CN116052660A - Method, device and storage medium for executing voice commands of intelligent equipment - Google Patents

Abstract

The application relates to an execution method, a device and a storage medium of a voice instruction of an intelligent device, wherein the method comprises the following steps: acquiring a voice instruction for controlling the intelligent equipment by a user; word segmentation is carried out on sentences corresponding to the voice instructions, and word segmentation results are compared with words in a preset database; comparing the sound wave diagram of the first word, which is not in the database, with the sound wave diagram of the second word in the database; and executing the voice instruction in combination with the semantics of the second word under the condition that the similarity of the sound wave diagram exceeds a first preset threshold. According to the method and the device, the problem that the intelligent home effect is poor due to the fact that voice recognition is deviated due to different pronunciations in the related art is caused.

Description

Method, device and storage medium for executing voice commands of intelligent equipment

technical field

The present application relates to the field of voice processing, and in particular to a method, device and storage medium for executing voice commands of smart devices.

Background technique

With the development of technology, smart home has been ubiquitous in the lives of most groups, and voice-operated smart products have become the mainstream trend. At the same time, it is also accompanied by problems caused by voice, such as everyone's pronunciation is different, causing the system to recognize voice When there is a deviation, or the recognition error caused by some homophones, the speech recognition efficiency is low, and the effect of controlling the smart home through voice commands is poor.

Contents of the invention

The present application provides a method, device, and storage medium for executing voice commands of smart devices to solve the problem in the related art that voice recognition has deviations due to different pronunciations, resulting in poor control of smart homes through voice commands.

In the first aspect, the present application provides a method for executing a voice command of a smart device, including: obtaining a voice command for the user to control the smart device; Comparing the words in the database; comparing the acoustic wave graph of the first word whose word segmentation result is not in the database with the acoustic wave graph of the second word in the database; the similarity in the acoustic graph exceeds the first preset In the case of a threshold value, execute the voice instruction in combination with the semantics of the second word.

In a second aspect, the present application provides an execution device for voice commands of smart devices, including: an acquisition module for acquiring voice commands for the user to control the smart device; a processing module for processing sentences corresponding to the voice commands Carrying out word segmentation, comparing the word segmentation result with the words in the preset database; comparison module, used to compare the sound wave of the first word whose word segmentation result is not in the database with the sound wave of the second word in the database The graphs are compared; an execution module is configured to execute the voice instruction in combination with the semantics of the second word when the similarity of the acoustic wave graph exceeds a first preset threshold.

In a third aspect, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete mutual communication through the communication bus;

memory for storing computer programs;

The processor is configured to implement the steps of the method for executing the voice command of the smart device described in any one embodiment of the first aspect when executing the program stored in the memory.

In the fourth aspect, there is provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for executing the voice command of the smart device as described in any embodiment of the first aspect is implemented. A step of.

Compared with the prior art, the above-mentioned technical solutions provided by the embodiments of the present application have the following advantages:

In the embodiment of the present application, after the user's voice command for controlling the smart device is obtained, the voice command will be segmented. If there is a word that is not in the preset database after the word segmentation, it needs to be compared with the database. Compare the words in the system to obtain the words with similarity in the waveform diagram, and then correct the words that are not in the preset database to obtain the correct semantics of the voice command, and execute the corrected voice command based on the correct semantics. That is to say, through this application, even if there is an accent problem that leads to inaccurate recognition of the voice information issued by the user, it can be corrected to obtain the correct voice command, so that the execution of subsequent voice commands is more accurate, avoiding misoperation or Unrecognizable situation solves the problem in the related art that voice recognition is biased due to different pronunciation, which leads to poor effect of controlling smart home through voice commands.

Description of drawings

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

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, In other words, other drawings can also be obtained from these drawings without paying creative labor.

FIG. 1 is one of the schematic flow diagrams of a method for executing a voice command of a smart device provided in an embodiment of the present application;

FIG. 2 is the second schematic flow diagram of a method for executing a voice command on a smart device provided in an embodiment of the present application;

FIG. 3 is a schematic flowchart of a semantic correction method based on sound wave recognition provided by an embodiment of the present application;

FIG. 4 is one of the structural schematic diagrams of an execution device for a smart device voice command provided by an embodiment of the present application;

FIG. 5 is the second structural schematic diagram of an execution device for a smart device voice command provided by an embodiment of the present application;

FIG. 6 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

Detailed ways

In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments It is a part of the embodiments of this application, but not all of them. Based on the embodiments in the present application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present application.

Fig. 1 is a schematic flow diagram of a method for executing a voice command of a smart device provided in an embodiment of the present application. As shown in Fig. 1, the steps of the method include:

Step 101, acquiring the user's voice command for controlling the smart device;

It should be noted that the smart device in the embodiment of the present application may be a smart home, such as an air conditioner, TV, washing machine, refrigerator, etc., or a smart office device, such as a printer, a smart seat, etc., or a personal smart device. Such as mobile phones, smart watches, etc.

Step 102, performing word segmentation on the sentence corresponding to the voice instruction, and comparing the word segmentation result with the words in the preset database;

In the embodiment of the present application, the voice command is voice information used to control the smart device, such as "query the life of the filter element of the water purifier", "query the maintenance time of the air conditioner", "turn on the printer" and so on. Taking the voice command as "query life of filter element of water purifier" as an example, the result after word segmentation is "query", "water purifier", "filter element", and "lifetime". "Water purifier filter element life" may be identified as "query water purifier travel life", then the result after analysis is "query", "water purifier", "travel", "life". It can be seen that word segmentation is a word segmentation with the conventional combination of words in Chinese to obtain the smallest unit that can recognize semantics.

Step 103, comparing the acoustic image of the first word in the database with the acoustic image of the second word in the database as a word segmentation result;

It should be noted that the second word may be any word in the database, or a word that is relatively error-prone. Furthermore, the voice command originally intended by the user is "query the life of the filter element of the water purifier", but because of the accent problem, the finally recognized voice command is "query the travel life of the water purifier". After word segmentation of the voice command, "travel " is obviously not the semantics originally intended to be expressed in the voice command, therefore, this needs to be corrected to obtain the correct voice command voice. For the voice control commands of smart devices, the more commonly used ones are "open...", "close...", "query...", so you can set up a database in advance, which stores commonly used voice commands. After the voice command issued by the user, it can be compared with the words in the database. For example, only the waveforms of "filter element" and "travel" in the database are relatively close, and the "filter element" is also in the preset voice command, then "Travel" could be corrected to "filter". That is to say, the voice command and the word segmentation result of the voice command can be pre-stored in the database in the embodiment of the present application, so that the voice command issued by the current user can be quickly identified.

Step 104, when the similarity of the acoustic wave pattern exceeds the first preset threshold, execute the voice instruction in combination with the semantics of the second word.

It should be noted that the first preset threshold in the embodiment of the present application can be set accordingly according to actual needs, for example, the results of the amplitude comparison and the frequency comparison in the compared waveform diagrams are both greater than 50% , or 60%, 80%, etc.

It can be seen that in the embodiment of the present application, after the user's voice command for controlling the smart device is acquired, the voice command will be segmented. If there are words that are not in the preset database after the word segmentation, they need to be Comparing with the words in the database, the words with waveform similarity are obtained, and then the words not in the preset database are corrected to obtain the correct semantics of the voice command, and the corrected voice command is executed based on the correct semantics. That is to say, through this application, even if there is an accent problem that leads to inaccurate recognition of the voice information issued by the user, it can be corrected to obtain the correct voice command, so that the execution of subsequent voice commands is more accurate, avoiding misoperation or Unrecognizable situation solves the problem in the related art that voice recognition is biased due to different pronunciation, which leads to poor effect of controlling smart home through voice commands.

In an optional implementation of the embodiment of the present application, the method of comparing the word segmentation result with the words in the preset database involved in the above step 102 may further include:

Step 11, count the words that exist in the database in the word segmentation results;

Step 12, the number of words existing in the database in the word segmentation result is compared with the number of all words in the word segmentation result to obtain the effective rate of the voice command;

In the specific example, take "query water purifier travel life", the result after analysis is "query", "water purifier", "travel", "life" as an example, then "query", "water purifier ", "life" has a corresponding word in the database, while "travel" does not exist in the database, so the number of words after the participle of "query water purifier travel life" is 4, which exists in the database The number of words is 3, then the effective rate of the "inquiry about the travel life of the water purifier" is 3/4=0.75, and the corresponding effective rate is 75%.

Step 13, in the case that the effective rate exceeds the second preset threshold, determine that the voice command is a valid command;

In the embodiment of the present application, the second preset threshold can be set correspondingly according to actual needs, for example, set to 60%, 65%, 70% and so on.

Step 14, if the effective rate does not exceed the second preset threshold, determine that the voice command is an invalid command; where the first word in the valid command that does not exist in the database is compared with the acoustic wave.

The second preset threshold is set to ensure that there are fewer words that are not normally recognized in the voice command, so that the correct semantics of the voice command can be obtained in subsequent corrections. If the second preset threshold is set smaller, Then, if the processing error rate of the current voice command recognition is high, it cannot be corrected, that is, the command is invalid, and the user needs to re-send the voice command.

In an optional implementation manner of the embodiment of the present application, for the method of comparing the acoustic wave image of the first word whose word segmentation result is not in the database with the acoustic wave image of the second word in the database involved in the above step 103, it can further be include:

Step 21, obtaining the waveform parameters of the acoustic wave diagram of the first word, and the waveform parameters of the acoustic wave diagram of the second word, wherein the waveform parameters include at least one of the following: amplitude, frequency;

Step 22, comparing the waveform parameters of the acoustic wave diagram of the first word and the waveform parameters of the acoustic wave diagram of the second word;

Step 23, acquiring the ratio between the waveform parameter of the acoustic wave diagram of the first word and the waveform parameter of the acoustic wave diagram of the second word, wherein the ratio is used to represent the similarity.

In the embodiment of the present application, the waveform diagram is obtained according to the pronunciation and intonation of words, that is, each word has a corresponding waveform diagram, and each waveform diagram has a corresponding amplitude and frequency. In this application, the problem of speech recognition errors caused by accent or pronunciation is mainly solved, but the errors caused by speech or accent are often the pronunciations of the two are relatively close and sound similar, such as "travel" Therefore, in this application, the similarity between two words can be determined according to the comparison of the amplitude and frequency in the waveform diagram. If the similarity is higher, it means that the second word is the correct word.

It should be noted that if there are multiple words whose similarity meets the first preset threshold in the current database, then semantic recognition can be performed on the multiple words, and the word whose semantics is most suitable for the current voice instruction is determined as the required word. The word that replaces the first word.

In an optional implementation of the embodiment of the present application, the manner of executing voice instructions in combination with the semantics of the second word involved in the above step 103 may further include:

Step 31, amending the first word in the word segmentation result to the second word;

Step 32, based on the semantics of the voice instruction corresponding to the modified word segmentation result, execute the voice instruction corresponding to the modified word segmentation result.

For the above step 31 and step 32, take "query water purifier travel life", then the result after analysis is "query", "water purifier", "travel", "life" as an example, then "query", "Water purifier" and "life" have corresponding words in the database, while "travel" does not exist in the database. Compare the waveform diagram of "travel" with the waveform diagram of words in the database, according to The waveform similarity finally identifies the "filter element" in the database, and the voice command is corrected based on the "filter element", and the correct voice command "query the life of the filter element of the water purifier" is executed. Of course, it is also possible to recognize multiple words that satisfy the waveform similarity. For example, the recognized result is "filter element", "aluminum core", "green heart", etc., and it can be further judged whether the semantics of the recognized result match the current voice command , in this example "filter element" matches the current voice command, but the semantics of "aluminum core" and "green heart" do not match, you need to correct the "filter element" voice command and execute the correct voice command "query Water purifier filter life".

In the embodiment of the present application, after executing the voice instruction in combination with the semantics of the second word, as shown in FIG. 2 , the method steps in the embodiment of the present application may also include:

Step 105, backing up the first word to the database.

It can be seen that in the embodiment of the present application, the first word can be added to the database. In a specific example, take "query the travel life of the water purifier" as an example, you can add "travel" to the database, and next time it is still "query the travel life of the water purifier", you can directly execute the "query "Water purifier travel life" voice command, there is no need to adjust it, because its correct semantics can already be known, that is, "query the life of the water purifier filter element". That is to say, in this application, even if the pronunciation is not standard, its correct semantics can be directly recognized, which improves the experience effect of voice command control smart devices.

The application will be explained below in conjunction with the specific implementation of the embodiment of the application. This specific implementation provides a semantic correction method based on sound wave recognition. The graph is compared with the data in the existing database. If the set value is reached, the word will be corrected, and the original word will be added to the backup prediction database. If the same word appears next time, the recognition will be successful. As shown in Figure 3, the steps of the semantic correction method based on sound wave recognition include:

Step 301, collecting voice instructions spoken by the user;

Among them, if the user's intention is: query the life of the filter element of the water purifier, and extract the voice into text.

Step 302, the extracted voice command is: query the travel life of the water purifier;

For this step 302, it may be because of the user's pronunciation or other factors that the extracted voice is inaccurate, that is, "query the life of the filter element of the water purifier" is recognized as "query the life of the travel water purifier".

Step 303, divide the sentence into "query", "water purifier", "travel", and "life" through word segmentation. At this time, the system can recognize the semantics of "query", "water purifier" and "life", but cannot Identify "travel". The current effective rate is 0.75.

Wherein, in a specific example, a threshold can be set to control resource consumption. If it exceeds this value, it is likely to be a valid instruction, and if it is lower than this threshold, it indicates that the current voice instruction is an invalid instruction.

Step 304, generate a sound wave map from "Travel", and compare it with the sound wave map in the existing database;

In a specific example, a threshold can be set to determine whether the similarity is up to standard, which can be set according to the acceptance level of a specific scene. The larger the value, the higher the accuracy of the system.

Step 305, if the set threshold is reached, the original word can be treated as a matched word, then the system can successfully identify the intent, and at the same time, the original word can be added to a backup database;

Wherein, adding the original word into a backup database can make the same sentence in the future, when reaching step 303, just can directly and successfully analyze the intention.

Through the above steps 301 to 305, for the homonyms that may exist when recognizing speech, or near-syllable words extracted due to pronunciation problems, resulting in a low speech recognition rate, in this application, some unrecognized characters are generated into sound waves Figure, by comparing with the data in the existing hot words, so as to correct the unrecognized content, thereby improving the efficiency of semantic recognition, and can automatically expand the database.

Corresponding to the above-mentioned Figure 1, the embodiment of the present application also provides an execution device for voice commands of smart devices, as shown in Figure 4, the device includes:

Obtaining module 42, is used for obtaining the voice instruction that the user controls smart device;

The processing module 44 is used to perform word segmentation on the sentence corresponding to the voice instruction, and compare the word segmentation result with the words in the preset database;

Contrast module 46, is used for comparing the acoustic wave figure of the first word in the database with the second word in the database with the word segmentation result;

The execution module 48 is configured to execute the voice instruction in combination with the semantics of the second word when the similarity of the acoustic wave pattern exceeds the first preset threshold.

It can be seen that in the embodiment of the present application, after the user's voice command for controlling the smart device is acquired, the voice command will be segmented. If there are words that are not in the preset database after the word segmentation, they need to be Comparing with the words in the database, the words with waveform similarity are obtained, and then the words not in the preset database are corrected to obtain the correct semantics of the voice command, and the corrected voice command is executed based on the correct semantics. That is to say, through this application, even if there is an accent problem that leads to inaccurate recognition of the voice information issued by the user, it can be corrected to obtain the correct voice command, so that the execution of subsequent voice commands is more accurate, avoiding misoperation or Unrecognizable situation solves the problem in the related art that voice recognition is biased due to different pronunciation, which leads to poor effect of controlling smart home through voice commands.

In an optional implementation of the embodiment of the present application, the processing module 44 in the embodiment of the present application may further include: a statistical unit for counting the words existing in the database in the word segmentation results; The number of words existing in the database is compared with the number of all words in the word segmentation result to obtain the effective rate of the voice command; the first determination unit is used to determine that the voice command is valid when the effective rate exceeds the second preset threshold instruction; the second determination unit is used to determine that the voice instruction is an invalid instruction when the effective rate does not exceed the second preset threshold; wherein, the first word in the effective instruction that does not exist in the database is compared with the acoustic wave operate.

In this regard, in the specific example, take "query water purifier travel life", then the result after analysis is "query", "water purifier", "travel", "life" as an example, then "query", " "Water purifier", "life" has corresponding words in the database, but "travel" does not exist in the database, so the number of words after the participle of "query water purifier travel life" is 4, in If the number of words in the database is 3, then the effective rate of "querying the travel life of the water purifier" is 3/4=0.75, and the corresponding effective rate is 75%. In addition, in the embodiment of the present application, the second preset threshold can be set correspondingly according to actual needs, for example, set to 60%, 65%, 70% and so on. The second preset threshold is set to ensure that there are fewer words that are not normally recognized in the voice command, so that the correct semantics of the voice command can be obtained in subsequent corrections. If the second preset threshold is set smaller, Then, if the processing error rate of the current voice command recognition is high, it cannot be corrected, that is, the command is invalid, and the user needs to re-send the voice command.

In an optional implementation of the embodiment of the present application, the comparison module 48 in the embodiment of the present application may further include: a first acquisition unit, configured to acquire the waveform parameters of the acoustic wave diagram of the first word, and the acoustic wave diagram of the second word The waveform parameters, wherein the waveform parameters include at least one of the following: amplitude, frequency; a comparison unit, used to compare the waveform parameters of the acoustic wave diagram of the first word and the waveform parameters of the acoustic wave diagram of the second word; the second acquisition unit, using The ratio between the waveform parameter of the acoustic wave image of the first word and the waveform parameter of the acoustic wave image of the second word is obtained, wherein the ratio is used to characterize the similarity.

In the embodiment of the present application, the waveform diagram is obtained according to the pronunciation and intonation of words, that is, each word has a corresponding waveform diagram, and each waveform diagram has a corresponding amplitude and frequency. In this application, the problem of speech recognition errors caused by accent or pronunciation is mainly solved, but the errors caused by speech or accent are often the pronunciations of the two are relatively close and sound similar, such as "travel" Therefore, in this application, the similarity between two words can be determined according to the comparison of the amplitude and frequency in the waveform diagram. If the similarity is higher, it means that the second word is the correct word.

It should be noted that if there are multiple words whose similarity meets the first preset threshold in the current database, then semantic recognition can be performed on the multiple words, and the word whose semantics is most suitable for the current voice instruction is determined as the required word. The word that replaces the first word.

In an optional implementation of the embodiment of the present application, the execution module 48 in the embodiment of the present application may further include: a correction unit, configured to modify the first word in the word segmentation result to a second word; an execution unit, configured to The semantics of the phonetic instruction corresponding to the word segmentation result after modification is corrected, and the phonetic instruction corresponding to the word segmentation result after modification is executed.

In this regard, taking "query water purifier travel life", the result after analysis is "query", "water purifier", "travel", "life" as an example, then "query", "water purifier" , "Life" has a corresponding word in the database, while "travel" does not exist in the database. Compare the waveform diagram of "travel" with the waveform diagram of the words in the database, and finally determine the If the "filter element" is identified in the database, the voice command is corrected based on the "filter element", and the correct voice command "query the life of the filter element of the water purifier" is executed. Of course, it is also possible to recognize multiple words that satisfy the waveform similarity. For example, the recognized result is "filter element", "aluminum core", "green heart", etc., and it can be further judged whether the semantics of the recognized result match the current voice command , in this example "filter element" matches the current voice command, but the semantics of "aluminum core" and "green heart" do not match, you need to correct the "filter element" voice command and execute the correct voice command "query Water purifier filter life".

In an optional implementation of the embodiment of the present application, as shown in FIG. 5 , the device in the embodiment of the present application further includes: a backup module 52, configured to convert the first word to Back up to the database.

It can be seen that in the embodiment of the present application, the first word can be added to the database. In a specific example, take "query the travel life of the water purifier" as an example, you can add "travel" to the database, and next time it is still "query the travel life of the water purifier", you can directly execute the "query "Water purifier travel life" voice command, there is no need to adjust it, because its correct semantics can already be known, that is, "query the life of the water purifier filter element". That is to say, in this application, even if the pronunciation is not standard, its correct semantics can be directly recognized, which improves the experience effect of voice command control smart devices.

As shown in Figure 6, the embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 are completed through the communication bus 114 mutual communication,

Memory 113, used to store computer programs;

In one embodiment of the present application, when the processor 111 is used to execute the program stored on the memory 113, it implements the execution method of the smart device voice command provided by any one of the above method embodiments, and its functions are similar , which will not be repeated here.

The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method for executing a voice command on a smart device as provided in any one of the above method embodiments are implemented. .

It should be noted that in this article, relative terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply these No such actual relationship or order exists between entities or operations. Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus comprising a set of elements includes not only those elements, but also includes elements not expressly listed. other elements of or also include elements inherent in such a process, method, article, or device. Without further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising said element.

The above descriptions are only specific embodiments of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Accordingly, the present invention will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims (10)

1. An execution method of an intelligent device voice command, characterized in that, comprising: Obtain the user's voice command to control the smart device; Segmenting the sentence corresponding to the voice instruction, and comparing the word segmentation result with the words in the preset database; comparing the acoustic wave image of the first word whose word segmentation result is not in the database with the acoustic wave image of the second word in the database; When the similarity of the acoustic wave pattern exceeds a first preset threshold, the voice instruction is executed in combination with the semantics of the second word. 2. The method according to claim 1, wherein the comparing the word segmentation result with the words in the preset database includes: Count the words that exist in the database in the word segmentation results; The number of words existing in the database in the word segmentation result is compared to the number of all words in the word segmentation result to obtain the effective rate of the voice command; When the effectiveness rate exceeds a second preset threshold, determining that the voice instruction is an effective instruction; In the case that the effective rate does not exceed the second preset threshold, it is determined that the voice instruction is an invalid instruction; wherein, the first word in the effective instruction that does not exist in the database is compared with the acoustic wave operation. 3. The method according to claim 1, characterized in that, comparing the acoustic wave pattern of the first word whose word segmentation result is not in the database with the acoustic wave pattern of the second word in the database includes : Obtaining waveform parameters of the acoustic wave diagram of the first word, and waveform parameters of the acoustic wave diagram of the second word, wherein the waveform parameters include at least one of the following: amplitude, frequency; comparing the waveform parameters of the sonogram of the first word with the waveform parameters of the sonogram of the second word; Acquiring a ratio between the waveform parameter of the acoustic wave diagram of the first word and the waveform parameter of the acoustic wave diagram of the second word, wherein the ratio is used to characterize the similarity. 4. The method according to claim 1, wherein, after performing the voice instruction in conjunction with the semantics of the second word, the method further comprises: The first word is backed up in the database. 5. The method according to claim 1, wherein the performing the voice instruction in conjunction with the semantics of the second word comprises: Amending the first word in the word segmentation result to the second word; Based on the semantics of the voice instruction corresponding to the modified word segmentation result, the voice instruction corresponding to the modified word segmentation result is executed. 6. An execution device for a smart device voice command, characterized in that it comprises: An acquisition module, configured to acquire a voice command for the user to control the smart device; A processing module, configured to perform word segmentation on the sentence corresponding to the voice instruction, and compare the word segmentation result with words in a preset database; A comparison module, used to compare the acoustic wave graph of the first word whose word segmentation result is not in the database with the acoustic wave graph of the second word in the database; An execution module, configured to execute the voice instruction in combination with the semantics of the second word when the similarity of the acoustic wave pattern exceeds a first preset threshold. 7. The device according to claim 6, wherein the processing module comprises: A statistical unit, used to count the words existing in the database in the word segmentation result; A processing unit, configured to compare the number of words existing in the database in the word segmentation result with the number of all words in the word segmentation result to obtain the effectiveness of the voice instruction; A first determining unit, configured to determine that the voice instruction is a valid instruction when the effective rate exceeds a second preset threshold; The second determining unit is configured to determine that the voice instruction is an invalid instruction when the effective rate does not exceed the second preset threshold; wherein, among the effective instructions that do not exist in the database The first word performs the operation of comparing the sonogram. 8. The device according to claim 6, wherein the comparison module comprises: The first acquisition unit is configured to acquire waveform parameters of the acoustic wave diagram of the first word and waveform parameters of the acoustic wave diagram of the second word, wherein the waveform parameters include at least one of the following: amplitude and frequency; A comparison unit, configured to compare the waveform parameters of the acoustic wave diagram of the first word with the waveform parameters of the acoustic wave diagram of the second word; The second acquisition unit is configured to acquire a ratio between the waveform parameter of the acoustic wave diagram of the first word and the waveform parameter of the acoustic wave diagram of the second word, wherein the ratio is used to characterize the similarity. 9. An electronic device, characterized in that it comprises a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory complete mutual communication through the communication bus; memory for storing computer programs; The processor is configured to implement the method steps of any one of claims 1-5 when executing the program stored in the memory. 10. A computer-readable storage medium, on which a computer program is stored, wherein, when the computer program is executed by a processor, the method steps according to any one of claims 1-5 are implemented.

Images