A voice control optimization method for in-vehicle intelligent audio

Abstract

The present invention discloses a voice control optimization method for vehicle-mounted intelligent audio, which relates to the field of vehicle internal control. The method includes: obtaining target driving state information; obtaining a first activation instruction, pre-analyzing the target driving state information based on the first activation instruction, generating a target audio application prediction, and the target audio application prediction includes multiple predicted application categories; obtaining a first selection instruction, and selecting and starting a first target application category among multiple predicted application categories based on the first selection instruction; obtaining a first control strategy for the first target application category, and performing voice control optimization on the first target application category based on the first control strategy. The method solves the technical problem that the existing vehicle-mounted intelligent audio voice control cannot adjust the control strategy according to the driving state and the specific needs of the user, resulting in a poor user experience, and achieves the technical effect of improving the control accuracy and response speed and enhancing the user experience by optimizing the voice control strategy.

Description

Technical Field

[0001] The present application relates to the field related to vehicle internal control, and in particular to a voice control optimization method for in-vehicle smart speakers. Background Art

[0002] With the rapid development of smart cars, in-vehicle entertainment systems, especially in-vehicle smart audio, have become an important part of improving the driving experience. Modern drivers not only expect to enjoy a high-quality audio experience while driving, but also hope that this experience can be intelligently adjusted according to personal preferences and current driving status. However, existing in-vehicle voice control systems can often only execute preset command sets and lack the ability to dynamically adjust functions according to the user's current driving status. They cannot meet the personalized needs of drivers and it is difficult to accurately recommend the audio applications that users are most likely to need at the moment, resulting in a less smooth and efficient user experience.

[0003] In the current relevant technologies, the voice control of in-vehicle smart speakers has the technical problem of being unable to adjust the control strategy according to the real-time driving status and the specific needs of the user, resulting in a poor user experience. Summary of the invention

[0004] The present application provides a voice control optimization method for in-vehicle smart speakers, which uses intelligent monitoring equipment to monitor the driving status of the target user in real time, generates target audio application predictions through preliminary analysis based on the driving status information and combined with the user's historical behavior data, recommends the most likely audio application categories to the user, and optimizes voice control for the selected audio application categories, thereby achieving the technical effect of optimizing the voice control strategy, improving the control accuracy and response speed, and thus improving the user experience.

[0005] The present application provides a voice control optimization method for an in-vehicle intelligent audio system, comprising: obtaining target driving state information, wherein the target driving state information refers to driving state information obtained by dynamically monitoring a target user through an intelligent monitoring device; obtaining a first activation instruction, and pre-analyzing the target driving state information based on the first activation instruction to generate a target audio application prediction, wherein the target audio application prediction includes multiple predicted application categories; obtaining a first selection instruction, and selecting and starting a first target application category from the multiple predicted application categories based on the first selection instruction; obtaining a first control strategy for the first target application category, and performing voice control optimization on the first target application category based on the first control strategy.

[0006] In a possible implementation, the following processing is performed: reading a predetermined monitoring dimension; the intelligent monitoring device dynamically monitors the target user's driving based on the predetermined monitoring dimension to obtain the target driving status information; wherein the predetermined monitoring dimension includes at least geographic location, driving speed, climbing grade, weather conditions and driving time.

[0007] In a possible implementation, the following processing is performed: obtaining a target historical audio application log, the target historical audio application log refers to the record information of the target user's historical application of the vehicle audio; extracting the first historical application record of the target historical audio application log, the first historical application record including the first historical driving state information and the first historical audio control path; comparing and analyzing the target driving state information with the first historical driving state information to obtain a first state similarity; when the first state similarity reaches a similarity limit, issuing a prediction instruction; performing union processing on the first historical audio control path based on the prediction instruction to obtain a first historical path set; analyzing the first historical path set to obtain the target audio application prediction.

[0008] In a possible implementation, the following processing is performed: reading a predetermined state quantization rule; quantizing the target driving state information and the first historical driving state information in turn according to the predetermined state quantization rule to obtain a target state vector and a first historical state vector, respectively; and comparing the target state vector and the first historical state vector to obtain the first state similarity.

[0009] In a possible implementation, the following processing is performed: obtaining a first nearest neighbor control strategy for the first target application category; traversing and matching the target driving state information in a predetermined control database based on the first control strategy to obtain a target predetermined control strategy; comparing the target predetermined control strategy with the first nearest neighbor control strategy to obtain a first real-time control fitness; when the first real-time control fitness does not reach the fitness limit, issuing an optimization query instruction; obtaining a first optimization decision for the target user based on the optimization query instruction; and performing voice control optimization on the first target application category according to the first optimization decision.

[0010] In a possible implementation, the following processing is performed: using the target driving state information as a control optimization constraint; using a control quality evaluation function as a control evaluation scheme; obtaining a predetermined control dimension, and extracting a first control dimension of the predetermined control dimension, wherein the first control dimension corresponds to a first control parameter threshold; establishing a control optimization space based on the first control parameter threshold; and under the control optimization constraint, optimizing in the control optimization space based on the control evaluation scheme to obtain the target predetermined control strategy.

[0011] In a possible implementation, the following processing is performed: the expression of the control quality evaluation function is as follows: ;in, Characterizing control strategies The control adaptability, Characterization Control strategy evaluation parameters, represents the number of control evaluation dimensions, and , Characterizing the Control strategy evaluation parameters With the control fitness The correlation of and Respectively represent the first weight feedback coefficient and the second weight feedback coefficient.

[0012] In a possible implementation, the following processing is performed: the predetermined control dimension includes at least a sound effect dimension and a volume dimension.

[0013] In a possible implementation, the following processing is performed: obtaining a control evaluation dimension, wherein the control evaluation dimension includes a stability dimension and a noise suppression dimension.

[0014] A voice control optimization method for in-vehicle smart audio proposed in this application first obtains target driving state information, which refers to driving state information obtained by dynamically monitoring the target user through intelligent monitoring equipment, then obtains a first activation instruction, and pre-analyzes the target driving state information based on the first activation instruction to generate a target audio application prediction, which includes multiple predicted application categories, then obtains a first selected instruction, and selects and starts a first target application category from multiple predicted application categories based on the first selected instruction, finally obtains a first control strategy for the first target application category, and optimizes voice control for the first target application category based on the first control strategy, thereby achieving the technical effect of improving control accuracy and response speed by optimizing the voice control strategy, thereby improving user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the method according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0016] Figure 1A flow chart of a method for optimizing voice control of an in-vehicle smart speaker provided in an embodiment of the present application.

[0017] Figure 2 A flow chart of voice control optimization for a first target application category according to a first optimization decision in a voice control optimization method for an in-vehicle smart speaker provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0019] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0020] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0021] The present application embodiment provides a voice control optimization method for an in-vehicle smart speaker, such as Figure 1 As shown, the method includes:

[0022] Step S100, obtaining target driving state information, which refers to driving state information obtained by dynamically monitoring the target user through intelligent monitoring equipment. Specifically, when the vehicle is started or the vehicle-mounted intelligent audio system is activated, the intelligent monitoring equipment connected thereto (various sensors, cameras and other equipment that can automatically sense and record the driving environment and the state of the target user) is also started, and these intelligent monitoring equipment enters the working state to collect real-time data of the driving environment and the target user (the user of the vehicle-mounted intelligent audio system and the monitored object, such as the driver). For example, the camera is used to capture visual information such as the driver's facial expression, sight direction, head posture, etc.; the acceleration sensor, gyroscope, steering wheel angle sensor, etc. are used to monitor the vehicle's speed, acceleration, steering angle, and possible driving behaviors such as sudden braking and sudden acceleration; the heart rate monitor (if integrated in the seat or steering wheel) is used to evaluate the driver's physiological state. The collected raw data is transmitted to the processor of the vehicle-mounted system or the cloud server for processing. The data processing includes steps such as noise removal, feature extraction, and pattern recognition to extract key driving state information. For example, by analyzing the driver's facial expression and line of sight, it is possible to determine whether the driver is distracted; by using acceleration sensor data, the stability of driving behavior is evaluated. After processing, target driving state information is generated, which is a data set reflecting the current driving state of the target user, including but not limited to attention level, fatigue level, and driving behavior stability.

[0023] In a possible implementation, the target driving state information is obtained, and step S100 further includes step S110, reading a predetermined monitoring dimension, wherein the predetermined monitoring dimension includes at least a geographical location, a driving speed, a climbing grade, a weather condition, and a driving time. Specifically, the intelligent monitoring device is started and connected to a corresponding data source (such as a GPS locator, a speed sensor, a tilt sensor, a weather station data interface, a clock, etc.). In the system settings or configuration files, the dimensions to be monitored are predefined and stored, including the geographical location (the latitude and longitude of the vehicle at the moment, used to determine whether the vehicle is driving in a specific area), the driving speed (used to evaluate the safety of driving behavior), the climbing grade (the slope change during vehicle driving, affecting the vehicle's handling and safety), the weather conditions (the impact of the external environment on driving, such as rainy and snowy weather will increase the slipperiness of the road surface and reduce visibility), and the driving time, etc. At the beginning of the monitoring task, the intelligent monitoring device reads these predetermined monitoring dimension parameters from the configuration file. Step S120, the intelligent monitoring device performs dynamic driving monitoring on the target user based on the predetermined monitoring dimension to obtain the target driving state information. Specifically, the intelligent monitoring device collects the current driving data of the target user from various sensors and data sources in real time, including location, speed, slope, weather conditions, time, etc. The collected data is compared and analyzed with the predetermined monitoring dimensions to determine whether the current driving status meets the safety standards or preset conditions. For example, check whether the speed exceeds the speed limit, whether the climbing gradient is too large, whether the weather conditions are bad, etc. Based on the data analysis results, the driving status of the target user is evaluated and the target driving status information is generated. This implementation method can more accurately understand the current driving environment and status of the target user by monitoring specific driving dimensions, thereby achieving a better technical effect of predicting and meeting the target user's needs for audio applications during driving.

[0024] Step S200, obtaining a first activation instruction, and pre-analyzing the target driving state information based on the first activation instruction, generating a target audio application prediction, wherein the target audio application prediction includes a plurality of predicted application categories. Specifically, the in-vehicle intelligent audio system receives the first activation instruction issued by the target user through a voice recognition module or other input device (such as a touch screen, a physical button, etc.), wherein the first activation instruction may be a simple voice command or a specific button operation, which is used to start the voice control function of the audio system or enter the voice interaction mode. After receiving the first activation instruction, the system analyzes context information such as the current driving environment, the vehicle state, and the historical usage habits of the target user to form a comprehensive user portrait and driving scene description. Using prediction models such as machine learning or deep learning, the information is analyzed and inferred to predict the audio applications that the target user may need in the current driving state. According to the prediction results, multiple possible audio application categories are determined, such as playing music, navigation broadcasting, audio book reading, etc. At the same time, the predicted application categories can be prioritized according to factors such as the target user's preferences, driving state, and the importance and urgency of the application. The prediction results can be presented to the target user in a visual manner (such as screen display) or in the form of voice prompts for the target user to select or confirm.

[0025] In a possible implementation, a first activation instruction is obtained, and based on the first activation instruction, the target driving state information is pre-analyzed to generate a target audio application prediction, and step S200 further includes step S210, obtaining a target historical audio application log, wherein the target historical audio application log refers to the record information of the target user's historical application of the vehicle audio. Specifically, the target user's historical audio application records are retrieved from the database, and these records are stored in the cloud or a local storage device. Among them, the target historical audio application log refers to all records of the target user using the vehicle audio in the past period of time, including but not limited to driving time, location, weather conditions, selected music type, radio channel, volume setting and other information. Step S220, extracting the first historical application record of the target historical audio application log, wherein the first historical application record includes the first historical driving state information and the first historical audio control path. Specifically, any one historical application record is selected from the target historical audio application log as the first historical application record, and the first historical driving state information and the first historical audio control path are extracted from the first historical application record. The first historical driving state information includes geographic location, weather conditions, driving time, etc., and the first historical audio control path refers to the sequence of audio control operations performed by the target user in the first historical driving state. Step S230, compare and analyze the target driving state information with the first historical driving state information to obtain the first state similarity. Specifically, compare the current target driving state information with the first historical driving state information to find similarities, calculate the similarity between the two by using algorithms such as cosine similarity and Euclidean distance, and obtain the first state similarity. Step S240, when the first state similarity reaches the similarity limit, issue a prediction instruction. Specifically, compare the calculated first state similarity with the preset similarity limit. If the similarity reaches or exceeds the limit, it is considered that the current driving state is similar to the state in the first historical application record, so a prediction instruction is issued to prepare for the prediction of the audio application. Step S250, perform a union process on the first historical audio control path based on the prediction instruction to obtain a first historical path set. Specifically, the first historical audio control path is subjected to a union process, that is, all control operations that have appeared in history are merged to form a first historical path set containing multiple possibilities. Step S260: Analyze the first historical path set to obtain the target audio application prediction. Specifically, analyze the control paths in the first historical path set to find out the audio application types that the target user is most likely to use under different similar driving conditions, generate the target audio application prediction based on the analysis result, and display it to the target user in a user interface (such as a display), such as listing possible application project numbers for the target user to select.This implementation provides personalized audio application predictions by analyzing users' historical usage records and learning their preferences and habits. By making predictions and preparations based on historical data, it can respond to user needs more quickly, achieving the technical effect of improving response speed and accuracy of voice control.

[0026] In a possible implementation, the target driving state information and the first historical driving state information are compared and analyzed to obtain the first state similarity, and step S230 further includes step S231, reading a predetermined state quantization rule. Specifically, based on the knowledge of domain experts, historical data analysis or experimental verification, a set of state quantization rules is predefined, and the state quantization rule is a set of rules for converting driving state information (such as geographic location, weather conditions, driving time, etc.) into a quantifiable and comparable format (such as a vector). The defined state quantization rules are stored in a database or configuration file accessible to the system, and these rules are read from the storage location when the state similarity needs to be calculated. Step S232, according to the predetermined state quantization rule, the target driving state information and the first historical driving state information are sequentially quantized to obtain a target state vector and a first historical state vector respectively. Specifically, according to the read predetermined state quantization rule, each element in the target driving state information (such as the current geographic location, weather conditions, etc.) is converted into a numerical value or vector form to obtain a target state vector. Similarly, the first historical driving state information is quantized to obtain a first historical state vector. Step S233, compare the target state vector and the first historical state vector to obtain the first state similarity. Specifically, use a vector similarity calculation method (such as cosine similarity, Euclidean distance, etc.) to compare the target state vector and the first historical state vector, and calculate the similarity value between the two vectors according to the similarity calculation method, that is, the first state similarity. This implementation method standardizes the quantization process by introducing a state quantization rule, thereby achieving the technical effect of improving the accuracy and efficiency of driving state similarity calculation.

[0027] Step S300, obtain the first selected instruction, and select and start the first target application category in the multiple predicted application categories based on the first selected instruction. Specifically, the target user sends a first selected instruction to the in-vehicle intelligent audio system through voice or other input methods (such as touch screen, physical buttons, etc.), and the first selected instruction is an instruction issued by the target user to select the required application from the predicted multiple audio application categories. The system recognizes the input of the target user through a voice recognition module or other input processing mechanism, extracts the audio application category selected by the target user, that is, the first target application category, allocates necessary computing resources, storage resources and network resources to the first target application category to ensure that it can be started and run smoothly, and then starts the first target application category, and performs corresponding operations according to the further instructions of the target user or the preset configuration. For example, if the target user selects "play music", the system automatically opens the music playlist and starts playing the songs that the target user likes.

[0028] Step S400, obtain the first control strategy of the first target application category, and optimize the voice control of the first target application category based on the first control strategy. Specifically, define or obtain specific control strategies for the first target application category, which include but are not limited to recognition rules of voice instructions, mapping relationships of in-application functions, priority sorting, exception handling, etc. The control strategy can come from multiple aspects, such as preset strategies during system design, user-defined settings, and strategies dynamically generated after the system learns user habits. The system obtains the first control strategy associated with the first target application category by querying the internal database, configuration files, or calling the corresponding API interface, and parses the obtained first control strategy to understand the specific rules, conditions, and execution actions contained therein. According to the current system status, user permissions, context environment, and other factors, the first control strategy is adapted and adjusted to ensure that it can be accurately and effectively applied to the first target application category. Based on the parsed and adapted first control strategy, the voice control function of the first target application category is optimized and implemented, including improving the accuracy of voice recognition, optimizing the response speed of voice instructions, and enhancing the linkage of in-application functions. Perform functional testing on the optimized first target application category to ensure that the voice control function works as expected and no new problems or vulnerabilities are introduced. Deploy the optimized first target application category and its control strategy to the actual operating environment to ensure that the target users can enjoy the optimized voice control function. Based on information such as user feedback, system logs, and data analysis results, the first control strategy is continuously optimized and updated to adapt to the ever-changing user needs and technical environment. The embodiment of the present application uses intelligent monitoring equipment to monitor the driving status of the target user in real time, based on the driving status information, combined with the user's historical behavior data, to generate a target audio application prediction through pre-analysis, recommend the most likely audio application category to the user, and perform voice control optimization for the selected audio application category. Technical means such as optimizing the voice control strategy, improving the control accuracy and response speed, and thus improving the user experience.

[0029] like Figure 2As shown, in a possible implementation, the first control strategy of the first target application category is obtained, and the voice control optimization of the first target application category is performed based on the first control strategy. Step S400 further includes step S410, obtaining the first nearest control strategy of the first target application category. Specifically, by accessing the user's history or configuration file, the control strategy of the last (or most recent) use of the first target application category, that is, the first nearest control strategy, is obtained, including volume, playlist, sound effect settings, etc. Step S420, based on the first control strategy, the target driving state information is traversed and matched in the predetermined control database to obtain the target predetermined control strategy. Specifically, according to the current first control strategy (which can be a default strategy set by the user or a strategy generated according to a certain rule), a target predetermined control strategy matching it is searched in the predefined control strategy database, and in the traversal process, the target driving state information (such as vehicle speed, road conditions, time, etc.) is combined to select the control strategy that best suits the current driving state. Among them, the predetermined control database is a database that stores control strategies recommended under multiple different driving states; the target predetermined control strategy is the control strategy that best matches the current first control strategy and the target driving state information in the predetermined control database. Step S430, compare the target predetermined control strategy with the first nearest control strategy to obtain the first real-time control fitness. Specifically, compare the parameters (such as volume, sound effects, playlist, etc.) of the target predetermined control strategy and the first nearest control strategy, calculate the similarity or difference between the two, and evaluate the first real-time control fitness, that is, the matching degree between the current control strategy (target predetermined control strategy) and the user's historical preference (first nearest control strategy) based on the above comparison results. Step S440, when the first real-time control fitness does not reach the fitness limit, issue an optimization query instruction. Specifically, if the first real-time control fitness does not reach the preset fitness limit, it is considered that the current control strategy may not meet the actual needs or expectations of the target user. At this time, an optimization query instruction is issued to the target user through voice or interface prompts to ask the target user whether he wants to perform further optimization. Step S450, obtain the first optimization decision of the target user based on the optimization query instruction. Specifically, the target user gives an optimization decision through voice or interface operation according to the inquiry of the system, that is, whether to agree to automatic optimization. Step S460: Optimize the voice control of the first target application category according to the first optimization decision. Specifically, if the target user agrees to perform automatic optimization, the system automatically adjusts the control strategy of the first target application category according to the current driving state, user preference and optimization algorithm to achieve a more optimized voice control experience.This implementation method provides target users with a personalized and intelligent voice control experience by combining the user's historical preferences, current driving status, and the system's automatic optimization mechanism, achieving the technical effect of meeting the user's personalized needs and realizing an intelligent automatic optimization process.

[0030] In a possible implementation, step S420 further includes step S421, using the target driving state information as a control optimization constraint. Specifically, the target driving state information is used as a constraint for control optimization, that is, the target driving state information is a condition or restriction that must be met in the process of optimizing the control strategy to ensure that the control strategy found meets the current driving environment and user needs. Step S422, using the control quality evaluation function as a control evaluation scheme. Specifically, one or more control quality evaluation functions are defined to quantitatively evaluate the quality or effect of different control strategies. The control quality evaluation function can formulate evaluation criteria based on multiple factors such as user satisfaction, driving safety, and system efficiency. Step S423, obtaining a predetermined control dimension, and extracting a first control dimension of the predetermined control dimension, wherein the first control dimension corresponds to a first control parameter threshold. Specifically, the predetermined control dimension is determined, and the predetermined control dimension represents a parameter that can be adjusted or optimized in the control strategy, such as volume, sound effect, playback mode, etc. The predetermined control dimension jointly determines the specific performance of the control strategy. Extract any one control dimension from the predetermined control dimensions as the first control dimension to be optimized, and extract the first control parameter threshold corresponding to the first control dimension, that is, the value range or restriction condition of the parameter in the first control dimension. Step S424, establish a control optimization space based on the first control parameter threshold. Specifically, according to the parameter threshold of the first control dimension, define a control optimization space, that is, search for the optimal control strategy in the control optimization space. The control optimization space can be a multidimensional continuous or discrete space, which can be divided into smaller subspaces or grids as needed to facilitate a more refined search. Step S425, under the control optimization constraint, optimize in the control optimization space based on the control evaluation scheme to obtain the target predetermined control strategy. Specifically, a search algorithm such as a genetic algorithm and a particle swarm optimization algorithm is used to search in the control optimization space. During the search process, each candidate control strategy is evaluated according to the control quality evaluation function, and the search direction and strategy are continuously adjusted based on the evaluation results to find the optimal solution. The search process continues until the preset convergence conditions are met (such as reaching the maximum number of iterations, the value of the control quality evaluation function no longer changes significantly, etc.). At this time, the control strategy obtained is the target predetermined control strategy. This implementation method only optimizes a single control dimension each time, which reduces the computational complexity and search difficulty, makes the search process more centralized and efficient, and achieves the technical effect of improving search efficiency.

[0031] In a possible implementation, step S422 further includes step S4221, and the expression of the control quality evaluation function is as follows:

[0032] ;

[0033] in, Characterizing control strategies The control adaptability, Characterization Control strategy evaluation parameters, represents the number of control evaluation dimensions, and , Characterizing the Control strategy evaluation parameters With the control fitness The correlation of and Respectively characterize the first weight feedback coefficient and the second weight feedback coefficient. Specifically, the control fitness is a comprehensive indicator used to measure the overall performance of the control strategy, reflecting the comprehensive performance of the control strategy in multiple dimensions. The control strategy evaluation parameter is an indicator used to measure the performance of the control strategy in a specific aspect. Different control strategy evaluation parameters represent the performance of the control strategy in different aspects. This implementation method comprehensively evaluates the performance of the control strategy by integrating multiple control strategy evaluation parameters, avoids the one-sidedness of a single-dimensional evaluation, and achieves the technical effect of improving the comprehensiveness and scientificity of the control strategy performance evaluation.

[0034] In a possible implementation, step S423 further includes step S4231, and the predetermined control dimension includes at least a sound effect dimension and a volume dimension. Specifically, the sound effect dimension refers to the specific aspects of the sound effect in the optimization of the control strategy, which can cover multiple sub-dimensions, such as the type of sound effect (such as ambient sound, music, prompt sound, etc.), the clarity of the sound effect, the layering of the sound effect, etc. The volume dimension refers to the specific aspects of the sound size or intensity in the optimization of the control strategy. During the optimization process, different control strategies are evaluated according to the specific parameters and evaluation criteria of the sound effect dimension and the volume dimension. Sound effect and volume are key factors affecting user experience. This implementation method includes at least a sound effect dimension and a volume dimension in the predetermined control dimensions. It is a reasonable choice based on user experience and actual application needs. By optimizing these two dimensions, the auditory enjoyment and overall satisfaction of the target users can be significantly improved, achieving the technical effect of improving user experience.

[0035] In a possible implementation, step S4221 further includes step S42211, obtaining a control evaluation dimension, wherein the control evaluation dimension includes a stability dimension and a noise suppression dimension. Specifically, the stability dimension refers to the ability of the system to maintain its output or state within a certain range when subjected to external interference or internal changes; the noise suppression dimension refers to the ability of the system to suppress or reduce noise interference and ensure that the output signal is clear and accurate. For each control evaluation dimension, a specific control strategy evaluation parameter is determined, for example, the stability dimension may include response time, fluctuation range, recovery time, etc.; the noise suppression dimension may include signal-to-noise ratio, noise suppression ratio, etc. This implementation focuses the control evaluation on the stability dimension and the noise suppression dimension, making the evaluation process more targeted. These two dimensions are key aspects in the control strategy performance evaluation and are directly related to the stability and reliability of the system. The stability and anti-interference ability of the system directly affect the user experience. By quantitatively evaluating the performance of the control strategy in the stability dimension and the noise suppression dimension, system failures and noise interference can be significantly reduced, and user satisfaction and trust can be improved. This targeted optimization avoids wasting computing resources in irrelevant or secondary directions, and achieves the technical effect of improving the optimization efficiency.

[0036] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A voice control optimization method for vehicle-mounted intelligent audio, characterized in that: include: Acquiring target driving state information, wherein the target driving state information refers to driving state information obtained by performing dynamic driving monitoring of a target user through an intelligent monitoring device; Acquire a first activation instruction, and pre-analyze the target driving state information based on the first activation instruction to generate a target audio application prediction, wherein the target audio application prediction includes a plurality of predicted application categories; Obtaining a first selected instruction, and selecting and starting a first target application category among the plurality of predicted application categories based on the first selected instruction; Acquire a first control strategy for the first target application category, and perform voice control optimization on the first target application category based on the first control strategy; Obtaining a first nearest neighbor control strategy for the first target application category; Based on the first control strategy, the target driving state information is traversed and matched in a predetermined control database to obtain a target predetermined control strategy; Comparing the target predetermined control strategy with the first nearest neighbor control strategy to obtain a first real-time control fitness; When the first real-time control fitness does not reach the fitness limit, issuing an optimization query instruction; Obtaining a first optimization decision for the target user based on the optimization query instruction; Performing voice control optimization on the first target application category according to the first optimization decision; Using the target driving state information as a control optimization constraint; Using the control quality assessment function as a control evaluation scheme; Acquire a predetermined control dimension, and extract a first control dimension of the predetermined control dimension, where the first control dimension corresponds to a first control parameter threshold; Establishing a control optimization space based on the first control parameter threshold; Under the control optimization constraint, optimizing in the control optimization space based on the control evaluation scheme to obtain the target predetermined control strategy; The expression of the control quality evaluation function is as follows: ； in, Characterizing control strategies The control adaptability, Characterization Control strategy evaluation parameters, represents the number of control evaluation dimensions, and , Characterizing the Control strategy evaluation parameters With the control fitness The correlation of and Respectively represent the first weight feedback coefficient and the second weight feedback coefficient.

2. The voice control optimization method for vehicle-mounted intelligent audio according to claim 1, characterized in that: include: Read the predetermined monitoring dimensions; The intelligent monitoring device performs dynamic driving monitoring on the target user based on the predetermined monitoring dimension to obtain the target driving state information; The predetermined monitoring dimensions include at least geographic location, driving speed, climbing grade, weather conditions and driving time.

3. The voice control optimization method for vehicle-mounted intelligent audio according to claim 1, characterized in that: include: Obtaining a target historical audio application log, wherein the target historical audio application log refers to the record information of the target user's historical application of the vehicle audio; Extracting a first historical application record of the target historical audio application log, wherein the first historical application record includes first historical driving state information and a first historical audio control path; Comparing and analyzing the target driving state information and the first historical driving state information to obtain a first state similarity; When the similarity of the first state reaches a similarity limit, issuing a prediction instruction; Performing a union process on the first historical sound control paths based on the prediction instruction to obtain a first historical path set; The first historical path set is analyzed to obtain the target audio application prediction.

4. The voice control optimization method for vehicle-mounted intelligent audio according to claim 3 is characterized in that: include: Reading predetermined state quantization rules; Quantizing the target driving state information and the first historical driving state information in sequence according to the predetermined state quantization rule to obtain a target state vector and a first historical state vector respectively; The target state vector and the first historical state vector are compared to obtain the first state similarity.

5. The voice control optimization method for vehicle-mounted intelligent audio according to claim 1, characterized in that: The predetermined control dimensions include at least a sound effect dimension and a volume dimension.

6. The voice control optimization method for vehicle-mounted intelligent audio according to claim 1, characterized in that: A control evaluation dimension is obtained, where the control evaluation dimension includes a stability dimension and a noise suppression dimension.

Images