Multimodal data driven multi-destination trip planning system

Abstract

The present application relates to the technical field of intelligent travel planning, specifically a multi-modal data-driven multi-destination travel planning system, which comprises a data integration and analysis module, a time series conversion module, a time window calculation module, a strategy development module, a route optimization module, and a real-time adjustment module. In the present application, by integrating GPS tracking, social media, and online travel platform data, the accuracy of capturing traveler preferences and associated locations during travel is significantly improved. The dynamic connectivity analysis between travel paths and locations is optimized. The application of Laplace transform processing time series data simplifies complexity, improves conversion accuracy and response speed. The dynamic adjustment of time windows matches travel demand, optimizes stay time allocation, improves travel efficiency, continuously updates travel plans and dynamically adjusts routes, responds to emergencies, and improves the flexibility and practicality of travel arrangements, thereby significantly improving user satisfaction and travel experience quality.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent travel planning, and in particular to a multi-modal data-driven multi-destination travel planning system. BACKGROUND

[0002] The technical field of intelligent travel planning involves the use of artificial intelligence and machine learning algorithms to optimize and personalize travel experiences by analyzing bulk data such as user preferences, historical travel data, weather conditions, traffic situations, and location-related information to develop optimal travel plans for users. Intelligent travel planning systems not only include route planning but also provide real-time traffic updates, recommendations for attractions, restaurants, hotel reservations, and other services. The field is also integrating more advanced features such as context analysis and context awareness capabilities to further improve the accuracy of travel planning and user satisfaction.

[0003] Among them, the multi-modal data-driven multi-destination travel planning system is an advanced travel planning tool that comprehensively utilizes data from multiple modalities such as images, text, and sound information, as well as users' personal preferences and historical behavior data to provide more comprehensive and customized travel planning services. This system can identify users' specific needs by analyzing multi-source data and plan travel routes that include multiple destinations while optimizing travel time and cost. The main purpose of this system is to provide a more intelligent and personalized travel planning method to make users' travel experience more convenient, efficient, and enjoyable.

[0004] Existing technologies have been able to use artificial intelligence and machine learning algorithms to optimize travel experiences, but there are limitations in handling complex data and real-time updates. The time series processing in existing technologies fails to effectively convert from the time domain to another frequency domain, thus not fully utilizing the frequency domain characteristics of data to simplify the processing and analysis process in some cases. Existing technologies lack sufficient flexibility and accuracy in dynamically adjusting travel plans, and cannot respond to sudden events or users' immediate demand changes in a timely manner, which leads to travel plans encountering inconvenience in actual travel. Existing systems also lack in-depth integration of multi-source data such as social media, GPS data, and user preferences, resulting in travel planning services that are not fully personalized, which affects user satisfaction and travel experience, and limits the ability of existing technologies to provide comprehensive, efficient, and personalized travel planning. SUMMARY

[0005] The purpose of the present application is to solve the shortcomings of the prior art, and a multi-modal data-driven multi-destination travel planning system is proposed.

[0006] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme: the multi-modal data-driven multi-destination travel planning system comprises:

[0007] The data integration and analysis module captures GPS tracking, social media and online travel platform data, performs data standardization, identifies traveler preferences and key locations, analyzes inter-location connectivity, determines travel paths, and obtains a travel preference database;

[0008] The time series conversion module converts time data in the travel preference database into a complex frequency domain, simplifies data processing, and obtains a complex frequency domain sequence;

[0009] The time window calculation module performs inverse transformation back to the time domain based on the complex frequency domain sequence, analyzes time data, adjusts time windows and matches requirements, and obtains adjusted time windows;

[0010] The strategy development module analyzes time delays and emergency risks through the adjusted time windows, updates and dynamically adjusts travel plans, and obtains a travel strategy draft;

[0011] The route optimization module performs detailed optimization of routes and time according to the travel strategy draft, adjusts travel time and cost of routes by referring to real-time traffic data and current user location, and obtains an optimized travel route map;

[0012] The real-time adjustment module continuously receives weather and traffic updates based on the optimized travel route map, adjusts details of multiple travel segments, and obtains a multi-destination travel schedule.

[0013] As a further scheme of the present application, the travel preference database includes preference types, activity frequencies and location preferences, the complex frequency domain sequence includes frequency components, amplitude values and phase information, the adjusted time windows include optimal stay durations, demand responsiveness and geographical adaptability, the travel strategy draft includes risk assessment, plan updating and route dynamic adjustment, the optimized travel route map includes path optimization, time efficiency and cost effectiveness, and the multi-destination travel schedule includes starting time, route order and estimated arrival time.

[0014] As a further scheme of the present application, the data integration and analysis module includes:

[0015] The location standardization submodule captures GPS tracking, social media and online travel platform data, identifies and extracts geographic location information, unifies location data to a standard format, matches and verifies the consistency of locations, and generates unified geographic landmarks;

[0016] The preference analysis submodule clusters user behavior data according to the unified geographic landmarks, identifies frequently visited locations and activity preferences, calibrates preference points of travelers, analyzes and calibrates high-frequency access locations, and obtains preference pattern records;

[0017] The path dynamic connection sub-module utilizes the preference mode record, analyzes the access sequence between multiple preference locations, calculates a travel route, dynamically analyzes the path, judges the connectivity and popularity, and generates a travel preference database.

[0018] As a further scheme of the present application, the time series conversion module comprises:

[0019] The time data extraction sub-module extracts time data according to a time label in the travel preference database, formats the extracted data, performs data calibration, and obtains formatted time data.

[0020] The Laplace conversion sub-module performs Laplace transform based on the formatted time data, including setting initial conditions of the transform and applying data processing measures, performs domain conversion of the time series, converts the time series from the time domain to the complex frequency domain, and obtains complex frequency domain basic data.

[0021] The frequency domain data processing sub-module performs data smoothing and noise filtering based on the complex frequency domain basic data, adjusts the correlation parameters, and optimizes the resolution and usability of the data, generates a complex frequency domain sequence through data arrangement and optimization.

[0022] As a further scheme of the present application, the time window calculation module comprises:

[0023] The sequence inverse transform sub-module performs inverse transform step by step based on the complex frequency domain sequence, maps the complex frequency domain data back to the time domain data structure item by item, sorts the time series points in the mapping result, adjusts the data order and meets the time continuity requirement, and obtains the restored time series.

[0024] The time window matching sub-module identifies the time interval based on the restored time series, marks the travel time length of each segment by decomposing the continuous time interval into differentiated independent segments, checks the adaptation degree of the time length of each segment to the travel demand, filters the combination of time intervals that meet the demand, arranges them into the required format, and generates matched time intervals.

[0025] The stay time optimization sub-module applies clustering analysis technology based on the matched time intervals, calculates the stay time of multiple travel segments, determines the stay time of multiple segments by analyzing the geographical interval corresponding to each time data, reallocates the stay distribution in the time window, adjusts the stay time of each segment and optimizes the connection effect, and generates an adjusted time window.

[0026] As a further scheme of the present application, the formula of the clustering analysis technology is as follows:

[0027]

[0028] The adjusted dwell time is calculated to obtain an optimized time allocation, wherein T adj represents the adjusted dwell time, T i represents the original dwell time of the ith travel segment, W geo,i represents the weight coefficient of the ith travel segment based on the geographical position, D i represents the geographical interval distance of the ith travel segment, represents the average geographical interval distance of the travel segments, and n represents the total number of travel segments.

[0029] As a further scheme of the present application, the strategy development module comprises:

[0030] The risk assessment submodule extracts key event sources of time intervals segment by segment based on the adjusted time window, calculates the delay probability, analyzes the burst situation in each time segment, records the delay risk segment, and summarizes the time difference of the differentiated delay scenarios to generate a potential risk analysis result;

[0031] The plan adjustment submodule identifies risk periods segment by segment based on the potential risk analysis result, allocates standby time, reorders time segments that affect the trip, adjusts risks, and allocates time intervals to dynamically optimize the trip and obtain an optimized travel plan;

[0032] The route updating submodule extracts new route nodes based on the optimized travel plan, adjusts the geographical position of differentiated key path nodes, reorders the path sequence, checks the continuity between travel segments, optimizes the connection order of nodes, and obtains a travel strategy draft.

[0033] As a further scheme of the present application, the route optimization module comprises:

[0034] The real-time data integration submodule gradually captures real-time traffic data based on the travel strategy draft, extracts key traffic information in each road section, identifies the congestion status of the road section, matches the current position of the user with the surrounding traffic information, integrates each road section, adjusts the congestion parameters of differentiated road sections, associates real-time information of multiple path segments, and obtains real-time path information;

[0035] The route adjustment submodule locates key road segments in the trip based on the real-time path information, performs segmented analysis on multiple traffic delay points, checks the traffic conditions of differentiated road segments within the time window, combines and filters sequentially adjacent road segments, reorders and sets alternative paths, selects road segment combinations with short expected times segment by segment, and obtains an optimized path structure;

[0036] The time and cost optimization sub-module applies Dijkstra algorithm based on the optimized path structure, calculates the time consumption and travel cost of each section, and adjusts the travel time to low-cost sections, so as to balance the time and cost ratio between sections, fine-tune part of the route, and generate an optimized travel route map.

[0037] As a further scheme of the application, the Dijkstra algorithm formula is as follows:

[0038]

[0039] The total travel cost of each vertex is calculated to obtain an optimal travel path, wherein T(v) represents the total travel cost, d v represents the distance from the vertex v to the starting point, c v represents the travel cost of the vertex v, p v represents the priority of the vertex v, w time , w cost and w priority respectively represent the influence of the adjusted time cost, the cost and the section priority.

[0040] As a further scheme of the application, the real-time adjustment module comprises:

[0041] The weather monitoring integration sub-module captures real-time weather data based on the optimized travel route map, extracts temperature, precipitation and wind speed information of each section, associates the parameters with the target section, identifies the potential influence of weather factors section by section, and selects the sections that affect the travel according to the weather conditions, marks the road condition on the corresponding route section, and generates real-time weather parameters;

[0042] The route detail review sub-module reviews the traffic and weather conditions of each route according to the real-time weather parameters, identifies the delayed sections, checks the road congestion and records the real-time traffic, adjusts the route sections that do not meet the expectations, resets the passing time period, updates and records the current time parameters of each section, and generates a detailed route arrangement;

[0043] The dynamic time arrangement sub-module adjusts the departure and arrival time of each section according to the current weather and traffic factors, compares the reasonableness of the connection time of each section, sets a reasonable departure time for each travel section, allocates buffer time, and generates a multi-destination travel arrangement table.

[0044] Compared with the prior art, the application has the advantages and positive effects that:

[0045] In the present application, the accuracy of capturing the correlation between traveler preferences and travel plans is significantly enhanced through the integrated analysis of GPS tracking, social media, and online travel platform data, which not only optimizes the determination of travel paths, but also realizes the efficient analysis of dynamic connectivity between locations, simplifies data complexity using Laplace transform processing of time series data, thereby improving processing efficiency and response speed, making the conversion from time domain to complex frequency domain more accurate, and further improving the accuracy and reliability of data recovery through inverse transform. By dynamically adjusting the time window to match travel demand and geographical interval, the allocation of stay time is optimized to ensure the optimization of travel efficiency and cost, and the continuous update of travel plans and dynamic adjustment of routes effectively reduces the inconvenience caused by unexpected events or time delays, enhances the flexibility and practicality of travel arrangements, not only improves the overall efficiency of travel planning, but also greatly enhances user satisfaction and the quality of travel experience. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 The system flowchart of the present application;

[0047] Figure 2 The system framework diagram of the present application;

[0048] Figure 3 The flowchart of the data integration and analysis module in the present application;

[0049] Figure 4 The flowchart of the time series conversion module in the present application;

[0050] Figure 5 The flowchart of the time window calculation module in the present application;

[0051] Figure 6 The flowchart of the strategy development module in the present application;

[0052] Figure 7 The flowchart of the route optimization module in the present application;

[0053] Figure 8 The flowchart of the real-time adjustment module in the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present application more clear and explicit, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0055] In the description of the present application, it should be understood that the terms "length", "width", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, in the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly and specifically limited.

[0056] Embodiment one

[0057] Please refer to Figures 1 to 2 , the multi-modal data-driven multi-destination travel planning system comprises:

[0058] The data integration and analysis module captures GPS tracking, social media and online travel platform data, standardizes the data, identifies traveler preferences and key places associated with the trip, analyzes the dynamic connectivity between places, determines the travel path, and obtains a travel preference database;

[0059] The time series conversion module uses the time data in the travel preference database to perform Laplace transform to convert the time series from the time domain to the complex frequency domain, simplifies the complexity of the time data, and obtains the complex frequency domain sequence;

[0060] The time window calculation module is based on the complex frequency domain sequence, performs inverse transform to restore the data to the time domain, analyzes the time data, adjusts the time window and matches the travel demand and geographical interval, determines the optimal stay time of each travel segment, and obtains the adjusted time window;

[0061] The strategy development module analyzes the potential time delay and the risk of sudden events through the adjusted time window, updates the travel plan, dynamically adjusts the travel plan, updates the travel route, and obtains a draft of the travel strategy;

[0062] The route optimization module optimizes the details of the route and time according to the draft of the travel strategy, adjusts the travel time and cost of the route by referring to real-time traffic data and the current location of the user, and obtains an optimized travel route map;

[0063] The real-time adjustment module continuously receives data updates from weather monitoring and traffic information services based on the optimized travel route map, conducts a detailed review of each travel route, adjusts the departure time, route and estimated arrival time of each travel segment according to real-time data, and obtains a multi-destination travel schedule.

[0064] The travel preference database includes preference types, activity frequencies, location preferences, the complex frequency domain sequence includes frequency components, amplitude values, and phase information, the adjusted time window includes optimal stay duration, demand responsiveness, and geographical adaptability, the travel strategy draft includes risk assessment, plan updating, and route dynamic adjustment, and the optimized travel route map includes path optimization, time efficiency, and cost effectiveness. The multi-destination travel schedule includes the starting time, route order, and estimated arrival time.

[0065] Specifically, as shown in Figure 2 、 3 The data integration and analysis module includes:

[0066] The location normalization submodule captures GPS tracking, social media, and online travel platform data, identifies and extracts geographic location information, unifies location data to a standard format, matches and verifies the consistency of locations, and generates a unified geographic sign. The execution process is as follows:

[0067] The standardization of location data involves the integration of multiple data sources, including GPS tracking, social media, and online travel platform data. Advanced algorithms are used to identify and extract geographic location information from each data source, and the location data is unified to a standard format. The unified processing steps include data cleaning, format standardization, and outlier processing, such as removing speed outliers from GPS data and extracting landmark property information from social media data. Then, a standardized geocoding is used to encode the extracted geographic information. The algorithm compares and verifies the consistency of locations to ensure the accuracy and usability of the information, generates a unified geographic sign, and ensures that each geographic location can be accurately matched in different data sources. This processing not only improves the accuracy of geographic data, but also facilitates subsequent data analysis and application.

[0068] The preference analysis submodule clusters user behavior data based on the unified geographic sign, identifies frequently visited locations and activity preferences, and labels the preference points of travelers. The execution process for analyzing and labeling high-frequency access locations and obtaining preference pattern records is as follows:

[0069] The user behavior data is clustered according to the formula:

[0070]

[0071] The clustering centers are calculated, where C represents the clustering center, w j represents the weight of each data point, x j represents the feature vector of the data point, and m represents the total number of data points.

[0072] When clustering, there are n data points, each data point x j has a corresponding weight w jThe weight can be determined according to the user access frequency and other factors, and the clustering center C is calculated by weighted average, for example, if there are three points x1=[1, 2], x2=[2, 3], x3=[3, 4], and the corresponding weights are w1=0.2, w2=0.5, and w3=0.3, then the calculation of the clustering center C is as follows:

[0073] C=0.2·[1,2]+0.5·[2,3]+0.3·[3,4]= [2.1,3.1]

[0074] This calculation process ensures that the user preferences are reflected, and points with high weights have a greater impact on the clustering center.

[0075] The path dynamic connection submodule uses the preference pattern record to analyze the access sequence between multiple preference locations, calculates the travel route, dynamically analyzes the path, judges the connectivity and popularity, and generates the execution flow of the travel preference database as follows:

[0076] The access sequence analysis and path planning process for multiple preference locations includes complex data analysis and mathematical calculation. The submodule first uses the preference pattern record, which contains the preference data of the user's most frequently visited locations and activities. According to the data in the database, the submodule analyzes the potential access sequence of the user from one preference location to another, calculates the travel route through an algorithm, and dynamically analyzes the path, including the evaluation of connectivity and popularity, to optimize the travel route and generate the travel preference database. This database can support the travel recommendation system, making the recommendation more personalized and accurate.

[0077] Specifically, as shown in Figure 2 , 4 The time series conversion module includes:

[0078] The time data extraction submodule extracts time data from the travel preference database based on the time tags, formats the extracted data, performs data calibration, and obtains the execution flow of the formatted time data as follows:

[0079] The time tag data is extracted from the travel preference database, processed, first read all the time tag data contained in the travel preference database, and check and adjust the format of the time tag to ensure that the time zone, date format, etc. meet the predetermined standard format requirements, then complete the invalid or missing time tag in the data, fill in the missing tag using interpolation algorithm, align the data to a fixed standard time axis to ensure time consistency, and finally correct the deviation through error analysis, and finally pass the formatted time data to the next module for use.

[0080] The Laplace transform sub-module performs Laplace transform based on the formatted time data, including setting initial conditions of the transform and applying data processing measures, and performs domain conversion of the time series to convert the time series from the time domain to the complex frequency domain, and the execution process of obtaining the complex frequency domain basis data is as follows:

[0081] The Laplace transform sub-module performs the conversion of the time series from the time domain to the complex frequency domain based on the formatted time data, according to the formula:

[0082]

[0083] The complex frequency domain basis data is calculated, wherein f(t) represents the formatted time data, t represents the time variable, s is the complex frequency variable, and is used to adjust the frequency domain characteristics of the time series, represents the function after the Laplace transform;

[0084] The Laplace transform formula realizes the conversion through integration of the time series, the parameter f(t) represents the preprocessed time data, the standardized time data obtained in seconds, and the value of the complex frequency variable s needs to be set according to the frequency domain analysis target, which is s = σ + iω, wherein σ is the time decay term, ω is the angular frequency, and the specific value is obtained from the frequency domain characteristic analysis of the observation data. For example, if the time data f(t) = e -2t , the Laplace transform is applied to obtain:

[0085]

[0086] The results show that by converting the time data to the complex frequency domain, the response characteristics of the data at different frequencies can be obtained, which is helpful for data smoothing and resolution optimization in subsequent frequency domain data processing.

[0087] The frequency domain data processing sub-module performs data smoothing and noise filtering based on the complex frequency domain basis data, adjusts the associated parameters, and optimizes the resolution and usability of the data. Through data arrangement and optimization, the execution process of generating the complex frequency domain sequence is as follows:

[0088] After receiving the complex frequency domain basis data, first, the data is smoothed to reduce the noise components in the data. A low-pass filter is often used to remove high-frequency noise. After data smoothing, the associated parameters of the processed data are adjusted one by one. The preliminary frequency domain signal is recovered through inverse Fourier transform to verify the accuracy of the smoothed results. The resolution of the data is further optimized according to the complex frequency domain characteristics of the data. The resolution corresponding to the sampling point number and sampling interval is set, and the smoothing degree is optimized through weighted average method. The finally generated complex frequency domain sequence is transmitted to the output module.

[0089] Specifically, as Figure 2 ,5 As shown, the time window calculation module includes:

[0090] The sequence inverse transformation submodule performs inverse transformation step by step based on the complex frequency domain sequence, maps the complex frequency domain data back to the time domain data structure item by item, and sorts the time sequence points in the mapping result, adjusts the data order and meets the time continuity requirement, and the execution process of the restored time sequence is as follows:

[0091] The inverse transformation is performed on the complex frequency domain sequence, each data point is recovered to the time domain structure step by step through the inverse discrete Fourier transform, the phase of each frequency component in the complex frequency domain is adjusted, the position of each frequency component in the time domain after inverse transformation is calculated, and the obtained time sequence data is sorted according to the time label, so that the data order strictly meets the continuity of the time axis. In the process, the complex frequency domain components of the data need to be discretely resampled to ensure the accuracy and stability of the calculation. The mapping and sorting are performed in turn, and the final result is arranged to obtain the restored time sequence.

[0092] The time window matching submodule is based on the restored time sequence, identifies the time interval, marks each travel time length by decomposing the continuous time interval into differentiated independent paragraphs, checks the adaptation degree of the time length of each paragraph to the travel demand, selects the time interval combination that meets the demand, and arranges it into the required format. The execution process of the matched time interval is as follows:

[0093] Based on the restored time sequence, first, the time period of each travel is identified through time interval identification, the demarcation points of the differentiated continuous time interval in the time sequence are determined, and each segment of time after the demarcation point is decomposed and processed according to the time window matching rule. Based on the duration of each segment, the time interval that does not meet the target time demand is removed, and the differentiated travel time interval is independently marked through segmentation processing. After formatting processing, the marked time interval is further cross-compared to select the time interval combination that meets the travel demand, and the matched time interval that meets the required format is generated.

[0094] The stay time optimization submodule is based on the matched time interval, applies clustering analysis technology, calculates the stay time of multiple travel segments, determines the stay time of multiple segments by analyzing the geographical interval corresponding to each time data, redistributes the stay distribution in the time window, adjusts the stay time of each segment and optimizes the connection effect, and the execution process of the adjusted time window is as follows:

[0095] The formula of the clustering analysis technology is as follows:

[0096]

[0097] The adjusted stay time is calculated, and the optimized time allocation is obtained, wherein T adj represents the adjusted stay time, T iW represents the original stay time of the ith trip geo,i D represents the weight coefficient of the ith trip based on geographical location i D represents the geographical interval distance of the ith trip n represents the average geographical interval distance of the trip segment, and n represents the total number of trip segments

[0098] Formula details and formula calculation derivation process:

[0099] The formula is used to calculate the stay time T adj optimized based on geographical location i W represents the original stay time of the ith trip geo,i D is the weight coefficient given based on geographical importance i D is the geographical interval distance of the ith segment, and n is the average geographical interval distance of the trip segment, and the division and root operations in the formula are used to standardize and balance the deviation degree of each segment geographical interval, so as to reasonably allocate the stay time;

[0100] It is assumed that there are three trip data:

[0101] The stay time of the first trip T1 is 30 minutes, and the geographical interval D1 is 5 kilometers;

[0102] The stay time of the second trip T2 is 45 minutes, and the geographical interval D2 is 3 kilometers;

[0103] The stay time of the third trip T3 is 60 minutes, and the geographical interval D3 is 8 kilometers;

[0104] The average value of the geographical interval distance Through calculation, we get:

[0105]

[0106] The weight coefficient W geo,i Based on the importance of geographical location, the weight coefficient obtained through actual monitoring and analysis is:

[0107] W geo,1 = 0.8, W geo,2 = 1.0, W geo,3 = 0.7

[0108] Then, the weighted stay time of each trip is calculated:

[0109]

[0110] The weighted minutes, the sum of the standard deviations of the geographical intervals of all trip segments is calculated:

[0111]

[0112] km, the adjusted dwell time is obtained by dividing the weighted dwell time by the square root of the standard deviation of the geographical interval:

[0113]

[0114] The results show that by weighting and standardizing the dwell time of each travel segment according to the geographical interval and importance, a more optimized overall dwell time allocation is obtained, which is about 48 minutes. This calculation method allows for detailed adjustment of differentiated geographical characteristics and actual dwell requirements, improving travel efficiency.

[0115] Specifically, as shown in Figure 2 、 6 The strategy development module includes:

[0116] The risk assessment submodule extracts key event sources in each time interval based on the adjusted time window, calculates the delay probability, analyzes the sudden situation in each time interval, records the delay risk segment, and summarizes the time difference of differentiated delay scenarios. The execution process of the potential risk analysis result is as follows:

[0117] Based on the adjusted time window, the key event sources in each time interval are extracted, first, the event source set is formed by marking the time nodes of the sudden events in the time sequence, where the data of the event source comes from the actual delay event record, and the suddenness of the event is determined by the anomaly detection algorithm. Further, the occurrence time, occurrence frequency, and influence range of the event source are quantified to construct a time-event matrix, and the sudden events in the differentiated time interval are integrated into the corresponding time interval. Then, the probability of causing delay for each event source is calculated. The calculation of the delay probability is based on the cumulative effect of the frequency of events in each time interval, correlation factors, and historical delay data parameters. The delay probability formula is obtained by statistical distribution fitting, and the calculation result is normalized. Finally, the delay risk distribution in each time interval is formed, and the delay risk segment is further marked on the time axis. The time difference generated by the differentiated delay scenario is summarized to obtain the potential risk analysis result.

[0118] The plan adjustment submodule identifies the risk period based on the potential risk analysis result, allocates the standby time, reorders the time interval that affects the trip, combines the risk adjustment, and performs time interval allocation to dynamically optimize the trip and obtain the optimized travel plan. The execution process is as follows:

[0119] Based on the potential risk analysis result, the risk period is identified and the standby time is allocated according to the formula:

[0120] Padj =∑(t n ·p n )

[0121] An optimized travel plan is calculated, wherein t n represents the length of the time period, and p n represents the risk adjustment parameter of the time period;

[0122] Referring to a simple case, the lengths of three time periods are 1 hour, 1.5 hours and 2 hours respectively, and the corresponding risk adjustment parameters are 0.2, 0.5 and 0.3 respectively, then the calculation process is as follows:

[0123] P adj = (1 · 0.2) + (1.5 · 0.5) + (2 · 0.3) = 0.2 + 0.75 + 0.6 = 1.55

[0124] This result shows that the total risk adjustment time of the entire trip is 1.55 hours after referring to the risk adjustment.

[0125] The route updating submodule is based on the optimized travel plan, extracts new route nodes, adjusts the geographic positions of the differentiated key path nodes, reorders the path order, checks the continuity between the travel segments, optimizes the node connection order, and the execution process of the travel strategy draft is as follows:

[0126] Based on the optimized travel plan, new route nodes are extracted and the geographic positions are adjusted. The selection of new paths and the analysis of differentiated key nodes of old paths are core tasks. The geographic positions of path nodes are optimized and adjusted according to the new travel strategy, while ensuring the reasonable ordering of path order and optimizing the node connection order to ensure the continuity between travel segments. The focus of the task in this stage is analyzed by algorithm to determine the optimal node connection order and geographic position adjustment, ensuring the feasibility and efficiency of the travel strategy in practical application. The final travel strategy draft will directly affect the smooth progress of the trip.

[0127] Specifically, as shown in Figure 2 , 7 , the route optimization module includes:

[0128] The real-time data integration submodule is based on the travel strategy draft, gradually captures real-time traffic data, extracts key traffic information in each road section, identifies the congestion status of the road section, matches the current position of the user with the surrounding traffic information, integrates each road section, adjusts the congestion parameters of the differentiated road sections, associates the real-time information of multiple path sections, and the execution process of obtaining real-time path information is as follows:

[0129] Based on the draft travel strategy, the real-time data integration submodule focuses on capturing and integrating real-time traffic data. Key tasks include capturing key traffic information and performing accurate extraction, identifying the congestion status of each road segment in detail, accurately matching the user's current location with surrounding traffic information to ensure real-time and accurate information, and adjusting congestion parameters in real time based on different time periods and road conditions to ensure that the provided data accurately reflects the current traffic situation. The submodule can integrate traffic information from different sources, including vehicle GPS data, traffic monitoring cameras, and traffic flow monitoring stations. Comprehensive analysis of information helps form a comprehensive traffic data model, providing users with the most up-to-date road conditions and supporting data for subsequent route adjustments to ensure that the travel strategy can quickly adapt to sudden changes in traffic conditions, ultimately improving travel efficiency and safety.

[0130] The route adjustment submodule is based on real-time path information and locates key road segments in the trip. It performs segmented analysis on multiple traffic delay points, checks the traffic conditions of different road segments within a time window, and combines and filters sequentially adjacent road segments. The execution process of obtaining the optimized path structure is as follows:

[0131] Based on real-time path information, the core work of the route adjustment submodule is to accurately locate key road segments in the trip and conduct in-depth analysis of traffic delays on road segments. The submodule first checks the traffic conditions of different road segments within a specific time window, and then combines and filters sequentially adjacent road segments. The purpose is to optimize the structure of the entire journey by reordering and setting alternative paths. In the execution process, the module uses advanced algorithms to predict traffic flow and delay on each road segment, ensuring that the selected path can be completed within the expected time, reducing unnecessary waiting and detours, and continuously updating traffic data on each road segment to ensure that all decisions are based on the latest traffic conditions, thereby achieving the optimal travel time. Through detailed analysis and adjustment, travelers can obtain a journey plan with the shortest predicted time and least traffic impact, improving travel efficiency and comfort.

[0132] The time and cost optimization submodule is based on the optimized path structure and applies Dijkstra's algorithm to calculate the time consumption and travel cost of each road segment. It prioritizes adjusting road segments with high time and cost, redistributes travel time to low-cost segments, and fine-tunes some routes by balancing the time and cost ratio between road segments. The execution process of generating the optimized travel route map is as follows:

[0133] The formula of Dijkstra's algorithm is as follows:

[0134]

[0135] The total travel cost of each vertex is calculated to obtain the optimal travel path, where T(v) represents the total travel cost, d v represents the distance from vertex v to the starting point, c v represents the travel cost of vertex v, and p v represents the priority of vertex v, and w time , w cost , and w priority represent the influence of adjusting time cost, cost, and road priority, respectively.

[0136] Formula details and formula calculation derivation process:

[0137] The parameters d v , c v , and p v in the formula represent the distance from vertex v to the starting point, the travel cost through the vertex, and the priority of the vertex, respectively. The parameters are obtained through actual road network data, such as GPS tracking and cost calculation software to monitor and collect data. The weight parameters w time , w cost , and w priority are used to adjust the influence of time cost, cost, and road priority.

[0138] If a specific route is set, the distance d v from the starting point to vertex v is 10 kilometers, the travel cost c v determined by the toll system is 50 yuan, and the priority p v of the vertex is obtained by analyzing traffic flow and road condition data, w time = 0.5, w cost = 0.3, and w priority = 0.2 are set by the traffic management department to balance time and cost.

[0139] Calculate the combination of distance and time cost:

[0140] d v · w time = 10 × 0.5 = 5

[0141] c v · w cost = 50 × 0.3 = 15

[0142] Add the two values:

[0143] TotalCost = 5 + 15 = 20

[0144] Calculate the weighted value of the priority and add 1 square root to adjust the overall cost:

[0145]

[0146] The total cost is divided by the result of step 3 to obtain the final adjusted total travel cost T(v):

[0147]

[0148] The result shows that the total travel cost of vertex v is approximately 16.95, which is the balanced result after combining travel distance, cost, and road priority for selecting the most cost-effective route in the path optimization module.

[0149] Specifically, as shown in Figure 2 , 8 , the real-time adjustment module includes:

[0150] The weather monitoring integration sub-module captures real-time weather data based on the optimized travel route map, extracts temperature, precipitation, and wind speed information for each road segment, associates the parameters with the target road segment, identifies the potential impact of weather factors segment by segment, and filters the road segments that affect the trip according to the weather conditions. The current road condition is marked on the corresponding route segment, and the execution process of real-time weather parameters is as follows:

[0151] During the operation based on the optimized travel route map, the weather monitoring integration sub-module plays a key role. This module first captures real-time weather data, including temperature, precipitation, and wind speed information for each road segment. Based on the collected data, it analyzes the weather conditions of each route segment by segment and associates the parameters with the target road segment. The operation ensures real-time updating and accuracy of information. Through the analysis of differentiated road weather conditions, it filters the road segments that affect the trip and marks the current road condition on the map, providing immediate navigation assistance for drivers. This not only improves travel safety but also ensures travel efficiency. Through weather monitoring of this system, traffic delays and accidents caused by adverse weather conditions can be effectively avoided or reduced, thereby providing a more stable and reliable travel plan.

[0152] The route detail review sub-module reviews the traffic and weather conditions of each route based on real-time weather parameters, identifies delayed road segments, checks road congestion and records real-time traffic flow, adjusts routes that do not meet expectations, resets the passing time period, updates and records the current time parameters for each segment, and generates the execution process of detailed route planning as follows:

[0153] According to real-time weather parameters, the traffic and meteorological conditions of each travel segment are reviewed in detail, the core includes identifying the road segments that cause delays, checking the current road congestion conditions, and recording the real-time traffic flow, the information helps the module to adjust those route segments that do not meet the expectations, the module resets the travel time of each route segment and updates and records the current time parameters of each route segment in real time, through this detailed review and adjustment, a more accurate and adaptive detailed route arrangement is generated according to the current weather and traffic conditions, this detailed process ensures the real-time and accuracy of the travel plan, reduces unpredictable delays, and optimizes the overall travel arrangement.

[0154] The dynamic time arrangement submodule adjusts the departure and arrival time of each segment based on the detailed route arrangement according to the current weather and traffic conditions, sets a reasonable departure time for each travel segment by comparing the connection time of each segment, and allocates buffer time, the execution process of the multi-destination travel arrangement table is as follows:

[0155] The dynamic time arrangement submodule adjusts the departure and arrival time of each segment according to the detailed route arrangement, according to the formula:

[0156] t new =t old +Δt

[0157] Calculate the new travel time, where t old represents the original planned departure or arrival time, Δt represents the time adjustment amount based on the current weather and traffic conditions, and t new represents the new travel time.

[0158] Consider a road trip originally scheduled to depart at 8:00 AM, due to unexpected traffic congestion and bad weather, it is expected to be delayed by 15 minutes, according to the formula, the new departure time is calculated as follows:

[0159] t new =8:00AM+0:15=8:15AM

[0160] This adjustment ensures reasonable allocation of travel time and reasonable setting of buffer time, generates a multi-destination travel arrangement table, and effectively responds to the cyclic changes in travel conditions.

[0161] The above is only a preferred embodiment of the present application, and does not limit the form of the present application, any person skilled in the art can use the above disclosed technical content to make changes or modifications as equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made according to the technical essence of the present application to the above embodiments without departing from the technical solution content of the present application, still belongs to the protection scope of the technical solution of the present application.

Claims

1. A multi-modal data driven multi-destination trip planning system, characterized by, The system comprises: The data integration and analysis module captures GPS tracking, social media and online travel platform data, performs data standardization, identifies traveler preferences and key locations, analyzes inter-location connectivity, determines travel paths, and obtains a travel preference database; The time series conversion module converts time series into complex frequency domain using time data in the travel preference database, simplifies data processing, and obtains complex frequency domain sequences; The time series conversion module comprises: The time data extraction submodule extracts time data according to the time label in the travel preference database, formats the extracted data, performs data calibration, and obtains formatted time data; The Laplace conversion submodule performs Laplace transform based on the formatted time data, including setting initial conditions for the transform and applying data processing measures, performs domain conversion of the time series, converts the time series from the time domain to the complex frequency domain, and obtains complex frequency domain basic data; The frequency domain data processing submodule performs data smoothing and noise filtering based on the complex frequency domain basic data, adjusts correlation parameters, and optimizes data resolution and usability, generates complex frequency domain sequences through data arrangement and optimization; The time window calculation module converts the complex frequency domain sequences back to the time domain through inverse transform based on the complex frequency domain sequences, analyzes time data, adjusts time windows and matches requirements, and obtains adjusted time windows; The time window calculation module comprises: The sequence inverse transform submodule performs inverse transform step by step based on the complex frequency domain sequences, maps complex frequency domain data back to time domain data structure item by item, sorts time series points in the mapping results, adjusts data order and meets time continuity requirements, and obtains restored time series; The time window matching submodule identifies time interval based on the restored time series, marks each travel duration by decomposing continuous time intervals into differentiated independent paragraphs, checks the adaptation degree of each time length to travel requirements, filters combinations of time intervals that meet requirements, arranges them into the required format, and generates matched time intervals; The stay time optimization submodule applies clustering analysis technology based on the matched time intervals, calculates the stay time of multiple travel segments, determines the stay time of multiple segments by analyzing the geographical interval corresponding to each time data, reallocates the stay distribution in the time window, adjusts each stay time and optimizes the connection effect, and generates adjusted time windows; The strategy development module analyzes time delays and unexpected event risks through the adjusted time windows, updates and dynamically adjusts the travel plan, and obtains a travel strategy draft; The route optimization module optimizes the details of the route and time according to the travel strategy draft, adjusts the travel time and cost of the route by referring to real-time traffic data and the current location of the user, and obtains an optimized travel route map; The real-time adjustment module continuously receives weather and traffic updates based on the optimized travel route map, adjusts the details of multiple travel segments, and obtains a multi-destination travel schedule.

2. The multi-modal data driven multi-destination trip planning system of claim 1, wherein: The travel preference database includes preference types, activity frequencies, location preferences, the complex frequency domain sequence includes frequency components, amplitude values, and phase information, the adjusted time window includes optimal stay durations, demand responsiveness, and geographical adaptability, the travel strategy draft includes risk assessments, plan updates, and route dynamic adjustments, and the optimized travel route map includes path optimization, time efficiency, and cost effectiveness.

3. The multi-modal data driven multi-destination trip planning system of claim 1, wherein: The data integration and analysis module includes: The location normalization submodule captures GPS tracking, social media, and online travel platform data, identifies and extracts geographic location information, unifies location data to a standard format, matches and verifies location consistency, and generates unified geographic markers; The preference analysis submodule clusters user behavior data based on the unified geographic markers, identifies frequently visited locations and activity preferences, calibrates traveler preference points, analyzes and calibrates high-frequency access locations, and obtains preference pattern records; The path dynamic connection submodule uses the preference pattern records to analyze access sequences between multiple preference locations, calculates travel routes, dynamically analyzes paths, judges connectivity and popularity, and generates a travel preference database.

4. The multi-modal data driven multi-destination trip planning system of claim 1, wherein: The formula of the cluster analysis technique is as follows: ； The adjusted dwell time is calculated to obtain an optimized time allocation, wherein, denotes the adjusted dwell time, denotes the original dwell time for the i-th trip, denotes the geographical position-based weight coefficient for the i-th trip, denotes the geographical separation distance for the i-th trip, denotes the average geographical separation distance for the trip segments, and n denotes the total number of trip segments.

5. The multi-modal data driven multi-destination trip planning system of claim 1, wherein: The strategy development module includes: The risk assessment submodule extracts key event sources in time intervals based on the adjusted time window, calculates delay probabilities, analyzes sudden situations in each time period, records delay risk segments, and summarizes time differences of differentiated delay scenarios to generate potential risk analysis results; The plan adjustment submodule identifies risk periods based on the potential risk analysis results, allocates standby time, reorders time periods that affect travel, adjusts risks, and allocates time intervals to dynamically optimize travel, and obtains an optimized travel plan; The route update submodule extracts new route nodes based on the optimized travel plan, adjusts geographic locations of differentiated key path nodes, reorders path sequences, checks the continuity between travel segments, optimizes node connection sequences, and obtains a travel strategy draft.

6. The multi-modal data driven multi-destination trip planning system of claim 1, wherein: The route optimization module includes: The real-time data integration submodule gradually captures real-time traffic data based on the travel strategy draft, extracts key traffic information in each road segment, identifies congestion conditions of road segments, matches the current user location with surrounding traffic information, integrates each road segment, adjusts congestion parameters of differentiated road segments, associates real-time information of multiple path segments, and obtains real-time path information; The route adjustment submodule locates key road segments in travel based on the real-time path information, analyzes multiple traffic delay points by segments, checks the traffic conditions of differentiated road segments within the time window, combines and filters sequentially adjacent road segments, reorders and sets alternative paths, selects road segment combinations with short estimated times by segments, and obtains an optimized path structure; The time and cost optimization submodule applies Dijkstra algorithm based on the optimized path structure, calculates the time consumption and travel cost of each section, and adjusts the travel time to low-cost sections to balance the time and cost ratio between sections, and fine-tunes part of the route to generate an optimized travel route map.

7. The multi-modal data driven multi-destination trip planning system of claim 6, wherein: The formula of the Dijkstra algorithm is as follows: ； The total travel cost of each vertex is calculated to obtain the optimal travel path, wherein, denotes the total travel cost, denotes the distance from the vertex v to the starting point, denotes the travel cost of the vertex v, denotes the priority of the vertex v, , and respectively denote the influence of the adjustment time cost, the fee cost and the link priority.

8. The multi-modal data driven multi-destination trip planning system of claim 1, wherein: The real-time adjustment module comprises: The weather monitoring integration submodule captures real-time weather data based on the optimized travel route map, extracts temperature, precipitation and wind speed information of each section, associates parameters with target sections, identifies potential weather factors for each section, and filters sections that affect travel according to weather conditions, marks road conditions on corresponding route sections, and generates real-time weather parameters; The route detail review submodule reviews traffic and weather conditions of each route based on the real-time weather parameters, identifies delayed sections, checks road congestion and records real-time traffic, adjusts sections that do not meet expectations, resets travel time, updates and records current time parameters of each section, and generates detailed route arrangements; The dynamic time scheduling submodule adjusts the departure and arrival time of each section based on the detailed route arrangement, sets reasonable departure time for each travel section by comparing the connection time of each section, allocates buffer time, and generates a multi-destination travel schedule.

Images