Intelligent sound box control method, device and equipment and storage medium

Abstract

The invention relates to the technical field of data processing, and provides an intelligent sound box control method, device and equipment and a storage medium, and the method comprises the steps: collecting environment sound data of an intelligent sound box, and carrying out the frequency analysis, and obtaining different components; generating user instruction data according to the low-frequency component, performing association analysis on the user instruction data and the intermediate-frequency component to obtain an instruction confirmation signal, performing noise filtering on the high-frequency component to obtain an equipment control instruction and a user interaction feedback instruction, and generating an updated playlist based on the equipment control instruction; and integrating the user interaction feedback instruction and the updated playlist to obtain a playlist responding to the user interaction, and controlling the intelligent sound box according to the playlist responding to the user interaction. Through frequency division analysis and association processing of the target sound, the sound recognition accuracy and operation responsivity of the intelligent sound box in a complex environment are improved, and the problem of recognition errors caused by background noise interference in the complex environment is improved.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and in particular to a smart speaker control method, device, equipment and storage medium. Background Art

[0002] In recent years, with the rapid development of artificial intelligence and Internet of Things technologies, smart speakers, as an important part of smart homes, have become increasingly popular among consumers. They can not only provide users with services such as music playback, information query and voice interaction, but also control other smart devices through voice commands. There are many types of smart speakers on the market, and major technology companies have launched their own products, striving to improve user experience through advanced voice recognition and processing technologies.

[0003] Among the related technical means, a method combining speech recognition and environmental sound analysis is used. This method usually generates user commands and matches them with previously defined device commands by picking up and processing voice signals in the environment. Through this method, smart speakers achieve relatively accurate speech recognition and the execution of corresponding functions. For example, after the user issues a command, the speaker can correctly understand and respond to perform related operations such as playing music and checking the weather.

[0004] Regarding the above technical solution, although voice recognition technology can be used to efficiently realize device control and service provision, in complex environments, there is background noise interference that leads to recognition errors. Especially in noisy environments or when the user's voice is not clear enough, the speaker will misunderstand the user's instructions or fail to perform specific operations. Summary of the invention

[0005] In order to improve the problem of recognition errors caused by background noise interference in complex environments, the present application provides a smart speaker control method, device, equipment and storage medium.

[0006] The present invention provides a smart speaker control method, comprising: collecting environmental sound data of the smart speaker, performing frequency analysis on the collected environmental sound data to obtain low-frequency components, medium-frequency components and high-frequency components; performing user voice recognition on the low-frequency components to generate user instruction data, and performing correlation analysis on the user instruction data and the medium-frequency components to obtain an instruction confirmation signal; performing noise filtering on the high-frequency components according to the instruction confirmation signal, and performing voice gain and separation on the instruction confirmation signal after noise filtering and the high-frequency components to generate a clear voice signal and a background music signal; performing semantic analysis on the clear voice signal to obtain a control command, performing situation matching on the control command and the background music signal to obtain a scene adaptation control signal and a situation music playback signal; performing logical judgment on the scene adaptation control signal to obtain a device control command and a user interaction feedback command, and using the device control command and the situation music playback signal to update a play list of the smart speaker to obtain an updated play list; integrating the user interaction feedback command and the updated play list to obtain a play list that responds to user interaction, and controlling the smart speaker according to the play list that responds to user interaction.

[0007] As a preferred solution, the steps of collecting ambient sound data of the smart speaker, performing frequency analysis on the collected ambient sound data, and obtaining low-frequency components, intermediate-frequency components, and high-frequency components include: using a microphone array to collect data on the environment in which the smart speaker is located to obtain ambient sound data; inputting the ambient sound data into a digital signal processor, performing frequency analysis on the ambient sound data through fast Fourier transform, taking frequency components below 300 Hz as low-frequency components, taking frequency components between 300 Hz and 3400 Hz as intermediate-frequency components, and taking frequency components above 3400 Hz as high-frequency components.

[0008] As a preferred embodiment, the step of performing user voice recognition on the low-frequency component, generating user command data, performing correlation analysis on the user command data and the intermediate frequency component, and obtaining a command confirmation signal includes: performing user voice recognition on the low-frequency component through a preset deep neural network model to obtain a speech feature vector, an acoustic model matching degree, and a confidence score, matching the speech feature vector to obtain a matching result, and using the matching result as an initial speech recognition result; performing phoneme adjustment on the initial speech recognition result according to the acoustic model matching degree to obtain an adjusted speech recognition result, fusing the adjusted speech recognition result with the confidence score to obtain user command data; inputting the user command data into a text semantic analysis module for feature extraction to obtain a semantic vector, performing time series correlation analysis on the semantic vector and the intermediate frequency component to obtain a time-frequency correlation matrix, performing principal component analysis on the time-frequency correlation matrix to obtain a command confirmation signal.

[0009] As a preferred solution, the steps of performing noise filtering on the high-frequency components according to the instruction confirmation signal, and performing voice gain and separation on the instruction confirmation signal and the high-frequency components after noise filtering to generate a clear voice signal and a background music signal include: inputting the instruction confirmation signal into a preset frequency filter, performing frequency selective filtering, and obtaining an initial noise suppression result and a filtered processed signal; performing energy detection and voice enhancement processing on the initial noise suppression result by a dynamic compression method to obtain a gain-adjusted processed signal and a noise suppression signal; performing time-frequency domain cross-correlation analysis on the gain-adjusted processed signal and the filtered processed signal by an adaptive filtering algorithm to obtain a gain adjustment signal, and optimizing the gain adjustment signal and the high-frequency components to obtain a clear voice signal; and performing background music separation processing on the noise suppression signal and the clear voice signal by a mixed signal processing module to obtain a background music signal.

[0010] As a preferred solution, the steps of semantically parsing the clear voice signal to obtain a control command, context-matching the control command and the background music signal to obtain a scene-adaptive control signal and a context music playback signal include: inputting the clear voice signal into a context analysis processor for semantic analysis to obtain a semantic feature vector; semantically classifying the semantic feature vector through a pattern recognition algorithm to obtain a preliminary semantic category and a detailed semantic description; inputting the preliminary semantic category into a preset semantic analysis model to perform intent recognition and semantic contradiction processing to obtain an intent recognition result; generating a preliminary control command based on the intent recognition result, inputting the preliminary control command and the detailed semantic description into a context understanding module for context modeling to obtain a scene perception result; performing context music adaptation processing on the scene perception result and the background music signal through a sound field adaptation processor to obtain a context music playback signal; performing audio encoding and sound effect enhancement processing on the context music playback signal using a sound effect encoding processor to obtain a music playback signal and a scene-adaptive control signal suitable for the current context.

[0011] As a preferred scheme, the step of performing logical judgment on the scene adaptation control signal to obtain device control instructions and user interaction feedback instructions, and using the device control instructions and the contextual music playback signal to update the play list of the smart speaker to obtain an updated play list includes: inputting the scene adaptation control signal into a decision logic processor, performing logical judgment and conditional decision on the scene adaptation control signal through conditional judgment and logical reasoning preset in the decision logic processor to obtain a logical judgment result; generating device control instructions and user interaction feedback instructions according to the logical judgment result; inputting the device control instructions into the playback management system of the smart speaker to perform play list update processing, and optimizing the audio arrangement in combination with the contextual music playback signal to generate a new play list; and integrating the new play list and user interaction feedback instructions using a data integration processor to obtain an updated play list.

[0012] As a preferred scheme, the step of integrating the user interaction feedback instruction and the updated playlist to obtain a playlist that responds to user interaction, and controlling the smart speaker according to the playlist that responds to user interaction includes: inputting the user interaction feedback instruction into an interaction response generator to generate a response strategy and behavior to obtain user response data; generating user interaction response data and demand management instructions according to the user response data; inputting the demand management instructions into a playlist optimization processor, optimizing the updated playlist through an optimization algorithm, and generating a playlist that responds to user interaction; inputting the user interaction response data and the playlist that responds to user interaction into a smart speaker control system to execute a program to complete the control of the smart speaker.

[0013] The present application also provides a smart speaker control device, including: a collection module, used to collect environmental sound data of the smart speaker, perform frequency analysis on the collected environmental sound data, and obtain low-frequency components, medium-frequency components, and high-frequency components; an identification module, used to perform user voice recognition on the low-frequency components, generate user command data, and perform correlation analysis on the user command data and the medium-frequency components to obtain a command confirmation signal; a filtering module, used to perform noise filtering on the high-frequency components according to the command confirmation signal, and perform voice gain and separation on the command confirmation signal after noise filtering and the high-frequency components to generate a clear voice signal and a background music signal; a matching module, used to The clear voice signal is semantically parsed to obtain a control command, and the control command and the background music signal are context-matched to obtain a scene adaptation control signal and a context music playback signal; an update module is used to perform logical judgment on the scene adaptation control signal to obtain a device control instruction and a user interaction feedback instruction, and use the device control instruction and the context music playback signal to update the play list of the smart speaker to obtain an updated play list; a control module is used to integrate the user interaction feedback instruction and the updated play list to obtain a play list that responds to user interaction, and control the smart speaker according to the play list that responds to user interaction.

[0014] The present application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the smart speaker control method described in any one of the above is implemented.

[0015] The present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the processor executes the smart speaker control method as described in any one of the above.

[0016] Compared with the prior art, the present application has the following beneficial effects: strong anti-interference and high applicability. Through frequency analysis and correlation processing of target sounds, the sound recognition accuracy and operation responsiveness of smart speakers in complex environments are improved. By distinguishing and independently processing sound components of different frequencies, the risk of misoperation is effectively reduced and the user experience is optimized. The combined use of noise filtering and voice gain makes the generation of clear voice signals more efficient, thereby improving the accuracy of command recognition and the overall response speed of the system, enabling smart speakers to better adapt to different environments and user needs. Through precise scene adaptation control signals and contextual music playback signals, the practicality and interactivity of smart speakers are enhanced, and the problem of recognition errors caused by background noise interference in complex environments is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0018] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0019] Figure 1 is a flow chart of a smart speaker control method provided by an embodiment of the present invention; Figure 2 is a schematic block diagram of the structure of the smart speaker control device provided by an embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0020] Description of reference numerals: 10. Smart speaker control device; 11. Acquisition module; 12. Identification module; 13. Filtering module; 14. Matching module; 15. Update module; 16. Control module; 20. Electronic device; 21. Memory; 22. Processor. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0023] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0024] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0025] The technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0026] Embodiment 1: like Figure 1 As shown, an intelligent speaker control method, apparatus, device and storage medium provided in an embodiment of the present application include steps S100 to S600.

[0027] Step S100: Collect ambient sound data of the smart speaker, perform frequency analysis on the collected ambient sound data, and obtain low-frequency components, mid-frequency components, and high-frequency components.

[0028] In this step, the smart speaker uses its built-in high-sensitivity microphone to collect ambient sound data, and then processes this data through a frequency analysis algorithm; specifically, the algorithm decomposes the sound data into low-frequency components, mid-frequency components, and high-frequency components to facilitate specialized processing in subsequent steps.

[0029] For example, environmental sound data includes voices, music, and TV sounds at family gatherings. These sounds will be subdivided into different frequency components for more accurate sound recognition and processing.

[0030] Step S200: Perform user voice recognition on the low-frequency component to generate user command data, perform correlation analysis on the user command data and the intermediate-frequency component to obtain a command confirmation signal.

[0031] In this step, the system first applies sound recognition technology to the low-frequency components to identify the user's command sound; specifically, the identified user command data will be correlated with the command-related sound information contained in the mid-frequency components to verify and confirm the accuracy of the user's command.

[0032] For example, if the user says "turn on the lights", the system not only recognizes the "turn on the lights" command in the low frequency, but also analyzes the related sound fluctuations in the mid-frequency components to confirm that this command is not misinterpreted due to ambient noise.

[0033] Step S300: noise filtering is performed on the high frequency components according to the instruction confirmation signal, and voice gain and separation are performed on the instruction confirmation signal and the high frequency components after noise filtering to generate a clear voice signal and a background music signal.

[0034] In this step, the command confirmation signal is used to guide the noise filtering of high-frequency components to ensure that only relevant high-frequency sounds are retained; specifically, the clarity of user commands is further improved through voice gain technology, and the clear user voice is effectively separated from the background music signal.

[0035] For example, this step can separate the greeting sound and music in the background of a house party, ensuring that the voice command is not interrupted by the music, while also keeping the background music playing continuously.

[0036] Step S400: semantically analyze the clear voice signal to obtain a control command, and context-match the control command with the background music signal to obtain a scene-adaptive control signal and a contextual music playback signal.

[0037] In this step, the system performs deep semantic analysis on the clear voice signal and converts it into a specific control command; specifically, the control command is combined with the background music signal to perform context matching to generate corresponding scene-adaptive control signals and contextual music playback signals.

[0038] For example, if the user's command is "get ready for bed", the system not only understands this command, but can also adjust the indoor lighting and play soft bedtime music.

[0039] Step S500: perform logical judgment on the scene adaptation control signal to obtain device control instructions and user interaction feedback instructions, and use the device control instructions and situational music playback signals to update the playlist of the smart speaker to obtain an updated playlist.

[0040] In this step, the system performs logical judgment according to the scene adaptation control signal and generates specific device control instructions and user interaction feedback instructions; specifically, the state of smart home devices is adjusted through device control instructions, and the music play list is updated using situational music playback signals.

[0041] For example, based on the user's bedtime command, the smart speaker not only dims the indoor lights, but also automatically adjusts the playlist to a bedtime mode containing soft music.

[0042] Step S600: Integrate the user interaction feedback instruction and the updated playlist to obtain a playlist that responds to the user interaction, and control the smart speaker according to the playlist that responds to the user interaction.

[0043] In this step, the system integrates the user interaction feedback instructions and the updated playlist to optimize the overall interactive experience; specifically, this integration not only ensures the instant update of the playlist, but also feeds back clear information about the operation status to the user.

[0044] For example, a playlist modified by user instructions will be automatically synchronized to the speaker, and the user will be notified of the latest playback status and the music to be played through voice feedback.

[0045] In this embodiment, the ambient sound data of the smart speaker is collected and the frequency analysis is performed on the ambient sound data to decompose it into low-frequency components, medium-frequency components and high-frequency components. Next, the low-frequency components are subjected to user voice recognition to generate user command data, and these user command data are correlated and analyzed with the medium-frequency components to obtain a command confirmation signal. According to this command confirmation signal, the high-frequency components are subjected to noise filtering, and the high-frequency components after noise filtering are combined with the command confirmation signal for voice gain and separation to generate a clear voice signal and a background music signal. Subsequently, the clear voice signal is semantically parsed to obtain a control command, which is context-matched with the background music signal to generate a scene adaptation control signal and a context music playback signal. Finally, the scene adaptation control signal is used for logical judgment to generate device control instructions and user interaction feedback instructions, and these signals are used to update the playlist of the smart speaker, integrating the user interaction feedback instructions and the updated playlist, thereby optimizing and responding to the user's interactive experience.

[0046] Through frequency analysis and correlation processing of target sounds, the sound recognition accuracy and operation responsiveness of smart speakers in complex environments are significantly improved. By distinguishing and independently processing sound components of different frequencies (such as user commands, background noise and background music), the risk of misoperation can be effectively reduced and the user experience can be optimized. At the same time, the combined use of noise filtering and voice gain makes the generation of clear voice signals more efficient, thereby improving the accuracy of command recognition and the overall response speed of the system. It enables smart speakers to better adapt to different environments and user needs, and enhances the practicality and interactivity of smart speakers through precise scene adaptation control signals and contextual music playback signals, and improves the problem of recognition errors caused by background noise interference in complex environments.

[0047] Embodiment 2: In step S100, a microphone array is used to collect data about the environment in which the smart speaker is located to obtain ambient sound data; the ambient sound data is input into a digital signal processor, and the frequency of the ambient sound data is analyzed by fast Fourier transform, and the frequency components below 300 Hz are taken as low-frequency components, the frequency components between 300 Hz and 3400 Hz are taken as intermediate-frequency components, and the frequency components above 3400 Hz are taken as high-frequency components.

[0048] The system uses a microphone array to accurately capture subtle changes in ambient sound, and uses a digital signal processor to implement fast Fourier transform to efficiently separate sound components of different frequencies. Specifically, this analysis helps the system accurately identify the low-frequency part of the human voice, the mid-frequency part of calls or background music, and high-frequency noise.

[0049] For example, this step can distinguish the sound of a television (mid-frequency components), a child's scream (high-frequency components), and the conversation of adults (low-frequency components) in a home environment.

[0050] In step S200, the user voice is recognized on the low-frequency component through a preset deep neural network model to obtain a speech feature vector, an acoustic model matching degree and a confidence score, the speech feature vector is matched to obtain a matching result, and the matching result is used as the initial speech recognition result.

[0051] By using deep neural networks to learn and compare speech features, the accuracy and efficiency of recognition can be improved; specifically, the acoustic model gives a confidence score by comparing the matching degree of feature vectors to ensure that the recognized command reflects the user's true intention.

[0052] For example, the system can learn to distinguish between the user's "turn on the lights" and "turn off the lights" commands and execute them accurately even in the presence of some background noise.

[0053] The initial speech recognition result is phoneme-adjusted according to the acoustic model matching degree to obtain an adjusted speech recognition result, and the adjusted speech recognition result is fused with the confidence score to obtain user command data.

[0054] The recognition results are optimized by adjusting phonemes to make them more in line with linguistic rules; specifically, the integration of confidence scores further verifies the reliability of the instructions and provides a solid foundation for subsequent processing.

[0055] For example, if a user says "play music," the system adjusts the phonemes to accurately identify "play" rather than "pause."

[0056] The user command data is input into the text semantic analysis module for feature extraction to obtain a semantic vector. The semantic vector and the intermediate frequency component are subjected to time series correlation analysis to obtain a time-frequency correlation matrix. The time-frequency correlation matrix is ​​subjected to principal component analysis to obtain a command confirmation signal.

[0057] Through the time-series analysis of text and sound, the correlation between user instructions and environmental sound data is ensured; specifically, the application of principal component analysis extracts the most important features from multidimensional data to form the final confirmation signal.

[0058] For example, when a user issues the command “turn up the volume” while watching a football game, the system can recognize and associate it with the cheers of the audience in the game, thereby confirming the user's intention.

[0059] In step S300, the instruction confirmation signal is input into a preset frequency filter to perform frequency selective filtering to obtain an initial noise suppression result and a filtered processed signal.

[0060] The clarity of speech signals is improved by selectively suppressing unnecessary frequency components through precise filtering technology; specifically, this process strengthens the frequencies related to user commands and reduces the interference of noise.

[0061] For example, if there is air conditioning sound or other mechanical noise in the environment, the system effectively suppresses these high-frequency noises through selective filtering.

[0062] The initial noise suppression result is subjected to energy detection and speech enhancement processing by a dynamic compression method to obtain a processing signal and a noise suppression signal after gain adjustment.

[0063] The dynamic range of the sound signal is adjusted through dynamic compression technology to ensure the consistency and clarity of the speech. Specifically, this step makes the speech part more prominent in the audio signal through gain adjustment, thereby improving intelligibility.

[0064] For example, for low-volume user commands, the system can automatically increase their volume so that they can still be clearly heard against a noisy background.

[0065] The adaptive filtering algorithm is used to perform time-frequency domain cross-correlation analysis on the processed signal after gain adjustment and the processed signal after filtering to obtain the gain adjustment signal, and the gain adjustment signal and the high-frequency component are optimized to obtain a clear speech signal.

[0066] Through the adaptive filtering algorithm, the cross-correlation of signals is effectively analyzed and processed in the time-frequency domain to extract the key components between each other; specifically, the algorithm adjusts the gain of the audio signal in real time, optimizes the processing signal according to the ambient noise level and the characteristics of the speech signal, and ensures that the output speech signal has the best clarity and consistency.

[0067] For example, in a family gathering scene with strong noise interference, the technology can extract the human voice from the background noise, making the sound clear and distortion-free.

[0068] The mixed signal processing module is used to separate the background music from the noise suppression signal and the clear speech signal to obtain the background music signal.

[0069] Through the mixed signal processing module, it is possible to distinguish and extract non-verbal background music components from complex soundscapes; specifically, the module effectively separates the music signal from the speech signal by analyzing the different properties of the signal and using a specific algorithm, allowing the background music to exist independently of the clear speech signal.

[0070] For example, when music is playing in a home environment and a conference call is taking place at the same time, the system is able to distinguish between the two and ensure that the music does not obscure the voice call.

[0071] In step S400, the clear speech signal is input into a context analysis processor for semantic analysis to obtain a semantic feature vector.

[0072] Through the context analysis processor, clear speech signals can be decoded into a set of representative feature vectors that fully describe the semantic information in the speech; specifically, this process converts multi-dimensional speech signals into easy-to-analyze semantic structures, which facilitates subsequent processing and understanding.

[0073] For example, when a user says “turn on music when it rains”, the system converts this sentence into a set of semantic features to identify weather intent and music playback requests.

[0074] The semantic feature vector is divided into semantic categories through pattern recognition algorithm to obtain preliminary semantic categories and detailed semantic descriptions.

[0075] By using pattern recognition algorithms, semantic feature vectors are classified to determine the semantic category to which they belong; specifically, this division provides a clear understanding of the individual components in a complex sentence and forms a detailed semantic description for use in further processing.

[0076] For example, recognizing that some sentences contain multiple requests, such as "turn off the lights and play music," allows the system to process each intent one by one.

[0077] The preliminary semantic categories are input into the preset semantic parsing model for intent recognition and semantic contradiction processing to obtain the intent recognition results.

[0078] Through the preset semantic parsing model, the system can identify the intention in natural language and deal with the semantic contradictions that arise; specifically, the parsing model uses contextual information and semantic rules to analyze and filter out the user's real needs.

[0079] For example, when a user issues multiple conflicting commands at the same time, such as “lower the volume” and “increase the volume,” the system can identify and parse the correct user intent.

[0080] A preliminary control command is generated based on the intention recognition result, and the preliminary control command and detailed semantic description are input into the context understanding module for situation modeling to obtain the scene perception result.

[0081] Through the context understanding module, the parsed commands are placed in their relevant contexts for modeling; specifically, this process not only considers the immediate user commands, but also combines environmental parameters (such as current time, location, etc.) to accurately perceive and predict scene requirements.

[0082] For example, in the instruction "play light music during dinner", the system can automatically adjust the appropriate music playback plan according to the time category.

[0083] The scene perception result and the background music signal are processed for situational music adaptation by a sound field adaptation processor to obtain a situational music playback signal.

[0084] Through the sound field adaptation processor, the background music can be optimally matched to the current scene perception results; specifically, this module will adjust various parameters of the music, such as volume, song style, etc., to adapt to the scene requirements, thereby optimizing the music experience according to the environment and user expectations.

[0085] For example, under warm ambient lighting, the speaker can be adjusted to play warm and soft background music to enhance the atmosphere.

[0086] The audio encoding processor is used to perform audio encoding and sound effect enhancement processing on the situational music playback signal to obtain a music playback signal and a scene adaptation control signal suitable for the current situation.

[0087] Through the sound effect coding processor, the music signal is encoded and the sound effect is enhanced to improve the sound quality and playback effect; specifically, this audio processing not only improves the sound quality, but also adjusts certain frequency components of the music to highlight situation-related details, making it fuller and more moving.

[0088] For example, during celebrations, the processor can make music have more powerful bass effects and clear high-frequency performance, thus enhancing the festive atmosphere.

[0089] In step S500, the scene adaptation control signal is input into the decision logic processor, and the scene adaptation control signal is subjected to logical judgment and conditional decision through conditional judgment and logical reasoning preset in the decision logic processor to obtain a logical judgment result.

[0090] By using conditional judgment and logical reasoning, the decision logic processor can efficiently parse the scene adaptation control signal and make accurate logical judgments based on the current environmental variables and preset rules; specifically, this process comprehensively considers user instructions, environmental characteristics and system status to ensure the rationality and accuracy of instruction execution.

[0091] For example, if a signal is received to increase the volume of background music in a restaurant environment, the system will first determine that the ambient noise level is within safe limits before increasing the volume.

[0092] Generate device control instructions and user interaction feedback instructions based on the logical judgment results.

[0093] By continuously analyzing the results of logical judgment, the system generates corresponding device control instructions and presentable user interaction feedback instructions; specifically, the device control instructions are directly used to adjust the state of the smart device, while the interaction feedback instructions are used to update the feedback information and operation prompts of the user interface.

[0094] For example, when the user asks to "dim the lights", the system will issue a device control instruction to adjust the light source, and at the same time feedback "the lights have been dimmed".

[0095] The device control instructions are input into the playback management system of the smart speaker to update the playlist, and the audio arrangement is optimized in combination with the contextual music playback signal to generate a new playlist.

[0096] By directing device control instructions to the playback management system, the system can instantly update the playlist and execute specific audio arrangement optimization strategies; specifically, the audio arrangement is adapted and organized in combination with the contextual music playback signal to facilitate a smooth and rational music experience.

[0097] For example, based on the user command "play classical music in the evening", the system will automatically adjust the playlist to prioritize soothing classical music.

[0098] The new playlist and the user interaction feedback instruction are integrated and processed by using a data integration processor to obtain an updated playlist.

[0099] Through the data integration processor, the effective integration of multiple signals improves the accuracy of the playlist and the consistency of user interaction feedback; specifically, the system's comprehensive analysis of user preferences and real-time feedback is reflected in the new playlist, making the music selection more in line with the current situation and user expectations.

[0100] For example, during family leisure time, the system automatically adds music that is different from that during work time based on the user's listening history.

[0101] In step S600, the user interaction feedback instruction is input into the interaction response generator to generate a response strategy and behavior, thereby obtaining user response data.

[0102] Through the interactive response generator, the system analyzes user interactive feedback instructions and generates targeted response strategies and behavior plans; specifically, this processing takes into account the user experience and the effectiveness of system feedback to ensure that users feel the real-time response of the smart speaker.

[0103] For example, when a user asks about the content of the currently playing song, the system will generate a voice prompt and display relevant information.

[0104] Generate user interaction response data and demand management instructions based on the user response data.

[0105] By analyzing user response data, interactive response data and demand management instructions that are highly relevant to the current situation are generated; specifically, this process combines user intentions and interaction history to optimize the generation and execution path of subsequent instructions.

[0106] For example, after the user chooses to skip a song, the system quickly adjusts the next song and updates the user's preference record.

[0107] The demand management instruction is input into the playlist optimization processor, and the updated playlist is optimized through the optimization algorithm to generate a playlist that responds to user interaction.

[0108] By applying optimization algorithms, demand management instructions are sorted and content adjusted in playlists to improve user satisfaction and system fluency; specifically, the optimization process improves the rationality of audio content selection and arrangement, making playback more personalized and predictable.

[0109] For example, the system will adjust the playback order to match the user's sports scene needs, giving priority to playing music with fast rhythm.

[0110] The user interaction response data and the playlist responding to the user interaction are input into the smart speaker control system to execute the program to complete the control of the smart speaker.

[0111] By inputting the integrated data into the control system, user commands and audio playback are fully and accurately executed; specifically, this process adjusts the operation method of the smart speaker in real time so that it can accurately represent user intentions and respond quickly to new instructions.

[0112] For example, when the user issues the "enter study mode" command, the system immediately switches to a focused music playlist and adjusts the ambient sound to reduce distractions.

[0113] In this embodiment, by combining advanced signal processing, logic analysis and user interaction strategies, the smart speaker significantly improves the operating efficiency and response accuracy in various scenarios. This embodiment uses multi-level data analysis and optimization technology to achieve full-chain intelligent control from sound capture to user interaction feedback. In particular, the use of decision logic processors enables the system to dynamically process a variety of scene adaptation signals, effectively respond to complex and changeable home environments, and seamlessly integrates contextual awareness and music playback management through intelligent optimization algorithms. This brings users an efficient and customized interactive experience, allowing the device to continuously learn and adapt to the unique needs of users, thereby enhancing the friendliness and practicality of human-computer interaction. This design not only ensures the precise execution of voice commands, but also improves overall user satisfaction and usage experience by continuously optimizing playback content and feedback information.

[0114] Embodiment 3: like Figure 2 As shown, the present application also provides a smart speaker control device 10, including a collection module 11, an identification module 12, a filtering module 13, a matching module 14, an update module 15 and a control module 16.

[0115] The acquisition module 11 is mainly used to collect the ambient sound data of the smart speaker, and perform frequency analysis on the collected ambient sound data to obtain low-frequency components, medium-frequency components and high-frequency components.

[0116] The recognition module 12 is mainly used to perform user voice recognition on low-frequency components, generate user instruction data, perform correlation analysis on the user instruction data and the intermediate-frequency components, and obtain an instruction confirmation signal.

[0117] The filtering module 13 is mainly used to perform noise filtering on the high frequency components according to the instruction confirmation signal, and to perform voice gain and separation on the instruction confirmation signal and the high frequency components after noise filtering to generate a clear voice signal and a background music signal.

[0118] The matching module 14 is mainly used to perform semantic analysis on the clear voice signal to obtain the control command, and to perform situation matching between the control command and the background music signal to obtain the scene adaptation control signal and the situation music playing signal.

[0119] The update module 15 is mainly used to perform logical judgment on the scene adaptation control signal, obtain device control instructions and user interaction feedback instructions, and use the device control instructions and situational music playback signals to update the play list of the smart speaker to obtain an updated play list.

[0120] The control module 16 is mainly used to integrate the user interaction feedback instructions and the updated playlist to obtain a playlist that responds to the user interaction and control the smart speaker according to the playlist.

[0121] In this embodiment, through the modular design of the smart speaker control device 10, each functional module works closely together to improve the overall performance and flexibility of the system. The acquisition module 11 accurately analyzes the frequency components of the ambient sound data, and the recognition module 12 confirms the validity of the user's instructions through detailed sound recognition and data association. The filtering module 13 timely processes high-frequency noise to ensure the clarity of the voice signal and the integrity of the background music. The matching module 14 identifies and adapts to the actual scene of the user through semantic analysis and context matching. The update module 15 is responsible for timely adjustment of the playlist so that the device can not only respond to user instructions in real time, but also optimize the music playback experience according to the current audio environment. The introduction of the control module 16 ensures smooth and efficient user interaction and device control. The combination of these modules enables the smart speaker to not only have a strong ability to adapt to the environment, but also better provide accurate and personalized services, significantly improving user experience and satisfaction.

[0122] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and conciseness of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned smart speaker control method embodiment, and will not be repeated here.

[0123] Embodiment 4: like Figure 3 As shown, the present application also provides an electronic device 20, including a memory 21 and a processor 22, the memory 21 stores a computer program that can be run on the processor 22, and the processor 22 implements the smart speaker control method of Example 1 when executing the computer program.

[0124] In this embodiment, by integrating the memory 21 and the processor 22 in the electronic device 20, the efficient operation of the smart speaker control method is achieved. The computer program stored in the memory 21 is clearly organized and optimized for different operating scenarios to ensure that the processor 22 can respond quickly and operate efficiently when executed. Such a device architecture enables the system to make full use of hardware resources and further improve the execution efficiency of the speaker control method. Not only does it shorten the response time of user commands, but it also increases the stability and reliability of the system. In terms of user experience, it is manifested as more immediate interaction and less processing delay, so users can get a smoother and more personalized smart speaker experience.

[0125] Embodiment 5: The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the processor executes the smart speaker control method as described in Example 1.

[0126] In this embodiment, the flexible performance of executing the smart speaker control method on the processor is achieved through a computer program stored on a computer-readable storage medium. The storage medium provides a reliable distribution and execution mechanism, so that complex speaker control systems can run on different hardware platforms and have wide applicability. This design ensures the popularization and sharing of smart speaker control functions, and can achieve a seamless operating experience on different types of electronic devices while maintaining high standards of performance and functional integration. Users can obtain a highly consistent control method experience through this architecture without being restricted by device diversity and hardware changes, which improves the compatibility and execution efficiency of smart speakers in smart home environments.

[0127] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart speaker control method, characterized in that: include: Collecting ambient sound data of the smart speaker, performing frequency analysis on the collected ambient sound data to obtain low-frequency components, medium-frequency components, and high-frequency components; Performing user voice recognition on the low-frequency component to generate user instruction data, and performing correlation analysis on the user instruction data and the intermediate-frequency component to obtain an instruction confirmation signal; Noise filtering is performed on the high frequency component according to the command confirmation signal, and voice gain and separation are performed on the command confirmation signal and the high frequency component after noise filtering to generate a clear voice signal and a background music signal; Performing semantic analysis on the clear voice signal to obtain a control command, and context-matching the control command with the background music signal to obtain a scene-adaptive control signal and a contextual music playback signal; Performing logical judgment on the scene adaptation control signal to obtain a device control instruction and a user interaction feedback instruction, and updating the playlist of the smart speaker using the device control instruction and the contextual music play signal to obtain an updated playlist; The user interaction feedback instruction and the updated playlist are integrated and processed to obtain a playlist that responds to the user interaction, and the smart speaker is controlled according to the playlist that responds to the user interaction.

2. The smart speaker control method according to claim 1, characterized in that: The step of collecting the ambient sound data of the smart speaker, performing frequency analysis on the collected ambient sound data, and obtaining low-frequency components, medium-frequency components, and high-frequency components includes: A microphone array is used to collect data about the environment in which the smart speaker is located to obtain ambient sound data; the ambient sound data is input into a digital signal processor, and frequency analysis is performed on the ambient sound data through fast Fourier transform, and frequency components below 300 Hz are used as low-frequency components, frequency components between 300 Hz and 3400 Hz are used as intermediate-frequency components, and frequency components above 3400 Hz are used as high-frequency components.

3. The smart speaker control method according to claim 1, characterized in that: The step of performing user voice recognition on the low-frequency component to generate user instruction data, and performing correlation analysis on the user instruction data and the intermediate-frequency component to obtain an instruction confirmation signal includes: Performing user voice recognition on the low-frequency component through a preset deep neural network model to obtain a voice feature vector, an acoustic model matching degree, and a confidence score, matching the voice feature vector to obtain a matching result, and using the matching result as an initial voice recognition result; Performing phoneme adjustment on the initial speech recognition result according to the acoustic model matching degree to obtain an adjusted speech recognition result, and fusing the adjusted speech recognition result with the confidence score to obtain user instruction data; The user instruction data is input into a text semantic analysis module for feature extraction to obtain a semantic vector, the semantic vector and the intermediate frequency component are subjected to time series correlation analysis to obtain a time-frequency correlation matrix, and the time-frequency correlation matrix is ​​subjected to principal component analysis to obtain an instruction confirmation signal.

4. The smart speaker control method according to claim 3, characterized in that: The step of filtering the high frequency component according to the instruction confirmation signal, and performing voice gain and separation on the instruction confirmation signal and the high frequency component after noise filtering to generate a clear voice signal and a background music signal includes: Inputting the instruction confirmation signal into a preset frequency filter to perform frequency selective filtering to obtain an initial noise suppression result and a filtered processed signal; Performing energy detection and speech enhancement processing on the initial noise suppression result by a dynamic compression method to obtain a gain-adjusted processing signal and a noise suppression signal; Using an adaptive filtering algorithm to perform time-frequency domain cross-correlation analysis on the gain-adjusted processed signal and the filtered processed signal to obtain a gain-adjusted signal, and optimizing the gain-adjusted signal and the high-frequency component to obtain a clear speech signal; A mixed signal processing module is used to perform background music separation processing on the noise suppression signal and the clear speech signal to obtain a background music signal.

5. The smart speaker control method according to claim 1, characterized in that: The step of semantically parsing the clear voice signal to obtain a control command, context-matching the control command with the background music signal to obtain a scene adaptation control signal and a context music playback signal comprises: Inputting the clear speech signal into a context analysis processor for semantic analysis to obtain a semantic feature vector; Semantically classify the semantic feature vectors using a pattern recognition algorithm to obtain preliminary semantic categories and detailed semantic descriptions; Inputting the preliminary semantic category into a preset semantic parsing model to perform intent recognition and semantic contradiction processing to obtain an intent recognition result; Generate a preliminary control command according to the intention recognition result, input the preliminary control command and the detailed semantic description into a context understanding module for situation modeling, and obtain a scene perception result; Performing contextual music adaptation processing on the scene perception result and the background music signal through a sound field adaptation processor to obtain a contextual music playback signal; The audio encoding and sound effect enhancement processing is performed on the situational music playback signal by using a sound effect encoding processor to obtain a music playback signal and a scene adaptation control signal suitable for the current situation.

6. The smart speaker control method according to claim 1, characterized in that: The step of performing logical judgment on the scene adaptation control signal to obtain a device control instruction and a user interaction feedback instruction, and updating the playlist of the smart speaker using the device control instruction and the situational music playback signal to obtain an updated playlist includes: Inputting the scene adaptation control signal into a decision logic processor, performing logical judgment and conditional decision on the scene adaptation control signal through conditional judgment and logical reasoning preset in the decision logic processor, and obtaining a logical judgment result; Generate device control instructions and user interaction feedback instructions according to the logic judgment result; Input the device control instruction into the playback management system of the smart speaker to update the playlist, and optimize the audio arrangement in combination with the contextual music playback signal to generate a new playlist; The new playlist and the user interaction feedback instruction are integrated by using a data integration processor to obtain an updated playlist.

7. The smart speaker control method according to claim 6, characterized in that: The step of integrating the user interaction feedback instruction and the updated playlist to obtain a playlist that responds to the user interaction, and controlling the smart speaker according to the playlist that responds to the user interaction includes: Inputting the user interaction feedback instruction into an interaction response generator to generate a response strategy and behavior, and obtaining user response data; generating user interaction response data and demand management instructions according to the user response data; Inputting the demand management instruction into a playlist optimization processor, optimizing the updated playlist by an optimization algorithm, and generating a playlist responsive to user interaction; The user interaction response data and the playlist responding to the user interaction are input into the smart speaker control system to execute the program to complete the control of the smart speaker.

8. An intelligent speaker control device, characterized in that: include: The acquisition module is used to collect the ambient sound data of the smart speaker, and perform frequency analysis on the collected ambient sound data to obtain low-frequency components, medium-frequency components, and high-frequency components; A recognition module, used to perform user voice recognition on the low-frequency component to generate user instruction data, and perform correlation analysis on the user instruction data and the intermediate-frequency component to obtain an instruction confirmation signal; A filtering module, used to perform noise filtering on the high-frequency component according to the instruction confirmation signal, and perform voice gain and separation on the instruction confirmation signal and the high-frequency component after noise filtering to generate a clear voice signal and a background music signal; A matching module, used for semantically analyzing the clear voice signal to obtain a control command, and contextually matching the control command with the background music signal to obtain a scene adaptation control signal and a context music playback signal; An updating module, configured to perform a logical judgment on the scene adaptation control signal, obtain a device control instruction and a user interaction feedback instruction, and update the play list of the smart speaker using the device control instruction and the contextual music playback signal to obtain an updated play list; The control module is used to integrate the user interaction feedback instruction and the updated playlist to obtain a playlist that responds to the user interaction, and control the smart speaker according to the playlist that responds to the user interaction.

9. An electronic device, characterized in that: It includes a memory and a processor, the memory stores a computer program that can be run on the processor, and the processor implements the smart speaker control method described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the processor executes the smart speaker control method as described in any one of claims 1 to 7.

Images