Intelligent socket control method and system

Abstract

The invention relates to an intelligent socket control method and system, and relates to the field of intelligent household electrical appliance technology, and the method comprises the steps: obtaining an external input audio of a current user; defining similar users according to the external input audio, and acquiring a registration recognition library according to the similar users; determining instruction texts and audio texts in a registration recognition library, constructing an overall instruction set according to all the instruction texts, and combining all the audio texts corresponding to the same instruction text to construct a single audio set; performing analysis according to each audio text to determine an audio representative parameter, and determining a single representative audio according to the audio representative parameter; constructing an initial recognition library of the current user according to the instruction text in the overall instruction set and the corresponding single representative audio; and controlling the socket to work after the initial identification library is constructed. The intelligent wall socket has the effect of improving the universality and user satisfaction of the intelligent wall socket.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent household appliance technology, in particular to an intelligent socket control method and system. BACKGROUND

[0002] With the rapid development of smart home technology, intelligent wall socket, as one of the core devices for household power control, has multiple intelligent functions such as remote control, voice interaction, and power statistics. Among them, voice control technology enables users to directly control the on-off of the socket through voice assistants, greatly improving the convenience of operation.

[0003] At present, some intelligent sockets have supported dialect recognition function to adapt to the local language habits of users. This kind of function usually relies on the pre-set dialect library built by manufacturers before leaving the factory, and analyzes specific dialects through cloud or local voice recognition engine, so as to realize voice control.

[0004] In the above related technology, due to the large number of dialects in China, the pronunciation, vocabulary and grammar of different regions are quite different, and it is difficult for manufacturers to completely cover all dialects. Therefore, users in regions not included in the dialect library cannot use dialect voice control, which limits the voice interaction experience and affects the universality and user satisfaction of the product. SUMMARY

[0005] In order to improve the universality and user satisfaction of intelligent wall socket, the present application provides an intelligent socket control method and system.

[0006] In the first aspect, the present application provides an intelligent socket control method, which adopts the following technical solution: An intelligent socket control method, comprising: Obtaining external input audio of a current user under a pre-set instruction question; Defining the user who is consistent with the external input audio as a same type user, and obtaining a registration recognition library according to the same type user; Determining instruction texts and audio texts in the registration recognition library, and constructing an overall instruction set according to all instruction texts, and combining all audio texts corresponding to the same instruction text to construct a single audio set; Analyzing each audio text in the single audio set to determine an audio representative parameter, and determining a single representative audio corresponding to the audio text of the largest audio representative parameter; Pairing and combining the instruction texts in the overall instruction set and the corresponding single representative audio to construct an initial recognition library of the current user; Real-time acquisition of demand control audio after initial identification library construction to input to the initial identification library for identification analysis to determine control instructions, and control the socket operation according to the control instructions.

[0007] After the same user is determined, the intelligent socket control method further comprises: Counting the same users to determine the same quantity; Determine whether the same quantity is greater than the preset reference quantity; If the same quantity is greater than the reference quantity, determine the initial identification library according to the current same user; If the same quantity is not greater than the reference quantity, determine the overall instruction set according to the current same user, and determine whether the overall instruction set contains the preset reference instruction set; If the overall instruction set contains the reference instruction set, determine the initial identification library according to the current same user; If the overall instruction set does not contain the reference instruction set, compare the overall instruction set and the reference instruction set to determine the missing instruction, and generate an instruction question according to the missing instruction and output it to wait for external input audio.

[0008] If the same quantity is greater than the reference quantity, the intelligent socket control method further comprises: When the same quantity is greater than the preset redundancy quantity, acquire the effective installation position of the current user and the comparison installation position of the same user; Draw an influence coverage range with a width of a preset influence distance centered on the effective installation position, and calculate the range same user proportion according to the same user and all users in the influence coverage range; Determine whether the range same user proportion is greater than the preset aggregation proportion; If the range same user proportion is not greater than the aggregation proportion, output a user independent signal; If the range same user proportion is greater than the aggregation proportion, define the same user whose comparison installation position is outside the current influence coverage range as canceled.

[0009] After the user independent signal is output, the intelligent socket control method further comprises: Randomly select one same user and draw a similar coverage range with a width of a preset similar distance according to the corresponding comparison installation position; Determine the similar same user proportion according to the similar coverage range, and define the same user corresponding to the similar same user proportion greater than the aggregation proportion as an aggregation user; Induce the aggregation user whose same user is overlapped in at least two similar coverage ranges to combine to form a user aggregation combination; Combine the same users contained in each similar coverage range in the user aggregation combination to construct a regional user combination; Counting the same kind of users in the regional user combination to determine the number of regions, and defining the same kind of users other than the same kind of users in the maximum number of regional user combinations as cancellation.

[0010] The step of analyzing each audio text in the single audio set to determine the audio representative parameter includes: Counting the same audio text under each audio text to determine the number of audio same; Counting all audio texts in the single audio set to determine the number of set audio; Calculating the number of set audio and the number of audio same to determine the audio representative parameter.

[0011] After the audio representative parameter is determined, the smart socket control method further includes: Defining the audio text with an audio representative parameter greater than the preset effective representative parameter as an effective audio; Judging whether the effective audio is unique; If the effective audio is unique, determining the single representative audio according to the audio representative parameter; If the effective audio is not unique, generating an instruction question according to the instruction text corresponding to the current single audio set and outputting to wait for external input audio.

[0012] After the initial recognition library is constructed, the smart socket control method further includes: Obtaining the job feedback state after controlling the socket job according to the control instruction; Outputting a control error instruction when the job feedback state is consistent with the preset error job state; Calculating the instruction error proportion according to the control error instruction and the control instruction; When the instruction error proportion is greater than the preset permitted error proportion and the number of control instructions is greater than the preset standard number, combining the instruction text and the audio text corresponding to the case when the job feedback state is consistent with the preset correct job state to generate a qualified combination; Defining the audio text corresponding to the instruction text in the qualified combination as the external input audio to redetermine the same kind of users, and redetermining the initial recognition library according to the same kind of users.

[0013] In a second aspect, the application provides a smart socket control system, which adopts the following technical solution: A smart socket control system, comprising: An acquisition module for acquiring external input audio of a current user under a preset instruction question; A processing module connected with the acquisition module for information storage and processing; The processing module defines a user who is consistent with the external input audio with the current external input audio as a same type user, and causes the acquisition module to acquire the registration recognition library according to the same type user; The processing module determines the instruction text and the audio text in the registration recognition library, constructs an overall instruction set according to all the instruction texts, and combines all the audio texts corresponding to the same instruction text to construct a single audio set; The processing module analyzes each audio text in the single audio set to determine the audio representative parameter, and determines the audio text corresponding to the maximum audio representative parameter as the single representative audio; The processing module pairs and combines the instruction text in the overall instruction set and the corresponding single representative audio to construct an initial recognition library of the current user; The processing module acquires the demand control audio in real time after the initial recognition library is constructed, inputs the demand control audio into the initial recognition library for recognition analysis to determine the control instruction, and controls the operation of the socket according to the control instruction.

[0014] In summary, the present application includes at least one of the following beneficial technical effects: During the use of the intelligent socket, the same voice user is matched according to the voice input by the user, and the recognition database of the corresponding user is constructed to construct the recognition database of the current user, so that each dialect can be effectively recognized, and the universality and user satisfaction of the intelligent wall socket are improved; During the recognition of the same voice user, the analysis of the location area is introduced, so that the accuracy of the sample data is improved, and the accuracy of the constructed recognition database is improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] Fig. 1 is a flowchart of an intelligent socket control method.

[0016] Fig. 2 is a module flowchart of an intelligent socket control method. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings Figs. 1-2 and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0018] The embodiments of the present application will be described in further detail below with reference to the accompanying drawings.

[0019] The embodiments of the present application disclose an intelligent socket control method, referring to Fig. 1 The method flow of the intelligent socket control method includes the following steps: Step S100: Obtain the external input audio of the current user under the preset instruction question.

[0020] The instruction question is a preset question that needs to be answered by voice when the user first uses the smart socket, for example, "please say "on" in your language", "please say "off" in your language", etc. The specific instruction question is set by the staff, and the external input audio is the audio information input by the user according to the instruction question.

[0021] Step S101: Define the user who is consistent with the current external input audio as a same type user, and obtain the registration recognition library according to the same type user.

[0022] The same type user is defined to identify users with the same language situation, and the registration recognition library is a voice library that can be recognized by the smart wall socket used by the same type user. The reading of the voice library requires the user to open the permission in advance.

[0023] Step S102: Determine the instruction text and the audio text in the registration recognition library, and construct the overall instruction set according to all instruction texts, and combine all audio texts corresponding to the same instruction text to construct the single audio set.

[0024] The instruction text is the text of the control instruction that needs to be implemented, and the audio text is the text of the audio corresponding to the instruction text so that the instruction text can output control. For example, the preface "on" is the instruction text, and the dialect corresponding to "on" is the audio text. The overall instruction set is a set composed of all instruction texts of all same type users, that is, a set of instructions that will be used in the process of controlling the smart wall socket. The single audio set is a set of all same type users corresponding to the audio text under a single instruction text.

[0025] Step S103: Analyze each audio text in the single audio set to determine the audio representative parameter, and determine the audio text corresponding to the maximum audio representative parameter as the single representative audio.

[0026] The audio representative parameter is the parameter value that the audio text as the standard audio of the current instruction text is feasible. The larger the value is, the more feasible the corresponding audio text is as the standard audio of the instruction text. For specific determination method, refer to steps S500-S502. The maximum audio representative parameter indicates that the corresponding audio text as the standard audio of the current instruction text is the most feasible, so it is defined as the single representative audio for identification, which is convenient for subsequent analysis.

[0027] Step S104: Pair and combine the instruction text in the overall instruction set and the corresponding single representative audio to construct the initial recognition library of the current user.

[0028] By pairing the instruction text and the monomer representative audio one by one, the identification corresponding relationship between the control instruction and the audio can be known, and the corresponding initial identification library can be constructed for the current user to perform voice recognition.

[0029] Step S105: After the initial identification library is constructed, the demand control audio is acquired in real time to input into the initial identification library for identification analysis to determine the control instruction, and the socket operation is controlled according to the control instruction.

[0030] The demand control audio is the audio input by the user, for example, the dialect corresponding to "turn on the socket", at this time the initial identification library can identify the audio to match the corresponding control instruction to control the socket, so as to realize the function that various dialects can start the socket voice control, and improve the universality and user satisfaction of the intelligent wall socket.

[0031] After the same type of user is determined, the intelligent socket control method further comprises: Step S200: Counting according to the same type of user to determine the same type of quantity.

[0032] The same type of quantity is the total quantity of the determined same type of user.

[0033] Step S201: Determine whether the same type of quantity is greater than a preset reference quantity.

[0034] The reference quantity is the minimum same type of quantity required when the staff sets the voice recognition sample with more samples, and the purpose of the determination is to know whether the current sample quantity meets the requirements.

[0035] Step S2011: If the same type of quantity is greater than the reference quantity, the initial identification library is determined according to the current same type of user.

[0036] When the same type of quantity is greater than the reference quantity, it means that there are more sample quantities, and at this time the initial identification library can be better constructed.

[0037] Step S2012: If the same type of quantity is not greater than the reference quantity, the overall instruction set is determined according to the current same type of user, and it is determined whether the overall instruction set contains a preset reference instruction set.

[0038] When the same type of quantity is not greater than the reference quantity, it means that there are fewer samples, which needs to be further analyzed; the overall instruction set is the set of instruction texts possessed by the same type of user, and the reference instruction set is the set of instruction texts required to realize the most basic control of the socket, and the purpose of the determination is to know whether the initial identification library constructed at present can cover the basic daily use requirements.

[0039] Step S20121: If the whole instruction set contains the benchmark instruction set, determine the initial recognition library according to the current same type user.

[0040] When the whole instruction set contains the benchmark instruction set, it indicates that the initial recognition library currently constructed can meet the daily basic use requirements, so the initial recognition library can be normally constructed.

[0041] Step S20122: If the whole instruction set does not contain the benchmark instruction set, compare the whole instruction set and the benchmark instruction set to determine the missing instruction, and generate an instruction question according to the missing instruction and output it to wait for external input audio.

[0042] When the whole instruction set does not contain the benchmark instruction set, it indicates that the initial recognition library constructed according to the current instruction text cannot meet the daily needs of the current user, so further analysis is required; the missing instruction is the instruction text that is in the benchmark instruction set but not in the whole instruction set, that is, the instruction text that needs to be added by the user when meeting the daily needs, so the corresponding instruction question is generated for the user to add voice, so that the initial recognition library constructed subsequently can meet the user's requirements.

[0043] If the same type quantity is greater than the benchmark quantity, the intelligent socket control method further comprises: Step S300: When the same type quantity is greater than the preset redundancy quantity, obtain the effective installation position of the current user and the comparison installation position of the same type user.

[0044] The redundancy quantity is the minimum same type quantity required by the staff to determine that the sample quantity is too large and requires a large amount of data analysis; when the same type quantity is greater than the redundancy quantity, it indicates that the current sample data is too large, in order to reduce the data analysis amount, the sample needs to be screened; the effective installation position is the actual position of the intelligent wall socket installed by the current user in the geographical space, which can be determined by carrying a positioning device on the socket; the comparison installation position is the position of the socket of the same type user.

[0045] Step S301: Draw an influence coverage range with a width of a preset influence distance centered on the effective installation position, and calculate the range same type proportion in the influence coverage range according to the same type user and all users.

[0046] The influence distance is a fixed distance value set by the staff to determine that the positions are relatively close, for example, the distance value between the positions of the same county, which is specifically set by the staff according to the actual situation; the influence coverage range is the area range drawn with the effective installation position as the center and the influence distance as the radius; the range same type proportion is the proportion value of the number of same type users in the current influence coverage range to the number of all users.

[0047] Step S302: determining whether the range same-class proportion is greater than a preset aggregation proportion.

[0048] The aggregation proportion is a minimum range same-class proportion set by a staff to determine that a large number of same-class users of a current user are in the range of the area, and the purpose of the determination is to know whether the position of the current user is a region of a commonly used dialect, i.e., whether it is a corresponding hometown.

[0049] Step S3021: outputting a user independent signal if the range same-class proportion is not greater than the aggregation proportion.

[0050] When the range same-class proportion is not greater than the aggregation proportion, it indicates that the user is not in an area with a large number of same-class users, and at this time, the user independent signal is outputted to mark this situation for subsequent analysis.

[0051] Step S3022: defining the same-class users outside the current influence coverage range at the comparison installation position to be cancelled if the range same-class proportion is greater than the aggregation proportion.

[0052] When the range same-class proportion is greater than the aggregation proportion, it indicates that the current user is in an area with a large number of same-class users, i.e., the user can be determined to be in a hometown, and at this time, the dialects of the same-class users in the area range and the current user should be most similar, so only the same-class users defined in the area can better and accurately perform data analysis.

[0053] After the user independent signal is outputted, the intelligent socket control method further comprises: Step S400: randomly selecting a same-class user and defining a similar coverage range with a width of a preset similar distance according to the corresponding comparison installation position.

[0054] The similar distance is a fixed value distance set by a staff to determine that the positions are relatively close, and the similar distance is slightly smaller than the influence distance to improve the accuracy of data analysis; and the similar coverage range is an area range defined according to the comparison installation position of the selected same-class user as a center and the similar distance as a radius.

[0055] Step S401: determining a similar same-class proportion according to the similar coverage range, and defining a same-class user corresponding to the similar same-class proportion greater than the aggregation proportion as an aggregation user.

[0056] The similar same-class proportion is a proportion value of the same-class users in the similar coverage range, and the aggregation user is defined to identify a large number of users in the same area for subsequent analysis.

[0057] Step S402: inducing aggregation users with at least one same-class user overlapping in two similar coverage ranges to combine to form a user aggregation combination.

[0058] Combining the user cluster combination to combine the clustered users in the same area.

[0059] Step S403: Combining the same type of users contained in each similar coverage range in the user cluster combination to construct an area user combination.

[0060] The area user combination is a combination of the same type of users in the same area.

[0061] Step S404: Counting the same type of users in the area user combination to determine the area number, and defining the same type of users in the area user combination other than the maximum area number as invalid.

[0062] The area number is the total number of the same type of users in the area user combination, and the maximum area number is determined to mark the area most likely to be the home, so only the same type of users in the combination are retained to improve the accuracy of data analysis.

[0063] The step of analyzing each audio text in the single audio set to determine the audio representative parameter includes: Step S500: Counting the same audio text under the single audio text to determine the audio same number.

[0064] The audio same number is the number value of the audio text with consistent pronunciation form obtained from the analysis of the single audio text.

[0065] Step S501: Counting all audio texts in the single audio set to determine the set audio number.

[0066] The set audio number is the number value of all audio texts in the single audio set.

[0067] Step S502: Calculating the set audio number and the audio same number to determine the audio representative parameter.

[0068] The audio representative parameter is the value obtained by dividing the set audio number by the audio same number.

[0069] After determining the audio representative parameter, the intelligent socket control method further includes: Step S600: Defining the audio text with an audio representative parameter greater than the preset valid representative parameter as a valid audio.

[0070] The valid representative parameter is the minimum audio representative parameter required by the staff to determine that it is qualified as a single representative audio under the current instruction text. Defining valid audio distinguishes different audio texts for subsequent analysis.

[0071] Step S601: judging whether the valid audio is unique.

[0072] The purpose of the judgment is to know whether there is only one audio text meeting the requirements.

[0073] Step S6011: if the valid audio is unique, determining the single representative audio according to the audio representative parameter.

[0074] When the valid audio is unique, it means that there is only one audio text meeting the requirements, so the single representative audio can be determined according to the audio representative parameter.

[0075] Step S6012: if the valid audio is not unique, generating an instruction question according to the instruction text corresponding to the current single audio set and outputting the instruction question to wait for external input audio.

[0076] When the valid audio is not unique, it means that there are multiple audio texts meeting the requirements, which need to be further analyzed. By re-generating an instruction question according to the current instruction text, the current user can perform voice input to increase the number and range of instruction questions, so that the generation of the initial recognition library is more accurate.

[0077] After the initial recognition library is constructed, the intelligent socket control method further comprises: Step S700: obtaining a job feedback state after controlling the socket operation according to the control instruction.

[0078] The job feedback state is information fed back by the user on the socket operation after the socket performs the operation according to the control instruction. For example, if the socket operation output receives the audio information of "error", it is determined that the job feedback state is an error operation state. When no other audio information representing the error operation state is input, it is determined that the job feedback state is a correct operation state.

[0079] Step S701: outputting a control error instruction when the job feedback state is consistent with a preset error operation state.

[0080] When the job feedback state is consistent with the error operation state, it means that the socket operation control is wrong, i.e., the voice recognition is wrong, so the control error instruction is output to identify this situation for subsequent analysis.

[0081] Step S702: calculating the instruction error proportion according to the control error instruction and the control instruction.

[0082] The instruction error proportion is a proportion value of the number of control error instructions to the number of all control instructions.

[0083] Step S703: when the instruction error ratio is greater than the preset allowable error ratio and the number of control instructions is greater than the preset number of standards, the instruction text and the audio text corresponding to the correct job state consistent with the preset correct job state are combined to generate a qualified combination.

[0084] The allowable error ratio is the maximum instruction error ratio allowed when it is determined that the initial recognition library can better recognize the user's dialect. The number of standards is the minimum number of control instructions required for defrosting recognition library identification analysis. When the instruction error ratio is greater than the allowable error ratio and the number of control instructions is greater than the number of standards, it indicates that the initial recognition library currently constructed cannot better recognize the current user's voice, i.e. the initial recognition library currently constructed has deviation and needs further analysis. When the job feedback state is consistent with the correct job state, it indicates that the corresponding instruction text and audio text can be successfully recognized, so the combination is defined as a qualified combination to identify the successful recognition, thereby facilitating subsequent analysis.

[0085] Step S704: defining the audio text corresponding to the instruction text in the qualified combination as external input audio to re-determine the same type of user, and re-determining the initial recognition library according to the same type of user.

[0086] By updating the instruction question and the corresponding external input audio, the same type of user can be re-matched, so that the initial recognition library can be better determined.

[0087] Reference Fig. 2 Based on the same inventive concept, the embodiments of the present application provide an intelligent socket control system, comprising: An acquisition module is configured to acquire external input audio of a current user under a preset instruction question; A processing module is connected to the acquisition module and is configured to store and process information; The processing module defines a user whose external input audio is consistent with the current external input audio as a same type of user, and causes the acquisition module to acquire a registration recognition library according to the same type of user; The processing module determines instruction texts and audio texts in the registration recognition library, and constructs an overall instruction set according to all instruction texts, and combines all audio texts corresponding to the same instruction text to construct a single audio set; The processing module analyzes each audio text in the single audio set to determine an audio representative parameter, and determines a single representative audio corresponding to the maximum audio representative parameter; The processing module pairs and combines the instruction texts in the overall instruction set and the corresponding single representative audio to construct an initial recognition library of the current user; The processing module acquires the demand control audio in real time after the initial recognition library is constructed to input to the initial recognition library to determine the control instruction, and controls the socket operation according to the control instruction; The problem supplementary analysis module analyzes and processes the situation that the number of similar users is insufficient; The aggregation situation analysis module is used for analyzing the aggregation situation of similar users; The similar user screening module is used for screening similar users; The audio representative parameter determination module is used for determining the audio representative parameters of each audio text; The multi-audio screening module is used for screening multiple audios meeting the requirements; The initial recognition library updating module is used for updating the initial recognition library in the case of inaccurate recognition.

[0088] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example for illustration, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

Claims

1. A smart socket control method, characterized in that: include: Get the external input audio of the current user under the preset command question; Define users whose external input audio is consistent with the current external input audio as similar users, and obtain a registration and identification library based on similar users; Determine the command text and audio text in the registration recognition database, and construct a whole command set based on all the command texts, and combine all the audio texts corresponding to the same command text to construct a single audio set; Analyze each audio text in the single audio set to determine an audio representative parameter, and determine the audio text corresponding to the largest audio representative parameter as the single representative audio; Pairing and combining the command texts in the overall command set and the corresponding single-unit representative audios to build an initial recognition library for the current user; After the initial recognition library is built, the required control audio is obtained in real time and input into the initial recognition library for recognition analysis to determine the control instructions, and the socket operation is controlled according to the control instructions.

2. The smart socket control method according to claim 1, characterized in that: After similar users are identified, the smart socket control method further includes: Count users of the same category to determine the number of categories; Determine whether the quantity of the same type is greater than the preset benchmark quantity; If the number of similar users is greater than the benchmark number, the initial identification library is determined based on the current similar users; If the number of similar users is not greater than the benchmark number, then determine the overall instruction set based on the current similar users and determine whether the overall instruction set includes the preset benchmark instruction set; If the overall instruction set includes the benchmark instruction set, then the initial recognition library is determined based on the current similar users; If the overall instruction set does not include the benchmark instruction set, a comparison is performed between the overall instruction set and the benchmark instruction set to determine the missing instruction, and a command question is generated based on the missing instruction and output to wait for obtaining external input audio.

3. The smart socket control method according to claim 2, characterized in that: If the number of the same type is greater than the benchmark number, the smart socket control method further includes: When the number of similar devices is greater than the preset redundant number, obtain the current user's valid installation location and the compared installation locations of similar users; Determine the impact coverage area with the effective installation location as the center and the width as the preset impact distance. Calculate the same type of users and all users within the impact coverage area to determine the same type of percentage within the scope. Determine whether the proportion of the same category in the range is greater than the preset aggregation proportion; If the proportion of the same type in the range is not greater than the proportion of the aggregate, the user's independent signal is output; If the proportion of the same category in the scope is greater than the proportion of the aggregation, the definition of the same category users whose installation locations are outside the current impact coverage will be cancelled.

4. The smart socket control method according to claim 3, characterized in that: After the user's independent signal is output, the smart socket control method further includes: Randomly select a user of the same type and define a similar coverage area with a preset width and a similar distance based on the corresponding comparison installation location; Determine the proportion of similar users based on similar coverage, and define the corresponding similar users when the proportion of similar users is greater than the aggregation proportion as aggregated users; Aggregate users with at least one overlapping user of the same type in two similar coverage areas to form a user aggregation combination; Combining similar users within similar coverage areas in the user aggregation combination to construct a regional user combination; The number of regions is determined by counting the same type of users in the regional user combination, and the definition of the same type of users other than the same type of users in the regional user combination with the largest number of regions is cancelled.

5. The smart socket control method according to claim 1, characterized in that: The steps of analyzing each audio text in a single audio set to determine audio representative parameters include: Count the same audio texts under a single audio text to determine the number of identical audios; Counting all audio texts in a single audio collection to determine the number of audios in the collection; The audio representative parameter is determined based on the number of aggregated audios and the number of identical audios.

6. The smart socket control method according to claim 5, characterized in that: After the audio representative parameters are determined, the smart socket control method further includes: An audio text whose audio representative parameter is greater than a preset valid representative parameter is defined as a valid audio; Determine whether the valid audio is unique; If the valid audio is unique, the single representative audio is determined according to the audio representative parameter; If the valid audio is not unique, an instruction question is generated according to the instruction text corresponding to the current single audio set and output to wait for obtaining external input audio.

7. The smart socket control method according to claim 1, characterized in that: After the initial recognition library is built, the smart socket control method further includes: Obtain operation feedback status after controlling the socket operation according to the control instruction; Outputting a control error instruction when the operation feedback status is consistent with the preset error operation status; Calculate the command error ratio based on the control error command and the control command; When the instruction error ratio is greater than a preset allowable error ratio and the number of control instructions is greater than a preset qualified number, the instruction text and audio text corresponding to the operation feedback status when it is consistent with the preset correct operation status are combined to generate a qualified combination; The audio text corresponding to the instruction text in the qualified combination is defined as the external input audio to re-determine the same type of users, and the initial recognition library is re-determined based on the same type of users.

8. A smart socket control system, characterized in that: include: The acquisition module is used to obtain the external input audio of the current user under the preset command question; A processing module, connected to the acquisition module, for storing and processing information; The processing module defines the user whose external input audio is consistent with the current external input audio as a similar user, and enables the acquisition module to acquire the registration recognition library based on the similar users; The processing module determines the command text and the audio text in the registration recognition library, and constructs a whole command set based on all the command texts, and combines all the audio texts corresponding to the same command text to construct a single audio set; The processing module analyzes each audio text in the single audio set to determine an audio representative parameter, and determines the audio text corresponding to the largest audio representative parameter as the single representative audio; The processing module pairs and combines the instruction texts in the overall instruction set with the corresponding single-unit representative audio to build an initial recognition library for the current user; After the initial recognition library is built, the processing module obtains the required control audio in real time and inputs it into the initial recognition library for recognition analysis to determine the control instructions, and controls the socket operation according to the control instructions.

Images