Speech recognition navigation method based on planning path data

Abstract

The present invention relates to the field of data processing technology, and in particular to a voice recognition navigation method based on planned path data. The method comprises the following steps: performing navigation demand analysis according to navigation request data to generate starting position data and navigation destination data; performing navigation route planning based on the starting position data and navigation destination data to generate planned path data; performing real-time voice collection according to the planned path data to generate voice data to be processed; performing optimization necessity evaluation on the voice data to be processed, and performing path replanning to generate path optimization data; transmitting the path optimization data to a cloud device to generate optimization record summary data; performing navigation optimization on a navigator device based on the optimization record summary data to generate navigator update data. The present invention realizes real-time and efficient processing of the user's navigation route change requirements by analyzing the user's voice segments during navigation.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a speech recognition navigation method based on planned path data. Background Art

[0002] Speech recognition allows users to interact with navigation systems through natural language without relying on keyboards or touch screens. It greatly improves the convenience of using portable navigation devices, especially in situations that require concentration, such as driving or walking. Voice navigation systems can better meet the personalized needs of users because they can adjust according to the user's accent, speaking speed, and word usage habits to provide a more personalized navigation experience. However, existing speech recognition navigation methods based on planned path data require specific voice commands to control the system, and the contextual connection of voice commands is weak, and voice commands cannot be understood in a timely and accurate manner. Summary of the invention

[0003] Based on this, it is necessary to provide a speech recognition navigation method based on planned path data to solve at least one of the above technical problems.

[0004] To achieve the above object, a speech recognition navigation method based on planned path data is provided, the method comprising the following steps:

[0005] Step S1: performing navigation request voice detection based on a voice acquisition device to generate navigation request data; performing device positioning based on the navigation request data to generate starting position data; performing travel mode and travel destination identification based on the navigation request data to generate travel mode data and destination information data; performing map position matching based on the travel mode data and destination information data to generate navigation destination data;

[0006] Step S2: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data; performing turn and stop state analysis based on the planned path data to generate turned and stopped data; performing voice segment screening based on the turned and stopped data to generate voice data to be processed;

[0007] Step S3: intercepting the voice data to be processed to meet the optimization requirements and generating segmented voice data; evaluating the necessity of optimization based on the segmented voice data and generating necessity evaluation data; replanning the planned path data based on the necessity evaluation data and generating path optimization data;

[0008] Step S4: Transmit the path optimization data to the cloud device to generate optimization record summary data; extract the causes of navigation defects based on the optimization record summary data to generate navigation defect feedback data; optimize the navigation equipment according to the navigation defect feedback data to generate navigator update data.

[0009] The present invention uses voice detection to issue navigation requests in a natural voice mode, thereby improving user experience and convenience. The complexity of user operations is reduced. Starting position data is generated according to device positioning, providing accurate starting position information, which is helpful for accurately planning navigation routes. By identifying travel modes and generating destination information data, the user's travel mode and destination information are understood, which is helpful for personalized navigation services and information push. Path planning based on starting position data and navigation destination data is helpful for generating the best path, saving time and resources, and providing efficient navigation guidance. By analyzing the turning and stopping states and screening voice fragments, key information is helpful for extracting key information, reducing the amount of voice data processing, and improving processing efficiency. Real-time interaction with users during navigation is ensured. The real-time and accuracy of the navigation system are improved, the user experience is enhanced, and it is ensured that users can get timely guidance and feedback during navigation. By optimizing the interception of demand voice and the necessity evaluation, the voice data is made more refined and targeted, redundant information is reduced, and data quality is improved. According to the evaluation results of the voice data, the planned path is adjusted and optimized. The intelligence and personalization of the navigation system are improved, the path planning is more in line with the actual needs of users, and the accuracy and efficiency of the navigation path are optimized. The optimized record summary data is used to record the optimization process. Based on the optimized record summary data, the navigation device is optimized and the navigation data is updated. The performance and adaptability of the navigation device are improved, so that it can better adapt to the changing road conditions and user needs, and the stability and accuracy of the navigation system are improved. Therefore, the speech recognition navigation method based on the planned path data of the present invention analyzes the user's travel status, filters the user's voice clips during navigation, so as to reduce the workload of speech analysis, and evaluates the necessity of optimizing the filtered voice clips to determine whether the user needs to optimize the navigation path, so as to promptly handle the changing needs of the user during navigation.

[0010] Preferably, step S1 comprises the following steps:

[0011] Step S11: collecting user voice based on the voice collection device to generate user voice data;

[0012] Step S12: performing navigation request voice detection on the user voice data to generate navigation request data;

[0013] Step S13: positioning the device based on the navigation request data to generate starting position data;

[0014] Step S14: classifying the sentence keywords according to the navigation request data to generate classified keyword data; identifying the travel mode and travel destination based on the classified keyword data to generate travel mode data and destination information data;

[0015] Step S15: extract keywords from the destination information data to generate destination keyword data; preset the current travel range according to the travel mode data to generate current travel range data; match the map position based on the current travel range data and the destination keyword data to generate navigation destination data.

[0016] By collecting user voice data, the present invention can better understand the user's voice characteristics and accent, providing a basis for subsequent voice recognition and interaction. It improves the accuracy and personalization of voice recognition, improves the interaction effect between the system and the user, and makes the system more user-friendly.

[0017] By detecting the user's voice data and identifying the navigation request information therein, the interaction efficiency and convenience between the user and the navigation system are improved. The ability to accurately identify the user's navigation needs from the voice data helps to respond to user needs more accurately and provide customized services. By obtaining the user's starting location information through device positioning, it helps to ensure that the navigation system can provide the most accurate and appropriate route planning. After obtaining the starting location data, the optimal navigation route can be planned for the user more accurately, saving time and resources. Extracting and classifying keywords from the sentences in the navigation request data helps to understand the user's specific needs. For example, the keyword in "I want to go to the hospital" can be classified as "hospital", so as to analyze the user's needs more accurately. Through keyword classification, the user's travel mode and destination information can be identified, providing a more accurate information basis for subsequent navigation services. Extracting keywords from the destination information helps to more accurately understand the specific place the user wants to go to, such as extracting the keyword "mall" from "go to the mall to buy things". Combining the user's current travel range and the extracted destination keywords, the possible destination range can be better preset and matched with the map location, providing support for the accurate determination of the navigation destination.

[0018] Preferably, step S14 comprises the following steps:

[0019] Step S141: performing word segmentation annotation on the navigation request data to generate word segmentation request data; performing keyword extraction on the word segmentation request data to generate initial keyword data;

[0020] Step S142: semantically analyzing the navigation request data to generate word association data; using the word association data to classify the initial keyword data to generate classified keyword data;

[0021] Step S143: Identify the travel mode based on the classified keyword data and generate travel mode data;

[0022] Step S144: performing location entity recognition based on the classified keyword data to generate location recognition result data; performing geographic location mapping based on the location recognition result data to generate mapping result data;

[0023] Step S145: performing recognition accuracy evaluation based on the mapping result data to generate recognition accuracy data;

[0024] Step S146: Perform data concatenation processing on the location recognition result data and the recognition accuracy data to generate destination information data.

[0025] The present invention extracts key information in the request by performing word segmentation and keyword extraction on navigation request data. The understanding and accuracy of the user's navigation intention are improved, and a basis for subsequent semantic analysis and keyword extraction is provided. Semantic analysis and keyword classification are performed to process the word association in the request data. The deep understanding of the navigation request data is improved, the classification and classification of keywords are optimized, and the intelligence and personalization of the navigation system are enhanced. The travel mode is identified based on the classified keyword data to identify the travel mode selected by the user. The user's travel mode is determined, which is conducive to the system to more accurately plan a navigation path suitable for the travel mode for the user. The location information in the navigation request is identified and the geographical location mapping is performed. Accurate location information is provided, and the accuracy and effectiveness of navigation are enhanced. The accuracy of the recognition result data is evaluated to evaluate the accuracy of the location recognition. The accuracy of the location recognition result is ensured, and the accurate interpretation and response capability of the navigation system to user needs is improved. Accurate destination information is generated by combining the location recognition and accuracy evaluation data. Complete and accurate navigation destination information is provided, ensuring the accuracy of navigation path planning and the optimization of user experience.

[0026] Preferably, step S15 comprises the following steps:

[0027] Step S151: Obtain navigation history data and preset frequently used address data;

[0028] Step S152: processing the destination information data based on a preset recognition accuracy threshold, and generating destination keyword data when the destination information data is greater than the preset recognition accuracy threshold;

[0029] Step S153: when the destination information data is less than or equal to the preset recognition accuracy threshold, fuzzy address matching is performed based on the preset common address data to generate destination keyword data;

[0030] Step S154: dividing the navigation history data into travel tools to generate tool division history data; formulating the travel distance range according to the tool division history data to generate travel range data;

[0031] Step S155: Preset the travel range of the current travel based on the travel mode data to generate current travel range data; match the map position of the destination keyword data based on the current travel range data to generate navigation destination data.

[0032] The present invention obtains the user's navigation history data and preset common address data. It provides an important data source for the navigation system to analyze, which is helpful to understand the user's common destinations and navigation preferences. The destination information data is processed for accuracy based on a preset threshold. The reliability of the destination information is ensured, and for data above the threshold, the accuracy and reliability of navigation are improved. When the destination information is insufficient, fuzzy matching is performed using the common address data. The destination information is supplemented, the system's ability to understand the user's intentions is enhanced, and the integrity and accuracy of navigation are improved. The navigation history data is divided into tools and the travel distance range is determined. The travel range data is preset for this trip based on the travel mode data, and the current travel range data is generated, which provides a more accurate reference range for the generation of subsequent destination data and improves the adaptability and personalization of the navigation path. The destination map position is matched based on the current travel range data. Accurate destination data is provided to ensure that the navigation system accurately interprets and responds to user needs, and enhances the effectiveness of navigation path planning and user experience.

[0033] Preferably, step S153 includes the following steps:

[0034] Step S1531: when the destination information data is less than or equal to a preset recognition accuracy threshold, generating fuzzy place entity data;

[0035] Step S1532: performing entity identification matching on the fuzzy location entity data based on the preset common address data to generate matching result data; when the matching result data is a successful match, generating destination keyword data;

[0036] Step S1533: when the matching result data indicates a matching failure, accurate address consultation is performed based on the mapping result data to generate address consultation result data;

[0037] Step S1534: performing location data association on the fuzzy location entity data according to the address consultation result data, and supplementing and correcting the preset common address data to generate common address data;

[0038] Step S1535: extract the destination address based on the commonly used address data and generate destination keyword data.

[0039] The present invention generates fuzzy place entity data. Incomplete destination information is supplemented to provide basic data for subsequent matching and query. Fuzzy place entities are matched and identified based on preset common address data. The accuracy of fuzzy place entities is improved. When the match is successful, more reliable destination keyword data is generated, and the accuracy of navigation is enhanced. When the match fails, precise address consultation is performed and consultation result data is generated. In the event of a matching failure, more accurate address information is obtained through consultation, which improves the completeness and accuracy of the destination information. The fuzzy place entity and common address data are associated and corrected according to the consultation results. The accuracy and completeness are improved, providing a more reliable data basis for the navigation system. Destination information is extracted based on the corrected common address data. More reliable destination keyword data is generated to ensure the accuracy of navigation path planning and the optimization of user experience.

[0040] Preferably, step S2 comprises the following steps:

[0041] Step S21: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data;

[0042] Step S22: performing path navigation according to the planned path data, and performing stay status analysis to generate stay status data;

[0043] Step S23: using a voice collection device and performing real-time voice collection based on the stay status data to generate voice data during the stay period;

[0044] Step S24: acquiring the movement direction based on the stay state data to generate stay direction data; extracting the turning direction from the stay direction data using the planned path data to generate turned stay data;

[0045] Step S25: filtering the voice data during the stop period based on the turn stop data to generate voice data to be processed.

[0046] The present invention generates the best path planning through the starting position and navigation destination data, provides the most optimized navigation solution, and saves time and resources. According to the starting position and destination data, a personalized navigation route can be provided, taking into account factors such as user preferences and traffic conditions. Through path navigation, real-time navigation guidance is provided to users, and through the stop state analysis, the user's stop situation during the navigation process is understood, the path planning is optimized, and the accuracy and practicality of the navigation route are ensured. The data analysis of the stop state is helpful to improve the navigation system and provide a more intelligent and real-time navigation service. Through the voice acquisition device, voice data is collected during the user's stay to understand the user's demand changes, feedback or other relevant information during the navigation process, which is helpful to improve the responsiveness of the navigation system and provide services that are closer to user needs. The user's movement direction is obtained through the stop state data, which is helpful to understand the relationship between the user's current position and the navigation path, and provide more accurate navigation guidance. Combined with the planning path data, the relevant information of the turn stop is extracted, which can provide basic information for the turn reminder, intersection guidance, etc. of the navigation system, and improve the real-time and accuracy of navigation. The voice data during the stop period is screened based on the turn stop data, which is helpful to extract the voice fragments related to the turn and reduce the amount of unnecessary voice information. The screened voice data to be processed is more likely to contain information about the reasons for turning and stopping, which improves the quality and relevance of the voice data and facilitates subsequent voice recognition and analysis.

[0047] Preferably, step S3 comprises the following steps:

[0048] Step S31: preprocessing the speech data to be processed to generate noise-reduced speech data;

[0049] Step S32: filtering out irrelevant voices from the noise reduction voice data to generate navigation-related voice data; converting the navigation-related voice data into text to generate text voice data;

[0050] Step S33: performing demand keyword recognition on the text and voice data to generate demand keyword data;

[0051] Step S34: selecting voice segments associated with the text voice data according to the required keyword data to generate segment voice data; performing optimization necessity evaluation according to the segment voice data to generate necessity evaluation data;

[0052] Step S35: screening the text and speech data for high necessity data according to the necessity evaluation data to generate data to be optimized and analyzed;

[0053] Step S36: Formulate an optimization strategy based on the analyzed data to be optimized to generate path optimization strategy data; replan the planned path data according to the path optimization strategy data to generate path optimization data.

[0054] The present invention performs pre-processing and noise reduction on the voice data to be processed. The quality of the voice data is improved and the influence of noise on voice recognition is reduced. The irrelevant voices are screened out and the processed voice data is converted into text data. The voice information related to navigation is extracted, and the text format conversion is helpful for the subsequent key information extraction and semantic analysis. The text voice data is identified with demand keywords. The navigation demand keywords proposed by the user in the voice are determined, which provides important navigation guidance for subsequent processing. The necessity of navigation optimization is evaluated to ensure the effectiveness and rationality of navigation path optimization. The text voice data is screened according to the necessity evaluation data to screen out the data that most needs to be optimized, thereby improving the pertinence and effectiveness of path optimization. According to the screened analysis data to be optimized, research and analysis are carried out, targeted optimization strategies are formulated, and path optimization strategy data are generated. The original planned path data is re-planned according to the path optimization strategy data to generate path optimization data. The optimization and effectiveness of the navigation path are ensured, and the accuracy of navigation and user experience are improved.

[0055] Preferably, step S34 includes the following steps:

[0056] Step S341: extracting feature information from the text and speech data to generate feature text data;

[0057] Step S342: context-associate the feature text data according to the demand keyword data to generate associated feature text data;

[0058] Step S343: selecting characteristic voice segments from the navigation-related voice data based on the associated characteristic text data to generate segment voice data;

[0059] Step S344: performing voice quality analysis on the segment voice data to generate voice quality data;

[0060] Step S345: Utilize the optimization necessity evaluation formula to perform optimization necessity evaluation on the voice quality data to generate necessity evaluation data.

[0061] The present invention generates feature text data by extracting feature information from text voice data. The key features of the voice data are extracted, and the information volume and accuracy of the data are improved. The feature text data is contextually associated according to the demand keyword data to generate associated feature text data. The voice data is associated with the demand keyword, which increases the relevance of the voice data and provides more comprehensive information for subsequent analysis. Feature voice fragments are selected for navigation-related voice data based on the associated feature text data, and voice fragments related to the demand keyword are screened out, which reduces the amount of data and improves the pertinence and efficiency of subsequent analysis. Voice quality analysis is performed on the fragment voice data to generate voice quality data. Evaluation data associated with the voice quality is obtained to provide a basis for subsequent optimization necessity evaluation. The voice quality data is evaluated using the optimization necessity evaluation formula. According to the evaluation result, the optimization requirements of the voice data are determined to ensure the importance and necessity of the voice data for navigation optimization.

[0062] Preferably, the optimization necessity evaluation formula in step S344 is as follows:

[0063]

[0064] Where E is the necessity evaluation value, t1 is the start time of the evaluation, t2 is the end time of the evaluation, n is the total number of speech segments at time t, and a i is the quality score of speech segment i, b i is the length of speech segment i, c i is the speech recognition accuracy of speech segment i, d i is the environmental sound recognition accuracy of speech segment i, e i is the ambient noise level of speech segment i, f i is the incoherence value of speech segment i, α is the length of the navigation route, and h t is the congestion level of the route at time t, g is the associated value of route congestion, v is the frequency of use of the route, θ is the historical optimization frequency, e is the base of the natural logarithm, k is the complexity of the route, and ω is the deviation correction value of the necessity evaluation value.

[0065] The present invention constructs an optimization necessity evaluation formula for optimizing the necessity evaluation of voice quality data and generating necessity evaluation data. The formula fully considers the start time t1 of the evaluation, the end time t2 of the evaluation, the total number of voice segments n at time t, the quality score a of the voice segment i, and the quality score t1 of the voice segment i. i , the length b of speech segment i i , the speech recognition accuracy c of speech segment i i , the environmental sound recognition accuracy d of speech segment i i , the ambient noise level e for speech segment i i, the incoherence value f of speech segment i i , navigation route length α, route congestion level value h at time t t , the associated value g of route congestion, the frequency of route use v, the historical optimization frequency θ, the base of the natural logarithm e, the complexity of the route k, the deviation correction value ω of the necessity evaluation value and the interaction relationship between the variables constitute the following functional relationship:

[0066]

[0067] By a i Represents the overall quality of the sound clip. A higher quality score means clearer, more accurate sound. i Affects the weight of the sound clip quality score. Longer sound clips may have more information, but may also contain more useless information. The recognition accuracy of the sound clip and the accuracy of the ambient sound will directly affect the clarity and effectiveness of the sound. The ambient noise level and the incoherence value affect the purity and coherence of the sound. Logarithmic processing is performed to make the relative effects of various influencing factors more balanced. When some factors are relatively large, the logarithmic function can mitigate their effects and prevent one factor from overly affecting the overall evaluation value. The actual road condition of the navigation route is evaluated based on the length of the navigation route. The actual status of the navigation route is evaluated by the correlation value between the congestion level and the route congestion. Different travel modes have different route congestion correlation values. For example, in the city, the route congestion correlation value of walking mode should be close to zero. The negative reference indicates that the contribution of this part is to reduce the overall evaluation value. The influence of the usage frequency and historical optimization frequency of the route on the overall evaluation is considered. The higher the usage frequency and historical optimization frequency, the greater the contribution to the overall necessity evaluation value. The overall formula comprehensively evaluates the navigation voice segments over a period of time through the integral term, taking into account the dynamic variability of the user's voice segments. It reflects the impact of path complexity and length on the evaluation. When the complexity of the route is relatively high or the route length is long, Will approach 0. This means that paths with greater complexity and length will have a greater negative impact on the overall evaluation. This functional relationship can accurately and quickly evaluate the necessity evaluation value and determine whether to optimize the navigation route based on the user's voice segment information. Improve the reliability of route optimization judgment. The deviation correction value ω of the necessity evaluation value is used to adjust and correct the functional relationship to reduce the error caused by the parameter error term, thereby more accurately generating the necessity evaluation value E, and improving the accuracy and reliability of the necessity evaluation of navigation route optimization. At the same time, the deviation correction value in the formula is adjusted according to the actual situation. For example, by considering acoustic characteristics, route characteristics, etc., ω is adjusted based on these field knowledge to be applied to different voice quality data for optimization necessity evaluation, thereby improving the flexibility and applicability of the algorithm.

[0068] Preferably, step S4 comprises the following steps:

[0069] Step S41: Transmitting the path optimization data to the cloud device to generate optimization record summary data;

[0070] Step S42: clustering the optimization record summary data according to similar optimization methods to generate optimization content division data;

[0071] Step S43: performing optimization reason analysis based on the optimization content division data to generate optimization reason data;

[0072] Step S44: extracting the cause of the navigation defect based on the optimized cause data, and generating navigation defect feedback data;

[0073] Step S45: Formulate equipment optimization strategy based on navigation defect feedback data and generate equipment optimization strategy data; optimize navigation equipment based on the equipment optimization strategy data and generate navigator update data.

[0074] The present invention allows centralized storage and processing of data by transmitting path optimization data to a cloud device. It is beneficial to form a comprehensive optimization record and provide a data basis for subsequent analysis and decision-making. By clustering similar methods of optimization records, the optimization methods can be classified. It helps to identify and understand the similarities and differences between different optimization methods, providing a basis for more in-depth analysis. Focus on the specific analysis of the optimization content and determine the reasons for the optimization. By deeply analyzing the optimization content of each category, the specific reasons that require optimization can be found to provide guidance for subsequent optimization. Extract equipment defect information from the optimization reasons. By analyzing the optimization reasons, potential defects or problems of the equipment can be identified. This information helps to formulate targeted improvement plans. Based on navigation defect feedback data, formulate corresponding equipment optimization strategies. According to the formulated equipment optimization strategy, the navigation equipment is optimized accordingly. This optimization will result in the generation of new navigator update data, thereby improving the performance and function of the navigation equipment and making it more adaptable to actual needs.

[0075] The beneficial effect of the present application is that, through voice detection, the present invention enables users to issue navigation requests in a natural voice manner, thereby improving user experience and interactivity. The starting position data is generated through device positioning, and accurate starting position information is provided, which is helpful for accurately planning navigation routes. Through travel mode identification and destination information data generation, the user's travel mode and destination information are understood, providing a basis for personalized navigation services and information push. Through navigation route planning, the best path planning is generated, saving time and resources, and providing efficient navigation guidance. Through turning stop state analysis and voice segment screening, it is helpful to extract key information, reduce the amount of voice data processing, and improve processing efficiency. By intercepting the voice data to be processed, the segmented voice data related to navigation is extracted, the amount of unnecessary information is reduced, and the effectiveness and efficiency of data processing are improved. Based on the segmented voice data, the navigation demand is evaluated and the necessity of its optimization is determined, so that it can be more accurately judged whether the planned path needs to be replanned. The planned path is replanned according to the necessity evaluation data, and the path optimization data is generated, which helps to provide a more accurate and user-friendly navigation route. The path optimization data is transmitted to the cloud device and the optimization record summary data is generated, providing detailed records and data support for the improvement of the navigation system. Analyzing the causes of defects in the navigation system based on the optimization record summary data helps to discover and solve system problems and extract navigation defect feedback data. Targeted equipment optimization is performed based on the navigation defect feedback data to generate navigator update data to improve the performance and function of the navigation device. Therefore, the speech recognition navigation method based on planned path data of the present invention analyzes the user's travel status, filters the user's voice clips during navigation to reduce the workload of speech analysis, and evaluates the necessity of optimizing the filtered voice clips to determine whether the user needs to optimize the navigation path, thereby promptly handling the changing needs of the user during navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 A schematic diagram of the steps of a speech recognition navigation method based on planned path data;

[0077] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0078] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0079] Figure 4 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0080] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0081] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0082] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0083] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0084] To achieve this, please refer to Figures 1 to 4 , a speech recognition navigation method based on planned path data, the method comprising the following steps:

[0085] Step S1: performing navigation request voice detection based on a voice acquisition device to generate navigation request data; performing device positioning based on the navigation request data to generate starting position data; performing travel mode and travel destination identification based on the navigation request data to generate travel mode data and destination information data; performing map position matching based on the travel mode data and destination information data to generate navigation destination data;

[0086] Step S2: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data; performing turn and stop state analysis based on the planned path data to generate turned and stopped data; performing voice segment screening based on the turned and stopped data to generate voice data to be processed;

[0087] Step S3: intercepting the voice data to be processed to meet the optimization requirements and generating segmented voice data; evaluating the necessity of optimization based on the segmented voice data and generating necessity evaluation data; replanning the planned path data based on the necessity evaluation data and generating path optimization data;

[0088] Step S4: Transmit the path optimization data to the cloud device to generate optimization record summary data; extract the causes of navigation defects based on the optimization record summary data to generate navigation defect feedback data; optimize the navigation equipment according to the navigation defect feedback data to generate navigator update data.

[0089] The present invention uses voice detection to issue navigation requests in a natural voice mode, thereby improving user experience and convenience. The complexity of user operations is reduced. Starting position data is generated according to device positioning, providing accurate starting position information, which is helpful for accurately planning navigation routes. By identifying travel modes and generating destination information data, the user's travel mode and destination information are understood, which is helpful for personalized navigation services and information push. Path planning based on starting position data and navigation destination data is helpful for generating the best path, saving time and resources, and providing efficient navigation guidance. By analyzing the turning and stopping states and screening voice fragments, key information is helpful for extracting key information, reducing the amount of voice data processing, and improving processing efficiency. Real-time interaction with users during navigation is ensured. The real-time and accuracy of the navigation system are improved, the user experience is enhanced, and it is ensured that users can get timely guidance and feedback during navigation. By optimizing the interception of demand voice and the necessity evaluation, the voice data is made more refined and targeted, redundant information is reduced, and data quality is improved. According to the evaluation results of the voice data, the planned path is adjusted and optimized. The intelligence and personalization of the navigation system are improved, the path planning is more in line with the actual needs of users, and the accuracy and efficiency of the navigation path are optimized. The optimized record summary data is used to record the optimization process. Based on the optimized record summary data, the navigation device is optimized and the navigation data is updated. The performance and adaptability of the navigation device are improved, so that it can better adapt to the changing road conditions and user needs, and the stability and accuracy of the navigation system are improved. Therefore, the speech recognition navigation method based on the planned path data of the present invention analyzes the user's travel status, filters the user's voice clips during navigation, so as to reduce the workload of speech analysis, and evaluates the necessity of optimizing the filtered voice clips to determine whether the user needs to optimize the navigation path, so as to promptly handle the changing needs of the user during navigation.

[0090] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a speech recognition navigation method based on planned path data of the present invention. In this example, the speech recognition navigation method based on planned path data includes the following steps:

[0091] Step S1: performing navigation request voice detection based on a voice acquisition device to generate navigation request data; performing device positioning based on the navigation request data to generate starting position data; performing travel mode and travel destination identification based on the navigation request data to generate travel mode data and destination information data; performing map position matching based on the travel mode data and destination information data to generate navigation destination data;

[0092] In the embodiment of the present invention, user voice data is collected by a voice collection device. Voice preprocessing, such as noise elimination and voice signal enhancement, is performed to ensure voice quality. Speech recognition technology or natural language processing is used to identify key information in the navigation request voice. Key data such as the navigation destination and travel mode in the user request is extracted. The user's current location information is obtained using positioning technology. The data obtained by positioning is converted into starting location information to ensure accuracy and availability. By analyzing the navigation request data, the user's travel mode and destination information are identified. Combined with map data, the travel mode and destination information are matched to generate the final navigation destination data.

[0093] Step S2: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data; performing turn and stop state analysis based on the planned path data to generate turned and stopped data; performing voice segment screening based on the turned and stopped data to generate voice data to be processed;

[0094] In an embodiment of the present invention, navigation route planning is performed using starting position data and navigation destination data, and factors such as road networks and traffic conditions are considered to generate planned path data. A navigation algorithm (such as Dijkstra algorithm or A* algorithm) is used to calculate the optimal path to ensure the accuracy and effectiveness of the path. Based on the planned path data, the turning and stopping states in the path are analyzed, turning points and stopping points are identified, and turning and stopping data are generated. For example, the identification and recording of intersections, crossovers, parking areas, etc. Based on the turning and stopping data, the voice data during the stop period is screened, and the voice segments related to the turn are extracted to generate voice data to be processed.

[0095] Step S3: intercepting the voice data to be processed to meet the optimization requirements and generating segmented voice data; evaluating the necessity of optimization based on the segmented voice data and generating necessity evaluation data; replanning the planned path data based on the necessity evaluation data and generating path optimization data;

[0096] In an embodiment of the present invention, noise-reduced voice data is generated by preprocessing the voice data to be processed, including noise reduction and filtering. On the basis of the noise-reduced voice data, irrelevant voices are screened out, and navigation-related voice data is extracted. The navigation-related voice data is converted into text to generate text voice data. Keyword information of navigation requirements is identified and extracted from the text voice data. According to the required keyword data, voice segments related to the navigation requirements are selected to generate segment voice data. Based on the segment voice data, an optimization necessity assessment is performed to determine high-value data that is helpful for navigation optimization, and to generate analysis data to be optimized. Based on the high-necessity analysis data to be optimized, a path optimization strategy is formulated to generate path optimization strategy data. According to the path optimization strategy data, the planned path data is replanned to generate path optimization data.

[0097] Step S4: Transmit the path optimization data to the cloud device to generate optimization record summary data; extract the causes of navigation defects based on the optimization record summary data to generate navigation defect feedback data; optimize the navigation equipment according to the navigation defect feedback data to generate navigator update data.

[0098] In an embodiment of the present invention, the path optimization data is uploaded to the cloud through a network or other transmission methods for storage and analysis to generate optimization record summary data. According to the optimization records, the optimization records are classified using clustering or similar methods to generate optimization content division data for analysis and identification of optimization reasons. The optimization content division data is analyzed to identify the specific reasons for the optimization, extract navigation defect data, and guide subsequent navigator optimization. Based on the navigation defect feedback data, the navigator equipment is optimized and updated to repair defects and improve navigation performance, and navigator update data is generated.

[0099] Preferably, step S1 comprises the following steps:

[0100] Step S11: collecting user voice based on the voice collection device to generate user voice data;

[0101] Step S12: performing navigation request voice detection on the user voice data to generate navigation request data;

[0102] Step S13: positioning the device based on the navigation request data to generate starting position data;

[0103] Step S14: performing navigation keyword recognition according to the navigation request data to generate navigation key information data, including travel mode data and destination information data;

[0104] Step S15: locating the destination according to the navigation key information data and generating navigation destination data;

[0105] Step S16: Associating the starting position data with the navigation destination data to generate navigation requirement data.

[0106] By collecting user voice data, the present invention can better understand the user's voice characteristics and accent, and provide a basis for subsequent voice recognition and interaction. The accuracy and personalization of voice recognition are improved, the interaction effect between the system and the user is improved, and the system is more user-friendly. By performing navigation request voice detection on user voice data, the user's navigation request can be accurately captured. The navigation system's accurate understanding of the user's intention is improved, ensuring the correctness and pertinence of the subsequent navigation process. Through device positioning, the starting point information of the navigation is provided. Identify the key information in the navigation request, including the travel mode and destination information. In order to better understand the user's travel needs and destination. Through destination positioning, the system can obtain the location data of the navigation destination. Provide the end point information of the navigation, providing key data support for the path planning and navigation process. Data association derives navigation demand data, including key information such as the starting point and the end point. Provide complete navigation demand information, which is conducive to the subsequent navigation system to perform path planning and optimization, and ensure the accuracy and efficiency of the navigation process.

[0107] In the embodiment of the present invention, the user voice data is recorded and collected by using the voice collection device pre-loaded in the navigator to ensure the voice quality and clarity. The collected voice data is stored in a safe and reliable media or server, and is properly managed and backed up. A suitable voice detection algorithm or voice recognition model is selected, such as a voice recognition model based on deep learning (such as CNN, RNN, Transformer, etc.). The user voice data is preprocessed, including denoising, audio format conversion, etc., to improve the accuracy of voice recognition. The voice detection algorithm is used to identify and extract voice fragments or keywords related to the navigation request to generate navigation request data. Or the user is guided to use a preset sentence to obtain navigation request data, such as a preset sentence such as "Xiaohang, I want to go to the mall." Select a suitable positioning method using technologies such as GPS, base station positioning or Wi-Fi positioning. The user's starting position information is obtained through the selected positioning technology to generate starting position data. The navigation request data is segmented, annotated, semantically analyzed and keyword extracted to obtain initial keyword data. The keywords are classified using semantic analysis and associated data to identify key information about the travel mode. The location information of the destination is obtained by identifying the location entity and the geographic location mapping. Evaluate the accuracy of the recognition results and generate detailed information data of the destination, such as location, name, etc., based on the evaluation results. Obtain the user's navigation history data and preset common address data, as well as related travel tools and distance range data. Filter and pre-process the destination information data according to the recognition accuracy threshold to generate destination keyword data. Perform map location matching based on the current travel range data, identify and locate the destination information, and generate navigation destination data.

[0108] Preferably, step S14 comprises the following steps:

[0109] Step S141: performing word segmentation annotation on the navigation request data to generate word segmentation request data; performing keyword extraction on the word segmentation request data to generate initial keyword data;

[0110] Step S142: semantically analyzing the navigation request data to generate word association data; using the word association data to classify the initial keyword data to generate classified keyword data;

[0111] Step S143: Identify the travel mode based on the classified keyword data and generate travel mode data;

[0112] Step S144: performing location entity recognition based on the classified keyword data to generate location recognition result data; performing geographic location mapping based on the location recognition result data to generate mapping result data;

[0113] Step S145: performing recognition accuracy evaluation based on the mapping result data to generate recognition accuracy data;

[0114] Step S146: Perform data concatenation processing on the location recognition result data and the recognition accuracy data to generate destination information data.

[0115] The present invention extracts key information in the request by performing word segmentation and keyword extraction on navigation request data. The understanding and accuracy of the user's navigation intention are improved, and a basis for subsequent semantic analysis and keyword extraction is provided. Semantic analysis and keyword classification are performed to process the word association in the request data. The deep understanding of the navigation request data is improved, the classification and classification of keywords are optimized, and the intelligence and personalization of the navigation system are enhanced. The travel mode is identified based on the classified keyword data to identify the travel mode selected by the user. The user's travel mode is determined, which is conducive to the system to more accurately plan a navigation path suitable for the travel mode for the user. The location information in the navigation request is identified and the geographical location mapping is performed. Accurate location information is provided, and the accuracy and effectiveness of navigation are enhanced. The accuracy of the recognition result data is evaluated to evaluate the accuracy of the location recognition. The accuracy of the location recognition result is ensured, and the accurate interpretation and response capability of the navigation system to user needs is improved. Accurate destination information is generated by combining the location recognition and accuracy evaluation data. Complete and accurate navigation destination information is provided, ensuring the accuracy of navigation path planning and the optimization of user experience.

[0116] In an embodiment of the present invention, by using natural language processing (NLP) technology, such as word segmentation tools (such as jieba, NLTK, etc.), the navigation request data is segmented and annotated to divide it into a vocabulary sequence. Based on the word segmentation result, keywords related to navigation are extracted from the word segmentation request data by a keyword extraction algorithm (such as TF-IDF, TextRank, etc.) to generate initial keyword data. Semantic analysis methods in NLP, such as word vector models (Word2Vec, GloVe), BERT, etc., are used to generate correlation data between words. Based on the semantic association data, the initial keyword data is classified by clustering, classification, etc. to form classified keyword data, and keywords with higher correlation are classified and summarized. Based on the classified keyword data, rule matching or machine learning models are used to identify information representing travel modes in keywords, such as "walking", "bicycle", "driving", etc., to generate travel mode data. Using NLP technology or entity recognition models, entity information representing locations, such as place names, addresses, etc., is identified from the classified keyword data to generate location recognition result data. The identified location information is matched with geographic location data or map services, mapped to specific geographic location information, and mapping result data is generated. Verify and evaluate the mapping result data, use indicators such as relevance and accuracy to evaluate the accuracy of location recognition, and generate recognition accuracy data. Combine location recognition results and recognition accuracy data, connect and filter data, and extract highly accurate destination information data for specifying navigation destinations.

[0117] Preferably, step S15 comprises the following steps:

[0118] Step S151: Obtain navigation history data and preset frequently used address data;

[0119] Step S152: processing the destination information data based on a preset recognition accuracy threshold, and generating destination keyword data when the destination information data is greater than the preset recognition accuracy threshold;

[0120] Step S153: when the destination information data is less than or equal to the preset recognition accuracy threshold, fuzzy address matching is performed based on the preset common address data to generate destination keyword data;

[0121] Step S154: dividing the navigation history data into travel tools to generate tool division history data; formulating the travel distance range according to the tool division history data to generate travel range data;

[0122] Step S155: Preset the travel range of the current travel based on the travel mode data to generate current travel range data; match the map position of the destination keyword data based on the current travel range data to generate navigation destination data.

[0123] The present invention obtains the user's navigation history data and preset common address data. It provides an important data source for the navigation system to analyze, which is helpful to understand the user's common destinations and navigation preferences. The destination information data is processed for accuracy based on a preset threshold. The reliability of the destination information is ensured, and for data above the threshold, the accuracy and reliability of navigation are improved. When the destination information is insufficient, fuzzy matching is performed using the common address data. The destination information is supplemented, the system's ability to understand the user's intentions is enhanced, and the integrity and accuracy of navigation are improved. The navigation history data is divided into tools and the travel distance range is determined. The travel range data is preset for this trip based on the travel mode data, and the current travel range data is generated, which provides a more accurate reference range for the generation of subsequent destination data and improves the adaptability and personalization of the navigation path. The destination map position is matched based on the current travel range data. Accurate destination data is provided to ensure that the navigation system accurately interprets and responds to user needs, and enhances the effectiveness of navigation path planning and user experience.

[0124] In an embodiment of the present invention, relevant data including the user's past navigation records, frequently visited places, and frequently used destinations are extracted and sorted from existing navigation records or user navigation history. The user can set some frequently used addresses in advance, such as home, office, frequently visited store, etc., and save them in the system in advance. The accuracy threshold of identifying the destination information is set to distinguish the accuracy of identification. The threshold can be preset by the expected matching result. The higher the accuracy threshold, the fewer matching results there are, but the higher the threshold, the correct matching results may be missed. The destination information data provided by the user is processed. If the recognition accuracy of the destination information exceeds the preset threshold, it is determined to be credible destination information. When the recognition accuracy of the destination information reaches or exceeds the preset threshold, key information such as address, place name, etc. is extracted from it to generate destination keyword data. These keyword data can be summaries or identifiers of accurately identified destination information. If its accuracy is lower than the threshold, the fuzzy matching process is entered. Based on the preset frequently used address data, fuzzy matching is performed on the destination information that fails to meet the accuracy requirements, and fuzzy place entity data is identified and generated. Use the fuzzy location entity data to match the entity identifier with the preset common address data. For example, if the mode location entity data is "home", it is associated with the data in the preset common address data. If the "home" identifier exists, it is considered a successful match, and the address corresponding to "home" is used as the destination keyword data; if the match fails, the address consultation process is executed. In the case of a match failure, address consultation is performed and address consultation result data is obtained. The address consultation result data is associated with the common address data, and the common address data is corrected to ensure the accuracy and completeness of the data. Based on the corrected common address data, the destination address information is extracted, and the destination keyword data is finally generated. The travel tools in different navigation records are identified and divided, which can be walking, bicycles, cars, etc. This process can be completed by recording the transportation tools used during navigation. Tool division history data is formed to record the use of each tool in the navigation history. A common travel distance range can be preset, such as 1 to 3 kilometers for walking and 5 to 10 kilometers for urban public transportation. Then, based on the tool division history data, the typical travel distance range of different travel tools is corrected, and the appropriate range of various travel tools for different users is formulated. For example, the walking distance is corrected to 1 to 5 kilometers, and the urban public transportation is corrected to 8 to 15 kilometers, etc., so as to generate travel range data. The travel range data is preset based on the travel mode data to generate the current travel range data. Based on the current travel range data, the destination keyword data is compared and screened with the map location to ensure that the destination is within the appropriate travel range.Based on the results of the map location matching, determine the destination data that meets the scope of the travel tool. When the matching results only contain one result, the navigation destination data is directly generated; when the matching results only contain one result, accurate address Q&A confirmation is required until the navigation destination data is generated.

[0125] Preferably, step S153 includes the following steps:

[0126] Step S1531: when the destination information data is less than or equal to a preset recognition accuracy threshold, generating fuzzy place entity data;

[0127] Step S1532: performing entity identification matching on the fuzzy location entity data based on the preset common address data to generate matching result data; when the matching result data is a successful match, generating destination keyword data;

[0128] Step S1533: when the matching result data indicates a matching failure, accurate address consultation is performed based on the mapping result data to generate address consultation result data;

[0129] Step S1534: performing location data association on the fuzzy location entity data according to the address consultation result data, and supplementing and correcting the preset common address data to generate common address data;

[0130] Step S1535: extract the destination address based on the commonly used address data and generate destination keyword data.

[0131] The present invention generates fuzzy place entity data. Incomplete destination information is supplemented to provide basic data for subsequent matching and query. Fuzzy place entities are matched and identified based on preset common address data. The accuracy of fuzzy place entities is improved. When the match is successful, more reliable destination keyword data is generated, and the accuracy of navigation is enhanced. When the match fails, precise address consultation is performed and consultation result data is generated. In the event of a matching failure, more accurate address information is obtained through consultation, which improves the completeness and accuracy of the destination information. The fuzzy place entity and common address data are associated and corrected according to the consultation results. The accuracy and completeness are improved, providing a more reliable data basis for the navigation system. Destination information is extracted based on the corrected common address data. More reliable destination keyword data is generated to ensure the accuracy of navigation path planning and the optimization of user experience.

[0132] In an embodiment of the present invention, the collected destination information data is evaluated. If its accuracy is less than or equal to a preset recognition accuracy threshold, it is identified as fuzzy data. The destination information data with accuracy lower than the threshold is marked as fuzzy place entity data, which may include fuzzy place names, incomplete addresses, etc. The fuzzy place entity data is matched using preset common address data. If the match is successful, it is considered to be valid destination information. The successfully matched fuzzy place entity data is matched with the common address data to generate matching result data. When the match is successful, key information is extracted from the matching result to generate destination keyword data. When the fuzzy place entity data fails to match, the precise address consultation of the mapping result data is performed. Based on the mapping result data, the precise address consultation result is obtained, and the corresponding address consultation result data is generated. The address consultation result data is used to associate with the fuzzy place entity data, and the precise address information is associated with the fuzzy place entity data to improve the accuracy of the data. The preset common address data is corrected and supplemented according to the associated precise address information. It may involve updating and revising information such as addresses, names, and landmarks. Based on the corrected common address data, the final destination address information is extracted. Extract key information from the destination address information, such as place name, address details, etc., to generate final destination keyword data.

[0133] Preferably, step S2 comprises the following steps:

[0134] Step S21: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data;

[0135] Step S22: performing path navigation according to the planned path data, and performing stay status analysis to generate stay status data;

[0136] Step S23: using a voice collection device and performing real-time voice collection based on the stay status data to generate voice data during the stay period;

[0137] Step S24: acquiring the movement direction based on the stay state data to generate stay direction data; extracting the turning direction from the stay direction data using the planned path data to generate turned stay data;

[0138] Step S25: filtering the voice data during the stop period based on the turn stop data to generate voice data to be processed.

[0139] The present invention generates the best path planning through the starting position and navigation destination data, provides the most optimized navigation solution, and saves time and resources. According to the starting position and destination data, a personalized navigation route can be provided, taking into account factors such as user preferences and traffic conditions. Through path navigation, real-time navigation guidance is provided to users, and through the stop state analysis, the user's stop situation during the navigation process is understood, the path planning is optimized, and the accuracy and practicality of the navigation route are ensured. The data analysis of the stop state is helpful to improve the navigation system and provide a more intelligent and real-time navigation service. Through the voice acquisition device, voice data is collected during the user's stay to understand the user's demand changes, feedback or other relevant information during the navigation process, which is helpful to improve the responsiveness of the navigation system and provide services that are closer to user needs. The user's movement direction is obtained through the stop state data, which is helpful to understand the relationship between the user's current position and the navigation path, and provide more accurate navigation guidance. Combined with the planning path data, the relevant information of the turn stop is extracted, which can provide basic information for the turn reminder, intersection guidance, etc. of the navigation system, and improve the real-time and accuracy of navigation. The voice data during the stop period is screened based on the turn stop data, which is helpful to extract the voice fragments related to the turn and reduce the amount of unnecessary voice information. The screened voice data to be processed is more likely to contain information about the reasons for turning and stopping, which improves the quality and relevance of the voice data and facilitates subsequent voice recognition and analysis.

[0140] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0141] Step S21: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data;

[0142] In the embodiment of the present invention, the optimal or most suitable navigation route is calculated based on the map data, the starting position data and the navigation destination data by using a suitable navigation algorithm (such as Dijkstra, A*, etc.). The navigation route calculated by the algorithm is converted into a data format, including route nodes, directions, distances, estimated time and other information.

[0143] Step S22: performing path navigation according to the planned path data, and performing stay status analysis to generate stay status data;

[0144] In the embodiment of the present invention, a navigation engine or algorithm (such as a GPS navigation system or a map application) is used to guide the user to the destination according to the planned path data. Navigation instructions, turn prompts, etc. are provided through map data and real-time location information. Based on the planned path data and real-time location information, the status of the user's current location is analyzed to identify the stop point or the area where the stay time is longer. Parking areas, traffic congestion points or other possible stop states are identified, relevant information is recorded and stop state data is generated.

[0145] Step S23: using a voice collection device and performing real-time voice collection based on the stay status data to generate voice data during the stay period;

[0146] In the embodiment of the present invention, by using a special voice collection device (such as a microphone, a voice recognizer, etc.), voice collection is started when the stay state is recognized. Voice collection is performed in real time based on the stay state data to capture voice information during the stay period.

[0147] Step S24: acquiring the movement direction based on the stay state data to generate stay direction data; extracting the turning direction from the stay direction data using the planned path data to generate turned stay data;

[0148] In the embodiment of the present invention, the direction of movement during the stay is calculated by using the position information or motion trajectory in the stay state data. The direction of movement during the stay is inferred by the change in position or direction, or by a gyroscope device, and the stay direction data is generated. Based on the navigation route information in the planned path data, the stay direction data is analyzed to determine whether a turn has occurred. It is identified whether the stay period occurs after a turn, and the type, angle or other relevant information of the turn is recorded to generate the turned stay data.

[0149] Step S25: filtering the voice data during the stop period based on the turn stop data to generate voice data to be processed.

[0150] In the embodiment of the present invention, the voice data during the stop period is screened based on the stop data, and the voice segments related to the turn are extracted. According to the time, location or other characteristic information of the turn, the voice segments related to the turn are selected to generate the voice data to be processed.

[0151] Preferably, step S3 comprises the following steps:

[0152] Step S31: preprocessing the speech data to be processed to generate noise-reduced speech data;

[0153] Step S32: filtering out irrelevant voices from the noise reduction voice data to generate navigation-related voice data; converting the navigation-related voice data into text to generate text voice data;

[0154] Step S33: performing demand keyword recognition on the text and voice data to generate demand keyword data;

[0155] Step S34: selecting voice segments associated with the text voice data according to the required keyword data to generate segment voice data; performing optimization necessity evaluation according to the segment voice data to generate necessity evaluation data;

[0156] Step S35: screening the text and speech data for high necessity data according to the necessity evaluation data to generate data to be optimized and analyzed;

[0157] Step S36: Formulate an optimization strategy based on the analyzed data to be optimized to generate path optimization strategy data; replan the planned path data according to the path optimization strategy data to generate path optimization data.

[0158] The present invention performs pre-processing and noise reduction on the voice data to be processed. The quality of the voice data is improved and the influence of noise on voice recognition is reduced. The irrelevant voices are screened out and the processed voice data is converted into text data. The voice information related to navigation is extracted, and the text format conversion is helpful for the subsequent key information extraction and semantic analysis. The text voice data is identified with demand keywords. The navigation demand keywords proposed by the user in the voice are determined, which provides important navigation guidance for subsequent processing. The necessity of navigation optimization is evaluated to ensure the effectiveness and rationality of navigation path optimization. The text voice data is screened according to the necessity evaluation data to screen out the data that most needs to be optimized, thereby improving the pertinence and effectiveness of path optimization. According to the screened analysis data to be optimized, research and analysis are carried out, targeted optimization strategies are formulated, and path optimization strategy data are generated. The original planned path data is re-planned according to the path optimization strategy data to generate path optimization data. The optimization and effectiveness of the navigation path are ensured, and the accuracy of navigation and user experience are improved.

[0159] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0160] Step S31: preprocessing the speech data to be processed to generate noise-reduced speech data;

[0161] In the embodiment of the present invention, the speech data to be processed is preprocessed by applying audio processing techniques, such as filtering, downsampling, etc. This includes operations such as removing background noise, eliminating interference, and smoothing audio to reduce interference in subsequent processing. Noise reduction algorithms or software, such as Fourier transform, wavelet transform, or machine learning model, are used to perform noise reduction processing on the preprocessed speech data to improve the quality of the speech signal.

[0162] Step S32: filtering out irrelevant voices from the noise reduction voice data to generate navigation-related voice data; converting the navigation-related voice data into text to generate text voice data;

[0163] In the embodiment of the present invention, by using technologies such as speech recognition and acoustic feature analysis, voice segments irrelevant to navigation, such as background noise and meaningless conversations, are screened out to generate voice data related to navigation. The screened navigation-related voice data is subjected to speech-to-text processing, and the speech is converted into text data using speech recognition technology to facilitate subsequent keyword recognition or analysis.

[0164] Step S33: performing demand keyword recognition on the text and voice data to generate demand keyword data;

[0165] In the embodiment of the present invention, keywords or phrases related to navigation requirements are extracted from text and speech data by applying natural language processing technology, such as text segmentation, keyword extraction algorithm, etc. Keywords expressing navigation requirements or instructions are identified and extracted using machine learning or rule matching methods to generate demand keyword data.

[0166] Step S34: selecting voice segments associated with the text voice data according to the required keyword data to generate segment voice data; performing optimization necessity evaluation according to the segment voice data to generate necessity evaluation data;

[0167] In the embodiment of the present invention, voice segments related to the demand keywords are screened out, and the segment voice data associated with the navigation demand can be generated by using technologies such as timestamp, semantic matching or keyword matching. The selected voice data associated with navigation is further analyzed for voice quality to determine whether the associated text information is reliable, and to generate optimization necessity evaluation data for the associated segments.

[0168] Step S35: screening the text and speech data for high necessity data according to the necessity evaluation data to generate data to be optimized and analyzed;

[0169] In the embodiment of the present invention, by using the evaluation value in the necessity evaluation data as a basis, the speech text data with high optimization necessity is selected to generate the analysis data to be optimized. The data that needs to be optimized the most can be selected by setting a threshold or using a sorting algorithm. The speech text data with high priority is marked or classified to prepare for analysis of path optimization.

[0170] Step S36: Formulate an optimization strategy based on the analyzed data to be optimized to generate path optimization strategy data; replan the planned path data according to the path optimization strategy data to generate path optimization data.

[0171] In an embodiment of the present invention, problems or opportunities for improvement in navigation are identified by conducting an in-depth analysis of the data to be optimized and analyzed. Areas that need to be improved in path planning are determined, such as traffic congestion, inaccurate routes, etc. Optimization strategies are formulated based on the data analysis results, and targeted path improvement plans are determined. Possible optimization strategies include: adjusting the route algorithm, considering real-time traffic information, selecting better roads, avoiding congested areas, etc. The formulated path optimization strategy is converted into a specific path planning adjustment plan. The original planned path data is adjusted and replanned using the optimized algorithm or rules. Based on the adjusted path planning, path optimization data is generated, including updated navigation path information. Ensure that the optimization data can be effectively recognized and applied by the navigation system.

[0172] Preferably, step S34 includes the following steps:

[0173] Step S341: extracting feature information from the text and speech data to generate feature text data;

[0174] Step S342: context-associate the feature text data according to the demand keyword data to generate associated feature text data;

[0175] Step S343: selecting characteristic voice segments from the navigation-related voice data based on the associated characteristic text data to generate segment voice data;

[0176] Step S344: performing voice quality analysis on the segment voice data to generate voice quality data;

[0177] Step S345: Utilize the optimization necessity evaluation formula to perform optimization necessity evaluation on the voice quality data to generate necessity evaluation data.

[0178] The present invention generates feature text data by extracting feature information from text voice data. The key features of the voice data are extracted, and the information volume and accuracy of the data are improved. The feature text data is contextually associated according to the demand keyword data to generate associated feature text data. The voice data is associated with the demand keyword, which increases the relevance of the voice data and provides more comprehensive information for subsequent analysis. Feature voice fragments are selected for navigation-related voice data based on the associated feature text data, and voice fragments related to the demand keyword are screened out, which reduces the amount of data and improves the pertinence and efficiency of subsequent analysis. Voice quality analysis is performed on the fragment voice data to generate voice quality data. Evaluation data associated with the voice quality is obtained to provide a basis for subsequent optimization necessity evaluation. The voice quality data is evaluated using the optimization necessity evaluation formula. According to the evaluation result, the optimization requirements of the voice data are determined to ensure the importance and necessity of the voice data for navigation optimization.

[0179] In the embodiment of the present invention, by using natural language processing techniques such as word segmentation and part-of-speech tagging of text content, feature text data in the speech text is extracted, and the feature text data is associated with the keyword based on the demand keyword data. That is, a specific keyword or phrase is associated with a specific acoustic or semantic feature. Generate associated feature text data, and associate the voice segment related to the navigation demand with its key information. According to the associated feature text data, select the feature voice segment related to the specific navigation demand. Algorithms or rules can be used to select voice segments closely related to the demand, which may involve considerations such as semantic matching, voice quality, and duration. Use acoustic analysis techniques such as signal-to-noise ratio, spectral smoothness, distortion, etc. to evaluate the quality of the segment voice data. Extract features of the relevant segment voice data, such as clarity, speech speed, pronunciation accuracy, etc., to quantify the voice quality. Take the voice quality data as input, and apply a preset optimization necessity evaluation formula, which fully considers the parameters such as the evaluation time interval, the total number of voice segments, and the voice segment quality score, so as to accurately and quickly evaluate the necessity evaluation value and determine whether to optimize the navigation route according to the user's voice segment information. That is, consider the impact of voice quality in navigation optimization needs. Use the calculated optimization necessity evaluation value to quantify the impact of voice quality on route optimization. Or through machine learning models: build models based on historical data and feature engineering to predict the impact of voice quality on route optimization; statistical analysis and regression methods: perform statistical analysis to find the correlation between voice quality and route optimization, etc. to evaluate the necessity of optimization and generate necessity evaluation data.

[0180] Preferably, the optimization necessity evaluation formula in step S344 is as follows:

[0181]

[0182] Where E is the necessity evaluation value, t1 is the start time of the evaluation, t2 is the end time of the evaluation, n is the total number of speech segments at time t, and a i is the quality score of speech segment i, b i is the length of speech segment i, c i is the speech recognition accuracy of speech segment i, d i is the environmental sound recognition accuracy of speech segment i, e i is the ambient noise level of speech segment i, f i is the incoherence value of speech segment i, α is the length of the navigation route, and h t is the congestion level of the route at time t, g is the associated value of route congestion, v is the frequency of use of the route, θ is the historical optimization frequency, e is the base of the natural logarithm, k is the complexity of the route, and ω is the deviation correction value of the necessity evaluation value.

[0183] The present invention constructs an optimization necessity evaluation formula for optimizing the necessity evaluation of voice quality data and generating necessity evaluation data. The formula fully considers the start time t1 of the evaluation, the end time t2 of the evaluation, the total number of voice segments n at time t, the quality score a of the voice segment i, and the quality score t1 of the voice segment i. i , the length b of speech segment i i , the speech recognition accuracy c of speech segment i i , the environmental sound recognition accuracy d of speech segment i i , the ambient noise level e for speech segment i i , the incoherence value f of speech segment i i , navigation route length α, route congestion level value h at time t t , the associated value g of route congestion, the frequency of route use v, the historical optimization frequency θ, the base of the natural logarithm e, the complexity of the route k, the deviation correction value ω of the necessity evaluation value and the interaction relationship between the variables constitute the following functional relationship:

[0184]

[0185] By a i Represents the overall quality of the sound clip. A higher quality score means clearer, more accurate sound. i Affects the weight of the sound clip quality score. Longer sound clips may have more information, but may also contain more useless information. The recognition accuracy of the sound clip and the accuracy of the ambient sound will directly affect the clarity and effectiveness of the sound. The ambient noise level and the incoherence value affect the purity and coherence of the sound. Logarithmic processing is performed to make the relative effects of various influencing factors more balanced. When some factors are relatively large, the logarithmic function can mitigate their effects and prevent one factor from overly affecting the overall evaluation value. The actual road condition of the navigation route is evaluated based on the length of the navigation route. The actual status of the navigation route is evaluated by the correlation value between the congestion level and the route congestion. Different travel modes have different route congestion correlation values. For example, in the city, the route congestion correlation value of walking mode should be close to zero. The negative reference indicates that the contribution of this part is to reduce the overall evaluation value. The influence of the usage frequency and historical optimization frequency of the route on the overall evaluation is considered. The higher the usage frequency and historical optimization frequency, the greater the contribution to the overall necessity evaluation value. The overall formula comprehensively evaluates the navigation voice segments over a period of time through the integral term, taking into account the dynamic variability of the user's voice segments. It reflects the impact of path complexity and length on the evaluation. When the complexity of the route is relatively high or the route length is long, Will approach 0. This means that paths with greater complexity and length will have a greater negative impact on the overall evaluation. This functional relationship can accurately and quickly evaluate the necessity evaluation value and determine whether to optimize the navigation route based on the user's voice segment information. Improve the reliability of route optimization judgment. The deviation correction value ω of the necessity evaluation value is used to adjust and correct the functional relationship to reduce the error caused by the parameter error term, thereby more accurately generating the necessity evaluation value E, and improving the accuracy and reliability of the necessity evaluation of navigation route optimization. At the same time, the deviation correction value in the formula is adjusted according to the actual situation. For example, by considering acoustic characteristics, route characteristics, etc., ω is adjusted based on these field knowledge to be applied to different voice quality data for optimization necessity evaluation, thereby improving the flexibility and applicability of the algorithm.

[0186] Preferably, step S4 comprises the following steps:

[0187] Step S41: Transmitting the path optimization data to the cloud device to generate optimization record summary data;

[0188] Step S42: clustering the optimization record summary data according to similar optimization methods to generate optimization content division data;

[0189] Step S43: performing optimization reason analysis based on the optimization content division data to generate optimization reason data;

[0190] Step S44: extracting the cause of the navigation defect based on the optimized cause data, and generating navigation defect feedback data;

[0191] Step S45: Formulate equipment optimization strategy based on navigation defect feedback data and generate equipment optimization strategy data; optimize navigation equipment based on the equipment optimization strategy data and generate navigator update data.

[0192] The present invention allows centralized storage and processing of data by transmitting path optimization data to a cloud device. It is beneficial to form a comprehensive optimization record and provide a data basis for subsequent analysis and decision-making. By clustering similar methods of optimization records, the optimization methods can be classified. It helps to identify and understand the similarities and differences between different optimization methods, providing a basis for more in-depth analysis. Focus on the specific analysis of the optimization content and determine the reasons for the optimization. By deeply analyzing the optimization content of each category, the specific reasons that require optimization can be found to provide guidance for subsequent optimization. Extract equipment defect information from the optimization reasons. By analyzing the optimization reasons, potential defects or problems of the equipment can be identified. This information helps to formulate targeted improvement plans. Based on navigation defect feedback data, formulate corresponding equipment optimization strategies. According to the formulated equipment optimization strategy, the navigation equipment is optimized accordingly. This optimization will result in the generation of new navigator update data, thereby improving the performance and function of the navigation equipment and making it more adaptable to actual needs.

[0193] As an example of the present invention, refer to Figure 4 As shown, in this example, step S4 includes:

[0194] Step S41: Transmitting the path optimization data to the cloud device to generate optimization record summary data;

[0195] In the embodiment of the present invention, the integrity and security of the data are ensured by transmitting the path optimization data to the cloud server through a secure communication protocol. The collected path optimization data is integrated and summarized in the cloud, and a database or data warehouse can be used to summarize the data for subsequent analysis and processing.

[0196] Step S42: clustering the optimization record summary data according to similar optimization methods to generate optimization content division data;

[0197] In the embodiment of the present invention, key features such as path optimization method, optimization parameters, effect evaluation and other information are extracted from the optimization record summary data. The records are grouped using a similarity measurement algorithm (such as a clustering algorithm), and similar optimization records are usually clustered into the same category using techniques such as K-means and hierarchical clustering. The clustering results are analyzed to identify the categories of different optimization modes or methods, and to understand the distribution and performance of similar optimization methods.

[0198] Step S43: performing optimization reason analysis based on the optimization content division data to generate optimization reason data;

[0199] In the embodiment of the present invention, by preparing clustered or classified optimization content partition data, including other data related thereto (such as effect evaluation data), an analysis method is selected, such as root cause analysis, flow chart, causal analysis or statistical method, to find out the root cause of the optimization demand. The data is analyzed to identify the specific reasons for the path optimization demand, and the reasons are sorted and recorded.

[0200] Step S44: extracting the cause of the navigation defect based on the optimized cause data, and generating navigation defect feedback data;

[0201] In the embodiment of the present invention, by optimizing the cause data for analysis, an in-depth analysis is performed to identify defects or problems in the navigation system. Possible defect types are identified, such as navigation errors, inaccurate information, path deviations, etc. Based on the identified defects, the specific causes of these defects are further extracted. Possible causes are found from the data, such as data source problems, algorithm logic, real-time updates, etc.

[0202] Step S45: Formulate equipment optimization strategy based on navigation defect feedback data and generate equipment optimization strategy data; optimize navigation equipment based on the equipment optimization strategy data and generate navigator update data.

[0203] In the embodiment of the present invention, a specific strategy for device optimization is formulated through navigation defect feedback data. A solution to each defect is determined, which may include improving the algorithm, adding data sources, optimizing the user interface, etc. The navigation device is actually improved or updated according to the formulated optimization strategy. Improvement solutions for defects are developed, tested, and deployed to ensure that the defects are resolved after the navigation device is updated.

[0204] The beneficial effect of the present application is that, through voice detection, the present invention enables users to issue navigation requests in a natural voice manner, thereby improving user experience and interactivity. The starting position data is generated through device positioning, and accurate starting position information is provided, which is helpful for accurately planning navigation routes. Through travel mode identification and destination information data generation, the user's travel mode and destination information are understood, providing a basis for personalized navigation services and information push. Through navigation route planning, the best path planning is generated, saving time and resources, and providing efficient navigation guidance. Through turning stop state analysis and voice segment screening, it is helpful to extract key information, reduce the amount of voice data processing, and improve processing efficiency. By intercepting the voice data to be processed, the segmented voice data related to navigation is extracted, the amount of unnecessary information is reduced, and the effectiveness and efficiency of data processing are improved. Based on the segmented voice data, the navigation demand is evaluated and the necessity of its optimization is determined, so that it can be more accurately judged whether the planned path needs to be replanned. The planned path is replanned according to the necessity evaluation data, and the path optimization data is generated, which helps to provide a more accurate and user-friendly navigation route. The path optimization data is transmitted to the cloud device and the optimization record summary data is generated, providing detailed records and data support for the improvement of the navigation system. Analyzing the causes of defects in the navigation system based on the optimization record summary data helps to discover and solve system problems and extract navigation defect feedback data. Targeted equipment optimization is performed based on the navigation defect feedback data to generate navigator update data to improve the performance and function of the navigation device. Therefore, the speech recognition navigation method based on planned path data of the present invention analyzes the user's travel status, filters the user's voice clips during navigation to reduce the workload of speech analysis, and evaluates the necessity of optimizing the filtered voice clips to determine whether the user needs to optimize the navigation path, thereby promptly handling the changing needs of the user during navigation.

[0205] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0206] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A speech recognition navigation method based on planned path data, characterized in that: The following steps are involved: Step S1: performing navigation request voice detection based on a voice acquisition device to generate navigation request data; Position the device according to the navigation request data and generate the starting position data; Identify the travel mode and travel destination according to the navigation request data, thereby generating travel mode data and destination information data; Performing map location matching based on travel mode data and destination information data to generate navigation destination data; Step S2: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data; Perform turning and stopping status analysis based on the planned path data to generate turning and stopping data; Screening voice segments based on the turning and stopping data to generate voice data to be processed; Step S3: intercepting the voice data to be processed to meet the optimization requirements and generating segment voice data; evaluating the necessity of optimization based on the segment voice data and generating necessity evaluation data; replanning the planned path data based on the necessity evaluation data and generating path optimization data; Step S3 includes the following steps: Step S31: preprocessing the speech data to be processed to generate noise-reduced speech data; Step S32: filtering out irrelevant voices from the noise reduction voice data to generate navigation-related voice data; converting the navigation-related voice data into text to generate text voice data; Step S33: performing demand keyword recognition on the text and voice data to generate demand keyword data; Step S34: selecting voice segments associated with the text voice data according to the required keyword data to generate segment voice data; performing optimization necessity evaluation according to the segment voice data to generate necessity evaluation data; Step S34 includes the following steps: Step S341: extracting feature information from the text and speech data to generate feature text data; Step S342: context-associate the feature text data according to the demand keyword data to generate associated feature text data; Step S343: selecting characteristic voice segments from the navigation-related voice data based on the associated characteristic text data to generate segment voice data; Step S344: performing voice quality assessment on the segment voice data to generate voice quality data; Step S345: Perform optimization necessity evaluation on the voice quality data using the optimization necessity evaluation formula to generate necessity evaluation data; the optimization necessity evaluation formula in step S345 is as follows: ； In the formula, is the necessity assessment value, is the start time of the evaluation, The end time of the evaluation. For time The total number of speech segments at For voice clips The quality rating of For voice clips Length, For voice clips The speech recognition accuracy of For voice clips The accuracy of environmental sound recognition is For voice clips The ambient noise level, For voice clips The incoherence value of is the length of the navigation route, For time The congestion level of the route at that time, is the irrelevance value of the route, is the frequency of route usage, Optimize frequency for history, is the base of natural logarithms, is the complexity of the route, It is the deviation correction value of the necessity assessment value; Step S35: screening the text and speech data for high necessity data according to the necessity evaluation data to generate data to be optimized and analyzed; Step S36: Formulate an optimization strategy based on the analyzed data to be optimized to generate path optimization strategy data; replan the planned path data according to the path optimization strategy data to generate path optimization data; Step S4: Transmit the path optimization data to the cloud device to generate optimization record summary data; extract the causes of navigation defects based on the optimization record summary data to generate navigation defect feedback data; optimize the navigation equipment according to the navigation defect feedback data to generate navigator update data.

2. The speech recognition navigation method based on planned path data according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: collecting user voice based on the voice collection device to generate user voice data; Step S12: performing navigation request voice detection on the user voice data to generate navigation request data; Step S13: positioning the device based on the navigation request data to generate starting position data; Step S14: classifying the sentence keywords according to the navigation request data to generate classified keyword data; identifying the travel mode and travel destination based on the classified keyword data to generate travel mode data and destination information data; Step S15: extract keywords from the destination information data to generate destination keyword data; preset the current travel range according to the travel mode data to generate current travel range data; match the map position based on the current travel range data and the destination keyword data to generate navigation destination data.

3. The speech recognition navigation method based on planned path data according to claim 2, characterized in that: Step S14 includes the following steps: Step S141: performing word segmentation annotation on the navigation request data to generate word segmentation request data; performing keyword extraction on the word segmentation request data to generate initial keyword data; Step S142: semantically analyzing the navigation request data to generate word association data; using the word association data to classify the initial keyword data to generate classified keyword data; Step S143: Identify the travel mode based on the classified keyword data and generate travel mode data; Step S144: performing location entity recognition based on the classified keyword data to generate location recognition result data; performing geographic location mapping based on the location recognition result data to generate mapping result data; Step S145: performing recognition accuracy evaluation based on the mapping result data to generate recognition accuracy data; Step S146: Perform data concatenation processing on the location recognition result data and the recognition accuracy data to generate destination information data.

4. The speech recognition navigation method based on planned path data according to claim 2, characterized in that: Step S15 includes the following steps: Step S151: Acquire preset frequently used address data and navigation history data; Step S152: processing the destination information data based on a preset recognition accuracy threshold, and generating destination keyword data when the destination information data is greater than the preset recognition accuracy threshold; Step S153: when the destination information data is less than or equal to the preset recognition accuracy threshold, fuzzy address matching is performed based on the preset common address data to generate destination keyword data; Step S154: dividing the navigation history data into travel tools to generate tool division history data; formulating the travel distance range according to the tool division history data to generate travel range data; Step S155: Preset the current travel range of the travel range data based on the travel mode data to generate the current travel range data; match the map position based on the current travel range data and the destination keyword data to generate the navigation destination data.

5. The speech recognition navigation method based on planned path data according to claim 4, characterized in that: Step S153 includes the following steps: Step S1531: when the destination information data is less than or equal to a preset recognition accuracy threshold, generating fuzzy place entity data; Step S1532: performing entity identification matching on the fuzzy location entity data based on the preset common address data to generate matching result data; when the matching result data is a successful match, generating destination keyword data; Step S1533: when the matching result data indicates a matching failure, accurate address consultation is performed based on the mapping result data to generate address consultation result data; Step S1534: performing location data association on the fuzzy location entity data according to the address consultation result data, and supplementing and correcting the preset common address data to generate common address data; Step S1535: extract the destination address based on the commonly used address data and generate destination keyword data.

6. The speech recognition navigation method based on planned path data according to claim 2, characterized in that: Step S2 includes the following steps: Step S21: performing navigation route planning based on the starting position data and the navigation destination data to generate planned path data; Step S22: performing path navigation according to the planned path data, and performing stay status analysis to generate stay status data; Step S23: using a voice collection device and performing real-time voice collection based on the stay status data to generate voice data during the stay period; Step S24: acquiring the movement direction based on the stay state data to generate stay direction data; extracting the turning direction from the stay direction data using the planned path data to generate turned stay data; Step S25: filtering the voice data during the stop period based on the turn stop data to generate voice data to be processed.

7. The speech recognition navigation method based on planned path data according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Transmitting the path optimization data to the cloud device to generate optimization record summary data; Step S42: clustering the optimization record summary data according to similar optimization methods to generate optimization content division data; Step S43: performing optimization reason analysis based on the optimization content division data to generate optimization reason data; Step S44: extracting the cause of the navigation defect based on the optimized cause data, and generating navigation defect feedback data; Step S45: Formulate equipment optimization strategy based on navigation defect feedback data and generate equipment optimization strategy data; optimize navigation equipment based on the equipment optimization strategy data and generate navigator update data.

Images