Intelligent voice recognition and control method and system for digital oscilloscope

Abstract

According to the intelligent voice recognition and control method and system for the digital oscilloscope, voice information of an instrument user is recognized, a recognition result is subjected to word segmentation processing, a word sequence obtained after word segmentation is corrected, then instrument instruction types are matched, and the accuracy of the instrument instruction types is improved. Performing keyword matching on the corrected text instruction and each instruction in the matched instrument instruction type, taking the text instruction, the program instruction and the instruction parameter of which the similarity is greater than a set threshold value as a calling instruction of the instrument, and calling a corresponding program control instruction according to the calling instruction so as to control the instrument. Compared with an existing mode of directly extracting keywords based on voice information to perform matching of each text, the scheme of the invention is short in time consumption and higher in recognition accuracy, and can recognize parameters in the voice information and apply the parameters to the call instruction.

Description

Technical Field

[0001] The present invention belongs to the technical field related to electronic measuring equipment, and in particular relates to an intelligent voice recognition and control method and system for a digital oscilloscope. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] When using a digital oscilloscope, users often test the circuit with both hands at the same time. In this case, another person is needed to operate the digital oscilloscope, which increases the difficulty and cost of debugging. At this time, if the tester sends instructions to the digital oscilloscope through voice, the digital oscilloscope can "understand" the instructions and make corresponding settings, which can greatly improve the flexibility of using the digital oscilloscope. Therefore, a method is invented to connect wireless headphones or microphones / speakers to the digital oscilloscope, and then the digital oscilloscope can be controlled by oral instructions without touching the screen or operating buttons, which frees both hands and provides testing flexibility and efficiency. For example, the user can wake up the instrument by saying "Hello, Xiaosi"; the user can turn on the analog channel 1 of the digital oscilloscope by saying "Please turn on channel 1"; the user can set the time base gear of the digital oscilloscope to 1ns by saying "Adjust the time base gear to 1ns".

[0004] In the past, if users wanted to control instruments through program control instructions, they needed to write a lot of program control instruction codes. The format and parameters of each program control instruction had to be compared one by one with the program control manual. If errors occurred, the program control could not be correct, which often took up a lot of engineers' time. If the commands that need to be program controlled can be input into the voice recognition module through voice, the module can accept multiple user-spoken voice control instructions and translate them into program control instructions. Users can click the C#, C++, and Python buttons on the interface to save the translated program control instructions as test scripts in the corresponding language. This can greatly facilitate the use of test projects and improve test efficiency.

[0005] The traditional operation methods of digital oscilloscopes include oscilloscope hardware panel button operation, touch screen operation, mouse and keyboard operation, etc. These operations require manual operation by the tester. In recent years, with the development of technologies such as deep learning, artificial intelligence, and speech recognition, many intelligent speech recognition applications have also entered people's lives. Voice intelligent control of electronic measuring instruments is also a good development direction, and many measuring instrument manufacturers are also launching their own intelligent products. Intelligent voice recognition and control functions can bring a new user experience to oscilloscope users, greatly improving ease of use, operating experience, and test efficiency.

[0006] At present, the intelligent voice interaction methods for electronic measuring instruments include intelligent keyword matching methods and voice chip recognition methods. The intelligent keyword matching method is to recognize the user's voice into text, extract the keywords in the text, and then match the command. This method does not have a semantic analysis function, and the user description must contain keywords to be recognized. If the user speaks in a colloquial way or changes the way he speaks, it is often not recognized. The voice chip recognition algorithm realizes the intelligent instrument voice system, which is essentially keyword matching. The first step is to enter the voice information and store the voice words required by the instrument in a certain order in the voice chip. The second step is to match the user's voice with the stored voice words. Compared with the intelligent keyword matching method, this method recognizes fewer words and has a lower recognition success rate.

[0007] The current intelligent voice control method for electronic measuring instruments has the following shortcomings: the tester's voice cannot carry parameters, and the parameter values ​​in the user's voice cannot be recognized; the instrument cannot be controlled through voice files; the control instructions after voice recognition cannot be converted into test scripts, etc. Summary of the invention

[0008] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides an intelligent voice recognition and control method and system for a digital oscilloscope. The solution of the present invention is time-saving, has a higher recognition accuracy, and can recognize parameters in voice information and apply them to call instructions.

[0009] To achieve the above object, the first aspect of the present invention provides an intelligent speech recognition and control method for a digital oscilloscope, comprising:

[0010] Recognize the voice information of the instrument user and obtain text instructions;

[0011] Segmenting the text instruction to obtain multiple word sequences, and inputting the multiple word sequences into the trained correction models respectively to obtain corresponding correction results;

[0012] Extract keywords from the corrected text instructions, match instrument instruction types according to the extracted keywords, and obtain text instructions, program instructions, and instruction parameters related to instrument control intent;

[0013] Calculate the semantic similarity between the corrected text instruction and each instruction in the matched instrument instruction type;

[0014] The text instructions, program instructions and instruction parameters corresponding to the semantic similarity degree greater than the set threshold are used as the calling instructions of the instrument, and the corresponding program control instructions are called according to the calling instructions to control the instrument.

[0015] A second aspect of the present invention provides an intelligent speech recognition and control system for a digital oscilloscope, comprising:

[0016] A recognition module is used to recognize the voice information of the instrument user and obtain text instructions;

[0017] A correction module, used for segmenting the text instruction to obtain multiple word sequences, and inputting the multiple word sequences into the trained correction model to obtain corresponding correction results;

[0018] A matching module is used to extract keywords from the corrected text instructions, match the instrument instruction types according to the extracted keywords, and obtain text instructions, program instructions and instruction parameters related to the instrument control intention;

[0019] A calculation module, used for calculating the semantic similarity between the corrected text instruction and each instruction in the matched instrument instruction type;

[0020] The calling control module is used to use the text instructions, program instructions and instruction parameters corresponding to the semantic similarity degree greater than the set threshold as the instrument calling instructions, and to call the corresponding program control instructions according to the calling instructions to control the instrument.

[0021] The third aspect of the present invention provides a computer device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, an intelligent speech recognition and control method for a digital oscilloscope is executed.

[0022] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, an intelligent speech recognition and control method for a digital oscilloscope is executed.

[0023] One or more of the above technical solutions have the following beneficial effects:

[0024] In the present invention, the voice information of the instrument user is recognized, the recognition result is segmented, the word sequence after segmentation is corrected, and then the instrument instruction type is matched, and then the corrected text instruction is matched with each instruction in the matched instrument instruction type for keywords, and the text instructions, program instructions and instruction parameters with similarity greater than a set threshold are used as instrument call instructions, and the corresponding program control instructions are called according to the call instructions to control the instrument. Compared with the existing method of directly extracting keywords based on voice information to match each text, the solution of the present invention is short in time, has a higher recognition accuracy, and can recognize parameters in voice information and apply them to call instructions.

[0025] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0027] Figure 1 This is an interactive diagram of an oscilloscope intelligent speech recognition and control system in Embodiment 1 of the present invention;

[0028] Figure 2 This is a flowchart of the intelligent speech recognition and control method in Embodiment 1 of the present invention;

[0029] Figure 3 This is a flowchart of semantic analysis in Embodiment 1 of the present invention;

[0030] Figure 4 This is a flowchart of instruction and parameter matching in Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0031] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0032] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0033] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0034] Embodiment 1

[0035] This embodiment is based on deep learning algorithms and data intelligent processing technology, and realizes a set of voice wake-up, voice recognition, semantic analysis, command matching and parameter matching, command voice broadcast, command interface display, and command list conversion into test scripts in multiple languages. By receiving the voice signal spoken by the tester, voice wake-up is first performed, and after successful wake-up, the voice signal is recognized as text information, and then the control command desired by the tester is obtained after the steps of semantic analysis, command and parameter matching, and after obtaining the control command, the control command is broadcast through the microphone and displayed on the interface, and after accepting the user's test script conversion operation, the selected control command is generated into a test script in multiple languages. The functions of voice wake-up, voice recognition, semantic analysis, command matching, voice control, command voice broadcast, command display, and voice control command to test script conversion of the digital oscilloscope are realized, and a voice operation mode is provided for the tester, which greatly improves the operational flexibility of the digital oscilloscope and the test efficiency of the user. Users can customize wake-up words to realize voice wake-up function; users can realize voice recognition and control by speaking through microphone and recording voice files in advance; the method and system have semantic analysis function, which can recognize the same voice control command in colloquial or different ways; it can recognize the user's voice parameters and directly splice the parameters into the control instructions; it can list the control commands said by the user and convert them into test scripts by selecting different languages ​​through visual buttons.

[0036] The intelligent voice recognition and control interaction diagram between the tester and the oscilloscope in this embodiment is shown in FIG. Figure 1 As shown. The tester records a voice file, and the oscilloscope reads the voice file to realize intelligent voice recognition and control. The oscilloscope displays the successful control command through the UI interface, and broadcasts the control command through the speaker or headphones. The tester can also dictate voice commands, connect the oscilloscope through a microphone or headphones to read the voice, and the oscilloscope reads the voice file to realize intelligent voice recognition and control. The oscilloscope displays the successful control command through the UI interface, and broadcasts the control command through the speaker or headphones.

[0037] This embodiment discloses intelligent voice recognition and control for a digital oscilloscope, including voice wake-up, voice input, voice recognition, semantic analysis, command and parameter matching, control command interface display and test script generation. The principle block diagram of the oscilloscope intelligent voice recognition and control system of this embodiment is as follows Figure 2 shown.

[0038] 1. Set the custom wake-up word "Hello, Xiaosi" for the oscilloscope voice wake-up. The wake-up module recognizes the set wake-up word, monitors the wake-up word in real time, and responds in time. When the system does not receive voice input within a certain time threshold or the wake-up module receives the sleep activation word, the system switches to sleep mode and starts to monitor the wake-up word, preparing for offline wake-up. The process of exiting the recognition system requires a period of time without input or sleep word activation.

[0039] 2. The voice input module can accept the voice control commands spoken by the tester, or it can record voice files through the UI interface and read the voice files to realize voice input.

[0040] 3. The function of the speech recognition module is to recognize the speech signal input into the instrument, extract the Chinese, digital and English semantic information in the speech, and output the recognition result. There are three steps to realize this function: 1) Dataset construction: According to the "Oscilloscope Programming Manual", a speech recognition data set is produced, and multiple samples containing common phrases are expanded for each instruction, and N samples are produced for each instruction. The samples are composed of speech signals and text information; 2) Model selection and training: In order to meet the time response requirements, a lightweight speech recognition model based on the Transformer structure is selected, which has a relatively ideal recognition rate while ensuring low time consumption, and the data set is trained based on the pre-trained model; 3) After the model is output, the results need to be classified and decoded to generate the corresponding spoken commands.

[0041] 4. The semantic analysis module mainly standardizes and corrects the text instructions obtained by voice recognition, so that the corrected instructions are more consistent with the description of the oscilloscope standard instructions. The main idea of ​​this module is to segment the voice recognition results to obtain multiple word sequences. Then, multiple words are input into the trained correction model to obtain the correction results of each word. The voice analysis flow chart is as follows: Figure 3 shown.

[0042] The semantic analysis process includes two steps: first, the results of speech recognition are segmented using common word segmentation tools, such as jieba word segmentation, ltp word segmentation, and ir word segmentation; second, word standardization and correction processing is performed based on the FastCorrect model to make it closer to the oscilloscope standardized instructions.

[0043] 5. The instruction and parameter matching module is mainly based on the standardized and corrected text instructions to query the corresponding program instructions required by the oscilloscope. The main idea of ​​this module is to extract the keywords in the text instructions, first query the type of instructions based on the keywords, and then calculate the similarity between each instruction and the recognized text instructions in this category, select the instructions with the highest similarity and higher than the set threshold and return them to the user, and call up the corresponding oscilloscope program instructions. If the similarity is not high, ask the user to describe it again. The instruction and parameter matching process is as follows: Figure 4 shown.

[0044] (1) Keyword matching: Since there are many complex instructions, it would take a lot of time to directly compare each text instruction with the speech recognition results obtained in step 3. Therefore, the text results of speech recognition are segmented using a word segmentation algorithm (Jieba word segmentation tool), and the corrected results after segmentation are matched with the obtained keyword dictionary to find the keywords contained in the text. The keyword search adopts a string matching method, that is, comparing whether the string of the word in the correction result is the same as a keyword in the keyword dictionary. If they are the same, the match is successful, thereby determining the type of instruction being queried. If the text does not contain the keyword in the dictionary, it is classified as other categories.

[0045] (2) Semantic similarity calculation: Based on keyword matching, the scope of voice input instructions can be queried. We calculate the similarity between the corrected segmented text and all segmented instructions in the category within this scope. The similarity calculation mainly chooses cosine similarity, which is calculated as shown in formula (1). Build a dictionary that contains the corrected results after segmentation and all the words in the segmented text instructions. According to the dictionary, onehot encode the results and instructions, that is, calculate the number of times each word in the dictionary appears in the two segmented sentences. Then we can use formula (1) to calculate the cosine distance between the two segmented sentences.

[0046]

[0047] Where, Cos(θ): the cosine distance between the corrected result after word segmentation and a certain instruction in the instruction dictionary; X i : The i-dimensional value of the onehot encoding of the corrected result after word segmentation; Y i : The i-dimensional value of a onehot encoding of an instruction in the instruction dictionary.

[0048] Select the text instruction in the instruction library with the greatest similarity and higher than the set threshold as our matching result. Query the corresponding program instruction based on the matched instruction, that is, the control statement that the required oscilloscope can recognize. If the obtained similarity is lower than the threshold, the speech recognition fails and the voice input by the user does not meet the instruction input requirements of the oscilloscope.

[0049] (3) Command completion: Since most commands contain parameters, and most of the queried program statements contain multiple parameter types, such as “:INPut <n>[:STATe] <bool>}" <n> 、 <bool>Each parameter needs to be supplemented according to the result of speech recognition. According to the instruction parameters in the Oscilloscope Programming Manual, it is found that there are no more than 5 parameters in each instruction. Therefore, the parameter completion can be achieved by the if-else query method.

[0050] 6. The command voice broadcast module realizes intelligent voice recognition and voice prompts of control status. When the recognition is successful, the standard statement of the current control command will be broadcast to the user. When the recognition is unsuccessful, the user will be informed that the input command was not executed successfully.

[0051] 7. The interface display module realizes the presentation of intelligent voice recognition and control status to the user in the form of a list in the software interface. When the recognition is successful, the standard program control commands and parameters of the current control command are updated to the list in the form of a text list. When the recognition is unsuccessful, the recognition error information is updated to the list in the form of a text list.

[0052] 8. The test script generation module converts the successfully recognized program control commands selected by the user into test scripts in multiple languages ​​such as C, C#, Python, etc.

[0053] In this embodiment, the user can customize the wake-up word to realize the voice wake-up function; the user can realize voice recognition and control by speaking through the microphone and recording the voice file in advance; it has a semantic analysis function, which can recognize the same voice control command in colloquial or different ways; it can recognize the user's voice parameters and directly splice the parameters into the control instructions; it can list the control commands said by the user, and select different languages ​​through the visual buttons to convert them into test scripts.

[0054] Embodiment 2

[0055] The purpose of this embodiment is to provide an intelligent speech recognition and control system for a digital oscilloscope, comprising:

[0056] A recognition module is used to recognize the voice information of the instrument user and obtain text instructions;

[0057] A correction module, used for segmenting the text instruction to obtain multiple word sequences, and inputting the multiple word sequences into the trained correction model to obtain corresponding correction results;

[0058] A matching module is used to extract keywords from the corrected text instructions, match the instrument instruction types according to the extracted keywords, and obtain text instructions, program instructions and instruction parameters related to the instrument control intention;

[0059] A calculation module, used for calculating the semantic similarity between the corrected text instruction and each instruction in the matched instrument instruction type;

[0060] The calling control module is used to use the text instructions, program instructions and instruction parameters corresponding to the semantic similarity degree greater than the set threshold as the instrument calling instructions, and to call the corresponding program control instructions according to the calling instructions to control the instrument.

[0061] Embodiment 3

[0062] The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the program.

[0063] Embodiment 4

[0064] The purpose of this embodiment is to provide a computer-readable storage medium.

[0065] A computer-readable storage medium stores a computer program, which executes the steps of the above method when executed by a processor.

[0066] The steps involved in the apparatuses of the above embodiments 2, 3 and 4 correspond to the method embodiment 1, and the specific implementation methods can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0067] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0068] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.< / bool> < / n> < / bool> < / n>

Claims

1. An intelligent speech recognition and control method for a digital oscilloscope, characterized in that: include: Recognize the voice information of the instrument user and obtain text instructions; Segmenting the text instruction to obtain multiple word sequences, and inputting the multiple word sequences into the trained correction models respectively to obtain corresponding correction results; Extract keywords from the corrected text instructions, match instrument instruction types according to the extracted keywords, and obtain text instructions, program instructions, and instruction parameters related to instrument control intent; Calculate the semantic similarity between the corrected text instruction and each instruction in the matched instrument instruction type; The text instructions, program instructions and instruction parameters corresponding to the semantic similarity degree greater than the set threshold are used as the calling instructions of the instrument, and the corresponding program control instructions are called according to the calling instructions to control the instrument.

2. The intelligent speech recognition and control method for a digital oscilloscope according to claim 1, characterized in that: The voice information of the instrument user is recognized to obtain text instructions, which are as follows: According to the instrument programming manual, each instruction is expanded into multiple samples containing common words to build a speech recognition data set; The constructed speech recognition dataset is used as a training sample to train the Transformer-based lightweight speech recognition model to obtain a trained lightweight speech recognition model; Based on the trained lightweight speech recognition model, the voice information of the instrument user is recognized to obtain text instructions.

3. The intelligent speech recognition and control method for a digital oscilloscope according to claim 1, characterized in that: Calculate the semantic similarity between the corrected text instruction and each instruction in the matched instrument instruction type, specifically: calculate the cosine distance between the corrected word sequence and each instruction in the instrument instruction type; use the calculated cosine distance as the similarity between the two segmented sentences.

4. The intelligent speech recognition and control method for a digital oscilloscope according to claim 1, characterized in that: It also includes instruction completion, specifically: according to the matched instruction parameters, parameter completion is achieved by using the if-else query method.

5. The intelligent speech recognition and control method for a digital oscilloscope according to claim 1, characterized in that: The text instruction is segmented by IEBA segmentation, LTP segmentation or IR segmentation.

6. The intelligent speech recognition and control method for a digital oscilloscope according to claim 1, characterized in that: Correct word sequences based on the FastCorrect model.

7. An intelligent speech recognition and control system for a digital oscilloscope, characterized in that: include: A recognition module is used to recognize the voice information of the instrument user and obtain text instructions; A correction module, used for segmenting the text instruction to obtain multiple word sequences, and inputting the multiple word sequences into the trained correction model to obtain corresponding correction results; A matching module is used to extract keywords from the corrected text instructions, match the instrument instruction types according to the extracted keywords, and obtain text instructions, program instructions and instruction parameters related to the instrument control intention; A calculation module, used for calculating the semantic similarity between the corrected text instruction and each instruction in the matched instrument instruction type; The calling control module is used to use the text instructions, program instructions and instruction parameters corresponding to the semantic similarity degree greater than the set threshold as the instrument calling instructions, and to call the corresponding program control instructions according to the calling instructions to control the instrument.

8. An intelligent speech recognition and control system for a digital oscilloscope as claimed in claim 7, characterized in that: In the recognition module, the voice information of the instrument user is recognized to obtain text instructions, which are specifically: According to the instrument programming manual, each instruction is expanded into multiple samples containing common words to build a speech recognition data set; The constructed speech recognition dataset is used as a training sample to train the Transformer-based lightweight speech recognition model to obtain a trained lightweight speech recognition model; Based on the trained lightweight speech recognition model, the voice information of the instrument user is recognized to obtain text instructions.

9. A computer device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, an intelligent speech recognition and control method for a digital oscilloscope as described in any one of claims 1 to 7 is executed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the intelligent speech recognition and control method for a digital oscilloscope according to any one of claims 1 to 7 is executed.

Images