Automobile driving behavior analysis method and system integrated with voice recognition

Abstract

The invention relates to the technical field, in particular to an automobile driving behavior analysis method and system integrated with voice recognition. The method comprises the following steps: acquiring voice input content, analyzing the voice input content, and identifying a vehicle driving problem; determining a driving link according to a vehicle driving problem, and calling real-time driving data according to the driving link; and analyzing the real-time driving data, determining problem relevance, and determining a solution according to the problem relevance. By extracting the voice input content, the problem described by the user is quickly positioned, and a clear direction is provided for diagnosis and analysis. The voice input content is associated with a specific driving link, the actual running state of the vehicle is provided for the occurrence of the vehicle driving problem by calling the real-time driving data, and the root cause of the vehicle driving problem is identified; and the real-time driving data is combined with factors such as environmental factors and vehicle part performance, so that the accuracy and comprehensiveness of diagnosis are improved.

Description

Technical Field

[0001] The present application relates to the technical field, and in particular to a method and system for analyzing automobile driving behavior with integrated speech recognition. Background Art

[0002] With the continuous development of modern automobile technology, driving behavior analysis has gradually become an important tool for improving driving safety, comfort and vehicle maintenance management. Driving behavior analysis can not only help drivers find potential vehicle problems, but also provide real-time feedback and correction suggestions when faults or abnormal driving behaviors occur.

[0003] Traditional driving behavior analysis methods usually rely on real-time data from multiple sensors in the vehicle, such as accelerator pedal opening, braking frequency, vehicle speed, acceleration, etc. Therefore, how to efficiently and accurately identify vehicle problems and provide solutions in a timely manner is still a major challenge facing current vehicle diagnostic systems. Summary of the invention

[0004] The present application provides a vehicle driving behavior analysis method and system integrated with speech recognition to solve the above-mentioned problems.

[0005] In a first aspect, the present application provides a method for analyzing automobile driving behavior with integrated speech recognition, the method comprising: Acquire voice input content, analyze the voice input content, and identify vehicle driving problems; Determine the driving link according to the vehicle driving problem, and retrieve real-time driving data according to the driving link; The real-time driving data is analyzed to determine problem relevance, and a solution is determined based on the problem relevance.

[0006] Through this solution, voice input provides users with a convenient way to report vehicle problems without manual operation, thereby reducing distractions during driving and improving driving safety. Convert voice input content into text content to improve the efficiency and accuracy of information processing. Through content extraction, quickly locate the problems described by the user, and provide a clear direction for diagnosis and analysis. Associating voice input content with specific driving links helps to more accurately analyze and solve problems in the driving link. By retrieving real-time driving data, the actual operating status of the vehicle is provided when vehicle driving problems occur, providing an important data basis for diagnosis. By analyzing real-time driving data, abnormal patterns in vehicle operation are discovered, which helps to identify the root cause of vehicle driving problems. Combine real-time driving data with environmental factors, vehicle component performance and other factors to improve the accuracy and comprehensiveness of diagnosis.

[0007] Optionally, the acquiring of voice input content includes: In response to a voice dialogue command triggered by a user, a voice signal collected by an onboard microphone array is acquired; Performing noise reduction and endpoint detection on the speech signal to extract valid speech segments; The valid voice segment is converted into text content to obtain the voice input content.

[0008] Through this solution, a natural interface for users to interact with vehicle equipment is provided, allowing users to trigger device functions through voice commands without distracting their attention, thereby improving driving safety. The on-board microphone array can capture sounds inside and outside the car, providing a sound source for the voice recognition device, so that users can issue voice commands at any location. Noise reduction processing can significantly reduce the interference of environmental noise and noise inside the car, and improve the accuracy of voice recognition equipment. Endpoint detection can accurately identify the starting and ending points of speech, extract valid voice segments, and avoid processing invalid information. By extracting valid voice segments, we focus on processing sound data related to voice commands, reduce the amount of calculation and processing time, and improve response speed. The voice recognition device converts voice signals into text content, which helps to analyze and respond to specific user needs. The final voice input content contains the vehicle driving problem described by the user, provides information for further analysis and diagnosis, and helps to provide more personalized solutions.

[0009] Optionally, analyzing the voice input content and identifying vehicle driving problems includes: Extracting a keyword set from the voice input content, the keyword set including abnormal description words, component name words, and driving scene words; Analyze the component name words and the driving scene words to determine the vehicle driving link; Analyze the abnormal description words to determine the link problem; The link problem and the vehicle driving link are regarded as the vehicle driving problem.

[0010] Through this solution, a keyword set is extracted from the user's voice input through natural language processing technology, which helps to identify the user's problem. Extracting a keyword set helps to quickly locate vehicle driving problems. Abnormal descriptors such as abnormal noise and shaking help to identify the nature of the vehicle driving problem; component name words such as engine and brake help to determine the vehicle components involved in the vehicle driving problem; driving scene words such as acceleration and turning help to determine the driving link where the problem occurs. By analyzing the keyword set, determining in which driving link the problem described by the user occurs, such as acceleration, deceleration, turning, etc., helps to narrow the scope of problem diagnosis and improve diagnostic efficiency. By analyzing the abnormal descriptors, determining the abnormal type described by the user, such as mechanical failure, performance degradation, etc., helps to provide more accurate solutions. Combining link problems with vehicle driving links to form a complete vehicle driving problem helps to provide a basis for real-time driving data retrieval, problem correlation analysis, and solution determination.

[0011] Optionally, the retrieving real-time driving data according to the driving link includes: Analyze the voice input content to determine a trigger timestamp; Determine the time window when the problem occurred based on the voice trigger timestamp; A multi-dimensional data stream within the time window is extracted from vehicle sensors and controllers as the real-time driving data.

[0012] This solution analyzes the user's voice input content and identifies the user's intentions and needs, thereby providing more personalized services. By determining the trigger timestamp of the user's voice command, the time window when the problem occurred is accurately recorded, which helps to diagnose and analyze vehicle driving problems. Determining the time window when the problem occurred based on the trigger timestamp helps to obtain sufficient multi-dimensional data for analysis, thereby improving the accuracy of diagnosis. Extracting multi-dimensional data streams within the time window from vehicle sensors and controllers provides an important data basis for problem diagnosis.

[0013] Optionally, the real-time driving data includes vehicle body data, and the step of analyzing the real-time driving data to determine the problem correlation includes: Acquire driving environment data during driving; Analyzing the vehicle driving problem based on the driving environment data to determine the abnormality type; Analyze the real-time driving data to determine frequency domain characteristics; Determining, based on the frequency domain characteristics, a vibration resonance peak position associated with the abnormal type; Determining a candidate abnormal component according to the vibration resonance peak position; The vehicle body data is analyzed to determine the problem correlation between the candidate abnormal component and the vehicle driving problem.

[0014] Through this solution, the driving environment data of the vehicle during driving is collected, which helps to identify the external conditions of vehicle operation and analyze the environmental impact of vehicle problems. By analyzing the driving environment data, the potential correlation between environmental factors and vehicle problems can be identified. For example, slippery road conditions lead to longer braking distances, which affects the braking performance of the vehicle. According to the determined abnormality type, preliminary diagnostic results are provided to provide users with corresponding solutions. Extracting the vehicle body's own data from vehicle sensors and controllers, such as speed, acceleration, engine speed, accelerator pedal opening, brake status, etc., helps reflect the actual operating status of the vehicle and helps analyze vehicle problems. Frequency domain features help diagnose vehicle driving problems related to vibration. The resonance peak position helps to determine the abnormal component that has failed. The abnormal component is a candidate abnormal component for further diagnosis and inspection. Determine the correlation between the candidate abnormal component and the vehicle driving problem described by the user, so as to provide users with accurate diagnostic results and solutions. Determining the problem correlation helps users to identify the vehicle condition in a timely manner and take corresponding maintenance measures.

[0015] Optionally, determining a solution according to the problem relevance includes: Analyze the driving environment data to determine the road slope and air humidity during driving; Analyze the road slope and the air humidity to determine the impact of environmental interference; Determining the coordinated abnormality degree of the candidate abnormal components during driving according to the problem correlation; A solution is determined based on the environmental interference impact and the coordination abnormality degree.

[0016] Through this solution, by analyzing driving environment data, the specific environmental conditions encountered by the vehicle during driving, such as road slope and air humidity, are identified. Driving environment data helps diagnose the environmental interference effects on vehicle problems. By analyzing road slope and air humidity, the environmental interference effects of road slope and air humidity on vehicle performance are determined, such as slope affecting the acceleration and braking performance of the vehicle, and humidity affecting the tire grip and braking of the vehicle. By analyzing the correlation of problems, the possibility of candidate abnormal components being abnormal at the same time under the influence of environmental interference is determined, which helps to narrow the scope of the problem and improve the accuracy of diagnosis. Combined with the environmental interference effects and the coordinated abnormality of candidate abnormal components, a solution is determined to provide users with corresponding maintenance suggestions or operation guidelines.

[0017] Optionally, the driving environment data includes a road bump index, and the analyzing the vehicle body data to determine the problem correlation between the candidate abnormal component and the vehicle driving problem includes: Analyzing the candidate abnormal component to determine the mechanical properties of the candidate abnormal component; According to the mechanical characteristics, a preset component feature database is retrieved to obtain a standard vibration frequency range of the candidate abnormal component; Analyzing the vibration resonance peak position in the frequency domain characteristics, and calculating the offset between the vibration resonance peak position and the standard vibration frequency range; According to the offset and the road bump index, the component abnormality confidence of each candidate abnormal component is determined and calculated using the following formula: ; in, Indicates the abnormal confidence level of the component; represents the offset; Indicates the influence weight of the offset on the abnormality of the component; Indicates the influence weight of the road bump index on the abnormality of the component; represents the road bumpiness index; Determining the current wear condition of each candidate abnormal component according to the preset component feature database; determining an abnormal threshold value according to the current wear condition; Comparing the component abnormality confidence with the abnormality threshold of the corresponding candidate abnormal component; If the component abnormality confidence exceeds the abnormality threshold of the corresponding candidate abnormal component, it is determined that the candidate abnormal component is associated with the vehicle driving problem.

[0018] Through this solution, the performance parameters and expected behaviors of the candidate abnormal parts are identified by analyzing the mechanical characteristics of the candidate abnormal parts. By calling the preset component feature database, the standard vibration frequency range of the candidate abnormal parts is obtained to provide a reference for frequency domain analysis. Through frequency domain analysis, the vibration resonance peak position is identified, and the offset from the standard vibration frequency range is calculated, which helps to determine whether the candidate abnormal parts are abnormal. The component abnormality confidence is calculated by formula to quantify the possibility of abnormality of each candidate component, thereby providing a more accurate basis for diagnosis. The current wear condition of the candidate abnormal parts is obtained through the preset component feature database, which helps to evaluate the aging and loss of the candidate abnormal parts. According to the current wear condition of the candidate abnormal parts, the abnormal threshold is determined, which helps to determine whether the performance of the candidate abnormal parts exceeds the normal range. By comparing the component abnormality confidence and the abnormal threshold, it is determined which candidate abnormal parts are abnormal, and diagnostic suggestions are provided to the user. If the component abnormality confidence exceeds the abnormal threshold, it is confirmed that the candidate abnormal part is related to the vehicle driving problem and is the cause of the problem, thereby providing maintenance suggestions or operation guidelines to the user.

[0019] Optionally, determining the coordinated abnormality degree of the candidate abnormal component during driving according to the problem correlation includes: Acquire sensor data related to the candidate abnormal component within the time window; analyzing the sensor data to determine phase synchronization of the candidate abnormal component; Calculating a mutual correlation coefficient between each sensor signal and each candidate abnormal component according to the phase synchronization; According to the mutual correlation coefficient, the coordinated abnormality degree of the candidate abnormal components during the driving process is determined.

[0020] Through this solution, sensor data related to candidate abnormal components in the time window is collected, which helps to analyze the basis of the performance of candidate abnormal components. By preprocessing the sensor data, the real-time working status and performance indicators of candidate abnormal components are identified. Analyzing the phase relationship between different sensor signals helps to determine whether the candidate abnormal components are working in the correct timing, thereby identifying potential collaborative abnormalities. By calculating the mutual correlation coefficient, the similarity and timing relationship between different sensor signals are quantified, which helps to evaluate the collaborative working between candidate abnormal components. Based on the mutual correlation coefficient, the collaborative abnormality of each candidate abnormal component during driving is determined, which helps to identify which candidate abnormal components jointly cause vehicle problems.

[0021] Optionally, the method further includes: After the solution is output, the vehicle status is monitored in real time, and according to the monitoring result, it is determined whether the user has solved the vehicle driving problem; If the problem is not solved, the vehicle driving problem and the real-time driving data are formed into a maintenance auxiliary data package.

[0022] Through this solution, the determined solution is clearly communicated to the user, including maintenance suggestions, adjustment methods or further inspection steps, so that the user can identify and take appropriate measures. By continuously monitoring the vehicle's operating status and sensor data, the performance changes of the vehicle are monitored in real time to ensure that the measures taken by the user are effective. Real-time driving data such as vibration, temperature, speed, etc. are collected and compared with the solutions output to the user to evaluate the effectiveness of the solution. Based on the real-time monitoring results and analysis results, it is judged whether the user has successfully solved the vehicle driving problem, so as to provide feedback or further guidance to the user. If the problem is not solved, the vehicle driving problem and real-time driving data will be collected and organized to form a maintenance assistance data package to provide the maintenance personnel with necessary diagnostic information.

[0023] In a second aspect, the present application provides a vehicle driving behavior analysis system integrated with speech recognition, the system comprising: A content analysis module, used to obtain voice input content, analyze the voice input content, and identify vehicle driving problems; A data retrieval module, used to determine the driving link according to the vehicle driving problem, and retrieve real-time driving data according to the driving link; The solution determination module is used to analyze the real-time driving data, determine the problem relevance, and determine the solution based on the problem relevance.

[0024] Optionally, when the content analysis module obtains the voice input content, it is used to: In response to a voice dialogue command triggered by a user, a voice signal collected by an onboard microphone array is acquired; Performing noise reduction and endpoint detection on the speech signal to extract valid speech segments; The valid voice segment is converted into text content to obtain the voice input content.

[0025] Optionally, when the content analysis module analyzes the voice input content and identifies vehicle driving problems, it is used to: Extracting a keyword set from the voice input content, the keyword set including abnormal description words, component name words, and driving scene words; Analyze the component name words and the driving scene words to determine the vehicle driving link; Analyze the abnormal description words to determine the link problem; The link problem and the vehicle driving link are regarded as the vehicle driving problem.

[0026] Optionally, when the data retrieval module retrieves the real-time driving data according to the driving link, it is used to: Analyze the voice input content to determine a trigger timestamp; Determine the time window when the problem occurred based on the voice trigger timestamp; A multi-dimensional data stream within the time window is extracted from vehicle sensors and controllers as the real-time driving data.

[0027] Optionally, the real-time driving data includes vehicle body data, and the solution determination module analyzes the real-time driving data to determine the problem relevance, and is used to: Acquire driving environment data during driving; Analyzing the vehicle driving problem based on the driving environment data to determine the abnormality type; Analyze the real-time driving data to determine frequency domain characteristics; Determining, based on the frequency domain characteristics, a vibration resonance peak position associated with the abnormal type; Determining a candidate abnormal component according to the vibration resonance peak position; The vehicle body data is analyzed to determine the problem correlation between the candidate abnormal component and the vehicle driving problem.

[0028] Optionally, when the solution determination module determines the solution according to the problem relevance, it is used to: Analyze the driving environment data to determine the road slope and air humidity during driving; Analyze the road slope and the air humidity to determine the impact of environmental interference; Determining the coordinated abnormality degree of the candidate abnormal components during driving according to the problem correlation; A solution is determined based on the environmental interference impact and the coordination abnormality degree.

[0029] Optionally, the driving environment data includes a road bump index, and the solution determination module analyzes the vehicle body data to determine the problem correlation between the candidate abnormal component and the vehicle driving problem, and is used to: Analyzing the candidate abnormal component to determine the mechanical properties of the candidate abnormal component; According to the mechanical characteristics, a preset component feature database is retrieved to obtain a standard vibration frequency range of the candidate abnormal component; Analyzing the vibration resonance peak position in the frequency domain characteristics, and calculating the offset between the vibration resonance peak position and the standard vibration frequency range; According to the offset and the road bump index, the component abnormality confidence of each candidate abnormal component is determined and calculated using the following formula: ; in, Indicates the abnormal confidence level of the component; represents the offset; Indicates the influence weight of the offset on the abnormality of the component; Indicates the influence weight of the road bump index on the abnormality of the component; represents the road bumpiness index; Determining the current wear condition of each candidate abnormal component according to the preset component feature database; determining an abnormal threshold value according to the current wear condition; Comparing the component abnormality confidence with the abnormality threshold of the corresponding candidate abnormal component; If the component abnormality confidence exceeds the abnormality threshold of the corresponding candidate abnormal component, it is determined that the candidate abnormal component is associated with the vehicle driving problem.

[0030] Optionally, when the solution determination module determines the degree of coordinated abnormality of the candidate abnormal component during driving according to the problem relevance, it is used to: Acquire sensor data related to the candidate abnormal component within the time window; analyzing the sensor data to determine phase synchronization of the candidate abnormal component; Calculating a mutual correlation coefficient between each sensor signal and each candidate abnormal component according to the phase synchronization; According to the mutual correlation coefficient, the coordinated abnormality degree of the candidate abnormal components during the driving process is determined.

[0031] Optionally, the automobile driving behavior analysis system integrated with speech recognition further includes a data formation module, which is used to: After the solution is output, the vehicle status is monitored in real time, and according to the monitoring result, it is determined whether the user has solved the vehicle driving problem; If the problem is not solved, the vehicle driving problem and the real-time driving data are formed into a maintenance auxiliary data package. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0033] Figure 1A schematic diagram of an application scenario provided for an embodiment of the present application; Figure 2 A flowchart of a method for analyzing automobile driving behavior with integrated speech recognition provided in one embodiment of the present application; Figure 3 A schematic diagram of the structure of a vehicle driving behavior analysis system with integrated voice recognition provided in one embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0035] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0036] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0037] Traditional driving behavior analysis methods usually rely on real-time data from multiple sensors in the vehicle, such as accelerator pedal opening, braking frequency, vehicle speed, acceleration, etc. Therefore, how to efficiently and accurately identify vehicle problems and provide solutions in a timely manner is still a major challenge facing current vehicle diagnostic systems.

[0038] Based on this, the present application provides a method and system for analyzing automobile driving behavior with integrated voice recognition, which obtains voice input content, analyzes the voice input content, and identifies vehicle driving problems; determines the driving link according to the vehicle driving problem, and retrieves real-time driving data according to the driving link; analyzes the real-time driving data, determines the problem correlation, and determines the solution according to the problem correlation. Voice input provides users with a convenient way to report vehicle problems without manual operation, thereby reducing distraction during driving and improving driving safety. Converting voice input content into text content improves the efficiency and accuracy of information processing. Through content extraction, quickly locate the problems described by the user, and provide a clear direction for diagnosis and analysis. Associating voice input content with specific driving links helps to more accurately analyze and solve problems in the driving link. By retrieving real-time driving data, the actual operating status of the vehicle is provided when vehicle driving problems occur, providing an important data basis for diagnosis. By analyzing real-time driving data, abnormal patterns in vehicle operation are found, which helps to identify the root cause of vehicle driving problems. Combining real-time driving data with environmental factors, vehicle component performance and other factors improves the accuracy and comprehensiveness of diagnosis.

[0039] Figure 1 A schematic diagram of an application scenario provided by the present application is used when analyzing driving behavior. Specifically, the method provided by the present application is applied to any vehicle-mounted chip, and the vehicle-mounted chip interacts with the user. Voice input provides a convenient way for users to report vehicle problems without manual operation, thereby reducing distraction during driving and improving driving safety. The voice input content is converted into text content to improve the efficiency and accuracy of information processing. Through content extraction, the problems described by the user can be quickly located to provide a clear direction for diagnosis and analysis. Associating the voice input content with the specific driving link helps to more accurately analyze and solve the problems in the driving link. By retrieving real-time driving data, the actual operating status of the vehicle is provided when the vehicle driving problem occurs, providing an important data basis for diagnosis. By analyzing real-time driving data, abnormal patterns in vehicle operation are found, which helps to identify the root cause of vehicle driving problems. Combining real-time driving data with environmental factors, vehicle component performance and other factors improves the accuracy and comprehensiveness of diagnosis.

[0040] For specific implementation methods, please refer to the following embodiments.

[0041] Figure 2 This is a flow chart of a method for analyzing vehicle driving behavior with integrated voice recognition provided in one embodiment of the present application. The method of this embodiment can be applied to the vehicle-mounted chip in the above scenario. Figure 2 As shown, the method includes: S201, obtaining voice input content, analyzing the voice input content, and identifying vehicle driving problems; The voice input content can be a description of the user's voice conversation commands, which is collected by the vehicle microphone array and converted into text content.

[0042] Vehicle driving problems may include various issues encountered during driving, such as vehicle performance, driving experience, safety warnings, etc.

[0043] Specifically, the user triggers the speech recognition device by pressing a button or saying a specific wake-up word. The onboard microphone array captures the user's voice signal. The collected voice signal is preprocessed by noise reduction and endpoint detection. The processed voice signal is converted into text content. The converted text content is semantically analyzed to extract a set of keywords such as abnormal description words, component name words, and driving scenario words. Based on the extracted keyword set, the vehicle driving problems described by the user, such as abnormal engine noise and abnormal brake pedal feeling, are identified.

[0044] S202, determining the driving link according to the vehicle driving problem, and retrieving real-time driving data according to the driving link; The driving phase can be driving behaviors or operations such as acceleration, deceleration, turning, and shifting encountered during the driving process.

[0045] Real-time driving data can be extracted from vehicle sensors and controllers, reflecting the current operating status of the vehicle.

[0046] Specifically, based on the vehicle driving problem, analyze the driving link where the vehicle driving problem occurs. Use natural language processing technology to extract acceleration, turning, braking and other content related to the driving link from the voice input content. According to the determined driving link, determine the time window when the problem occurs. Extract multi-dimensional data streams within the time window from vehicle sensors and controllers, that is, real-time driving data.

[0047] S203: Analyze the real-time driving data, determine the problem relevance, and determine a solution based on the problem relevance.

[0048] Problem correlation can be determined by analyzing real-time driving data to determine the correlation between vehicle driving problems and abnormal components or driving behaviors.

[0049] The solution can be a solution to the problem provided to the user based on the results of the problem correlation analysis.

[0050] Specifically, the extracted data is preprocessed by data cleaning and normalization. The vehicle body data such as accelerator pedal opening change rate, braking frequency, gear shifting operation, engine speed, steering wheel angle, tire pressure, etc. are analyzed. The frequency domain characteristics of real-time driving data are analyzed by Fourier transform signal processing technology to identify frequency domain features such as vibration resonance peaks and frequency mutations. Based on driving environment data and vehicle body data, vehicle driving problems are analyzed to determine the abnormality type. The frequency domain characteristics of real-time driving data are analyzed to identify the position of vibration resonance peaks related to the abnormality type. According to the position of the resonance peak, candidate abnormal components are determined. Driving environment data is analyzed to determine environmental factors such as road slope and air humidity during driving. The impact of environmental interference is analyzed. According to the correlation of the problems, the degree of coordinated abnormality of candidate abnormal components during driving is determined. The final solution is determined by combining the impact of environmental interference and the degree of coordinated abnormality.

[0051] Through this solution, voice input provides users with a convenient way to report vehicle problems without manual operation, thereby reducing distractions during driving and improving driving safety. Convert voice input content into text content to improve the efficiency and accuracy of information processing. Through content extraction, quickly locate the problems described by the user, and provide a clear direction for diagnosis and analysis. Associating voice input content with specific driving links helps to more accurately analyze and solve problems in the driving link. By retrieving real-time driving data, the actual operating status of the vehicle is provided when vehicle driving problems occur, providing an important data basis for diagnosis. By analyzing real-time driving data, abnormal patterns in vehicle operation are discovered, which helps to identify the root cause of vehicle driving problems. Combine real-time driving data with environmental factors, vehicle component performance and other factors to improve the accuracy and comprehensiveness of diagnosis.

[0052] In some embodiments, in response to a voice dialogue command triggered by a user, a voice signal collected by a vehicle-mounted microphone array is obtained; noise reduction and endpoint detection are performed on the voice signal to extract valid voice segments; the valid voice segments are converted into text content to obtain voice input content.

[0053] The voice dialogue command may be a command issued by the user to the vehicle-mounted device via voice.

[0054] A vehicle microphone array may be an array of multiple microphones installed inside a vehicle to capture sounds inside and outside the vehicle.

[0055] The voice signal can be the sound wave emitted by the user through the vehicle-mounted microphone array, which is converted into an electrical signal by the microphone and then transmitted to the voice recognition device for processing.

[0056] The valid speech segment may be a speech signal portion that is confirmed to contain useful information after endpoint detection and noise reduction processing.

[0057] The text content may be content in text form converted by a voice recognition device from a valid voice segment.

[0058] Specifically, a wake-up word is pre-set, and a wake-up word detection algorithm is used to identify the user's voice dialogue instructions. Once the wake-up word is detected, the voice recognition device is activated and ready to receive and process the user's voice dialogue voice instructions. The voice signal emitted by the user is captured through the vehicle microphone array. The collected voice signal is subjected to noise reduction processing, and the endpoint detection algorithm is used to determine the starting and ending points of the voice signal, so as to extract the valid voice segment. The extracted valid voice segment is input into the voice recognition device, and the voice signal is converted into text content through the voice recognition engine to generate voice input content.

[0059] Through this solution, a natural interface for users to interact with vehicle equipment is provided, allowing users to trigger device functions through voice commands without distracting their attention, thereby improving driving safety. The on-board microphone array can capture sounds inside and outside the car, providing a sound source for the voice recognition device, so that users can issue voice commands at any location. Noise reduction processing can significantly reduce the interference of environmental noise and noise inside the car, and improve the accuracy of voice recognition equipment. Endpoint detection can accurately identify the starting and ending points of speech, extract valid voice segments, and avoid processing invalid information. By extracting valid voice segments, we focus on processing sound data related to voice commands, reduce the amount of calculation and processing time, and improve response speed. The voice recognition device converts voice signals into text content, which helps to analyze and respond to specific user needs. The final voice input content contains the vehicle driving problem described by the user, provides information for further analysis and diagnosis, and helps to provide more personalized solutions.

[0060] In some embodiments, a set of keywords is extracted from the voice input content; component name words and driving scenario words are analyzed to determine the vehicle driving link; abnormal description words are analyzed to determine link problems; link problems and vehicle driving links are treated as vehicle driving problems.

[0061] The keyword set may be a set of key information extracted from the voice input content, and the keyword set includes abnormal description words, component name words, and driving scenario words.

[0062] Abnormal description words may be words such as abnormal noise, shaking, and jamming, etc., used to describe abnormal phenomena of the vehicle.

[0063] Part name words can be the names of vehicle parts such as engines, brakes, tires, etc. mentioned when describing the problem.

[0064] Driving scenario words can be driving environments such as acceleration, turning, and high-speed driving mentioned when describing the problem.

[0065] Link problems may be specific problems such as weak acceleration, abnormal brake pedal feeling, etc. that occur during the driving process.

[0066] Specifically, natural language processing technology is used to perform semantic analysis on the user's voice input content to identify a keyword set including abnormal description words, component name words, and driving scenario words. The vehicle component mentioned by the user is determined through the identified component name words. The driving link where the problem occurs is determined through the identified driving scenario words. The identified component name words are associated with the driving scenario words, and the vehicle driving link is determined based on the association results. The abnormal description words are analyzed to determine the nature of the vehicle driving problem. The abnormal description words and the determined vehicle driving link are combined to determine the link problem. The determined link problem is combined with the vehicle driving link to form a complete vehicle driving problem.

[0067] Through this solution, a keyword set is extracted from the user's voice input through natural language processing technology, which helps to identify the user's problem. Extracting a keyword set helps to quickly locate vehicle driving problems. Abnormal descriptors such as abnormal noise and shaking help to identify the nature of the vehicle driving problem; component name words such as engine and brake help to determine the vehicle components involved in the vehicle driving problem; driving scene words such as acceleration and turning help to determine the driving link where the problem occurs. By analyzing the keyword set, determining in which driving link the problem described by the user occurs, such as acceleration, deceleration, turning, etc., helps to narrow the scope of problem diagnosis and improve diagnostic efficiency. By analyzing the abnormal descriptors, determining the abnormal type described by the user, such as mechanical failure, performance degradation, etc., helps to provide more accurate solutions. Combining link problems with vehicle driving links to form a complete vehicle driving problem helps to provide a basis for real-time driving data retrieval, problem correlation analysis, and solution determination.

[0068] In some embodiments, the voice input content is analyzed to determine the trigger timestamp; based on the voice trigger timestamp, the time window in which the problem occurred is determined; and the multi-dimensional data stream within the time window is extracted from the vehicle sensors and controllers as real-time driving data.

[0069] The trigger timestamp may be the exact time point when the user issues the voice command, and may be expressed in a date and time format.

[0070] The time window may be a time period before and after the trigger timestamp and may be expressed in the form of a time range.

[0071] The multi-dimensional data stream can be real-time driving data such as speed, acceleration, accelerator pedal opening, brake status, engine speed, etc. extracted from vehicle sensors and controllers.

[0072] Specifically, natural language processing technology is used to perform semantic analysis on the user's voice input content. The exact time point when the user issued the voice command, i.e., the trigger timestamp, is determined through the voice input content. According to the trigger timestamp, the time window in which the problem occurred is determined. Multi-dimensional data streams such as speed, acceleration, accelerator pedal opening, brake status, engine speed, etc. within the time window are extracted from vehicle sensors and controllers, and the multi-dimensional data streams are used as real-time driving data.

[0073] This solution analyzes the user's voice input content and identifies the user's intentions and needs, thereby providing more personalized services. By determining the trigger timestamp of the user's voice command, the time window when the problem occurred is accurately recorded, which helps to diagnose and analyze vehicle driving problems. Determining the time window when the problem occurred based on the trigger timestamp helps to obtain sufficient multi-dimensional data for analysis, thereby improving the accuracy of diagnosis. Extracting multi-dimensional data streams within the time window from vehicle sensors and controllers provides an important data basis for problem diagnosis.

[0074] In some embodiments, driving environment data is acquired during driving; based on the driving environment data, the vehicle driving problem is analyzed to determine the abnormality type; the real-time driving data is analyzed to determine the frequency domain characteristics; based on the frequency domain characteristics, the position of the vibration resonance peak associated with the abnormality type is determined; based on the vibration resonance peak position, candidate abnormal components are determined; and the vehicle body data itself is analyzed to determine the problem correlation between the candidate abnormal components and the vehicle driving problem.

[0075] Driving environment data can be external environmental information such as road slope, road conditions, air humidity, temperature, and lighting conditions collected by the vehicle during driving.

[0076] The abnormality type may be a classification of abnormal phenomena that occur during the driving of the vehicle.

[0077] The frequency domain features may be features obtained after frequency domain analysis of the vehicle body vibration signal.

[0078] The vibration resonance peak position may be a frequency position related to the vehicle body vibration in the frequency domain analysis result.

[0079] The candidate abnormal components may be a list of components that are determined to have failed based on frequency domain characteristics and abnormality types.

[0080] The vehicle's own data can be data collected during the vehicle's driving that reflects the vehicle's own operating status, such as speed, acceleration, engine speed, accelerator pedal opening, and brake status.

[0081] Specifically, driving environment data such as road slope, road condition, air humidity, and temperature are obtained from vehicle sensors. Driving environment data is analyzed to analyze the impact of environmental factors on vehicle performance. Based on the analysis results, the abnormal type of vehicle driving problems is determined. Vehicle body data is extracted from vehicle sensors and controllers. The vehicle body data is analyzed in the frequency domain using fast Fourier transform technology to identify frequency domain features such as vibration mode and resonance frequency. Based on the frequency domain features, the position of the vibration resonance peak related to vehicle body vibration is identified. Based on the position of the vibration resonance peak, the faulty component is matched with the preset component feature database to determine the candidate abnormal component. Based on the vehicle's own data, the standard vibration frequency range of the candidate abnormal component is analyzed and compared with the actual measured vibration frequency. Based on the analysis results, the problem correlation between the candidate abnormal component and the vehicle driving problem is determined.

[0082] Through this solution, the driving environment data of the vehicle during driving is collected, which helps to identify the external conditions of vehicle operation and analyze the environmental impact of vehicle problems. By analyzing the driving environment data, the potential correlation between environmental factors and vehicle problems can be identified. For example, slippery road conditions lead to longer braking distances, which affects the braking performance of the vehicle. According to the determined abnormality type, preliminary diagnostic results are provided to provide users with corresponding solutions. Extracting the vehicle body's own data from vehicle sensors and controllers, such as speed, acceleration, engine speed, accelerator pedal opening, brake status, etc., helps reflect the actual operating status of the vehicle and helps analyze vehicle problems. Frequency domain features help diagnose vehicle driving problems related to vibration. The resonance peak position helps to determine the abnormal component that has failed. The abnormal component is a candidate abnormal component for further diagnosis and inspection. Determine the correlation between the candidate abnormal component and the vehicle driving problem described by the user, so as to provide users with accurate diagnostic results and solutions. Determining the problem correlation helps users to identify the vehicle condition in a timely manner and take corresponding maintenance measures.

[0083] In some embodiments, driving environment data is analyzed to determine the road slope and air humidity during driving; the road slope and air humidity are analyzed to determine the impact of environmental interference; based on the correlation of problems, the coordinated abnormality of candidate abnormal components during driving is determined; based on the impact of environmental interference and the coordinated abnormality, a solution is determined.

[0084] Road grade can be the degree of inclination of the road surface and can be expressed as an angle.

[0085] Air humidity can be the amount of water vapor in the atmosphere and can be expressed as relative humidity (%).

[0086] Environmental interference effects can be a potential negative impact of driving environment data on vehicle performance and driving experience.

[0087] The collaborative abnormality degree may be the possibility that the candidate abnormal components simultaneously become abnormal under the influence of environmental interference.

[0088] Specifically, collect driving environment data such as road slope and air humidity. Analyze the environmental interference effects of road slope and air humidity on vehicle performance. Use speech recognition technology to analyze vehicle driving problems in voice input content, combine vehicle driving problems with environmental interference effects and real-time driving data, and analyze problem relevance. According to the problem relevance, determine candidate abnormal components that may fail. Analyze the coordinated abnormality of candidate abnormal components during driving, that is, the possibility that candidate abnormal components will be abnormal at the same time under the influence of environmental interference. Determine a solution based on the environmental interference effects and the coordinated abnormality of candidate abnormal components.

[0089] Through this solution, by analyzing driving environment data, the specific environmental conditions encountered by the vehicle during driving, such as road slope and air humidity, are identified. Driving environment data helps diagnose the environmental interference effects on vehicle problems. By analyzing road slope and air humidity, the environmental interference effects of road slope and air humidity on vehicle performance are determined, such as slope affecting the acceleration and braking performance of the vehicle, and humidity affecting the tire grip and braking of the vehicle. By analyzing the correlation of problems, the possibility of candidate abnormal components being abnormal at the same time under the influence of environmental interference is determined, which helps to narrow the scope of the problem and improve the accuracy of diagnosis. Combined with the environmental interference effects and the coordinated abnormality of candidate abnormal components, a solution is determined to provide users with corresponding maintenance suggestions or operation guidelines.

[0090] In some embodiments, the candidate abnormal components are analyzed to determine the mechanical characteristics of the candidate abnormal components; based on the mechanical characteristics, a preset component feature database is retrieved to obtain the standard vibration frequency range of the candidate abnormal components; the vibration resonance peak position in the frequency domain characteristics is analyzed to calculate the offset between the vibration resonance peak position and the standard vibration frequency range; based on the offset and the road bump index, the component abnormality confidence of each candidate abnormal component is determined, and the calculation is performed using the following formula: (1) in, Indicates the confidence level of component abnormality; Indicates the offset; Indicates the weight of the impact of the offset on the abnormality of the component; Indicates the weight of the impact of the road bump index on the abnormality of the component; Indicates a road bumpiness index; determines the current wear condition of each candidate abnormal component according to a preset component feature database; determines an abnormal threshold according to the current wear condition; compares the component abnormality confidence with the abnormal threshold of the corresponding candidate abnormal component; if the component abnormality confidence exceeds the abnormal threshold of the corresponding candidate abnormal component, it is determined that the candidate abnormal component is associated with the vehicle driving problem.

[0091] Mechanical characteristics can be the physical and mechanical performance parameters of each component of the vehicle.

[0092] The preset component characteristic database may be a preset database containing characteristic parameters of each component of the vehicle under normal working conditions, which is pre-stored in the vehicle chip and called when in use.

[0093] The standard vibration frequency range may be the vibration frequency range expected for each component of the vehicle under normal working conditions.

[0094] The offset can be the difference between the actual measured data and the standard value in the preset component feature database.

[0095] The road roughness index can be a quantitative indicator of the degree of road roughness.

[0096] The component abnormality confidence level may be the confidence level of the device in the judgment of the component abnormality.

[0097] The impact weight can be the relative importance of different factors such as environmental factors, driving behavior, component status, etc. on vehicle performance.

[0098] The current wear condition may be the current wear degree of each component of the vehicle.

[0099] The abnormality threshold may be a critical value of component performance abnormality defined in a preset component feature database.

[0100] Specifically, the candidate abnormal parts are analyzed in detail for their mechanical properties, such as hardness, elastic modulus, and strength. Based on the mechanical properties of the candidate abnormal parts, the standard vibration frequency range of the candidate abnormal parts under normal working conditions is obtained from the preset part feature database. The vehicle body vibration signal is analyzed in the frequency domain to identify the vibration resonance peak position, and compared with the standard vibration frequency range to calculate the offset ( ). The road bump index is measured by the vehicle's suspension equipment sensor. The component abnormality confidence is calculated using formula (1) ( ). The current wear conditions of the candidate abnormal parts, such as the degree of wear and service life, are obtained from the preset part feature database. According to the current wear conditions of the candidate abnormal parts, the critical value of the abnormal performance of the parts, i.e., the abnormal threshold, is determined. The calculated confidence of the abnormality of the parts is compared with the abnormal threshold to determine whether the parts are beyond the normal range. If the confidence of the abnormality of the parts exceeds the abnormal threshold, there is a correlation between the candidate abnormal parts and the vehicle driving problem.

[0101] Through this solution, the performance parameters and expected behaviors of the candidate abnormal parts are identified by analyzing the mechanical characteristics of the candidate abnormal parts. By calling the preset component feature database, the standard vibration frequency range of the candidate abnormal parts is obtained to provide a reference for frequency domain analysis. Through frequency domain analysis, the vibration resonance peak position is identified, and the offset from the standard vibration frequency range is calculated, which helps to determine whether the candidate abnormal parts are abnormal. The component abnormality confidence is calculated by formula to quantify the possibility of abnormality of each candidate component, thereby providing a more accurate basis for diagnosis. The current wear condition of the candidate abnormal parts is obtained through the preset component feature database, which helps to evaluate the aging and loss of the candidate abnormal parts. According to the current wear condition of the candidate abnormal parts, the abnormal threshold is determined, which helps to determine whether the performance of the candidate abnormal parts exceeds the normal range. By comparing the component abnormality confidence and the abnormal threshold, it is determined which candidate abnormal parts are abnormal, and diagnostic suggestions are provided to the user. If the component abnormality confidence exceeds the abnormal threshold, it is confirmed that the candidate abnormal part is related to the vehicle driving problem and is the cause of the problem, thereby providing maintenance suggestions or operation guidelines to the user.

[0102] In some embodiments, sensor data related to the candidate abnormal components within a time window is acquired; the sensor data is analyzed to determine the phase synchronization of the candidate abnormal components; based on the phase synchronization, a mutual correlation coefficient between each sensor signal and each candidate abnormal component is calculated; based on the mutual correlation coefficient, the coordinated abnormality of the candidate abnormal components during driving is determined.

[0103] Sensor data may be data collected from various sensors of the vehicle, including throttle opening, brake pressure, and engine speed.

[0104] Phase synchronicity may be the phase relationship between different sensor signals, ie, the relative positions of different sensor signals in time.

[0105] The sensor signal can be an electrical signal output by the sensor and is a specific form of sensor data.

[0106] The cross-correlation coefficient may be a measure of the degree of correlation between two sensor signals.

[0107] Specifically, sensor data such as vibration sensors and temperature sensors related to candidate abnormal components within a time window are collected from the vehicle's sensors. The collected sensor data are preprocessed by data cleaning and normalization. The sensor data and the component vibration signal are Hilbert transformed to calculate the standard deviation of the instantaneous phase difference. According to the standard deviation, the phase synchronization of the sensor data is analyzed, that is, the phase relationship between different sensor signals is analyzed. According to the phase synchronization, the sensor signals are time synchronized to ensure that different sensor signals are compared within the same time window. The cross-correlation function of each pair of sensor signals and the candidate abnormal component is calculated. The maximum value, i.e., the cross-correlation coefficient, is extracted from the cross-correlation function. The calculated cross-correlation coefficient is analyzed to evaluate the collaborative working between the candidate abnormal component and other components. A weighted scoring model including the cross-correlation coefficient peak value, time delay, and signal energy ratio is established to output a normalized score in the range of 0-1. The degree of collaborative abnormality of each candidate abnormal component is determined based on the evaluation situation and the normalized score.

[0108] Through this solution, sensor data related to candidate abnormal components in the time window is collected, which helps to analyze the basis of the performance of candidate abnormal components. By preprocessing the sensor data, the real-time working status and performance indicators of candidate abnormal components are identified. Analyzing the phase relationship between different sensor signals helps to determine whether the candidate abnormal components are working in the correct timing, thereby identifying potential collaborative abnormalities. By calculating the mutual correlation coefficient, the similarity and timing relationship between different sensor signals are quantified, which helps to evaluate the collaborative working between candidate abnormal components. Based on the mutual correlation coefficient, the collaborative abnormality of each candidate abnormal component during driving is determined, which helps to identify which candidate abnormal components jointly cause vehicle problems.

[0109] In some embodiments, after the solution is output, the vehicle status is monitored in real time, and based on the monitoring results, it is determined whether the user has solved the vehicle driving problem; if not, the vehicle driving problem and real-time driving data are formed into a maintenance assistance data package.

[0110] The vehicle status may be the real-time operating status of the vehicle during driving, such as engine speed, vehicle speed, accelerator pedal position, brake status, and suspension device working status.

[0111] The monitoring result may be the output after real-time monitoring and analysis of the vehicle status.

[0112] The maintenance assistance data package may be a file or information collection containing vehicle driving problems and real-time driving data that is collected and organized when a user fails to solve a vehicle driving problem.

[0113] Specifically, solutions such as maintenance suggestions, adjustment methods, and further inspection steps are output to the user in the form of text or voice. After the user attempts to solve the vehicle driving problem, continue to monitor the vehicle's operating status and sensor data in real time. Collect real-time driving data such as vehicle operating status and sensor data, and compare and analyze them with the solutions output to the user. Based on the real-time monitoring results and analysis results, evaluate whether the user has successfully solved the vehicle driving problem. If the user fails to solve the vehicle driving problem, the vehicle driving problem and real-time driving data collected and organized will form a maintenance assistance data package.

[0114] Through this solution, the determined solution is clearly communicated to the user, including maintenance suggestions, adjustment methods or further inspection steps, so that the user can identify and take appropriate measures. By continuously monitoring the vehicle's operating status and sensor data, the performance changes of the vehicle are monitored in real time to ensure that the measures taken by the user are effective. Real-time driving data such as vibration, temperature, speed, etc. are collected and compared with the solutions output to the user to evaluate the effectiveness of the solution. Based on the real-time monitoring results and analysis results, it is judged whether the user has successfully solved the vehicle driving problem, so as to provide feedback or further guidance to the user. If the problem is not solved, the vehicle driving problem and real-time driving data will be collected and organized to form a maintenance assistance data package to provide the maintenance personnel with necessary diagnostic information.

[0115] Figure 3 A structural diagram of a vehicle driving behavior analysis system integrated with speech recognition provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, the automobile driving behavior analysis system 300 integrated with speech recognition in this embodiment includes: a content analysis module 301 , a data retrieval module 302 , and a solution determination module 303 .

[0116] The content analysis module 301 is used to obtain the voice input content, analyze the voice input content, and identify the vehicle driving problem; The data retrieval module 302 is used to determine the driving link according to the vehicle driving problem, and retrieve the real-time driving data according to the driving link; The solution determination module 303 is used to analyze the real-time driving data, determine the problem relevance, and determine a solution based on the problem relevance.

[0117] Optionally, when the content analysis module 301 obtains the voice input content, it is used to: In response to a voice dialogue command triggered by a user, a voice signal collected by an onboard microphone array is acquired; Performing noise reduction and endpoint detection on the speech signal to extract valid speech segments; The valid voice segment is converted into text content to obtain the voice input content.

[0118] Optionally, when the content analysis module 301 analyzes the voice input content and identifies vehicle driving problems, it is used to: Extracting a keyword set from the voice input content, the keyword set including abnormal description words, component name words, and driving scene words; Analyze the component name words and the driving scene words to determine the vehicle driving link; Analyze the abnormal description words to determine the link problem; The link problem and the vehicle driving link are regarded as the vehicle driving problem.

[0119] Optionally, when the data retrieval module 302 retrieves the real-time driving data according to the driving link, it is used to: Analyze the voice input content to determine a trigger timestamp; Determine the time window when the problem occurred based on the voice trigger timestamp; A multi-dimensional data stream within the time window is extracted from vehicle sensors and controllers as the real-time driving data.

[0120] Optionally, the real-time driving data includes vehicle data. When the solution determination module 303 analyzes the real-time driving data and determines the problem relevance, it is used to: Acquire driving environment data during driving; Analyzing the vehicle driving problem based on the driving environment data to determine the abnormality type; Analyze the real-time driving data to determine frequency domain characteristics; Determining, based on the frequency domain characteristics, a vibration resonance peak position associated with the abnormal type; Determining a candidate abnormal component according to the vibration resonance peak position; The vehicle body data is analyzed to determine the problem correlation between the candidate abnormal component and the vehicle driving problem.

[0121] Optionally, when the solution determination module 303 determines a solution according to the problem relevance, it is used to: Analyze the driving environment data to determine the road slope and air humidity during driving; Analyze the road slope and the air humidity to determine the impact of environmental interference; According to the problem relevance, the coordinated abnormality degree of the candidate abnormal component during driving is determined; according to the environmental interference impact and the coordinated abnormality degree, a solution is determined.

[0122] Optionally, the driving environment data includes a road bump index, and the solution determination module 303 analyzes the vehicle body data to determine the problem correlation between the candidate abnormal component and the vehicle driving problem, and is used to: Analyzing the candidate abnormal component to determine the mechanical properties of the candidate abnormal component; According to the mechanical characteristics, a preset component feature database is retrieved to obtain a standard vibration frequency range of the candidate abnormal component; Analyzing the vibration resonance peak position in the frequency domain characteristics, and calculating the offset between the vibration resonance peak position and the standard vibration frequency range; According to the offset and the road bump index, the component abnormality confidence of each candidate abnormal component is determined and calculated using the following formula: ; in, Indicates the abnormal confidence level of the component; represents the offset; Indicates the influence weight of the offset on the abnormality of the component; Indicates the influence weight of the road bump index on the abnormality of the component; represents the road bumpiness index; Determining the current wear condition of each candidate abnormal component according to the preset component feature database; determining an abnormal threshold value according to the current wear condition; Comparing the component abnormality confidence with the abnormality threshold of the corresponding candidate abnormal component; If the component abnormality confidence exceeds the abnormality threshold of the corresponding candidate abnormal component, it is determined that the candidate abnormal component is associated with the vehicle driving problem.

[0123] Optionally, when the solution determination module 303 determines the degree of coordinated abnormality of the candidate abnormal component during driving according to the problem relevance, it is used to: Acquire sensor data related to the candidate abnormal component within the time window; analyzing the sensor data to determine phase synchronization of the candidate abnormal component; Calculating a mutual correlation coefficient between each sensor signal and each candidate abnormal component according to the phase synchronization; According to the mutual correlation coefficient, the coordinated abnormality degree of the candidate abnormal components during the driving process is determined.

[0124] Optionally, the automobile driving behavior analysis system integrated with speech recognition further includes a data forming module 304, which is used to: After the solution is output, the vehicle status is monitored in real time, and according to the monitoring result, it is determined whether the user has solved the vehicle driving problem; If the problem is not solved, the vehicle driving problem and the real-time driving data are formed into a maintenance auxiliary data package.

[0125] The system of this embodiment can be used to execute the method of any of the above embodiments. The implementation principles and technical effects are similar and will not be described in detail here.

Claims

1. A method for analyzing automobile driving behavior with integrated speech recognition, characterized in that: include: Acquire voice input content, analyze the voice input content, and identify vehicle driving problems; Determine the driving link according to the vehicle driving problem, and retrieve real-time driving data according to the driving link; The real-time driving data is analyzed to determine problem relevance, and a solution is determined based on the problem relevance.

2. The method according to claim 1, characterized in that: The obtaining of voice input content includes: In response to a voice dialogue command triggered by a user, a voice signal collected by an onboard microphone array is acquired; Performing noise reduction and endpoint detection on the speech signal to extract valid speech segments; The valid voice segment is converted into text content to obtain the voice input content.

3. The method according to claim 1, characterized in that The analyzing the voice input content and identifying the vehicle driving problem includes: Extracting a keyword set from the voice input content, the keyword set including abnormal description words, component name words, and driving scene words; Analyze the component name words and the driving scene words to determine the vehicle driving link; Analyze the abnormal description words to determine the link problem; The link problem and the vehicle driving link are regarded as the vehicle driving problem.

4. The method according to claim 1, characterized in that: The step of retrieving real-time driving data according to the driving link includes: Analyze the voice input content to determine a trigger timestamp; Determine the time window when the problem occurred based on the voice trigger timestamp; A multi-dimensional data stream within the time window is extracted from vehicle sensors and controllers as the real-time driving data.

5. The method according to claim 4, characterized in that The real-time driving data includes vehicle body data, and the analyzing the real-time driving data to determine the problem correlation includes: Acquire driving environment data during driving; Analyzing the vehicle driving problem based on the driving environment data to determine the abnormality type; Analyze the real-time driving data to determine frequency domain characteristics; Determining, based on the frequency domain characteristics, a vibration resonance peak position associated with the abnormal type; Determining a candidate abnormal component according to the vibration resonance peak position; The vehicle body data is analyzed to determine the problem correlation between the candidate abnormal component and the vehicle driving problem.

6. The method according to claim 5, characterized in that Determining a solution according to the correlation of the problems includes: Analyze the driving environment data to determine the road slope and air humidity during driving; Analyze the road slope and the air humidity to determine the impact of environmental interference; Determining the coordinated abnormality degree of the candidate abnormal components during driving according to the problem correlation; A solution is determined based on the environmental interference impact and the coordination abnormality degree.

7. The method according to claim 5, characterized in that The driving environment data includes a road bump index, and the analyzing the vehicle body data to determine the problem correlation between the candidate abnormal component and the vehicle driving problem includes: Analyzing the candidate abnormal component to determine the mechanical properties of the candidate abnormal component; According to the mechanical characteristics, a preset component feature database is retrieved to obtain a standard vibration frequency range of the candidate abnormal component; Analyzing the vibration resonance peak position in the frequency domain characteristics, and calculating the offset between the vibration resonance peak position and the standard vibration frequency range; According to the offset and the road bump index, the component abnormality confidence of each candidate abnormal component is determined and calculated using the following formula: ； in, Indicates the abnormal confidence level of the component; represents the offset; Indicates the influence weight of the offset on the abnormality of the component; Indicates the influence weight of the road bump index on the abnormality of the component; represents the road bumpiness index; Determining the current wear condition of each candidate abnormal component according to the preset component feature database; determining an abnormal threshold value according to the current wear condition; Comparing the component abnormality confidence with the abnormality threshold of the corresponding candidate abnormal component; If the component abnormality confidence exceeds the abnormality threshold of the corresponding candidate abnormal component, it is determined that the candidate abnormal component is associated with the vehicle driving problem.

8. The method according to claim 6, characterized in that Determining the coordinated abnormality degree of the candidate abnormal component during driving according to the problem correlation includes: Acquire sensor data related to the candidate abnormal component within the time window; analyzing the sensor data to determine phase synchronization of the candidate abnormal component; Calculating a mutual correlation coefficient between each sensor signal and each candidate abnormal component according to the phase synchronization; According to the mutual correlation coefficient, the coordinated abnormality degree of the candidate abnormal components during the driving process is determined.

9. The method according to claim 1, characterized in that: The method further comprises: After the solution is output, the vehicle status is monitored in real time, and according to the monitoring result, it is determined whether the user has solved the vehicle driving problem; If the problem is not solved, the vehicle driving problem and the real-time driving data are formed into a maintenance auxiliary data package.

10. A vehicle driving behavior analysis system integrated with speech recognition, characterized in that: The method as claimed in any one of claims 1 to 9 comprises: A content analysis module, used to obtain voice input content, analyze the voice input content, and identify vehicle driving problems; A data retrieval module, used to determine the driving link according to the vehicle driving problem, and retrieve real-time driving data according to the driving link; The solution determination module is used to analyze the real-time driving data, determine the problem relevance, and determine the solution based on the problem relevance.

Images