device; obtaining second operating status information reported by all home appliances, wherein all home appliances include the control device; detecting environmental information of the area through all home appliances; inputting the second operating status information, the environmental information and the fuzzy intention into the intention recognition model to identify the fuzzy intention through the intention recognition model.

[0009] In an exemplary embodiment, the fuzzy intent of the target object is identified through an intention recognition model to obtain an intention result corresponding to the fuzzy intent, including: extracting preset keywords from the fuzzy intent, and determining the control device from all the home devices based on the preset keywords; determining the first operating status information of the control device from the second operating status information; and determining the adjustment direction indicated by the first adjustment parameter based on the first operating status information and the environmental information.

[0010] In an exemplary embodiment, before identifying the fuzzy intent of the target object through the intention recognition model, the method also includes: obtaining a voice instruction input by the target object, and performing intent recognition on the voice instruction through a preset algorithm; if the preset algorithm fails to identify the intention result corresponding to the voice instruction, determining the voice instruction as a fuzzy intent.

[0011] In an exemplary embodiment, a second adjustment parameter for controlling the control device is determined based on the first operating status information of the control device, the historical operating data of the control device, and the environmental information of the area where the control device is located, including: determining the habit data of the target object in using the control device based on the historical operating data, and determining the adjustment base for controlling the control device based on the habit data and the first operating status information; determining the adjustment coefficient for controlling the control device based on the historical environmental information and current environmental information of the area where the control device is located, wherein the environmental information of the area where the control device is located includes the historical environmental information and the current environmental information; and determining the adjustment base and the adjustment coefficient as the second adjustment parameter.

[0012] In an exemplary embodiment, a control instruction for controlling the control device is generated based on the intention result and the second adjustment parameter, including: calculating the target adjustment parameter for the control device based on the first adjustment parameter, the adjustment base and the adjustment coefficient; and generating the control instruction based on the first operating status information of the control device and the target adjustment parameter.

[0013] In an exemplary embodiment, after generating a control instruction for controlling the control device based on the intention result and the second adjustment parameter, the method further includes: generating a prompt voice based on the control instruction, and sending the prompt voice to the intelligent interactive device for playback; receiving the feedback result of the target object based on the prompt voice; and sending the control instruction to the control device when the feedback result is a positive result.

[0014] According to another embodiment of the present invention, a control instruction generation device is also provided, including: an identification module, used to identify the fuzzy intention of the target object through an intention recognition model, and obtain an intention result corresponding to the fuzzy intention, wherein the intention result is used to indicate the control device and the first adjustment parameter of the control device; a determination module, used to determine the second adjustment parameter for controlling the control device based on the first operating status information of the control device, the historical operation data of the control device and the environmental information of the area where the control device is located; a generation module, used to generate a control instruction for controlling the control device based on the intention result and the second adjustment parameter.

[0015] According to another aspect of the embodiments of the present invention, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is configured to execute the above-mentioned method for generating control instructions when running.

[0016] According to another aspect of an embodiment of the present invention, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the control instruction generation method through the computer program.

[0017] In an embodiment of the present application, the fuzzy intention of the target object is first identified through an intention recognition model to determine the control device to be controlled indicated by the fuzzy intention and the first adjustment parameter of the control device; then, the second adjustment parameter of the control device is determined based on the first operating status information of the control device, the historical operation data of the target object on the control device, and the environmental information of the area where the control device is located; finally, a control instruction for the control device is generated based on the identified intention result and the second adjustment parameter; the above technical solution solves the problem in the prior art that the adjustment parameters of the device are set according to characteristics such as time and region, and cannot meet the different preferences of different users for the device; and achieves the technical effect of adjusting the device control parameters in combination with the internal characteristics of the device to meet the personal preferences of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 1 is a schematic diagram of a hardware environment for an optional method for generating a control instruction according to an embodiment of the present application;

[0021] Figure 2 is a flow chart of an optional method for generating a control instruction according to an embodiment of the present invention;

[0022] Figure 3 is a data flow diagram of an optional method for generating control instructions according to an embodiment of the present invention;

[0023] Figure 4 is a system framework diagram of an optional control instruction generation method according to an embodiment of the present invention;

[0024] Figure 5 is a schematic diagram of an optional working principle of XGBoost according to an embodiment of the present invention;

[0025] Figure 6 A structural block diagram of an optional control instruction generation device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] According to one aspect of the embodiment of the present application, a method for generating a control instruction is provided. The method for generating a control instruction is widely used in smart home (Smart Home), smart home, smart home device ecology, smart residential (Intelligence House) ecology and other whole-house intelligent digital control application scenarios. Optionally, in this embodiment, the above-mentioned method for generating a control instruction can be applied to Figure 1 In the hardware environment shown in FIG. 1 , which is composed of a terminal device 102 and a server 104. Figure 1 As shown, the server 104 is connected to the terminal device 102 via a network, and can be used to provide services (such as application services, etc.) for the terminal or a client installed on the terminal. A database can be set up on the server or independently of the server to provide data storage services for the server 104. Cloud computing and / or edge computing services can be configured on the server or independently of the server to provide data computing services for the server 104.

[0029] The aforementioned network may include, but is not limited to, at least one of the following: a wired network and a wireless network. The aforementioned wired network may include, but is not limited to, at least one of the following: a wide area network, a metropolitan area network, and a local area network. The aforementioned wireless network may include, but is not limited to, at least one of the following: Wi-Fi (Wireless Fidelity) and Bluetooth. The terminal device 102 may include, but is not limited to, a PC, a mobile phone, a tablet computer, a smart air conditioner, a smart range hood, a refrigerator, a smart oven, a smart stove, a smart washing machine, a smart water heater, a smart washing machine, a smart dishwasher, a smart projector, a smart TV, a smart clothes drying rack, smart curtains, a smart audio / video system, a smart socket, a smart speaker, a smart fresh air system, smart kitchen and bathroom equipment, smart bathroom equipment, a smart robot vacuum, a smart window cleaning robot, a smart robot mop, a smart air purifier, a smart steamer, a smart microwave oven, a smart kitchen appliance, a smart purifier, a smart water dispenser, a smart door lock, and the like.

[0030] In this embodiment, a method for generating a control instruction is provided, which is applied to the above-mentioned visualization system. Figure 2 4 is a flow chart of an optional method for generating a control instruction according to an embodiment of the present invention, the flow chart comprising the following steps:

[0031] Step S202: Identify the fuzzy intention of the target object using an intention recognition model to obtain an intention result corresponding to the fuzzy intention, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device;

[0032] Step S204, determining a second adjustment parameter for controlling the control device according to the first operating state information of the control device, the historical operation data of the control device, and the environmental information of the area where the control device is located;

[0033] Step S206: Generate a control instruction for controlling the control device according to the intention result and the second adjustment parameter.

[0034] In an embodiment of the present application, the fuzzy intention of the target object is first identified through an intention recognition model to determine the control device to be controlled indicated by the fuzzy intention and the first adjustment parameter of the control device; then, the second adjustment parameter of the control device is determined based on the first operating status information of the control device, the historical operation data of the target object on the control device, and the environmental information of the area where the control device is located; finally, a control instruction for the control device is generated based on the identified intention result and the second adjustment parameter; the above technical solution solves the problem in the prior art that the adjustment parameters of the device are set according to characteristics such as time and region, and cannot meet the different preferences of different users for the device; and achieves the technical effect of adjusting the device control parameters in combination with the internal characteristics of the device to meet the personal preferences of users.

[0035] Optionally, step S202: identifying the fuzzy intention of the target object through the intention recognition model can be achieved through the following scheme, including: obtaining the fuzzy intention input by the target object through the intelligent interactive device; obtaining the second operating status information reported by all household devices, wherein all household devices include the control device; detecting the environmental information of the area through all household devices; inputting the second operating status information, the environmental information and the fuzzy intention into the intention recognition model to identify the fuzzy intention through the intention recognition model.

[0036] To more accurately identify the meaning of a user's fuzzy intent, the intent recognition model first obtains the fuzzy intent input by the user and obtains the second operating status information reported by all home devices. This second status information indicates the internal characteristics of all home devices, including the indoor environment and the operating status of the home devices, including the current operating mode and target state. The model then uses all home devices to detect the environmental information of the area where these home devices are located, that is, the environmental information of the user's area, including weather, time, and geographic location. Finally, the obtained second operating status information, environmental information, and fuzzy intent are input into the intent recognition model to identify the meaning corresponding to the fuzzy intent, namely the intent result. This approach, combining external and internal features, allows for more accurate identification of user intent, providing better service.

[0037] Optionally, execute step S202: identify the fuzzy intention of the target object through the intention recognition model to obtain the intention result corresponding to the fuzzy intention, including: extracting preset keywords from the fuzzy intention, and determining the control device from all the home devices based on the preset keywords; determining the first operating status information of the control device from the second operating status information; determining the adjustment direction indicated by the first adjustment parameter based on the first operating status information and the environmental information.

[0038] Identifying fuzzy intents through an intent recognition model involves the following steps: first, extracting preset keywords from the fuzzy intent to identify the control device to be controlled from all home devices; then, using the identified control device, determining the first operating state information of the control device from all second operating state information; and finally, determining the adjustment direction indicated by the first adjustment parameter based on the first operating state information and environmental information. This approach, by extracting preset keywords, collecting environmental information, and the operating state information of the control device, accurately determines the user's intent and provides a better user experience.

[0039] For example, the preset keywords can be "cold" and "hot". For example, if the user says "it's too cold", and the word "cold" is extracted from the sentence, it can be determined that the device to be controlled can be an air conditioner or a heating device; then the first operating status information of the air conditioner and heating device is determined from a large amount of second operating status information; combined with the first operating status information and the environmental information, it can be determined that it is a summer night, the temperature outside is 30 degrees Celsius, the air conditioner is running, and the target temperature is "16 degrees Celsius", so the adjustment direction of the device is "raising the temperature".

[0040] It should be noted that the preset keywords can also be "bright", "dark", "dry", "humid", etc., which are used to indicate different home appliances respectively.

[0041] Optionally, the adjustment direction indicated by the first adjustment parameter can also be determined by the following scheme, including: determining whether there is a second preset keyword in the fuzzy intention, wherein the second preset keyword is used to indicate that the fuzzy intention is a negative intention, and the negative intention is used to indicate that the adjustment direction of controlling the control device for adjustment is negative; when it is determined that the second preset keyword exists in the fuzzy intention, determining that the adjustment direction indicated by the first adjustment parameter is negative.

[0042] For example, if the user's fuzzy intention is "not warm enough" and the second preset keyword "not enough" is recognized, it means that the fuzzy intention is a negative intention. Originally, when warm is recognized, the temperature of the air conditioner should be lowered or the heating device should be cooled down, but because it is a negative intention, the temperature of the air conditioner will be raised or the temperature of the heating will be increased.

[0043] Optionally, the above-mentioned step S202 is performed: before the fuzzy intention of the target object is identified through the intention recognition model, the method also includes: obtaining the voice instruction input by the target object, and performing intent recognition on the voice instruction through a preset algorithm; if the preset algorithm fails to identify the intention result corresponding to the voice instruction, the voice instruction is determined to be a fuzzy intention.

[0044] Current smart home devices generally have the function of receiving user voice commands, that is, they can recognize some user commands, such as "turn on the air conditioner to 26 degrees Celsius" and "turn on the lights". For these commands, there is no need to use this solution to identify fuzzy intent. Therefore, before enabling this solution, it is necessary to first obtain the voice command input by the target object and use a preset algorithm to identify the intent of the voice command. If the preset algorithm fails to generate the intent result corresponding to the voice command, the voice command is determined to be a fuzzy intent. Through the above solution, the work objectives of this application can be made more precise, providing users with more accurate services and improving the user experience.

[0045] Optionally, executing the above-mentioned step S204: determining the second adjustment parameter for controlling the control device based on the first operating status information of the control device, the historical operation data of the control device and the environmental information of the area where the control device is located can be achieved through the following steps, specifically including: determining the habit data of the target object in using the control device based on the historical operation data, and determining the adjustment base for controlling the control device based on the habit data and the first operating status information; determining the adjustment coefficient for controlling the control device based on the historical environmental information and current environmental information of the area where the control device is located, wherein the environmental information of the area where the control device is located includes the historical environmental information and the current environmental information; and determining the adjustment base and the adjustment coefficient as the second adjustment parameter.

[0046] The method for determining the second adjustment parameter includes: first determining the habit data of the target object in using the control device through historical operation data, and the historical operation data may be historical operation data whose similarity with the current environmental information exceeds a preset threshold, and then determining the adjustment base for controlling the control device based on the obtained habit data and the first operating status information of the control device, and then determining the adjustment coefficient for controlling the control device based on the historical environmental information and current environmental information of the area where the control device is located, and finally determining the adjustment base and the adjustment coefficient as the second adjustment parameter.

[0047] For example, based on historical operation data, it is determined that the user's operating habit for the air conditioning equipment in the past when the environment was similar or the same as the current environment was 26 degrees Celsius for cooling, and the current operating temperature of the air conditioner is 16 degrees Celsius for cooling, so the adjustment base is 10 degrees Celsius; then, based on the recorded historical environmental information and the current environmental information, it is determined that the temperature has dropped today, and then the equipment adjustment coefficient model is used for calculation, and the obtained adjustment coefficient is 0.1; the second adjustment parameter is obtained as [10, 0.1].

[0048] Optionally, the above-mentioned step S206: generating a control instruction for controlling the control device based on the intention result and the second adjustment parameter, includes the following steps: calculating the target adjustment parameter for the control device based on the first adjustment parameter, the adjustment base and the adjustment coefficient; generating the control instruction based on the first operating status information of the control device and the target adjustment parameter.

[0049] After obtaining the intention result and the second adjustment parameter, the target adjustment parameter for the control device is calculated based on the first adjustment parameter, the adjustment base and the adjustment coefficient, and then a control instruction is generated based on the first operating status information of the control device and the target adjustment parameter.

[0050] For example, if the first adjustment parameter is to increase the temperature, the adjustment base is 10 degrees Celsius, and the adjustment coefficient is 0.1, then the target adjustment parameter = 10 * (1 + 0.1) = 11. Therefore, the target adjustment parameter is to increase the temperature by 11 degrees Celsius. Considering the current operating temperature of the air conditioner is 16 degrees Celsius, the generated control instruction is to adjust the operating temperature of the air conditioner to 27 degrees Celsius. This solution comprehensively considers the user's personal habits and environmental changes when adjusting home appliances, providing more personalized and diversified services to users and bringing them a better user experience.

[0051] It should be noted that the above historical operation data can be statistics of the usage status of the device in the current hour in the past 15 days, and determine the most used status, such as: the most used working mode of the air conditioner, the most used wind speed, or the most used working mode of the light.

[0052] It's important to note that the aforementioned device adjustment coefficient model is used to control continuous device parameters, allowing for more personalized device adjustments rather than relying solely on historical operating data. For example, if the temperature drops today, historical air conditioner usage statistics are based on days without a drop in temperature. Therefore, the user's desired temperature increase should be higher than the historical data. The final continuous parameter control result is: historical data * (1 + k), where k is the adjustment coefficient.

[0053] It should be noted that the value range of the adjustment coefficient output by the equipment adjustment coefficient model is usually [-1, 1], and can also be other value ranges, which is not limited in this application.

[0054] In an exemplary embodiment, after executing step S206 above: generating a control instruction for controlling the control device based on the intention result and the second adjustment parameter, the method further includes: generating a prompt voice based on the control instruction, and sending the prompt voice to the intelligent interactive device for playback; receiving the feedback result of the target object based on the prompt voice; and sending the control instruction to the control device when the feedback result is a positive result.

[0055] After generating a control command, a voice prompt needs to be sent to the user for confirmation to avoid misidentification or incorrect recognition that could lead to direct device control, which could result in a poor user experience. Therefore, a voice prompt needs to be generated based on the control command, and then sent to the intelligent interactive device for playback. The system then receives feedback from the user based on the voice prompt. If the feedback is positive, such as "OK," the control command is sent to the control device. If the feedback is negative, such as "No," it indicates that the recognition was misidentified or incorrect, and this information is used as negative evidence to train the model. This approach can avoid the poor user experience caused by misidentification.

[0056] In this embodiment, an optional method for generating a control instruction is provided. Figure 3 is a data flow chart of an optional method for generating control instructions according to an embodiment of the present invention, such as Figure 3 As shown, specifically including:

[0057] During use, the device will report its current status, namely internal features, in real time. The internal features mainly include the working status of the home device and the environmental status inside the room that can be measured by the home device, such as the current air conditioning working mode, target temperature, current temperature, wind speed, blowing direction, light brightness, light mode, curtain opening, oxygen concentrator operating mode, indoor oxygen content, humidifier working mode, humidifier gear, indoor humidity, etc., the frequency of device usage in the current time period within a week, the time interval from the last use to the current time, the time period with the most or least use within a week, etc.

[0058] External features can include:

[0059] (1) Weather characteristics: weather conditions, wind direction, wind level, air pressure, sunrise time, sunset time, temperature, etc.;

[0060] (2) Time characteristics: season, month, hour, whether it is a weekday, etc.;

[0061] (3) Location characteristics: longitude, latitude, distance from the 33rd parallel north, etc.

[0062] The fuzzy recognition model (equivalent to the above-mentioned intention recognition model) mainly combines the user input corpus (equivalent to the above-mentioned fuzzy intent) and internal features (equivalent to the above-mentioned first operating status information) and external features (equivalent to the above-mentioned environmental information) to predict the device that the user needs to control (equivalent to the above-mentioned control device) and the adjustment direction of the device (equivalent to the above-mentioned first adjustment parameter), for example (control device: air conditioning, adjustment direction: heating), (control device: light, adjustment direction: increase brightness).

[0063] Device historical data (equivalent to the above-mentioned historical operation data): Statistics on the usage status of each controlled device within the current hour over the past 15 days, including the most used status, such as the most used operating mode of the air conditioner, the most used wind speed, and the most used operating mode of the light. Device historical data is the average value of continuous data, such as the average temperature of the air conditioner, the average brightness of the light, and the average opening and closing degree of the curtains.

[0064] Equipment adjustment coefficient model: The output result is the adjustment coefficient k. The value range of k is usually [-1, 1]. It is used to adjust the continuous parameters of the equipment, making the equipment adjustment more personalized, rather than relying solely on historical operation data. For example, if the temperature drops today, the air conditioning usage data based on historical data is collected on weather without cooling. Therefore, the temperature that the user wants to increase should be higher than the historical data. The final control result of the continuous parameter is: historical data * (1 + k).

[0065] It should be noted that the data output in the above models must be converted into vectors or features that the model can recognize in advance, including:

[0066] User input: A sentence spoken by the user, such as "The room is a bit cold," "The water is a bit cold," "It's too dark," "A bit dry," "There's a lot of smoke," etc., is converted into a 128-dimensional vector representation using the word2vec model.

[0067] It should be noted that the word2vec model is a group of related models used to generate word vectors. These models are shallow, two-layer neural networks that are trained to reconstruct linguistic word texts and can be used to map each word to a vector.

[0068] Internal features:

[0069] (1) Indoor environment:

[0070] Temperature: continuous feature, used directly as a feature;

[0071] CO2 content: continuous feature, used directly as a feature;

[0072] PM2.5 content: continuous feature, used directly as a feature;

[0073] O2 content: continuous feature, used directly as a feature;

[0074] Humidity: Continuous feature, used directly as a feature.

[0075] (2) Air conditioning status:

[0076] Target temperature: continuous feature, used directly as a feature;

[0077] Working mode: Category features are converted into vector representation using onehot encoding.

[0078] (3) Light status:

[0079] Brightness: continuous feature, used directly as a feature;

[0080] Mode: Categorical features, converted to vector representation using onehot encoding.

[0081] (4) Curtain status:

[0082] Opening degree: continuous feature, used directly as a feature.

[0083] (5) Water heater status:

[0084] Current water temperature: a continuous feature, used directly as a feature;

[0085] Target water temperature: a continuous feature, used directly as a feature.

[0086] (6) Status of the sweeping robot:

[0087] Electricity: Continuous feature, used directly as a feature;

[0088] Is it working?: Categorical feature, only two states, represented by 01.

[0089] (7) Humidifier status:

[0090] Working or not: categorical feature, only two states, represented by 01;

[0091] Working mode: Categorical features are converted into vector representation using onehot encoding;

[0092] Target humidity: Continuous feature, used directly as a feature.

[0093] (8) Oxygen concentrator status:

[0094] Working or not: categorical feature, only two states, represented by 01;

[0095] Working mode: Category features are converted into vector representation using onehot encoding.

[0096] (9) Range hood working status:

[0097] Working or not: categorical feature, only two states, represented by 01;

[0098] Working mode: Categorical features are converted into vector representation using onehot encoding;

[0099] Wind speed: categorical feature, converted into a vector representation using onehot encoding.

[0100] (10) Electric fan working status:

[0101] Working or not: categorical feature, only two states, represented by 01;

[0102] Working mode: Categorical features are converted into vector representation using onehot encoding;

[0103] Wind speed: categorical feature, converted into a vector representation using onehot encoding.

[0104] (11) Window pusher:

[0105] Position: Continuous feature, used directly as a feature.

[0106] External features:

[0107] (1) Weather characteristics:

[0108] Weather: categorical features, converted into vector representation using onehot encoding;

[0109] Wind direction: categorical feature, converted into vector representation using onehot encoding;

[0110] Wind level: categorical feature, converted into vector representation using onehot encoding;

[0111] Air pressure: continuous feature, used directly as a feature;

[0112] Sunrise time: continuous feature, used directly as a feature;

[0113] Sunset time: a continuous feature, used directly as a feature;

[0114] Temperature: Continuous feature, used directly as a feature.

[0115] (2) Time characteristics:

[0116] Season: categorical feature, converted into vector representation using onehot encoding;

[0117] Month: categorical feature, converted into vector representation using onehot encoding;

[0118] Hour: Category features, converted into vector representation using onehot encoding;

[0119] Whether it is a weekday: categorical feature, converted into a vector representation using onehot encoding.

[0120] (3) Location characteristics:

[0121] Longitude: continuous feature, used directly as a feature;

[0122] Latitude: continuous feature, used directly as a feature;

[0123] Distance from the 33rd parallel north: Continuous feature, used directly as a feature.

[0124] In order to further understand the above data flow diagram, this application also proposes another embodiment. Figure 4 To illustrate, Figure 4 As shown, Figure 4 It is a system framework diagram of an optional control instruction generation method.

[0125] The inputs of the fuzzy intent recognition model include: user input, internal features and external features of all home devices, such as {user input text, room temperature, CO2, ..., humidity, air conditioning target temperature, air conditioning working mode, ..., window pusher position, weather, wind direction, ..., latitude}, which are converted into feature vectors to represent and then input into the fuzzy intent recognition model; the final output result is (device, adjustment direction), for example: (air conditioning, increase); the input of the device adjustment coefficient model is the internal and external features of the control device, such as {room temperature , CO2,…, humidity, air conditioning target temperature, air conditioning working mode,…, window pusher position, weather, wind direction,…, latitude}, and convert them into feature vectors respectively. The output of the final model is the adjustment coefficient k, and the value range of k is [-1,1], for example, 0.5; the final air conditioning historical preference temperature is determined to be 24 degrees Celsius, and the current room temperature is set at 22 degrees Celsius. Referring to the historical preference data, the temperature needs to be increased by 2 degrees Celsius; and the proportional adjustment coefficient is 0.5, and the final temperature rise is (1+0.5)*2=3, so the final air conditioning temperature is set to 25 degrees Celsius.

[0126] In the existing technology, for prediction problems, the academic community usually tends to use classifiers such as "decision trees" and "logistic regression" to solve them. Among them, the "decision tree" is to classify and judge a single sample by globally establishing a tree-like algorithm description logic. The training of this tree is usually also trained with sampled global user usage history data to add different data dimensions to achieve information gain and entropy reduction, and then obtain a global classification prediction model; and "logistic regression" is also trained by sampling global user history data to obtain a decision plane. The prediction of a single sample is to determine whether it falls above or below the decision plane to determine whether it belongs to a certain category. At the academic level, both methods have relatively good logical explanations and theoretical support.

[0127] As the IoT continues to expand, algorithms that once seemed mature in academia are facing a series of challenges in their implementation. Many theoretically feasible algorithms, however, are cost-prohibitive to implement in practice due to the sheer volume of data and the sheer number of samples involved. Workarounds are necessary to address this issue. To achieve approximate predictions, a better approach is to move away from classification algorithms like decision trees and logistic regression and instead use a kNN search method. Because it doesn't require explicit training, this approach can achieve relatively accurate predictions with very limited computing resources. Furthermore, the vector search function used in this process is easy to implement, offering significant advantages in mitigating implementation risks and reducing computational costs.

[0128] It should be noted that the above fuzzy intent recognition model and device adjustment coefficient model are both obtained based on XGBoost training. The working principle of XGBoost is as follows: After training, XGBoost will obtain multiple trees. Each tree will split nodes according to different features and feature values, and finally obtain the score of the category in the current tree at the leaf node. During the prediction process, the input features will be classified to the final leaf node according to the features on each tree, and the score of the tree will be obtained. Therefore, the tree score multiplied by the weight of each tree is the score of the category, such as Figure 5 As shown, Figure 5 This is a schematic diagram of the working principle of XGBoost.

[0129] Assuming there are n trees in total, the score of the current category is:

[0130] Score = μ*f1(x) +μ*f2(x) +μ*f3(x) + …+μ*fn(x).

[0131] Through the description of the above embodiments, those skilled in the art will clearly understand that the methods according to the above embodiments can be implemented using software plus the necessary general-purpose hardware platform. Of course, hardware can also be used, but in many cases the former is the more preferred embodiment. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, or optical disk) and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0132] In this embodiment, a device for generating control instructions is also provided. The device for generating control instructions is used to implement the above-mentioned embodiments and preferred embodiments. The details already described will not be repeated here. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0133] Figure 6 A structural block diagram of an optional control instruction generating device according to an embodiment of the present invention; Figure 6 including:

[0134] an identification module 62 for identifying the fuzzy intent of the target object using an intention recognition model to obtain an intention result corresponding to the fuzzy intent, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device;

[0135] a determination module 64 configured to determine a second adjustment parameter for controlling the control device based on the first operating state information of the control device, historical operation data of the control device, and environmental information of an area where the control device is located;

[0136] The generating module 66 is configured to generate a control instruction for controlling the control device according to the intention result and the second adjustment parameter.

[0137] Through the above-mentioned device, the fuzzy intention of the target object is first identified through the intention recognition model to determine the control device to be controlled indicated by the fuzzy intention and the first adjustment parameter of the control device; then the second adjustment parameter of the control device is determined according to the first operating status information of the control device, the historical operation data of the target object on the control device, and the environmental information of the area where the control device is located; finally, the control instruction of the control device is generated according to the identified intention result and the second adjustment parameter; the above-mentioned technical solution solves the problem in the prior art that the adjustment parameters of the device are set according to characteristics such as time and region, and cannot meet the different preferences of different users for the device; and achieves the technical effect of adjusting the device control parameters in combination with the internal characteristics of the device to meet the personal preferences of users.

[0138] Optionally, the identification module 62 is further used to obtain the fuzzy intention input by the target object through the intelligent interactive device; obtain the second operating status information reported by all household devices, wherein all the household devices include the control device; detect the environmental information of the area through all the household devices; input the second operating status information, the environmental information and the fuzzy intention into the intention recognition model to identify the fuzzy intention through the intention recognition model.

[0139] Optionally, the identification module 62 is also used to extract preset keywords from the fuzzy intention, and determine the control device from all the home devices based on the preset keywords; determine the first operating status information of the control device from the second operating status information; and determine the adjustment direction indicated by the first adjustment parameter based on the first operating status information and the environmental information.

[0140] Optionally, the recognition module 62 is also used to obtain the voice instruction input by the target object before identifying the fuzzy intention of the target object through the intention recognition model, and perform intent recognition on the voice instruction through a preset algorithm; if the preset algorithm fails to identify the intention result corresponding to the voice instruction, the voice instruction is determined to be a fuzzy intention.

[0141] Optionally, the determination module 64 is also used to determine the habit data of the target object in using the control device based on the historical operation data, and determine the adjustment base for controlling the control device based on the habit data and the first operating status information; determine the adjustment coefficient for controlling the control device based on the historical environmental information and current environmental information of the area where the control device is located, wherein the environmental information of the area where the control device is located includes the historical environmental information and the current environmental information; and determine the adjustment base and the adjustment coefficient as the second adjustment parameter.

[0142] Optionally, the generation module 66 is further used to calculate the target adjustment parameter for the control device based on the first adjustment parameter, the adjustment base and the adjustment coefficient; and generate the control instruction based on the first operating status information of the control device and the target adjustment parameter.

[0143] In an exemplary embodiment, the generation module 66 is also used to generate a control instruction for controlling the control device based on the intention result and the second adjustment parameter, generate a prompt voice based on the control instruction, and send the prompt voice to the intelligent interactive device for playback; receive the feedback result of the target object based on the prompt voice; and send the control instruction to the control device when the feedback result is a positive result.

[0144] An embodiment of the present invention further provides a storage medium, which includes a stored program, wherein the program executes any of the above methods when it is run.

[0145] Optionally, in this embodiment, the storage medium may be configured to store program codes for executing the following steps:

[0146] S1, identifying the fuzzy intent of the target object through an intention recognition model to obtain an intention result corresponding to the fuzzy intent, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device;

[0147] S2, determining a second adjustment parameter for controlling the control device according to the first operating state information of the control device, historical operation data of the control device, and environmental information of the area where the control device is located;

[0148] S3: Generate a control instruction for controlling the control device according to the intention result and the second adjustment parameter.

[0149] An embodiment of the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0150] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0151] Optionally, in this embodiment, the processor may be configured to execute the following steps through a computer program:

[0152] S1, identifying the fuzzy intent of the target object through an intention recognition model to obtain an intention result corresponding to the fuzzy intent, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device;

[0153] S2, determining a second adjustment parameter for controlling the control device according to the first operating state information of the control device, historical operation data of the control device, and environmental information of the area where the control device is located;

[0154] S3: Generate a control instruction for controlling the control device according to the intention result and the second adjustment parameter.

[0155] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0156] Optionally, specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0157] Obviously, those skilled in the art will appreciate that the various modules or steps of the present invention described above can be implemented using a general-purpose computing device, centralized on a single computing device, or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. In some cases, the steps shown or described can be performed in a different order than that shown, or can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0158] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for generating a control instruction, characterized in that: include: Identifying the fuzzy intention of the target object through the intention recognition model to obtain an intention result corresponding to the fuzzy intention, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device; determining a second adjustment parameter for controlling the control device according to first operating state information of the control device, historical operation data of the control device, and environmental information of an area where the control device is located; generating a control instruction for controlling the control device according to the intention result and the second adjustment parameter; The method of determining a second adjustment parameter for controlling the control device according to the first operating state information of the control device, the historical operation data of the control device, and the environmental information of the area where the control device is located includes: determining habit data of the target object in using the control device based on the historical operation data, and determining an adjustment base for controlling the control device based on the habit data and the first operating state information; Calculating an adjustment coefficient for controlling the control device based on historical environmental information and current environmental information of the area where the control device is located, wherein the environmental information of the area where the control device is located includes the historical environmental information and the current environmental information; The adjustment base and the adjustment coefficient are determined as the second adjustment parameters.

2. The method for generating a control instruction according to claim 1, wherein: The fuzzy intention of the target object is identified through the intention recognition model, including: Acquiring the fuzzy intention input by the target object through an intelligent interactive device; Obtaining second operating status information reported by all home appliances, wherein all home appliances include the control device; Detecting environmental information of the area through all the home devices; The second operating state information, the environmental information and the fuzzy intention are input into the intention recognition model to recognize the fuzzy intention through the intention recognition model.

3. The method for generating a control instruction according to claim 2, wherein: The fuzzy intent of the target object is identified through the intent recognition model to obtain the intent result corresponding to the fuzzy intent, including: Extracting preset keywords from the fuzzy intent, and determining the control device from all the home devices according to the preset keywords; determining first operating state information of the control device from the second operating state information; An adjustment direction indicated by the first adjustment parameter is determined according to the first operating state information and the environmental information.

4. The method for generating a control instruction according to claim 1, wherein: Before identifying the fuzzy intent of the target object using the intent recognition model, the method further includes: Acquire the voice command input by the target object, and perform intent recognition on the voice command using a preset algorithm; If the preset algorithm fails to identify the intention result corresponding to the voice instruction, the voice instruction is determined to be an ambiguous intention.

5. The method for generating a control instruction according to claim 1, wherein: Generating a control instruction for controlling the control device according to the intention result and the second adjustment parameter includes: Calculating a target adjustment parameter for the control device according to the first adjustment parameter, the adjustment base, and the adjustment coefficient; The control instruction is generated according to the first operating state information of the control device and the target adjustment parameter.

6. The method for generating a control instruction according to claim 1, wherein: After generating a control instruction for controlling the control device according to the intention result and the second adjustment parameter, the method further includes: Generate a prompt voice according to the control instruction, and send the prompt voice to the intelligent interactive device for playback; receiving a feedback result from the target object according to the prompt voice feedback; When the feedback result is a positive result, the control instruction is sent to the control device.

7. A control instruction generating device, characterized in that: include: an identification module, configured to identify the fuzzy intent of the target object using an intention recognition model, and obtain an intention result corresponding to the fuzzy intent, wherein the intention result is used to indicate a control device and a first adjustment parameter of the control device; a determination module, configured to determine a second adjustment parameter for controlling the control device based on the first operating state information of the control device, historical operation data of the control device, and environmental information of an area where the control device is located; a generating module, configured to generate a control instruction for controlling the control device according to the intention result and the second adjustment parameter; Among them, the determination module is also used to determine the habit data of the target object in using the control device based on the historical operation data, and determine the adjustment base for controlling the control device based on the habit data and the first operating status information; calculate the adjustment coefficient for controlling the control device based on the historical environmental information and current environmental information of the area where the control device is located, wherein the environmental information of the area where the control device is located includes the historical environmental information and the current environmental information; and determine the adjustment base and the adjustment coefficient as the second adjustment parameter.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the method according to any one of claims 1 to 6 is executed when the program is executed.

9. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 6 through the computer program.

Images