AI voice control method and system for intelligent household electrical appliance

By dividing smart home appliance groups and personnel permissions, combining voice recognition and control capability index, the problem of inaccurate recognition and error triggering of voice control system is solved, and the safety and convenience of smart home appliances are improved, especially for the use of the elderly and children. Security.

CN119028346APending Publication Date: 2024-11-26STATE GRID ZHEJIANG ELECTRIC POWER CO MARKETING SERVICE CENT +1
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202411369110.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2024-07-18
Filing Date
2024-09-29
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

When the existing voice control system deals with non-standard Mandarin or sound insulation problems, the voice command recognition is inaccurate and easily triggered by mistake, resulting in inconvenient use and safety hazards of smart home appliances, especially in families with young children or elderly people with limited mobility. middle.

Method used

By obtaining the functional information and personnel information of smart home appliances, it is divided into ordinary home appliance groups and risk home appliance groups, and calculates the control capability index, matches voice data for identification and classification processing, restricts usage permissions, and ensures that only authorized users can trigger device instructions, reduces Incorrectly trigger risk.

Benefits of technology

It improves the recognition accuracy and use safety of voice commands, reduces the risk of false triggering of smart home appliances, and enhances user convenience and privacy, especially for the elderly and children.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119028346A_ABST
    Figure CN119028346A_ABST
Patent Text Reader

Abstract

The invention discloses an AI voice control method and system for an intelligent household electrical appliance, and relates to the technical field of voice intelligent control. At present, a voice control system has the problems of inaccurate recognition, potential safety hazards of easily touching household appliances by mistake and the like. The method comprises the following steps: dividing intelligent household appliances into a common household appliance group and a risk household appliance group, and enabling the risk household appliance group to have use permission; dividing the users into an independent execution group and a risk execution group, and dividing use permissions for the users; therefore, the use safety is standardized; the voice information is improved, so that the accuracy of voice information recognition is improved; the use permission of the user to the home appliances in the risk home appliance group is limited through a preposed false triggering step; only authorized users can use the system, so that the risks of mistaken touch use, illegal invasion, privacy disclosure and the like are greatly reduced; according to the invention, misoperation of children or improper use among family members can be prevented, potential safety hazards are reduced, and accidents caused by use of intelligent household electrical appliances are avoided.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of voice intelligent control technology, and specifically to an AI voice control method and system for smart home appliances. Background Art

[0002] With the rapid development of science and technology and the improvement of people's living standards, smart home devices have gradually penetrated into people's daily lives, bringing great convenience and comfort to people. Among them, smart home appliances are an important part of smart homes, and their intelligence directly affects the user experience.

[0003] In terms of the control methods of smart home appliances, traditional methods such as buttons and remote controls can no longer meet users' needs for convenience and efficiency. Therefore, with the continuous development of artificial intelligence technology, AI voice control technology has gradually become a new trend in the control of smart home appliances. AI voice control technology uses technologies such as voice recognition and natural language processing to enable users to control smart home appliances through voice commands. This method is not only convenient and fast, but also can greatly improve the user experience.

[0004] In the Chinese invention application with application publication number CN113851130A, a universal intelligent device voice control method and system based on gateway devices are disclosed. The method includes initializing the newly connected intelligent devices and completing the device registration and service registration, and converting the user voice command into the corresponding device instruction when receiving it, each intelligent device parses the device instruction and feedbacks the device instruction that can be processed by the device, and selects and controls the intelligent device according to the feedback result; the system includes a network access management subsystem and a voice command processing subsystem; it can network and manage the intelligent devices of heterogeneous manufacturers, collect user instructions in a standardized manner and automatically distribute them to the corresponding intelligent devices according to the application layer interface protocol, so as to achieve unified management.

[0005] In combination with the above invention, the prior art still has the following deficiencies: The voice commands of existing voice control systems are generally based on Mandarin. When the user's Mandarin is not standard or there is a dialect problem, the voice command recognition is inaccurate; and in the case of poor sound insulation, it is easy to passively trigger voice commands, resulting in inconvenience in use; and with the increase in the types of smart home appliances, the use of some home appliances is dangerous, such as smart switches, sockets, smart kitchen appliances, smart curtains and window openers, smart water heaters, etc., especially for families with young children or elderly people with limited mobility. If dangerous home appliances are accidentally triggered, there will be great safety hazards. Summary of the invention

[0006] 1. Technical issues to be resolved In view of the shortcomings of the prior art, the present invention provides an AI voice control method and system for smart home appliances, which increases the safety of use by restricting the use rights of some users to smart home appliances, and solves the problems of inaccurate voice command recognition and high false triggering rate by identifying voice information and false triggering steps.

[0007] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: an AI voice control method for smart home appliances: comprising the following steps: Step 1: Obtain functional information of smart home appliances; connect smart home appliances to smart gateways respectively, and divide them into ordinary home appliance group and risk home appliance group according to functional information; Step 2: Obtain personnel information and calculate the personnel control ability index based on the personnel information ; Preset control capability index threshold , the personnel control ability index and controllability index threshold Compare and divide the personnel information into independent execution group and risk execution group according to the judgment results; Step 3: Obtain a voice data set, which includes multiple voice data packets matching the personnel information; perform voice recognition and classification on the data in the voice data set to form a personnel label matching the voice data set, and associate and match the personnel label with the personnel information of the independent execution group and the risk execution group respectively; Step 4: Obtain a voice command data set, and set voice commands according to the voice command data set; Step 5: Acquire voice information, identify the voice information, and determine whether the voice information contains voice commands; if it contains voice commands, perform false trigger identification to determine whether the device has the trigger permission; if it has the trigger permission, proceed to the next step; Step 6: Convert the voice command into the corresponding device command. If the person tag of the person who issues the command meets the conditions for enabling the device command, the device command will be sent to the corresponding smart home appliance. If the person tag of the person who issues the command does not meet the conditions for enabling the device command, a voice feedback prompt will be issued. Determine whether to continue executing the device command based on the voice feedback information.

[0008] In the preferred embodiment of the AI ​​voice control method for smart home appliances: in step 2, the personnel information includes at least age ,height and intellectual stage ; Preprocess the data of personnel information separately, and mark the results as , and ; Then substitute into the following formula:

[0009] in: is the control ability index; for The weight coefficient is between 0.2 and 0.4; for The weight coefficient is between 0.2 and 0.4; for The weight coefficient is between 0.3 and 0.5, and .

[0010] In the preferred embodiment of the AI ​​voice control method for smart home appliances: in step 2, if the control capability index is satisfied ≥Control Ability Index Threshold , the corresponding personnel information will be assigned to the independent execution group; if the control capability index is met <Control Ability Index Threshold , the corresponding personnel information will be assigned to the risk execution group.

[0011] In the preferred embodiment of the above-mentioned AI voice control method for smart home appliances: in step 3: The multiple voice data packets matching each person's information respectively correspond to voice data of different speaking speeds and different intonations; Preprocessing the voice data of the voice data packet; extracting the sound quality, sound length, sound intensity, pitch data and Mel cepstral coefficients from the processed voice signal, combining the sound quality, sound length, sound intensity, pitch data and Mel cepstral coefficients to form a multidimensional feature vector for describing the characteristics of the sound signal; The classifier is trained using the multidimensional feature vector 1 formed by different speech signals and the matching person labels, so that the classifier maps the multidimensional feature vector 1 to the corresponding person label.

[0012] In the preferred embodiment of the AI ​​voice control method for smart home appliances, in step 4, the specific method of setting the voice command is: According to the voice command setting prompts, in a noise-free state and a background noise state, the content of the voice command prompts is read aloud using multiple speaking speeds, multiple intonations, Mandarin and dialects to generate a voice command data set; the data of the voice command data set are preprocessed and converted into multiple voice feature sequences one; then the multiple voice feature sequences one are converted into multiple text sequences one, and the multiple text sequences one are associated with corresponding voice commands.

[0013] In the preferred embodiment of the above-mentioned AI voice control method for smart home appliances: in step 4, the voice commands include single commands, all-in-one commands and compound commands; a single command is only used to match one function of a smart home appliance; an all-in-one command can be used to match multiple functions of a smart home appliance; a compound command can be used to match several functions of multiple smart home appliances.

[0014] In the preferred embodiment of the AI ​​voice control method for smart home appliances, in step 5, the voice information is preprocessed and converted into a voice feature sequence 2, and then converted into a text sequence 2 according to the voice feature sequence 2, and the text sequence 2 is compared with multiple text sequences 1. If the text sequence 2 contains the content of one of the text sequences 1, it is considered to be the same as the included text sequence 1 and is judged to be a voice command; if it does not contain the content of any of the text sequences 1, it is judged not to be a voice command. When the voice information meets the condition that it is a voice command, false trigger identification is performed, specifically: Extract the multidimensional feature vector 2 of the voice information, input the multidimensional feature vector 2 into the classifier trained in step 3, and map it to the personnel label; if there is a matching personnel label, mark it as having trigger authority and proceed to the next step; if there is no matching personnel label, mark it as a false trigger action and terminate the program.

[0015] In the preferred embodiment of the AI ​​voice control method for smart home appliances, in step 6, the method for determining the activation condition of the device command is: The voice command is converted into the corresponding device command, and it is determined whether the smart home appliance executing the device command belongs to the ordinary home appliance group or the risky home appliance group; if it belongs to the ordinary home appliance group, the device command is sent to the smart home appliance that matches it; if it belongs to the risky home appliance group, it is further determined whether the personnel tag with triggering authority in step 5 belongs to the independent execution group or the risky execution group. If it belongs to the independent execution group, the device command is sent to the smart home appliance that matches it; if it belongs to the risky execution group, a voice feedback prompt is issued; if this voice command is a compound command, and some of the smart home appliances that execute this compound command belong to the risky home appliance group, a voice feedback prompt is issued.

[0016] In the preferred embodiment of the AI ​​voice control method for smart home appliances, the specific method for determining whether to continue to execute the device command according to the voice feedback information is: Extract the multidimensional feature vector three of the voice feedback information, input the multidimensional feature vector three into the classifier trained in step 3, and map it to the personnel label; if there is a matching personnel label, determine whether the personnel label belongs to the independent execution group or the risk execution group; if it belongs to the independent execution group, determine whether to continue to execute the device instruction based on the voice feedback information; if it belongs to the risk execution group or there is no matching personnel label for feature vector three, do not refer to the voice feedback information and terminate the program.

[0017] An AI voice control system for smart home appliances, used to implement the above-mentioned smart management method: including: Data analysis module 1 is used to obtain functional information of smart home appliances; connect smart home appliances to smart gateways respectively, and divide them into ordinary home appliance group and risk home appliance group according to the functional information; Data analysis module 2 is used to obtain personnel information and calculate the personnel control ability index based on the personnel information ; Preset control capability index threshold , the personnel control ability index and controllability index threshold Compare and divide the personnel information into independent execution group and risk execution group according to the judgment results; The speech recognition module is used to obtain a speech data set, perform speech recognition and classification processing on the data in the speech data set, form a personnel label that matches the speech data set, and associate and match the personnel label with the personnel information of the independent execution group and the risk execution group respectively; A voice command setting module, used to obtain a voice command data set and set a voice command according to the voice command data set; The anti-mistouch module is used to obtain voice information, identify the voice information, and determine whether the voice information contains voice commands; if it contains voice commands, it will perform mistouch identification and determine whether it has the trigger authority; if it has the trigger authority, it will proceed to the next step; The execution authority judgment module can convert voice commands into corresponding device commands. If the personnel tag of the person who issues the command meets the conditions for enabling the device command, the device command will be sent to the corresponding smart home appliance. If the personnel tag of the person who issues the command does not meet the conditions for enabling the device command, a voice feedback prompt will be issued. The voice feedback information will be used to determine whether to continue executing the device command.

[0018] (III) Beneficial effects The present invention provides an AI voice control method and system for smart home appliances, which has the following beneficial effects: (1) By connecting several smart home appliances to a smart gateway, multiple smart home appliances can be controlled through a unified interface or application. Users no longer need to control each device separately, but can centrally manage all smart home appliances through one device or application, which greatly improves the convenience of use. At the same time, the smart gateway can support new devices and functions through software upgrades or adding new hardware devices to ensure the continuous development and upgrading of the smart home system. By setting the permissions of ordinary home appliance groups and risky home appliance groups, it is conducive to independent management and improved security. (2) By limiting the user's access to appliances in the risky appliance group, it ensures that only authorized users can use them, thereby greatly reducing the risks of accidental use, illegal intrusion, and privacy leakage. It can prevent children from operating them by mistake or family members from using them improperly, reduce safety hazards, and avoid accidents caused by the use of smart appliances. (3) By performing voice recognition on each personnel tag, each member can be matched with the personnel information of the independent execution group and the risk execution group. When the voice information is detected to correspond to a certain personnel tag, it is matched with the personnel information of the independent execution group or the risk execution group, thereby realizing the addition of permissions; (4) By processing the voice data of family members corresponding to different personal labels separately and forming associations with voice commands respectively, it is possible to avoid the problem of inaccurate command recognition caused by accents, intonations, or dialects of different users, thereby improving the convenience of use; by customizing the voice commands, a single command can be used to control the program actions of multiple smart home appliances, and the operation is easier, which is more friendly to the elderly, children or people with limited mobility. The voice commands set according to usage habits can provide services to users more intelligently, improving the comfort and convenience of home life; (5) Recognize language information, filter language commands, improve accuracy, and through further false trigger identification, ensure that only the user who enters the personnel information has the right to use it, which greatly avoids the problem of other people triggering voice commands due to poor sound insulation and other problems, improves privacy, and also improves security; (6) By setting usage permissions for smart home appliances, it is possible to effectively prevent users who are not capable of using smart home appliances from using them incorrectly, thereby reducing safety hazards for users, and helping to prevent accidents and improve safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of the steps of the AI ​​voice control method for smart home appliances of the present invention; Figure 2 A schematic diagram of the connection between the smart gateway and the smart home appliance in the AI ​​voice control method for smart home appliances of the present invention; Figure 3 This is a flowchart of the instruction execution of a common home appliance group in the AI ​​voice control method for smart home appliances of the present invention; Figure 4 This is a flow chart of language command recognition in the AI ​​voice control method for smart home appliances of the present invention; Figure 5 This is a flow chart of false trigger identification in the AI ​​voice control method for smart home appliances of the present invention; Figure 6 A flowchart of personnel division in the AI ​​voice control method for smart home appliances of the present invention; Figure 7 The present invention is a flow chart of voice feedback in the AI ​​voice control method for smart home appliances. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] See also Figure 1-7 The present invention provides an AI voice control method for smart home appliances, comprising the following steps: Step 1: Obtain functional information of smart home appliances; connect smart home appliances to smart gateways respectively, and divide them into ordinary home appliance group and risk home appliance group according to functional information; Step one includes the following steps: Step 101: Acquire function information of a smart home appliance. For example, the smart home appliance is a smart air conditioner, which has a start function, a stop function, a temperature adjustment function, a mode adjustment function, etc.; It should be noted that smart home appliances may include smart air conditioners, smart refrigerators, smart lamps, smart switches, sockets, smart curtains and window openers, and smart water heaters.

[0022] Step 102: Connect the smart home appliance to the smart gateway; It should be noted that the smart gateway selected here can be an existing product with voice control function and compatible with smart home appliances; when using it, the smart home appliances need to be added to the management APP of the smart gateway; Step 103: According to the functional information of smart home appliances, the smart home appliances are divided into a common home appliance group and a risky home appliance group; for example, smart lamps have the functions of controlling light switches and adjusting light brightness, and smart air conditioners have the functions of controlling air conditioner switches, adjusting operation modes, and adjusting air temperature; such functions do not pose actual safety hazards to children and the elderly, and can be divided into the common home appliance group; smart curtain machine window openers can control the function of opening and closing windows in a house, and smart water boilers have the function of boiling water, etc. When a child plays by the window and accidentally opens the window, there is a risk of falling; when playing by the water boiler and turning on the water boiling function, there is a risk of the child touching the hot water, etc., so such smart home appliances are divided into the risky home appliance group; When using, combine the contents of step 101 to step 103: Connecting several smart home appliances to the smart gateway allows multiple smart home appliances to be controlled through a unified interface or application. This means that users no longer need to control each device separately, but can centrally manage all smart home appliances through one device or application, which greatly improves the convenience of use; users can also remotely access and control smart home appliances at home; no matter where the user is, as long as there is a network connection, they can remotely monitor and control the home appliances through devices such as mobile phones, tablets or computers. This remote access function makes the smart home system more flexible and convenient; automation and scene setting functions can also be achieved through the smart gateway; users can customize different scenes, such as "away mode" and "sleep mode", and set different home appliance states in these scenes; when the user switches to a certain scene, the smart gateway will automatically adjust the working status of the relevant equipment to bring users a more intelligent life experience; the smart gateway has good scalability. With the continuous development of smart home technology, new smart home appliances and functions are constantly emerging; at the same time, the smart gateway can support new devices and functions through software upgrades or adding new hardware devices to ensure the continuous development and upgrading of the smart home system; and by setting the permissions of ordinary home appliance groups and risk home appliance groups, it is conducive to independent management and improved security.

[0023] Step 2: Obtain personnel information and calculate the personnel control ability index based on the personnel information ; Preset control capability index threshold , the personnel control ability index and controllability index threshold Compare and divide the personnel information into independent execution group and risk execution group according to the judgment results; Step 2 includes the following: Step 201: Obtain the personal information of family members, which at least includes age. ,height and intellectual stage wait; It should be noted that the intellectual stage The intelligence level of specific users can be set to the 0-14 years old stage, the 15-20 years old stage, and the stage above 20 years old; the way professional institutions test the management level can be used as a reference.

[0024] Step 202: Change the age ,height and intellectual stage The data were normalized and preprocessed, and the results were marked as , and ; Then substitute into the following formula:

[0025] in: is the control ability index; for The weight coefficient of is between 0.2 and 0.4, and here it is 0.3; for The weight coefficient of is between 0.2 and 0.4, and here it is 0.3; for The weight coefficient of is between 0.3 and 0.5, and here it is 0.4; and ; , and The value can be adjusted according to the actual situation of family members or the passage of time.

[0026] Step 203: Preset the control capability index threshold ; Personnel control ability index and controllability index threshold Compare; if the control ability index is met ≥Control Ability Index Threshold , the corresponding personnel information will be assigned to the independent execution group; if the control capability index is met <Control Ability Index Threshold , the corresponding personnel information will be assigned to the risk execution group; It should be noted that the control capability index threshold The specific method can be: collecting event information of injuries or accidents of different levels caused by smart home appliances or ordinary home appliances; obtaining information such as age, height and intelligence stage of the injured people; analyzing and counting the proportion of people of different ages, heights and intelligence stages in the total number of people in the incident, and selecting the personnel information of the people with the largest proportion as the estimated basic parameters; normalizing the basic parameters separately; then determining the weight coefficients of the three basic parameters according to the frequency of events caused by age, height and intelligence stage in the event statistics; multiplying the corresponding values ​​of age, height and intelligence stage after normalization by the corresponding weight coefficients, and then adding the three products to obtain the estimated control ability index threshold ; Calculate the control capability index threshold When calculating the control ability index threshold, the user can adjust the control ability index threshold according to the specific situation of the user. The parameters can be selectively adjusted to make them more targeted. For example, if more attention is paid to children's risks, the proportion of age parameters can be adjusted.

[0027] By combining the contents of step 201 to step 203: By limiting the user's access to appliances in the risky appliance group and ensuring that only authorized users can use them, the risks of accidental use, illegal intrusion, privacy leakage, etc. are greatly reduced; it can prevent children from operating them by mistake or improper use among family members, reduce safety hazards, and avoid accidents caused by the use of smart appliances.

[0028] Step 3: Obtain a voice data set, which includes multiple voice data packets matching the personnel information; perform voice recognition and classification on the data in the voice data set to form a personnel label matching the voice data set, and associate and match the personnel label with the personnel information of the independent execution group and the risk execution group respectively; Step 3 includes the following steps: Step 301: Acquire a voice data set of each family member, wherein the voice data set matching the personal information of each family member includes multiple voice data packets; the multiple voice data packets may include voice data of different speech speeds, voice data of different intonations, etc.; Step 302: preprocessing the voice data of the voice data packet, wherein the preprocessing may perform noise reduction, reverberation removal, framing and windowing processing on the voice data; extracting sound quality, sound length, sound intensity, pitch data and Mel cepstral coefficients from the processed voice signal, combining the sound quality, sound length, sound intensity, pitch data and Mel cepstral coefficients to form a multidimensional feature vector 1 for describing the characteristics of the sound signal; It should be noted that, through the above preprocessing and extraction of sound signal features, the data packets in the voice data sets of different family members can form a multi-dimensional feature vector 1 matching different family members, which is used for further distinguishing personnel information in subsequent steps.

[0029] Step 303: training a classifier using the multidimensional feature vector 1 formed by different speech signals and the matching person label, so that the classifier maps the multidimensional feature vector 1 to the corresponding person label; Specifically, the classifier uses machine learning algorithms, such as Gaussian mixture model GMM, support vector machine SVM, neural network NN, etc., to model the extracted features. During the training process, the classifier will learn how to map the multi-dimensional feature vector to the corresponding person label to describe the distribution of speech features of different speakers and realize the recognition of person labels; Step 304: Match the multiple personnel tags with the personnel information of the independent execution group and the risk execution group respectively to form a link; By combining the contents of step 301 to step 304: By performing voice recognition on each personnel tag, each member can be matched with the personnel information of the independent execution group and the risk execution group. When it is detected that the voice information corresponds to a certain personnel tag, it is matched with the personnel information of the independent execution group or the risk execution group, thereby realizing the addition of permissions.

[0030] Step 4: Get the voice command dataset and set the voice command according to the voice command dataset: Step 4 mainly includes the following steps: Step 401: Acquire voice command data sets of different family members; specifically, according to the voice command setting prompt of the intelligent gateway, in a noise-free state and a background noise state, the family members corresponding to each person tag read the content of the voice command prompt in the form of multiple speech speeds, multiple intonations, Mandarin and dialects to generate a voice command data set; Step 402: pre-processing the data of the voice command data set respectively and converting them into a plurality of voice feature sequences 1; then converting the plurality of voice feature sequences 1 into a plurality of text sequences 1, and associating the plurality of text sequences 1 with corresponding voice commands; In step 402, an acoustic model is built through a convolutional neural network and trained through a voice command data set. After the training is completed, the received voice information can be converted into a voice feature sequence one; the language model is continued to be used to convert the voice feature sequence one converted by the acoustic model into a text sequence one; then the text sequence one is received through a machine learning model or a classifier, and the text sequence one is associated with the voice command; building an acoustic model, a language model, and its learning model and classifier, etc. are all existing technologies.

[0031] In step 402: the voice command includes a single command, an all-in-one command and a compound command; a single command is only used to match one function of a smart home appliance; an all-in-one command can be used to match multiple functions of a smart home appliance; a compound command can be used to match several functions of multiple smart home appliances; specifically, a single command of a smart air conditioner can be set to "turn on the air conditioner", "turn off the air conditioner", "adjust the air conditioner to cooling", "adjust the air conditioner to 25°C", etc., and the air conditioner performs the corresponding function; the all-in-one command can be set to "adjust the air conditioner to 25°C in summer", corresponding to the smart air conditioner's "turn on function, adjust cooling mode, adjust temperature to 25°C"; set to "adjust the air conditioner to 28°C in winter" corresponding to the smart air conditioner's "turn on function, adjust heating mode, adjust temperature to 28°C", etc.; the compound command can be set to "I'm home", corresponding to "turn on smart lamps, turn on smart air conditioners, turn on fresh air fans to purify air, turn on water heaters", set to "I'm going to work" corresponding to "turn off smart lamps, turn off smart air conditioners, turn on fresh air fans to purify air, turn on water heaters", etc.; the voice command can be customized according to user needs to improve the comfort of use.

[0032] By combining the contents of step 401 to step 402: By processing the voice data of family members corresponding to different personal labels separately and forming associations with voice commands respectively, the problem of inaccurate command recognition due to accents, intonations, or dialects of different users can be avoided, thereby improving the convenience of use; by customizing the voice commands, a single command can be used to control the program actions of multiple smart home appliances, and the operation is less difficult, which is more friendly to the elderly, children or people with mobility difficulties. Voice commands set according to usage habits can provide services to users more intelligently and improve the comfort and convenience of home life.

[0033] Step 5: Get the voice information, identify the voice information, and determine whether the voice information contains voice commands; if it contains voice commands, perform false trigger identification to determine whether the trigger permission is available: Step 5 mainly includes the following steps: Step 501: Acquire the operator's voice information; Step 502: performing conventional language preprocessing on the language information, such as noise removal, voice endpoint detection, voice segmentation and channel conversion, and pre-emphasis; Step 503: pre-process the processed voice information in the same manner as in step 402, and convert it into voice feature sequence 2, and then convert it into text sequence 2 based on the voice feature sequence 2, and compare the text sequence 2 with multiple text sequences 1. If the text sequence 2 contains the content of one of the text sequences 1, it is considered to be equal to the included text sequence 1 and is determined to be a voice command; if it does not contain the content of any of the text sequences 1, it is determined not to be a voice command; Step 504: When the voice information is a voice command, false trigger identification is performed; specifically: Extract the multidimensional feature vector 2 of the voice information by the method of step 3, input the multidimensional feature vector 2 into the classifier trained in step 3, and map it to the personnel label; if there is a matching personnel label, it is marked as having the trigger authority and proceeding to the next step; if there is no matching personnel label, it is marked as a false trigger action and the program is terminated; By combining the contents of step 501 to step 504: First, the language information is recognized and the language commands are screened to improve accuracy. Through further false trigger identification, it can ensure that only the user who enters the personnel information has the right to use the system. This greatly avoids the problem of other people triggering voice commands due to poor sound insulation, thereby improving privacy and security.

[0034] Step 6: Convert the voice command into the corresponding device command. If the person tag of the person who issues the command meets the conditions for opening the device command, the device command is sent to the corresponding smart home appliance. If the person tag of the person who issues the command does not meet the conditions for opening the device command, a voice feedback prompt is issued. Determine whether to continue to execute the device command based on the voice feedback information. Step 6 mainly includes the following steps: Step 601: The intelligent gateway converts the voice command into a device command; Step 602: Determine whether the conditions for opening the device command are met, specifically: convert the voice command into the corresponding device command, and determine whether the smart home appliance executing the device command belongs to the ordinary home appliance group or the risky home appliance group; if it belongs to the ordinary home appliance group, send the device command to the smart home appliance that matches it; if it belongs to the risky home appliance group, further determine whether the personnel tag with triggering authority in step 5 belongs to the independent execution group or the risky execution group, if it belongs to the independent execution group, send the device command to the smart home appliance that matches it, if it belongs to the risky execution group, issue a voice feedback prompt; if the voice command is a compound command, and some of the smart home appliances that execute the compound command belong to the risky home appliance group, issue a voice feedback prompt; Step 602: After the voice feedback prompt is issued, determine whether to continue to execute the device instruction according to the voice feedback information; the specific method is: According to the method of step 3, extract the multidimensional feature vector three of the voice feedback information, input the multidimensional feature vector three into the classifier trained in step 3, and map it to the personnel label; if there is a matching personnel label, continue to judge whether the personnel label belongs to the independent execution group or the risk execution group; if it belongs to the independent execution group, judge whether to continue to execute the device instruction according to the voice feedback information; if it belongs to the risk execution group or there is no matching personnel label for the feature vector three, do not refer to the voice feedback information and terminate the program; In step 602, the voice feedback information can be set to express affirmative language, such as "agree to open" or negative language "disagree to open" to assist in determining whether to continue executing the program action.

[0035] By combining the contents of step 601 to step 602: By adding a feedback mechanism, people with permission can supervise unauthorized users, which helps to further reduce usage risks and improve usage security.

[0036] On the other hand, the present invention also discloses an AI voice control system for smart home appliances, which is used to implement the above-mentioned smart management method, including: Data analysis module 1 is used to obtain functional information of smart home appliances; connect smart home appliances to smart gateways respectively, and divide them into ordinary home appliance group and risk home appliance group according to the functional information; Data analysis module 2 is used to obtain personnel information and calculate the personnel control ability index based on the personnel information ; Preset control capability index threshold , the personnel control ability index and controllability index threshold Compare and divide the personnel information into independent execution group and risk execution group according to the judgment results; The speech recognition module is used to obtain a speech data set, perform speech recognition and classification processing on the data in the speech data set, form a personnel label that matches the speech data set, and associate and match the personnel label with the personnel information of the independent execution group and the risk execution group respectively; A voice command setting module, used to obtain a voice command data set and set a voice command according to the voice command data set; The anti-mistouch module is used to obtain voice information, identify the voice information, and determine whether the voice information contains voice commands; if it contains voice commands, it will perform mistouch identification and determine whether it has the trigger authority; if it has the trigger authority, it will proceed to the next step; The execution authority judgment module can convert voice commands into corresponding device commands. If the personnel tag of the person who issues the command meets the conditions for enabling the device command, the device command will be sent to the corresponding smart home appliance. If the personnel tag of the person who issues the command does not meet the conditions for enabling the device command, a voice feedback prompt will be issued. The voice feedback information will be used to determine whether to continue executing the device command.

[0037] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware or in combination with computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0038] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0039] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. An AI voice control method for smart home appliances, characterized in that: The following steps are involved: Step 1: Get the function information of smart home appliances; Connect smart home appliances to smart gateways respectively and divide them into ordinary home appliance group and risk home appliance group according to their functional information; Step 2: Obtain personnel information and calculate the personnel control ability index based on the personnel information ; Preset control capability index threshold , the personnel control ability index and controllability index threshold Compare and divide the personnel information into independent execution group and risk execution group according to the judgment results; Step 3: Obtain a voice data set, which includes multiple voice data packets matching the personnel information; perform voice recognition and classification on the data in the voice data set to form a personnel label matching the voice data set, and associate and match the personnel label with the personnel information of the independent execution group and the risk execution group respectively; Step 4: Obtain a voice command data set, and set voice commands according to the voice command data set; Step 5: Acquire voice information, identify the voice information, and determine whether the voice information contains voice commands; if it contains voice commands, perform false trigger identification to determine whether the device has the trigger permission; if it has the trigger permission, proceed to the next step; Step 6: Convert the voice command into a corresponding device command. If the person tag of the person who issues the command meets the conditions for opening the device command, the device command is sent to the corresponding smart home appliance. If the person tag of the person who issues the command does not meet the conditions for opening the device command, a voice feedback prompt is issued. Determine whether to continue executing device instructions based on the voice feedback information.

2. The AI ​​voice control method for smart home appliances according to claim 1, characterized in that: In step 2, the personnel information includes at least age ,height and intellectual stage ; The data of personnel information are preprocessed separately, and the results are marked as , and ; Then substitute into the following formula: in: is the control ability index; for The weight coefficient is between 0.2 and 0.4; for The weight coefficient is between 0.2 and 0.4; for The weight coefficient is between 0.3 and 0.5, and .

3. The AI ​​voice control method for smart home appliances according to claim 2, characterized in that: In step 2, if the control capability index is satisfied ≥Control Ability Index Threshold , the corresponding personnel information will be assigned to the independent execution group; if the control capability index is met <Control Ability Index Threshold , the corresponding personnel information will be assigned to the risk execution group.

4. The AI ​​voice control method for smart home appliances according to claim 3, characterized in that: In step 3: The multiple voice data packets matching each person's information respectively correspond to voice data of different speaking speeds and different intonations; Preprocessing the voice data of the voice data packet; extracting the sound quality, sound length, sound intensity, pitch data and Mel cepstral coefficients from the processed voice signal, combining the sound quality, sound length, sound intensity, pitch data and Mel cepstral coefficients to form a multidimensional feature vector for describing the characteristics of the sound signal; The classifier is trained using the multidimensional feature vector 1 formed by different speech signals and the matching person labels, so that the classifier maps the multidimensional feature vector 1 to the corresponding person label.

5. The AI ​​voice control method for smart home appliances according to claim 4, characterized in that: In step 4, the specific method of setting the voice command is as follows: According to the voice command setting prompts, in a noise-free state and a background noise state, the content of the voice command prompts is read aloud using multiple speaking speeds, multiple intonations, Mandarin and dialects to generate a voice command data set; the data of the voice command data set are preprocessed and converted into multiple voice feature sequences one; then the multiple voice feature sequences one are converted into multiple text sequences one, and the multiple text sequences one are associated with corresponding voice commands.

6. The AI ​​voice control method for smart home appliances according to claim 5, characterized in that: In step 4, the voice commands include single commands, all-in-one commands and compound commands; a single command is only used to match one function of a smart home appliance; an all-in-one command can be used to match multiple functions of a smart home appliance; a compound command can be used to match several functions of multiple smart home appliances.

7. The AI ​​voice control method for smart home appliances according to claim 6, characterized in that: In step 5, the voice information is preprocessed and converted into a voice feature sequence 2, and then converted into a text sequence 2 based on the voice feature sequence 2. The text sequence 2 is compared with multiple text sequences 1. If the text sequence 2 contains the content of one of the text sequences 1, it is considered to be the same as the included text sequence 1 and is judged to be a voice command; if it does not contain the content of any of the text sequences 1, it is judged not to be a voice command; When the voice information meets the condition that it is a voice command, false trigger identification is performed, specifically: Extract the multidimensional feature vector 2 of the voice information, input the multidimensional feature vector 2 into the classifier trained in step 3, and map it to the personnel label; if there is a matching personnel label, mark it as having trigger authority and proceed to the next step; if there is no matching personnel label, mark it as a false trigger action and terminate the program.

8. The AI ​​voice control method for smart home appliances according to claim 7, characterized in that: In step 6, the method for determining the start condition of the device instruction is: Convert the voice command into the corresponding device command, and determine whether the smart home appliance executing the device command belongs to the ordinary home appliance group or the risky home appliance group; if it belongs to the ordinary home appliance group, send the device command to the matching smart home appliance; If it belongs to the risky appliance group, it is further determined whether the personnel tag with triggering authority in step 5 belongs to the independent execution group or the risky execution group. If it belongs to the independent execution group, the device command is sent to the smart appliance that matches it. If it belongs to the risky execution group, a voice feedback prompt is issued; if this voice command is a compound command, and some of the smart appliances that execute this compound command belong to the risky appliance group, a voice feedback prompt is issued.

9. The AI ​​voice control method for smart home appliances according to claim 8, characterized in that: The specific method for determining whether to continue executing device instructions based on voice feedback information is as follows: Extract the multidimensional feature vector three of the voice feedback information, input the multidimensional feature vector three into the classifier trained in step 3, and map it to the personnel label; if there is a matching personnel label, determine whether the personnel label belongs to the independent execution group or the risk execution group; if it belongs to the independent execution group, determine whether to continue to execute the device instruction based on the voice feedback information; if it belongs to the risk execution group or there is no matching personnel label for feature vector three, do not refer to the voice feedback information and terminate the program.

10. An AI voice control system for smart home appliances, used to implement an AI voice control method for smart home appliances according to any one of claims 1 to 9, characterized in that: include: Data analysis module 1, used to obtain functional information of smart home appliances; Connect smart home appliances to smart gateways respectively and divide them into ordinary home appliance group and risk home appliance group according to their functional information; Data analysis module 2 is used to obtain personnel information and calculate the personnel control ability index based on the personnel information ; Preset control capability index threshold , the personnel control ability index and controllability index threshold Compare and divide the personnel information into independent execution group and risk execution group according to the judgment results; The speech recognition module is used to obtain a speech data set, perform speech recognition and classification processing on the data in the speech data set, form a personnel label that matches the speech data set, and associate and match the personnel label with the personnel information of the independent execution group and the risk execution group respectively; A voice command setting module, used to obtain a voice command data set and set a voice command according to the voice command data set; The anti-mistouch module is used to obtain voice information, identify the voice information, and determine whether the voice information contains voice commands; if it contains voice commands, it will perform mistouch identification and determine whether it has the trigger authority; if it has the trigger authority, it will proceed to the next step; The execution authority judgment module can convert voice commands into corresponding device commands. If the person tag of the person who issues the command meets the conditions for opening the device command, the device command will be sent to the corresponding smart home appliance; if the person tag of the person who issues the command does not meet the conditions for opening the device command, a voice feedback prompt will be issued; Determine whether to continue executing device instructions based on the voice feedback information.

Citation Information

Cited By

  • Numerical control machine tool interaction control method and system based on voice recognition

    CN119559935A

  • Internet of Things equipment control method based on large language model

    CN119851667A

  • A method for controlling IoT devices based on large language model

    CN119851667B