A voice recognition-based intelligent household appliance control method and system, and a storage medium

Abstract

The application discloses a kind of based on speech recognition's intelligent household appliance control method, system and storage medium, belong to intelligent household appliance control field, method includes the following steps: establishing control network;When mobile terminal access control network, the authentication of mobile terminal is carried out;When multiple intelligent household appliances receive control voice, control voice received by each other is sent to cloud module by central control module, and cloud module converts control voice into text information, and compares whether the text information converted by different control voice is identical, determines the correct text information as target information;If target information does not include appliance name, based on the working mode of each intelligent household appliance, the demand of target information is identified to determine the target equipment in intelligent household appliance.The application can control intelligent household appliance without including wake-up word, so as to improve the convenience of appliance control.

Description

Technical Field

[0001] This invention belongs to the field of smart home appliance control technology, specifically relating to a smart home appliance control method, system, and storage medium based on voice recognition. Background Technology

[0002] The rise of smart homes has greatly facilitated people's lives. Currently, smart home appliances are generally controlled via individual voice commands that activate them, allowing for direct human-computer interaction, or remotely via a mobile app. However, one-to-one voice control of individual smart appliances is inconvenient. Users must activate the appliance with a specific activation command before issuing the corresponding action command. When there are multiple types of smart appliances, users may not be able to recall the activation command, resulting in the inability to control them via voice. Controlling via a mobile app requires opening the app, adding an intermediate step and reducing the convenience of voice control. Summary of the Invention

[0003] To address the aforementioned issues, a method, system, and storage medium for controlling smart home appliances based on voice recognition are proposed, aiming to resolve the inconvenience of voice control of smart home appliances in existing technologies.

[0004] In a first aspect, the present invention also provides a method for controlling smart home appliances based on voice recognition, the method comprising: The central control module establishes a control network and connects mobile terminals and smart home appliances to the control network; After the mobile terminal accesses the control network, the authentication module authenticates the mobile terminal. If the authentication is successful, and the mobile terminal is the first device to access the control network, the voice module located in each of the smart home appliances is activated. When multiple smart home appliances receive control voice commands, they send their respective received control voice commands to the central control module. The central control module then sends the control voice commands to the cloud module. The cloud module converts the control voice commands into text information and compares whether the text information converted from different control voice commands is the same. If it is the same, it selects one of the text information as the target information. If not, it determines the correct text information as the target information. The cloud module sends the target information to the central control module. If the target information does not include the name of the home appliance, the module identifies the need for the target information based on the current working mode of each smart home appliance to determine the target device among the smart home appliances. The target device is the smart home appliance that the control voice directly wants to control. The central control module generates a first instruction to control the target device and a second instruction to control non-target devices. Each smart home appliance performs a corresponding action based on the first instruction and the second instruction.

[0005] Furthermore, the cloud module converts the control voice into the text information by including the following steps: The central control module acquires the identifiers of each target device located in the control network, and obtains the type and working mode of the target device based on the identifiers, with each working mode corresponding to a pre-assigned identification code; The central control module generates an indoor map, which marks the location of the smart home appliances. After receiving the control voice, it locates the user's voice position and obtains the current working mode of the smart home appliances located in the same room as the voice position. Based on the identification code corresponding to the current working mode, it generates a noise sequence. The central control module sends the noise sequence and the control voice together to the cloud module. The cloud module has multiple separation models pre-set. After receiving the control voice, the cloud module calls the corresponding separation model based on the noise sequence. After using the called separation model to separate the human voice track from the control voice, it calls the recognition model to convert the human voice track into the text information.

[0006] Furthermore, the central control module generates the indoor map by including the following steps: The location of the central control module is preset. The central control module generates virtual commands and sends the virtual commands to each of the smart home appliances so that the smart home appliances can perform virtual actions. The central control module and each of the smart home appliances acquire the sound generated by the virtual actions. Based on the volume of the sound, the positional relationship between each of the smart home appliances and the central control module, as well as the positional relationship between the smart home appliances, is determined. Based on the positional relationship, a first initial map is generated. If no designated device is connected within the control network, the central control module generates a notification message and sends it to the mobile terminal to remind the user to manually draw the second initial map. If the designated device is connected, the second initial map stored therein is retrieved from the designated device. The first initial map is combined with the second initial map to generate the indoor map.

[0007] Further, determining the target device in the smart home appliance includes the following steps: The central control module is equipped with a function table, which includes the function of each smart home appliance and multiple trigger characters corresponding to the function. The trigger characters include positive characters and negative characters. The trigger characters are extracted from the text information and matched with the corresponding function based on the function table to define the target function. The smart home appliance that contains the target function and is in operation is identified as the initial target. If there is only one initial target, it is directly set as the target device. If there are multiple initial targets, the current working mode of each initial target is acquired. The central control module pre-stores the contribution value of each working mode, acquires the contribution value of the current working mode, and determines the target device among the initial targets by combining the type of the extracted trigger character and the size of the contribution value.

[0008] Further, generating the second instruction for control of a device other than the target device includes the following steps: The smart home appliance currently located in the same room as the sound source and having a positive impact on the functions related to the trigger character is defined as the first home appliance, and the smart home appliance having a negative impact is defined as the second home appliance. The central control module generates a positive second instruction for controlling the first home appliance and a negative second instruction for controlling the second home appliance. The positive second instruction is used to control the functions of the first home appliance related to the trigger character to increase the impact on the environment, and the negative second instruction is used to control the second home appliance to reduce the impact on the environment.

[0009] Furthermore, authenticating the mobile terminal includes the following steps: During authentication, the user is prompted to issue an authentication voice. The central control module determines whether the user belongs to the control network based on the tone and timbre of the authentication voice. It also distinguishes between adult users and child users based on the tone and timbre of the voice. Adult users have control rights over all smart home appliances, while child users have control rights over non-kitchen smart home appliances.

[0010] Furthermore, the central control module stores the names of various smart home appliances and the wake words used by family members for each smart home appliance. If the central control module recognizes that the text information includes the wake word, it skips the judgment process for the target device.

[0011] Secondly, the present invention provides a smart home appliance control system based on voice recognition, comprising: A central control module establishes a control network and connects mobile terminals and smart home appliances to the control network. The central control module also receives the control voice and sends it to the cloud module, and receives target information from the cloud module. If the target information does not include the name of the home appliance, the central control module identifies the needs of the target information based on the current working mode of each smart home appliance to determine the target device among the smart home appliances. The target device is the smart home appliance that the control voice directly wants to control. The central control module generates a first instruction to control the target device and a second instruction to control other than the target device. The authentication module authenticates the mobile terminal after it connects to the control network. If the authentication is successful and the mobile terminal is the first device to connect to the control network, the voice modules located in each of the smart home appliances are activated. A voice module is installed in each of the smart home appliances. The voice module is used to receive the control voice and send the control voice to the central control module. The cloud module converts the control voice into text information, compares whether the text information converted from different control voices is the same. If they are the same, it selects one of the text information as the target information; otherwise, it determines the correct text information as the target information and sends the target information to the central control module.

[0012] Thirdly, the present invention provides a storage medium storing program instructions, wherein the program instructions, when running, control the device where the storage medium is located to execute the aforementioned intelligent home appliance control method based on voice recognition.

[0013] Compared with the prior art, the beneficial effects of the present invention are at least as follows: In this invention, when a control network is established, each smart home appliance receives a control voice message and sends it to the central control module via the network. The central control module then packages and sends the message to the cloud module, enabling the cloud module to collectively recognize all control voice messages for easy comparison of subsequent text information. By comparing the text information, the correct text information can be extracted as the target information. Finally, the central control module analyzes the target information to determine the requirements of the control voice messages, thus enabling control of the corresponding smart home appliances even when the control voice messages do not contain the names of the appliances.

[0014] The central control module in this invention also monitors the number of mobile terminals in the control network. If there are mobile terminals in the control network, the voice recognition function of each smart home appliance connected to the control network is activated. If there are no mobile terminals in the control network, the voice recognition function of the smart home appliances is turned off. This can save energy and also prevent smart home appliances from accidentally starting when they receive voice signals from the external environment when no one is home. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0016] Figure 1 A flowchart of a smart home appliance control method based on voice recognition provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the principle of an indoor map according to an embodiment of the present invention; Figure 3 This is a block diagram of a smart home appliance control system based on voice recognition, provided for an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the specific embodiments described herein are merely illustrative of the invention and represent only a portion, not all, of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0018] like Figure 1 This is a flowchart of a voice recognition-based smart home appliance control method provided by the present invention. The process specifically includes: Step S1: The central control module establishes a control network and connects mobile terminals and smart home appliances to the control network.

[0019] Specifically, the central control module can be a Bluetooth gateway, the established control network can be a Zigbee or Bee network, and the mobile terminal can be a smartphone, tablet, laptop, etc. Smart home appliances include smart refrigerators, smart water heaters, smart water dispensers, smart TVs, smart air conditioners, smart lights, smart curtains, smart toilets, robot vacuum cleaners, etc.

[0020] Step S2: After the mobile terminal accesses the control network, the authentication module authenticates the mobile terminal. If the authentication is successful, and the mobile terminal is the first device to access the control network, the voice module located in each smart home appliance is activated.

[0021] Specifically, when the central control module detects a mobile terminal accessing the network, it sends an authentication reminder to the mobile terminal. Upon receiving the reminder, the mobile terminal authenticates using a preset method, such as password authentication, voiceprint authentication, fingerprint authentication, or facial recognition authentication. When no mobile terminal is connected to the control network, the voice recognition modules in each smart home appliance are in a sleep state by default, not detecting control voice commands, thus saving power. Afterward, if the central control module detects a mobile terminal connected to the control network, it will activate the voice module.

[0022] Step S3: After multiple smart home appliances receive control voice commands, they send their respective received control voice commands to the central control module. The central control module then sends the control voice commands to the cloud module. The cloud module converts the control voice commands into text information and compares whether the text information converted from different control voice commands is the same. If they are the same, it selects one of the text information as the target information; otherwise, it determines the correct text information as the target information.

[0023] When a smart home appliance receives a control voice command, it sends the command to the central control module. The central control module then transmits the command via the internet to a cloud module for recognition and comparison. The reason for this comparison is that, for example... Figure 2 As shown, a TV (TV), a stereo (TC), a central control module (CL), and a robot vacuum cleaner (SR) are placed in the living room; a washing machine (HW) is placed on the balcony; and a refrigerator (RF) is placed in the kitchen. When the user issues a control voice command from location P1, the volume of the voice received by each smart home appliance varies due to the distance from P1, and the volume is generally low. It is impossible to guarantee which device receives the best voice quality, and therefore impossible to determine which converted text information is correct. Thus, comparison is required. A voting method is used for comparison. If two of the three pieces of text information are identical, then the content of those two pieces of text information is set as the target information.

[0024] Step S4: The cloud module sends the target information to the central control module. If the target information does not include the name of the home appliance, the system identifies the target information based on the current working mode of each smart home appliance to determine the target device among the smart home appliances. The target device is the smart home appliance that the voice control is directly used to control.

[0025] Step S5: The central control module generates a first instruction to control the target device and a second instruction to control the non-target device. Each smart home appliance executes the corresponding action based on the first instruction and the second instruction.

[0026] In this embodiment, the control voice includes two types: with and without appliance names. For example, the control voice without appliance names is "Too noisy," while the control voice with appliance names is "Air conditioner, adjust the temperature to ** degrees" or "Little *, please adjust the air conditioner temperature for me." When the converted text information does not include appliance names, the device to which the target information points is determined based on the current operating mode of each appliance. For example, "Too noisy" might point to a TV or stereo. It is necessary to determine which specific device is the target device, and then generate a first instruction for the target device and a second instruction for the non-target device. The specific methods for determining the target device and generating the first and second instructions will be described in detail later.

[0027] In this invention, when a control network is established, each smart home appliance receives a control voice message and sends it to the central control module via the network. The central control module then packages and sends the message to the cloud module, enabling the cloud module to collectively recognize all control voice messages for easy comparison of subsequent text information. By comparing the text information, the correct text information can be extracted as the target information. Finally, the central control module analyzes the target information to determine the requirements of the control voice messages, thus enabling control of the corresponding smart home appliances even when the control voice messages do not contain the names of the appliances.

[0028] The central control module in this invention also monitors the number of mobile terminals in the control network. If there are mobile terminals in the control network, the voice recognition function of each smart home appliance connected to the control network is activated. If there are no mobile terminals in the control network, the voice recognition function of the smart home appliances is turned off. This can save energy and also prevent smart home appliances from accidentally starting when they receive voice signals from the external environment when no one is home.

[0029] Of particular note is that this invention enables the control of smart home appliances without the need for a wake word, thereby improving the convenience of appliance control.

[0030] In this embodiment, the cloud module converts control speech into text information by including the following steps: The central control module obtains the identifiers of each target device located in the control network, and based on the identifiers, obtains the type of the target device and the included working modes. Each working mode corresponds to a pre-assigned identification code.

[0031] Specifically, the central control module has a working mode lookup table, which includes the identifier, working mode, and identification code of each target device. Specifically, the central control module determines the type of smart home appliance in the control network through the identifier, and determines the working mode according to the current identification code of the smart home appliance. For example, identifier 10000 is an air conditioner, and the corresponding identification code 01 is the cooling mode and 02 is the heating mode. 01000 is a washing machine, and identification code A is the washing mode and B is the spin-drying mode.

[0032] The central control module generates an indoor map, which marks the locations of smart home appliances. After receiving a control voice, it locates the user's voice position and obtains the current operating mode of the smart home appliances in the same room as the voice position. Based on the identification code corresponding to the current operating mode, it generates a noise sequence. The central control module sends the noise sequence and the control voice together to the cloud module.

[0033] Specifically, such as Figure 2 The image shows an indoor map. After receiving a control voice command, each smart home appliance first locates the user's position P2. The central control module determines the name of the smart home appliance that issued the control voice command based on the identifier attached to the voice command. Since the TV (TV), speaker (TC), and robot vacuum cleaner (SR) are located in the same room as the user and are all in operation mode, the central control module then obtains the operating mode of the three appliances based on their identification codes and generates corresponding noise sequences. For example, the TV identification code is W12, representing normal playback mode with a volume of 12; the speaker identification code is Y20, representing rock mode with a volume of 20; and the robot vacuum cleaner SR's identification code is U, representing powerful cleaning mode. Therefore, the noise sequence W12Y20U is generated and sent to the cloud module.

[0034] The cloud module has multiple pre-set separation models. After receiving the control voice, the cloud module calls the corresponding separation model based on the noise sequence. After using the called separation model to separate the human voice track from the control voice, it calls the recognition model to convert the human voice track into text information.

[0035] The cloud module contains pre-trained separation models built on neural network technology. These models can be BP, RBF, or GAN, and there are multiple separation models, each generated from different training data. For example, separation model 1 is trained using a sample set containing human voice, television volume between 10-20, noise from a speaker in rock mode at 10-20, and noise from a robot vacuum cleaner during heavy cleaning. Separation model 2 is trained using a sample set containing human voice and noise from a robot vacuum cleaner during normal cleaning. Therefore, upon receiving a noise sequence, the cloud module calls the corresponding separation model for processing. Furthermore, the recognition model in this embodiment is also built on neural network technology, such as an RNN model. The advantage of this approach is that different models are built for different noise conditions. As is well known, models trained on proprietary sample sets perform better than general-purpose models. Therefore, calling different models based on different noise environments makes the separated human voice clearer, thereby increasing the accuracy of subsequent speech-to-text conversion.

[0036] In this embodiment, the central control module generates an indoor map by including the following steps: The location of the central control module is preset. The central control module generates virtual commands and sends them to each smart home appliance to make the smart home appliances perform virtual actions. The central control module and each smart home appliance obtain the sound generated by the virtual actions. Based on the sound volume, the positional relationship between each smart home appliance and the central control module, as well as the positional relationship between smart home appliances, is determined. Based on the positional relationship, a first initial map is generated.

[0037] Specifically, the central control module's location is first manually marked on a blank map. Then, the central control module generates virtual commands and sends them to smart home appliances. These virtual commands include instructing the washing machine to enter spin-dry mode, starting the fan, and making the refrigerator emit a warning sound. The smart home appliances then locate other appliances by analyzing the sounds and directions they emit. There are various methods for sound-based location, such as sound wave propagation speed and voiceprint localization. Notably, devices with voice modules inevitably include microphone arrays, which are the fundamental hardware for voice recognition. With a microphone array, the device can... The relative direction between two devices is determined using the speed of sound. For example, if a washing machine (HW) emits sound, and the central control module contains a circular microphone array, the direction of the washing machine (HW) relative to the central control module is determined by the line drawn between the microphone that first receives the sound and the center of the array. By multiplying the time difference between when the washing machine emits the sound and when the central control module receives it by the speed of sound, the approximate distance between the two devices can be determined. This gives the relative position of the washing machine (HW) and the central control module. The method for determining the position of other appliances relative to the central control module is similar and will not be elaborated here. Once the relative positions of each appliance and the central control module are known, the positional relationships between the appliances are also determined. After determining the relative positions of each appliance, an initial map can be generated.

[0038] If no designated device is connected within the control network, the central control module generates a notification message and sends it to the mobile terminal, reminding the user to manually draw the second initial map. If a designated device is connected, the second initial map stored in the designated device is retrieved.

[0039] The first initial map is combined with the second initial map to generate an indoor map.

[0040] The second initial map is a floor plan of the room, with the designated device being a robot vacuum cleaner. Since the robot vacuum cleaner will inevitably traverse the room and create a room map during use, the second initial map can be directly obtained from the room if a robot vacuum cleaner is present. By combining the first and second initial maps, an indoor map including the location of smart home appliances can be generated. If a robot vacuum cleaner is not present, the user needs to be reminded to manually draw the floor plan on the first initial map to generate the specific second initial map.

[0041] In this embodiment, determining the target device in a smart home appliance includes the following steps: The central control module contains a function table, which includes the function of each smart home appliance and multiple trigger characters corresponding to the function. The trigger characters include positive characters and negative characters. The trigger characters are extracted from the text information and matched with the corresponding function based on the function table to define the target function.

[0042] In the function list, the functions corresponding to the air conditioner include heating, cooling, and dehumidification. The trigger characters for the heating function include cold and hot. The speakers and TVs include audio playback functions, and the corresponding trigger characters include loud, low, noisy, and inaudible. Among them, the positive trigger characters are low and inaudible, and the negative trigger characters are loud and noisy. After receiving the text information, the central control module extracts the trigger characters from the text information. For example, if the text information is "It's so noisy", extracting the "noisy" will determine that the corresponding function is audio playback, which is defined as the target function.

[0043] The system acquires smart home appliances that contain the target function and are in operation, defining them as initial targets. If there is only one initial target, it directly sets the initial target as the target device. If there are multiple initial targets, it continues to acquire the current working mode of each initial target. The central control module pre-stores the contribution value of each working mode, acquires the contribution value of the current working mode, and determines the target device among the initial targets by combining the type of the extracted trigger character and the size of the contribution value.

[0044] Then, obtain the smart home appliances that are currently in operation, refer to... Figure 2 When a user at location P2 says "It's so noisy," and the TV and speakers are running, the TV and speakers are defined as the initial targets. The operating modes of the TV and speakers are then determined. The contribution values ​​for each initial target and its operating mode are pre-set. For example, the speaker's contribution value is 30 in rock mode at volume 20, and 15 in normal TV mode at volume 12. Since the trigger character "noisy" is negative, the initial device with the highest contribution value is set as the target device. A first instruction to lower the speaker volume is then generated, indicating that the speaker volume is too loud and interfering with the user's TV viewing. Conversely, if the user says "The volume is too low," the initial device with the lowest contribution value is set as the target device, and a first instruction to increase the TV volume is generated, indicating that the TV volume is too low. Finally, if the user says "The volume is too low," but only the TV is running, a first instruction to increase the TV volume is directly generated.

[0045] In this embodiment, generating a second instruction to control a non-target device includes the following steps: The smart home appliance currently located in the same room as the sound source and having a positive impact on the functions related to the trigger character is defined as the first appliance, and the smart home appliance having a negative impact is defined as the second appliance. The central control module generates a positive second instruction to control the first appliance and a negative second instruction to control the second appliance. The positive second instruction is used to control the functions of the first appliance related to the trigger character to increase the impact on the environment, and the negative second instruction is used to control the second appliance to reduce the impact on the environment.

[0046] Specifically, if the user issues a voice command saying "It's too noisy," the system will detect non-target devices located in the room, such as... Figure 2 As shown, if the target device is the speaker TC (meaning the speaker TC's volume is too loud and affects the user's TV viewing), and the non-target devices include the TV, the robot vacuum SR (meaning the robot vacuum SR is too loud and affects the user's TV viewing), then a positive second instruction is generated for the TV and the robot vacuum SR, and a negative second instruction is generated for the robot vacuum SR. Specifically, by increasing the volume of the TV, the user can hear the TV more easily, and by controlling the robot vacuum SR to close, external sounds are blocked, thus increasing the impact on the environment. By reducing the cleaning intensity of the robot vacuum SR, the noise it produces is reduced, thus reducing the impact on the environment. In addition, if the user repeatedly issues the "too noisy" command, a further instruction is generated to move the robot vacuum SR out of the room, thereby further reducing the impact on the living room environment.

[0047] In this embodiment, authenticating a mobile terminal includes the following steps: During authentication, the user is prompted to send an authentication voice message. The central control module confirms whether the user belongs to a member of the control network based on the tone and timbre of the authentication voice message. It also distinguishes between adult users and child users based on the tone and timbre of the voice message. Adult users have control permissions for all smart home appliances, while all child users have control permissions for non-kitchen smart home appliances.

[0048] In this embodiment, the central control module stores the names of various smart home appliances and the wake words used by family members for these appliances. If the central control module detects that the text information includes a wake word, it skips the process of judging the target device.

[0049] like Figure 3 As shown, the present invention provides a smart home appliance control system based on voice recognition, comprising: The central control module establishes a control network and connects mobile terminals and smart home appliances to the control network. The central control module also sends the received control voice to the cloud module and receives target information from the cloud module. If the target information does not include the name of the home appliance, it identifies the target information based on the current working mode of each smart home appliance to determine the target device among the smart home appliances. The target device is the smart home appliance that the control voice directly wants to control. The central control module generates a first instruction to control the target device and a second instruction to control the non-target device. The authentication module authenticates the mobile terminal after it connects to the control network. If the authentication is successful and the mobile terminal is the first device to connect to the control network, the voice modules located in each smart home appliance are activated. The voice module is installed in each smart home appliance. The voice module is used to receive control voice commands and send them to the central control module. After the cloud module converts the control voice into text information, it compares whether the text information converted from different control voices is the same. If they are the same, it selects one of the text information as the target information. If they are not the same, it determines the correct text information as the target information and sends the target information to the central control module.

[0050] The present invention provides a storage medium storing program instructions, wherein, when the program instructions are executed, the device where the storage medium is located is controlled to perform the above-described intelligent home appliance control method based on voice recognition.

[0051] The above embodiments merely illustrate preferred embodiments of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for controlling smart home appliances based on voice recognition, characterized in that, Includes the following steps: The central control module establishes a control network and connects mobile terminals and smart home appliances to the control network; When the mobile terminal accesses the control network, the authentication module authenticates the mobile terminal. If the authentication is successful, and the mobile terminal is the first device to access the control network, the voice module located in each of the smart home appliances is activated. When multiple smart home appliances receive control voice commands, they send their respective received control voice commands to the central control module. The central control module then sends the control voice commands to the cloud module. The cloud module converts the control voice commands into text information and uses a voting method to compare whether the text information converted from different control voice commands is the same. If they are the same, one of the text information is selected as the target information; otherwise, the correct text information is determined as the target information. The cloud module sends the target information to the central control module. If the target information does not include the name of the home appliance, the module identifies the need for the target information based on the current working mode of each smart home appliance to determine the target device among the smart home appliances. The target device is the smart home appliance that the control voice directly wants to control. The central control module generates a first instruction to control the target device and a second instruction to control non-target devices, and each smart home appliance executes a corresponding action based on the first instruction and the second instruction. The cloud module converts the control voice into the text information by the following steps: The central control module acquires the identifiers of each target device located in the control network, and obtains the type and working mode of the target device based on the identifiers, with each working mode corresponding to a pre-assigned identification code; The central control module generates an indoor map, which marks the location of the smart home appliances. After receiving the control voice, it locates the user's voice position and obtains the current working mode of the smart home appliances located in the same room as the voice position. Based on the identification code corresponding to the current working mode, it generates a noise sequence. The central control module sends the noise sequence and the control voice together to the cloud module. The cloud module is pre-set with multiple separation models, which are built based on neural networks. After receiving the control voice, the cloud module calls the corresponding separation model based on the noise sequence. After using the called separation model to separate the human voice track from the control voice, it calls the recognition model to convert the human voice track into the text information. Determining the target device among the smart home appliances includes the following steps: The central control module is equipped with a function table, which includes the function of each smart home appliance and multiple trigger characters corresponding to the function. The trigger characters include positive characters and negative characters. The trigger characters are extracted from the text information and matched with the corresponding function based on the function table to define the target function. The system acquires a smart home appliance that contains the target function and is in operation, and defines it as an initial target. If there is only one initial target, the initial target is directly set as the target device. If there are multiple initial targets, the system continues to acquire the current working mode of each initial target. The central control module pre-stores the contribution value of each working mode, acquires the contribution value of the current working mode, and determines the target device among the initial targets by combining the type of the extracted trigger character and the size of the contribution value. Generating the second instruction for control of a device other than the target device includes the following steps: The smart home appliance currently located in the same room as the sound source and having a positive impact on the functions related to the trigger character is defined as the first home appliance, and the smart home appliance having a negative impact is defined as the second home appliance. The central control module generates a positive second instruction for controlling the first home appliance and a negative second instruction for controlling the second home appliance. The positive second instruction is used to control the functions of the first home appliance related to the trigger character to increase the impact on the environment, and the negative second instruction is used to control the second home appliance to reduce the impact on the environment.

2. The intelligent home appliance control method based on voice recognition according to claim 1, characterized in that, The central control module generates the indoor map by the following steps: The location of the central control module is preset. The central control module generates virtual commands and sends the virtual commands to each of the smart home appliances so that the smart home appliances can perform virtual actions. The central control module and each of the smart home appliances acquire the sound generated by the virtual actions. Based on the volume of the sound, the positional relationship between each of the smart home appliances and the central control module, as well as the positional relationship between the smart home appliances, is determined. Based on the positional relationship, a first initial map is generated. If no designated device is connected within the control network, the central control module generates a notification message and sends it to the mobile terminal to remind the user to manually draw the second initial map. If the designated device is connected, the second initial map stored therein is retrieved from the designated device. The first initial map is combined with the second initial map to generate the indoor map.

3. The intelligent home appliance control method based on voice recognition according to claim 1, characterized in that, Authenticating the mobile terminal includes the following steps: During authentication, the user is prompted to issue an authentication voice. The central control module determines whether the user belongs to the control network based on the tone and timbre of the authentication voice. It also distinguishes between adult users and child users based on the tone and timbre of the voice. Adult users have control rights over all smart home appliances, while child users have control rights over non-kitchen smart home appliances.

4. The intelligent home appliance control method based on voice recognition according to claim 3, characterized in that, The central control module stores the names of various smart home appliances and the wake words that family members use for each smart home appliance. If the central control module recognizes that the text information includes the wake word, it skips the judgment process for the target device.

5. A voice recognition-based intelligent home appliance control system, used to implement the voice recognition-based intelligent home appliance control method as described in any one of claims 1-4, characterized in that, include: A central control module establishes a control network and connects mobile terminals and smart home appliances to the control network. The central control module also receives the control voice and sends it to the cloud module, and receives target information from the cloud module. If the target information does not include the name of the home appliance, the central control module identifies the needs of the target information based on the current working mode of each smart home appliance to determine the target device among the smart home appliances. The target device is the smart home appliance that the control voice directly wants to control. The central control module generates a first instruction to control the target device and a second instruction to control other than the target device. The authentication module authenticates the mobile terminal after it connects to the control network. If the authentication is successful and the mobile terminal is the first device to connect to the control network, the voice modules located in each of the smart home appliances are activated. A voice module is installed in each of the smart home appliances. The voice module is used to receive the control voice and send the control voice to the central control module. The cloud module converts the control voice into text information, compares whether the text information converted from different control voices is the same. If they are the same, it selects one of the text information as the target information; otherwise, it determines the correct text information as the target information and sends the target information to the central control module.

6. A computer storage medium, characterized in that, The computer storage medium stores program instructions, wherein, when the program instructions are executed, the device where the computer storage medium is located is controlled to execute the intelligent home appliance control method based on voice recognition as described in any one of claims 1-4.

Images