Navigation trajectory data processing method and system
By employing a multi-dimensional comprehensive evaluation method, the problem of decreased navigation system accuracy caused by trajectory data contamination in existing technologies has been solved. This method enables the identification and filtering of high-quality trajectory data, thereby improving the accuracy and stability of the navigation system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- IDEAPOOL CULTURE & TECH CO LTD
- Filing Date
- 2025-12-24
- Publication Date
- 2026-05-12
AI Technical Summary
Existing trajectory data suffers from group data contamination, which reduces the accuracy and reliability of navigation systems. Existing methods struggle to effectively identify abnormal trajectories in complex terrain and large-scale data scenarios.
By employing a multi-dimensional comprehensive evaluation method, including trajectory and physical obstacle layer overlay analysis, user credibility model construction, path rationality assessment, and weighted fusion, abnormal trajectories are identified and eliminated, generating a high-quality, effective trajectory dataset.
It improves the accuracy and reliability of navigation route recommendation and map updates, dynamically responds to the trajectory data characteristics of different regions and user groups, and enhances system stability and robustness.
Smart Images
Figure CN122019516A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of geospatial data processing technology, and in particular to a method and system for processing navigation trajectory data. Background Technology
[0002] With the popularization of mobile internet, positioning technology (GPS, BeiDou) and smart terminals, a large number of users continuously generate trajectory data when using navigation applications. This data is an important foundation for map building, route planning and travel analysis.
[0003] However, existing trajectory data generally suffers from the problem of group data pollution. Some trajectories have positional deviations due to signal drift or upload delays, while others cross physical obstacles such as rivers, buildings, and railway barriers. Individual users may also upload abnormal data due to violations such as crossing lines or driving against traffic. When a large number of abnormal trajectories are collected by the system, it can easily lead to map misjudgments and navigation misleading. Existing technologies mostly use methods such as trajectory smoothing, speed filtering, or anomaly detection based on single trajectories. However, these methods lack a comprehensive consideration of the physical rationality of the trajectory, user credibility, and group consistency, making it difficult to effectively identify abnormal trajectories in complex terrain and large-scale data scenarios, thus affecting the accuracy and reliability of the navigation system. Summary of the Invention
[0004] To overcome the above shortcomings, this invention provides a navigation trajectory data processing method and system that can address the problem of group data pollution in existing trajectory data by performing multi-dimensional comprehensive evaluation and anomaly screening of trajectory data, thereby improving the reliability of the navigation system and the accuracy of path planning.
[0005] In a first aspect, the present invention provides the following technical solution: a navigation trajectory data processing method, comprising:
[0006] Obtain the trajectory dataset uploaded by the user. Each trajectory data includes a timestamp sequence, a location coordinate sequence, and the corresponding user identification information.
[0007] The trajectory data is overlaid and analyzed with a preset physical obstacle layer to identify trajectory segments that cross physical obstacles. Based on the number of crossings, crossing distance, and obstacle severity, a physical violation score is calculated for each trajectory.
[0008] Based on the compliance of each user's historical trajectory and the quality of the trajectory data, a user credibility model is constructed, and a corresponding user credibility score is generated to reflect the reliability of the user's uploaded trajectory.
[0009] Based on the preset road network, terrain data, land use information and traffic rule constraints, the path rationality index is calculated for each trajectory, and a path rationality score is generated for each trajectory.
[0010] Based on the user credibility score, the trajectory data is clustered to identify trajectory groups with similar physical violation characteristics or similar path behavior characteristics, thereby uncovering abnormal trajectory patterns or suspicious path segments.
[0011] The physical violation score, user credibility score, and path rationality score are weighted and fused to generate a comprehensive rationality score for each trajectory, wherein the weighting coefficients are adaptively adjusted according to the trajectory clustering category;
[0012] Based on the comprehensive rationality score, a threshold is set, and trajectory data below the threshold are removed or marked as abnormal, thereby obtaining a valid trajectory dataset that has been quality-screened.
[0013] Preferably, the step of calculating the physical violation score for each trajectory includes:
[0014] Spatial overlay analysis of the trajectory and physical obstacle layer is performed to identify trajectory segments that traverse obstacles;
[0015] Extract the feature parameters for each trajectory segment, which include at least the number of crossings, crossing distance, obstacle type, and crossing speed;
[0016] The feature parameters are input into a pre-trained physical violation assessment model, and the model outputs the corresponding physical violation score.
[0017] Preferably, the steps for constructing a user credibility model include:
[0018] Perform compliance analysis on the trajectory data to calculate the physical violation rate and data integrity of the trajectory;
[0019] The physical violation rate, data integrity, and trajectory upload frequency are used as user characteristic parameters.
[0020] The user characteristic parameters are input into the user credibility assessment model, and the model outputs a user credibility score.
[0021] Preferably, the step of generating a path rationality score for each trajectory includes:
[0022] The location information of the trajectory is matched with the preset road network, and the spatial matching degree between the trajectory and the road centerline is calculated.
[0023] Based on terrain data analysis, the slope changes of the trajectory are analyzed to determine whether the trajectory is reasonable for the terrain.
[0024] Determine whether the trajectory meets the traffic rule compliance requirements based on traffic rule constraints;
[0025] The path rationality assessment model is input as feature parameters, including spatial matching degree, terrain rationality, land use compliance, and traffic rule compliance. The model outputs the corresponding path rationality score.
[0026] Preferably, the step of clustering the trajectory data includes:
[0027] The physical violation score, path rationality score, and user credibility score of each trajectory are used as feature parameters, and the feature parameters are standardized to eliminate the difference in units.
[0028] The distance matrix between trajectories is calculated based on the feature similarity between trajectories, and a clustering algorithm is used to divide the trajectory data into multiple trajectory groups.
[0029] Preferably, the step of weightedly fusing the physical violation score, user credibility score, and path rationality score to generate a comprehensive rationality score for each trajectory includes:
[0030] The physical violation score, user credibility score, and path rationality score corresponding to each trajectory are used as input parameters;
[0031] The scores are calculated by fusion according to a preset weighted fusion model, wherein the weight coefficient of each score is adaptively adjusted according to the cluster category to which the trajectory belongs.
[0032] Preferably, the step of adaptively adjusting the weight coefficients of each score according to the cluster category to which the trajectory belongs includes:
[0033] The distribution characteristics of physical violation scores, user credibility scores, and path rationality scores in each trajectory cluster category were statistically analyzed.
[0034] The confidence level of each rating dimension is determined based on the standard deviation of the ratings within different cluster categories.
[0035] The weight coefficients in the weighted fusion model are dynamically adjusted based on the stated confidence level.
[0036] Preferably, the step of setting a threshold based on the comprehensive rationality score, and removing or marking trajectory data below the threshold as abnormal to obtain a quality-screened valid trajectory dataset includes:
[0037] The distribution characteristics of the overall rationality score for all trajectories were statistically analyzed;
[0038] The abnormal threshold range is calculated based on the mean and standard deviation of the scores;
[0039] When the trajectory score is lower than the standard deviation of the mean minus a preset multiple, the trajectory is marked as abnormal.
[0040] All abnormal trajectories are removed to obtain a valid trajectory dataset.
[0041] Secondly, the present invention provides the following technical solution: a navigation trajectory data processing system for implementing the above-mentioned navigation trajectory data processing method, the system comprising:
[0042] The data acquisition module is used to acquire the trajectory dataset uploaded by the user. Each trajectory data includes a timestamp sequence, a location coordinate sequence, and the corresponding user identification information.
[0043] The physical verification module is used to overlay and analyze the trajectory data with a preset physical obstacle layer, identify trajectory segments that cross physical obstacles, and calculate the physical violation score of each trajectory based on the number of crossings, crossing distance, and obstacle severity.
[0044] The user credibility assessment module is used to build a user credibility model based on the compliance of each user's historical trajectory and the quality of the trajectory data, and generate a corresponding user credibility score to reflect the reliability of the user's uploaded trajectory.
[0045] The trajectory rationality assessment module is used to calculate the path rationality index for each trajectory based on the preset road network, terrain data, land use information and traffic rule constraints, and generate a path rationality score for each trajectory.
[0046] The trajectory clustering analysis module is used to cluster trajectory data based on the user credibility score, identify trajectory groups with similar physical violation characteristics or similar path behavior characteristics, and thus discover abnormal trajectory patterns or suspicious path segments.
[0047] The weighted fusion module is used to weight and fuse the physical violation score, user credibility score and path rationality score to generate a comprehensive rationality score for each trajectory. The weighting coefficients are adaptively adjusted according to the trajectory clustering category.
[0048] The abnormal trajectory filtering module is used to set a threshold based on the comprehensive rationality score, and to remove or mark trajectory data below the threshold as abnormal, thereby obtaining a valid trajectory dataset that has been quality filtered.
[0049] The present invention has the following beneficial effects:
[0050] The navigation trajectory data processing method of this invention enables multi-dimensional comprehensive evaluation of trajectory data, integrating physical constraints, user credibility, and path rationality. It effectively identifies abnormal trajectories and abnormal trajectory groups, eliminates low-credibility or illegal trajectories, and obtains a high-quality, effective trajectory dataset, thereby improving the accuracy and reliability of navigation path recommendation, map updates, and travel analysis. Furthermore, through adaptive weight adjustment and threshold mechanisms, this method can dynamically respond to the characteristics of trajectory data from different regions and user groups, improving the overall stability and robustness of the system. It also possesses good scalability and is applicable to navigation systems in large-scale trajectory data environments, achieving intelligent and efficient trajectory data processing. Attached Figure Description
[0051] Figure 1 This is a flowchart of a navigation trajectory data processing method proposed in this invention;
[0052] Figure 2 This is a framework diagram of a navigation trajectory data processing system proposed in this invention. Detailed Implementation
[0053] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0054] Example 1
[0055] In a first embodiment of the present invention, the present invention provides a navigation trajectory data processing method and system, such as... Figure 1 As shown, it includes the following steps:
[0056] S1. Obtain the trajectory dataset uploaded by the user. Each trajectory data includes a timestamp sequence, a location coordinate sequence, and the corresponding user identification information.
[0057] Specifically, user terminals (such as smartphones, in-vehicle devices, or wearable positioning devices) periodically sample location points through a positioning module. Each location point includes a timestamp, latitude and longitude coordinates, and optional speed and direction information. The sampling frequency can be dynamically adjusted according to the application scenario; for example, the default interval is 1 to 5 seconds. The collected raw location points undergo preliminary cleaning on the terminal or server side, including removing points with abnormal time intervals, deleting points that are repetitive or have not changed for a long time in a static state, and smoothing or interpolating abnormal coordinate points that deviate significantly. The cleaned location information is then arranged into a trajectory data record in chronological order. Each trajectory data record includes a timestamp sequence, a location coordinate sequence, and user identification information.
[0058] S2. Overlay the trajectory data with a preset physical obstacle layer for analysis, identify trajectory segments that cross physical obstacles, and calculate the physical violation score of each trajectory based on the number of crossings, crossing distance, and obstacle severity.
[0059] Preferably, the step of calculating the physical violation score for each trajectory includes:
[0060] Spatial overlay analysis of the trajectory and physical obstacle layer is performed to identify trajectory segments that traverse obstacles;
[0061] Extract the feature parameters for each trajectory segment, which include at least the number of crossings, crossing distance, obstacle type, and crossing speed;
[0062] The feature parameters are input into a pre-trained physical violation assessment model, and the model outputs the corresponding physical violation score.
[0063] Specifically, the server pre-loads a physical obstacle layer data set, which may include road fences, building outlines, rivers, green belts, non-motorized vehicle isolation zones, etc. Each trajectory consists of continuous position coordinate points. The system performs spatial overlay analysis on the trajectory and the obstacle layer to detect whether the trajectory segments intersect or overlap with the obstacle boundaries. This analysis can quickly identify the sections of the trajectory that cross obstacles based on the R-tree indexing algorithm and geometric calculation functions.
[0064] For each detected trajectory segment crossing an obstacle, the system extracts the following feature parameters: number of crossings, i.e., the number of times the trajectory crosses the obstacle area; crossing distance, i.e., the distance the trajectory moves within the obstacle area, in meters; obstacle type, i.e., the category representing the obstacle, such as "building", "river", "road divider", etc.; crossing speed, i.e., the average speed during crossing, which can be calculated from the position and time difference of adjacent sampling points.
[0065] The aforementioned feature parameters are combined to form a feature vector, which is then input into a pre-trained physical violation assessment model. This assessment model can be a regression or classification model built based on machine learning algorithms, such as a random forest model, support vector regression model, or a neural network-based scoring model. The model's training data comes from manually labeled trajectory samples, where the violation level of each sample trajectory is predefined by geographic experts or system rules. The model output is the raw score for the physical violation. The range is 0 to 1.
[0066] Through the above steps, the present invention can quantitatively assess whether a trajectory violates geographical reality constraints based on the spatial relationship between the trajectory and the physical obstacle layer, thereby achieving automatic verification of the trajectory's physical feasibility.
[0067] S3. Based on the compliance of each user's historical trajectory and the quality of the trajectory data, construct a user credibility model and generate a corresponding user credibility score to reflect the reliability of the user's uploaded trajectory.
[0068] Preferably, the steps for constructing a user credibility model include:
[0069] Perform compliance analysis on the trajectory data to calculate the physical violation rate and data integrity of the trajectory;
[0070] The physical violation rate, data integrity, and trajectory upload frequency are used as user characteristic parameters.
[0071] The user characteristic parameters are input into the user credibility assessment model, and the model outputs a user credibility score.
[0072] Specifically, the system uses the calculated physical violation score of the trajectory to statistically analyze the violations of each user's historically uploaded trajectories. For users... Let the set of trajectories it uploads be... Then the user's physical violation rate can be expressed as: ,in Representing the trajectory The physical violation score is calculated for each trajectory. One or more combinations of sampling density, temporal continuity, and positioning stability metrics are used to measure data acquisition quality. The temporal continuity metric indicates whether the time interval between adjacent sampling points in the trajectory is stable; the spatial continuity metric indicates whether there are abnormal jumps in the trajectory's velocity changes; and the data loss rate indicates the proportion of missing points in the trajectory. For users... The completeness of its historical trajectory data It can be represented as: ,in Representing the trajectory The system extracts trajectory upload frequency features from users' historical upload records to reflect user activity and data contribution stability. The statistical time window is set as follows: (For example, the upload frequency in the last 30 days) ,in This indicates the number of user-uploaded tracks within that time window.
[0073] The above three features are combined into a user feature vector:
[0074] ;
[0075] The system can normalize the feature vector to ensure that the values of each dimension are on the same scale, so as to avoid the difference in the dimensions between features from affecting the model calculation.
[0076] User feature vectors are input into the user credibility assessment model. This model can be a regression or classification model built based on machine learning algorithms, such as random forest models, support vector regression models, or neural network-based scoring models. The model's training data can come from labeled user trajectory datasets, where "high-credibility users" and "low-credibility users" are defined by manual or system rules (e.g., users with long-term stable trajectories and low physical violation rates are labeled as high-credibility samples). The model learns the mapping relationship between user features and credibility through training and outputs a user credibility score. Its range is between 0 and 1.
[0077] Through the above steps, the present invention can construct a quantitative user credibility model based on the user's historical trajectory performance and data quality characteristics, and realize dynamic evaluation of the reliability of user-uploaded data.
[0078] S4. Based on the preset road network, terrain data, land use information and traffic rule constraints, calculate the path rationality index for each trajectory and generate a path rationality score for each trajectory.
[0079] Preferably, the step of generating a path rationality score for each trajectory includes:
[0080] The location information of the trajectory is matched with the preset road network, and the spatial matching degree between the trajectory and the road centerline is calculated.
[0081] Based on terrain data analysis, the slope changes of the trajectory are analyzed to determine whether the trajectory is reasonable for the terrain.
[0082] Determine whether the trajectory meets the traffic rule compliance requirements based on traffic rule constraints;
[0083] The path rationality assessment model is input as feature parameters, including spatial matching degree, terrain rationality, land use compliance, and traffic rule compliance. The model outputs the corresponding path rationality score.
[0084] Specifically, based on the location information of the trajectory, the system spatially matches the trajectory coordinate sequence with pre-set road network data. The road network may include road centerlines, road grades, directional attributes, and lane constraint information. The matching of the trajectory and the road network can employ a map matching algorithm, combining time series data and road topology relationships to output the mapping path of the trajectory within the road network. The system calculates the shortest distance to the road centerline for each trajectory point. Based on this, the spatial matching degree index of the trajectory is calculated. ,in The maximum allowable offset distance set for the system.
[0085] The system calls digital elevation model data, calculates the corresponding elevation values based on the latitude and longitude of the trajectory points, and obtains the elevation sequence. The slope change rate of the trajectory is calculated based on the high-precision program sequence:
[0086] ;
[0087] in, This represents the horizontal distance between two trajectory points. The system sets reasonable slope thresholds based on different modes of transportation, such as ≤ 30% for walking and ≤ 10% for vehicles. If the slope of the trajectory exceeds this threshold range, it is considered to have terrain anomalies. Based on this, the system calculates a terrain rationality score for the trajectory; a higher score indicates that the path better conforms to the terrain features. A higher score indicates that the path better matches the terrain features.
[0088] The system verifies the legality of trajectory paths based on urban land use data and traffic regulations. Specifically, it determines whether the trajectory crosses non-road land, such as building land, water areas, or green belts. If the trajectory is located within a legally accessible area, land use compliance is granted. A higher score is awarded for a higher score, while a lower score is awarded for a lower score. The system verifies trajectory behavior based on road attribute data, such as one-way streets, speed limits, and restricted areas. If the trajectory direction is consistent with the prescribed road direction and the speed does not exceed the limit, it is considered to comply with traffic rules. A traffic rule compliance index is calculated based on the above results. .
[0089] The spatial matching degree, terrain rationality, land use compliance, and traffic rule compliance are used as feature parameters input into the path rationality assessment model. The model can be a regression or classification model built based on machine learning algorithms, such as a random forest model, support vector regression model, or a neural network-based scoring model. The model can be trained using known rational and abnormal trajectory samples to learn the contribution relationship of each feature parameter to path rationality. The model outputs a path rationality score for each trajectory. The value range is 0 to 1.
[0090] Through the above steps, the present invention can comprehensively analyze geospatial matching, terrain constraints, land use and traffic rules information to quantitatively evaluate the path rationality of each trajectory, effectively eliminating abnormal trajectories that deviate from roads, cross non-traffic areas and violate traffic constraints.
[0091] S5. Based on the user credibility score, cluster the trajectory data to identify trajectory groups with similar physical violation characteristics or similar path behavior characteristics, thereby mining abnormal trajectory patterns or suspicious path segments.
[0092] Preferably, the step of clustering the trajectory data includes:
[0093] The physical violation score, path rationality score, and user credibility score of each trajectory are used as feature parameters, and the feature parameters are standardized to eliminate the difference in units.
[0094] The distance matrix between trajectories is calculated based on the feature similarity between trajectories, and a clustering algorithm is used to divide the trajectory data into multiple trajectory groups.
[0095] Specifically, the system constructs a feature vector using the physical violation score, path rationality score, and user credibility score of each trajectory as feature parameters, i.e., each trajectory Corresponding to a three-dimensional feature vector To avoid the impact of differences in the dimensions of different feature dimensions on the clustering results, the system standardizes the above feature parameters. The standardization method can be Z-score standardization or Min-Max normalization.
[0096] The system calculates similarity based on standardized feature vectors between trajectories and constructs a distance matrix from the similarity results. In a preferred embodiment, the similarity metric can be Euclidean distance. The system selects an appropriate clustering algorithm to group the trajectory data based on the distribution characteristics of the trajectory features. In a preferred embodiment, the system uses the K-means clustering algorithm and automatically determines the optimal number of clusters K based on clustering performance evaluation metrics. The system performs statistical analysis on the clustering results, calculating the average physical violation and path rationality distribution for each group. When a group as a whole exhibits high violation and low rationality characteristics, the system marks that group as an abnormal trajectory group.
[0097] Through the above steps, the present invention achieves cluster analysis of multidimensional trajectory features, which can automatically identify trajectories with similar behavioral characteristics. This process not only improves the structuring of trajectory data, but can also be used to discover potential abnormal group behaviors and suspicious path distributions.
[0098] S6. The physical violation score, user credibility score and path rationality score are weighted and fused to generate a comprehensive rationality score for each trajectory, wherein the weighting coefficient is adaptively adjusted according to the trajectory clustering category;
[0099] Preferably, the step of weightedly fusing the physical violation score, user credibility score, and path rationality score to generate a comprehensive rationality score for each trajectory includes:
[0100] The physical violation score, user credibility score, and path rationality score corresponding to each trajectory are used as input parameters;
[0101] The scores are calculated by fusion according to a preset weighted fusion model, wherein the weight coefficient of each score is adaptively adjusted according to the cluster category to which the trajectory belongs.
[0102] Preferably, the step of adaptively adjusting the weight coefficients of each score according to the cluster category to which the trajectory belongs includes:
[0103] The distribution characteristics of physical violation scores, user credibility scores, and path rationality scores in each trajectory cluster category were statistically analyzed.
[0104] The confidence level of each rating dimension is determined based on the standard deviation of the ratings within different cluster categories.
[0105] The weight coefficients in the weighted fusion model are dynamically adjusted based on the stated confidence level.
[0106] Specifically, the system uses a weighted fusion model to comprehensively calculate the three types of scores and generate a comprehensive rationality score for the trajectory. ,in , and These are the weight coefficients corresponding to physical violation, user credibility, and path rationality, respectively. To improve the adaptability and stability of the scoring, the system dynamically adjusts the weight coefficients of each score based on the cluster category to which the trajectory belongs. First, the statistical characteristics of the three scores in each cluster category are analyzed. For a certain scoring dimension (e.g., user credibility), the smaller its standard deviation in the current cluster, the more stable and reliable that dimension is in that group. Therefore, a confidence level is defined. ,in, Let the standard deviation of the score be the value within the cluster. To prevent constants with zero denominators, the confidence scores are then standardized to weights. This yields the adaptive weight vector for each cluster category.
[0107] Through the above steps, this invention establishes an adaptive multidimensional fusion evaluation mechanism at the trajectory data level, which can dynamically adjust the balance between physical constraints, user reliability, and path rationality according to the characteristics of the group to which the trajectory belongs, thereby generating a more objective and robust comprehensive rationality score.
[0108] S7. Set a threshold based on the comprehensive rationality score, and remove or mark trajectory data below the threshold as abnormal to obtain a valid trajectory dataset after quality screening.
[0109] Preferably, the step of setting a threshold based on the comprehensive rationality score, and removing or marking trajectory data below the threshold as abnormal to obtain a quality-screened valid trajectory dataset includes:
[0110] The distribution characteristics of the overall rationality score for all trajectories were statistically analyzed;
[0111] The abnormal threshold range is calculated based on the mean and standard deviation of the scores;
[0112] When the trajectory score is lower than the standard deviation of the mean minus a preset multiple, the trajectory is marked as abnormal.
[0113] All abnormal trajectories are removed to obtain a valid trajectory dataset.
[0114] Specifically, the system obtains a comprehensive rationality score for all trajectories from the aforementioned steps. To form a rating set The system performs statistical analysis on the set of ratings and calculates its mean.
[0115] with standard deviation Simultaneously, the system can generate a score distribution histogram to visualize the overall distribution of trajectory quality. Anomaly detection thresholds are set based on the mean and standard deviation of the scores. ,in, This is a preset multiplier, typically ranging from 1.0 to 2.5. For each trajectory... The system compares its overall rationality score. With abnormal threshold ,like Trajectories that are marked as abnormal are identified. The system iterates through all trajectories, filters them based on the anomaly labels, removes the trajectories marked as abnormal from the main dataset, and retains the unmarked trajectories as the valid trajectory dataset.
[0116] Through the above steps, this invention uses statistical methods to perform distribution analysis on the comprehensive rationality score, and uses the mean and standard deviation to calculate the threshold, thereby achieving adaptive identification and elimination of abnormal trajectories.
[0117] Example 2
[0118] In a central urban area, the road network is dense and the terrain complex, with multiple railway barriers and buildings providing additional obstacles. Within this area, navigation applications frequently encounter problems when collecting user trajectories, including data discrepancies, trajectories crossing physical obstacles, and traffic violations. Existing trajectory data processing methods struggle to effectively identify abnormal paths taken by groups, impacting the accuracy of navigation route recommendations and map data updates.
[0119] To solve the above problems, this invention provides a navigation trajectory data processing method system, the structure of which is as follows: Figure 2 As shown, the system includes a data acquisition module, a physical verification module, a user credibility assessment module, a trajectory rationality assessment module, a trajectory clustering analysis module, a weighted fusion module, and an abnormal trajectory screening module. The specific implementation process of the system is as follows:
[0120] Data Acquisition Module: This module retrieves trajectory datasets uploaded by users in the city center area from the navigation application server. Each trajectory includes a timestamp sequence, a latitude and longitude coordinate sequence, and corresponding user identification information.
[0121] The physical verification module performs spatial overlay analysis on the trajectory and physical obstacle layers such as railways, buildings, and rivers to identify segments in the trajectory that traverse physical obstacles. For each trajectory segment, the system extracts feature parameters such as the number of crossings, crossing distance, obstacle type, and crossing speed. Subsequently, the system inputs these features into a pre-trained physical violation assessment model and outputs a physical violation score for each trajectory.
[0122] The user credibility assessment module generates a user feature vector based on the compliance, data integrity, and upload frequency of the user's historical trajectory. This feature vector is then input into the user credibility assessment model, which outputs a user credibility score to reflect the reliability of the user's uploaded trajectories and the overall data quality.
[0123] The trajectory rationality assessment module matches the location information of the trajectory with the urban road network, calculating the spatial matching degree between the trajectory and the road centerline. Combining road slope, land use information, and traffic regulations, the system evaluates the terrain rationality and traffic compliance of the trajectory. Finally, the system inputs the spatial matching degree, terrain rationality, land use compliance, and traffic regulation compliance into the path rationality assessment model to generate a path rationality score for each trajectory.
[0124] The trajectory clustering analysis module standardizes the physical violation score, user credibility score, and path rationality score to construct trajectory feature vectors. The system calculates a distance matrix based on the feature similarity between trajectories and uses a clustering algorithm to divide the trajectories into multiple groups.
[0125] The weighted fusion module takes physical violation score, user credibility score, and path rationality score as input, and calculates a comprehensive rationality score for each trajectory using a weighted fusion model. The weight coefficients are adaptively adjusted according to the cluster category to which the trajectory belongs, ensuring that high violation groups receive higher weights for physical violation and high-credibility user groups receive higher weights for path rationality, thereby achieving adaptive scoring across different groups and scenarios.
[0126] The system includes an abnormal trajectory filtering module. This module performs statistical analysis on the overall reasonableness score of all trajectories, calculates the mean and standard deviation, and sets a threshold, such as the mean minus a preset multiple of the standard deviation. Trajectories below the threshold are marked as abnormal and removed from the dataset, ultimately generating a high-quality, valid trajectory dataset. This dataset can be used for navigation route recommendation, map updates, and travel pattern analysis, significantly improving the system's accuracy and reliability.
[0127] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for processing navigation trajectory data, characterized in that, include: Obtain the trajectory dataset uploaded by the user. Each trajectory data includes a timestamp sequence, a location coordinate sequence, and the corresponding user identification information. The trajectory data is overlaid and analyzed with a preset physical obstacle layer to identify trajectory segments that cross physical obstacles. Based on the number of crossings, crossing distance, and obstacle severity, a physical violation score is calculated for each trajectory. Based on the compliance of each user's historical trajectory and the quality of the trajectory data, a user credibility model is constructed, and a corresponding user credibility score is generated to reflect the reliability of the user's uploaded trajectory. Based on the preset road network, terrain data, land use information and traffic rule constraints, the path rationality index is calculated for each trajectory, and a path rationality score is generated for each trajectory. Based on the user credibility score, the trajectory data is clustered to identify trajectory groups with similar physical violation characteristics or similar path behavior characteristics, thereby uncovering abnormal trajectory patterns or suspicious path segments. The physical violation score, user credibility score, and path rationality score are weighted and fused to generate a comprehensive rationality score for each trajectory, wherein the weighting coefficients are adaptively adjusted according to the trajectory clustering category; Based on the comprehensive rationality score, a threshold is set, and trajectory data below the threshold are removed or marked as abnormal, thereby obtaining a valid trajectory dataset that has been quality-screened.
2. The navigation trajectory data processing method according to claim 1, characterized in that, The steps for calculating the physical violation score for each trajectory include: Spatial overlay analysis of the trajectory and physical obstacle layer is performed to identify trajectory segments that traverse obstacles; Extract the feature parameters for each trajectory segment, which include at least the number of crossings, crossing distance, obstacle type, and crossing speed; The feature parameters are input into a pre-trained physical violation assessment model, and the model outputs the corresponding physical violation score.
3. The navigation trajectory data processing method according to claim 1, characterized in that, The steps to build a user credibility model include: Perform compliance analysis on the trajectory data to calculate the physical violation rate and data integrity of the trajectory; The physical violation rate, data integrity, and trajectory upload frequency are used as user characteristic parameters. The user characteristic parameters are input into the user credibility assessment model, and the model outputs a user credibility score.
4. The navigation trajectory data processing method according to claim 1, characterized in that, The steps for generating a path rationality score for each trajectory include: The location information of the trajectory is matched with the preset road network, and the spatial matching degree between the trajectory and the road centerline is calculated. Based on terrain data analysis, the slope changes of the trajectory are analyzed to determine whether the trajectory is reasonable for the terrain. Determine whether the trajectory meets the traffic rule compliance requirements based on traffic rule constraints; The path rationality assessment model is input as feature parameters, including spatial matching degree, terrain rationality, land use compliance, and traffic rule compliance. The model outputs the corresponding path rationality score.
5. The navigation trajectory data processing method according to claim 1, characterized in that, The steps for clustering trajectory data include: The physical violation score, path rationality score, and user credibility score of each trajectory are used as feature parameters, and the feature parameters are standardized to eliminate the difference in units. The distance matrix between trajectories is calculated based on the feature similarity between trajectories, and a clustering algorithm is used to divide the trajectory data into multiple trajectory groups.
6. The navigation trajectory data processing method according to claim 1, characterized in that, The steps for weighted fusion of the physical violation score, user credibility score, and path rationality score to generate a comprehensive rationality score for each trajectory include: The physical violation score, user credibility score, and path rationality score corresponding to each trajectory are used as input parameters; The scores are calculated by fusion according to a preset weighted fusion model, wherein the weight coefficient of each score is adaptively adjusted according to the cluster category to which the trajectory belongs.
7. A navigation trajectory data processing method according to claim 6, characterized in that, The steps for adaptively adjusting the weight coefficients of each score based on the cluster category to which the trajectory belongs include: The distribution characteristics of physical violation scores, user credibility scores, and path rationality scores in each trajectory cluster category were statistically analyzed. The confidence level of each rating dimension is determined based on the standard deviation of the ratings within different cluster categories. The weight coefficients in the weighted fusion model are dynamically adjusted based on the stated confidence level.
8. The navigation trajectory data processing method according to claim 1, characterized in that, The steps for setting a threshold based on the comprehensive rationality score, removing or marking trajectory data below the threshold as abnormal, and thus obtaining a quality-filtered valid trajectory dataset include: The distribution characteristics of the overall rationality score for all trajectories were statistically analyzed; The abnormal threshold range is calculated based on the mean and standard deviation of the scores; When the trajectory score is lower than the standard deviation of the mean minus a preset multiple, the trajectory is marked as abnormal. All abnormal trajectories are removed to obtain a valid trajectory dataset.
9. A navigation trajectory data processing system, characterized in that, The system for the navigation trajectory data processing method according to any one of claims 1-8 comprises: The data acquisition module is used to acquire the trajectory dataset uploaded by the user. Each trajectory data includes a timestamp sequence, a location coordinate sequence, and the corresponding user identification information. The physical verification module is used to overlay and analyze the trajectory data with a preset physical obstacle layer, identify trajectory segments that cross physical obstacles, and calculate the physical violation score of each trajectory based on the number of crossings, crossing distance, and obstacle severity. The user credibility assessment module is used to build a user credibility model based on the compliance of each user's historical trajectory and the quality of the trajectory data, and generate a corresponding user credibility score to reflect the reliability of the user's uploaded trajectory. The trajectory rationality assessment module is used to calculate the path rationality index for each trajectory based on the preset road network, terrain data, land use information and traffic rule constraints, and generate a path rationality score for each trajectory. The trajectory clustering analysis module is used to cluster trajectory data based on the user credibility score, identify trajectory groups with similar physical violation characteristics or similar path behavior characteristics, and thus discover abnormal trajectory patterns or suspicious path segments. The weighted fusion module is used to weight and fuse the physical violation score, user credibility score and path rationality score to generate a comprehensive rationality score for each trajectory. The weighting coefficients are adaptively adjusted according to the trajectory clustering category. The abnormal trajectory filtering module is used to set a threshold based on the comprehensive rationality score, and to remove or mark trajectory data below the threshold as abnormal, thereby obtaining a valid trajectory dataset that has been quality filtered.