Roaming intelligent driving method and system and readable storage medium

By leveraging natural language understanding and voiceprint recognition technologies, combined with data interfaces from multiple travel platforms and preset scoring rules, the system dynamically adjusts intelligent driving parameters, solving the problems of cumbersome user interaction and fixed parameter settings in existing technologies, and enhancing the personalization and flexibility of the intelligent driving system.

CN121973798APending Publication Date: 2026-05-05ZHIJI AUTOMOTIVE TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202512006954.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-29
Publication Date
2026-05-05

AI Technical Summary

Technical Problem

Existing intelligent driving systems present cumbersome user interaction methods when dealing with scenarios where there is no clear destination or where the destination needs to be dynamically adjusted. They are unable to dynamically adapt driving parameters according to travel intentions and real-time road conditions, which affects driving comfort and personalized experience.

Method used

By extracting travel intent keywords through natural language understanding and voiceprint feature recognition technologies, calling data interfaces of multiple travel platforms to obtain destination information, recommending destinations based on preset scoring rules, and dynamically adjusting intelligent driving parameters such as vehicle speed, following distance, suspension height, lane selection, and lane change responsiveness.

Benefits of technology

It enables users to receive personalized travel suggestions without having to manually enter their destination, improving the accuracy and convenience of destination selection, allowing driving behavior to better adapt to different travel scenarios, and enhancing driving comfort, safety, and personalized experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121973798A_ABST
    Figure CN121973798A_ABST
Patent Text Reader

Abstract

The invention relates to a roaming intelligent driving method and system and a readable storage medium, and relates to the field of automatic driving, and the method comprises the steps: receiving the voice input of a user, and extracting a travel intention keyword through natural language understanding and voiceprint feature recognition; calling data interfaces of a plurality of travel platforms based on the travel intention keyword, and obtaining destination information matched with the travel intention keyword; sorting the destination information according to a preset scoring rule, and recommending a destination to the user in a voice interaction mode; and dynamically adjusting intelligent driving parameters according to the destination confirmed by the user and the travel intention keyword. According to the invention, dynamic destination matching and driving parameter adaptive adjustment based on user intention can be realized, so that the flexibility, comfort and personalized experience of travel are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of autonomous driving, and in particular to a roaming intelligent driving method, system, and readable storage medium. Background Technology

[0002] With the rapid development of technology, the integration of intelligent driving and large-scale models has become a research hotspot and development trend in the automotive industry. Traditional intelligent driving systems typically rely on the user having a clear destination and generating a specific route based on navigation for autonomous driving. While this approach improves driving convenience, it still requires the user to have a clear travel intention and destination before setting off. Existing intelligent driving systems primarily plan routes by manually inputting the destination and control the vehicle using fixed or manually adjustable driving parameters. Some systems display highly popular points of interest on the navigation map, guiding the user to select them as destinations, or use a default driving strategy when no navigation route is available.

[0003] However, existing technologies suffer from insufficient adaptability when dealing with travel scenarios without a clear destination or requiring dynamic destination adjustments. User interaction methods are relatively cumbersome, unable to understand travel intentions through natural language and match suitable destinations accordingly, especially lacking flexibility in complex tourism and business scenarios. Furthermore, intelligent driving parameters are typically fixed, unable to dynamically adapt to travel intentions and real-time traffic conditions, impacting driving comfort and personalized experiences. These limitations make it difficult for existing systems to meet the broader and more diverse travel needs of users. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a roaming intelligent driving method, system and readable storage medium that can realize dynamic destination matching and adaptive adjustment of driving parameters based on user intent, thereby improving the flexibility, comfort and personalized experience of travel.

[0005] To achieve the above objectives, the present invention adopts the following technical solution.

[0006] Firstly, the present invention provides a method for intelligent driving while roaming, which adopts the following technical solution: It receives users' voice input and extracts travel intent keywords through natural language understanding and voiceprint feature recognition; Based on the travel intent keywords, the data interfaces of multiple travel platforms are invoked to obtain destination information that matches the travel intent keywords; The destination information is sorted according to preset scoring rules, and destinations are recommended to users through voice interaction; and The intelligent driving parameters are dynamically adjusted based on the user's confirmed destination and the keywords indicating their travel intent.

[0007] Furthermore, in the aforementioned intelligent driving roaming method, the step of receiving the user's voice input and extracting travel intent keywords through natural language understanding and voiceprint feature recognition includes: Noise filtering and frame segmentation are performed on the speech signal; Voiceprint features are extracted using a voiceprint recognition model, and emotion recognition is performed. Use a speech recognition model to convert speech signals into text; Semantic analysis of transcribed text was performed using a large language model to extract keywords related to travel intent; and The results of emotion recognition, intent recognition, number of passengers in the vehicle, and real-time vehicle location are fused into multimodal data.

[0008] Furthermore, in the aforementioned intelligent driving roaming method, the travel intention keywords include at least one of the following: working, picking up and dropping off children, flower viewing, dining, shopping, hiking, mountain climbing, seeing a doctor, and exercising.

[0009] Furthermore, in the above-mentioned intelligent driving roaming method, the step of calling the data interfaces of multiple travel platforms based on the travel intent keywords includes: Combine time information, location information, and keywords related to travel intent as search criteria; Calling application programming interfaces (APIs) of multiple travel platforms to obtain user-generated content data; and Destination information and corresponding geographic coordinates are extracted from the user-generated content data.

[0010] Furthermore, in the aforementioned intelligent driving roaming method, the preset scoring rules are based on a comprehensive scoring calculation using the number of likes, the number of favorites, the matching degree of travel mode, and the matching degree of travel duration.

[0011] Furthermore, in the above-mentioned roaming intelligent driving method, the dynamic adjustment of intelligent driving parameters includes: Fill in the road segment attribute information based on the point of interest attributes in the navigation route; Based on the vehicle's real-time location and the corresponding point of interest attributes, and the road type identified by perception, a pre-set parameter matching table is queried; and Dynamically adjust at least one of the following parameters: speed limit, following distance, suspension height, lane selection, and lane change responsiveness.

[0012] Furthermore, in the above-mentioned intelligent driving roaming method, the point of interest attributes include at least one of schools, hospitals, tourist attractions, business districts, airports, night markets, mountain roads, tunnels, and residential areas.

[0013] Furthermore, the aforementioned intelligent driving roaming method also includes: Based on the location of the optimal route, search for similar popular points of interest in the travel popularity database; The aforementioned popular interests will be used as recommendation points to provide feedback to users; and The navigation route plan is updated based on user confirmation.

[0014] Furthermore, in the above-mentioned intelligent driving roaming method, the search range for nearby popular points of interest is within a preset distance threshold, and the preset distance threshold is dynamically adjusted according to the road type and vehicle speed.

[0015] Secondly, the intelligent driving system provided by this invention adopts the following technical solution: The voice interaction module is used to receive users' voice input and extract travel intent keywords through natural language understanding and voiceprint feature recognition; The data acquisition module is used to call the data interfaces of multiple travel platforms based on the travel intent keywords to obtain destination information that matches the travel intent keywords; The recommendation module is used to sort the destination information according to preset rating rules and recommend destinations to users through voice interaction; and The parameter adjustment module is used to dynamically adjust intelligent driving parameters based on the user-confirmed destination and the travel intent keywords.

[0016] Thirdly, the present invention provides a readable storage medium, which adopts the following technical solution: A readable storage medium storing computer instructions that, when executed by a processor, implement the roaming intelligent driving method as described in any one of the first aspects above.

[0017] In summary, compared with the prior art, the present invention has at least one of the following beneficial technical effects: The intelligent driving method of this invention utilizes natural language understanding and voiceprint recognition technologies to intelligently identify user travel intentions, providing personalized travel suggestions without requiring users to manually input a specific destination. By accessing data interfaces from multiple travel platforms and combining them with preset scoring rules, it can recommend destinations that better match the user's travel intentions, improving the accuracy and convenience of destination selection. Simultaneously, by dynamically adjusting intelligent driving parameters based on the user-confirmed destination and travel intention keywords, the vehicle's driving behavior can better adapt to different travel scenarios, thereby enhancing driving comfort, safety, and personalized experience, and meeting users' diverse travel needs in tourism, business, and other scenarios. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 The overall flowchart of the intelligent driving roaming method is shown.

[0020] Figure 2 The flowchart of the travel intention recognition module is shown.

[0021] Figure 3 The flowchart illustrates the process of identifying popular destinations based on travel intentions.

[0022] Figure 4 A flowchart of intelligent driving parameter matching is shown.

[0023] Figure 5 The system architecture diagram of the intelligent driving system is shown. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Furthermore, it should be understood that the specific embodiments described herein are only for illustration and explanation of this application and are not intended to limit this application.

[0025] It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments of this application. Furthermore, the descriptions of each embodiment in the following embodiments have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0026] The method steps described in this embodiment of the invention can be executed in the order described in the specific implementation, or the execution order of each step can be adjusted according to actual needs, provided that the technical problem can be solved. These are not listed one by one here.

[0027] The present invention will be further described in detail below with reference to the accompanying drawings.

[0028] Reference Figure 1This invention provides a method for intelligent driving roaming, which uses user voice input as a starting point and a voice acquisition device to capture the voice commands of the driver or passenger. In some embodiments, the voice acquisition device is a microphone array installed in the vehicle's ceiling or dashboard.

[0029] After receiving the user's voice input, the method performs a travel intent keyword recognition step. This step analyzes and processes the user's voice input using natural language understanding and voiceprint recognition technologies. In some implementations, natural language understanding technology includes speech recognition and semantic analysis, converting the speech signal into text and extracting semantic information. Voiceprint recognition technology analyzes the user's voice characteristics to identify the user's identity and emotional state. Through the combination of these two technologies, the method can accurately extract travel intent keywords, such as "office work," "picking up children," "flower viewing," "dining," "shopping," "hiking," "mountain climbing," "medical treatment," and "fitness."

[0030] Based on the extracted travel intent keywords, the method proceeds to a social media keyword search step. This step calls the data interfaces of multiple travel platforms to obtain destination information matching the travel intent keywords. In some implementations, the travel platforms include social and travel platforms. The method combines time information, location information, and travel intent keywords as query conditions, obtains user-generated content data through application programming interfaces, and extracts destination information and corresponding geographic coordinates from it.

[0031] After obtaining destination information, the method executes the influencer destination recommendation step. This step sorts the destination information according to preset scoring rules and recommends destinations to the user via voice interaction. In some implementations, the preset scoring rules are based on a comprehensive score calculation based on the number of likes, the number of favorites, the matching degree of travel mode, and the matching degree of travel duration. The method arranges the destinations in descending order of score and confirms with the user one by one via voice, such as asking "Do you want to go to [location name]?"

[0032] After the user confirms their destination, the method proceeds to the navigation planning and location tagging step, which considers the user's travel intentions. This step combines the user-confirmed destination with the vehicle's real-time location for route planning. The vehicle's real-time location information is simultaneously input into the navigation planning step and the subsequent intelligent driving parameter dynamic adaptation step to ensure the coordination between route planning and parameter adjustment.

[0033] The final step of the method is dynamic adaptation of intelligent driving parameters. Based on the user-confirmed destination and travel intent keywords, the intelligent driving parameters are dynamically adjusted. This step receives route information, location tag information, and real-time vehicle location information from the navigation planning step. In some implementations, the dynamically adjusted intelligent driving parameters include speed limits, following distance, suspension height, lane selection, and lane change responsiveness. The method queries a pre-defined parameter matching table based on the point of interest attributes in the navigation route and the perceived road type to achieve intelligent adjustment of driving parameters.

[0034] The entire intelligent driving approach achieves a personalized intelligent driving experience based on user intent through the collaborative work of each step. From voice input to intent recognition, and then to destination recommendation and driving parameter adjustment, each step forms a complete closed-loop processing flow, enabling the vehicle to understand the user's travel needs and automatically adapt the corresponding driving strategies.

[0035] Reference Figure 2 The process of receiving user voice input and extracting travel intent keywords through natural language understanding and voiceprint feature recognition involves multiple processing steps. This process begins with voice input reception and preprocessing, using a microphone array mounted on the vehicle's ceiling or dashboard to collect voice signals from the driver and passengers.

[0036] In the speech input reception and preprocessing stage, the method performs noise filtering and framing processing on the acquired speech signal. The noise filtering step filters background noise from the speech signal to improve speech quality and ensure the accuracy of subsequent processing. Framing processing divides the speech signal into small time frames and applies windowing. Speech framing and windowing processing reduce boundary effects and prepare for speech feature extraction.

[0037] like Figure 2 As shown, the preprocessed speech signal enters two parallel processing branches. The first branch extracts voiceprint features and performs emotion recognition using a voiceprint recognition model. The voiceprint recognition model analyzes the user's voice characteristics to determine if the user is in an emergency. The emotion recognition process analyzes the user's emotional state using voiceprint features, providing auxiliary information for subsequent intent recognition.

[0038] The second parallel processing branch uses a speech recognition model to convert the speech signal into text. In some implementations, the speech recognition model employs a Transformer model architecture. The Transformer model converts the processed speech signal into text and outputs confidence scores and detailed information for each word. The speech recognition process generates transcribed text, providing the textual basis for subsequent semantic analysis.

[0039] After the transcribed text is generated, the method uses a large language model to perform semantic analysis on the transcribed text and extract travel intent keywords. The large language model performs semantic analysis on the speech-transcribed text to identify information such as the user's travel intent and time. In some implementations, the large language model uses specific prompt words for intent recognition, such as: "You are a smart assistant that can extract the corresponding travel intent based on the user's dialogue input."

[0040] Continue to refer to Figure 2 The method integrates emotion recognition results, intent recognition results, vehicle passenger numbers, and real-time vehicle location during the multimodal data fusion stage. The multimodal data fusion process cross-validates the emotion recognition results extracted from voiceprint features with the intent recognition results output by the large language model. Vehicle passenger information and real-time vehicle location information are incorporated as vehicle body signals into the fusion process, improving the accuracy of travel intent recognition.

[0041] Travel intent keywords include at least one of the following: going to work, picking up children, flower viewing, dining, shopping, hiking, mountain climbing, seeing a doctor, and exercising. These travel intent keywords cover the main scenarios of users' daily travel. In some implementations, the method asks the user via voice to confirm the accuracy of the recognition results based on the travel intent keywords with high confidence.

[0042] The entire travel intent recognition process forms a closed-loop control through travel intent output and feedback. The method feeds back high-confidence travel intents to the user via voice for confirmation, and the user's feedback is then returned to the voice input reception and preprocessing stage, enabling continuous interaction optimization. A multimodal information fusion approach, combining voiceprint features, semantic understanding, and vehicle status information, improves the reliability of travel intent recognition.

[0043] Reference Figure 3 The process of calling multiple travel platform data interfaces based on travel intent keywords begins with three input pieces of information. The method combines time information, location information, and travel intent keywords as query conditions to form structured search parameters. Time information is provided at the seasonal or monthly granularity, location information is determined at the city granularity, and travel intent keywords represent the specific intent type confirmed by the user.

[0044] like Figure 3 As shown, after the three input pieces of information are combined, the API call step is initiated. The method calls the application programming interfaces (APIs) of multiple travel platforms to obtain user-generated content data. In some implementations, the travel platforms include social and tourism platforms. The APIs receive combined query conditions, search for matching content in the databases of each platform, and return the relevant user-generated content data.

[0045] After the API call returns the address information of the top destinations, the method extracts the destination information and corresponding geographic coordinates from the user-generated content data. User-generated content data includes travelogues, reviews, images, and location tags posted by platform users. The method uses data parsing techniques to identify and extract specific destination names, address descriptions, and GPS coordinates from this content.

[0046] Continue to refer to Figure 3 After extracting destination information, the method proceeds to the scoring and sorting step. This step calculates a comprehensive score for the destination information based on preset scoring rules. The preset scoring rules are based on four dimensions: number of likes, number of favorites, travel mode matching degree, and travel duration matching degree.

[0047] The overall score is calculated using a weighted summation method, specifically: Overall Score = Number of Likes × 0.4 + Number of Favorites × 0.3 + Travel Mode Matching Score × 0.15 + Travel Duration Matching Score × 0.15. In this formula, the number of likes and favorites reflects the destination's popularity, accounting for 40% and 30% of the weight, respectively. Travel mode matching assesses the degree to which the destination matches the user's travel intentions, while travel duration matching assesses the degree to which the time required to reach the destination matches the user's expectations; both scores account for 15% of the weight.

[0048] After the ratings are calculated, the method sorts the popular destinations from highest to lowest based on their overall ratings. This sorting process ensures that destinations with higher ratings are displayed to users first. In some implementations, the method selects the top-rated destinations as recommendation candidates.

[0049] like Figure 3 As shown, after the destinations are sorted, the user confirmation step begins. The method provides voice feedback to the user for confirmation for each sorted destination. The voice feedback uses natural language to ask the user if they want to select a specific destination. The user confirms or declines the recommended destinations through voice responses, and the method determines the final travel destination based on the user feedback.

[0050] The entire process of calling data interfaces from multiple travel platforms based on travel intent keywords achieves intelligent matching from user intent to destination recommendations. By acquiring rich candidate destination information through multi-platform data interfaces, combining it with a multi-dimensional scoring mechanism for intelligent sorting, and allowing users to confirm their final selection through interactive methods, this process improves the accuracy of destination recommendations and user experience.

[0051] Reference Figure 4The process of dynamically adjusting intelligent driving parameters begins with two inputs: the navigation route and points of interest (POIs). This process analyzes the POIs of various points along the navigation route to provide foundational data for the dynamic adjustment of intelligent driving parameters. POIs include at least one of the following: schools, hospitals, tourist attractions, commercial areas, airports, night markets, mountain roads, tunnels, and residential areas.

[0052] The process of filling road segment attribute information based on the point of interest (POI) attributes in the navigation route extends the POI attributes of a point to the corresponding road segment range. In some implementations, the method defaults to filling the POI attributes of the road segment where the point is located with the POI attributes of the point in the navigation route. When the coverage area of ​​a point is less than 500 meters, the method extends the attribute filling range to road segments within 500 meters before and after it to ensure the integrity and continuity of the road segment attribute information.

[0053] like Figure 4 As shown, after the road segment attribute information is filled in, the method processes the data based on the point of interest attributes corresponding to the vehicle's real-time location and the perceived road type. The vehicle-side sensing devices include cameras, millimeter-wave radar, and lidar, used to identify environmental information such as road type, lane lines, and traffic signs. Based on the road information acquired by the sensing devices, the method classifies road types into four main categories: school / commercial / construction area, mountain road / muddy road, tourist attraction, and residential area.

[0054] Based on the identified road type, the method queries a pre-defined parameter matching table to obtain the corresponding driving parameter configuration. The pre-defined parameter matching table stores intelligent driving parameter settings corresponding to different point-of-interest attributes and road types. The method searches for a matching parameter configuration in the parameter matching table based on the vehicle's real-time location, the corresponding point-of-interest attributes, and the perceived road type.

[0055] Continue to refer to Figure 4 The method dynamically adjusts at least one parameter among speed limit, following distance, suspension height, lane selection, and lane change responsiveness. For roads in school / commercial / construction areas, the method sets the speed limit to 10% lower than the road speed limit, the lane change strategy to prohibit overtaking and lane changing, and the suspension height to standard mode. This parameter configuration takes into account the pedestrian density in school and commercial areas as well as the safety requirements of construction areas.

[0056] For mountain roads / muddy roads, different parameter adjustment strategies are employed. The lane-changing strategy is set to prohibit overtaking and lane changes to avoid dangerous lane-changing maneuvers in complex road conditions. Following distance is increased by 10% from the existing mode, providing more sufficient safe distance for braking on mountain roads or muddy surfaces. The suspension height is adjusted to a raised mode to improve vehicle passability and stability.

[0057] like Figure 4As shown, roads designated for scenic spots employ a specific parameter configuration strategy. The speed limit is set to 10% lower than the road speed limit, allowing passengers to better enjoy the scenery. The lane selection strategy prioritizes the right lane for better views and to avoid other vehicles. The suspension height is adjusted to a raised mode for improved visibility and ride comfort.

[0058] For residential roads, the method implements targeted parameter adjustments. Steering parameters are set to a wide-angle mode to accommodate sharp turns and narrow roads that may exist in residential areas. The lane selection strategy is set to prioritize the leftmost lane to minimize impact on roadside parking and pedestrian activity. Suspension height is maintained in standard mode to ensure normal driving performance on flat roads.

[0059] The entire process of dynamically adjusting intelligent driving parameters achieves intelligent configuration of driving parameters through real-time analysis of road type and point-of-interest attributes. The method automatically adjusts the vehicle's driving behavior according to different driving scenarios, improving driving comfort and adaptability while ensuring driving safety. The coordinated adjustment of various parameters ensures that the vehicle can adapt to diverse driving environments, from urban roads to mountain roads, and from commercial areas to residential areas.

[0060] Reference Figure 1 and Figure 3 The intelligent driving roaming method further includes a step in the navigation route planning process of searching for nearby popular points of interest in a travel popularity database based on the locations along the optimal route. This step is executed after the system completes navigation planning and location tagging that takes into account travel intentions. By analyzing the various locations along the optimal route, it identifies and retrieves popular points of interest similar to the route in the travel popularity database.

[0061] The travel popularity database stores information on popular points of interest (POIs) in various regions, including their geographical coordinates, popularity ratings, user reviews, and visit frequency. In some implementations, the travel popularity database is maintained and updated through a cloud service platform, regularly obtaining the latest POI popularity data from multiple social media and travel platforms. The method uses the geographical coordinates of each point on the optimal route to perform a spatial range query in the travel popularity database, identifying popular POIs located within a preset distance threshold.

[0062] like Figure 1 As shown, the searched popular points of interest (POIs) are presented to the user as recommended route points. The method displays the identified POI information to the user via voice interaction, including the POI's name, type, distance from the main route, and estimated additional travel time. In some implementations, the method sorts the recommended route points according to their popularity rating and their relevance to the user's travel intentions, prioritizing POIs with higher ratings and better matching.

[0063] After a user confirms or declines a recommended waypoint via voice response, the method updates the navigation route plan based on the user's confirmation. When a user confirms adding a popular point of interest as a waypoint, the method recalculates the optimal route including that waypoint, updates navigation instructions and estimated arrival time. In some implementations, the route planning update process considers dwell time at waypoints, road congestion, and the user's time constraints to generate an adjusted complete itinerary plan.

[0064] Continue to refer to Figure 3 The search range for nearby popular online interest points is within a preset distance threshold, which is dynamically adjusted based on road type and vehicle speed. The method identifies the current road type using vehicle-mounted sensing devices, including different types such as highways, urban roads, rural roads, and mountain roads. Simultaneously, the method acquires real-time vehicle speed information as a reference parameter for adjusting the distance threshold.

[0065] In highway driving scenarios, due to higher vehicle speeds and limitations on lane changes and U-turns, the method sets a relatively large preset distance threshold, typically within the range of 5-10 kilometers. This setting ensures that recommended popular points of interest are located within a reasonable exit range from the highway, avoiding the need for users to make complex route adjustments. In urban road driving scenarios, the method sets a medium preset distance threshold, typically within the range of 2-5 kilometers, balancing the accessibility of waypoints with the reasonableness of route deviations.

[0066] For rural and mountain roads, due to the relatively complex road conditions and lower vehicle speeds, the method sets a small preset distance threshold, typically within the range of 1-3 kilometers. This setting takes into account the traffic conditions of rural and mountain roads, ensuring that the recommended waypoints are within a reasonable detour range and avoiding excessively prolonged travel time or increased driving difficulty.

[0067] The dynamic adjustment process of the preset distance threshold also considers real-time changes in vehicle speed. When the vehicle speed is high, the method appropriately increases the search range to provide users with more waypoint options. When the vehicle speed is low, the method narrows the search range to avoid recommending points of interest that are too far away. In some implementations, the distance threshold adjustment uses a linear or non-linear function relationship, calculating the final search range based on road type coefficients and speed coefficients.

[0068] The entire process of searching and recommending popular interest points dynamically adjusts the search scope, enabling more flexible and personalized waypoint recommendations. Based on actual driving conditions and user preferences, the method provides users with a wide range of waypoint options while ensuring the rationality of the trip, thus improving the quality and personalization of the intelligent driving experience.

[0069] This invention also discloses a roaming intelligent driving system.

[0070] Reference Figure 5 The intelligent driving system 500 described in this embodiment of the invention includes a voice interaction module 502, a data acquisition module 504, a recommendation module 506, and a parameter adjustment module 508. Through the coordinated operation of these modules, the intelligent driving system 500 achieves a complete functional flow from voice input to intelligent driving parameter adjustment.

[0071] like Figure 5 As shown, the voice interaction module 502 receives user voice input and extracts travel intent keywords through natural language understanding and voiceprint feature recognition. The voice interaction module 502 includes a microphone array mounted on the vehicle's roof or dashboard to collect voice signals from the driver and passengers. The voice interaction module 502 performs noise filtering and framing processing on the collected voice signals, extracts voiceprint features using a voiceprint recognition model, performs emotion recognition, converts the voice signals into text using a speech recognition model, and performs semantic analysis on the transcribed text using a large language model.

[0072] The data acquisition module 504 is connected to the voice interaction module 502 and is used to call the data interfaces of multiple travel platforms based on travel intent keywords to obtain destination information matching the travel intent keywords. The data acquisition module 504 combines time information, location information, and travel intent keywords as query conditions, calls the application programming interfaces of the travel platforms to obtain user-generated content data, and extracts destination information and corresponding geographic coordinates from it.

[0073] Continue to refer to Figure 5 The recommendation module 506 receives destination information from the data acquisition module 504, sorts the destination information according to preset scoring rules, and recommends destinations to the user through voice interaction. The recommendation module 506 calculates a comprehensive score based on the number of likes, favorites, travel mode matching degree, and travel duration matching degree, sorts popular destinations from highest to lowest score, and confirms the recommended destinations to the user one by one via voice.

[0074] The parameter adjustment module 508 is connected to the recommendation module 506 and is used to dynamically adjust intelligent driving parameters based on the user-confirmed destination and travel intent keywords. The parameter adjustment module 508 fills in road segment attribute information based on the point of interest attributes in the navigation route, and based on the point of interest attributes corresponding to the vehicle's real-time location and the perceived road type, queries a preset parameter matching table to dynamically adjust at least one parameter among speed limit, following distance, suspension height, lane selection, and lane change responsiveness.

[0075] like Figure 5As shown, the recommendation module 506 and the voice interaction module 502 form a feedback connection to achieve closed-loop control of user interaction and system response. The user confirms or rejects the recommended destination via voice response, and the feedback information is returned to the voice interaction module 502 through this connection. The system then adjusts subsequent recommendation strategies and parameter configurations based on the user feedback.

[0076] The Roaming Intelligent Driving System 500 includes an in-vehicle terminal, which comprises an in-vehicle computing unit and a storage unit. The in-vehicle computing unit runs the travel intent recognition module, destination recommendation module, and intelligent driving parameter matching module, executing the computational tasks of each functional module. The storage unit stores corresponding model parameters and rule tables, including voiceprint recognition model parameters, large language model parameters, scoring rule tables, and parameter matching tables.

[0077] The intelligent driving system 500 includes a vehicle-mounted sensing device, which comprises cameras, millimeter-wave radar, lidar, and other equipment to identify environmental information such as road types, lane lines, and traffic signs. The vehicle-mounted sensing device provides real-time road environment data to the parameter adjustment module 508, supporting dynamic adjustment decisions for intelligent driving parameters.

[0078] The Roaming Intelligent Driving System 500 includes a communication module that interacts with the cloud service platform via 4G / 5G or vehicle-to-everything (V2X) networks. This module retrieves user-generated content data from social media platforms, an interest point popularity database, and updated large language model parameters from the cloud, providing real-time data support for various functional modules.

[0079] The Roaming Intelligent Driving System 500 includes a cloud service platform. This platform is used for cross-platform data aggregation and offline model training, and periodically sends updated models and point-of-interest databases to the vehicle. The cloud service platform integrates data sources from multiple mobility platforms, performs data cleaning and model optimization, and pushes the latest model parameters and database updates to the in-vehicle terminal through the communication module, ensuring continuous optimization of system functions and real-time data.

[0080] This invention also discloses a readable storage medium.

[0081] A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the intelligent driving method described in any of the above embodiments. The computer-readable storage medium may include any entity or device capable of carrying a computer program, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc. The computer program includes computer program code. The computer program code may be in the form of source code, object code, an executable file, or some intermediate form, etc. The computer-readable storage medium may include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc.

[0082] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0083] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a system including a processing module or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0084] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent driving roaming, characterized in that, include: It receives users' voice input and extracts travel intent keywords through natural language understanding and voiceprint feature recognition; Based on the travel intent keywords, the data interfaces of multiple travel platforms are invoked to obtain destination information that matches the travel intent keywords; The destination information is sorted according to preset scoring rules, and destinations are recommended to users through voice interaction. as well as The intelligent driving parameters are dynamically adjusted based on the user's confirmed destination and the keywords indicating their travel intent.

2. The roaming intelligent driving method according to claim 1, characterized in that, The process of receiving user voice input and extracting travel intent keywords through natural language understanding and voiceprint feature recognition includes: Noise filtering and frame segmentation are performed on the speech signal; Voiceprint features are extracted using a voiceprint recognition model, and emotion recognition is performed. Use a speech recognition model to convert speech signals into text; Semantic analysis of transcribed text was performed using a large language model to extract keywords related to travel intent; and The results of emotion recognition, intent recognition, number of passengers in the vehicle, and real-time vehicle location are fused into multimodal data.

3. The roaming intelligent driving method according to claim 2, characterized in that, The travel intention keywords include at least one of the following: going to work, picking up and dropping off children, flower viewing, dining, shopping, hiking, mountain climbing, seeing a doctor, and exercising.

4. The roaming intelligent driving method according to claim 1, characterized in that, The method of calling data interfaces of multiple travel platforms based on the travel intent keywords includes: Combine time information, location information, and keywords related to travel intent as search criteria; Calling application programming interfaces (APIs) of multiple travel platforms to obtain user-generated content data; and Destination information and corresponding geographic coordinates are extracted from the user-generated content data.

5. The roaming intelligent driving method according to claim 4, characterized in that, The preset scoring rules are based on a comprehensive score calculation using the number of likes, the number of favorites, the matching degree of travel mode, and the matching degree of travel duration.

6. The roaming intelligent driving method according to claim 1, characterized in that, The dynamic adjustment of intelligent driving parameters includes: Fill in the road segment attribute information based on the point of interest attributes in the navigation route; Based on the vehicle's real-time location and the corresponding point of interest attributes, and the road type identified by perception, a pre-set parameter matching table is queried; and Dynamically adjust at least one of the following parameters: speed limit, following distance, suspension height, lane selection, and lane change responsiveness.

7. The roaming intelligent driving method according to claim 6, characterized in that, The points of interest attributes include at least one of the following: schools, hospitals, tourist attractions, commercial districts, airports, night markets, mountain roads, tunnels, and residential areas.

8. The roaming intelligent driving method according to claim 1, characterized in that, Also includes: Based on the location of the optimal route, search for similar popular points of interest in the travel popularity database; The aforementioned popular interests will be used as recommendation points to provide feedback to users. as well as The navigation route plan is updated based on user confirmation.

9. The roaming intelligent driving method according to claim 8, characterized in that, The search range for similar online celebrity points of interest is within a preset distance threshold, and the preset distance threshold is dynamically adjusted according to road type and vehicle speed.

10. A roaming intelligent driving system, characterized in that, include: The voice interaction module is used to receive users' voice input and extract travel intent keywords through natural language understanding and voiceprint feature recognition; The data acquisition module is used to call the data interfaces of multiple travel platforms based on the travel intent keywords to obtain destination information that matches the travel intent keywords; The recommendation module is used to sort the destination information according to preset scoring rules and recommend destinations to users through voice interaction. as well as The parameter adjustment module is used to dynamically adjust intelligent driving parameters based on the user-confirmed destination and the travel intent keywords.

11. A readable storage medium, characterized in that, The readable storage medium stores computer instructions that, when executed by a processor, implement the roaming intelligent driving method as described in any one of claims 1-9.