Itinerary scheduling methods, devices, electronic equipment and storage media

Abstract

This invention provides a trip scheduling method, apparatus, electronic device, and storage medium, relating to the field of data processing technology. The method includes: acquiring multi-source trip information and performing structured processing on the multi-source trip information to obtain structured trip data; wherein the structured trip data includes point-of-interest information and time information; extracting scene features from the structured trip data to generate scene tags; generating a set of candidate solutions based on conflict detection results, and determining conflict solutions from the set of candidate solutions through an interactive decision-making process and a content decision-making process; and dynamically scheduling the structured trip data based on the conflict solutions to obtain scheduled trip information.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a route scheduling method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the rapid development of intelligent vehicles and vehicle networking technology, the intelligent cockpit has become the core carrier of automotive intelligence, and users face a large number of trip management needs in their daily travel.

[0003] However, existing in-vehicle navigation and trip management solutions typically only support route planning for a single destination or provide basic, single-dimensional time reminders. Faced with the complex multiple trips users encounter in their daily commutes, existing systems cannot perceive the differentiated needs of different travel scenarios or the real-time driving status of the vehicle, nor can they automatically detect and handle potential complex conflicts between multiple trips. This forces users to manually coordinate and rearrange trips during driving when faced with overlapping trips or unexpected situations, resulting in low management efficiency, significant distraction from driving, and obvious driving safety hazards. Summary of the Invention

[0004] This invention provides a trip planning method, apparatus, electronic device, and storage medium to solve the problem that existing systems cannot perceive the differentiated needs of different travel scenarios when faced with complex multiple trip arrangements in users' daily travel.

[0005] This invention provides a itinerary scheduling method, comprising: Multi-source travel information is acquired and structured to obtain structured travel data; wherein, the structured travel data includes point-of-interest information and time information; Scene features are extracted from the structured trip data to generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The structured travel data is subjected to multi-dimensional conflict detection according to the conflict detection configuration determined by the scene label to obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold, and detection frequency; Based on the conflict detection results, a set of candidate solutions is generated, and a conflict solution is determined from the set of candidate solutions through an interactive decision-making process and a content decision-making process. The interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution. The content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier. The structured itinerary data is dynamically rearranged based on the conflict resolution method to obtain rearranged itinerary information.

[0006] According to a trip scheduling method provided by the present invention, the sources of the multi-source trip information include at least two of the following: calendar applications, in-vehicle navigation commands, AI agent output, and user natural language input; the multi-source trip information is subjected to structured processing to obtain structured trip data, including performing the following structured processing steps: The multi-source trip information, along with the current time context and the current location context, is input into a large language model. The large language model then performs multi-intent segmentation, interest point semantic reasoning, and relative time normalization on the multi-source trip information to obtain initial structured data. The initial structured data is subjected to duplicate travel deduplication and data field completion processing to obtain the structured travel data; wherein, the data field completion processing is used to complete the field by calling an external point of interest interface when the point of interest information is missing a preset core field; When the reasoning time of the large language model exceeds a preset time threshold, the system switches to a rule-based processing method based on regular expressions and named entity recognition to perform the structured processing on the multi-source trip information to obtain the structured trip data.

[0007] According to a trip planning method provided by the present invention, the point of interest information includes at least one of point of interest name, point of interest coordinates, point of interest category, and point of interest opening hours; The step of extracting scene features from the structured travel data and generating scene labels includes performing the following label generation steps: Based on the points of interest categories, the event keywords in the structured itinerary data, and the sources of the multi-source itinerary information, a static scene identifier is generated; wherein, the static scene identifier includes at least two of the following: time urgency identifier, itinerary distance span identifier, and travel purpose identifier; the time urgency identifier includes one of rigid urgency, flexible urgency, and no time constraint; the itinerary distance span identifier includes one of short-distance within the city, medium-distance between cities, and long-distance across provinces; and the travel purpose identifier includes one of business travel, family travel, medical travel, leisure travel, and daily commuting. The dynamic scene identifier is generated based on vehicle motion signals; wherein the vehicle motion signals include at least one of vehicle speed signals, gear signals, and positioning signals, and the dynamic scene identifier includes one of parking, low-speed driving, and high-speed driving.

[0008] According to a trip scheduling method provided by the present invention, the step of performing multi-dimensional conflict detection on the structured trip data according to the conflict detection configuration determined by the scene label to obtain conflict detection results includes performing at least one of the following specific scene detection steps: When the travel purpose identifier indicates business travel or medical travel and the point of interest category is a transportation hub, the latest departure time is calculated by reverse time calculation, the allowable stay time window for intermediate trips is calculated based on the latest departure time, and then a prompt message to shorten or skip the intermediate trip is generated based on the allowable stay time window. The latest departure time and the prompt message are used as part of the conflict detection result. When the trip distance span indicates a long-distance inter-provincial journey, the driving fatigue dimension and energy replenishment dimension are added to the multiple conflict dimensions. When the continuous driving time of the vehicle exceeds a preset fatigue threshold, an unskippable rest point insertion instruction is generated. When the ratio of the vehicle's remaining driving range to the remaining trip distance is lower than a preset replenishment threshold, an energy replenishment prompt is generated. The rest point insertion instruction and the energy replenishment prompt are used as part of the conflict detection result.

[0009] According to a trip scheduling method provided by the present invention, determining conflict solutions from the candidate solution set through an interactive decision-making process and a content decision-making process includes performing at least one of the following interactive configuration steps during the interactive decision-making process: When the time urgency indicator indicates rigid urgency and the dynamic scenario indicator indicates low-speed or high-speed driving, the output channel is limited to a voice channel, and the user response processing strategy is configured to automatically execute a single conflict resolution. When the time urgency indicator indicates rigid urgency and the dynamic scene indicator indicates parking, the output channel is configured as a visual channel, the number of conflict solutions is configured to be no less than two, and the user response processing strategy is configured to automatically select a preset default solution when no user selection is obtained within the first timeout period. When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scenario indicator indicates low-speed driving or high-speed driving, the user response processing strategy is configured to mark the corresponding conflict as pending when no user response is obtained within the second timeout period, and to prompt again when the dynamic scenario indicator changes to parking. When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scene indicator indicates parking, the output channel is configured as a visual channel, the conflict solution is configured to include complete conflict information and alternative point of interest information, and the user response processing strategy is configured to support user modification and order adjustment operations.

[0010] According to a trip planning method provided by the present invention, the content decision process includes: When the travel purpose identifier indicates business travel, the candidate solution set is limited to the fastest route strategy type, and the trip cancellation type is excluded from the candidate solution set; When the travel purpose identifier indicates family travel, alternative points of interest that do not have children's facilities are filtered from the candidate solution set, and a soft time window is reserved in the time arrangement of the target solution content; wherein, the soft time window includes lunch time and children's nap time; When the travel purpose identifier indicates leisure travel, an option to adjust the itinerary order is added to the candidate solution set; wherein, the option to adjust the itinerary order is used to exchange the order of outdoor and indoor itineraries according to weather distribution; When the travel destination identifier indicates daily commuting, a route preference set is generated based on the user's historical commuting routes, and the target route content is determined from the route preference set; When the travel destination identifier indicates medical travel, the candidate solution set is limited to the fastest route strategy type, and when the time taken for the regular route exceeds a preset value, an alternative hospital recommendation type is added to the candidate solution set.

[0011] According to a trip scheduling method provided by the present invention, when the conflict detection result includes at least two conflicts existing simultaneously in multiple conflict dimensions, the step of determining a conflict solution from the candidate solution set through an interactive decision-making process and a content decision-making process includes performing the following composite conflict resolution steps: When at least two conflicts include driving safety-related conflicts, the processing priority of the driving safety-related conflicts is placed at the highest level, and the intervention of the scene label is blocked; wherein, the driving safety-related conflicts include at least one of the combination of extreme weather and high-speed driving and continuous fatigue driving; When the solution generated for the first conflict among the at least two conflicts would trigger a second conflict, the associated impact information of the second conflict shall be output together with the conflict solution. When the number of same-day trips corresponding to the structured trip data exceeds a first threshold and the number of at least two conflicts exceeds a second threshold, at least two global orchestration schemes are generated at once based on the large language model, which integrates the multiple conflict dimensions and the constraints of the structured trip data.

[0012] According to a trip scheduling method provided by the present invention, the step of dynamically scheduling the structured trip data based on the conflict resolution to obtain the scheduled trip information includes performing the following scheduling processing steps: Extract the trip distance span identifier, the trip purpose identifier, the time urgency identifier, and the number of trips from the structured trip data; Based on the extracted feature data, the target arrangement strategy is determined from the preset spatial clustering arrangement strategy, the approximate shortest path arrangement strategy, the historical route learning arrangement strategy, and the endpoint backward arrangement strategy. According to the target orchestration strategy, route planning and time node calculation are performed on the structured trip data to obtain the orchestrated trip information; wherein, the process of determining the target orchestration strategy includes at least one of the following processing branches: When the trip distance span identifier indicates a short trip within the city and the number of trips exceeds a preset number, the target scheduling strategy is determined to be the spatial clustering scheduling strategy; wherein, the spatial clustering scheduling strategy is used to cluster based on the geographical location corresponding to the point of interest information, and uses points of interest with clear time requirements as fixed anchor points; When the trip distance span identifier indicates a long-distance inter-provincial trip or an intercity trip and the trip destination identifier indicates a leisure trip, the target scheduling strategy is determined as the approximate shortest path scheduling strategy; wherein, the approximate shortest path scheduling strategy is used to solve for the approximate shortest path for the nodes corresponding to the point of interest information, and to forcibly insert rest stops and energy replenishment points into the scheduled trip information; When the travel destination identifier indicates daily commuting, the target scheduling strategy is determined as the historical route learning scheduling strategy; wherein, the historical route learning scheduling strategy is used to generate a route preference probability distribution based on the user's historical commuting data, and select a travel route from the route preference probability distribution in combination with real-time traffic conditions; When the time urgency indicator indicates rigid urgency, the target scheduling strategy is determined as the endpoint backward scheduling strategy; wherein, the endpoint backward scheduling strategy is used to calculate the latest departure time and the available stay time window for each point of interest in the structured itinerary data in reverse from the deadline time.

[0013] According to a trip scheduling method provided by the present invention, after obtaining the scheduled trip information, the method further includes performing the following trip auxiliary service steps: When a trip interruption event is detected, the unfinished trips in the structured trip data are queried and real-time status detection is performed to obtain the real-time status detection results; When the real-time status detection result indicates no change, the interruption point navigation information is loaded and navigation is automatically resumed if no user response is obtained after a timeout; when the real-time status detection result indicates a change, the conflict resolution is regenerated and navigation resumes after obtaining user confirmation. During the trip, the triggering time and scope of recommended service information are determined based on the scenario tags; wherein, when the travel destination identifier indicates business travel, the scope of recommended content is limited to destination parking and business assistance information; when the travel destination identifier indicates daily commuting, the regular service recommendation function is turned off and replenishment service information is generated when the vehicle's remaining energy is lower than a preset value.

[0014] The present invention also provides a travel scheduling device, comprising: The acquisition module is used to acquire multi-source travel information and perform structured processing on the multi-source travel information to obtain structured travel data; wherein, the structured travel data includes point of interest information and time information; The generation module is used to extract scene features from the structured trip data and generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The detection module is used to perform multi-dimensional conflict detection on the structured travel data according to the conflict detection configuration determined by the scene label, and obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold and detection frequency; The determination module is used to generate a set of candidate solutions based on the conflict detection results, and to determine the conflict solution from the set of candidate solutions through an interactive decision-making process and a content decision-making process; wherein, the interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution; the content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier; The orchestration module is used to dynamically orchestrate the structured trip data based on the conflict resolution to obtain orchestrated trip information.

[0015] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the schedule arrangement method described above.

[0016] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the schedule arrangement method as described above.

[0017] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the schedule arrangement method described above.

[0018] This invention provides a trip scheduling method, apparatus, electronic device, and storage medium. By structuring multi-source trip information and extracting scene features, the system generates scene tags containing static and dynamic scene identifiers. This allows the system to perceive the travel scene of the current trip before conflict detection. Based on these scene tags, the system dynamically determines conflict detection configurations such as detection priority, judgment thresholds, and detection frequencies for multiple conflict dimensions. This enables multi-dimensional conflict detection to employ differentiated detection strategies for different travel scenarios, improving the timeliness and accuracy of identifying key conflicts. Simultaneously, in the conflict resolution stage, the system determines the output channel and user response processing strategy by cross-combining static and dynamic scene identifiers. Target solution content is filtered based on trip attribute features, ensuring that conflict solutions adapt to both the safety interaction requirements of the current driving state and the user's actual concerns under specific travel purposes. Based on this, the trip is dynamically scheduled, achieving refined and adaptive trip conflict management and scheduling for different travel scenarios. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the itinerary scheduling method provided by the present invention; Figure 2 A schematic diagram of the travel scheduling device provided by the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] The trip scheduling method provided in this invention can be applied to trip management systems in intelligent cockpit scenarios. The executing entity of this method can be a processor, controller, server, in-vehicle terminal, cloud computing platform, or other computing devices with data processing capabilities.

[0023] In one feasible implementation, the method can be executed collaboratively by a cloud server and an in-vehicle terminal. The cloud server handles computationally intensive tasks such as large language model inference, conflict detection calculation, and orchestration optimization, while the in-vehicle terminal handles tasks with high real-time requirements, such as user interaction, state awareness, and navigation execution. In the subsequent description of embodiments of this invention, the processor will be used as the execution entity.

[0024] It should be noted that the trip scheduling method involved in the embodiments of the present invention aims to provide users with intelligent multi-trip overall management capabilities in driving travel scenarios.

[0025] Users typically face multiple travel needs during their daily travels, such as commuting, business meetings, leisure travel, and life services. This travel information is often scattered across different terminals and applications, and there are potential conflicts in multiple dimensions such as time, space, and weather. Traditional navigation and schedule management tools struggle to achieve cross-application information integration, automatic detection of multi-dimensional conflicts, and scenario-based differentiated processing.

[0026] This invention achieves refined management of different travel scenarios by introducing a scenario-label-driven conflict detection and resolution mechanism.

[0027] Figure 1 This is a flowchart illustrating the itinerary scheduling method provided by the present invention, as shown below. Figure 1 As shown, the method includes the following: Step S110: Obtain multi-source trip information and perform structured processing on the multi-source trip information to obtain structured trip data, wherein the structured trip data includes point of interest information and time information; In this step, the processor obtains trip information related to the user's travel from multiple different information sources, including but not limited to terminal calendar applications, in-vehicle navigation systems, artificial intelligence agents, and natural language commands input by the user via voice or text.

[0028] Multi-source travel information refers to information gathered from two or more different types of channels. Because travel information from different sources differs in data format, field completeness, and representation, the processor needs to perform structured processing on this information so that subsequent stages can analyze and calculate based on a unified data format.

[0029] Structured travel data refers to a collection of travel information with a unified data structure after standardization processing, which includes at least point-of-interest information and time information.

[0030] Points of interest (POIs) refer to geographic location information related to the destination of a trip, such as location name, geographic coordinates, address description, and place category.

[0031] Time information refers to information related to the execution time of the trip, such as the estimated departure time, estimated arrival time, and the business hours window of the destination.

[0032] In some implementation scenarios, structured trip data can also include task description information and route preference information. Task description information refers to the specific tasks a user needs to complete at the destination, such as picking up documents, attending a meeting, or having a meal. Route preference information refers to the user's preferred route settings, such as prioritizing highways or choosing the shortest distance.

[0033] The structured processing can be implemented in various ways. In one implementation, the processor can use a large language model to perform semantic understanding on the natural language text input by the user, extract the travel intention and key information, and convert it into structured data in a preset format.

[0034] In another implementation, the processor can directly read existing structured data from a calendar application or navigation system and perform format conversion through a preset data interface.

[0035] During structured processing, the processor also needs to perform data cleaning and improvement operations. For example, when multiple sources provide itinerary information pointing to the same destination and at the same time, the processor needs to perform deduplication to avoid redundant data. Another example is when a point of interest in a itinerary is missing a necessary field, the processor can complete it by calling an external data interface.

[0036] As an example, suppose a user inputs a command containing multiple travel destinations via voice while in a car, and the user's phone calendar also contains several scheduled events. The processor obtains travel information from both the voice input and the calendar application. It performs semantic understanding on the voice input to extract points of interest and time information, and performs format conversion on the calendar events to unify data fields. Finally, it merges all travel information into structured travel data in a unified format.

[0037] Each piece of structured travel data is assigned a unique travel identifier and marked with its initial travel state, forming a set of travel data that can be used in subsequent processing stages.

[0038] Step S120: Extract scene features from the structured travel data and generate scene labels.

[0039] In this step, the processor extracts scene feature information related to the current trip based on the structured trip data obtained in step S110, and generates corresponding scene labels based on the scene feature information.

[0040] Scene tags are identifying information that characterizes the travel scene category of the current trip, and their role is to drive the differentiated configuration of subsequent conflict detection and orchestration strategies. Scene tags consist of two components: static scene identifiers and dynamic scene identifiers.

[0041] Static scene identifiers refer to scene identifiers generated based on the trip attribute features in structured trip data, which are determined at the time of trip creation and do not change in real time during trip execution.

[0042] Trip attribute features may include the category of the trip destination, the keyword features in the description of the trip items, the strength of the trip time constraints, and the spatial span of the trip.

[0043] The specific classification dimensions of static scene identifiers may include, but are not limited to: time urgency dimension, used to characterize the sensitivity of the trip to time constraints; distance span dimension, used to characterize the spatial range of the trip; and travel purpose dimension, used to characterize the main purpose category of the user's trip.

[0044] The processor can analyze features such as point of interest categories, event keywords, and time representation methods in structured trip data to determine the specific category values ​​for each dimension.

[0045] Dynamic scene identifiers refer to scene identifiers generated based on the vehicle's real-time operating status and continuously updated during the trip. The vehicle's real-time status can include the vehicle's speed, gear position, and location information.

[0046] The classification dimensions of dynamic scene identifiers can include, but are not limited to, the vehicle motion state dimension, used to characterize whether the vehicle is currently stationary or moving, and its speed. The processor determines the value of the dynamic scene identifier by reading the vehicle's sensor signals or the vehicle's infotainment system status information in real time.

[0047] Existing in-vehicle navigation and trip management solutions employ a uniform processing strategy for all trips, failing to consider the differences in user priorities across various travel scenarios. This invention introduces a scenario classification step before conflict detection, enabling subsequent conflict detection and resolution strategies to adaptively adjust based on specific scenarios, thereby achieving refined trip management.

[0048] Step S130: Perform multi-dimensional conflict detection on the structured trip data according to the conflict detection configuration determined by the scene label, and obtain the conflict detection results.

[0049] In this step, the processor determines the collision detection configuration applicable to the current scene based on the scene label generated in step S120, and performs multi-dimensional collision detection on the structured travel data according to the collision detection configuration to obtain the collision detection results.

[0050] Conflict detection configuration refers to a set of parameters used to control the conflict detection process, including at least one of the following: detection priority ranking, decision threshold, and detection frequency for multiple conflict dimensions.

[0051] Multiple conflict dimensions refer to the analytical dimensions that detect potential contradictions or infeasibility between itineraries from different perspectives.

[0052] These dimensions may include, but are not limited to: the time dimension, used to detect whether there is time overlap or insufficient time between multiple trips; Points of interest dimension is used to detect whether the accessibility and business status of the destination meet the travel requirements; The weather dimension is used to detect whether weather conditions at the destination or along the route affect the safety or quality of the trip. The route dimension is used to detect whether the route planning of the trip has problems such as serious detours, congestion or road closures; the event dimension is used to detect whether there are semantic logical contradictions between trips.

[0053] Detection priority ranking refers to the order in which the aforementioned conflict dimensions are detected during the detection process. In different scenarios, users pay varying degrees of attention to conflicts across different dimensions; therefore, the detection priority ranking needs to be dynamically adjusted based on scenario tags.

[0054] For example, in travel scenarios with strict time constraints, the detection priority of the time dimension should be higher than that of the weather dimension; while in travel scenarios where experience quality is the primary concern, the detection priority of the weather dimension may need to be increased.

[0055] The decision threshold refers to the quantitative standard used to determine whether a conflict has occurred in each conflict dimension. The decision threshold for the same dimension can differ in different scenarios. For example, the buffer time in the time dimension should be set more generously in urgent scenarios, while it can be appropriately reduced in lenient scenarios.

[0056] The detection frequency refers to the time interval or triggering condition at which the system performs real-time monitoring and detection of each conflict dimension during the execution of the process.

[0057] The detection frequency can be set differently for different scenarios to achieve a balance between detection timeliness and system resource consumption.

[0058] Conflict detection results refer to the output information after multi-dimensional detection, including whether a conflict exists, the specific type of conflict, the severity of the conflict, and the trip identifiers involved in the conflict.

[0059] Step S140: Generate a set of candidate solutions based on the conflict detection results, and determine the conflict solution from the set of candidate solutions through an interactive decision-making process and a content decision-making process.

[0060] In this step, when the conflict detection result of step S130 indicates that at least one conflict exists, the processor generates multiple possible solutions based on the type and severity of the conflict, forming a set of candidate solutions.

[0061] Subsequently, the processor determines the final conflict resolution from the set of candidate solutions through interactive decision-making processes and content decision-making processes.

[0062] A candidate solution set refers to a collection of alternative handling schemes generated for a detected conflict. The processor can automatically generate multiple feasible solutions by leveraging the reasoning capabilities of a large language model, combined with the constraints of the current itinerary and available resources. Each solution can include different types of adjustment measures such as adjusting itinerary time, replacing the destination, switching routes, and changing the itinerary sequence.

[0063] The interactive decision-making process refers to the decision-making process of determining how conflict solutions should be presented to users and how user feedback should be obtained. The core of this process lies in the cross-combination of static and dynamic scenario identifiers to determine the output channels of the conflict solution and the user response handling strategy.

[0064] Output channels refer to the interactive methods by which conflict solutions are presented to users, and can include voice channels, visual channels, or a combination of voice and visual channels.

[0065] The voice channel refers to broadcasting solution information to users through voice synthesis technology; the visual channel refers to displaying solution information to users through pop-up windows or card interfaces on the in-vehicle display screen.

[0066] Different combinations of static and dynamic scene identifiers correspond to different output channel selections. For example, when the vehicle is traveling at high speed, the voice channel should be prioritized for driving safety; when the vehicle is stationary, the visual channel can be used to display more detailed solution information.

[0067] User response handling strategy refers to the rules for processing when the system is waiting for user feedback, including whether user confirmation is required, how to handle timeouts, and the default handling method when no response is received.

[0068] Different scenarios require different user response handling strategies. For example, in urgent driving scenarios, the system can automatically execute the plan without user confirmation; in relaxed parking scenarios, the system can wait for the user to fully review the plan before execution.

[0069] The content decision-making process refers to the decision-making process of determining which specific adjustments should be included in a conflict resolution solution. This process, based on the travel attribute characteristics in a static scenario identifier, filters the target solution content from a set of candidate solutions to best match the current scenario. Different travel purposes correspond to different solution content preferences.

[0070] For example, in travel scenarios where punctuality is the primary goal, the plan should focus on selecting the route with the shortest travel time; in travel scenarios where safety and comfort are the primary goals, the plan should focus on selecting a smooth driving route and a safe alternative destination.

[0071] Step S150: Dynamically arrange the structured trip data based on the conflict resolution to obtain the arranged trip information.

[0072] In this step, the processor optimizes and adjusts the access order, timing, and route planning of each trip in the structured trip data according to the conflict resolution determined in step S140, and finally obtains the arranged trip information.

[0073] Dynamic orchestration refers to the process of optimizing the order and timing of a trip in real time based on current constraints and conflict resolution results. Unlike static, one-time orchestration, dynamic orchestration can continuously adjust based on real-time factors such as changes in road conditions and weather during trip execution.

[0074] The orchestrated itinerary information is the final itinerary plan after optimization, sorting, and time planning. It can include the visit order of each destination, estimated departure time, estimated arrival time, recommended route, and estimated time for each segment of the route. After completing the orchestration, the processor can send the orchestrated itinerary information to the in-vehicle terminal for display and continuously monitor and dynamically adjust it during the execution of the itinerary.

[0075] The selection of orchestration strategies can be differentiated based on the characteristics of the travel scenario. Different types of travel scenarios are suitable for different orchestration algorithms and parameter settings. The processor can maintain a set of preset orchestration strategies and select the most suitable strategy based on the scenario label when performing orchestration.

[0076] This invention implements a complete trip scheduling method. This method eliminates information silos between different trip sources through unified collection and structured processing of multi-source trip information; the introduction of scene tags enables conflict detection and resolution strategies to adaptively adjust according to actual travel scenarios; a dual decision-making mechanism of interactive and content-based decision-making processes balances driving safety and user decision-making freedom; and a dynamic scheduling mechanism allows trip plans to continuously adapt to real-time changing external conditions. Compared to existing technologies that use a uniform and fixed strategy for all trips, this invention enables differentiated and refined management for different travel scenarios, significantly improving the intelligence level of trip management and user experience.

[0077] Optionally, the sources of the multi-source trip information include at least two of the following: calendar applications, in-vehicle navigation commands, AI agent output, and user natural language input; the multi-source trip information is then processed in a structured manner to obtain structured trip data, including performing the following structured processing steps: The multi-source trip information, along with the current time context and the current location context, is input into a large language model. The large language model then performs multi-intent segmentation, interest point semantic reasoning, and relative time normalization on the multi-source trip information to obtain initial structured data. The initial structured data is subjected to duplicate travel deduplication and data field completion processing to obtain the structured travel data; wherein, the data field completion processing is used to complete the field by calling an external point of interest interface when the point of interest information is missing a preset core field; When the reasoning time of the large language model exceeds a preset time threshold, the system switches to a rule-based processing method based on regular expressions and named entity recognition to perform the structured processing on the multi-source trip information to obtain the structured trip data.

[0078] In this application, the calendar application source refers to the schedule event information stored in calendar applications installed on the user's mobile phone or vehicle's infotainment system. The processor obtains schedule events through the calendar synchronization interface and extracts events containing specific location information as trip candidate data. It should be noted that for schedule events that only contain vague descriptions and do not contain specific location information, such as events simply marked as "to be done on a certain afternoon," the processor does not synchronize them to the trip management module to avoid introducing incomplete data that cannot be used for route planning and conflict detection.

[0079] The source of in-vehicle navigation commands refers to the navigation request initiated by the user within the vehicle cabin through the basic navigation function. When the user sets a destination and initiates navigation through the navigation system, the processor automatically collects the navigation destination information as trip data.

[0080] The output sources of AI agents refer to the itinerary recommendations or planning results generated by various AI agents deployed in the cockpit system. For example, a scenic spot recommendation agent can generate sightseeing plans based on user preferences, while an itinerary planning agent can generate multi-day travel plans based on user needs. The processor obtains the itinerary information output by these agents through a pre-defined interface.

[0081] User natural language input refers to the description of travel needs directly entered by the user through voice or text. User natural language input often contains multiple travel intentions, vague location descriptions, and relative time expressions, which require the semantic understanding capabilities of a large language model for parsing.

[0082] The structured processing includes the following steps: The first step involves the processor inputting multi-source trip information along with the current time context (system's current date and time) and the current location context (vehicle's current latitude and longitude coordinates) into the large language model. The large language model performs three processes: multi-intent decomposition, which identifies and separates multiple independent trip intents from a single user input (up to five independent intents per input); point-of-interest semantic reasoning, which performs semantic analysis on ambiguous location descriptions, such as user aliases, descriptive language, or characteristic expressions, mapping them to specific points of interest; and relative time normalization, which converts relative time expressions combined with the current time into absolute timestamps. After processing, the initial structured data is obtained.

[0083] The second step involves the processor performing duplicate trip deduplication on the initial structured data, merging records with the same time, the same point of interest, and the same event, retaining the latest input, and performing data field completion processing to obtain the final structured trip data.

[0084] In addition, when the inference time of the large language model exceeds the preset time threshold, the processor automatically switches to a rule-based processing method based on regular expressions and named entity recognition for fallback processing, ensuring that the system can still complete the basic process extraction function when the model response is delayed.

[0085] Continuing with the aforementioned scenario: The user's voice input contains instructions for three destinations. After speech recognition and transcription, the processor performs intent decomposition, semantic reasoning, and time normalization on the text and the current context input large language model. Then, it performs deduplication and completion to obtain three complete structured itinerary data.

[0086] This invention enables unified collection and standardized processing of multi-source heterogeneous travel information. The large language model enables the system to understand natural language and fuzzy descriptions, and the degradation mechanism ensures the system's stability and availability under various conditions.

[0087] Optionally, the point of interest information includes at least one of the following: point of interest name, point of interest coordinates, point of interest category, and point of interest opening hours; The step of extracting scene features from the structured travel data and generating scene labels includes performing the following label generation steps: Based on the points of interest categories, the event keywords in the structured itinerary data, and the sources of the multi-source itinerary information, a static scene identifier is generated; wherein, the static scene identifier includes at least two of the following: time urgency identifier, itinerary distance span identifier, and travel purpose identifier; the time urgency identifier includes one of rigid urgency, flexible urgency, and no time constraint; the itinerary distance span identifier includes one of short-distance within the city, medium-distance between cities, and long-distance across provinces; and the travel purpose identifier includes one of business travel, family travel, medical travel, leisure travel, and daily commuting. The dynamic scene identifier is generated based on vehicle motion signals; wherein the vehicle motion signals include at least one of vehicle speed signals, gear signals, and positioning signals, and the dynamic scene identifier includes one of parking, low-speed driving, and high-speed driving.

[0088] In this application, the name of a point of interest (POI) refers to the official name or common designation of a destination, used for display in the user interface and for retrieval and matching in the POI database. The coordinates of a POI refer to the latitude and longitude geographic coordinates of the destination, used for route planning and distance calculation.

[0089] Points of interest (POIs) categories refer to the functional classification of a destination, such as transportation hubs (airports, train stations, bus stations), business and office areas (office buildings, convention centers), leisure and entertainment areas (scenic spots, restaurants, cinemas), healthcare areas (hospitals, pharmacies), and daily service areas (supermarkets, shopping malls). POI operating hours refer to the start and end times of operation of the destination, used to determine whether the arrival time of the trip falls within the valid service window.

[0090] The process of generating scene tags includes two parts: the generation of static scene tags and the generation of dynamic scene tags. The processor generates static scene identifiers based on a comprehensive analysis of point-of-interest categories, event keywords in structured travel data, and the sources of multi-source travel information.

[0091] Static scene identifiers include at least two of the following: time urgency identifier, trip distance identifier, and trip purpose identifier.

[0092] The time urgency indicator is used to characterize the sensitivity of a trip to time constraints, including three values: rigid urgency, flexible urgency, and no time constraints.

[0093] Rigidity and urgency indicate that the destination of the trip has an insurmountable deadline, and lateness will result in trip failure. The processor determines rigidity and urgency in the following ways: when the point of interest is a transportation hub and the trip is associated with known transportation information; when the keywords of the event contain words with strict time constraints, such as interviews, exams, visas, and court hearings; and when the user uses statements expressing mandatory time requirements in the input. The processor performs semantic analysis on the user input using a large language model to extract the rigidity of the time constraint, and makes a comprehensive judgment by combining the importance labels of the point of interest category and calendar events.

[0094] "Flexible urgency" indicates that the trip has a general time preference but allows for some degree of flexibility. The processor determines flexible urgency in the following ways: when the user uses vague time modifiers to describe time; when the point of interest category is a place with long operating hours (such as shopping malls, supermarkets, parks); and when the calendar event is not marked as high importance. The processor can use a large language model to score the flexibility of time modifiers in the user input, and judge it as flexible urgency when the score is lower than a preset flexibility threshold.

[0095] No time constraint means that the trip has no time limit and is entirely arranged by the system based on the gaps between other trips. The processor determines that there is no time constraint when the user only expresses the destination without mentioning any time information.

[0096] The trip distance span indicator represents the spatial range of the trip, including three values: short-distance within the city, medium-distance between cities, and long-distance across provinces. The processor determines this based on the straight-line distance from the starting point to the farthest point of interest and the estimated total travel time. Short-distance within the city indicates a trip with a short total distance and estimated total travel time, characterized by a high proportion of road condition variations affecting the total time. Medium-distance between cities indicates a trip with a moderate total distance and estimated time, characterized by the need to consider highway conditions and weather differences between cities. Long-distance across provinces indicates a trip with a long total distance or estimated time, characterized by driver fatigue and refueling being important considerations.

[0097] The travel purpose identifier is used to characterize the user's main purpose for this trip, including five possible values: business travel, family travel, medical travel, leisure travel, and daily commuting. The processor uses a large language model to perform semantic analysis on the user's input description of the event, the category of points of interest, and the source of the trip information to identify the travel purpose.

[0098] The criteria for determining business travel include: the description of the event contains keywords such as meeting, visiting clients, business trip, and signing contracts; the categories of points of interest are office buildings, convention centers, government service centers, etc.; and the source of the itinerary information is calendar events synchronized from the company email.

[0099] The criteria for determining family travel include: the description of the trip contains keywords such as bringing children, bringing the elderly, or the whole family, and the categories of points of interest are such as amusement parks, zoos, parks, and children's hospitals.

[0100] The criteria for determining medical travel include: the point of interest category being hospitals, emergency centers, pharmacies, etc., and the description of the matter containing keywords such as seeing a doctor, emergency room, vaccination, and picking up reports.

[0101] The criteria for determining leisure travel include: the point of interest is a scenic spot, restaurant, cinema, shopping mall, etc.; the description of the activity includes verbs expressing leisure activities such as play, stroll, eat, and watch; and the travel date is a non-working day.

[0102] The criteria for determining daily commuting include: the origin and destination points of the trip matching the user's historically frequently used fixed route patterns, such as fixed round trips between home and the company, and travel times falling during weekday morning and evening rush hours.

[0103] The processor generates dynamic scene identifiers based on vehicle motion signals. Vehicle motion signals include at least one of vehicle speed, gear position, and positioning signals. Dynamic scene identifiers include three values: parked, low-speed driving, and high-speed driving.

[0104] Parking indicates that the vehicle is stationary. The determination is made when the vehicle speed signal remains zero for a preset stationary duration threshold and the gear signal is in parking mode. While parked, the user can focus their attention on screen interaction; the system can display detailed information and support touch operation.

[0105] Low-speed driving indicates that the vehicle is traveling at a low speed. The determination rule is that the vehicle speed signal is continuously below a preset low-speed threshold and the gear signal is in driving gear. In low-speed driving mode, the user can briefly look at the screen, and the system should use brief information in conjunction with voice interaction.

[0106] High-speed driving indicates that the vehicle is traveling at a relatively high speed. The determination rule is that the vehicle speed signal is continuously higher than a preset low-speed threshold. When driving at high speed, the user should not be distracted, and all interactions should be primarily conducted via voice.

[0107] When each trip is created, the processor generates a scene label containing multiple identifiers based on the classification results of the aforementioned dimensions. Among them, the time urgency identifier, trip distance identifier, and trip purpose identifier are determined and persistently stored during trip creation, while the dynamic scene identifier is read in real time during conflict detection. The scene label serves as the input parameter for the conflict detection engine, driving the adaptive adjustment of subsequent detection priorities, judgment thresholds, and resolution strategies.

[0108] Continuing with the aforementioned scenario, suppose a user, on Saturday morning at 10:00 AM while parked, expresses their travel needs via voice: first visit a memorial hall, then an ancient town, and finally watch the sunset at a lake. The processor extracts scenario features from this itinerary: Regarding time urgency, the user did not express a rigid deadline, classifying it as flexible and urgent; regarding distance, the total distance calculated based on the coordinates of each destination is approximately 80 kilometers, classifying it as an intercity journey; regarding travel purpose, the current day is a non-working day, and all destinations are scenic spots, with the description containing verbs like "play" and "see," classifying it as leisure travel; regarding dynamic scenario identifiers, the vehicle is currently parked. The resulting scenario tags are: Flexible and Urgent, Intercity Journey, Leisure Travel, Parked.

[0109] This invention enables automatic identification and accurate classification of the travel scenario, providing a precise decision-making basis for subsequent conflict detection and differentiated adjustment of scheduling strategies. The multi-dimensional scenario classification system covers key features such as time, space, purpose, and real-time status, and can comprehensively characterize the features of travel scenarios.

[0110] Optionally, the plurality of conflict dimensions includes at least three of the following: time dimension, point of interest dimension, weather dimension, route dimension, and event dimension; the method for determining the conflict detection configuration includes performing at least one of the following dimension configuration steps based on the scene label: When the time urgency indicator indicates rigid urgency and the travel purpose indicator indicates business travel or medical travel, the detection priority of the time dimension is placed at the highest position among the multiple conflict dimensions, the judgment threshold of the time dimension is adjusted to a first time buffer amount that is more stringent than the default time buffer amount, and the detection frequency is adjusted to a first frequency that is higher than the default frequency. When the travel purpose identifier indicates leisure travel, the detection priority of the weather dimension is placed at the highest position among the multiple conflict dimensions, and the judgment threshold of the weather dimension is expanded to include at least one comfort judgment condition among the perceived temperature, UV index and wind force level. When the travel destination identifier indicates daily commuting, the detection priority of the route dimension is placed at the highest position among the multiple conflict dimensions, and a normal time baseline is generated based on the user's historical commuting data. The judgment threshold of the route dimension is adjusted to the deviation ratio threshold relative to the normal time baseline. When the travel purpose identifier indicates family travel, the detection priority of the weather dimension is placed at the highest position among the multiple conflict dimensions, and a suitability judgment condition is introduced in the point of interest dimension; wherein, the suitability judgment condition is used to determine whether the target point of interest corresponding to the point of interest information has at least one of the following: children's facilities, mother and baby room, and barrier-free access.

[0111] In this application, the time dimension is used to detect whether there are conflicts in the time arrangement between multiple trips, including but not limited to: the expected arrival time is later than the closing time of the point of interest, the time periods of the two trips overlap, and the interval between adjacent trips is insufficient to complete the movement.

[0112] The Points of Interest (POI) dimension is used to detect whether a destination is reachable and usable during the execution of the itinerary plan, including but not limited to: the POI being in a non-operational state, multiple POIs pointing to the same destination constituting duplication, and the real-time queue time of the POI being too long, affecting subsequent itinerary arrangements, etc.

[0113] The weather dimension is used to detect whether weather conditions at the destination or along the route will adversely affect the trip, including but not limited to: extreme weather affecting driving safety, and weather at the destination being unsuitable for outdoor activities.

[0114] The route dimension is used to detect whether there are any unreasonable aspects in the route planning of the trip, including but not limited to: the order of the trip leading to serious detours, and traffic obstruction caused by congestion or closure of road sections along the way.

[0115] The event dimension is used to detect whether there are contradictions at the semantic and logical level between multiple trips, including but not limited to: events with conflicting destinations scheduled at the same time, and trip content that is inconsistent with user intent.

[0116] For scenarios involving urgent, business, or medical travel, when the time urgency indicator signifies urgent and the travel purpose indicator signifies business or medical travel, the processor prioritizes the time dimension detection as the highest among multiple conflict dimensions. This means that during conflict detection, the time dimension detection will be performed first, and its results will be processed preferentially.

[0117] In this scenario, the processor adjusts the time-dimensional threshold to a first time buffer, which is more stringent than the default time buffer. The default time buffer refers to the safety margin the system reserves for the trip under normal circumstances. A first time buffer larger than the default time buffer means that in urgent scenarios, the system requires a more generous lead time to reduce the risk of lateness. For example, if the default time buffer is set to fifteen minutes, the first time buffer can be set to thirty minutes, meaning the estimated arrival time must be at least thirty minutes before the deadline.

[0118] Simultaneously, the processor adjusts the detection frequency to a first frequency, which is higher than the default frequency. The default frequency is the time interval at which the system performs real-time conflict monitoring under normal scenarios. A shorter interval at the first frequency means that the system checks for changes in external conditions such as road conditions and weather more frequently under pressure. For example, if the default frequency is once every five minutes, the first frequency can be set to once every two minutes.

[0119] The design logic of this configuration is that in travel scenarios with rigid deadlines and business or medical purposes, the consequences of being late are extremely serious (such as missing a flight or delaying medical treatment). Therefore, the time dimension becomes the overwhelming primary concern, requiring the strictest judgment criteria and the highest frequency of monitoring.

[0120] When the travel destination identifier indicates leisure travel, the processor prioritizes weather-related detection as the highest among multiple conflict dimensions.

[0121] In this scenario, the processor expands the weather-related threshold to include at least one comfort criterion from factors such as perceived temperature, UV index, and wind speed. This means that weather conflict detection is no longer limited to extreme weather conditions (such as heavy rain or snowstorms) but extends to meteorological conditions that affect outdoor comfort. For example, when perceived temperature exceeds the comfort range, UV index is too high, or wind speed is too strong, even if it does not pose a threat to driving safety, the system will still classify it as a weather conflict and generate adjustment suggestions.

[0122] The design logic of this configuration is that the core demand for leisure travel is the quality of experience rather than punctuality. Unsuitable weather conditions will seriously affect the experience of outdoor activities. Therefore, the weather dimension should receive the highest detection priority in leisure scenarios.

[0123] When the travel destination identifier indicates daily commuting, the processor prioritizes route dimension detection as the highest among multiple conflict dimensions.

[0124] In this scenario, the processor generates a normal commuting time baseline based on the user's historical commuting data and adjusts the route dimension's judgment threshold to a deviation percentage threshold relative to this baseline. The normal commuting time baseline refers to the normal commuting time level statistically derived from the user's recent commuting records, such as the average or median commuting time during the same period over the last thirty days. The deviation percentage threshold triggers a route conflict alarm when the actual estimated commuting time deviates from the normal commuting time baseline by a preset value. For example, when the deviation percentage threshold is set to 20%, if the normal commuting time is thirty minutes, an alarm will be triggered when the estimated commuting time reaches thirty-six minutes or more.

[0125] The design logic behind this configuration is that daily commuters are most sensitive to routes, and changes in traffic conditions directly affect their departure decisions and arrival time estimates. By establishing a personalized baseline for normal travel time, the system can accurately determine whether current traffic conditions are abnormal, avoiding unnecessary interference within normal fluctuation ranges.

[0126] When the travel purpose identifier indicates family travel, the processor prioritizes the weather dimension detection the highest among multiple conflict dimensions and introduces suitability judgment conditions in the point of interest dimension.

[0127] Suitability determination criteria is a sub-dimension of interest detection specifically introduced for family travel scenarios, used to determine whether the target interest point has the facilities to meet the needs of family members (especially children and the elderly).

[0128] Specifically, this involves determining whether the target point of interest has at least one of the following: children's facilities (such as children's play areas and children's restrooms), a mother and baby room (a dedicated space for changing diapers and breastfeeding infants), and accessible pathways (accessible facilities for wheelchairs or strollers). When the target point of interest lacks any of these facilities, the processor marks it as an appropriateness conflict.

[0129] Meanwhile, the threshold for weather-related judgments has also been adjusted accordingly for family travel scenarios. Since children and the elderly are more sensitive to weather conditions, the trigger threshold for weather-related conflicts should be lower than for ordinary leisure scenarios. For example, the criteria for judging conditions such as excessively high or low temperatures, or excessively strong winds, should be more stringent.

[0130] The design logic behind this configuration is that safety and comfort are paramount when traveling with family, especially with children. Inclement weather not only affects the experience but can also impact children's health, and the suitability of facilities at the destination directly affects the convenience of family travel. Therefore, weather and the suitability of points of interest require greater attention in family settings.

[0131] Continuing with the aforementioned scenario, in a leisure travel scenario, the processor prioritizes weather detection. The system queries the 24-hour weather forecast for the three destination areas and finds that one of them is expected to experience afternoon rainfall. Since the current scenario is leisure travel, even though light rain wouldn't constitute a serious weather conflict in a normal scenario, it would affect the outdoor experience in a leisure setting. Therefore, the system classifies it as an experience-related weather conflict and generates the corresponding conflict detection result.

[0132] This invention enables fine-tuning of conflict detection parameters under different travel scenarios. Different priority rankings and threshold parameters are used for different scenarios, allowing conflict detection to accurately match the user's actual focus in a specific scenario, avoiding the false alarms or false negatives caused by the one-size-fits-all detection strategy in existing technologies.

[0133] Optionally, performing multi-dimensional conflict detection on the structured trip data according to the conflict detection configuration determined by the scene label to obtain conflict detection results includes performing at least one of the following specific scene detection steps: When the travel purpose identifier indicates business travel or medical travel and the point of interest category is a transportation hub, the latest departure time is calculated by reverse time calculation, the allowable stay time window for intermediate trips is calculated based on the latest departure time, and then a prompt message to shorten or skip the intermediate trip is generated based on the allowable stay time window. The latest departure time and the prompt message are used as part of the conflict detection result. When the trip distance span indicates a long-distance inter-provincial journey, the driving fatigue dimension and energy replenishment dimension are added to the multiple conflict dimensions. When the continuous driving time of the vehicle exceeds a preset fatigue threshold, an unskippable rest point insertion instruction is generated. When the ratio of the vehicle's remaining driving range to the remaining trip distance is lower than a preset replenishment threshold, an energy replenishment prompt is generated. The rest point insertion instruction and the energy replenishment prompt are used as part of the conflict detection result.

[0134] When the travel destination identifier indicates business or medical travel and the point of interest category is transportation hub, the processor performs conflict detection using a reverse time-based calculation method. Transportation hub points of interest include locations with transportation schedule constraints, such as airports, train stations, and bus stations.

[0135] The reverse time estimation method refers to starting from the departure time of the transportation service, gradually subtracting each time consumption item in reverse, and finally calculating the user's latest departure time, and assessing whether there is any time risk in the current travel arrangement.

[0136] The specific calculation process is as follows: The processor first obtains the travel time associated with the trip, such as flight departure time or train departure time, and then subtracts the buffer time of the corresponding transportation hub type to obtain the latest time that the user must arrive at the transportation hub.

[0137] Buffer times are set differently depending on the specific type of transportation hub. For example, airport trips require a longer buffer time for check-in and security checks, train station trips require a medium buffer time for entry and security checks, and bus station trips require a shorter buffer time.

[0138] The processor then subtracts the estimated travel time from the current location to the transportation hub from the latest arrival time to obtain the latest departure time.

[0139] When a trip includes intermediate stops, where the user needs to visit other locations before heading to a transportation hub, the processor needs to further calculate the dwell time window for each intermediate stop.

[0140] The dwell time window refers to the maximum length of time a user can stay at a certain intermediate point of interest. It is calculated as follows: the latest departure time minus the current time, minus the travel time required to reach the intermediate point of interest, and minus the travel time required to travel from the intermediate point of interest to the transportation hub.

[0141] When the available time window is insufficient for the user to complete their scheduled activity at an intermediate point of interest, the processor generates a prompt message to shorten or skip that intermediate journey.

[0142] For example, when the permitted stay time window is lower than a preset urgency alert threshold (e.g., 15 minutes), the system generates a voice alert to inform the user that time is running out and suggests shortening the stay; when the permitted stay time window drops to zero or a negative value, the system suggests skipping this intermediate journey and going directly to the transportation hub. The latest departure time and alert information are output as part of the conflict detection results.

[0143] When the trip distance span indicates a long-distance inter-provincial trip, the processor adds the driver fatigue dimension and energy replenishment dimension to the conflict detection dimension set in addition to the usual five conflict dimensions.

[0144] The driver fatigue dimension is used to detect whether there is a safety risk of the driver driving continuously for too long. The processor monitors the continuous driving time of the vehicle since the start of the trip, which can be defined as the cumulative driving time when the vehicle is in continuous motion and has not stopped at service areas or other locations for more than a preset rest time. When the continuous driving time exceeds the preset fatigue threshold, the processor generates a non-skippable rest point insertion instruction.

[0145] The preset fatigue threshold can be set to two hours. The rest stop insertion instruction requires the mandatory insertion of a rest stop along the route, such as a highway service area, in the trip planning. This instruction is marked as a safety mandatory constraint and cannot be canceled or skipped by the user.

[0146] Meanwhile, the processor uses a voice synthesis module to remind the driver of the duration of continuous driving and suggests stopping for a rest stop ahead.

[0147] The energy replenishment dimension is used to detect whether the vehicle's remaining range is sufficient to cover the remaining distance. The processor estimates the remaining range by reading the vehicle's energy status information and compares it with the remaining distance.

[0148] When the ratio of estimated remaining driving range to remaining distance is lower than a preset replenishment threshold, the processor generates an energy replenishment prompt. The preset replenishment threshold has two levels: when the ratio is lower than the first replenishment threshold, the processor generates a normal level of refueling or charging recommendation; when the ratio further decreases to below the second replenishment threshold, the processor generates an emergency warning and automatically searches for the nearest replenishment station along the route and inserts it into the trip planning.

[0149] Rest stop insertion instructions and energy replenishment prompts are output as part of the conflict detection results, and are used in subsequent conflict resolution and itinerary scheduling steps.

[0150] Let's continue with the long-distance self-driving example from the previous scenario. Suppose a user plans a self-driving trip with a total distance of approximately 300 kilometers, and the distance is identified as a long-distance inter-provincial trip. In addition to performing the standard five-dimensional conflict detection, the processor adds a driver fatigue dimension and an energy replenishment dimension.

[0151] The system estimates the total driving time to be about four hours and automatically generates a mandatory rest stop instruction at a service area about two hours into the trip, roughly in the middle of the journey.

[0152] Simultaneously, the system reads the vehicle's battery state of charge and estimates that the remaining battery level will be less than 20% upon arrival at the destination, and the ratio of remaining range to distance will be below the first replenishment threshold. Therefore, it generates charging recommendations, suggesting fast charging at rest service areas. The system further optimizes the plan by merging rest stops and charging points into the same service area, reducing the number of additional stops.

[0153] This invention expands upon conventional conflict detection dimensions by providing specialized detection capabilities for specific scenarios. The reverse time estimation mechanism in transportation hub scenarios can accurately calculate users' remaining time and provide timely warnings when time is insufficient. The fatigue and resupply detection mechanisms in long-distance driving scenarios can proactively ensure driving safety and travel endurance. These specialized detection dimensions further enhance the comprehensiveness and practicality of conflict detection.

[0154] Optionally, determining conflict solutions from the set of candidate solutions through an interactive decision-making process and a content decision-making process includes performing at least one of the following interactive configuration steps during the interactive decision-making process: When the time urgency indicator indicates rigid urgency and the dynamic scenario indicator indicates low-speed or high-speed driving, the output channel is limited to a voice channel, and the user response processing strategy is configured to automatically execute a single conflict resolution. When the time urgency indicator indicates rigid urgency and the dynamic scene indicator indicates parking, the output channel is configured as a visual channel, the number of conflict solutions is configured to be no less than two, and the user response processing strategy is configured to automatically select a preset default solution when no user selection is obtained within the first timeout period. When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scenario indicator indicates low-speed driving or high-speed driving, the user response processing strategy is configured to mark the corresponding conflict as pending when no user response is obtained within the second timeout period, and to prompt again when the dynamic scenario indicator changes to parking. When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scene indicator indicates parking, the output channel is configured as a visual channel, the conflict solution is configured to include complete conflict information and alternative point of interest information, and the user response processing strategy is configured to support user modification and order adjustment operations.

[0155] When the time urgency indicator indicates rigid urgency and the dynamic scenario indicator indicates low-speed or high-speed driving, the processor will limit the output channel to the voice channel and configure the user response processing strategy to automatically execute a single conflict resolution.

[0156] In this scenario, the user is driving and facing tight time constraints, making it a high-risk situation in terms of both safety and time. The processor selects the voice channel as the sole output method, avoiding visual pop-ups on the in-vehicle display to prevent distracting the driver.

[0157] Meanwhile, the user response handling strategy is configured to automatically execute the system's recommended optimal solution without user confirmation. The system uses a speech synthesis module to briefly inform the user of the adjustments that have been made, such as announcing that the user has been switched to the fastest route and the estimated arrival time.

[0158] In emergency situations involving driving, requiring user interaction poses a safety hazard, and the time spent waiting for a response further reduces the already limited time available. The system should make decisive, automatic decisions and execute the most conservative solution as quickly as possible.

[0159] When the time urgency indicator indicates rigid urgency and the dynamic scene indicator indicates parking, the processor will configure the output channel as the visual channel, configure the number of conflict solutions to be no less than two, and configure the user response handling strategy to automatically select the preset default solution if no user selection is obtained within the first timeout period.

[0160] In this scenario, although the user faces a tight time constraint, the vehicle is stationary, allowing the user to view the solution details on the screen. The processor displays a conflict pop-up window on the in-vehicle display screen via the visual channel, showing two or three candidate solutions. Each solution includes key comparative information, and the user can click to view a route map preview of each solution.

[0161] The first timeout period refers to the maximum waiting time the system can take for the user to make a choice among multiple options. Considering the time constraints, the first timeout period should be set to a short value, such as thirty seconds. If the user has not made a choice by the end of the first timeout period, the system will automatically select the default preset option to execute. The default preset option is usually the one with the shortest execution time or the lowest risk.

[0162] When the car is parked, users have the attention span to compare and decide on options, and offering multiple options provides them with a degree of decision-making freedom. However, considering the time constraints, a timeout fallback mechanism is implemented to ensure that the plan is not delayed due to user hesitation.

[0163] When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scenario indicator indicates low-speed driving or high-speed driving, the processor will configure the user response handling strategy to mark the corresponding conflict as pending if no user response is obtained within the second timeout period, and to prompt again when the dynamic scenario indicator changes to parking.

[0164] In this scenario, the urgency of the conflict is relatively low, but users are unable to perform complex interactions while driving. For low-risk conflicts (such as minor detours or light rain warnings), the processor can handle them automatically and provide a brief voice notification; for high-risk conflicts (such as extreme weather or destination closures), the processor will announce the key points of the conflict and two candidate solutions via voice, allowing users to choose through voice responses.

[0165] The second timeout period refers to the maximum waiting time for the system to receive a user's voice response, which can be set to fifteen seconds. If the user does not respond after the second timeout period, the system will not execute the plan but will instead mark the conflict as pending and store it in the local database. When the subsequent dynamic scene indicator changes to "parking," the system will re-display a visual pop-up to remind the user to handle the pending conflict.

[0166] Conflicts in non-emergency scenarios do not immediately cause irreversible consequences, so they can be safely delayed until they are convenient for the user to handle, avoiding unnecessary interaction interference during driving.

[0167] When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scene indicator indicates parking, the processor configures the output channel as the visual channel, configures the conflict solution to include complete conflict information and alternative points of interest information, and configures the user response handling strategy to support user modification and order adjustment operations.

[0168] In this scenario, with ample time and the vehicle stationary, it is the safest and most suitable setting for users to make thorough decisions. The processor displays a comprehensive interactive interface on the in-vehicle screen, showcasing a detailed description of the conflict, a complete comparison of multiple candidate solutions, and details of alternative points of interest. Users can freely modify their itinerary, manually drag and drop to adjust the order of visits to points of interest, and view detailed information on each alternative solution. No timeout is set in this scenario, allowing users to make decisions after careful consideration.

[0169] The rationale behind this configuration is that the non-emergency and parked state is the safest and most relaxed scenario, and users should be given the greatest freedom of decision-making and the richest information presentation so that they can make the most satisfactory choice.

[0170] This invention dynamically adjusts the presentation of conflict solutions and user response handling rules based on different combinations of urgency and motion state, achieving an optimal balance between driving safety and user decision-making freedom. In emergency scenarios, the system makes decisive and automatic decisions to save time, while in relaxed scenarios, the system fully displays information to respect user preferences. This differentiated interaction strategy significantly improves the human-machine interaction experience in in-vehicle scenarios.

[0171] Optionally, the content decision process includes: When the travel purpose identifier indicates business travel, the candidate solution set is limited to the fastest route strategy type, and the trip cancellation type is excluded from the candidate solution set; When the travel purpose identifier indicates family travel, alternative points of interest that do not have children's facilities are filtered from the candidate solution set, and a soft time window is reserved in the time arrangement of the target solution content; wherein, the soft time window includes lunch time and children's nap time; When the travel purpose identifier indicates leisure travel, an option to adjust the itinerary order is added to the candidate solution set; wherein, the option to adjust the itinerary order is used to exchange the order of outdoor and indoor itineraries according to weather distribution; When the travel destination identifier indicates daily commuting, a route preference set is generated based on the user's historical commuting routes, and the target route content is determined from the route preference set; When the travel destination identifier indicates medical travel, the candidate solution set is limited to the fastest route strategy type, and when the time taken for the regular route exceeds a preset value, an alternative hospital recommendation type is added to the candidate solution set.

[0172] When the travel destination identifier indicates business travel, the processor limits the candidate solution set to the fastest route strategy, including switching to the real-time fastest route and departing in advance to avoid congestion.

[0173] The processor excludes trip cancellation solutions from the candidate solution set because trip cancellation typically incurs high business costs in business scenarios.

[0174] When the conflict is a time conflict, the processor prioritizes generating a solution that starts earlier and uses the fastest route.

[0175] If a meeting is detected in the calendar, alternative remote participation options can be generated. For route conflicts, the fastest route will be forced, and scenic or shortest routes will not be recommended.

[0176] When the travel destination indicates a family trip, the processor filters alternative points of interest (POIs) from the candidate solution set that lack child-friendly facilities. In other words, when the system recommends alternative destinations, it automatically excludes places without child-friendly amenities. Simultaneously, it reserves soft time windows in the timing of the target solution's content, including lunchtime and children's nap time.

[0177] The processor automatically identifies these time windows during scheduling and avoids scheduling long-distance travel within them. For example, it allocates lunch time between 11:30 AM and 1:00 PM, and nap time for children between 1:00 PM and 3:00 PM. For weather-related conflicts, the solution prioritizes alternative indoor points of interest, automatically filtering locations without child-friendly facilities when searching for alternatives. For route conflicts, the solution favors smooth driving routes, avoiding routes with frequent curves or road changes to reduce the risk of passenger discomfort.

[0178] When the trip destination identifier indicates leisure travel, the processor adds an option to adjust the trip order to the candidate solution set. This option allows for the exchange of outdoor and indoor trips based on weather distribution.

[0179] Specifically, when the system detects that the weather is unsuitable for outdoor activities during a certain period, it can suggest swapping outdoor activities originally scheduled for that period with indoor activities scheduled for a better weather period, thereby optimizing weather adaptability without canceling any activities.

[0180] For conflicts of interest, the solution prioritizes alternative attractions with shorter queues or during less popular times, providing user reviews and popularity indexes to aid decision-making. For conflicts of route, in addition to the fastest route, the solution also offers route options prioritizing scenery, indicating scenic ratings and available viewpoints along the way.

[0181] When the travel destination identifier indicates daily commuting, the processor generates a set of route preferences based on the user's historical commuting routes, recording the frequency and condition preferences of the user's route selection at different times and under different weather conditions in the past thirty days.

[0182] When frequently used routes become congested, the processor selects the user's second-most-preferred route from the route preference set as an alternative, rather than recommending a new, unfamiliar route, and only switches the result via voice notification. The solution for time conflicts in commuting scenarios primarily involves pushing accurate estimated arrival times. If the delay exceeds fifteen minutes and the user has a meeting scheduled in their calendar, the system additionally suggests sending a delay notification.

[0183] When the travel destination identifier indicates medical travel, the processor will limit the set of candidate solutions to the fastest route strategy and will not recommend postponement or cancellation.

[0184] When the time taken for the regular route exceeds a preset value, the processor adds alternative hospital recommendations to the candidate solution set, actively searching for nearby alternative hospitals with the same departments. In emergency scenarios, the system automatically selects the fastest route without any user confirmation.

[0185] This invention provides customized conflict solutions based on the characteristics of different travel purposes, making the recommended solutions more aligned with users' actual needs and preferences in specific scenarios.

[0186] Optionally, when the conflict detection result includes at least two conflicts existing simultaneously across multiple conflict dimensions, determining the conflict solution from the candidate solution set through an interactive decision-making process and a content decision-making process includes performing the following composite conflict resolution steps: When at least two conflicts include driving safety-related conflicts, the processing priority of the driving safety-related conflicts is placed at the highest level, and the intervention of the scene label is blocked; wherein, the driving safety-related conflicts include at least one of the combination of extreme weather and high-speed driving and continuous fatigue driving; When the solution generated for the first conflict among the at least two conflicts would trigger a second conflict, the associated impact information of the second conflict shall be output together with the conflict solution. When the number of same-day trips corresponding to the structured trip data exceeds a first threshold and the number of at least two conflicts exceeds a second threshold, at least two global orchestration schemes are generated at once based on the large language model, which integrates the multiple conflict dimensions and the constraints of the structured trip data.

[0187] In this application, when at least two conflicts include driving safety-related conflicts, the processor prioritizes the driving safety-related conflicts and disables the scene label's interference with the conflict processing priority.

[0188] Driving safety-related conflicts include combinations of extreme weather and high-speed driving, as well as continuous fatigue driving. Extreme weather refers to weather conditions that severely affect driving safety, such as heavy rain, heavy snow, dense fog, and icy roads. Continuous fatigue driving refers to a driver's continuous driving time exceeding the safety threshold without rest. The intervention of blocking scene labels means that regardless of the current trip's scene label value, driving safety-related conflicts are always handled with the highest priority.

[0189] For example, even if the current scenario is leisure travel and the weather dimension has the highest priority under normal logic, when the detected weather conflict escalates to the driving safety level, the handling method no longer follows the experience-first logic of leisure scenarios, but directly adopts a safety-first handling strategy. This design reflects the basic principle of safety first.

[0190] When a solution generated for the first conflict would trigger a new second conflict, the processor outputs the associated impact information of the second conflict along with the conflict solution. This associated impact information describes the potential cascading effects that may occur after resolving the first conflict.

[0191] For example, when the solution to a weather conflict involves changing the destination of an outdoor trip to an alternative indoor point of interest, if there are route detours between this alternative point of interest and other destinations in the existing trip, the processor will inform the user of the increased distance or time due to the changed destination when presenting the solution. The purpose of this step is to avoid introducing another, more serious problem while resolving one conflict, allowing users to fully understand the pros and cons of each solution when making a decision.

[0192] When the number of same-day trips corresponding to structured trip data exceeds a first threshold and the number of at least two conflicts exceeds a second threshold, the processor triggers a global batch optimization mode. The first threshold can be set to five, and the second threshold can be set to three. In batch optimization mode, the processor inputs all trip information, all conflict information, and all constraints into a large language model. The large language model then comprehensively considers all factors and generates at least two globally optimal orchestration schemes at once. Each scheme includes information on which trips were adjusted, the specific reasons for each adjustment, and a comparison of key differences between the schemes, facilitating quick understanding and selection by the user.

[0193] For example, a user schedules six trips on a certain day, and conflict detection reveals three conflicts: one time conflict, one weather conflict, and one route conflict. The number of trips on the same day (six) exceeds the first threshold of five, and the number of conflicts (three) reaches the second threshold of three. The processor then triggers batch optimization mode.

[0194] The large language model integrates all constraints and conflicts, and generates two global solutions at once: Solution 1 focuses on optimizing time efficiency, and Solution 2 focuses on optimizing user experience quality. Each solution is marked with specific adjustments and their reasons.

[0195] Through the aforementioned composite conflict resolution mechanism, this invention ensures the inviolability of safety baseline constraints, avoids chain reactions in the conflict resolution process, and improves global optimization efficiency in dense travel scenarios.

[0196] Optionally, the step of dynamically orchestrating the structured itinerary data based on the conflict resolution to obtain orchestrated itinerary information includes: Extract the trip distance span identifier, the trip purpose identifier, the time urgency identifier, and the number of trips from the structured trip data; Based on the extracted feature data, the target arrangement strategy is determined from the preset spatial clustering arrangement strategy, the approximate shortest path arrangement strategy, the historical route learning arrangement strategy, and the endpoint backward arrangement strategy. According to the target orchestration strategy, route planning and time node calculation are performed on the structured trip data to obtain the orchestrated trip information; wherein, the process of determining the target orchestration strategy includes at least one of the following processing branches: When the trip distance span identifier indicates a short trip within the city and the number of trips exceeds a preset number, the target scheduling strategy is determined to be the spatial clustering scheduling strategy; wherein, the spatial clustering scheduling strategy is used to cluster based on the geographical location corresponding to the point of interest information, and uses points of interest with clear time requirements as fixed anchor points; When the trip distance span identifier indicates a long-distance inter-provincial trip or an intercity trip and the trip destination identifier indicates a leisure trip, the target scheduling strategy is determined as the approximate shortest path scheduling strategy; wherein, the approximate shortest path scheduling strategy is used to solve for the approximate shortest path for the nodes corresponding to the point of interest information, and to forcibly insert rest stops and energy replenishment points into the scheduled trip information; When the travel destination identifier indicates daily commuting, the target scheduling strategy is determined as the historical route learning scheduling strategy; wherein, the historical route learning scheduling strategy is used to generate a route preference probability distribution based on the user's historical commuting data, and select a travel route from the route preference probability distribution in combination with real-time traffic conditions; When the time urgency indicator indicates rigid urgency, the target scheduling strategy is determined as the endpoint backward scheduling strategy; wherein, the endpoint backward scheduling strategy is used to calculate the latest departure time and the available stay time window for each point of interest in the structured itinerary data in reverse from the deadline time.

[0197] First, the processor extracts trip distance span identifiers, trip purpose identifiers, time urgency identifiers, and the number of trips from the structured trip data to form the feature data required for scheduling decisions.

[0198] Then, based on the extracted feature data, the processor determines the target orchestration strategy from the preset spatial clustering orchestration strategy, the approximate shortest path orchestration strategy, the historical route learning orchestration strategy, and the endpoint backward orchestration strategy.

[0199] Finally, the processor performs route planning and time node calculations on the structured trip data according to the target orchestration strategy to obtain the orchestrated trip information. Route planning can be achieved by calling the route planning interface of the map service, and the route preference parameters are adaptively set according to the scene labels.

[0200] Time node calculation refers to calculating the estimated arrival and departure times of each point of interest based on the estimated travel time of each route segment and the expected stay time at each point of interest.

[0201] When the trip distance span indicates short-distance trips within the city and the number of trips exceeds the preset limit, the processor will determine the target orchestration strategy as a spatial clustering orchestration strategy. The preset limit can be set to three.

[0202] Spatial clustering orchestration strategy is an orchestration method that optimizes the access order of points of interest based on geographic location clustering. Its core idea is to group geographically close points of interest together, allowing consecutive visits to points of interest within the same group, and connecting groups in the optimal direction of the overall route, thereby avoiding inefficient routes that require traversing across areas and then turning back in congested urban areas.

[0203] The specific implementation is as follows: The processor first performs spatial clustering on all points of interest (POIs) based on their geographic coordinates. The clustering algorithm can employ density-based clustering, using a preset neighborhood radius as the clustering parameter to group POIs with similar geographical locations into the same cluster. POIs within the same cluster are visited in the order of proximity, and different clusters are connected sequentially according to the optimal overall route.

[0204] When performing spatial clustering orchestration, the processor also needs to consider time window constraints. Points of interest with specific time requirements serve as fixed anchors, and their access times cannot be arbitrarily adjusted. Other points of interest are arranged around these fixed anchors. For example, if a point of interest requires arrival at 2 PM, then the access time of the cluster to which that point of interest belongs needs to cover 2 PM, and other points of interest are arranged in the time gaps before and after that anchor.

[0205] To illustrate with a specific scenario: Suppose a user has five in-city trips, each heading to one of five points of interest (POIs) located in different areas of the city. One of the POIs requires arrival at 2 PM. The processor performs spatial clustering, dividing the five POIs into three clusters based on their geographical location. The cluster containing the POI requiring 2 PM arrival is scheduled for the corresponding afternoon time slot. The other clusters are arranged sequentially based on their spatial relationship with the current cluster and the optimal route direction. Within each cluster, POIs are arranged in proximity order. Individual POIs not belonging to any cluster are inserted into the most suitable time slot.

[0206] When the trip distance span identifier indicates a long-distance inter-provincial trip or an intercity trip, and the trip destination identifier indicates a leisure trip, the processor will determine the target orchestration strategy as the approximate shortest path orchestration strategy.

[0207] The approximate shortest path arrangement strategy is an arrangement method that draws on the solution approach of the Traveling Salesman Problem. Its core idea is to treat all destination cities or scenic spots as nodes in graph theory, and use the travel distance or time between nodes as edge weights to find the approximate shortest access path.

[0208] The specific implementation method is as follows: the processor takes the current position as the starting point, takes all destination points of interest as nodes to be visited, and uses the greedy nearest neighbor algorithm to generate an initial path, that is, starting from the current node, the nearest unvisited node is always selected, and then the path is improved through local optimization methods to reduce the total travel distance.

[0209] Based on the approximate shortest path arrangement, the processor also needs to insert mandatory constraints. These constraints include: a rest stop must be inserted after each preset driving duration; the total daily driving time must not exceed a preset limit; and no new outdoor scenic area trips will be scheduled after sunset. Furthermore, if the estimated end time of the last trip of the day is late, the processor automatically recommends accommodation information near the last point of interest for the day and inserts it into the scheduling scheme. Energy replenishment points also need to be inserted at appropriate locations based on the vehicle's remaining range.

[0210] When the destination identifier indicates daily commuting, the processor determines the destination orchestration strategy as the historical route learning orchestration strategy.

[0211] The core of the historical route learning and scheduling strategy lies not in the ranking of points of interest, but in the intelligent selection of routes. Based on users' historical commuting data, the processor learns the user's route selection patterns at different times and under different weather conditions, forming a probability distribution of route preferences.

[0212] The route preference probability distribution records the user's historical selection frequency and conditional preferences for each candidate commuting route. For example, users most often choose route A during the morning rush hour on sunny days, followed by route B; on rainy days, they tend to choose route C. At departure, the processor selects the route with the optimal estimated travel time from the route preference probability distribution based on the current real-time traffic information and weather conditions.

[0213] When a user has an additional task during their commute, such as buying breakfast or picking up a package, the processor searches for points of interest that meet the criteria within a preset range along the frequently used route and selects the one that deviates the least from the frequently used route to insert into the commute route.

[0214] When the time urgency indicator indicates rigid urgency, the processor will determine the target orchestration strategy as the endpoint backward orchestration strategy.

[0215] The destination-based reverse scheduling strategy is a scheduling method that uses the deadline as a starting point for backward time planning. Its core idea is to start from the final destination's deadline, progressively subtracting the travel time and dwell time for each segment in reverse, calculating the latest departure time and available dwell time window for each intermediate point of interest, thereby determining whether the entire itinerary chain is time-feasible.

[0216] The specific implementation is as follows: The processor takes the deadline as the starting point, subtracts the buffer time and the estimated time for the last leg of the journey, to obtain the latest departure time from the last intermediate point of interest. This process is repeated segment by segment in reverse to obtain the latest departure time for each intermediate point of interest. The dwell time window equals the latest departure time minus the estimated arrival time. When the dwell time window for any point of interest is lower than a preset minimum dwell threshold, the processor suggests skipping that point of interest or replacing it with a more convenient point of interest of the same type along the route.

[0217] After the routing is completed, the processor marks the entire trip chain as a time-sensitive identifier, triggering high-frequency real-time monitoring. Route planning adopts the fastest arrival strategy, prioritizing the route with the shortest travel time and actively avoiding congested sections.

[0218] After the itinerary is arranged, the processor generates standardized itinerary card information, adding scene-aware display content in addition to basic information. For example, the top of the card for rigid and urgent itineraries displays countdown information until the latest departure time; the card for family travel itineraries is marked with child-friendly facility icons; the card for long-distance inter-provincial travel itineraries marks the locations of rest stops and energy supply points; and the card for leisure travel itineraries displays interest ratings and queue estimation information, etc.

[0219] This invention achieves differentiated scheduling optimization for different travel scenarios. For densely packed urban journeys, spatial clustering reduces unnecessary travel; for long-distance travel, near-shortest paths combined with mandatory rest and resupply ensure safety and efficiency; for daily commutes, historical route learning enables personalized route recommendations; and for time-sensitive scenarios, reverse-engineering the destination precisely manages every minute of remaining time. This multi-strategy collaborative scheduling system offers significant advantages in scheduling quality compared to the single proximity-based sorting strategy in existing technologies.

[0220] Optionally, after obtaining the arranged itinerary information, the method further includes performing the following itinerary assistance service steps: When a trip interruption event is detected, the unfinished trips in the structured trip data are queried and real-time status detection is performed to obtain the real-time status detection results; When the real-time status detection result indicates no change, the interruption point navigation information is loaded and navigation is automatically resumed if no user response is obtained after a timeout; when the real-time status detection result indicates a change, the conflict resolution is regenerated and navigation resumes after obtaining user confirmation. During the trip, the triggering time and scope of recommended service information are determined based on the scenario tags; wherein, when the travel destination identifier indicates business travel, the scope of recommended content is limited to destination parking and business assistance information; when the travel destination identifier indicates daily commuting, the regular service recommendation function is turned off and replenishment service information is generated when the vehicle's remaining energy is lower than a preset value.

[0221] Trip interruption events may occur during the execution of the trip. Trip interruption events refer to events that cause the ongoing trip to be temporarily interrupted, including but not limited to: the user getting off the vehicle midway through the trip and the vehicle's infotainment system being turned off, the need to connect to the next trip after the previous trip has reached its destination, and the need to continue the next day for a trip that spans multiple days.

[0222] When the processor detects a trip interruption event, it first queries the structured trip data for incomplete trips with an ongoing status. The query scope is recent trip records cached in the local database.

[0223] Upon detecting an incomplete trip, the processor performs real-time status detection and obtains the results. Real-time status detection involves requesting the latest environmental status information for the incomplete trip from the cloud server, including road conditions along the route, destination weather information, and the operational status of points of interest at the destination. The processor compares the latest status returned with the status at the time the trip was interrupted to determine if any changes have occurred that affect the feasibility of the trip.

[0224] When the real-time status detection result indicates no change, the processor loads the navigation information from the interrupted point and restores the navigation state to its pre-interruption state. The system briefly prompts the user via the voice synthesis module that it is resuming the previous trip and asks if they wish to continue navigation. If the user does not respond after a preset response waiting time, the processor automatically resumes navigation to ensure the trip is not delayed due to waiting.

[0225] When real-time status detection results indicate a change, the processor re-executes the conflict detection and resolution process based on the changed status information, generating a new conflict resolution solution. The processor simultaneously informs the user of the detected abnormal change and the recommended new solution through speech synthesis and the in-vehicle display screen.

[0226] After obtaining user confirmation, the processor continues navigation according to the new plan. If a feasible plan cannot be generated, the processor marks the incomplete trip as pending, waiting for subsequent changes in conditions or manual intervention by the user.

[0227] For a trip chain containing multiple trips, the processor sequentially connects them according to the order in the arranged trip information. Upon reaching the current destination, the system automatically displays information about the next trip via a floating window and voice prompts. After user confirmation, the next segment of navigation begins.

[0228] During the trip, the processor determines the timing and scope of service recommendations based on scenario tags. Service recommendations refer to suggestions for related services relevant to the current trip scenario, such as parking recommendations, restaurant recommendations, ticket recommendations, and refueling / charging recommendations.

[0229] Triggering timing refers to the time point at which service recommendation information is displayed to the user, which can include three stages: pre-trial triggering, in-trial triggering, and post-trial triggering. The scope of recommended content refers to the service categories and information richness covered by the service recommendation information.

[0230] When the travel destination is identified as business travel, the processor will limit its recommended content to destination parking and business support information. Business support information includes parking reservations near the destination, coffee delivery services, and other services directly related to business activities. The processor will not recommend dining and entertainment services to avoid interfering with the user's experience in a business setting.

[0231] When the travel destination indicates daily commuting, the processor disables regular service recommendations by default. Users have the lowest tolerance for service recommendations during commutes, and frequent recommendations are considered intrusive. The processor only generates replenishment service information and recommends gas stations or charging stations along the route when the vehicle's remaining energy is below a preset value. Furthermore, when a user actively inquires via voice, the processor responds to the user's query and provides corresponding service information.

[0232] When the travel destination indicates a family trip, the pre-trip recommendations include family package information and a selection of child-friendly restaurants. During the trip, in addition to parking guidance, information on facilities such as mother-and-baby rooms and children's rest areas is added. Recommended alternative points of interest and restaurants have all been screened for child-friendliness.

[0233] When the travel destination is identified as leisure travel, the recommended content is the most comprehensive. Before the trip, it provides complete information on tickets, restaurants, and accommodations; during the trip, it offers audio guides and information on popular tourist attractions; and after the trip, it recommends nearby restaurants and dessert shops. The recommendation frequency is relatively flexible, allowing a maximum of two recommendations per point of interest.

[0234] When the travel destination is identified as medical, the system recommends real-time parking availability at the destination hospital before the trip. No non-medical related services are recommended during or after the trip.

[0235] All service recommendations offer a one-click operation button and support voice command operation. Booking information can be synchronized to the mobile device in real time to generate electronic vouchers. When a user rejects a certain type of recommendation, the system will not trigger similar recommendations on the same day to avoid repeated interference. The system can also record the user's recommendation acceptance rate in different scenarios and continuously optimize the recommendation strategy and frequency settings for each scenario.

[0236] This invention achieves intelligent management capabilities throughout the entire trip lifecycle, building upon trip planning. The intelligent seamless navigation mechanism resolves the issue of user experience disruption caused by the need for manual navigation reset after trip interruptions, ensuring the continuity of trip execution. The scenario-based proactive service recommendation mechanism precisely controls the recommended content and frequency based on the travel purpose, providing rich value-added services in scenarios requiring them and remaining restrained in scenarios where interference is unnecessary, thus achieving precise and user-friendly service recommendations.

[0237] The travel scheduling device provided by the present invention is described below. The travel scheduling device described below and the travel scheduling method described above can be referred to in correspondence.

[0238] Figure 2 This is a schematic diagram of the travel scheduling device provided by the present invention, as shown below. Figure 2 As shown, it includes: The acquisition module 210 is used to acquire multi-source travel information and perform structured processing on the multi-source travel information to obtain structured travel data; wherein, the structured travel data includes point of interest information and time information; The generation module 220 is used to extract scene features from the structured trip data and generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The detection module 230 is used to perform multi-dimensional conflict detection on the structured travel data according to the conflict detection configuration determined by the scene label, and obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold and detection frequency; The determination module 240 is used to generate a set of candidate solutions based on the conflict detection results, and to determine a conflict solution from the set of candidate solutions through an interactive decision-making process and a content decision-making process; wherein, the interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution; the content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier; The orchestration module 250 is used to dynamically orchestrate the structured trip data based on the conflict resolution to obtain orchestrated trip information.

[0239] In this application, by structuring multi-source trip information and extracting scene features to generate scene tags containing static and dynamic scene identifiers, the system can perceive the travel scenario to which the current trip belongs before conflict detection. Then, based on the scene tags, it dynamically determines the conflict detection configuration, such as the detection priority ranking, judgment threshold, and detection frequency for multiple conflict dimensions. This allows multi-dimensional conflict detection to adopt differentiated detection strategies for different travel scenarios, improving the timeliness and accuracy of key conflict identification. Simultaneously, in the conflict resolution stage, by cross-combining static and dynamic scene identifiers to determine the output channel and user response processing strategy, and by filtering target solution content based on trip attribute features, the conflict resolution solution can adapt to both the safety interaction requirements of the current driving state and the actual concerns of the user under specific travel purposes. Based on this, the trip is dynamically orchestrated, thereby achieving refined and adaptive trip conflict management and orchestration for different travel scenarios.

[0240] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 3 As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340. The processor 310, communications interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute a trip scheduling method. This method includes: acquiring multi-source trip information and performing structured processing on the multi-source trip information to obtain structured trip data; wherein the structured trip data includes point-of-interest information and time information. Scene features are extracted from the structured trip data to generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The structured travel data is subjected to multi-dimensional conflict detection according to the conflict detection configuration determined by the scene label to obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold, and detection frequency; Based on the conflict detection results, a set of candidate solutions is generated, and a conflict solution is determined from the set of candidate solutions through an interactive decision-making process and a content decision-making process. The interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution. The content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier. The structured itinerary data is dynamically rearranged based on the conflict resolution method to obtain rearranged itinerary information.

[0241] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0242] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program, the computer program being able to be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer is able to execute the trip scheduling method provided by the above methods, the method including: acquiring multi-source trip information and performing structured processing on the multi-source trip information to obtain structured trip data; wherein, the structured trip data includes point of interest information and time information; Scene features are extracted from the structured trip data to generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The structured travel data is subjected to multi-dimensional conflict detection according to the conflict detection configuration determined by the scene label to obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold, and detection frequency; Based on the conflict detection results, a set of candidate solutions is generated, and a conflict solution is determined from the set of candidate solutions through an interactive decision-making process and a content decision-making process. The interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution. The content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier. The structured itinerary data is dynamically rearranged based on the conflict resolution method to obtain rearranged itinerary information.

[0243] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the trip scheduling method provided by the above methods, the method comprising: acquiring multi-source trip information and performing structured processing on the multi-source trip information to obtain structured trip data; wherein the structured trip data includes point of interest information and time information; Scene features are extracted from the structured trip data to generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The structured travel data is subjected to multi-dimensional conflict detection according to the conflict detection configuration determined by the scene label to obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold, and detection frequency; Based on the conflict detection results, a set of candidate solutions is generated, and a conflict solution is determined from the set of candidate solutions through an interactive decision-making process and a content decision-making process. The interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution. The content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier. The structured itinerary data is dynamically rearranged based on the conflict resolution method to obtain rearranged itinerary information.

[0244] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0245] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0246] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; 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; and these 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 scheduling travel routes, characterized in that, include: Multi-source travel information is acquired and structured to obtain structured travel data; wherein, the structured travel data includes point-of-interest information and time information; Scene features are extracted from the structured trip data to generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The structured travel data is subjected to multi-dimensional conflict detection according to the conflict detection configuration determined by the scene label to obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold, and detection frequency; Based on the conflict detection results, a set of candidate solutions is generated, and a conflict solution is determined from the set of candidate solutions through an interactive decision-making process and a content decision-making process. The interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution. The content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier. The structured itinerary data is dynamically rearranged based on the conflict resolution method to obtain rearranged itinerary information.

2. The itinerary scheduling method according to claim 1, characterized in that, The sources of the multi-source trip information include at least two of the following: calendar applications, in-vehicle navigation commands, AI agent output, and user natural language input; the multi-source trip information is processed in a structured manner to obtain structured trip data, including performing the following structured processing steps: The multi-source trip information, along with the current time context and the current location context, is input into a large language model. The large language model then performs multi-intent segmentation, interest point semantic reasoning, and relative time normalization on the multi-source trip information to obtain initial structured data. The initial structured data is subjected to duplicate travel deduplication and data field completion processing to obtain the structured travel data; wherein, the data field completion processing is used to complete the field by calling an external point of interest interface when the point of interest information is missing a preset core field; When the reasoning time of the large language model exceeds a preset time threshold, the system switches to a rule-based processing method based on regular expressions and named entity recognition to perform the structured processing on the multi-source trip information to obtain the structured trip data.

3. The itinerary scheduling method according to claim 1, characterized in that, The point of interest information includes at least one of the following: point of interest name, point of interest coordinates, point of interest category, and point of interest opening hours; The step of extracting scene features from the structured travel data and generating scene labels includes performing the following label generation steps: Based on the points of interest categories, the event keywords in the structured itinerary data, and the sources of the multi-source itinerary information, a static scene identifier is generated; wherein, the static scene identifier includes at least two of the following: time urgency identifier, itinerary distance span identifier, and travel purpose identifier; the time urgency identifier includes one of rigid urgency, flexible urgency, and no time constraint; the itinerary distance span identifier includes one of short-distance within the city, medium-distance between cities, and long-distance across provinces; and the travel purpose identifier includes one of business travel, family travel, medical travel, leisure travel, and daily commuting. The dynamic scene identifier is generated based on vehicle motion signals; wherein the vehicle motion signals include at least one of vehicle speed signals, gear signals, and positioning signals, and the dynamic scene identifier includes one of parking, low-speed driving, and high-speed driving.

4. The itinerary scheduling method according to claim 3, characterized in that, The multiple conflict dimensions include at least three of the following: time dimension, point of interest dimension, weather dimension, route dimension, and event dimension; the method for determining the conflict detection configuration includes performing at least one of the following dimension configuration steps based on the scene label: When the time urgency indicator indicates rigid urgency and the travel purpose indicator indicates business travel or medical travel, the detection priority of the time dimension is placed at the highest position among the multiple conflict dimensions, the judgment threshold of the time dimension is adjusted to a first time buffer amount that is more stringent than the default time buffer amount, and the detection frequency is adjusted to a first frequency that is higher than the default frequency. When the travel purpose identifier indicates leisure travel, the detection priority of the weather dimension is placed at the highest position among the multiple conflict dimensions, and the judgment threshold of the weather dimension is expanded to include at least one comfort judgment condition among the perceived temperature, UV index and wind force level. When the travel destination identifier indicates daily commuting, the detection priority of the route dimension is placed at the highest position among the multiple conflict dimensions, and a normal time baseline is generated based on the user's historical commuting data. The judgment threshold of the route dimension is adjusted to the deviation ratio threshold relative to the normal time baseline. When the travel purpose identifier indicates family travel, the detection priority of the weather dimension is placed at the highest position among the multiple conflict dimensions, and a suitability judgment condition is introduced in the point of interest dimension; wherein, the suitability judgment condition is used to determine whether the target point of interest corresponding to the point of interest information has at least one of the following: children's facilities, mother and baby room, and barrier-free access.

5. The itinerary scheduling method according to claim 4, characterized in that, The step of performing multi-dimensional conflict detection on the structured travel data according to the conflict detection configuration determined by the scene label, and obtaining the conflict detection result, includes performing at least one of the following specific scene detection steps: When the travel purpose identifier indicates business travel or medical travel and the point of interest category is a transportation hub, the latest departure time is calculated by reverse time calculation, the allowable stay time window for intermediate trips is calculated based on the latest departure time, and then a prompt message to shorten or skip the intermediate trip is generated based on the allowable stay time window. The latest departure time and the prompt message are used as part of the conflict detection result. When the trip distance span indicates a long-distance inter-provincial journey, the driving fatigue dimension and energy replenishment dimension are added to the multiple conflict dimensions. When the continuous driving time of the vehicle exceeds a preset fatigue threshold, an unskippable rest point insertion instruction is generated. When the ratio of the vehicle's remaining driving range to the remaining trip distance is lower than a preset replenishment threshold, an energy replenishment prompt is generated. The rest point insertion instruction and the energy replenishment prompt are used as part of the conflict detection result.

6. The itinerary scheduling method according to claim 3, characterized in that, The step of determining conflict solutions from the candidate solution set through an interactive decision-making process and a content decision-making process includes performing at least one of the following interactive configuration steps during the interactive decision-making process: When the time urgency indicator indicates rigid urgency and the dynamic scenario indicator indicates low-speed or high-speed driving, the output channel is limited to a voice channel, and the user response processing strategy is configured to automatically execute a single conflict resolution. When the time urgency indicator indicates rigid urgency and the dynamic scene indicator indicates parking, the output channel is configured as a visual channel, the number of conflict solutions is configured to be no less than two, and the user response processing strategy is configured to automatically select a preset default solution when no user selection is obtained within the first timeout period. When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scenario indicator indicates low-speed driving or high-speed driving, the user response processing strategy is configured to mark the corresponding conflict as pending when no user response is obtained within the second timeout period, and to prompt again when the dynamic scenario indicator changes to parking. When the time urgency indicator indicates flexible urgency or no time constraint and the dynamic scene indicator indicates parking, the output channel is configured as a visual channel, the conflict solution is configured to include complete conflict information and alternative point of interest information, and the user response processing strategy is configured to support user modification and order adjustment operations.

7. The itinerary scheduling method according to claim 3, characterized in that, The content decision-making process includes: When the travel purpose identifier indicates business travel, the candidate solution set is limited to the fastest route strategy type, and the trip cancellation type is excluded from the candidate solution set; When the travel purpose identifier indicates family travel, alternative points of interest that do not have children's facilities are filtered from the candidate solution set, and a soft time window is reserved in the time arrangement of the target solution content; wherein, the soft time window includes lunch time and children's nap time; When the travel purpose identifier indicates leisure travel, an option to adjust the itinerary order is added to the candidate solution set; wherein, the option to adjust the itinerary order is used to exchange the order of outdoor and indoor itineraries according to weather distribution; When the travel destination identifier indicates daily commuting, a route preference set is generated based on the user's historical commuting routes, and the target route content is determined from the route preference set; When the travel destination identifier indicates medical travel, the candidate solution set is limited to the fastest route strategy type, and when the time taken for the regular route exceeds a preset value, an alternative hospital recommendation type is added to the candidate solution set.

8. The itinerary scheduling method according to claim 1, characterized in that, When the conflict detection result includes at least two conflicts existing simultaneously across multiple conflict dimensions, the step of determining the conflict solution from the candidate solution set through an interactive decision-making process and a content decision-making process includes: When at least two conflicts include driving safety-related conflicts, the processing priority of the driving safety-related conflicts is placed at the highest level, and the intervention of the scene label is blocked; wherein, the driving safety-related conflicts include at least one of the combination of extreme weather and high-speed driving and continuous fatigue driving; If the solution generated for the first conflict among the at least two conflicts would trigger a second conflict, the associated impact information of the second conflict shall be output together with the conflict solution. When the number of same-day trips corresponding to the structured trip data exceeds a first threshold and the number of at least two conflicts exceeds a second threshold, at least two global orchestration schemes are generated at once based on the large language model, which integrates the multiple conflict dimensions and the constraints of the structured trip data.

9. The itinerary scheduling method according to claim 3, characterized in that, The step of dynamically orchestrating the structured itinerary data based on the conflict resolution to obtain orchestrated itinerary information includes: Extract the trip distance span identifier, the trip purpose identifier, the time urgency identifier, and the number of trips from the structured trip data; Based on the extracted feature data, the target arrangement strategy is determined from the preset spatial clustering arrangement strategy, the approximate shortest path arrangement strategy, the historical route learning arrangement strategy, and the endpoint backward arrangement strategy. According to the target orchestration strategy, route planning and time node calculation are performed on the structured trip data to obtain the orchestrated trip information; wherein, the process of determining the target orchestration strategy includes at least one of the following processing branches: When the trip distance span identifier indicates a short trip within the city and the number of trips exceeds a preset number, the target scheduling strategy is determined to be the spatial clustering scheduling strategy; wherein, the spatial clustering scheduling strategy is used to cluster based on the geographical location corresponding to the point of interest information, and uses points of interest with clear time requirements as fixed anchor points; When the trip distance span identifier indicates a long-distance inter-provincial trip or an intercity trip and the trip destination identifier indicates a leisure trip, the target scheduling strategy is determined as the approximate shortest path scheduling strategy; wherein, the approximate shortest path scheduling strategy is used to solve for the approximate shortest path for the nodes corresponding to the point of interest information, and to forcibly insert rest stops and energy replenishment points into the scheduled trip information; When the travel destination identifier indicates daily commuting, the target scheduling strategy is determined as the historical route learning scheduling strategy; wherein, the historical route learning scheduling strategy is used to generate a route preference probability distribution based on the user's historical commuting data, and select a travel route from the route preference probability distribution in combination with real-time traffic conditions; When the time urgency indicator indicates rigid urgency, the target scheduling strategy is determined as the endpoint backward scheduling strategy; wherein, the endpoint backward scheduling strategy is used to calculate the latest departure time and the available stay time window for each point of interest in the structured itinerary data in reverse from the deadline time.

10. The itinerary scheduling method according to claim 3, characterized in that, After obtaining the arranged itinerary information, the method further includes performing the following itinerary support service steps: When a trip interruption event is detected, the unfinished trips in the structured trip data are queried and real-time status detection is performed to obtain the real-time status detection results; When the real-time status detection result indicates no change, the interruption point navigation information is loaded and navigation is automatically resumed if no user response is obtained after a timeout; when the real-time status detection result indicates a change, the conflict resolution is regenerated and navigation resumes after obtaining user confirmation. During the trip, the triggering time and scope of recommended service information are determined based on the scenario tags; wherein, when the travel destination identifier indicates business travel, the scope of recommended content is limited to destination parking and business assistance information; when the travel destination identifier indicates daily commuting, the regular service recommendation function is turned off and replenishment service information is generated when the vehicle's remaining energy is lower than a preset value.

11. A travel scheduling device, characterized in that, include: The acquisition module is used to acquire multi-source travel information and perform structured processing on the multi-source travel information to obtain structured travel data; wherein, the structured travel data includes point of interest information and time information; The generation module is used to extract scene features from the structured trip data and generate scene tags; wherein, the scene tags include static scene identifiers generated based on the structured trip data and dynamic scene identifiers generated based on the real-time vehicle status; The detection module is used to perform multi-dimensional conflict detection on the structured travel data according to the conflict detection configuration determined by the scene label, and obtain conflict detection results; wherein, the conflict detection configuration includes at least one of multiple conflict dimensions, such as detection priority ranking, judgment threshold and detection frequency; The determination module is used to generate a set of candidate solutions based on the conflict detection results, and to determine the conflict solution from the set of candidate solutions through an interactive decision-making process and a content decision-making process; wherein, the interactive decision-making process includes cross-combining the static scene identifier and the dynamic scene identifier to determine the output channel and user response processing strategy of the conflict solution; the content decision-making process includes filtering the target solution content of the conflict solution from the set of candidate solutions based on the trip attribute features in the static scene identifier; The orchestration module is used to dynamically orchestrate the structured trip data based on the conflict resolution to obtain orchestrated trip information.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the schedule arrangement method as described in any one of claims 1 to 10.

13. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the schedule arrangement method as described in any one of claims 1 to 10.

Images